Towards the electrification of city logistics?

Philippe Lebeau

Towards the electrification of city logistics?

Abstract City logistics is facing an important challenge. It is one of the most polluting segments of the transport sector but policy makers want it to become one of the cleanest in the future. The European Commission, for example, has set the goal of reaching CO2 free city logistics by 2030. Battery electric vehicles represent in that context a potential solution. They can indeed reduce CO2 emissions, especially if electricity is generated from renewables. Moreover, they can improve air quality and reduce noise generated by traffic in cities. However their adoption by freight transport operators remains limited despite the recent development of electric vans and trucks on the market. The objective of this thesis is therefore to investigate the feasibility of introducing battery electric vehicles in city logistics.

The PhD is structured around three main research questions that address (1) the potential adoption of battery electric vehicles in city logistics, (2) the strategies to reduce or solve their economic and operational constraints and (3) the stakeholders’ support regarding a shift from conventional to battery electric vehicles. By tackling these three aspects, the thesis demonstrates that an electrification of city logistics is possible. The different stakeholders of city logistics are indeed found to support that transition. But the adoption of battery electric vehicles remains limited because of their economic and operational constraints. The thesis identifies therefore the different conditions where battery electric vehicles can become profitable for freight transport operators. It recommends also a range of policies that can further stimulate the adoption of battery electric vehicles.

Philippe Lebeau is working as a research associate at Vrije Universiteit Brussel, within the research group MOBI (Mobility, Logistics and Automotive Technology Research Centre). His PhD was supervised by his promotor Prof. Dr. Cathy Macharis and his co-promotor Prof. Dr. Joeri Van Mierlo.

FACULTY OF ECONOMICS, SOCIAL AND POLITICAL SCIENCES AND SOLVAY BUSINESS SCHOOL

Towards the electrification of city logistics?

Thesis submitted in fulfilment of the requirements for the award of the degree of Doctor in Toegepaste Economische Wetenschappen by

Philippe Lebeau

Promotors:

Prof. Dr. Cathy Macharis Prof. Dr. ir. Joeri Van Mierlo

Date:

21st of March 2016

ii

Acknowledgements This thesis represents the result of four inspiring years of research where I could develop different solutions to stimulate the electrification of city logistics. I hope that the findings developed in this book will inspire and help the readers to further improve the sustainability of our cities and of our transport systems. The name standing on the cover of this book represents however only the emerged part of the iceberg. There are a lot of people behind this thesis that made these four years of research possible. They contributed also to this work from a way or another and I would like to start this book by expressing my gratitude to them. First and foremost, I want to thank my two promotors. Cathy has been a wonderful promotor and has remained very supportive through the different critical steps of the PhD adventure. Without her, I would probably not even have thought about doing a thesis. Joeri also has been very present through my research and very responsive to my concerns. Thanks to his detailed and constructive feedbacks, he helped me to improve my research skills which allowed me to finally defend my thesis. I also wish to thank the members of my jury for the time they invested in this thesis and the interesting suggestions they gave to ensure the quality of the thesis. Thank you to Prof. Dr. Michael Browne, Prof. Dr. Jean Charles Jacquemin, Prof. Dr. Thierry Coosemans and especially Prof. Dr. Wouter Verbeke for being the president of my jury. I want also to thank Innoviris, the institution in the Brussels-Capital Region encouraging scientific research and innovation. It made that PhD possible thanks to its financial support through the Prospective Research for Brussels entitled “a strategy for the implementation of a sustainable logistics concept for the city distribution of the Brussels-Capital Region”. The motivation might be however the most important ingredient for a PhD. And I believe that colleagues and researchers have been playing a great role in that. The MOBI research group was an amazing place to conduct my PhD. Besides everyone being wonderful colleagues, I want to thank Kenneth Lebeau who introduced me to the research world by demonstrating the benefits of the ping-pong effect; Sara Verlinde and Bram Kin for the really interesting talks and exchanges around city logistics; Koen Mommens and Dries Meers for the great teambuilding we organised;

iii Jeroen Bulckaen, Imre Keseru, Geert te Boveldt, Sheida Hadavi, Paul Otuyalo and Nils Wuytens to have been such nice roommates; Floris De Vriendt, Heleen Buldeo Rai, Georges Petrides Koen Van Raemdonck and Tom van Lier for the nice lunch times we could share together talking about everything else but the work; Tom Matthijs and Katleen Cornelis for their help in every important moments; Frederic Dobruszkes, Ethem Pekin, Astrid De Witte and Olivier Mairesse for the nice memories they left after leaving the research group; the group of ETEC for the two enjoyable EVS conferences and their extended knowledge about electric vehicles and last but not least the members of the MOBI football team for their great team spirit. They all have been making the Vrije Universiteit Brussel, the place to be, the reason I came every day to the office. I also need to thank the other researchers and people I met and worked with during my PhD. In particular, I want to thank Milena Janjevic for being a fantastic researcher to work with within my Prospective Research for Brussels. I want to thank Charlotte Debroux, Christophe De Voghel and Marianne Thys. They have brought an essential support to this PhD by recommending some good contacts, by being available for interesting discussions around freight transport in Brussels or by supporting my research in many other various ways. I want also to thank Joeri De Ridder for his help with the TCO analysis and his very inspiring knowledge about electric vehicles. The support I received during these four years of research goes however beyond the professional sphere. Friends have been playing a critical role to breathe and relax before coming back on the thesis with a recharged brain. I want therefore to give a special big up to the friends that were present during these four years of research. Thank you to my good friends of the 55eme unite, to the RUN flat in Namur, to the Tabellion flat in Ixelles, to the Ernest Laude House in Schaerbeek, to the AS Basilique, to the Tango classes and all the others that do not fit in any category ! I want to give also a warm thank to my parents for their support during these four years but also for all their attention they have given to me and to my beloved sister and brothers. My thanks go also to the whole family of the Lebeau, the Scorneau, the Bachelerie and the Bonnet. And thank you also to the Veithen and the Denié who have welcomed me so greatly in their family these last years. But the most important thanks should go to my girlfriend, the love of my life, who was there in every nice but also difficult moment. She makes my life so much more beautiful, as she did for the cover of this thesis with her wonderful drawing. Thank you!

Brussels, the 21st of March 2016

iv

v

Table of Contents Acknowledgements .............................................................................................................. ii Table of Contents .................................................................................................................. v List of Figures ................................................................................................................... viii List of Tables ........................................................................................................................x List of Abbreviations ........................................................................................................... xi 1 Introduction ................................................................................................................... - 1 1.1 The challenges of the transport sector ....................................................................................... - 1 1.2 The role of city logistics ............................................................................................................... - 5 1.3 Electric vehicles as a solution? .................................................................................................... - 9 1.4 Research questions ......................................................................................................................- 17 1.5 Structure of the dissertation ......................................................................................................- 18 References...........................................................................................................................................- 22 2 Freight transport in Brussels and its impact on road traffic....................................... - 30 2.1 Introduction .................................................................................................................................- 30 2.2. Goods transported by road in Brussels: data and observations ..........................................- 30 2.3. Solutions for Brussels ................................................................................................................- 39 2.4 Conclusion....................................................................................................................................- 42 References...........................................................................................................................................- 44 3 Exploring the choice of battery electric vehicles in city logistics: a conjoint based choice analysis ............................................................................................................... - 47 3.1 Introduction .................................................................................................................................- 47 3.2 Drivers and barriers to BEV adoption.....................................................................................- 48 3.3 Methodology ................................................................................................................................- 51 3.4 Results – Choice behaviour .......................................................................................................- 53 3.5 Results – Attitudes of respondents ...........................................................................................- 59 3.6 Discussion ....................................................................................................................................- 61 3.7 Conclusions ..................................................................................................................................- 63 -

vi References...........................................................................................................................................- 65 4 Electrifying light commercial vehicles for city logistics? A total cost of ownership analysis ........................................................................................................................... - 68 4.1 Introduction .................................................................................................................................- 68 4.2 Methodology ................................................................................................................................- 70 4.3 Results ...........................................................................................................................................- 75 4.4 Sensitivity analysis .......................................................................................................................- 78 4.5 Conclusion....................................................................................................................................- 84 References...........................................................................................................................................- 87 5 Diesel or/and battery electric vehicles: Which mixed technology fleet for city logistics? ......................................................................................................................... - 90 5.1 Introduction .................................................................................................................................- 90 5.2 Literature review ..........................................................................................................................- 91 5.3 Methodology ................................................................................................................................- 94 5.4 Results ........................................................................................................................................ - 101 5.5 Sensitivity analysis .................................................................................................................... - 103 5.6 Conclusions ............................................................................................................................... - 109 References........................................................................................................................................ - 111 6 Strategic Scenarios for Sustainable Urban Distribution in the Brussels-Capital Region Using Urban Consolidation Centres............................................................................. - 114 6.1 Introduction .............................................................................................................................. - 114 6.2 Methodology ............................................................................................................................. - 115 6.3 Results ........................................................................................................................................ - 127 6.4 Conclusion................................................................................................................................. - 129 References........................................................................................................................................ - 130 7 Operationalising a bottom-up approach through a multi-actor multi-criteria analysis: an application to city logistics ........................................................................................... - 133 7.1 Introduction .............................................................................................................................. - 133 7.2 Stakeholders involvement in city logistics ............................................................................ - 134 7.3 Relevance of MAMCA for a bottom up approach in strategy formulation .................... - 136 7.4 Application to a stakeholder consultation in Brussels ........................................................ - 137 7.5 Discussion ................................................................................................................................. - 150 7.6 Conclusions ............................................................................................................................... - 152 References........................................................................................................................................ - 154 Conclusions ................................................................................................................... - 160 -

vii Battery electric vehicle: a solution for sustainable city logistics? ............................................. - 160 Findings ............................................................................................................................................ - 161 Contributions and implications of findings ................................................................................ - 167 Limitations and future research .................................................................................................... - 169 References........................................................................................................................................ - 172 Annexes ......................................................................................................................... - 173 Annex 1: Publications .................................................................................................................... - 173 Annex 2: City logistics projects involving battery electric vehicles ......................................... - 176 -

viii

List of Figures Figure 1: Greenhouse gas emissions by sector for the EU28 ........................................................... - 2 Figure 2: Percentage of the urban population in the EU exposed to air pollutant concentrations above the EU and WHO reference levels (2009-2011) ..................................................................... - 4 Figure 3: External costs due to urban transport ................................................................................. - 6 Figure 4: Impact of different vehicle technologies on CO2 emissions. .........................................- 10 Figure 5: Evolution of electric passenger cars in the European Union (combining battery electric vehicles, fuel cell electric vehicles and plug-in hybrid vehicles) .....................................................- 11 Figure 6: The diffusion of innovations ..............................................................................................- 13 Figure 7: Overview of city logistics projects involving BEVs ........................................................- 14 Figure 8: the Cargohopper in Utrecht ................................................................................................- 15 Figure 9: Evolution of freight transport in the Brussels-Capital Region ......................................- 31 Figure 10: Sections of motorway with structural congestion in 2009 ...........................................- 32 Figure 11: Number of freight vehicles entering and leaving Brussels according to the day of the week, excluding motorways .................................................................................................................- 33 Figure 12: Number of freight vehicles entering and leaving Brussels according to the time of the day, excluding motorways ....................................................................................................................- 34 Figure 13: Ranking of main entrances and exits in the Brussels-Capital Region for freight vehicles, motorways included ..............................................................................................................................- 35 Figure 14: Impact of freight vehicles on traffic during weekday peak hours, excluding motorways 36 Figure 15: Evolution in the number of vans and lorries in Belgium .............................................- 36 Figure 16: Distribution of vans and lorries in each Region according to their category ............- 37 Figure 17: Share of pollution caused by vans and lorries traffic in the total emissions from road transport in the Brussels-Capital Region ...........................................................................................- 38 Figure 18: Distribution of freight vehicles registered in Brussels according to different types of technology ..............................................................................................................................................- 39 Figure 19: Screenshot of a choice task submitted to respondents .................................................- 53 Figure 20: average utilities of purchase-cost levels ...........................................................................- 55 Figure 21: average utilities of operating-cost levels ..........................................................................- 55 Figure 22: average utilities of range levels..........................................................................................- 55 Figure 23: average utilities of refuelling- and charging-time levels ................................................- 55 Figure 24: average utilities of volume levels ......................................................................................- 55 Figure 25: average utilities of Ecoscore levels ...................................................................................- 55 Figure 26: Choice in business as usual ...............................................................................................- 58 Figure 27: Choice with a subsidy.........................................................................................................- 58 -

ix Figure 28: Choice with road pricing ...................................................................................................- 58 Figure 29: Disadvantages of BEVs .....................................................................................................- 59 Figure 30: Advantages of BEVs ..........................................................................................................- 60 Figure 31: Respondent support for measures to stimulate BEVs ..................................................- 61 Figure 32: Average kilometres driven per year by LCVs according their age ..............................- 72 Figure 33: Total cost of ownership for diesel, petrol and battery electric vehicles .....................- 77 Figure 34: TCO sensitivity on the kilometres driven per year by the quadricycles .....................- 79 Figure 35: TCO sensitivity on the kilometres driven per year by the LCVs (N1) .......................- 79 Figure 36: TCO sensitivity on the years of ownership of the quadricycles ..................................- 79 Figure 37: TCO sensitivity on the years of ownership of the LCVs (N1) ....................................- 79 Figure 38: The paradox of the battery electric vehicle .....................................................................- 91 Figure 39: Energy losses from plug to the wheels of battery electric vehicles .............................- 95 Figure 40 : Observed and expected energy consumption of electric vehicle trips ......................- 96 Figure 41: Observed and expected regenerated energy of electric vehicle trips ..........................- 97 Figure 42: Results of the FSMVRPTW ........................................................................................... - 102 Figure 43: Impact of time windows on costs of the last mile ...................................................... - 104 Figure 44: Impact of congestion on costs of the last mile ........................................................... - 105 Figure 45: Impact of shops density on costs of the last mile....................................................... - 106 Figure 46: Impact of the demand on costs of the last mile.......................................................... - 107 Figure 47: Impact of the costs of the fleet on costs of the last mile .......................................... - 109 Figure 48: Overall modeling structure............................................................................................. - 125 Figure 49: Optimal locations of the UCCs for the Scenario 1, 2 and 3 respectively ................ - 127 Figure 50: The seven steps of the MAMCA methodology (Source: Macharis, 2007).............. - 138 Figure 51: Scenarios for the Brussels-Capital Region ................................................................... - 139 Figure 52: Multi-actor view of stakeholder’s preferences for the scenarios .............................. - 148 Figure 53: Uni-actor view of LSPs evaluations for the scenarios................................................ - 148 -

x

List of Tables Table 1: Overview of battery electric freight vehicles ......................................................................- 16 Table 2: Attributes and levels considered in the design of the conjoint experiment. .................- 52 Table 3: Characteristics of the two versions of the Peugeot Partner.............................................- 57 Table 4: Input parameters of the vehicles..........................................................................................- 73 Table 5: Sensitivity analysis of TCO results on battery prices, fuel prices and residual value of the battery......................................................................................................................................................- 82 Table 6: Sensitivity analysis of TCO results on the level of subsidies, BEV deductibility, city access toll, urban kilometre toll and road taxes. ...............................................................................- 83 Table 7: Overview of literature review ...............................................................................................- 93 Table 8: Evaluation of decision scenarios with regards to business as usual ............................ - 127 Table 9: Identification of stakeholders in city logistics. ................................................................ - 141 Table 10: Objectives of the stakeholder’s groups and their importance .................................... - 142 Table 11: Summary of indicators and evaluation used for estimating the criteria .................... - 145 Table 12 : Evaluation of the scenarios compared to the business as usual scenario ................ - 146 -

xi

List of Abbreviations BAU

Business As Usual

BE

Belgium

BEV

Battery Electric Vehicle

CBC

Conjoint Based Choice

CNG

Compressed Natural Gas

CO

Carbon monoxide

CO2

Carbon dioxide

DK

Denmark

EC

European Commission

ES

Spain

EU

European Union

FI

Finland

FR

France

GE

Germany

GHG

Greenhous Gas

GVW

Gross Vehicle Weight

HGV

Heavy Good Vehicle

HoReCa

Hotels, Restaurants and Cafés

IT

Italy

LCV

Light Commercial Vehicle

xii Li-Ion

Lithium Ion

LSP

Logistics Service Providers

MAMCA

Multi Actor Multi Criteria Analysis

MRT

Maintenance, repair and tyre replacements

NEDC

New European Driving Cycle

NL

Netherlands

NMVOC

Non-Methane Volatile Organic Compounds

NO

Norway

NOx

Nitrogen Oxide

PM

Particulate Matters

SE

Sweden

SK

Slovakia

SOx

Sulphur Oxide

TCO

Total Cost of Ownership

UCC

Urban Consolidation Centres

UK

United Kingdom

VRP

Vehicle Routing Problem

WHO

World Health Organization

Introduction

-1-

1 Introduction 1.1 The challenges of the transport sector The transport system is a key driver of our economies. It enables trade and the movement of people. It reduces geographical constraints and makes regions more accessible. Better-connected locations benefit from lower transportation costs, which increases their attractiveness. It also improves the competitiveness of businesses located in such areas, creating additional wealth for these regions. An efficient transport system has, therefore, often been considered to play a critical role in economic development (Banister and Berechman, 2000). However, the transport sector exhibits market failures. The sector generates negative externalities that are not adequately reflected in transport prices (Friedrich and Bickel, 2001). Because prices guide markets towards equilibrium, the presence of externalities in the transport sector can lead to inefficient resource allocations. When externalities are negative, the volumes that are produced in the market are theoretically too high. Congestion might be the most observable example of market inefficiency in the transport sector. However, Stern (2007) argues that “climate change is the greatest market failure the world has ever seen”. But local pollution should not be overlooked given the considerable effects on public health through traffic noise and the emission of pollutants. Correcting these various externalities therefore represents important challenges for the transport sector in the future. Shifting away from oil is another key challenge for the transport sector. Europe’s dependence on oil creates indeed another type of market failure that can distort prices (Parry et al., 2007). To provide some background for this thesis, the externalities from global pollution, local pollution and congestion are introduced in this section, as is the oil dependency of the transport sector. Their recent evolution and the future plans at the European level are presented.

Introduction

-2-

Global pollution Anthropogenic greenhouse gas emissions are considered the dominant cause of the climate change observed since the mid-20th century (IPCC, 2014). These emissions are responsible for changing natural and human systems. Because the benefits of early action can outweigh the economic costs of not acting (Stern, 2008), signatory nations agreed to reduce greenhouse gas emissions in the “Kyoto protocol”, an international treaty that entered into force in 1997. The European Union is committed to that treaty and has integrated such objectives in its “Europe 2020” strategy. Greenhouse gas emissions should be reduced by at least 20% relative to 1990 levels (EC, 2010). In the long term, the objective is a minimum reduction of 80% by 2050 relative to 1990 levels (EC, 2011a). These achievements should contribute to the efforts needed to limit global warming to an average increase of 2°C. Reducing greenhouse gas emissions in the transport sector, however, is particularly challenging, as Figure 1 shows: transport was the only sector that did not contribute to a reduction in greenhouse gas emissions prior to 2007. Given the considerable share of such emissions attributable to road, the EU imposed specific objectives on automotive manufacturers to ameliorate that trend. The average CO2 emissions of all new cars registered in the EU should have a maximum of 130 g/km in 2015 and 95 g/km in 2020 (EU, 2009). That system was then extended to light commercial vehicles (LCVs), with the requirement of maximum average CO 2 emissions of 175 g/km by 2017 and 147 g/km by 2020 (EU, 2011). As Figure 1 indicates, there has been a reduction in greenhouse gas emissions in the transport sector since 2007. However, it remains unclear whether these new regulations are responsible for these reductions (EC, 2015) or they are primarily due to the global recession, meaning that emissions could increase again as the economy recovers (Hill et al., 2012). As of 2012, the transport sector remained the second greatest contributor to greenhouse gas emissions in Europe. Given the complexity of decarbonising the transport sector, the European objective for greenhouse gas emissions in the transport sector is less ambitious than the overall goal. They should be reduced by 60% in 2050 relative to 1990 levels (EC, 2011b). Intermediary steps were also devised: by 2030, the sector is required to achieve a 20% reduction in emissions relative to 2008 levels. Figure 1: Greenhouse gas emissions by sector for the EU28

Source: EC (2014)

Introduction

-3-

Air pollution In addition to the global challenge of reducing greenhouse gas emissions, transport also needs to address more local challenges. Exhaust gas emissions represent another type of externality in transport. Their impacts on air quality are particularly notable during smog alerts, when the concentration of pollutants reaches peak levels. Road transport is the main source of local pollutants from the transport sector, except for sulphur oxides (SOx), which are mostly generated by maritime transport. Of the total emissions of local pollutants generated in Europe in 2013, road transport was responsible for 12% of fine particulate matter (PM10), 13% of fine particulate matter (PM2.5), 22% of carbon monoxide (CO) and 39% of nitrogen oxide emissions (NOx) (EEA, 2015). Particulate matter (PM) is a solid core of carbon that can vary depending on particle size. PM with a size above 10 µm has less of an effect on human health because it is large enough to be filtered before entering the human lungs. However, when the size of PM decreases, the emissions penetrate more easily into human lungs, which can cause cardiovascular and lung diseases. PM2.5 is therefore considered particularly harmful to human health. Carbon Monoxide (CO) is a toxic gas coming from the incomplete combustion of diesel. At small concentrations, CO can affect the respiratory tract and cause cardiovascular disorders. At high concentrations, it can lead to death. Finally, nitrogen oxides (NOx) describe a combination of Nitrogen Monoxide and Nitrogen Dioxide. Emissions of Nitrogen Dioxide are particularly harmful to human health. High concentrations affect the respiratory tract and can reduce lung function (EEA, 2013). Given the impact of these pollutants on human health, Europe has established maximum emission levels to be respected by each member state (EC, 2008; EU, 2001). Recognising the contribution of road transport to such pollution, Europe has also established a set of maximum emissions levels for road vehicles. That system is called the EURO norms and has been updated regularly since its introduction in 1990. Manufacturers need to comply with these increasingly stringent benchmarks when applying for EC or national approval (EU, 2007). However, EURO norms are based on the New European Driving Cycle (NEDC), which has frequently been criticized. It represents a specific driving pattern with a time series of vehicle speeds that is employed to assess the emissions of each vehicle. However, the NEDC generally underestimates the real emissions of vehicles because of, among other factors, its assumption of a smooth acceleration profile (Pelkmans and Debal, 2006). As a result, EURO norms might have been effective in reducing various road pollutants (EEA, 2015), but these reductions have been insufficient to meet air quality standards in cities. As Figure 2 shows, pollution levels are above the World Health Organisation’s recommended levels and above EU targets. Figure 2 shows in particular that PM emissions, NOx emissions and CO emissions are among the most problematic pollutants to which road transport is an important contributor. Noise pollution Noise can also be considered a form of local pollution. In urban areas, road traffic is considered the dominant source of noise emissions (EEA, 2014). Persistent or high levels of noise can affect human health in various ways such as sleep disturbance, cardiovascular and physiological effects, mental health effects, annoyance or cognitive impairment. As 65% of Europeans living in cities are exposed to high noise levels (55db during the day or 50db at night), Europe is planning to

Introduction

-4-

implement a union-wide noise policy to approach the levels recommended by the WHO (EU, 2013). European legislation is however regulating already since 1970 the noise levels of motor vehicles given the important impact of road traffic on noise levels. That scheme was recently updated in 2014 (EU, 2014). Depending on the mass and the power of the vehicle, a maximum sound level should be respected by new vehicles, which will decrease over time (phase 1 starting in 2016, phase 2 in 2020 and phase 3 in 2024). Maximum values for cars range from 72 to 75 dB in 2016 and should be reduced to levels ranging from 68 to 72 dB in 2024. Higher levels are allowed for heavier vehicles: they can reach up to 82 dB in 2016 and should be reduced to 79 dB by 2024 (freight vehicles with gross vehicle weight of more than 12 tonnes and a rated engine power of 250kW).

Figure 2: Percentage of the urban population in the EU exposed to air pollutant concentrations above the EU and WHO reference levels (2009-2011)

Source: EEA (2013)

Oil dependence In addition to the environmental impact of transport, the transport sector must also face the future challenges of energy supply. In 1998, Europe was dependent on energy imports to approximately 36% of its total needs, a figure expected to rise to 60% by 2030 due to the exhaustion of North Sea oil and gas reserves, the steady or possibly declining share of nuclear energy and the lack of competitiveness of the European coal industry (European Commission, 2000). Disaggregating this 60% figure, Europe is expected to be dependent on foreign sources for 80% of its oil, 70% of its natural gas and 50% of its coal (European Commission, 2000).

Introduction

-5-

That rising dependence on foreign sources is particularly risky for the transport sector given that 96% of its energy needs relies on oil products (EC, 2011b). These suppliers will benefit from an increasing market power, which allows them to manipulate prices (Parry et al., 2007). In addition, energy prices are expected to increase in the long term given the increasing energy demand from fast-growing economies and the depletion of oil and gas reserves (Gnansounou, 2008). As a result, Europe and its transport sector in particular will face increasing costs due to energy imports. The EU’s import bill for oil was estimated at approximately €210 billion in 2010 according the European Commission (EC, 2011b) and at approximately €385 billion in 2012 according Cambridge Econometrics (2013). Recognising the strategic importance of energy, the EU is developing a strategy that promotes a shift from oil to more renewables and resource efficient solutions (European Commission, 2000). Congestion Congestion might be the market failure in the transport sector that receives the greatest attention in the media. It is also considered a priority in the various strategic documents published by the European Commission (EC, 2011b, 2010). Congestion is a major inefficiency in current transport systems. It reduces the accessibility of congested areas by increasing travel times. Congestion has also important environmental impacts. It primarily does so by exacerbating the impacts described above. It increases fuel consumption, greenhouse gas emissions, local pollutants and noise. As an example, it is estimated that congestion increases the EU fuel bill by 6% (EC, 2011c). Nevertheless, the European Commission offers no specific objectives to alleviate this problem. Instead, the total cost of congestion is often calculated. It is estimated to €110 billion annually in Europe (EC, 2012) and is expected to increase by approximately 50% by 2050 (EC, 2011b).

1.2 The role of city logistics The importance of city logistics The market failures of the transport sector that were described above affects mostly urban areas. Cities are indeed considered to suffer the most from congestion, poor air quality and noise exposure (EC, 2011b). As a result, these negative externalities reduce the attractiveness and competitiveness of urban areas (Crainic et al., 2004). However, cities are also geographical zones that should be priority supported given their critical importance in Europe. They contribute to 85% of the EU’s gross domestic product (European Investment Bank, 2011) and they concentrate 75% of the European population (World Bank, 2012). Urban transport systems deserve therefore a particular attention in the context of the future challenges of the transport sector. In cities, passenger transport captures most of the attention of local authorities when developing solutions for the urban transport sector (Anderson et al., 2005). According a survey of various municipalities in Sweden (Lindholm and Blinge, 2014), most of the time spent in transport planning is on car traffic, public transport and bike & pedestrian topics. Conversely, the time invested in freight transport issues is substantially more limited: 45% of the municipalities dedicate 10% of their time to freight transport, while 43% of the municipalities do not spend any time on freight transport issues. However, logistics represent an important dimension of urban transport on which authorities can act to improve the sustainability of a city. Freight vehicles also contribute to congestion, greenhouse gas emissions and local pollution (OECD, 2003).

Introduction

-6-

Cities are often equipped with different modes of transport. But road is the most preferred for goods distribution (Verlinde, 2015). On the road network, freight vehicles represent between 10 and 15% of a city’s traffic (Dablanc, 2007). However, their contribution to congestion is more pronounced because delivery operations typically occur on the street (Aiura and Taniguchi, 2006; Browne et al., 2010b; OECD, 1999). This common practice has considerable effects on traffic flow and exacerbates the contribution of freight vehicles to the congestion level of a city. In addition to affecting congestion, freight vehicles are also affected by congestion. They are entailed by among others an increased fuel consumption. As diesel is the standard fuel for freight vehicles, their fuel consumption is further affected by their short-distance trips and high stopping frequency, which are not the optimal conditions for diesel motors. As a result, freight transport is considered responsible for approximately 20% of energy consumption in road transport (Russo and Comi, 2012). Logistics has therefore an important impact on the oil dependency of the transport sector. However, the most important impact of city logistics is in the form of emissions. Freight vehicles are responsible for approximately one fourth of CO2 emissions, one third of NOx emissions and half of the particulate matter generated by the transport sector in large cities (Dablanc, 2011). This large share of emissions can be partly attributed to the predominance of diesel in freight vehicles. Freight vehicles also significantly contribute to noise emissions in the city: their traffic is estimated to generate an average of 2 to 10 times higher noise levels than that of passenger cars depending on the traffic conditions, the type of roads and the type of vehicles (Schoemaker et al., 2006). By monetarising the costs of these external impacts, Figure 3 shows the importance of freight vehicles relative to passenger cars.

External costs of urban transport

Figure 3: External costs due to urban transport 30 20 10 0 Greenhouse

Pollutants

Passengers (eurocents/pass-km)

Noise

Congestion

Freight (eurocents/t-km)

Source: adapted from Russo and Comi (2012)

The future of city logistics Future developments in the logistics sector are expected to intensify the pressure of city logistics on these externalities. The traffic of freight vehicles will grow at a much faster rate than that of private vehicles, which will increase the negative impacts of city logistics (Regué, 2013). According Schade and Krail (2010), forecasts predict 39% growth in freight transport in Europe by 2030 relative to 2006 (in ton-kilometres), while the growth of passenger transport should be limited to 16% (in passenger kilometres). This phenomenon is similar in other regions of the world such as in Australia (Taylor, 2005). Several trends can explain the growing importance of city logistics:

Introduction 





-7-

In a city of a developed country, 1 person is estimated to generate an average of 0.1 deliveries or pickups per day, a traffic of 0.3 to 0.4 trucks per day and a volume of 30 to 50 tonnes of goods per year (Dablanc, 2009). A growing population therefore implies more volumes, traffic and deliveries. However, the growth of urban economic activities will also contribute to increased freight traffic. The large national French programme “Marchandises en ville” estimated that 1 job generates 1 delivery or pick up per week on average. However, that impact will depend on the evolution of the economic structure of the city because that ratio varies according to sectors: tertiary, large retailers, industry, small retailers and wholesalers have ratios of 0.2, 0.8, 1, 1.8 and 4.1, respectively (Gerardin et al., 2000). The tertiarisation of urban economies might therefore alleviate somewhat the growth of freight traffic. Freight vehicles are expected to drive longer distances to deliver to urban establishments (Browne et al., 2010b). The delocalisation of logistics activities from the city to the periphery can explain this trend. Logistics activity consumes considerable space, which is becoming scarce and expensive in cities. However, the sector cannot cope with these increasing real estate costs because the activity does not generate substantial added value. As a result, the logistics platforms are moved further from the area they serve, entailing more vehicle kilometres driven by freight vehicles in a city (Dablanc and Rakotonarivo, 2010). Freight flows will likely be increasingly fragmented, which will multiply the number of freight vehicles in urban traffic. Just-in-time deliveries are requiring transporters to achieve more frequent and smaller deliveries to keep inventories at a minimum (Crainic et al., 2004; Dablanc, 2007; OECD, 2003). Additionally, the rise of e-commerce in B2B and B2C will also contribute to that trend (Taylor, 2005). The fine distribution required for delivering to households implies multiple small freight vehicles and reduces opportunities for consolidation (Crainic et al., 2004). As a result, LCVs have already become the dominant type of vehicle used in city logistics (Browne et al., 2010a; Verlinde, 2015) and are expected to absorb most of the growth in volume generated by the growing population (Zanni and Bristow, 2010).

In response to the growing negative impact of city logistics, local authorities have become increasingly interested in a sector that was traditionally left to the private sector (Allen and Wild, 2008). The role of local authorities has typically been limited to regulations such as time windows, vehicle weight restrictions access and street access limitations (Lindholm and Browne, 2013). These measures have, however, perverse effects. They increase freight traffic because of additional constraints imposed on transporters (Quak and Dekoster, 2007). Solutions to the negative impacts of city logistics should therefore not be considered solely from a regulatory perspective. Solutions should also consider the positive impacts of city logistics. The sector represents a critical foundation for urban economic vitality (Anderson et al., 2005) and the inhabitants’ lifestyle (Browne et al., 2007). Freight vehicles ensure that deliveries reach stores and maintain the connection between suppliers and customers. They contribute to the accessibility of the city and, thereby, the competitiveness of the urban economy (Russo and Comi, 2012). An efficient freight transport system guarantees an attractive environment for businesses and that consumer goods are offered to citizens at a competitive price. Moreover, city logistics is a major employer. It represents 2 to 5% of the employed urban workforce (Dablanc, 2011). The

Introduction

-8-

challenge for local authorities is therefore to implement solutions that stimulate the economic development of urban areas on the one hand and reduce the negative impacts of freight vehicles on the other. Towards sustainability To increase the attractiveness of cities, local authorities are increasingly integrating the concept of sustainability into their policies. In the context of transport, sustainable mobility can be defined as “a form of mobility that doesn’t hold any danger to human health or ecosystems and still complies with the need for accessibility, but in such a way that the consumption of renewable energy sources is smaller than its regeneration speed and the consumption of exhaustible energy sources is slower than the development of renewable substitutes” (Macharis and Van Mierlo, 2013). Sustainable mobility therefore essentially focuses on solutions that reduce the environmental impacts of transport (Macharis and Van Mierlo, 2013). The integration of sustainability can also be observed in city logistics. Potentially the most important evidence in this regard is the objective established by the European Commission in its white paper on transport: major urban areas should achieve CO2-free city logistics by 2030 (EC, 2011b). To achieve this ambitious objective, greater attention in that area is required. At present city logistics is recognised as a specific aspect of urban transport. The OECD created a working group dedicated to the topic of city logistics, which issued its first report in 2003 (OECD, 2003). Every two years, the Institute for City Logistics organises an international conference specifically on that topic. Additionally, the World Conference on Transportation Research (WCTR) has gathered a specific interest group on urban freight transport. Finally, numerous doctoral theses confirm city logistics as a specific research field including, among others, the theses of Dablanc (1997), van Binsbergen and Visser (1999), Quak (2011), Behrends (2011), Lindholm (2012), Anand (2015) and Verlinde (2015). The development of sustainable city logistics has resulted in, among other outcomes, a portfolio of solutions (Arvidsson et al., 2013; Muñuzuri et al., 2005). These solutions can be classified according the four ‘A’s of Macharis (2013): Awareness, Avoidance, Act & shift and Anticipation. The first A of Awareness refers to the actions that promote the need for sustainable logistics in this sector. These actions are intended to improve the understanding of the total impact of transport activities. Including externalities in transport decisions is an example of solution that increases awareness. It facilitates making better-informed decisions that recognise the social and environmental costs of transport. Congestion pricing is, in this respect, an example of integration of external costs from congestion (van Lier, 2014). However, its impact on freight transport patterns remains contested (Quak and van Duin, 2010; Russo and Comi, 2012). The second A is Avoidance and refers to the solutions that reduce transport by optimising the load rate of vehicles. In that category, the urban consolidation centre (UCC) is among the most popular solutions in city logistics (Woodburn et al., 2005). However, other forms of freight consolidation are possible in the absence of a consolidation centre (Verlinde et al., 2011). The third A is Act & shift and refers to solutions that reduce the impact of transport by using other modes of transport. The barge and rail modes offer better environmental performance compared to road transport (van Lier and Macharis, 2011). Although intermodal transport is mostly used for longdistance trips, barge and rail were shown to have some potential in city logistics (Mommens et al., 2014). Freight trams are also investigated as a means of decreasing the share of road transport in cities (Regué, 2013; Strale, 2014). Within road transport, the use of cyclocargos represents an

Introduction

-9-

interesting alternative to motorised vehicles (Schliwa et al., 2015). The European project “Cyclelogistics” was dedicated to that specific solution (2011-2014). Act & shift also refers to the potential of off-hour deliveries, which has been demonstrated in various trials (Verlinde and Macharis, 2015). Finally, the last A, Anticipation, refers to the future technologies that can improve the sustainability of city logistics. Environmentally friendly vehicles represent an important potential in that category for city logistics.

1.3 Electric vehicles as a solution? Relevance of electric vehicles to transport challenges Given the sustainable objectives imposed on manufacturers in Europe, new environmentally friendly vehicles are appearing on the market. These vehicles are defined as a “complex ensemble of technologies that contribute to a lower environmental impact along the total chain of energy conversations from the well to the wheel” (Van Mierlo et al., 2006). Among these technologies, electric vehicles are believed to have achieved momentum since 2005 (Dijk et al., 2013). The number of companies producing electric vehicles has increased significantly in recent, reaching 76 in 2011 (Sierzchula et al., 2012). These vehicles are attracting particular attention for several reasons. The electric motor has the benefit of producing zero emissions. Still, to assess the real environmental impact of electric vehicles, a comprehensive approach needs to be used. Indeed, electric vehicles generate emissions through the electricity they consume. As a result, the electric vehicle is sometimes referred as a “displaced emission vehicle” instead of a “zero emission vehicle” (OECD, 2012). Its environmental impact will therefore depend on the energy source used to produce electricity. Hence, a full life-cycle analysis should be used to consider the total impacts from the well to the wheel. Messagie et al. (2014) used that approach to compare the environmental impact of diesel and petrol vehicles with that of electric vehicles. As Figure 4 shows, electric vehicles benefit from a lower impact on CO2 emissions, especially when the energy source used to produce the electricity is renewable. That figure thus stresses the importance of shifting to more sustainable electricity production in order to fully benefit from the environmental performance of electric vehicles. In so doing, electric vehicles can contribute more effectively to reducing greenhouse gas emissions in the transport sector. The fact that the electric motor does not generate emissions is especially interesting for cities. Since traffic is an important contributor to local emissions, electric vehicles can contribute to a better air quality. Soret et al. (2014) estimated that the electrification of the fleet would particularly decrease the emissions of NOx and CO in cities. Fully electrified vehicles decrease these emissions from the tank to the wheel by 100% relative to conventional vehicles (Rangaraju et al., 2015). However, their contribution to reducing PM emissions is less drastic. Vehicles produce also non-exhaust emissions that are generated by the suspension and brakes, and tyre and road abrasion. As a result, the reduction in PM emissions from the tank to the wheel is more limited for a fully electrified vehicle: it generates 33% less PM than a diesel vehicle (Rangaraju et al., 2015).

Introduction

- 10 Figure 4: Impact of different vehicle technologies on CO2 emissions.

Source: Messagie et al. (2014)

Electric vehicles also have the potential to reduce the noise level in cities. An electric motor is much quieter than internal combustion engines (Van Mierlo and Maggetto, 2007). Verheijen and Jabben (2010) estimated that the noise could be reduced by 4 dB in cities if 90% of cars and 80% of heavy trucks were electrified. Some argue, however, that the silence of electric vehicles could be dangerous, especially at low speeds. As a result, continuous warning sounds have been imposed on electric vehicles in America, and a similar regulation is coming to Europe: by the 1 st of July 2019, manufacturers should have installed an acoustic vehicle alerting system on all electric vehicles that “should sound similar to the sound of a vehicle of the same category equipped with an internal combustion engine” (EU, 2014). This continuous warning signal thus somewhat offsets the noise advantage. However, that regulation applies until vehicle speed reaches 20 km/h. That regulation might also be further adapted for specific applications such as night deliveries. However, one of the most important benefits of the electric vehicle might be the solution it provides to the energy challenges facing the European economy. On the one hand, the electric vehicle contributes to reducing the transport sector’s dependence on oil products as long as the electricity production is not dependent on fossil fuels. On the other hand, the electric motor is considerably more efficient than are the engines of the conventional vehicles, thereby ensuring a better utilization of resources. It is estimated that the efficiency rate of an electric motor is approximately 80 to 90% (van Vliet et al., 2011), while the internal combustion engines used in conventional vehicles have an efficiency rating below 15% in urban environments (Van Mierlo and Maggetto, 2007). The ability of the electric vehicle to recover energy through regenerative braking is an additional energy efficiency that benefits electric vehicles. Early market development Since 2010, the market has witnessed the introduction of several electric vehicles (Lebeau, 2013). The vehicles currently proposed by manufacturers can be classified according their degree of electrification, which can be defined by the motorisation and energy used by the vehicle (MIT,

Introduction

- 11 -

2010). A vehicle with 0% electrification can be classified as a conventional vehicle, while a vehicle with 100% electrification can be classified as a battery electric vehicle (BEV) or a fuel cell electric vehicle. In between, hybrid forms employ both a combustion engine and an electric motor. These hybrid forms include plug-in hybrid electric vehicles and extended-range electric vehicles, but Westbrook (2001) recognises that this balance offers an almost infinite number of future innovative designs. Following the development of the products supplied by the manufacturers, the sales of electric vehicles increased rapidly in the European Union, as shown in Figure 5. The market penetration of electric vehicles remains however limited. The sales are driven by the largest European countries but particularly by countries with policies supporting the introduction of electric vehicles. Figure 5 shows that in 2013, the Netherlands had the largest fleet of electric vehicles in the European Union (EU28) despite the small size of the country. The market share of electric vehicles reached 5.43% of national new passenger car registrations (ICCT, 2014). However, this number is mostly attributed to the sales of plug-in hybrid vehicles. The market share of BEVs is limited to approximately 1%. This limited performance of BEVs in the most electrified vehicle fleet in Europe can be explained by the barriers to their adoption by consumers. According to a survey conducted among households by Lebeau et al. (2013), the most critical disadvantages of BEVs were identified as their high purchase price and limited driving range. Only 10% of the respondents were satisfied with a range of 200 km. Additionally, only 27% of the respondents are willing to pay more for a BEV. They instead expect a price similar to that of a conventional car. The lack of charging infrastructure, the long charging time and the impossibility of charging at home because of the absence of a garage are the set of second most critical barriers. However, the expected evolution of BEV technology should reduce these barriers and increase the adoption of BEVs (Lebeau et al., 2012). Figure 5: Evolution of electric passenger cars in the European Union (combining battery electric vehicles, fuel cell electric vehicles and plug-in hybrid vehicles) 0.5%

50,000

0.4%

40,000 0.3% 30,000 0.2%

20,000

0.1%

10,000 0 2009 Germany

2010 UK

France

2011 Italy

Netherlands

2012 Sweden

Source: adapted from ICCT (2014)

0.0% 2013 Other EU countries

EV market share

Total sales/registrations

60,000

Introduction

- 12 -

Nevertheless, BEVs are likely to continue to face more operational constraints in the future than are conventional vehicles. In addition to improvements in the technology and government support, the introduction of BEVs requires also a change in the mentality of consumers (Lebeau et al., 2013). The diffusion of new technologies in the market requires indeed time. BEVs are therefore expected to also follow the process of the cycle of innovations described by Rogers (1983), which includes four important steps in the diffusion of innovations and is depicted in Figure 6. An innovation is first adopted by consumers who are obsessed with acquiring new technological products. They enjoy experimenting with new ideas and sharing them with other innovators in social networks. They are called innovators. Although they represent only a small part of the customer base, they are critical in the cycle of innovation diffusion because they introduce the technology in the marketplace and show that the technology is reliable. Early adopters are also characterised by their interest in new technological products. However, they do not purchase them first because they are new – the reason that innovators do so – but because they understand and appreciate the benefits of this innovation. They have sufficient financial resources to take the risk of investing in a new technology. They are believed to make judicious innovation decisions and are respected for that by their peers. As a result, they become references who can advise potential adopters. Their most important characteristic is therefore their opinion leadership. They reduce the uncertainty in the cycle of innovation diffusion. That will help to convince the early majority, which represents an important part of the customer base, and ensure a profitable business for the innovation. Members of the early majority interact frequently with early adopters but do not hold leadership positions. This group of consumers prefer to see the benefits of an innovation from its use by early adopters before investing in it. The late majority is another important part of the customer base. They adopt the innovation later because of economic necessity or increasing social pressure. They are less comfortable with new technologies and therefore prefer to wait until the innovation becomes standard to acquire it. Finally, laggards are consumers who resist acquiring new technologies. They will do so only if they are forced to. According a survey among BEV users (Trommer et al., 2015), their profile suggests that the diffusion process of BEVs is reaching the early adopters phase: BEV users have higher incomes and are interested in the innovative aspect of BEVs, but their purchase was also motivated by the reduced environmental impact of BEVs. Efforts should therefore be made in identifying new early adopters to stimulate the diffusion process of BEVs. In this context, Sierzchula (2014) stresses the importance of fleet managers and organisations such as taxi fleets, car sharing companies or public agencies. They have been actively supporting the diffusion of BEVs during the early stage of these vehicles’ introduction.

Introduction

- 13 Figure 6: The diffusion of innovations

Source: Rogers (1983)

BEVs in city logistics In the quest of searching for new early adopters, fleets in city logistics represent a potential group that deserves a greater attention (Lebeau, 2013; Van Mierlo and Maggetto, 2007; den Boer et al., 2013). City logistics is indeed considered to offer a suitable environment for BEVs where their barriers are relaxed and their benefits are more valorised compared to the passenger car segment. Limited range in particular is assumed to be less constraining for three main reasons. First, city logistics is characterised by smaller distances: it has been estimated that more than 80% of freight trips in European cities are shorter than 80 km which is compatible with the limited range of BEVs (BESTUFS II, 2008a). Second, range can be more easily controlled by the structured and time-based environment of the logistics chain (E-Mobility NSR, 2013). Finally, daily range can be extended by recharging batteries during vehicles’ loading/unloading operations at the depot (Macharis et al., 2007). Long charging times are therefore also less constraining because it takes place at the depot. On the other hand, transport companies can profit from BEVs’ low operating costs, a consideration that can outweigh their high purchase costs. These considerations have led to a number of research projects. Between 1994 and 1998, the first European research project dedicated to city logistics, called COST321, investigated a large set of possible solutions to improve the sustainability of city logistics and identified BEVs to be one of the most promising alternatives. Given that assessment, new European research projects in city logistics were launched with a focus on BEVs. ELCIDIS and EVD-Post (1998-2002) were the first two projects testing the operations of BEVs in different urban supply chains such as postal services and UCC. Since then, numerous city logistics projects supporting sustainable trials have included BEVs as part of their solutions such as CO2NeuTrAlp (2008-2011) CIVITAS (20022016), STRAIGHTSOL (2011-2014), ENCLOSE (2012-2014), and SMARTFUSION (20122015). Currently, FREVUE (2013-2017) can be considered an update of the ELCIDIS project

Introduction

- 14 -

made necessary by the evolution of the technology. These projects stimulated the emergence of new logistics operators offering green transport services with BEVs. Traditional actors in the sector also began to use BEVs for their last mile operations such as DHL, UPS and DPD (Schneider et al., 2014). To track these developments, other European research projects have collected the different cases to promote them and learn from them. The following EU-funded projects can be reported in this regard: BESTUFS (2004-2008), SUGAR (2008-2012), TURBLOG (2010-2013), C-LIEGE (2011-2013), E-Mobility NSR (2011-2014) and BESTFACT (2012-2015). Based on these research projects, Figure 7 offers an overview of the various cases referenced by these documents in Europe where BEVs were used in the logistics. Appendix 2 gives more detailed information on these different demonstrations. However, this overview remains non-exhaustive because other projects are also conducted at the national or even regional level. Figure 7: Overview of city logistics projects involving BEVs

Source: map generated by www.espatial.com (see the details in Annex 2)

These projects have mostly assessed their contribution in reducing externalities of freight transport in cities. The most studied case is the UCC of La Rochelle that was implemented in 2001 under the European research project ELCIDIS and remains operational. It employs an allelectric fleet for the last mile, which currently comprises 6 LCVs (Citroen Berlingo) and 1 HGV (Modec). According to the evaluations of the ADEME (2004), the new logistics organisation decreased the noise level by 61%, energy consumption was reduced by 48%, and the emissions were cut by approximately 48%. In greater detail, NOx, CO, PM and greenhouse gas emissions were reduced by 48% and SO2 by 50%. However, congestion increased by 174%. This negative

Introduction

- 15 -

impact in terms of congestion, however, cannot be attributed to BEVs. Their limited payload is reducing the potential for consolidation, which increased vehicle kilometres of the scheme in La Rochelle. The project was more focused on environmental performance than on the optimization of freight flows. The case of Gnewt in London is a more extreme case as observed by Browne et al. (2011). That UCC uses electric quadricycles (Aixam Mega) and electrically assisted tricycles. This UCC thus operates vehicles with payloads that are even more limited than those in La Rochelle. As a result, the freight captured by the UCC generated a 349% increase in vehicle kilometres in the city. Nevertheless, the total environmental performance of the scheme achieved a CO2 reduction of 54% on the flows captured by the UCC. However, other city logistics using BEVs in the last mile exhibited a decrease in congestion. The Cargohopper, for example, in Utrecht resulted in a reduction of approximately 4,000 freight vehicle trips in the city (van Rooijen and Quak, 2014). It saved 24.000 litres of diesel fuel per year and 34 tonnes of CO2 emissions per year (C-Liege, 2012). Similarly, the Stadsleveransen service in Gothenburg was estimated to reduce congestion and noise levels and to improve the attractiveness of the city centre (Widegren, 2014). However, in both of these schemes, the payload of the BEVs remained as limited as in the distribution schemes of Gnewt in London and ELCIDIS in La Rochelle. The difference with these projects relies in the towing capacity of the vehicle which was used at its full potential by using trailers, as shown in Figure 8. This solution alleviated the limited payload of the quadricycles, avoiding the growth in congestion observed in London and La Rochelle. Figure 8: the Cargohopper in Utrecht

Source: EV World (2009)

The number of projects involving larger BEVs is more limited. The UCC serving the cities of Bristol and Bath, for example, successfully introduced a 9 t battery electric truck (Smith Newton). It achieved a reduction of 55.7% in energy consumption over equivalent diesel delivery vehicles, also leading to reduced emissions (CIVITAS, 2015). The UCC of Siena also operates 6 electric LCVs of 3.5 tonnes. It showed a reduction of 37% in the number of freight vehicles in the city (Woodburn et al., 2005). However, these schemes use a mixed fleet with conventional or other

Introduction

- 16 -

environmentally friendly vehicles. As a result, it remains difficult to evaluate the net contribution of BEVs to the performance of such schemes. In that context, the evaluation of the Chronopost project in Paris is particularly interesting because it measured the specific contribution of BEVs to the sustainability of logistics operations. It estimated that, within a UCC, two thirds of CO2 emissions reductions are attributable to the use of BEVs and one third to the new logistics organization (C-Liege, 2012; SUGAR, 2011). In that project, BEVs saved 41,000 km of trips by fuel-powered vehicles. Nevertheless, the important lack of evaluations on the specific contribution of BEVs can be stressed in city logistics. FREVUE (2013) recognises also that the majority of the literature and information available on BEVs in city logistics is rather focused on the description of the context and of the trial experience. Furthermore, the limited sales do not indicate whether city logistics is a suitable environment for BEVs. Indeed, a recent state of the art of our research field achieved by Samuel Pelletier et al. (2014) could not conclude on the potential of BEVs in city logistics. That research gap should therefore be addressed urgently. Addressing this research gap is indeed pressing given the recent development of BEVs by manufacturers. Since 2010, some important brands are securing a leading position with the introduction of BEVs in the segment of LCVs. The new European regulation on the average CO2 emissions of LCVs might have contributed to this evolution (EU, 2011). However, the expected evolution of city logistics might have been another incentive for manufacturers to extend their portfolios of LCVs to electric vehicles for two major reasons as described in section 1.2. First, freight vehicles are expected to generate a faster growth than passenger cars, particularly in the LCV segment. Manufacturers are therefore developing their portfolio of freight vehicles to better meet the various needs of that growing customer base. The second reason comes from the sustainability objectives developed by authorities. Cities are indeed increasingly regulated areas where the environmental performance of vehicles becomes a competitive advantage. A non-exhaustive overview of freight BEVs available in Europe is provided in Table 1. Table 1: Overview of battery electric freight vehicles Vehicle

GOUPIL G3-2

GOUPIL G5E

MEGA e-worker

MEGA e-truck

Availability (in Belgium)

Yes

Yes

Yes

Yes

Segment

L7

L7

L7

L7

GVW

Battery type

Battery capacity

Range

Lead Acid (24 cells of 2 V)

8,6kWh

53 km*


11,5kWh

70 km*


15,4kWh

95 km*


19,2kWh

110 km*


8,6kWh

40 km*


11,5kWh

55 km*


14,4kWh

70 km*


8,6kWh

60 km


11,5kWh

80 km


17,3kWh

110 km

Lithium-Ion

9,2kWh

70 km

2075 kg

1950 kg

1640 kg

1110 kg

Introduction

ALKE ATX210E

ALKE XT320

- 17 -

Yes

Yes

L7

N1

1500 kg 1500 kg 2800 kg 2800 kg 2800 kg


8,7kWh

70 km

Pure Lead (4 cells of 12 V)

9kWh

70 km


13kWh

75 km


18kWh

110 km


26kWh

140 km

MUSES Mooville

Yes

N1

2400 kg


13,7kWh

70 km*

KANGOO Express ZE

Yes

N1

2126 kg

Lithium-Ion (leasing)

22kWh

80-125 km*

PEUGEOT Partner Electric

Yes

N1

2225 kg

Lithium-Ion

22,5kWh

170 km**

CITROEN Berlingo Electric

N.A.

N1

2225 kg

Lithium-Ion

22,5kWh

170 km**

Yes

N1

2220 kg

Lithium-Ion

24kWh

163 km**

Yes

N1

2220 kg

Lithium-Ion (leasing)

24kWh

163 km**

Sodium Nickel Chloride

42,3kWh

90 km

N.A.

N1

3500 kg Sodium Nickel Chloride

63,4kWh

120 km

Lithium-Ion

36kWh

110 km

Lithium-Ion

51kWh

145 km

Lithium-Ion

80kWh

115 km

Lithium-Ion

120kWh

160 km

NISSAN e-NV200

IVECO EcoDaily

SMITH Edison

SMITH Newton

2016-2017

2016-2017

N1 N1

3500 kg

N2

7500 kg

RENAULT Maxity

Yes

N2

4500 kg

Lithium-Ion

40kWh

100 km*

FUSO Canter E-cell

2018

N2

6000 kg

N.A.

N.A.

100 km

* Range estimated by the manufacturer ** Range according the NEDC

1.4 Research questions The previous sections have demonstrated the important challenge that city logistics face. Although the negative impacts of freight transport are expected to grow in the future, logistics should become CO2 free by 2030 in major urban areas. In that context, BEVs represent an interesting solution. The objective of this thesis is therefore to explore the feasibility of introducing battery electric vehicles into urban supply chains. To do so, three main research questions will be addressed throughout the thesis. RQ1: What is the potential of battery electric vehicles in city logistics? The introduction showed the limited penetration of BEVs in the European automotive market despite their environmental performance. However, the diffusion of innovation requires time. This new technology demands an important change in consumer mentality, in which the behaviour of early adopters is key. Captive fleets owned by taxi companies, car sharing companies or public agencies have been observed to be critical players in the early development of the market. Fleets in city logistics might be another future adopter. The first research question will therefore explore the potential demand for BEVs in that segment. The context of that sector will be first presented in chapter 2 to show the need to implement sustainable alternatives in city

Introduction

- 18 -

logistics. Then, the potential of BEVs will be investigated in chapter 3 by exploring the behaviour of transport operators. RQ2: What are the factors that can stimulate a shift from diesel to battery electric vehicles? The limited potential of BEVs in city logistics can be explained by the choice behaviour of transport operators that is investigated in chapter 3. Financial and operational constraints are mentioned as the most challenging barriers. In the context of city logistics, where performance is key and competitiveness is high, these constraints are indeed hindering particularly the adoption of BEVs. The second research question will therefore address these barriers through the development of tools. They will help transport operators to better identify the factors that can reduce (or even eliminate) these barriers. Chapter 4 will investigate the financial constraints from a total cost of ownership perspective. Then, operational constraints will be addressed in chapter 5 by integrating the results of chapter 4 to a vehicle routing problem. RQ3: What is the level of support from stakeholders for the electrification of city logistics? Authorities are often considered key actors in supporting the introduction of BEVs. However, previous projects in city logistics have shown that a top-down approach might not be the most effective method to drive change in this sector. Hence, this last research question will be less focused on the perspective of freight transport operators. Instead, the focus will be on the behaviour of stakeholders in city logistics. Their position regarding the electrification of city logistics will be assessed based on a set of scenarios. Chapter 6 will develop these scenarios and estimate their impacts based on the tools developed in the two previous chapters. Chapter 7 will then supplement the analysis of chapter 6 by incorporating the preferences of stakeholders in city logistics through a multi-actor multi-criteria analysis. It will show the support that the various stakeholder groups can bring to the electrification of city logistics.

1.5 Structure of the dissertation To address these research questions, the thesis was divided into six main chapters, grouped in three parts. Each part addresses a research question. Each chapter was accepted or submitted to peer-reviewed journals and peer-reviewed scientific conference proceedings. In this section, the role of each chapter in the thesis is described. Then, the contributions of the co-authors to the different papers are detailed. Annex 1 gives the full list of the publications produced during the PhD dissertation.

The objective of the first part is to address the first research question. It will introduce the context of the thesis and investigate the purchase behaviour of transport operators. Chapter 2: Freight transport in Brussels and its impact on road traffic The first chapter of the thesis will introduce the context in which the research is performed. The challenges of city logistics will be presented for the Brussels-Capital Region. The available data in the Region are collected and merged to provide the most comprehensive overview of the sector possible. The chapter presents the recent evolution of the sector, the current patterns of freight flows and the impact of freight vehicles on the sustainability of the city. Based on that analysis,

Introduction

- 19 -

the various solutions that are being developed in Brussels are presented. It will show in particular that BEVs are considered by the authorities as one of the possible solutions to improve the sustainability of city logistics. Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a conjoint-based choice analysis. After presenting the context of city logistics, the potential of BEVs in city logistics will be investigated in chapter 3. It will explore the purchase behaviour of urban transport operators. Based on a review of various experiences using BEVs in logistics, the different attributes influencing the choice of BEVs are identified. The most relevant attributes are then used to model the preferences of transporters through a conjoint-based choice analysis. These preferences are collected through a survey conducted among freight transport operators in the Brussels-Capital Region. The survey also explores the attitudes of the transporters to obtain further insights into their purchase behaviour. A discussion compares these results and highlights the most important aspects influencing BEVs adoption in city logistics. It identifies also the most important measures to implement in order to improve the potential of BEVs in city logistics.

The objective of the second part is to address the second research question. A chapter will be dedicated to the two most important barriers identified in the first part: the financial and operational constraints of BEVs in city logistics. Tools will be developed to better understand and address these barriers. Chapter 4: Electrifying light commercial vehicles for city logistics? A total cost of ownership analysis. Chapter 3 indicated that costs represented the second most important barrier to the adoption of BEVs. They also represent the most efficient driver for policy makers to stimulate BEVs’ adoption. As a result, chapter 4 will investigate the competitive position of BEVs relative to conventional technologies. A model is developed to compare the total cost of ownership of 15 vehicles in a city logistics context. A sensitivity analysis of the different cost components in the total cost of ownership will identify the breakeven points at which BEVs become competitive to their conventional counterparts. Based on these findings, different policies are recommended to effectively support the competitiveness of BEVs. Chapter 5: Diesel or/and battery electric vehicles: Which mixed technology fleet for city logistics? Chapter 3 also identified operational constraints as the most critical aspects in the adoption of BEVs in city logistics. Chapter 5 therefore investigates the constraints related to long charging time and limited range. It develops a fleet size and mix vehicle routing problem in which the distribution of a fleet with different vehicle technologies is modelled. It considers the cost structure of the different vehicles, the capacity constraints of the vehicles, the time windows of the customers receiving the goods and the battery constraints of BEVs. The results demonstrate the operational and economic feasibility of introducing BEVs in urban distribution.

Introduction

- 20 -

The objective of the third part is to address the last research question. It will estimate the impact of the introduction of BEVs on city logistics and evaluate the support that the various stakeholder groups can bring to the electrification of city logistics. Chapter 6: Strategic scenarios for sustainable urban distribution in Brussels-Capital Region using urban consolidation centers. The introduction of BEVs in city logistics was investigated in chapter 6 in the context of a UCC. Strategic scenarios are designed to improve the sustainability of the logistics in the BrusselsCapital Region. They present different combinations of urban distribution centre(s) with BEVs, road pricing and night distribution. To better evaluate the impact of the introduction of BEVs in city logistics, some scenarios differ only in the type of fleet operated for the last mile of the UCC. The impacts of these scenarios are estimated based on the tools developed in the first parts of the thesis. Chapter 7: Implementing an urban distribution centre: involving stakeholders in a bottom-up approach The last chapter focuses on the stakeholders in city logistics in the Brussels-Capital Region. Their support for the various scenarios developed in chapter 6 is evaluated through a multi-actor multicriteria analysis. The chapter presents the results of a workshop conducted with the stakeholders and identifies the most implementable scenarios that improve the sustainability of city logistics. Because some scenarios differ only in the type of fleet operating the last mile of the UCC, the support of the stakeholders for an electrification of city logistics can be highlighted. The conclusion will finally use the findings developed in the chapters to answer the research questions that were described in the previous section of this introduction. The contributions of these findings will also be highlighted. Finally, the limitations of these conclusions will be stressed. It will provide the opportunity to suggest avenues for future research.

The different chapters of the thesis are identical to the text that was either published or submitted to journals. Their statuses at the time of the publication of this thesis are presented in the footnotes on the first page of each chapter. Only chapter 5 was published and then updated for this thesis such that it better contributes to addressing the research questions. This thesis was conducted under the supervision of Prof. Dr. Cathy Macharis as promoter and Prof. Dr. Joeri Van Mierlo as co-promoter, both leading the research group MOBI at the Vrije Universitiet Brussel. They ensured the quality of this thesis and of the various publications by providing feedback. However, other authors also worked on the publications included in this thesis. Below, the contribution of each author is described. 

Prof. Dr. Thierry Coosemans is affiliated with the Vrije Universiteit Brussel and a member of the research group MOBI. He contributed to the chapter 5 by providing feedback on the manuscript.

Introduction 

 





- 21 -

Prof. Dr. Wouter Verbeke is affiliated with the Vrije Universiteit Brussel and a member of the research group MOBI. He contributed to the development of the algorithm that solves the FSMVRPTW-EV in chapter 5. Prof. Dr. Alassane Ndiaye is affiliated with the Université Libre de Bruxelles. He contributed to chapter 6 by revising the manuscript. Dr. Kenneth Lebeau was affiliated with the Vrije Universiteit Brussel and is a former member of the research group MOBI. He contributed to the development of the methodology in chapter 4 and provided feedback on the manuscript. Milena Janjevic is affiliated with the Université Libre de Bruxelles. She collaborated with Philippe Lebeau on the Prospective Research for Brussels entitled “A strategy for the implementation of a sustainable logistics concept for the city distribution of the BrusselsCapital Region” and financed by Innoviris. She summarized in chapter 6 the results of the discussions between the various authors on the design of the scenarios. The assessment of the scenarios was jointly performed by Milena Janjevic and Philippe Lebeau. Milena Janjevic also contributed to chapter 7 by providing feedback on the manuscript. Cedric De Cauwer is affiliated with the Vrije Universiteit Brussel and a member of the research group MOBI. He contributed in chapter 5 to the development of the energy consumption model. He pursued that work subsequently in the following paper: De Cauwer et al. (2015). He provided also feedback on the manuscript.

Introduction

- 22 -

References ADEME, 2004. Espaces logistiques urbains de monaco et la rochelle. Afroditi, A., Boile, M., Theofanis, S., Sdoukopoulos, E., Margaritis, D., 2014. Electric Vehicle Routing Problem with Industry Constraints: Trends and Insights for Future Research. Transp. Res. Procedia 3, 452–459. Aiura, Taniguchi, 2006. Planning On-Street Loading-Unloading Spaces Considering the Behaviour of PickupDelivery Vehicles and Parking Enforcement, in: Taniguchi, E., Thompson, R. (Eds.), 4th International Conference on City Logistics. Elsevier, Langkawi, Malaysia, 12-24 July 2005, pp. 107–116. Ajzen, 1985. From intentions to actions: A theory of planned behavior, in: Kuhl, J., Beckmann, J. (Eds.), ActionControl: From Cognition to Behavior. Heidelberg, pp. 11–39. Allen, J., Wild, D., 2008. BESTUFS II Policy and Research Recommendations II on Environmental Zones in European Cities, Accommodating the needs of passenger and freight transport in cities, and BESTUFS Project Recommendations. Amburg, B. Van, Pitkanen, W., 2012. Best Fleet Uses , Key Challenges and the Early Business Case for E-Trucks : Findings and Recommendations of the E-Truck Task Force, in: EVS26. Los Angeles, California, May 6-9, pp. 1–12. Anand, N., 2015. An Agent Based Modelling Approach for Multi- Stakeholder Analysis of City Logistics Solutions An Agent Based Modelling Approach for Multi-Stakeholder Analysis of City Logistics Solutions. Delft University of Technology. Anderson, S., Allen, J., Browne, M., 2005. Urban logistics––how can it meet policy makers’ sustainability objectives? J. Transp. Geogr. 13, 71–81. Arvidsson, N., Woxenius, J., Lammgård, C., 2013. Review of Road Hauliers’ Measures for Increasing Transport Efficiency and Sustainability in Urban Freight Distribution. Transp. Rev. 33, 107–127. Banister, D., Berechman, J., 2000. Transport investment and economic development. UCL press, London. Bard, J.F., Huang, L., Dror, M., Jaillet, P., 1998. A branch and cut algorithm for the VRP with satellite facilities. IIE Trans. 30, 821–834. Behrends, S., 2011. Urban freight transport sustainability: The interaction of urban freight and intermodal transport. Chalmers University of Technology, Gothenburg. Belfiore, P., Yoshizaki, H.T.Y., 2013. Heuristic methods for the fleet size and mix vehicle routing problem with time windows and split deliveries. Comput. Ind. Eng. 64, 589–601. BESTUFS II, 2006. Quantification of Urban Freight Transport Effects II. BESTUFS II, 2008a. Best Practice Update (2008) - Public Private Partnerships (PPP) in urban freight transport. BESTUFS II, 2008b. Best Practice Update (2008) on Waste transport logistics in urban areas & Experiments and incentives for environment-friendly vehicles. Blanco, S., 2013. 2013 Nissan Leaf gets 75-mile range (actually 84) in new EPA test [WWW Document]. Autobloggreen. URL http://green.autoblog.com/2013/02/21/2013-nissan-leaf-revealed-gets-75-mile-rangeactually-84-in-n/ (accessed 5.20.14). Bräysy, O., Dullaert, W., Hasle, G., Mester, D., Gendreau, M., 2008. An Effective Multi-restart Deterministic Annealing Metaheuristic for the Fleet Size and Mix Vehicle Routing Problem with Time Windows. Transp. Sci. 42, 371–386. Browne, M., Allen, J., Leonardi, J., 2011. Evaluating the use of an urban consolidation centre and electric vehicles in central London. IATSS Res. 35, 1–6. Browne, M., Allen, J., Nemoto, T., Visser, J., 2010a. Light goods vehicles in urban areas. Procedia - Soc. Behav. Sci. 2, 5911–5919. Browne, M., Allen, J., Steele, S., Cherrett, T., McLeod, F., 2010b. Analysing the results of UK urban freight studies. Procedia - Soc. Behav. Sci. 2, 5956–5966. Browne, M., Piotrowska, M., Woodburn, A., Allen, J., 2007. Literature Review WM9 : Part I - Urban Freight Transport, Transport. London.

Introduction

- 23 -

Bruxelles Environnement, 2014. Emissions de polluants liées au trafic routier de marchandises en RBC (relativement à l’ensemble du trafic routier). Bruxelles Mobilité, 2012. Comptages poids lourds - Juin 2012. Cambridge Econometrics, 2013. Fuelling Europe’s Future. Christofides, N., Mingozzi, A., Toth, P., 1981. Exact algorithms for the vehicle routing problem, based on spanning tree and shortest path relaxations. Math. Program. 20, 255–282. CIVITAS, 2015. CASE STUDY - Freight consolidation in Bath. Clarke, G., Wright, J.W., 1964. Scheduling of vehicles from a central depot to a number of delivery points. Oper. Res. 12, 568–581. C-Liege, 2012. Good Practices Database (Excel-ZIP-Archive). Conrad, R.G., Figliozzi, M.A., 2011. The Recharging Vehicle Routing Problem, in: Doolen, T., Van Aken, E. (Eds.), Proceedings of the 2011 Industrial Engineering Research Conference. Crainic, T.G., Ricciardi, N., Storchi, G., 2004. Advanced freight transportation systems for congested urban areas. Transp. Res. Part C Emerg. Technol. 12, 119–137. Crowe, P., 2013. European-Specific Nissan Leaf To Be Unveiled In Geneva [WWW Document]. Hybrid Cars. URL http://www.hybridcars.com/european-specific-nissan-leaf-to-be-unveiled-in-geneva/ (accessed 5.20.14). Dablanc, L., 1997. Entre police et service: l’action publique sur le transport de marchandises en ville. Le cas des métropoles de Paris et New York. Ecole national des ponts et chaussées - ENPC Paris / Marne-la-Vallée: 2006. Dablanc, L., 2007. La notion de développement urbain durable appliquée au transport des marchandises. Les Cah. Sci. du Transp. 97–126. Dablanc, L., 2009. FREIGHT TRANSPORT FOR DEVELOPMENT TOOLKIT : Urban Freight. Dablanc, L., 2011. City distribution, a key element of the urban economy: guidelines for practitioners, in: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 13–36. Dablanc, L., Rakotonarivo, D., 2010. The impacts of logistics sprawl: How does the location of parcel transport terminals affect the energy efficiency of goods’ movements in Paris and what can we do about it? Procedia Soc. Behav. Sci. 2, 6087–6096. Dantzig, G.B., Ramser, J.H., 1959. The truck dispatching problem. Manage. Sci. 6, 80–91. den Boer, E., Aarnik, S., Kleiner, F., Pagenkopf, J., 2013. Zero emissions trucks - An overview of state-of-the-art technologies and their potential. CE Delft. 151p. De Cauwer, C., Van Mierlo, J., Coosemans, T., 2015. Energy Consumption Prediction for Electric Vehicles Based on Real-World Data. Energies 8, 8573–8593. De Vroey, L., Jahn, R., El Baghdadi, M., Van Mierlo, J., 2013. Plug-to-wheel energy balance - Results of a two years experience behind the wheel of electric vehicles, in: EVS27. Barcelona, pp. 1–5. DELIVER, 2012. Deliverable D1.1 - Report on Technology, Market and Urban Logistics Roadmap from 2020 and Beyond. Aachen, Germany. DGSIE, 2011. Transports routiers de marchandises. Dijk, M., Orsato, R.J., Kemp, R., 2013. The emergence of an electric mobility trajectory. Energy Policy 52, 135–145. EC, 2008. DIRECTIVE 2008/50/EC. EC, 2010. Europe 2020: A strategy for smart, sustainable and inclusive growth. Brussels. EC, 2011a. A Roadmap for moving to a competitive low carbon economy in 2050. EC, 2011b. White paper: Roadmap to a Single European Transport Area. Brussels. EC, 2011c. European Green Cars Initiative.

Introduction

- 24 -

EC, 2012. Measuring road congestion. EC, 2014. EU transport in figures - Statistical pocketbook. EC, 2015. Evaluation of Regulations 443/2009 and 510/2011 on CO2 emissions from light-duty vehicles - Final Report. EEA, 2013. Air quality in Europe — 2013 report. EEA, 2014. Noise in Europe 2014. EEA, 2015. European Union emission inventory report 1990–2012 under the UNECE Convention on Long-range Transboundary Air Pollution (LRTAP). El Baghdadi, M., De Vroey, L., Coosemans, T., Van Mierlo, J., Jahn, R., 2013. Electric Vehicle Performance and Consumption Evaluation, in: EVS27. Barcelona, pp. 1–8. ELCIDIS, 2002. Electric Vehicle City Distribution, Final Report. Electrification Coalition, 2010. Electrification Roadmap. Washington. E-Mobility NSR, 2013. Comparative Analysis of European Examples of Schemes for Freight Electric Vehicles Compilation report. Erdoğan, S., Miller-Hooks, E., 2012. A Green Vehicle Routing Problem. Transp. Res. Part E Logist. Transp. Rev. 48, 100–114. EU, 2001. Directive 2001/81/EC. EU, 2007. REGULATION (EC) No 715/2007. EU, 2009. REGULATION (EC) No 443/2009 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL. EU, 2011. Regulation (EU) No 510/2011 setting emission performance standards for new light commercial vehicles as part of the Union’s integrated approach to reduce CO2 emissions from light-duty vehicles. EU, 2013. Decision N° 1386/2013/EU - 7th EAP programme to 2020. EU, 2014. Regulation (EU) No 540/2014 on the sound level of motor vehicles and of replacement silencing systems, and amending Directive 2007/46/EC and repealing Directive 70/157/EEC, Official Journal of the European Union. European Investment Bank, 2011. Shaping sustainable cities. Luxembourg. EV

World, 2009. Cargohopper: Utrecht’s Electric Logistic http://evworld.com/news.cfm?newsid=21396 (accessed 10.9.15).

Mover

[WWW

Document].

URL

Feng, W., Figliozzi, M. a., 2012. Conventional vs Electric Commercial Vehicle Fleets: A Case Study of Economic and Technological Factors Affecting the Competitiveness of Electric Commercial Vehicles in the USA. Soc. Behav. Sci. 39, 702–711. FREVUE, 2013. State of the art of the electric freight vehicles implementation in city logistics. Friedrich, R., Bickel, P., 2001. Environmental External Costs of Transport. Springer, Stuttgart. Gelman, A., Carlin, J.B., Stern, H.S., Rubin, D.B., 2009. Bayesian Data Analysis, Second. ed. Chapman & Hall/CRC. Gerardin, B., Patier, D., Routhier, J.L., Segalou, E., 2000. Programme national “Marchandises en ville.” Goeke, D., Schneider, M., 2014. Routing a Mixed Fleet of Electric and Conventional Vehicles. Gonçalves, F., Cardoso, S.R., Relevas, S., Povoa, A.P.F.D.B., 2011. Optimization of a distribution network using electric vehicles: a VRP problem. Lisboa. Green, P.E., Srinivasan, V., 1978. Conjoint Analysis in Consumer Research: Issues and Outlook. J. Consum. Res. 5, 103–123. Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., 2010. Multivariate Data Analysis: a global perspective, Seventh. ed. Pearson, New Jersey. Hayes, J.G., de Oliveira, R.P.R., Vaughan, S., Egan, M.G., 2011. Simplified electric vehicle power train models and

Introduction

- 25 -

range estimation. IEEE Veh. Power Propuls. Conf. 1–5. Hiermann, G., Puchinger, J., Hartl, R.F., 2014. The Electric Fleet Size and Mix Vehicle Routing Problem with Time Windows and Recharging Stations. Submitt. to Transp. Sci. Hill, N., Brannigan, C., Smokers, R., Schroten, A., Essen, H. Van, Skinner, I., 2012. EU Transport GHG: Routes to 2050 II Developing a better understanding of the secondary impacts and key sensitivities for the decarbonisation of the EU ’ s transport sector by 2050. ICCT, 2014. European vehicle market statistics - Pocketbook 2014. IPCC, 2014. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva, Switzerland. Laporte, G., Nobert, Y., Desrochers, M., 1982. Two exact algorithms for the distance constrained vehicle routing problem. Cah. du GERAD G-82-05. Laporte, G., Nobert, Y., Desrochers, M., 1985. Optimal Routing under Capacity and Distance Restrictions. Oper. Res. 33, 1050–1073. Laporte, G., Nobert, Y., Taillefer, S., 1987. A Branch-and-Bound algorithm for the asymmetrical distanceconstrained vehicle routing problem. Math. Model. 9, 857–868. Lebeau, K., 2013. Electrifying cars : the economic potential of electric vehicles. Vrije Universiteit Brussel. Lebeau, K., Van Mierlo, J., Lebeau, P., Mairesse, O., Macharis, C., 2012. The market potential for plug-in hybrid and battery electric vehicles in Flanders: A choice-based conjoint analysis. Transp. Res. Part D Transp. Environ. 17, 592–597. Lebeau, K., Van Mierlo, J., Lebeau, P., Mairesse, O., Macharis, C., 2013. Consumer attitudes towards battery electric vehicles : a large-scale survey. Int. J. Electr. Hybrid Veh. 5, 28–41. Lin, C., Choy, K.L., Ho, G.T.S., Chung, S.H., Lam, H.Y., 2014. Survey of Green Vehicle Routing Problem: Past and future trends. Transp. Res. Procedia 3, 452–459. Lindholm, M., Browne, M., 2013. Local Authority Cooperation with Urban Freight Stakeholders: A Comparison of Partnership Approaches 13, 20–38. Lindholm, M.E., 2012. Enabling sustainable development of urban freight from a local authority perspective. Chalmers University of Technology. Lindholm, M.E., Blinge, M., 2014. Assessing knowledge and awareness of the sustainable urban freight transport among Swedish local authority policy planners. Transp. Policy 32, 124–131. Litschke, D.-I.A., Knitschky, D.-V.G., 2012. Future development in road freight transport regarding more environmentally friendly vehicle technologies. Procedia - Soc. Behav. Sci. 48, 1557–1567. Macharis, C., Melo, S., Woxenius, J., Van Lier, T., 2014. Sustainable Logistics. Emerald, Bingley, UK. Macharis, C., Van Mierlo, J., 2013. Sustainable Mobility and Logistics. VUBPress. Macharis, C., Van Mierlo, J., Van Den Bossche, P., 2007. Combining Intermodal Transport With Electric Vehicles: Towards More Sustainable Solutions. Transp. Plan. Technol. 30, 311–323. Massiani, J., 2014. Stated preference surveys for electric and alternative fuel vehicles : are we doing the right thing ? Transp. Lett. Int. J. Transp. Res. 6, 152–160. Melo, S., Costa, A., 2011. Definition of a set of indicators to evaluate the performance of urban goods distribution initiatives, in: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 120–147. Messagie, M., Boureima, F., Matheys, J., Sergeant, N., Timmermans, J.-M., Macharis, C., Van Mierlo, J., 2010. Environmental performance of a battery electric vehicle : a descriptive Life Cycle Assessment approach. World Electr. Veh. J. 4, 782–786. Messagie, M., Boureima, F.-S., Coosemans, T., Macharis, C., Van Mierlo, J., 2014. A Range-Based Vehicle Life Cycle Assessment Incorporating Variability in the Environmental Assessment of Different Vehicle Technologies and Fuels. Energies 7, 1467–1482. MIT, 2010. Electrification of the Transportation System.

Introduction

- 26 -

Mommens, K., Lebeau, P., Macharis, C., 2014. A modal shift of palletized fast moving consumer goods to the inland waterways: a viable solution for the Brussels-Capital Region?, in: Brebbia, C.A. (Ed.), Urban Transport XX. WIT Press, pp. 359–371. Moniteur belge, 2009. C- 2009/31231. Moniteur belge, 2013. C-2013/31464. Muñuzuri, J., Larrañeta, J., Onieva, L., Cortés, P., 2005. Solutions applicable by local administrations for urban logistics improvement. Cities 22, 15–28. OECD, 1999. Freight transport and the city. OECD, 2003. Delivering the Goods: 21st century challenges to urban goods transport. OECD, 2012. Electric Vehicles Revisited – Costs, Subsidies and Prospects. Omar, N., 2012. Assessment of Rechargeable Energy Storage Systems for Plug-In Hybrid Electric Vehicles. Vrije Universiteit Brussel. Orme, B.K., 2011. Personal communication on the Sawtooth software support forum [WWW Document]. URL https://sawtoothsoftware.com/forum/6206/rlh-benchmark?show=6206#q6206 Parry, I.W.H., Walls, M., Harrington, W., 2007. Automobile Externalities and Policies. J. Econ. Lit. 45, 373–399. Pelkmans, L., Debal, P., 2006. Comparison of on-road emissions with emissions measured on chassis dynamometer test cycles. Transp. Res. Part D Transp. Environ. 11, 233–241. Pelletier, S., Jabali, O., Laporte, G., 2014. Battery Electric Vehicles for Goods Distribution: A Survey of Vehicle Technology, Market Penetration, Incentives and Practices. Port de Bruxelles, 2014. Masterplan du port de Bruxelles à l’horizon 2030. Quak, H., 2011. Urban freight transport: the challenge of sustainability, in: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 37–55. Quak, H., Dekoster, M., 2007. Exploring retailers’ sensitivity to local sustainability policies. J. Oper. Manag. 25, 1103–1122. Quak, H., van Duin, J.H.R., 2010. The influence of road pricing on physical distribution in urban areas. Procedia Soc. Behav. Sci. 2, 6141–6153. Rangaraju, S., De Vroey, L., Messagie, M., Mertens, J., Van Mierlo, J., 2015. Impacts of electricity mix, charging profile, and driving behavior on the emissions performance of battery electric vehicles: A Belgian case study. Appl. Energy 148, 496–505. Regué, R., 2013. Appraising Freight Tram Schemes: A Case Study of Barcelona. Eur. J. Transp. Infrastruct. Res. 13, 56–78. Rogers, E.M., 1983. Diffusion of Innovations, Third edit. ed. Th free press, New York. Russo, F., Comi, A., 2012. City Characteristics and Urban Goods Movements: A Way to Environmental Transportation System in a Sustainable City. Procedia - Soc. Behav. Sci. 39, 61–73. Schade, W., Krail, M., 2010. iTREN-2030: Experiences and results for integrated technology, energy and transport policy assessment - Final Report. Schliwa, G., Armitage, R., Aziz, S., Evans, J., Rhoades, J., 2015. Sustainable city logistics — Making cargo cycles viable for urban freight transport. Res. Transp. Bus. Manag. 15, 50–57. Schneider, M., Stenger, A., Goeke, D., 2014. The Electric Vehicle Routing Problem with Time Windows and Recharging Stations. Transp. Sci. 48, 500–520. Schoemaker, J., Allen, J., Huschebeck, M., Monigl, J., 2006. Quantification of Urban Freight Transport Effects I. Ségalou, E., Routhier, J., de Rham, C., Albergel, A., 2006. Mise en place d’une méthodologie pour un bilan environnemental physique du transport de marchandises en ville. Lyon. Sierzchula, W., 2014. Factors influencing fleet manager adoption of electric vehicles. Transp. Res. Part D Transp. Environ. 31, 126–134. Sierzchula, W., Bakker, S., Maat, K., Van Wee, B., 2012. The competitive environment of electric vehicles: An

Introduction

- 27 -

analysis of prototype and production models. Environ. Innov. Soc. Transitions 2, 49–65. Soret, A., Guevara, M., Baldasano, J.M., 2014. The potential impacts of electric vehicles on air quality in the urban areas of Barcelona and Madrid ( Spain ). Atmos. Environ. 99, 51–63. SPF Mobilité et Transports, 2011. Recensement général de la circulation 2009 (n°52). SPF Mobilité et Transports, 2014. Parc des véhicules utilitaires. Stern, N., 2008. The Economics of Climate Change. Am. Econ. Rev. Pap. Proc. 98, 1–37. Stern, N.H., 2007. The Economics of Climate Change: The Stern Review. Cambridge University Press, Cambridge, UK. Strale, M., 2014. The Cargo Tram: Current Status and Perspectives, the Example of Brussels, in: Macharis, C., Melo, S., Woxenius, J., van Lier, T. (Eds.), Sustainable Logistics. Emerald, pp. 245–263. SUGAR, 2011. Sustainable Urban Goods Logistics Achieved by Regional and Local Policies, City. Bologna, Italy. Taniguchi, E., Kawakatsu, S., Tsuji, H., 2000. New co-operative system using electric vans for urban freight transport. Transport. Taylor, M.A.P., 2005. The City Logistics paradigm for urban freight transport 1–19. Toth, P., Vigo, D., 2002. Models, relaxations and exact approaches for the capacitated vehicle routing problem. Discret. Appl. Math. 123, 487–512. Trommer, S., Jarass, J., Kolarova, V., 2015. Early adopters of electric vehicles in Germany unveiled, in: EVS28. pp. 1–11. van Binsbergen, A., Visser, J., 1999. New urban goods distribution systems, in: Conference on Urban Transport Systems. Van Duin, J.H.R., Tavasszy, L.A., Quak, H.J., 2013. Towards E ( lectric ) - urban freight : first promising steps in the electric vehicle revolution 1–19. van Lier, T., 2014. The development of an external cost calculator framework for evaluating the sustainability of transport solutions. Vrije Universiteit Brussel. van Lier, T., Macharis, C., 2011. Transport of goods to and from the center of Brussels : using the port to improve sustainability, in: City Distribution and Urban Freight Transport : Multiple Perspectives. pp. 176–199. Van Mierlo, J., Maggetto, G., 2007. Fuel Cell or Battery: Electric Cars are the Future. Fuel Cells 7, 165–173. Van Mierlo, J., Maggetto, G., Lataire, P., 2006. Which energy source for road transport in the future? A comparison of battery, hybrid and fuel cell vehicles. Energy Convers. Manag. 47, 2748–2760. Van Mierlo, J., Maggetto, G., Vereecken, L., Favrel, V., Meyer, S., Hecq, W., 2003. How to define clean vehicles? Environmental impact rating of vehicles. Int. J. Automot. Technol. 4, 77–86. van Rooijen, T., Quak, H., 2014. City Logistics in the European CIVITAS Initiative. Procedia - Soc. Behav. Sci. 125, 312–325. van Vliet, O., Brouwer, A.S., Kuramochi, T., van den Broek, M., Faaij, A., 2011. Energy use, cost and CO2 emissions of electric cars. J. Power Sources 196, 2298–2310. Verheijen, E., Jabben, J., 2010. Effect of electric cars on traffic noise and safety, Public Health. Verlinde, S., 2015. Promising but challenging urban freight transport solutions. Vrije Universiteit Brussel. Verlinde, S., Macharis, C., 2015. Who is in favour of off-hour deliveries to Brussels supermarkets? The multi actor multi criteria analysis (MAMCA) applied to measure overall stakeholder support, in: The 9th International Conference on City Logistics. Tenerife, Canary Islands. Verlinde, S., Macharis, C., Witlox, F., 2011. How to Consolidate Urban Flows of Goods Without Setting up an Urban Consolidation Centre ?, in: City Logistics VII. Mallorca, Spain. Wellman, G.S., Vidican, C., 2008. Pilot study of a hierarchical Bayes method for utility estimation in a choice-based conjoint analysis of prescription benefit plans including medication therapy management services. Res. Social Adm. Pharm. 4, 218–30.

Introduction

- 28 -

Westbrook, M.H., 2001. The Electric Car: Development and Future of Battery, Hybrid and Fuel-Cell Cars. Society of Automotive Engineers. London. Widegren, C., 2014. Success stories of small and clean vehicle use in Gothenburg, in: BESTFACT Workshop Growth of Small Scale Solutions in Urban Freight, Green Logistics and E-Freight. Woodburn, A.G., Browne, M., Sweet, M., Allen, J., 2005. Urban freight consolidation centres 190. World Bank, 2012. Urban Development http://data.worldbank.org/topic/urban-development

Data

[WWW

Document].

2010.

URL

Zanni, A.M., Bristow, A.L., 2010. Emissions of CO2 from road freight transport in London: Trends and policies for long run reductions. Energy Policy 38, 1774–1786.

PART I

Chapter 2: Freight transport in Brussels and its impact on road traffic

- 30 -

2 Freight transport in Brussels and its impact on road traffic1 2.1 Introduction The mobility debate in Brussels is not new. The recent publication by Hubert, Lebrun, Huynen, & Dobruszkes (2013) has highlighted the key elements of this problem, essentially in the context of the mobility of people. But, this contribution hardly deals with the impact of freight transport. Earlier on, Hubert, Dobruszkes, & Macharis (2008) had, however, identified the logistics sector as one of the key elements of the transport policy in the Brussels-Capital Region. Freight transport also contributes to traffic and air quality problems in the capital. And this is likely to increase in the future: the Federal Planning Bureau (2012) foresees a 68% increase in tonnekilometres for freight in Belgium between 2008 and 2030, while the increase in passengerkilometres should not exceed 20%. The logistics sector therefore deserves greater attention in the mobility debate. The aim of this chapter is to provide an overview of the significance of freight transport in Brussels based on the different existing sources of information, and to present the different solutions developed by the Region to tackle these issues.

2.2. Goods transported by road in Brussels: data and observations The volume of goods and modes of transport Freight transport is essential to the economic vitality of cities (Allen et al., 2000). It provides consumer goods to inhabitants and supports urban economic activity. Dablanc (2009) estimates that urban logistics generates approximately one delivery per job per week, and 30 to 50 tonnes of goods per inhabitant per year. These estimates were confirmed for Brussels by the study 1

This chapter is based on the following paper: Lebeau, P. and Macharis, C. 2014. “Freight transport in Brussels and its impact on road traffic”. Brussel Studies (80).


- 31 -

conducted by STRATEC (2002), which mentions volumes of just over 40 million tonnes for a population of just under one million inhabitants at the time (IBSA, 2014). The resulting for Brussels is 41.5 tonnes of goods per inhabitant per year. However, the existing data regarding freight transport in Brussels indicate much lower quantities. Figure 9 illustrates the evolution of freight flows (in tonnage) per mode of transport. Rail is represented only partially given the difficulty in obtaining data on this sector. The total volume of goods transported in 2002 in Brussels is about 18 million tonnes according to the available data, whereas the STRATEC (2002) report mentioned almost 40 million tonnes. In reality, the data presented in Figure 9 greatly underestimate road transport. The Direction générale Statistique et Information économique (DGSIE), which gathers these data each year, restricts its observations to vehicles whose load capacity equals one tonne or more. Though, it is estimated that in European cities, the volume transported by vans represents more than half of the total volume transported (PORTAL, 2003). Given that 80% of vans have a load capacity of less than one tonne (SPF Mobilité et Transports, 2011), this flow is not included in the DGSIE data, thus explaining the low volume of goods recorded. Figure 9: Evolution of freight transport in the Brussels-Capital Region 40,000

Number of thousands tonnes

35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 1994

1996 Train

1998

2000

Barge

2002

2004

Road (> 1 ton)

2006

2008

2010

2012

2014

Road theoric (< 1 ton)

Source: VUB-MOBI calculations based on Dablanc (2009), DGSIE (2011), Port de Bruxelles (2014)

This widening gap in the observations illustrates the tendency towards a fragmentation of the flow of goods. The volume transported by vehicles with a load capacity of more than one tonne is decreasing while the population is increasing (and water and rail transport do not explain this drop), which is explained by the fact that an increasing proportion of these goods is transferred to vehicles whose load capacity is less than one tonne, i.e. vans. As they have a lower transport capacity, the number of freight vehicles has increased tenfold. Zunder (2011) notes that the


- 32 -

quantity of deliveries per person in French cities is increasing, while the volume per person remains constant. This trend may be explained in part by the structural change in urban economies, which are evolving towards services to the detriment of industry. This trend is also seen in the Brussels-Capital Region. Service activities make use of more light goods vehicles than industry activities do. Consequently, the Brussels-Capital Region should expect a rise in freight transport in its territory. On the one hand, the population boom expected in the Brussels Region should cause an increase in the volume of transported goods which is proportionate to the population, i.e. a 28% increase between 2010 and 2050 (Federal Planning Bureau, 2013). On the other hand, the sector is likely to make more use of light goods vehicles. The combined effects of these two trends greatly increase the impact of logistics due to an increase in freight vehicle traffic. This is why, despite a 28% increase in volume between 2010 and 2050, the Region foresees an 80% increase in freight vehicle traffic by 2050 (Bruxelles mobilité, 2013). The impact of freight vehicles on traffic congestion The increase in the number of freight vehicles on the roads is problematic with respect to the traffic situation in Brussels. Several rankings have positioned Brussels among the most congested cities in the world (Inrix, 2014; Tomtom, 2012). Network saturation has indeed increased. During working days outside the school holidays, the SPF Mobilité et Transports (2011) has observed that the number of kilometres of motorways with more than 75% saturation (i.e. 1,500 vehicles per hour per lane) has risen from 178km in 1990 to 735km in 2009. This structural congestion is above all concentrated around Brussels, as shown in Figure 10. Figure 10: Sections of motorway with structural congestion in 2009

Morning peak hour

(traffic direction from left to right)

Evening peak hour (traffic direction from left to right)

Morning peak hour (traffic direction from right to left)

Evening peak hour (traffic direction from right to left)

Source: SPF Mobilité et Transports (2011)

Counting efforts carried out at the borders of the Brusels-Capital Region in June 2012 have allowed an evaluation of the impact of freight transport on traffic (Bruxelles mobilité, 2012b). Let us note that buses and coaches have been counted as lorries. According to the distribution of vehicles in Belgium, lorries actually represent 85% of this category (SPF Mobilité et Transports, 2014).


- 33 -

The intensity of freight traffic varies above all according to the day of the week. Figure 11 shows the evolution in the number of vehicles entering or leaving the Region during a week in June (excluding motorways). The difference in activity between weekday and weekend is clear: at weekends, vans represent only 4 to 5% of traffic, and heavy goods vehicles, 1 to 2%, whereas during the week, the share of vans reaches 8%, and that of heavy goods vehicles, 5 to 6%. Figure 11 also shows that Tuesday is the busiest day of the week in terms of traffic. It is often recognised as being the most critical day (Stern, 2004). The counting analysis was therefore focused on this day of the week. By adding the countings achieved on the motorways the following Tuesday, we could observe a volume of approximately 45,000 freight vehicles entering and 45,000 leaving the capital on a Tuesday. Figure 11: Number of freight vehicles entering and leaving Brussels according to the day of the week, excluding motorways 35,000

250,000

200,000 25,000 150,000

20,000 15,000

100,000

Number of cars

Number of vans and lorries

30,000

10,000 50,000

5,000 0

0 MO TU WE TH

FR

SA

SU

IN Vans

Lorries (buses and coaches included)

MO TU WE TH

FR

SA

SU

OUT Lorries with trailer

Articulated lorries

Cars

Source: VUB-MOBI calculations based on data from Bruxelles Mobilité (2012)

The share of freight vehicles in traffic also varies according to the time of day. Figure 12 summarises the distribution per hour of all vehicles entering and leaving the Region. Freight traffic differs from car traffic in that its distribution is less concentrated during peak hours. The highest peak for freight vehicles entering the Region is at 6am, when traffic reaches its busiest level immediately. Traffic then decreases gradually throughout the day. Traffic leaving the Region is more balanced and remains stable between 6am and 4pm (approximately 3,000-3,500 vehicles per hour). Figure 12 shows that overall, the logistics sector is not greatly affected by the level of car traffic. Thus, the greatest conflict between passenger transport and freight transport occurs during the morning peak. This is due to the fact that deliveries are prepared during the night and must be delivered in the morning, before the first customers arrive. Finally, the level of freight traffic varies according to the location. The counting efforts have shown that the Region's main entrances and exits for freight vehicles are Boulevard Industriel in the


- 34 -

south of the Region, the A12 in the northeast and the A3 in the northwest. Figure 13 the importance of freight transport at different locations in Brussels on a Tuesday. Some of them, however, are used more by freight vehicles than by cars. For example, Chaussée de Vilvorde and Avenue de Tyras represent 48% and 32% of incoming freight traffic and 44% and 31% of outgoing freight traffic respectively. This particularly high proportion is explained firstly by the fact that these roads are less appealing to cars and, secondly, by the many heavy goods vehicles which use them (close to 20%). Figure 12: Number of freight vehicles entering and leaving Brussels according to the time of the day, excluding motorways 4,500

25,000 20,000

3,500 3,000

15,000

2,500 2,000

10,000

1,500 1,000

Number of cars


4,000

5,000

IN

04h 05h 06h 07h 08h 09h 10h 11h 12h 13h 14h 15h 16h 17h 18h 19h 20h 21h

0

04h 05h 06h 07h 08h 09h 10h 11h 12h 13h 14h 15h 16h 17h 18h 19h 20h 21h

500 0

OUT

Vans


Lorries with trailer

Articulated lorries

Cars Source: VUB-MOBI calculations based on data from Bruxelles Mobilité (2012)

Figure 11, Figure 12 and Figure 13 allow us to identify the organisation of freight transport traffic in Brussels. Yet in order to evaluate its impact on road traffic, it is necessary to consider this traffic in its context, with the available road infrastructure and car traffic. The method2 used by the SPF Mobilité et Transports (2011) allows this information to be integrated in a traffic index. By comparing the proportion of freight traffic with the level of congestion in Figure 14, it is possible to identify the roads where freight traffic has the most impact on traffic jams during peak hours. For example, Figure 14 shows that traffic entering via Chaussée de Vilvorde is made up of a significant share of freight vehicles, but their impact on congestion is limited given the low score on the traffic index. Conversely, Boulevard Industriel as an entrance into the capital has the highest score on the traffic index for the morning peak, partly because of the limited number of lanes compared to the flow of vehicles. Given that freight vehicles represent 17% of traffic, Boulevard Industriel may be considered as the road where freight transport contributes most to the congestion of incoming traffic in the morning. 2

The traffic index is defined as the relationship between the number of vehicles observed and the conventional limit of 2000 vehicles per hour and per lane.


- 35 -

IN

OUT

Figure 13: Ranking of main entrances and exits in the Brussels-Capital Region for freight vehicles, motorways included Bld industriel A12 A3 Ch Ninove Av Tyras E411 A10 Ch Mons Av Vilvorde Rue de Stalle Bld Woluwe Av Exposition A201 Ch Vilvorde Av Strooper Bld Dupuis Av Tervueren Ch Louvain Ch Haecht Bld 2nd armée b Rue Chateau d or Ch Waterloo Ch La Hulpe Av. Vandervelde Bld Careme Bld industriel A3 A12 A10 E411 Ch Ninove Av Tyras Av Vilvorde Ch Mons Ch Vilvorde Bld Woluwe Rue de Stalle Av Strooper Bld Dupuis A201 Ch Haecht Bld 2nd armée b Av Tervueren Rue Chateau d or Av Exposition Ch Louvain Av. Vandervelde Ch Waterloo Ch La Hulpe Bld Careme

0

1,000

2,000

3,000

4,000

5,000

6,000

Number of vans and lorries Vans



Articulated lorries

Source: VUB-MOBI calculations based on data from Bruxelles Mobilité (2012)

Let us note an alternative approach to this graph, which allows us to identify the roads where freight vehicles are affected most by congestion due to cars. For example, the traffic index for Avenue de Tervuren in the morning ranks fourth, with the proportion of cars reaching 94%. In this case, congestion due to cars has an impact on transport companies in particular. The time and fuel wasted are therefore reflected in transport costs. Unfortunately, the counting efforts were not able to integrate motorways in the analysis, given that cars were not included. Yet these roads are probably the most problematic, as seen in Figure 10.


- 36 -

Ch Vilvorde Av Tyras Ch La Hulpe Av Vilvorde Av Strooper Bld Woluwe Bld Dupuis Ch Haecht Ch Mons Av. Vandervelde Ch Louvain Ch Ninove Rue Chateau d or Bld 2nd armée b Bld Careme Ch Waterloo Av Tervueren Rue de Stalle Bld industriel Av Exposition

IN 7h-10h

80% 70% 60% 50% 40% 30% 20% 10% 0%

Congestion index

50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0%

Ch Vilvorde Av Tyras Ch Louvain Ch La Hulpe Bld Dupuis Av Vilvorde Ch Waterloo Rue de Stalle Av. Vandervelde Bld Careme Ch Ninove Ch Mons Ch Haecht Bld Woluwe Av Strooper Av Tervueren Bld 2nd armée b Rue Chateau d or Av Exposition Bld industriel

Share of vans and lorries in the traffic

Figure 14: Impact of freight vehicles on traffic during weekday peak hours, excluding motorways

OUT 16h-19h

Vans



Articulated lorries

Congestion index Source: VUB-MOBI calculcations based on Bruxelles Mobilité (2012)

Figure 15: Evolution in the number of vans and lorries in Belgium

900,000

6,000,000

800,000 5,000,000

600,000 500,000

3,000,000

400,000 300,000

2,000,000

200,000 1,000,000

100,000

Articulated lorries

Lorries

Vans

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

0 2001

0

Cars

Source: VUB-MOBI calculations based on SPF Mobilité et Transports (2014)

Number of cars

4,000,000

2000


700,000


- 37 -

The rise in the number of vans As there have not been any counting efforts using the same methodology as those in June 2012, it is impossible to see how freight traffic has evolved over time. But in the first section we have identified an increase in the use of vans. This trend may also be observed through the evolution in the number of freight vehicles in Belgium. Figure 15 shows that there has been a higher increase in the use of vans than heavy goods vehicles or even cars: between 1997 and 2011, the number of vans rose by 6.5% per year on average, while the number of lorries and articulated lorries remained stable. The growth of cars was limited to 1.5%. Hence, the number of freight vehicles is rising more quickly than the number of cars. It is, however, necessary to remain cautious when analysing the numbers of vehicles, given that cars are also used for freight transport and conversely, vans are used for purposes other than freight transport. Nevertheless, these observations are in line with the conclusions of the Federal Planning Bureau, which forecasts a greater increase of freight traffic than for passenger traffic (Federal Planning Bureau, 2012). This also confirms that an increasingly significant proportion of freight vehicles are not included in the observations of the DGSIE national survey. Among the motorised freight transport vehicles in Belgium, in 2013, Brussels represented 9.5% of vehicles, compared to 62.2% in Flanders and 28.3% in Wallonia (SPF Mobilité et Transports, 2014). For Brussels, this represents a total of 74,562 motorised vehicles. Figure 16 presents a comparison of these vehicles in Brussels and those in Flanders and Wallonia. As thecategory of vans represents 81% of all motorised freight transport vehicles in the country, it has been divided into sub-categories. The distribution is similar in the Regions, although the Brussels-Capital Region does have a smaller share of articulated lorries, and instead, more vans. This profile may be explained by the purely urban aspect of the Region and more distribution activities. Regulations also encourage the use of light goods vehicles, such as the ban on tunnel access for heavy goods vehicles due to fire safety reasons. Figure 16: Distribution of vans and lorries in each Region according to their category 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

3% 13% 2% 27%

7% 13% 1%

5% 12% 2%

27%

24%

Articulated lorries Lorries Vans (unspecified)

34%

32%

33%

Vans (GVW over 3t) Vans (GVW between 2,001 and 3t)

21%

20%

23%

Brussels

Flanders

Wallonia

Vans (GVW between 1,001 and 2t)

Source: VUB-MOBI calculations based on SPF Mobilité et Transports (2014)

A particularly polluting sector The data on air quality in Brussels are gathered by Bruxelles Environnement. The pollution caused by freight traffic in the Brussels-Capital Region is presented proportionately with respect to road traffic in general (Figure 17). Freight vehicles (vans and heavy goods vehicles together) are


- 38 -

responsible for 25% of CO2 emissions, 31% of NOx emissions, 33% of PM 2.5 emissions and 32% PM 10 emissions. When considering these figures, one must bear in mind that freight vehicles represent an average of 14% of road traffic. This parallel makes it clear that vehicles used in logistics cause much more pollution that those used for personal transport. Heavy goods vehicles cause a particularly high pollution level in terms of NOx emissions, and vans, in terms of particle and fine particle emissions (PM10 and PM2.5). Figure 17: Share of pollution caused by vans and lorries traffic in the total emissions from road transport in the Brussels-Capital Region

PM10 all Vans

PM2.5 all

Lorries NOx

Cars

Buses

CO2 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Source: VUB-MOBI calculations based on Bruxelles Environnement (2014)

The fuel used by freight vehicles explains these findings. Diesel represents a 93% market share of freight vehicles registered in 2013 in the Brussels-Capital Region. As a comparison, the use of diesel for cars in Belgium as a whole reached a historical high with 62.3%. Petrol is much less common, being used by 4% of freight vehicles in the Region. As seen in Figure 18, it is chosen mainly for vans whose maximum authorised mass (the GVW is the sum of the weight of the empty vehicle, the driver and the vehicle's load capacity) is between 1 and 2 tonnes. LPG is the most commonly used alternative fuel in the Brussels Region, representing 0.5% of the market share. This is followed by electricity, with a market share of 0.1%: 51 electric vehicles are registered in the GVW category of 2 to 3 tonnes, and 20 in the GVW category of under 1 tonne. Finally, there are 17 vehicles which run on natural gas, i.e. a market share of 0.02%. These are mainly lorries, as seen in Figure 18.


- 39 -

Figure 18: Distribution of freight vehicles registered in Brussels according to different types of technology 100% 90% 80%

Articulated lorries

70%

Lorries

60% 50%

Vans (Unspecified)

40%

Vans (GVW between 3,001 and 3,5t)

30%


20%


10%

Vans (GVW lower or equal to 1t)

0% Petrol

Diesel

GPL

Electricity Natural gas

Others

Source: VUB-MOBI based on data from SPF Mobilité et Transports (2014)

A logistics area beyond the regional borders The data from the national survey on goods transported by road (DGSIE, 2011) show that the logistics areas which the Brussels economy depends on is spread beyond its borders. Of the goods loaded in the Brussels-Capital Region, 26% remain in the Region whereas 44% are sent to Flanders, 29% to Wallonia and 2% to another country. The geographic distribution of the origins of deliveries to the Brussels-Capital Region is similar: 23% originate in the Region, 47% in Flanders, 28% in Wallonia and 2% in another country. Besides, we can stress that Brabant and the Brussels-Capital Region together receive 61% of goods loaded in the Region and send 61% of goods delivered to the Brussels Region. This economic integration of the Brussels-Capital Region with its outskirts should continue to strengthen. Strale (2013) has shown that there is a trend in the relocation of urban logistics activities to the outskirts of Brussels. This phenomenon is also observed in Paris. Dablanc & Rakotonarivo (2010) have presented their consequences. This trend has the effect of increasing the distances travelled, increasing freight traffic in the road network and therefore increasing the emissions caused by road transport. But this growing dependency of the Brussels Region on its outskirts stresses especially the need for cooperation between the three Regions in the country. The problems discussed in this section of the chapter cannot be resolved effectively by the unilateral decisions made by the different Regions.

2.3. Solutions for Brussels The various issues involved with goods transported by road converge and lead to an increase in freight traffic: population growth in Brussels increases the volumes to be delivered, the fragmentation of these volumes increases the number of vehicles per delivery and the relocation of logistics activities to the outskirts increases the distances to be covered. The imbalance between supply and demand in road infrastructure risks therefore to become more and more problematic. The current excessive demand causes congestion and damages the air quality in Brussels. There is, however, a range of solutions. The objective of this section is to present some


- 40 -

of the solutions to the problems identified in the first part of this article. They are aimed at decreasing this excessive demand through more efficient transport of goods. The possibility of increasing the road infrastructure is not considered a priority in this section, as public space must also serve purposes other than traffic (Hubert et al., 2013). Modal shift Given the increasing relocation of logistics activities to the outskirts, a first solution would consist in sending goods to the Brussels-Capital Region via train or waterway. Brussels has indeed the advantage of benefiting from highly developed multimodal accessibility. Railway is a possible alternative to roads. The Region has 163 kilometres of railways, thus making Brussels one of the regions with the densest network in Europe (BECI, 2010). Furthermore, the star configuration of the national network centred round Brussels offers access to many different destinations. However, Figure 9 shows the gradual disappearance of this mode in Brussels. Earlier in 2014, when the grain transporter CERES stopped using the railway, the Audi factory was the only remaining company to use it. Poor service is responsible for this: the trains are often late and are not always available (STRATEC, 2002). The network is dedicated as a priority to passenger transport. Hence, freight transport via rail is indeed difficult given the current saturation of the railway network in Brussels. The future of railway transport lies in the project at Schaerbeek Formation, where a trimodal platform is planned. Currently, the modal shift to barge transport is the most interesting one in the Brussels-Capital Region, thanks to the canal which links the Port of Brussels to Antwerp in five hours. Van Lier & Macharis (2011) have estimated that the use of the waterway in 2007 decreased the number of lorries in the city by 255,000, thus making the port an essential stakeholder in sustainable development in Brussels. For the moment, the logistics activity in the port is quite specialised. It involves above all the import of construction materials, petroleum products and agricultural products in the north section of the Brussels canal. In particular, the port plans to reinforce its position in the transport of construction materials via a 'construction village'. This platform will allow bigger quantities to be brought together, which is necessary in order to justify the cost of transport via waterway. But the port is also diversifying its activities. A container terminal has allowed a new type of traffic in the port since 2003. The port's new master plan includes further diversification of activities (Port of Brussels, 2014). A first project is aimed at organising ro-ro traffic (roll-on/rolloff) via waterway in 2017. There is a considerable second hand vehicle market in Brussels (100,000 cars per year). These vehicles are gathered here from across Europe before they are sent to Africa from Antwerp (Rosenfeld, 2009). The project plans to gather the different stakeholders in the outer port in order to avoid lorry traffic, which has a particular impact on the neighbourhood around Delacroix underground station. A second project in the master plan consists in creating a network of transshipment points for urban distribution. In this way, the port will provide a transport solution for most urban goods, namely parcels and pallets. Deliveries with staggered hours An alternative type of modal shift involves a change in timing. This solution has received much attention but is still barely used. By authorising deliveries before 6am, freight vehicles may handle their morning distribution earlier when the road network is quieter. Freight vehicles would no


- 41 -

longer be in traffic jams and would therefore not waste time and fuel, and would create less emissions. But the concept is difficult to implement (Holguin-veras and Polimeni, 2006). The main obstacle is in the Brussels-Capital Region government decree of 21 November 2002 regarding the fight against noise and vibrations caused by classified installations. To allow these types of delivery, transporters must adapt their delivery operations in order to limit noise pollution for local residents. On the other hand, authorities must adapt the current regulations. In order for the different stakeholders to be able to evaluate the feasibility of this concept, a pilot test was conducted in Brussels with the distributors Colruyt and Delhaize to test deliveries with staggered hours. The results are not available yet, but they will be published in the report by the European Straightsol project around the end of 2014. An optimisation of road traffic The analysis of freight transport in Brussels has shown that there is an increase in the use of light goods vehicles. They mainly have the advantage of providing a higher level of service thanks to more regular deliveries. This phenomenon is probably more common in an urban environment given the cost of real estate: as the space reserved for stock must be minimised, the need for frequency is greater in order to ensure the availability of goods. However, it is possible to ensure the same frequency while reducing the number of vehicles required for transport thanks to the consolidation of goods. Transport may be optimised by grouping the loads for several vans into a single vehicle with greater capacity. There are many different types of consolidation, as explained by Verlinde, Macharis and Witlox (2011). TNT tested the use of a Mobile Depot in Brussels: an articulated lorry was sent from the depot in the outskirts to a parking area in the city, with the remaining kilometres covered by cargo bicycles. Thus, a constant flow of vans between the city and the outskirts was avoided. The test allowed a 24% decrease in CO2 emissions and up to a 78% decrease in NOx emissions, while maintaining a comparable level of service (Verlinde et al., 2014). The costs are higher, however, but may still be optimised. A second way to consolidate goods is to set up an urban consolidation centre, i.e. a logistics platform located in the city, allowing the deliveries of different transporters to be grouped. The new Regional Sustainable Development Plan (PRDD) for Brussels plans to establish several of these centres throughout Brussels, but the most likely site would be Schaerbeek Formation. In the meantime, an urban consolidation centre pilot test will be implemented in 2014 at the TIR centre along the Vergote dock ('LaMiLo' 2014). Finally, other measures will encourage transporters to consolidate better and even to collaborate. The introduction of a toll is an example: given that the price per kilometre is higher, the grouping of goods becomes more worthwhile financially. This type of measure is planned for 2016 in Brussels but it will only cover vehicles over 3.5 tonnes. As long as this toll for heavy goods vehicles is not accompanied by a toll for light commercial vehicles, the opposite effect may occur, namely that there may be a rise in the use of vans because lorries will become more expensive. The electrification of urban logistics New technologies may also provide an answer to the problem of freight transport, in particular regarding air quality and CO2 objectives. Urban logistics must be entirely 'decarbonised' by 2030 in major urban centres (EC, 2011). The electric solution seems particularly interesting in this respect as it does not cause any emissions in an urban environment. Besides, electric vehicles are


- 42 -

often appreciated in the context of distribution operations due to their performance with respect to acceleration, comfortable driving and silent engines (Dasburg and Schoemaker, 2006; SUGAR, 2011). Finally, certain people value the greater availability of these vehicles because they require less maintenance and their batteries may be charged at the depot. However, several obstacles restrain the development of electric vehicles. The high purchase cost is the main obstacle (Van Amburg and Pitkanen, 2012). But this is offset by lower operational costs: electric energy is cheaper, less maintenance is required and insurance premiums are lower. Furthermore, the Brussels-Capital Region and the federal government offer incentives for the use of electric vehicles in logistics operations (Fisconetplus, 2012; Moniteur Belge, 2009). Hence, electric vehicles which have a maximum weight of 2.3 tonnes may be more competitive than diesel vehicles (Lebeau et al., 2013). Certain express transporters are therefore developing green distribution solutions. In Brussels, for example, Ecopostale delivers parcels using electric vehicles and cargo bicycles. Delivery areas Transporters are confronted regularly with a lack of delivery areas, which force them to double park (AVCB, 2003). The increase in delivery frequencies may worsen this phenomenon, which disrupts the flow of traffic. STRATEC (1998) estimated that by reinforcing the repressive measures against illegal parking in Brussels (in particular at major crossroads in the regional network), the average speed of vehicles would increase by 30%, and the distances covered as well as fuel consumption would decrease by 2% and 12% respectively. Ensuring enough space for loading and unloading operations is therefore a priority in order to improve freight transport in Brussels. Certain municipalities have already taken the initiative by implementing a new category of parking space for deliveries (AVCB, 2010): yellow areas. One must pay to use these parking spaces, except for delivery operations. They are clearly marked and and kept under strict control by specific agents. They have been integrated into the regional parking policy, ensuring the future development of this solution in other municipalities of Brussels (Bruxelles mobilité, 2013). Furthermore, the Brussels-Capital Region has published a guide to help municipalities with the development of areas for street deliveries (Bruxelles mobilité, 2012a).

2.4 Conclusion Traffic jams and mediocre air quality are well-known problems in Brussels. But the contribution of freight transport to these problems is less known. This chapter has placed logistics in Brussels at the heart of the road traffic issue. We have underlined the fact that today, freight vehicles represent 14% of traffic entering and leaving the Region, a majority of which are vans, whose share is increasing while the one of lorries and articulated lorries is decreasing. Although heavy goods vehicles may create more problems in traffic, the preference for vans means more freight vehicles on the roads. Combined with a growing demand for goods and an increase in distances covered, the evolution in the freight vehicles used will contribute more and more to road congestion.


- 43 -

Freight transport is now becoming an issue integrated in the mobility debate. Data are still lacking, however, in order to have a better influence on this sector. The problem of freight transport using vans must be better understood in particular. Bruxelles mobilité is aware of the challenges of urban logistics. A freight transport plan was adopted by the Region in 2013 to answer those challenges. Several projects to consolidate flows are planned and a modal shift as well as the use of cleaner vehicles are encouraged. New parking areas are also being tested to solve the crucial problem of delivery operations. However, other solutions still face regulatory obstacles, such as off-peak deliveries, while tests undergoing evaluation will allow the compatibility of this logistics system with the Brussels environment to be determined. The development of these different types of solution – none of which, however, focus on the demand, strictly speaking – witnesses the positive dynamics in the Brussels-Capital Region. But there are many steps to take before a global implementation is possible, starting with interregional cooperation on the subject.


- 44 -

References AIURA, Taniguchi, 2006. Planning On-Street Loading-Unloading Spaces Considering the Behaviour of PickupDelivery Vehicles and Parking Enforcement. In: Taniguchi, E., Thompson, R. (Eds.), 4th International Conference on City Logistics. Langkawi, Malaysia: Elsevier. pp. 107–116. ALLEN, J., ANDERSON, S., BROWNE, M., JONES, P., 2000. A framework for considering policies to encourage sustainable urban freight traffic and goods / service flows; A Research Project Funded by the EPSRC as part of the Sustainable Cities Programme, Transport Studies Group. London. VAN AMBURG, B., PITKANEN, W., 2012. Best Fleet Uses , Key Challenges and the Early Business Case for ETrucks : Findings and Recommendations of the E-Truck Task Force. In: EVS26. Conference. Los Angeles, California, May 6-9. AVCB, 2003. Opération pilote Goulet Louise. Bruxelles. AVCB, 2010. Etat de la question en ce qui concerne le stationnement et la politique en matière de stationnement en Région de Bruxelles-Capitale: Enquête 2009. Bruxelles. BECI, 2010. Elements d’analyse concernant le péage urbain. Bruxelles. BRUXELLES ENVIRONEMENT, 2014. Emissions de polluants liées au trafic routier de marchandises en RBC (relativement à l’ensemble du trafic routier). Département « Planification air, énergie et climat ». BRUXELLES MOBILITE, 2012a. Organisation des livraisons en voirie. Bruxelles. BRUXELLES MOBILITE, 2012b. Comptages poids lourds - Juin 2012. Bruxelles. BRUXELLES MOBILITE, 2013. Plan régional de politique du stationement. Brussels. BRUXELLES MOBILITE, 2014. Enquète sur 3000 entreprises bruxelloise (projet Lamilo). Bruxelles. BUREAU FEDERAL DU PLAN, 2012. Destination 2030 : autoroute du chaos ou itinéraires alternatifs ? Bruxelles. BUREAU FEDERAL DU PLAN, 2013. Perspectives de population 2012-2060. Bruxelles. DABLANC, L., 2009. Freight transport for development - toolkit. Washington. DABLANC, L., Rakotonarivo, D., 2010. The impacts of logistics sprawl: How does the location of parcel transport terminals affect the energy efficiency of goods’ movements in Paris and what can we do about it? In: Procedia Social and Behavioral Sciences. Vol. 2, pp. 6087–6096. DASBURG, N., SCHOEMAKER, J., 2006. Quantification of Urban Freight Transport Effects II. DGSIE, 2011. Statsitiques sur le transports routiers de marchandises. Bruxelles. EC, 2011. White paper: Roadmap to a Single European Transport Area. Brussels. FISCONETPLUS, 2012. Codes des impôts sur les revenus 1992 - exercice d’imposition 2013 (revenus 2012). In : http://ccff02.minfin.fgov.be. [Consulté le 22 avril 2014]. Disponible à l’adresse: http://ccff02.minfin.fgov.be/KMWeb/document.do?method=view&nav=1&id=bdf4e90d-fbb7-4216-a99fdb12efe9a34c&disableHighlightning=true#findHighlighted. GETIS, A., ORD, J. K., 1992. The Analysis of Spatial Association. In: Geographical Analysis, Volume 24, Numéro 3, pp. 189–206.


- 45 -

HOLGUIN-VERAS, J., POLIMENI, J., 2006. Potential for off-peak freight deliveries to congested urban areas. New-York. HUBERT, M., DOBRUSZKES, F., MACHARIS, C., 2008. La mobilité à, de, vers et autour de Bruxelles. In : Brussels Studies. Note de synthèse n°71, pp. 1-14. HUBERT, M., LEBRUN, K., HUYNEN, P., DOBRUSZKES, F., 2013. La mobilité quotidienne à Bruxelles: défis, outils et chantiers prioritaires. In : Brussels Studies. Numéro 71, pp. 1-28. IBSA, 2014. Evolution annuelle de la population. In : http://www.ibsa.irisnet.be/. [Consulté le 22 avril 2014]. Disponible à l’adresse : http://www.ibsa.irisnet.be/themes/population#.U1Zbuvl_vz4. INRIX, 2014. Inrix index. In: http://www.inrix.com/. [Consulté le 22 avril 2014]. Disponible à l’adresse: http://www.inrix.com/scorecard/. LAMILO, 2014. Lamilo project – Sustainable city logistics – Last Mile Logistics. In : http://www.lamiloproject.eu . [Consulté le 4 juillet 2014]. Disponible à l’adresse: http://www.lamiloproject.eu/brussels/. LEBEAU, P., MACHARIS, C., VAN MIERLO, J., LEBEAU, K., 2013. Electric vehicles for logistics: a total cost of ownership analysis. In: HESSE, et al. (Eds.), Proceedings of the BIVEC-GIBET Transport Research Days 2013. Walferdange, Luxemburg-City. pp. 307–318. LEBRUN, K., HUBERT, M., HUYNEN, P., 2012. Observatoire de la Mobilité de la RBC. Bruxelles. MONITEUR BELGE, 2009. C- 2009/31231. Bruxelles. PORT DE BRUXELLES, 2010. Plan stratégique. Bruxelles. PORT DE BRUXELLES, 2013. Statistiques sur le transport fluvial et maritime de marchandises. Bruxelles. PORT DE BRUXELLES, 2014. Masterplan du port de Bruxelles à l’horizon 2030. Bruxelles. PORTAL, 2003. Inner Urban Freight Transport and city logistics. Brussels ROSENFELD, M., 2009. Le commerce d’exportation de voitures d'occasion entre Bruxelles et Cotonou. In : Cahiers de l’URMIS. Volume 12. SPF MOBILITE ET TRANSPORTS, 2011. Recensement général de la circulation 2009 (n°52). Bruxelles. SPF MOBILITE ET TRANSPORTS, 2014. Parc des véhicules utilitaires. Bruxelles. STERN, E., 2004. Spatio-temporal patterns of subjectively reported congestion in Tel Aviv metropolitan area. In: Journal of Transport Geography. Vol. 12, n°1, pp. 63–71. STRALE, M., 2013. La logistique comme outil de reconversion des ports urbains : application au port de Bruxelles. Bruxelles. STRATEC, 1998. Urban freight transport strategy in Brussels. Brussels. STRATEC, 2002. WP1: Belgium, Methods. Brussels. SUGAR, 2011. Sustainable Urban Goods Logistics Achieved by Regional and Local Policies, City. Bologna, Italy. TOMTOM, 2012. TomTom European Congestion Index. In: http://www.tomtom.com/ [consulté le 22 avril 2014]. Disponible à l’adresse : http://www.tomtom.com/lib/doc/trafficindex/20131101%20TomTomTrafficIndex2013Q2EUR-mi.pdf


- 46 -

VAN LIER, T., MACHARIS, C., 2011. Transport of goods to and from the center of Brussels : using the port to improve sustainability. In: MACHARIS, Cathy and MELO, SANDRA (ed.), City Distribution and Urban Freight Transport : Multiple Perspectives. Cheltenham: Edward Elgar. pp. 176–199. VERLINDE, S., MACHARIS, C., MILAN, L., KIN, B., 2014. Does a Mobile Depot make urban deliveries faster, more sustainable and more economically viable: results of a pilot test in Brussels. Bruxelles: VUB – MOBI. Working paper. VERLINDE, S., MACHARIS, C., WITLOX, F., 2011. How to Consolidate Urban Flows of Goods Without Setting up an Urban Consolidation Centre ? In: City Logistics VII. Mallorca, Spain. WINCKELMANS, K., (2014), E-commerce zet de toon, de vastgoedmarkt volgt. In Logistics Management, mars 2014. ZUNDER, T., 2011. Urban Freight: Myths, Good Practices, Best Practices. In European Freight Conference. Conférence. Newcastle. 22 février 2011

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 47 -

3 Exploring the choice of battery electric vehicles in city logistics: a conjoint based choice analysis3 3.1 Introduction City logistics faces a major challenge. According the European Commission (EC, 2011), freight transport should become CO2-free in major urban centres by 2030. However, it is currently responsible for 25% of CO2 emissions in cities (Dablanc, 2011), and forecasts project an increasing number of freight vehicles in city traffic (Zanni and Bristow, 2010). In that context, battery electric vehicles (BEVs) could represent an interesting solution. If electricity is generated from renewable energy sources, BEVs’ operations could contribute to decarbonising city logistics (Browne et al., 2011). But, electrifying city logistics brings also benefits at a more local level. Because the electric motor does not emit exhaust gas emissions, BEVs can contribute to a better air quality in cities (Soret et al., 2014). This solution has therefore attracted the attention from local authorities (Litschke and Knitschky, 2012). That notwithstanding, the perceptions of freight transport companies regarding BEVs seem to be less enthusiastic than that of authorities. Sales of electric freight vehicles remains indeed limited. However, city logistics is considered a suitable environment for BEVs (Lebeau, 2013; Van Mierlo and Maggetto, 2007). Greater attention must therefore be devoted to transport companies, particularly with respect to how they value BEVs.

3

This chapter is based on the following paper: Lebeau, P., Macharis, C. and Van Mierlo, J. 2015. “Exploring the choice of battery electric vehicles in city logistics: a conjoint-based choice analysis”. Transportation Research PART E (submitted).

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 48 Conjoint-based choice analysis offers an interesting approach in this context. It is a stated preference technique that simulates a choice situation involving a set of competing alternatives. By observing the alternatives that the respondents prefer, it estimates the trade-offs that respondents make among various attributes. Conjoint-based choice is a traditional approach to investigating the choice behaviour of respondents in the context of BEVs (Lebeau et al., 2012). However, analyses have been limited to the passenger car segment. This chapter will therefore extend these analyses to the preferences of urban transport operators for electric freight vehicles. That is a significant contribution given the state of the art in city logistics. As next section will show, the literature is mostly limited to identifying the barriers and drivers of BEVs. Amburg and Pitkanen (2012) have provided some additional insights through a survey in the logistics industry where respondents were asked to rate the importance of the different BEVs attributes. But we still find a gap in understanding how these attributes interact between each other and how they influence the final choice. The objective of this chapter is therefore to assess the trade-offs that urban freight transporters make between BEVs’ attributes. Based on their preference structure, we will highlight the most effective policies that can stimulate their adoption. This chapter will first present the main factors that influence the value of BEVs based on a review of various experiences using BEVs in logistics. That review will be used in the subsequent section to select the most relevant attributes to consider in the design of the conjoint-based choice analysis. Their influence on choice behaviour is assessed through a survey conducted among freight transport operators in Brussels, and the results are presented in section 4. The respondents’ attitudes regarding these attributes are estimated in the subsequent section 5. A discussion will compare the results of section 4 and section 5 with the findings of the literature review in order to identify the most effective policies that can stimulate the introduction of BEVs into city logistics.

3.2 Drivers and barriers to BEV adoption Given the interest expressed by authorities, numerous research projects have been conducted in Europe to explore the feasibility of BEVs in city logistics. They include projects such as EVDPOST (2001), ELCIDIS (2002), ENCLOSE (2014), STRAIGHTSOL (2015) and the current FREVUE (2013). Based on the literature and on the lessons learnt from the various demonstrations, this section reviews the different drivers and barriers experienced in the context of BEVs in city logistics. Compatibility with the logistics The limited range of BEVs is often considered the most important barrier to BEV adoption. That perception might be reinforced by the ranges advertised by manufacturers that are often not achieved under everyday conditions (FREVUE, 2013). Ranges are indeed usually assessed based on the NEDC which assumes a more energy efficient driving cycle than in real conditions (Pelkmans and Debal, 2006). Still, the limited range of BEVs can be compatible with specific applications in city logistics. Indeed, the last mile in urban areas is characterised by small distances: it has been estimated that more than 80% of freight trips in European cities are shorter than 80 km, which is compatible with the limited range of BEVs (BESTUFS II, 2008a). That constraint is then no longer perceived as a barrier. This can be the case in a hub-and-spoke structure such as an intermodal chain in which BEVs can operate the last miles (Macharis et al., 2007). Urban consolidation centres are also commonly identified as a suitable logistics concept in

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 49 which BEVs can achieve city-centre deliveries (ELCIDIS, 2002). Other experiments, however, have been using BEVs in more traditional concepts of urban distribution (E-Mobility NSR, 2013). These various experiments often share a common characteristic: BEVs depend on a home base at which they can be recharged during their inactive period (usually at night). In that logistics environment, the lack of a public charging infrastructure and the long charging time become less important (ELCIDIS, 2002). The possibility of charging at “home” is even considered a benefit. The operator saves time and fuel by not driving to the pump to refuel a conventional vehicle. Moreover, BEV daily range can be increased by charging during the loading/unloading period at the depot (E-Mobility NSR, 2013). Good performance In a test of electric light commercial vehicles (LCVs) in Osaka, 73% of users reported that the vehicles’ performance was the same or better than that of conventional vehicles (Taniguchi et al., 2000). The review of BEV best practices conducted by Nesterova et al. (2013) also shows that feedback from private operators and demonstrators was mainly positive. Drivers are the most surprised by the benefits of BEVs. They are the most reluctant to change, but they recognise after testing that BEVs improve their working conditions and make driving more comfortable (ADEME, 2004; BESTUFS II, 2006; SUGAR, 2011). They appreciate BEVs’ good acceleration, smoothness when driving and quietness in the driver’s cabin (ELCIDIS, 2002). These aspects are particularly convenient for deliveries in urban areas (ADEME, 2004). However, some experiences indicated a lack of comfort caused by the effort required to reduce electric consumption from auxiliaries (ELCIDIS, 2002). Nevertheless, experiencing a BEV can be considered an important step towards their adoption in logistics. This observation is also true for the management staff: those who had tested BEVs were satisfied with their performance, whereas those who had not remained reluctant because their perception of BEVs was still limited to cost considerations (BESTUFS II, 2008b). The most important performance of BEVs for companies is however their low environmental impact. For transport operators, BEVs can contribute to a corporate strategy of reducing CO2 emissions (E-Mobility NSR, 2013). BEVs can also be used as a marketing tool. Experience has shown that companies highly value a positive attitude on the part of the general public and customers. In some cases, companies even received media coverage or environmental awards. These benefits can help the organisation to create a positive corporate image, which is argued to attract new customers (E-Mobility NSR, 2013). It can also provide a solution for potential customers seeking to reduce their environmental footprint. Ambiguous costs Logistics projects that show environmental benefits are however difficult to implement if they do not demonstrate in advance their ability to generate profits (Melo and Costa, 2011). High purchase cost is therefore perceived by transport companies as the main barrier to adopting BEVs (Amburg and Pitkanen, 2012). However, when analysing their financial viability, a total cost of ownership analysis would be more relevant. Indeed, BEVs also benefit from low operating costs (E-Mobility NSR, 2013). If the vehicle is used optimally, the total cost of ownership of BEVs might be more advantageous than a comparable conventional vehicle (Lebeau et al., 2015).

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 50 The total cost of ownership of BEVs primarily depends on the batteries for three major reasons. First, batteries are expensive, which mainly explains the higher purchase cost of BEVs. Second, that important cost might further affect the total cost of ownership with the replacements required by the ageing of the battery. Finally, the uncertainty regarding these batteries increases the risk of projects involving BEVs. Previous projects have experienced several battery-related breakdowns which have seriously affected the costs of the project (E-Mobility NSR, 2013). To reduce these uncertainties for BEV operators, manufacturers have already developed solutions by providing a warranty on the battery or by offering the possibility of leasing the battery. With such systems, the risk of the battery is transferred to manufacturers. However, the uncertainty concerning these batteries still affects their residual value which impacts also the total cost of ownership of BEVs. The future evolution of battery technology should improve BEVs’ reliability and competitiveness. Their costs are expected to be nearly halved between 2010 and 2020 (Nykvist & Nilsson, 2015). As a result, BEVs could become the cheapest alternative in the LCV segment beginning in 2017 (DELIVER, 2012). However, this rapid development of battery technology might restrain some adopters from purchasing BEVs today because they would prefer to wait for the next generation of BEVs (FREVUE, 2013). Nevertheless, experience has shown that companies that operate BEVs in the early stage of their introduction secure a leading position: they are considered innovators and pioneers in new technologies (E-Mobility NSR, 2013). Early market development The limited maintenance service is considered one of the most important struggles perceived by transport operators (FREVUE, 2013; SUGAR, 2011). In previous projects, some manufacturers were unable to provide rapid assistance in the event of breakdowns, resulting in vehicles being unavailable for a long period. The development of maintenance or service facilities is therefore identified to be of “the utmost importance” (ELCIDIS, 2002). The early market development phase of BEVs also implies a limited choice of electric LCVs. This is identified as an important barrier given the diversity of supply chains in city logistics (FREVUE, 2013). The portfolio of electric vehicles proposed by manufacturers should be more developed in order to better meet the payload needs of the various customers. The need for policy support Given the environmental performance of BEVs, some authorities are stimulating a shift from conventional vehicles to BEVs in logistics. Experiences in Rotterdam and Stockholm have shown that the vehicle choice of transporters can indeed be influenced towards BEVs when access restrictions for conventional vehicles are enforced or when delivery hours are extended specifically for BEVs (BESTUFS II, 2008b; ELCIDIS, 2002). In that context, night distribution can be regarded as a more extreme extension of delivery hours. It represents an important potential application of BEVs in city logistics because that technology provides a solution to the noise problem associated with off-hour deliveries (Cavar et al., 2011). Exemptions from road charges and taxes are also policies that stimulate the introduction of BEVs, as we can observe in London with the congestion charge scheme (E-Mobility NSR, 2013). However, subsidies for BEV purchases represent the most common support as shown by the review of E-Mobility NSR (2013).

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 51 Experiences in Amsterdam have also revealed the negative influence of some policies on BEV adoption because of the heavier weight of the batteries. Weight restrictions that are common in urban areas limit the access of some BEVs that would have been authorised if they had been diesel powered. Also, BEVs that exceed a gross vehicle weight of 3.5 tonnes are penalised relative to their conventional counterparts because they require a driver with a C1 license and the equipment of a tachograph. Consequently, policies related to gross vehicle weight are sometimes relaxed for BEVs, as in Amsterdam (E-Mobility NSR, 2013). Overall, transport companies regard these policies as an advantage given by the authorities to compensate for their effort to switch to clean and silent vehicles (ELCIDIS, 2002). Adapting logistics organisation is indeed another challenge for BEV adoption (FREVUE, 2013). These measures should be promoted by the authorities at a central information point, such as the website www.klimabiler.no in Norway, to improve knowledge about BEVs, which is still considered a barrier to their adoption (E-Mobility NSR, 2013).

3.3 Methodology Conjoint-Based Choice (CBC) Conjoint-Based Choice (CBC) is a methodology arising out of consumer research. It allows the modelling of consumer preferences based on multi-attribute alternatives (Green and Srinivasan, 1978). The idea of CBC is to simulate a choice in the marketplace and observe the respondents’ decisions. Various products (or services) are presented to the respondents based on a common set of attributes. Given the description of the products, the respondents have to select the option that best matches their preferences. Based on these observations, it is then possible to evaluate consumers’ trade-offs among the various attributes. The selection of attributes is therefore critical in CBC design. The literature review showed that many aspects influence the choice of BEVs. However, Hair et al. (2010) recommend a maximum of 6 attributes for an efficient CBC design. Following the recommendations of Massiani (2014), we selected daily range, charging time, environmental performance, type of vehicle, purchase cost and operating cost as the most relevant attributes to consider for modelling BEV choice behaviour. The emissions of noise, pollutants and greenhouse gases were grouped in one attribute and measured by the Ecoscore indicator, with 0 being the worst performance and 100 the best (Van Mierlo et al., 2003). The extension of delivery hours was initially considered among the relevant attributes but was ultimately not included. When validating the design with professionals from the sector, it appeared that this was a difficult attribute for respondents to understand. Others policies, such as subsidies and toll exemptions, are considered through the purchase- and operating-cost attributes. Limited charging infrastructure is also not included as a single attribute. Instead, we integrated that aspect with the charging time attribute by stating that charging time was spent at the depot. According to the literature review, charging is indeed usually performed at the depot. Finally, reliability was not considered because the literature showed that it should be a condition for purchasing BEVs, not a preference attribute. Keeping BEVs out of operation for a long time because of slow maintenance service could only affect BEVs’ reputation (ELCIDIS, 2002).

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 52 Table 2 shows the different attributes and their possible values that we call “levels”. These levels were selected such that they can reflect either the characteristics of conventional vehicles or the characteristics of BEVs. However, like attributes, levels must be limited to ensure an efficient CBC design. They should also be both actionable and communicable measures (Hair et al., 2010): communicable because every respondent must understand them in the same way to evaluate them correctly, and actionable because levels need to be put into practice for the analysis. In other words, levels must capture the desired effect to be measured as accurately as possible. Because the “maximum volume” attribute could be difficult to understand, we depicted the levels with a picture as shown in Figure 19. Table 2: Attributes and levels considered in the design of the conjoint experiment.

Purchase costs

Operating costs

Ecoscore

Maximum range with a full charge/tank

€10,000 €25,000 €50,000

€5/100km €10/100km €25/100km

40 60 90

100km 200km 500km

Refuelling or charging time at the depot 5 min 30 min 8h

Maximum volume 3 m³ 8 m³ 20 m³

Survey design To generate the various choice tasks that will be presented to the respondent, it was necessary to make several decisions concerning the CBC design. Figure 19 shows an example of a choice task included in the survey resulting from these decisions. The first decision is the number of alternatives that should be presented in a choice task. Using more alternatives provides richer trade-offs. However, a growing number of alternatives in the task choice make the design less efficient. Indeed, the unchosen products do not provide much additional information because the CBC typically observes only the preferred product. Moreover, the design of the CBC needs to avoid burdening the respondent: evaluating more alternatives in a choice task increases the effort required by the respondent. Thus, as recommend by Hair et al. (2010), we included three alternatives per choice task in the CBC. The second decision is the number of choice tasks to be presented to each respondent. According to Hoogerbrugge and van der Wagt (2006), asking more than 10-15 choice tasks does not provide additional insights into the preference structure of respondents. Because the survey includes additional questions besides the CBC experiment, we limited the number of choice tasks to 10 to avoid burdening respondents. Finally, the method for generating the choice tasks needed to be selected. To use an appropriate CBC design, two principles should be respected: orthogonality and balance (Hair et al., 2010). Orthogonality of the design means that no correlation should exist among the levels of an attribute. The balance of the design means that each level of an attribute should appear the same number of times. However, when respecting these principles, the design would never present one level twice in a choice task because we consider three alternatives and each attribute has three possible levels. In this case, it is useful to relax these rules to allow some overlap of the levels in the same choice task. Doing so allows us, for example, to generate a choice task in which two

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 53 alternatives have the same purchase cost, and hence we can better observe the effects of the other attributes on the choice behaviour. Therefore, we used the balanced overlap option provided by Sawtooth Software to generate the choice tasks. Figure 19: Screenshot of a choice task submitted to respondents

Computing the utilities Once the data are collected, different estimation methods are available to model transporters’ preferences. Traditional methods involve multiple regressions and multinomial logit models for more complicated preferences. Recently, the Hierarchical Bayes method has attracted interest and has proven more accurate in CBC experiments (Wellman and Vidican, 2008). Moreover, that method considers the heterogeneity of the population because it estimates the utilities at an individual level (Gelman et al., 2009). Hence, the utilities of our preference model were estimated using that method. In order to compare the utilities of each level across attributes, we rescaled them on a common interval scale. The difference in utilities assesses therefore the utility gap. Levels with lower utilities are less preferred than levels with higher utilities.

3.4 Results – Choice behaviour The sample The CBC analysis was based on a survey conducted among transport companies located in the Brussels-Capital Region. This region was selected to capture the preferences of companies that are active in an urban context. The geographic coverage of the Brussels-Capital Region is essentially defined by the ring road around Brussels. Among the 818 companies that are referenced by the “golden pages” in the freight transport sector of the region, we did not consider companies in the import/export subsector because they organise but do not perform the transport service. We also did not consider transport companies lacking a phone number. We contacted the remaining 427 companies by phone to obtain their participation in the survey, and 128 agreed. They received the link to the online survey through their email, and two reminders were sent. After one month, 105 questionnaires were answered. However, some of those

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 54 questionnaires were excluded from the sample for several reasons. First, they did not complete the questionnaire. Second, they were not purchasing vehicles. Third, the respondents were not involved in the purchase decision. Consequently, the sample was reduced to 45 respondents, which Hair et al. (2010) consider satisfactory for revealing basic trends in the respondents’ preference structure. The sample can be considered representative of the population of freight transporters active in city logistics. We observe that the sample is primarily composed of small firms with staffs of fewer than 10 employees (82%), which is consistent with the characteristics of city logistics. Dablanc (2011) mentions indeed that 85% of short-distance companies have fewer than 5 employees. Therefore, a majority of transporters operate small fleets: 58% of our respondents operate fewer than 5 vehicles, while ADSEI, DIV and MOBI (2015) estimate that 56% of third account freight carriers located in the Brussels-Capital Region (the city) and in its surrounding municipalities (the periphery) operate fewer than 5 vehicles. A minority of the transporters own a depot (40%). For these firms, the location of the depot is in the city (56%), in the periphery (39%) or outside the city (6%). These locations confirm the focus of respondents on urban activities. Note, however, that depots are not often used for vehicle parking. Among all of the respondents, only 27% use a depot for parking. Most of the firms park their vehicles on streets (38%). Others use private parking (29%). Among the respondents in our sample, 82% usually make the purchase decision alone, 16% make the decision based on a first selection made by colleagues and 2% do not make the decision but instead prepare an initial selection for the final decision. That a majority of respondents take the purchase decision alone is in line with the findings of Nesbitt & Sperling (2001). They estimate that most of the small companies have autocratic decision-making structures in which one individual is typically responsible for taking decisions. As a result, most of respondents do not have any formal rules for purchasing vehicles (58%). However, when they do, most of them state that they include environmental criteria in their purchase rules (84%). The average utilities The Hierarchical Bayes method estimated the individual utilities of the 45 respondents based on 450 choice tasks. The goodness of fit of these estimations was assessed with the percent certainty of the model, which was first proposed by Hauser (1978). It indicates that the log likelihood of our model is 70.7% better than the log likelihood of the chance model (Kurz and Binner, 2012). According to Orme (2011), the average percent certainty is 71%, the minimum is 60%, and the maximum is 83%, which confirms the quality of our model. In this section, we report the average utilities of all of the respondents with their 95% confidence intervals and a trend line when relevant. Purchase costs The attribute purchase costs were specified to the respondents to include the purchase price of the vehicle (excluding VAT), governmental support and the registration tax. Figure 20 shows the expected evolution of the purchase costs attribute: lower purchase costs are preferred. The curve shows also a negative diminishing marginal utility. It suggests that transporters become less sensitive to purchase costs differences when their values increase. The effect of a subsidy is therefore reduced when the purchase cost of the subsidised vehicle increases and vice versa. Given the higher purchase costs of BEVs relative to conventional vehicles, a subsidy might therefore not be the most efficient measure to influence transporters towards BEVs.

60

40

40

40

20

20

20

0 €0

€ 20,000

€ 40,000

€ 60,000

0

10

20

30

0 0

-20

-20

-20

-40

-40

-40

-60

-60

Purchase costs (€)

Figure 20: average utilities of purchase-cost levels

Figure 21: average utilities of operating-cost levels 60

60

40

40

40

20

20

20

8hours

30min

5min

-20

0 3m³

8m³

20m³

400

600

Range (km)

Figure 22: average utilities of range levels

60

0

200

-60

Operating costs (€/100 km)

Utilities

Utilities

0

Utilities

60

Utilities

60

Utilities

Utilities


0 0

-20

-20

-40

-40

20

40

60

80

100

-40 -60

Charging or refuelling time at the depot

Figure 23: average utilities of refuelling- and charging-time levels

-60

Volume capacity of the vehicle

Figure 24: average utilities of volume levels

-60

Ecoscore

Figure 25: average utilities of Ecoscore levels


Operating costs Operating costs are expressed in euros per kilometre to compare fuel prices and electricity prices. Respondents were informed that this attribute also considered maintenance costs. As in the case of the purchase costs attribute, Figure 21 shows that lower operating costs are preferred. However, the curve shows in that case a negative increasing marginal utility. It indicates that transporters will be more sensitive to increasing operating costs than to decreasing operating costs. The low operating costs of BEVs are therefore poorly valorised by transporters. Conversely, an increase of operating costs for conventional vehicles should influence more strongly the choice of transporters towards BEVs. Range The range attribute was described as the number of kilometres that the vehicle can drive on a full charge or a full tank. The trend in Figure 22 shows that a higher range is preferred, as expected. Still, we observe a positive increasing marginal utility. It suggests that transporters are less sensitive to range differences when their values are low. The difference between a BEV that has a range of 100km or a range of 200km will therefore not be much valorised: the limited range is still constraining. Note, however, that the spread of the confidence interval is large for the range of 100 kilometres. It indicates a greater heterogeneity in the preferences of the respondents for that level. Some transport companies can indeed be very sensitive to lower ranges, as these might no longer be compatible with their transport activity. Conversely, some transporters might have an activity compatible with low ranges which impact then less negatively their valorisation of BEVs. Refuelling or charging time To reduce the number of attributes, we considered the lack of charging infrastructure and the charging time in a single attribute. We described this attribute to the respondents as the time needed to achieve a full charge or a full tank at the depot. Figure 23 shows that charging a vehicle for 8 hours is not an option for transporters. This is the level that receives the lowest utility across all attributes. However, fast charging at the depot could be a real solution for BEVs, according to transporters. The utility of that level is very similar to the 5-minute level. Transporters spend indeed more than 5 minutes at the depot to unload and load vehicles. A longer charging/refuelling time would therefore not affect their operations. Volume The type of vehicle was described by its available payload in terms of volume. A picture was also attached, as depicted in Figure 19, to allow the respondents to have a better understanding of the volume measurement. Figure 24 depicts the average utility for each type of vehicle. A vehicle with a payload of 3 m³ receives the lowest utility due to its limited capacity. However, the vehicle with the largest capacity (volume of 20 m³) is not the most preferred type of vehicle. We observe a clear preference for a vehicle with a volume of 8 m³. The estimation of that attribute might reflect the fact that respondents consider both manoeuvrability and the legislation associated with freight vehicles presenting a gross vehicle weight over 3.5 tonnes. However, we can stress the large spread of the confidence intervals, which are the most important across all attributes. That strong heterogeneity in the preferences for the vehicle type reflects the large diversity of needs in city logistics. This is particularly the case for small vans.

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 57 Some transporters might value small vans more because they specialise in delivering small packages such as mail. They do not need large capacity, and they prefer the higher level of accessibility that the small vehicle can provide. Conversely, a small capacity might not meet the requirements of other transport activities, such as that of a moving company. In that case, we expect that the level with the highest capacity will be the most preferred. Ecoscore Finally, the Ecoscore was described to the respondents as the vehicle’s environmental performance. The Ecoscore was explained as a measure that combined greenhouse gas emissions, local pollutants and the vehicle’s noise level. They were also informed that a higher score means better environmental performance. Figure 25, however, shows a weak trend, indicating the limited impact of these levels on utility, although the trend line is positive. This shows that respondents do not highly value the environmental impact of the vehicles relative to the other attributes. Therefore, the BEVs’ environmental benefit might not have a strong influence on transporters’ purchase decisions. The trade-offs To illustrate transporters’ choice behaviour when comparing a BEV to a conventional vehicle, we used the average preference structure in a real choice situation. In this section, we apply the utility functions presented in the previous section to the characteristics of the electric and diesel version of the Peugeot Partner, as presented in Table 3. By summing the utility of each attribute, we can estimate the total average utility of the vehicle. We observe that the diesel version is much more preferred by transporters. Table 3: Characteristics of the two versions of the Peugeot Partner

Purchase costs Operating costs Range Charging time Volume Ecoscore Total average utility

Electric Partner

Diesel Partner

€26,000

€13,960

€5/100 km

€11.7/100 km

100 km

500 km

30 min

5 min

3 m³

3 m³

90

60

-6.86

+65.75

To clarify the results in Table 3, Figure 26 highlights the contribution of each attribute to the total average utility of the vehicles. It shows the trade-offs made by the respondents and identifies the most important aspects that influence the purchase decision. We first observe that the utilities of both vehicles are entailed by the Partner’s limited volume. However, because both vehicles have the same value, volume will not influence the vehicle choice. The influence of the Ecoscore is also very limited. We have indeed shown that the environmental performance was poorly valued by respondents. The attribute that contributes the most to the utility of the BEV relative to that of the diesel vehicle is low operating costs. However, that advantage cannot

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 58 compensate for the utility gap created by the BEV’s higher purchase costs. But the attribute that most restrains the choice of a BEV appears to be its limited range. Conversely, longer charging times are not observed to have a significant impact on the choice of the BEV due to the possibility of fast charging at the depot. To stimulate the purchase of BEVs, various policy measures have been identified in the literature review as influencing the choice behaviour of freight transport operators. Based on the preference structure estimated by our model, we can also estimate the effect of such policies on the average choice behaviour. In Figure 27, we show the influence of the €5,000 subsidy provided by the Brussels-Capital Region on the choice in a business-as-usual scenario (Moniteur belge, 2013, 2009). In that situation, the utility gap induced by high purchase costs is reduced but the BEV remains less attractive from a pure cost perspective. As a result, the total utility of the BEV is still below that of the diesel vehicle.

Utility

Figure 26: Choice in business as usual

Figure 27: Choice with a subsidy

Figure 28: Choice with road pricing

100

100

100

80

80

80

60

60

60

40

40

40

20

20

20

0 -20

0 Electric

Diesel

-20

0 Electric

Diesel

-20

-40

-40

-40

-60

-60

-60

-80

-80

-80

-100

-100

-100

Electric

Diesel

In 2016, the Brussels-Capital Region will introduce a limited road pricing scheme whereby all heavy goods vehicles will pay a minimal price of 25 eurocents per kilometre. We therefore assessed the influence of such a scheme if the road pricing was extended to light commercial vehicles with a fee of 15 eurocents per kilometre, except for BEVs that are exempted. Figure 28 shows that the total utility gap between the electric and diesel vehicles is substantially reduced. The BEV’s utility is not affected relative to that in Figure 26 because it is exempted from the scheme. However, the diesel’s utility is reduced to an extent where the valuation of the BEV becomes close to that of the diesel vehicle. If both policy measures are combined, the average utility of the BEV will be higher than that of the diesel vehicle.


3.5 Results – Attitudes of respondents To better interpret the results of the CBC analysis, it is useful to investigate also the attitudes of the respondents. Choices are indeed influenced by intentions (Ajzen, 1985). Respondents were therefore asked to report their attitudes on different statements concerning BEVs. The results in this section present the distribution of the commitment level to these statements across the sample. Disadvantages The two most critical disadvantages of BEVs pertain to range anxiety. Figure 29 shows that the scant public charging infrastructure and the limited range represent the most critical barriers to adopting BEVs. Long charging times, however, appear to contribute less to range anxiety. The high purchase cost of BEVs is indeed considered a more critical disadvantage than long charging times. The other disadvantages of BEVs are more closely associated with the nascent stage of the market. They are considered less critical. Still, we can stress that the lack of BEV maintenance expertise in garages is identified by the respondents to be a disadvantage that is nearly as important as the high purchase cost of the vehicle. Information on BEV technology and available models are also lacking given the importance associated with those barriers. Finally, we note that respondents consider uncertainty about residual value as the least important barrier, although it can influence significantly the total cost of ownership of BEVs. Figure 29: Disadvantages of BEVs Uncertainty on residual value Few models of BEVs available Few information available Few garages with the required expertise Long charging time without fast charger High purchase cost Limited Range Low public charging infrastructure 0%

Critical disadvantage

10%

20%

30%

40%

Important disadvantage

50%

60%

70%

80%

90% 100%

Less important disadvantage

Advantages Figure 30 confirms the results depicted in Figure 29. Aspects affecting transport operations are considered a priority. We observe that the most critical advantage of BEVs over conventional vehicles is the ability to charge at the depot. This improves operations because transporters save time and distance by not having to drive to the fuel station. Figure 30 confirms also that cost is the second most important aspect after operational issues: low operational costs are considered the second most critical advantage. BEVs’ environmental friendliness is ranked as the third most critical advantage, after operational and costs constraints. The other advantages are considered much less important, particularly that of quick acceleration. In a city with traffic, we recognise that such an advantage will not enable a

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 60 driver to gain a significant amount of time. Silence and ease of driving, however, are two advantages that might be more valued by transporters. Figure 30: Advantages of BEVs Easy to drive in the city Quick acceleration until 60km/h (Almost) silent Environmentally friendly Low operational cost Possibility to charge at the depot

0%

Critical advantage

10%

20%

30%

40%

Important advantage

50%

60%

70%

80%

90% 100%

Less important advantage

Support from authorities Transporters have however a limited knowledge about electric freight vehicles. Among our respondents, 62% reported that they are unfamiliar with any electric freight vehicles on the market and that they do not know any models on the market. Still, 20% stated that they are aware of models but they do not know about their characteristics. In fact, only 13% of respondents stated that they are aware of some models and are well informed about their characteristics. Nevertheless, transporters show a positive attitude towards BEVs despite this limited knowledge: 85% of the respondents agree that the government should support the introduction of BEVs. Respondents were therefore asked about the most important policy measures that the authorities should implement. The results depicted in Figure 31 show that the respondents primarily support measures that reduce the costs of BEVs. In particular, we observe that respondents regard an exemption for BEVs from a kilometre tax as a priority. Subsidies for the purchase of BEVs, fiscal deductions for BEVs and exemption from an urban toll at the city entrance for BEVs are other measures that are strongly supported by respondents. However, addressing operational constraints should not be neglected given the importance of that barrier. Indeed, Figure 31 shows that transporters regard installing fast chargers at existing fuel stations as the second highest priority measure that the authorities should adopt to support the introduction of BEVs. We also note that this measure is the only one that every respondent considers to be at least an important measure. Conversely, implementing slow chargers in parking facilities is not considered a popular solution to address range anxiety. Finally, we stress that measures modifying the accessibility of certain areas in the city are less preferred by transporters. For example, a ban on polluting vehicles delivering in the city centre or the whole city is not considered an important measure. However, providing BEVs with an exclusive right to access bus lanes or to deliver at night in the city is regarded as a more important measure to stimulate BEVs, according the respondents.

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 61 Figure 31: Respondent support for measures to stimulate BEVs Ban polluting vehicles from delivering the whole city Ban polluting vehicles from delivering the city center Allow BEVs to deliver at night between 22h and 6h Implement slow chargers in parkings Allow BEV to access bus lanes Exemption from an urban toll at the city entrance for… Provide subsidies for the purchase of BEVs Deduct fiscally costs related to BEVs Implement fast chargers in gas stations Exemption from a kilometer tax for BEVs

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Prior measure

Important measure

Low important measure

3.6 Discussion In section 4, we assessed the influence of BEVs’ attributes on the choice behaviour of transporters, and in section 5, we investigated their attitudes. In this discussion, we compare those results in order to provide some policy recommendations. The role of authorities The respondents’ attitudes showed that a vehicle’s environmental friendliness is one of the most important advantages of BEVs over conventional vehicles. The literature revealed the same findings from transporters operating BEVs. However, the CBC analysis shows that the Ecoscore attribute is little considered in the purchase decision. That difference suggests that the influence of the environmental attribute on vehicle choices is seriously reduced when it is compared to other attributes. The large attitude-action gap observed by Mairesse et al. (2012) among households for the purchase of BEVs can therefore also be identified among freight transporters. Nevertheless, 85% of the respondents agreed that the government should stimulate the introduction of BEVs. That support from transporters regarding regulations stimulating BEVs confirms that they have a positive attitude towards BEVs and that they consider these vehicles as a possible solution for city logistics. However, some regulations are less preferred than others. In particular, authorities should be attentive when implementing measures that restrict access of some urban areas to BEVs only. We observed that a ban on polluting vehicles for entering the city was the least supported. Developing charging infrastructure In the literature review, we found that an interesting benefit of BEVs is the ability to charge in the depot. Our results highlight this benefit as the most important advantage of BEVs. It enables transporters to save time and distance because they no longer need to drive to fuel stations. Additionally, the constraint of a slower charging relative to conventional vehicles is reduced given the time needed to load/unload the vehicles. The results of the CBC confirmed this finding because obtaining a full charge in 30 minutes at the depot was valued similarly to a refuelling time

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 62 of 5 minutes. However, the CBC revealed that charging time could also become the worst attribute of a BEV if no fast charging were available at the depot. The profile of respondents can explain that behaviour. We have indeed a large number of transporters that do not park their vehicles in a depot. Instead, they leave their vehicles in a private parking or on streets where chargers are difficult to find. They therefore expressed in the CBC a strong preference for fast charging at the depot. The installation of fast chargers in fuel stations was also expressed by respondents to be the second most important measure to implement by authorities. Charging infrastructure is therefore also an important barrier that needs to be solved in city logistics that is less recognised by the literature. The lack of public charging infrastructure is indeed regarded by respondents as the most critical disadvantage of BEVs. Developing knowledge Developing a public network of fast chargers might however quickly become expensive. Transporters nevertheless support that measure to a greater extent than the development of slow chargers in parking facilities. But they might not be aware of the costs of such an infrastructure. They might also not realise that exclusively using fast charging might degrade more rapidly the batteries and thereby increase the costs of battery replacements (Omar, 2012). We found indeed that 82% of respondents did not have a good understanding of the characteristics of BEVs. Better information should therefore be communicated to freight operators. The literature review has stressed indeed that the lack of knowledge was another important barrier to BEV adoption. Existing charging infrastructure should be in particular well reported. Improving that information could already reduce the barrier of the lack of infrastructure. More emphasis should also be made on the good performances of BEVs. The literature review has stressed that testing a BEV is important to counterbalance the negative costs aspects of BEVs. Available government support for freight transport companies should also be better communicated. Developing financial incentives After charging infrastructure, the results indicate that limited range is the second most critical disadvantage of BEVs in city logistics. The CBC revealed the important utility gap that BEVs must overcome because of their range constraints relative to conventional vehicles. The profile of respondents can again explain this behaviour. We observe that most of them operate small fleets. Thus, such transporters might have a higher valuation of vehicles that are flexible to operate. Nevertheless, the CBC revealed some heterogeneity in how low ranges are valued, which leaves space for potential early adopters. Solutions to the range barrier are however limited because this constraint is inherent to BEVs. And respondents do not valorise sufficiently the environmental performance as well as the low operating costs to overcome the barriers of BEVs. Our analysis suggests therefore that the most effective strategy is to use a combination of financial incentives to compensate for range constraints and high purchase costs of BEVs. The most popular measure is an exemption for BEVs from a kilometre tax. The CBC showed that it was also the type of measure that influences the most the choice of transporters towards BEVs. Other supported measures involve an exemption for BEVs from urban tolls at the city entrance; a fiscal deduction for costs related to BEVs; and subsides for the purchase of BEVs.

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 63 The role of manufacturers In addition to authorities, manufacturers also have an important role to play in stimulating the adoption of BEVs in city logistics, especially in terms of product development. In the literature review, we observed that the limited choice of BEVs was an important barrier (ELCIDIS, 2002). The results concerning the respondents’ attitudes suggest that this barrier remains problematic, despite that the number of companies producing electric vehicles increased significantly between 2000 and 2011 (Sierzchula et al., 2012). Perhaps more electric models should be developed in the large-van segment. The results of the CBC stressed indeed an important preference for those vehicles. The respondents’ attitudes showed also the importance of developing a better maintenance service for BEVs. Ensuring a good reliability of BEVs has been stressed in the literature to be of “the utmost importance”. We did not assess the importance of that attribute as we considered it to be a condition to consider BEVs in the choice of vehicles and not a preference. That is another challenge that the manufacturers should address to reduce the barriers to BEV adoption. Finally, the literature review has stressed the importance of reducing uncertainty related to batteries of BEVs. Manufacturers have already provided solutions regarding the uncertainty around battery replacements with warranty or battery leasing systems. But solutions related to the residual value of batteries can be another important improvement for BEVs although the respondent’s attitudes have considered it as the least critical disadvantage of BEVs. Perhaps, they do not realise the impact of that aspect on the total costs of ownership given their limited knowledge on BEVs.

3.7 Conclusions This chapter investigated the choice behaviour of transport companies for BEVs in city logistics. To capture their preference structure, a survey was conducted in the Brussels-Capital Region using a CBC analysis. Respondents were asked to choose between hypothetical vehicles that were described according to a set of attributes. Based on a literature review, we identified the following attributes as the most relevant for modelling the choice between BEVs and conventional vehicles: daily range, charging time, environmental performance, purchase cost, operating cost and type of vehicle. The survey collected the preferences from 45 respondents, which represents approximately 10% of the transporters registered in the Brussels-Capital Region. They each completed 10 choice tasks, resulting in a total of 450 choice tasks that we could analyse. Additional questions regarding the advantages and disadvantages of BEVs over conventional vehicles were asked to the respondents to gain further insights into their preferences. Based on the results, we discussed solutions that could stimulate the introduction of BEVs into city logistics. We found that environmental performance and low operating costs are insufficiently valorised by transport operators to influence their choice towards BEVs. Still, they generally agree that the authorities should support the introduction of BEVs. Policies should therefore be implemented in order to achieve the objective that city logistics faces in major urban areas: become CO2 free by 2030. Operational constraints are considered the most important barriers for transport operators, although city logistics is often considered in the literature as a suitable environment for BEVs. We identified that charging might be particularly difficult because transporters mostly park their vehicles on streets and less in depots. The respondents

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 64 thus regard the low public charging infrastructure as the most critical disadvantage of BEVs. This barrier could however become an advantage over conventional vehicles. Indeed, charging at the depot is identified as the most important benefit of BEVs. The CBC showed in particular the importance of developing fast charging solutions at the depot. Still, the limited range should remain the most complex operational constraint to solve. The CBC revealed that a combination of financial incentives could be an effective strategy to compensate for the limited range of BEVs and their high purchase costs. These measures are also the most supported by transporters, with a particular preference for a BEV exemption from a kilometre tax. These conclusions are however limited to the attributes that we investigated. The CBC did not consider other attributes that positively influence choice behaviour for BEVs, such as smooth and quick acceleration, low maintenance costs and access advantages for BEVs to some areas of the city. The CBC also did not consider some attributes that negatively influence choice behaviour for BEVs such as the limited speed, battery replacement costs or limited availability of maintenance services for BEVs. Moreover, the limited size of the sample also limits the analyses to transport companies’ average choice behaviour. The confidence intervals indicated however that some transporters value the attributes of the vehicles differently, which might leave room for potential early adopters. Future research could therefore address the choice behaviour of different segments in city logistics. Future research could also specifically address the barriers of range constraints, which appear to be particularly challenging to solve. Finally, because measures addressing the costs of BEVs seem to be the most effective policies to stimulate BEV adoption, future research could also investigate the impact of measures such as fiscal incentives, subsidies and toll schemes on the competitive position of BEVs in a city logistics context.


References ADEME, 2004. Espaces logistiques urbains de monaco et la rochelle. ADSEI, DIV and MOBI, 2015. Calculations based on the development of the TRABAM model in the research group MOBI. Ajzen, 1985. From intentions to actions: A theory of planned behavior, in: Kuhl, J., Beckmann, J. (Eds.), ActionControl: From Cognition to Behavior. Heidelberg, pp. 11–39. Amburg, B. Van, Pitkanen, W., 2012. Best Fleet Uses, Key Challenges and the Early Business Case for E-Trucks  : Findings and Recommendations of the E-Truck Task Force, in: EVS26. Los Angeles, California, May 6-9, pp. 1–12. BESTUFS II, 2006. Quantification of Urban Freight Transport Effects II. BESTUFS II, 2008a. Best Practice Update (2008) - Public Private Partnerships (PPP) in urban freight transport. BESTUFS II, 2008b. Best Practice Update (2008) on Waste transport logistics in urban areas & Experiments and incentives for environment-friendly vehicles. Browne, M., Allen, J., Leonardi, J., 2011. Evaluating the use of an urban consolidation centre and electric vehicles in central London. IATSSR 35, 1–6. Cavar, I., Kavran, Z., Jolic, N., 2011. Intelligent transportation system and night delivery schemes for city logistics. Comput. Technol. Appl. 2, 782–787. Dablanc, L., 2011. City distribution, a key element of the urban economy: guidelines for practitioners, in: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport  : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 13–36. De Vroey, L., Jahn, R., El Baghdadi, M., Van Mierlo, J., 2013. Plug-to-wheel energy balance - Results of a two years experience behind the wheel of electric vehicles, in: EVS27. Barcelona, pp. 1–5. DELIVER, 2012. Deliverable D1.1 - Report on Technology, Market and Urban Logistics Roadmap from 2020 and Beyond. Aachen, Germany. EC, 2011. White paper: Roadmap to a Single European Transport Area. Brussels. ELCIDIS, 2002. Electric Vehicle City Distribution, Final Report. ENCLOSE. (2014). ENergy efficiency in City LOgistics Services for small and mid-sized European Historic Towns Final Publishable Report. EVD-POST. (2001). Electric Vehicle Deliveries in Postal Services. E-Mobility NSR, 2013. Comparative Analysis of European Examples of Schemes for Freight Electric Vehicles Compilation report. FREVUE, 2013. State of the art of the electric freight vehicles implementation in city logistics. Gelman, A., Carlin, J.B., Stern, H.S., Rubin, D.B., 2009. Bayesian Data Analysis, Second. ed. Chapman & Hall/CRC. Green, P.E., Srinivasan, V., 1978. Conjoint Analysis in Consumer Research: Issues and Outlook. J. Consum. Res. 5, 103–123. Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., 2010. Multivariate Data Analysis: a global perspective, Seventh. ed. Pearson, New Jersey. Hauser, J. R., 1978. Testing and Accuracy, Usefulness, and Significance of Probabilistic Choice Models: An Information-Theoretic Approach. Operations Research 26, 406-421. Hoogerbrugge, M., van der Wagt, K., 2006. How many choice tasks should we ask?, in: Proceedings of the Sawtooth Software Conference. pp. 97–109. Kurz, P., Binner, S., 2012. “The individual choice task threshold” Need for Variable Number of Choice Tasks, in: Proceedings of the Sawtooth Software Conference. pp. 111–127. Lebeau, K., Van Mierlo, J., Lebeau, P., Mairesse, O., Macharis, C., 2012. The market potential for plug-in hybrid and battery electric vehicles in Flanders: A choice-based conjoint analysis. Transp. Res. Part D Transp. Environ. 17, 592–597.

Chapter 3: Exploring the choice of battery electric vehicles in city logistics: a CBC analysis - 66 Lebeau, K., 2013. Electrifying cars  : the economic potential of electric vehicles. Vrije Universiteit Brussel. Lebeau, P., Macharis, C., Van Mierlo, J., Lebeau, K., 2015. Electrifying light commercial vehicles for city logistics? A total cost of ownership analysis. The European Journal of Transport and Infrastructure Research 15 (4), 551569. Litschke, D.-I.A., Knitschky, D.-V.G., 2012. Future development in road freight transport regarding more environmentally friendly vehicle technologies. Procedia - Soc. Behav. Sci. 48, 1557–1567. Macharis, C., Van Mierlo, J., Van Den Bossche, P., 2007. Combining Intermodal Transport With Electric Vehicles: Towards More Sustainable Solutions. Transp. Plan. Technol. 30, 311–323. Mairesse, O., Macharis, C., Lebeau, K., Turcksin, L., 2012. Understanding the attitude-action gap: Functional integration of environmental aspects in car purchase intentions. Psicologica 33, 547–574. Massiani, J., 2014. Stated preference surveys for electric and alternative fuel vehicles  : are we doing the right thing  ? Transp. Lett. Int. J. Transp. Res. 6, 152–160. Melo, S., Costa, A., 2011. Definition of a set of indicators to evaluate the performance of urban goods distribution initiatives, in: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport  : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 120–147. Moniteur belge, 2009. C- 2009/31231. Moniteur belge, 2013. C-2013/31464. Nesbitt, K., & Sperling, D. (2001). Fleet purchase behavior: decision processes and implications for new vehicle technologies and fuels. Transportation Research Part C: Emerging Technologies, 9(5), 297–318. doi:10.1016/S0968-090X(00)00035-8 Nykvist, B., & Nilsson, M. (2015). Rapidly falling costs of battery packs for electric vehicles. Nature Climate Change, 5(4), 329–332. doi:10.1038/nclimate2564. Omar, N., 2012. Assessment of Rechargeable Energy Storage Systems for Plug-In Hybrid Electric Vehicles. Vrije Universiteit Brussel. Orme, B.K., 2011. Personal communication on the Sawtooth software support forum [WWW Document]. URL https://sawtoothsoftware.com/forum/6206/rlh-benchmark?show=6206#q6206 Pelkmans, L., Debal, P., 2006. Comparison of on-road emissions with emissions measured on chassis dynamometer test cycles. Transp. Res. Part D Transp. Environ. 11, 233–241. Sierzchula, W., Bakker, S., Maat, K., Van Wee, B., 2012. The competitive environment of electric vehicles: An analysis of prototype and production models. Environ. Innov. Soc. Transitions 2, 49–65. STRAIGHTSOL. (2015). Straightsol - Final publishable report. Soret, A., Guevara, M., Baldasano, J.M., 2014. The potential impacts of electric vehicles on air quality in the urban areas of Barcelona and Madrid ( Spain ). Atmos. Environ. 99, 51–63. SUGAR, 2011. Sustainable Urban Goods Logistics Achieved by Regional and Local Policies, City. Bologna, Italy. Taniguchi, E., Kawakatsu, S., Tsuji, H., 2000. New co-operative system using electric vans for urban freight transport. Transport. Van Mierlo, J., Maggetto, G., Vereecken, L., Favrel, V., Meyer, S., Hecq, W., 2003. How to define clean vehicles? Environmental impact rating of vehicles. Int. J. Automot. Technol. 4, 77–86. Van Mierlo, J., Maggetto, G., 2007. Fuel Cell or Battery: Electric Cars are the Future. Fuel Cells 7, 165–173. Wellman, G.S., Vidican, C., 2008. Pilot study of a hierarchical Bayes method for utility estimation in a choice-based conjoint analysis of prescription benefit plans including medication therapy management services. Res. Social Adm. Pharm. 4, 218–30. Zanni, A.M., Bristow, A.L., 2010. Emissions of CO2 from road freight transport in London: Trends and policies for long run reductions. Energy Policy 38, 1774–1786.

PART II

Chapter 4: Electrifying light commercial vehicles for city logistics? A TCO analysis

- 68 -

4 Electrifying light commercial vehicles for city logistics? A total cost of ownership analysis4 4.1 Introduction Freight transport is expected to increase in cities the coming years because of converging trends: the current urbanization process generates more freight volumes in cities, transport is increasingly fragmented due to the success of light commercial vehicles (LCVs) and distances are stretching out due to the delocalisation of logistics platforms in the periphery (Dablanc and Rakotonarivo, 2010; Browne et al., 2010). Even though that growth should support the development of the urban economy, it might also increase the environmental burden of freight transport in cities. Freight vehicles are already responsible for about one fourth of CO 2 emissions, one third of NOx emissions and half of particulate matters generated by the transport sector in cities (Dablanc, 2011). Recognizing the need to find solutions, the European Commission has given the objective of reaching free CO2 city logistics in major urban areas by 2030 (EC, 2011). A wide range of logistics concepts, regulations and technologies have been developed to fulfil the city logistics carbon free target (Quak, 2011). Among them, battery electric vehicles (BEVs) are considered to be an answer to the negative impacts listed above (Crainic et al., 2004; Van Mierlo & Maggetto, 2005). Browne, Allen & Leonardi (2011) have shown the positive contributions 4

This chapter is based on the following paper: Lebeau, P., Macharis, C., Van Mierlo, J. and Lebeau, K. 2015. “Electrifying light commercial vehicles for city logistics? A total cost of ownership analysis”. The European Journal of Transport and Infrastructure Research 15 (4), 551-569.


- 69 -

related to the use of BEVs in an urban consolidation centre. Their analysis showed an important reduction in terms of CO2 and congestion on the whole urban area. The logistic concept analysed showed also that the limited driving range of BEVs is not problematic. This constraint is easily controlled by the structured and time-based environment of the logistics chain. Besides, the frequent use of the vehicles in city logistics makes the low running costs of BEVs particularly interesting to reduce transport costs. Indeed, BEVs benefit from both lower energy prices and a more efficient energy consumption compared to conventional vehicles (Martensson, 2005). However, purchase cost tends to be the most important criteria in the vehicle choice of fleet managers (Nesbitt and Sperling, 2001). A survey conducted by Van Amburg and Pitkanen (2012) confirms this behaviour since fleet managers consider the high purchase cost of a BEV as the main barrier for switching from a conventional vehicle. But when deciding on the purchase of a vehicle, a rational fleet manager should consider every cost related to the vehicle choice, and not only the purchase cost. The difference of cost structure between electric and conventional vehicles makes such an analysis particularly important: estimating the breakeven points where BEVs become competitive with conventional vehicles should be therefore clarified. Previous research on the competitive position of BEVs has overlooked the segment of freight vehicles. Most of the attention has been instead on the passenger vehicles (Lebeau, 2013; Turcksin, 2011). Literature reveals only a few studies. Feng and Figliozzi (2012, 2013) approached the question by developing a fleet replacement model considering an electric and a conventional truck. The results show that electric trucks are competitive only with high utilisation: the breakeven point is estimated around 45,000 kilometres per year per truck. The market position of electric trucks was further investigated by Davis and Figliozzi (2013), who integrated four models to evaluate the optimal routing parameters, the energy needed for operating the deliveries, the range constraints and the associated ownership costs. The results illustrate that electric trucks become more competitive when more customers are served, the energy costs increase and when longer distances are travelled. On the other hand, the competitive position decreases with lower purchase costs for diesel trucks and reduced payload. As a result, Davis and Figliozzi (2013) conclude that electric trucks are viable in an environment where daily distances approach 160 kilometres, low speeds and congestion are present, stops are frequent and where authorities are supporting electric vehicles with tax incentives or technological breakthroughs. The planning horizon of the vehicle should also be beyond 10 years. These results are partly confirmed by Lee et al. (2013) who conducted a total cost of ownership analysis comparing an electric and a conventional truck in an urban environment (with frequent stops and low average speed) and in a suburban environment (with less frequent stops and higher average speed). In the urban environment, they found that the BEV has a TCO of 22% less than the conventional truck while, in a sub-urban environment, the TCO of the BEV was 1% more than the conventional vehicle. These studies consider however a limited selection of vehicles to describe the competitive position of BEVs. Often, the analyses compare one electric with one conventional freight vehicle. Also, the scope of the analysed vehicles covers the freight vehicles with a gross vehicle weight of more than 3.5 tonnes. However, these heavy good vehicles (HGVs) only represent 12% of the trips in city logistics, while 88% of the trips is done with LCVs (PORTAL, 2003). Given the


- 70 -

expected growth of the LCVs (Zanni & Bristow, 2010), that segment deserves a detailed analysis. Hence, the objective of this chapter is to extend the previous competitive analyses of BEVs to the segment of LCVs. Consequently, the main research questions addressed in this chapter are: 1. What is the competitive position of electric LCVs according transport companies compared to their conventional counterparts? 2. What are the breakeven points for electric LCVs to be competitive at current market conditions and future market conditions? 3. What are the minimum levels of market interventions needed in order for electric LCVs to be competitive? We developed a total cost of ownership (TCO) model for LCVs with a gross vehicle weight of maximum 3.5 tonnes. We analysed the costs from 7 BEVs, 5 diesel vehicles and 3 petrol vehicles available on the Belgian market. As we are interested in city logistics, we used as case study the institutional context and the geographical scope of the Brussels-Capital Region. After detailing the methodology, the chapter presents the results of the model in order to solve the first research question. Then, the different breakeven points are identified through a sensitivity analysis in order to answer the two last research questions. They allow to extend the results to other urban contexts than the one of the Brussels-Capital Region.

4.2 Methodology Owning and operating a vehicle is associated with costs that occur at different moments in time. To compare these costs across time, the TCO methodology uses the financial formula of the present discounted value. This way, the present value of every cost can be summed to obtain the full cost of one alternative. The TCO is defined as “a purchasing tool and philosophy which is aimed at understanding the true cost of buying a particular good or service from a particular supplier”(Ellram, 1995, p. 4). To calculate the present value of future one-time costs, the following formula is used (Mearig et al., 1999):

PV  At 

1 (1  I ) t

(1)

Where:

PV

= Present value

At

= Amount of one-time cost at a time t

I

= Real discount rate

t

= Time (expressed as number of years)

The financial formula of the present discounted value is divided into three parameters: (1) the period of time over which these costs occurred, (2) the discount rate applied on future costs to actualize them and (3) the costs of ownership. This section describes how these parameters were collected. Finally, once the present values are calculated, the sum of the actualized costs related to the use of a vehicle is divided by the number of kilometres driven during the ownership period in order to get the TCO of each vehicle per kilometre.


- 71 -

Period of ownership Because of the intense competition in the logistics sector, transport operators tend to operate their vehicles as long as possible in order to limit investments (Dablanc, 2007). According the data of the Belgian technical control (GOCA, 2015), the average end life of the LCVs in Belgium is 9.89 years. The period of ownership used in the TCO model is therefore 10 years. But the sensitivity analysis will investigate how the TCO change with different ownership periods. Discount rate The discount rate can be defined as “the rate of interest reflecting the investor’s time value of money” (Mearig et al., 1999, p. 6). By actualizing the future costs with the discount rate, the TCO considers that paying tomorrow is preferred than paying today. The discount rate reflects the return that the money spent in a more expensive vehicle could have brought if it was invested in the financial markets. In order to eliminate the additional return for the risk, the discount rate is often based on long-term interest rate of state bonds. Since we consider a period of ownership of 10 years, we use the Belgian long-term bounds at 10 years which showed a rate of 1.15% in July 2015 (ECB, 2015). The discount rate can be either real (excluding inflation) or nominal (including inflation). We prefer in our analysis the real discount rate because it eliminates complex accounting for inflation over the future costs. We extract from the interest rate the 1.1% of expected average inflation between 2014 and 2020 in the euro zone (Federal Planning Bureau, 2015) to find a real discounted rate of 0.05%. Costs of ownership The costs for each vehicle were retrieved from the standard vehicle versions by contacting the manufacturers, the distributors, the car dealers, the insurance companies and the regulatory bodies. The selection of BEVs took into account the diversity of the market supply: we included vehicles from different segments according the European vehicle classification (European Commission, 2007, 2002), with different gross vehicle weight and different business models (battery leasing and purchasing). When different battery packs are available, we considered always the option that allows a minimum range of 70km. To be able to compare the BEVs with conventional vehicles, the most similar versions of the selected BEVs were also analysed (5 diesel and 3 petrol vehicles). The TCO analysis considers all costs associated with the use of the vehicle, except for the investments in charging infrastructure as they should be diluted according to the size of the fleet. The following costs streams are considered: annual car inspection, fuel (and electricity) costs, “maintenance, repair and tyre replacements” (MRT), insurance, vehicle purchase costs, battery costs, governmental support, fiscal incentives and road taxes. All costs are excluding VAT. They are detailed in Table 4 and in the following sub sections. The following sub sections describe also the assumptions applied on these costs. Fuel and electricity costs According GOCA (2015), LCVs drive on average 147,281 kilometres on 10 years. But that distance depends on their age as shown in Figure 32. That variation will be considered in the


- 72 -

sensitivity analysis. In order to calculate the running costs, the model uses the following energy prices (excl. VAT): €1.542/l 5 for petrol, €1.281/l6 for diesel and €0.1636/kWh 7 for electricity. Energy costs are assumed not to increase more than inflation. Hence, the TCO model does not simulate changes in fuel prices since we use the real discount rate. Though, the effect of different fuel prices on the TCO of conventional vehicles is investigated in the sensitivity analysis given the depletion of oil resources in the future. The influence of electricity prices will also be estimated. Figure 32: Average kilometres driven per year by LCVs according their age 25,000

Kilometers per year

20,000 15,000

10,000 5,000 0 1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

Age of the vehicle

Source: GOCA (2015)

Maintenance, repair and tyre replacements (MRT) The costs for MRT are based on the estimates of van Vliet et al. (2010). They recognise that maintenance and repair increase with the age of the drivetrains, so they consider that petrol engines have an MRT cost of 3.8€/100km for the first 60,000km, which increases to 4.3€/100km afterwards. On the other hand, they consider that the MRT cost of diesel engine remains constant over time with 4.3€/100km. Estimating the MRT cost for BEVs is more challenging, since the data are still limited nowadays. However, it is assumed that these costs are lower compared to conventional vehicles since they do not have an internal combustion engine: BEVs have less moving components, they face less temperature stress and they do not need oil and filter replacements (Fischer et al., 2009). According Davis and Figliozzi (2013), the MRT costs of the BEVs are half of the conventional cars. Hence, the model considers a MRT cost of 2.15€/100km for BEVs.

Source : www.petrolfed.be (price of “petrol 95 oct”, Consulted on 5th of July, 2015) Source : www.petrolfed.be (price of “diesel 10ppm”, Consulted on 5th of July, 2015) 7 Source : average of electricity prices in 2014 from Electrabel for a SME with a consumption of 50.000kWh per month (Brugel, 2014). 5 6


- 73 -

Table 4: Input parameters of the vehicles GVW (kg)

Vehicle name

Ligier Flex

Capacity (m³)

Motor type

Battery type

Purchase cost (€)

Battery cost *purchase (€) **leasing (€/month)

Battery warranty

Urban Consumption (l/100km or kWh/100km)

Insurance (€/year)

755

2.3

Diesel

N.A

11,649

N.A

N.A

5

828.16

1

2.8

Diesel

N.A

12,374

N.A

N.A

4.2

828.16

1,64

2.7

BEV

Lead Acid

20,075

2,961*

4 years

15

828.16

2,075

2.8

BEV

Lead Acid

20,370

3,248*

4 years

19.5

828.16

1,5

2.7

BEV

Lead Acid

21,500

2,115*

1 year

13.8

828.16

1,92

3

Petrol

N.A

13,950

N.A

N.A

7.2

1129.62

1,95

3

Diesel

N.A

13,600

N.A

N.A

5.2

941.06

2,126

3

BEV

Li-ion

21,150

from 73**

N.A

15.5

871.92

Peugeot Partner

1,96

3.3

Petrol

N.A

12,940

N.A

N.A

8.3

1047.91

Peugeot Partner

1,96

3.3

Diesel

N.A

13,960

N.A

5.8

941.06

Peugeot Partner Electric

2,225

3.3

BEV

Li-ion

26,000

6,000*

17.6

903.35

Nissan NV200

2

4.2

Petrol

N.A

14,100

N.A

N.A 8 years / 100.00km N.A

8.8

1104.48

Nissan NV200

2

4.2

Diesel

N.A

15,400

N.A

N.A

5.7

1010.20

Nissan e-NV200

2,22

4.2

BEV

Li-ion

25,652

From 73**

16.5

1098.20

Nissan e-NV200

2,22

4.2

BEV

Li-ion

31,550

6,000*

N.A 5 years / 100.000km

16.5

1098.20

Mega d-truck (fourgon) Mega e-Worker (fourgon court - 11,5kWh) Goupil G3-2 (L) (fourgon standard; 11,5kWh) Alke ATX210E (van box - 8,7kWh) Renault New Kangoo Express Renault New Kangoo Express Renault Kangoo ZE


- 74 -

Insurance costs The insurance costs are calculated by the insurance company AXA in 2014 using the following criteria. The vehicle belongs to a company based in Brussels (postcode 1000). It is operated “frequently”, mainly by a driver of 30 years, having a driver license for 10 years and with no accidents in the last 5 years. The insurance is limited to the civil liability. No cost difference as such is applied between electric and conventional vehicles. The differences in the insurance premiums between the vehicles depend on the power of the motors. Still, every vehicle with a power of less than 25kW have the same insurance premiums. Vehicle purchase costs The vehicle purchase cost reflects the depreciation of the vehicle during the period of ownership: the purchase price is reduced by the residual value that can be retrieved as a capital gain realised on fixed asset. We considered an annual depreciation rate of 18.57%8 on the value of the diesel and petrol vehicles. For battery electric vehicles, we used a higher depreciation rate which is set at 24.43%9. This faster depreciation of BEVs is explained by the less mature market of second hand BEVs and by the fast development of BEV supply observed these last years. Let us note that this annual depreciation rate is not applied on the battery costs but only on the vehicle purchase costs. When the battery costs are included in the initial purchase costs, they are deduced from the purchase costs category and affected to the battery costs category. Battery costs The batteries have a limited lifetime in transport applications. Once they reach 80% of their initial energy capacity, they need to be replaced. Old batteries could be used in second life applications but, since this market is not developed yet, it is not clear if such batteries can be valorised. As a result, the TCO analysis does not consider a residual value for BEV batteries but the sensitivity analysis will investigate the influence of such effects. The lifetime of the batteries can change to a large extent depending on many factors such as the depth of discharge, operating temperature and the charging method (Omar, 2012). But in order to estimate their number of cycles, we used figures based on the standardized lifecycle methodology. These figures differ according to the type of batteries. The lead-acid is the least performing battery with a lifetime of 500 cycles (Van den Bossche et al., 2006). Lithium-ion batteries have a longer lifetime that can be estimated to 1,500 cycles (Omar, 2012). Once the number of cycles is reached, the model considers that the battery is replaced by a new one. Since batteries are assumed to be charged once a day during 260 days a year, lead acid batteries are replaced after 2 years and lithium-ion batteries are replaced after 6 years. Replacement costs are assumed to be in charge of the customer only once the warranty provided by the manufacturer is over. As we could not find the cost of the battery of the Peugeot Partner, we used the same cost as the e-NV200 since they have both the same type of battery (Lithium Ion of about 24kWh). Replacement costs are however not supported by the customer if the battery is leased. Instead, a monthly fee covers the costs of battery. This is an average of the annual depreciation rates of the company LeasePlan between the Kangoo 1.5dci, Caddy 1.6tdi, Trafic 27 2.0dci L1H1, Transporter 2.0tdi swb, Master 35 2.3 dci L3H3, Crafter 35 2.0tdi lwb. 9 This is based on the Kangoo ZE annual depreciation rate of the company LeasePlan. 8


- 75 -

Regional subsidies The Brussels-Capital Region introduced a financial support for electric commercial vehicles, which consists of 20% for large firms, 30% for medium firms and 40% for small firms on the investment costs with a maximum of 5,000 euros (Moniteur belge, 2009, 2013). Since most of the firms in city logistics are small (Dablanc, 2011), the TCO model considers a 40% support. In order to evaluate the effectiveness of the government support, the sensitivity analysis will investigate the effect of different levels of subsidies. Fiscal incentives The Belgian fiscal system allows deductibility from corporate income taxes up to 120% for BEVs on every cost related to the use of the vehicle. Conversely, every cost related to a conventional vehicle supports a deductibility rate ranging from 50% to 100% depending on their CO2 emissions (Fisconetplus, 2012a). However, that fiscal system does not apply for vehicles that are designed for the transport of goods. In that case, deductibility rate remains at 100% for every type of vehicles. Hence, the model does not make a difference between electric and conventional vehicles from a fiscal point of view. But the sensitivity analysis will explore the influence of that system on the results. Road taxes and car inspection The segment of the quadricycles is exempted of road taxes and car inspection cost. Conversely, LCVs have to annually be inspected and pay road taxes. These taxes depend on their gross vehicle weight (SPF finances, 2015). No differences are however made between electric and conventional vehicles. The sensitivity analysis will explore the impact of an exemption of road taxes for BEVs. A distance based road pricing scheme is also being discussed in the Region. That scheme should be implemented in Belgium by 2016 for heavy goods vehicles but nothing is planned yet for light commercial vehicles. Still, the sensitivity analysis will investigate the effect of such a system on the competitive position of BEVs when they are exempted of such a scheme.

4.3 Results Based on the TCO model developed for quadricycles and light commercial vehicles (LCVs) in the Brussels-Capital Region, the cost structure of the different vehicles is shown in Figure 33. The results are consistent with the current market observation as diesel vehicles have a lower TCO compared to petrol vehicles, mainly due to lower fuel costs. This difference explains the market dominance of diesel vehicles within the Brussels-Capital Region. Diesel owns indeed a market share of 93% of the registered freight vehicles in the Region (Lebeau & Macharis, 2014). Competitive position of BEVs The competitiveness position of BEVs is more intermediate in their respective segments. In the light commercial vehicle segment, the results in Figure 33 show that the TCO of battery electric vehicles is located between the TCO of diesel and petrol vehicles. Still, battery electric vehicles can sometimes have a lower TCO than diesel vehicles. That is the case when we compare the Electric Kangoo ZE with the diesel NV200. But when we compare the diesel and electric versions of the Kangoo, diesel remains the most competitive vehicle. Conversely, BEVs might


- 76 -

have a higher TCO than petrol vehicles. That is the case when comparing the Electric e-NV200 with the petrol Kangoo. But the electric e-NV200 remains more competitive than its petrol counterpart. From a financial point of view, BEVs can therefore be more interesting than petrol vehicles for light commercial vehicles. Still, diesel vehicles remain today the most competitive technology in that segment. That observation can also be confirmed in the segment of quadricycles. Diesel vehicles are also more competitive than BEVs. However, it is interesting to notice that electric quadricycles have a lower TCO than diesel LCVs. Even though these vehicles cannot really be compared because of different performances, they might represent an affordable option to transporters that want to adopt BEVs. The limited power of quadricycles might indeed not be an important drawback in city logistics compared to their environmental benefits. Cost structure of BEVs When analysing the cost structure of the BEVs with the conventional vehicles in Figure 33, we observe that the main cost advantage of the BEVs over conventional vehicles (and especially petrol vehicles) is their lower operational costs. Diesel and petrol are entailed with high fuel prices and larger MRT costs while BEVs benefit from low maintenance, low electricity prices and efficient energy consumption. These advantages are however reduced by the more important purchase costs. BEVs are indeed entailed with a faster depreciation than conventional vehicles. Also, the low maturity of the market of BEVs does not ensure a competitive market and economies of scale which might explain the higher initial purchase costs for these vehicles (battery excluded). Battery costs reduce further the advantage from the low operational costs of the BEVs. Different business models are being proposed by the manufacturers regarding these costs, namely the “battery ownership model” and the “battery leasing model”. In the battery ownership model, the customer purchases the battery with the vehicle and supports the risks of replacement costs once the warranty on the battery is over. This model is mainly used in the quadricycles segment where lead acid is the standard battery. The battery ownership model is also used sometimes for light commercial vehicles. However, the standard battery in that segment is Lithium Ion which is more expensive but has a longer lifetime than lead acid. The customer faces therefore more important one-time costs with lithium Ion than with lead-acid batteries. In order to avoid the burden of a few important costs events, manufacturers propose also in that segment the battery leasing model. In that case, the battery is owned by the manufacturer and costs of battery are spread for the customer monthly through a renting system. When comparing both models in the light commercial vehicle segment, Figure 33 shows that the battery leasing model reduces the TCO compared to the battery ownership model. However, the sensitivity analysis on the years of ownership brings in the next section more insights into the impact of these two business models on the TCO.


- 77 -

Figure 33: Total cost of ownership for diesel, petrol and battery electric vehicles


- 78 -

The role of authorities When the different costs are aggregated, Figure 33 shows that the battery costs and high purchase costs of BEVs offset their low running costs advantage compared to conventional vehicles. However, the authorities of the Brussels-Capital Region provide a subsidy of maximum 5.000 euros to the purchase of freight electric vehicles. These incentives are shown to support effectively the competitive position of BEVs: the TCO of the battery electric vehicles is triggered down thanks to the subsidies. As a result, BEVs become more competitive than their petrol versions. A slightly better support however might improve further the competitive position of BEVs with diesel vehicles. The sensitivity analysis in the next section will show the subsidy levels required to support the competitiveness of BEVs compared to diesel vehicles.

4.4 Sensitivity analysis The results shown in the previous section are however limited to the context of the BrusselsCapital Region. The assumptions developed in the methodology section might also differ according to the transport operator or market changes. In order to generalize the findings, the authors relaxed these assumptions in a sensitivity analysis where we collected more insights on the evolution of the competitiveness position of BEVs. A first section of this sensitivity analysis show how the competitive position of BEVs can change with different utilisation of the vehicles. Then, the impact of market changes are presented in the second section. Finally, the last section is focused on the policies that the government can use to influence the competitiveness of BEVs. Breakeven points at current market conditions Sensitivity on the number of kilometres The results of the sensitivity analysis on the number of kilometres are shown in Figure 34 for the quadricycles and in Figure 35 for LCVs. In both segments, the diesel vehicles receive a lower TCO than the BEVs when the number of kilometres driven per year is low. But when the kilometres increase, the BEVs benefit from their lower operational costs and become more competitive than their conventional counterparts. The first electric vehicle to reach the breakeven point with a diesel vehicle in the quadricycles segment is around 16,000 kilometres per year. It becomes more competitive than all diesel vehicles when it drives a distance of about 18,000 kilometres per year. Since the other electric quadricycles are more costly, their breakeven points with the diesel quadricycles are located further: the most expensive electric quadricycle crosses the TCO of the most competitive diesel vehicle at around 25,000 kilometres per year. In the second segment, Figure 35 depicts more breakeven points since petrol vehicles are also analysed. Petrol vehicles have the lowest TCO when kilometres are small but become quickly less competitive compared to diesel vehicles. BEVs become competitive later. Their first breakeven points with a petrol vehicle can be located at a distance of about 7,500 kilometres per year. The last breakeven point between BEVs and petrol vehicles is located at around 18,000 kilometres per year. After that point, petrol is found to be the option with the highest TCO.

50

50

45

45

TCO (€urocents per km)



40 35 30 25 20 15 5,000

10,000

15,000

20,000

40 35 30 25 20 15 5,000

25,000

Number of kilometres per year per vehicle

50

45

45

35 30 25 20

15,000

20,000

25,000

Figure 35: TCO sensitivity on the kilometres driven per year by the LCVs (N1)

50

40

10,000

Number of kilometres per year per vehicle



Figure 34: TCO sensitivity on the kilometres driven per year by the quadricycles

- 79 -

40

35 30 25 20

15

15 0

2

4

6

8

10

12

14

16

Number of years of ownership per vehicle

0

2

4

6

8

10

12

14

16

Number of years of ownership per vehicle

Figure 36: TCO sensitivity on the years of ownership of the quadricycles

Figure 37: TCO sensitivity on the years of ownership of the LCVs (N1)

Legend quadricycles

Legend LCVs (N1)


- 80 -

Regarding the competitiveness with diesel vehicles, BEVs reach their first breakeven points later in the LCV segment than in the quadricycle segment: the yearly distance should be around 13.000 kilometres for the first BEV to become more competitive than a diesel vehicle. When analysing the breakeven points between similar vehicle versions, the Electric Kangoo is the first one to reach the TCO of its diesel counterpart (at around 18,000km/y), then the electric Partner (at around 20,000km/y) and finally the e-NV200 with battery leased (at around 23,500km/y). These breakeven points are significantly lower than what is reported in other journal papers. Feng and Figliozzi (2013) had indeed estimated a breakeven point at 45,000 kilometres per year for electric trucks. BEVs in the heavy goods vehicle segment might therefore compete less with their diesel counterparts than in the light commercial vehicles segment. Sensitivity on the years of ownership The results of the sensitivity analysis on the years of ownership are given in Figure 36 for the quadricycles and in Figure 37 for LCVs. We observe first important variations in the TCO of BEVs across time. They can be explained by the important costs that these vehicles face when the battery has to be replaced. These variations are less extreme but more frequent in the quadricycle segment than in the LCV segment. That difference is due to the type of batteries used. Lead acid batteries (used mostly for quadricycles) are less expensive but require more frequent replacements. Conversely, lithium ion batteries (used mostly in electric LCVs) require less frequent but more expensive battery replacements. Hence, BEVs with Lithium Ion battery show more extreme variations in the TCO according the years of ownership than BEVs with lead acid batteries. These variations are strengthened by the residual value of batteries that is considered to be null. The TCO of battery electric vehicles can therefore be optimised by operating them until the battery reaches its end life in transport applications (when the energy capacity of the battery is below 80% of its initial capacity). The sensitivity analysis shows indeed several minimum levels of the TCO across the years of ownership. Hence, we do not identify a minimum number of years of ownership as suggested by Davis and Figliozzi (2013). Instead, we recommend to sell the vehicle when the battery has to be replaced and the owner is not planning to use the BEV for another battery lifetime. These observations do not however apply on BEVs with a battery leasing model. Instead of supporting important sporadic battery costs, these are spread on a monthly basis. In that case, Figure 37 shows that it not possible to optimise the battery lifetime in order to reduce the TCO of these vehicles. As a result, they remain less competitive than diesel vehicles. Still, battery leasing model is shown to be the most interesting model for BEVs. That model remains indeed more competitive than the battery ownership model, even when the battery lifetime is optimised. In that regards, we should stress that a battery replacement every 6 years for Lithium Ion can be considered conservative. If the battery shows a longer lifetime, Figure 37 suggests that the TCO of the battery ownership model can indeed become more interesting than the battery leasing model. An interesting observation of the sensitivity of the TCO in function of the ownership period is the particularly low TCO in the first year depicted in Figure 36 and Figure 37 for BEVs. This effect can be attributed to the subsidies granted by the Brussels-Capital Region. If the cost of the first battery purchased with the vehicle is small enough, the subsidy can cover the battery costs


- 81 -

and a part of the depreciation of the vehicle. Authorities should therefore follow carefully subsidy demands in order to avoid abuses: transport operators could theoretically purchase a BEV, receive the subsidy, sell the vehicle after one year and purchase a new BEV with a new subsidy in order to operate vehicles with a low TCO. Breakeven points at future market conditions Sensitivity on the price of the battery Among the different market changes expected, studies forecast that the cost of batteries for BEVs will decrease significantly in the upcoming decades (Nykvist and Nilsson, 2015). To understand such a change on the competitiveness of BEVs, we analysed the TCO of the vehicles in function of decreasing battery prices. The sensitivity was found to be linear and the results are therefore presented in Table 5. Since conventional vehicles do not have batteries, their TCO remain unchanged. Regarding BEVs, their sensitivity follow a similar trend: a decrease of 1% in battery costs, reduce their TCO by around 0.07 eurocents per kilometre. If battery prices are divided by two in 2020 as expected by DELIVER (2012) and Electrification Coalition (2010), the TCO of battery electric vehicles should therefore be reduced by 3.5 eurocents. By comparing that reduction with the results of the TCO in Figure 33, we can observe that such an evolution would change the competitive position of BEVs compared to the diesel vehicles. Only the e-NV200 would need a further decrease in battery prices: the breakeven point of the battery leasing version with the diesel version is located at a decrease of about 70% of battery costs. On the other hand, the first BEV to become more competitive with its diesel counterpart is the Kangoo ZE in the LCV segment and the e-worker in the quadricycle segment when battery costs drop by about 25%. The Partner Electric becomes more competitive than its diesel counterpart after a cost reduction of about 30%. Sensitivity on the residual value from the battery The sensitivity analysis has also investigated the possibility to recover a residual value from the used batteries. There is indeed a potential for a second hand market of BEV batteries since 80% of the battery capacity remains when it reaches its end life for transport applications. Table 5 shows how the TCO results are influenced by different battery residual values. The effects are similar to the sensitivity on battery prices but slightly less important because of the real discount rate: the money invested in the battery cannot be used for other investments before the residual value is recovered. Still, the major difference between these two sensitivity analyses can be observed on the BEVs with a battery leasing model. These vehicles do not benefit from an increased residual value. Since the battery is owned by the manufacturer, the residual value is not recovered by the customer. As a result, the TCO is not reduced, unless the manufacturer shares the reduced costs of batteries with the customers through a lower battery leasing price. Sensitivity on energy prices An important factor influencing the TCO results is the fuel prices of petrol and diesel. The results in Table 5 show that petrol LCVs would be the most affected vehicles by rising fuel prices. Indeed, petrol vehicles have a higher consumption than diesel vehicles and the fuel price is also more expensive than diesel. On the other hand, BEVs are not influenced since they are operated with electricity. Hence, BEVs become more competitive with rising fuel prices. The first breakeven point of BEVs with their diesel counterpart is reached by the Kangoo ZE in the LCV


- 82 -

segment and by the e-worker in the quadricycle segment when fuel prices increase by about 20%. The electric Partner becomes competitive later with its diesel counterpart, when diesel prices increase by 35%. Finally, the e-NV200 (battery leasing) receives a lower TCO than its diesel versions after an increase of minimum 55% of diesel prices. But electricity prices might also change in the future. Hence, the sensitivity analysis explored also their impact on the TCO. Table 5 shows that the TCO of electric vehicles are entailed by an added cost of between 0.15 and 0.19 eurocents per kilometres when electricity price rise by 1 eurocent per kWh. This sensitivity on electricity prices shows the importance of charging BEVs at the best time of the day in order to optimise their TCO with the lowest electricity rates.

Table 5: Sensitivity analysis of TCO results on battery prices, fuel prices and residual value of the battery On battery prices (variation of TCO for every increase of 1%)

On residual value of the battery (variation of TCO for every increase of 1%)

(variation of TCO for every increase of 1%)

On electricity prices (variation of TCO for every increase of 1eurocents/kWh)

Diesel Mega D-truck

0

Diesel Ligier Flex

0

0

0.05

0

0

0.06

0

Electric Mega e-Worker

0.06

-0.06

0

0.15

Electric Goupil G3

0.07

-0.07

0

0.19

Electric Alke ATX210E

0.07

-0.07

0

0.14

Diesel Renault Kangoo Express

0

0

0.07

0

Diesel Peugeot Partner

0

0

0.07

0

Diesel Nissan NV200

0

0

0.07

0

Electric Renault Kangoo ZE (b. leased)

0.06

0

0

0.15

Electric Peugeot Partner

0.08

-0.08

0

0.18

Electric Nissan e-NV200 (b. leased)

0.06

0

0

0.16

Electric Nissan e-NV200

0.08

-0.08

0

0.16

Petrol Renault Kangoo Express

0

0

0.11

0

Petrol Peugeot Partner

0

0

0.13

0

Petrol Nissan NV200

0

0

0.14

0

On fuel prices

Breakeven points with market intervention Sensitivity on the subsidy level for BEVs In order to support the competitiveness of BEVs, a traditional instrument used by authorities is subsidies. The sensitivity analysis investigated the impact of different subsidy levels for BEVs on the TCO results. Table 6 shows that it influences every BEV to the same extent: an increase of 1,000 euros subsidy to BEVs decreases their TCO by 0.68 eurocents per kilometre. As a result, the regional subsidy level should be set at around 10,000 euros to have a similar effect than a decrease of the battery costs by 50% (a decrease of the TCO of 3.4 eurocents per kilometre). In that case, almost every BEVs were competitive with their diesel counterpart. The sensitivity analysis shows also that the current regional subsidy supports the competitiveness of the BEVs: without the subsidy of 5,000 euros, the TCO of electric vehicles would increase by


- 83 -

3.4 eurocents per kilometre. In such a case, petrol vehicles would become more competitive than BEVs. Under the current market conditions, subsidies are therefore required in order to support the competitiveness of BEVs with their conventional vehicles. Sensitivity on the deductibility rate for BEVs In Belgium, the corporate taxable base can be reduced by deducting 120% of costs related to the BEVs. However, that fiscal regime does not apply on vehicles designed for goods transport. In this sensitivity analysis, we investigated the impact of that incentive, extended to freight vehicles. We considered for that analysis a company making a profit between 1 and 25,000 euros, implying a corporate tax rate of 24.25% (Fisconetplus, 2012b). The sensitivity analysis evaluated the impact of different deductibility rate for BEVs on the TCO results. Table 6 summarises the results. Conversely to the subsidy scheme, the fiscal system is more flexible: more expensive BEVs are better supported. If the fiscal system applied on passenger cars was extended to freight vehicles, the TCO could be reduced by about 1.6 eurocents per kilometre. In that case, the Kangoo ZE and e-worker would become more competitive than their diesel counterparts. But higher deductibility rates are required to support the competitiveness of the other BEVs. The breakeven points of the Electric Partner and the e-NV200 (battery leased) with their diesel counterparts are located respectively at a deductibility rate of about 130% and 150%. Table 6: Sensitivity analysis of TCO results on the level of subsidies, BEV deductibility, city access toll, urban kilometre toll and road taxes. On subsidies

On BEV deductibility

On city access toll (variation of TCO for every increase of 1€ per day)

On urban kilometre toll (variation of TCO for every increase of 1€cents per km)

On BEV road taxes

(variation of TCO for every additional subsidy of 1000€)

(variation of TCO for every increase of 1%)

Diesel Mega D-truck

0

0

1.76

1.00

0

Diesel Ligier Flex

0

0

1.76

1.00

0

Electric Mega e-Worker

-0.68

-0.07

0

0

0

Electric Goupil G3

-0.68

-0.07

0

0

0

Electric Alke ATX210E

-0.68

-0.07

0

0

0

Diesel Renault Kangoo Express

0

0

1.76

1.00

0

Diesel Peugeot Partner

0

0

1.76

1.00

0

Diesel Nissan NV200

0

0

1.76

1.00

0

Electric Renault Kangoo (b. leased)

-0.68

-0.08

0

0

-0.01

Electric Peugeot Partner

-0.68

-0.08

0

0

-0.01

Electric Nissan e-NV200 (b. leased)

-0.68

-0.09

0

0

-0.01

Electric Nissan e-NV200

-0.68

-0.09

0

0

-0.01

Petrol Renault Kangoo Express

0

0

1.76

1.00

0

Petrol Peugeot Partner

0

0

1.76

1.00

0

Petrol Nissan NV200

0

0

1.76

1.00

0

(variation of TCO for every decrease of 1%)


- 84 -

Sensitivity on an urban toll fee Another way to support the competitiveness of BEVs in cities is to introduce a toll for accessing the urban area, except for zero emission vehicles. The sensitivity analysis investigated this option by introducing a daily fee in the TCO model. The variation in the TCO results are summarised in Table 6. We observe that the influence is the same for every conventional vehicle: their TCO increases by 1.76 eurocents per kilometre when the access fee increases by 1 euro per day. As a result, with a fee of 1 euro per day, the electric Kangoo ZE and the e-worker become more competitive than their diesel counterparts. The breakeven point of the Partner Electric with its diesel version is located at a fee of 1.5 euros per day and the electric e-NV200 (battery leased) at a fee of 2.5 euros per day. Sensitivity on a kilometre toll fee The toll can also be based on a kilometre basis rather than on a daily basis. Table 6 shows the results of the sensitivity analysis where a fee per kilometre has been added for conventional vehicles only. Since the fee is expressed in the same units than the TCO, the sensitivity analysis is straightforward: an increase of 1 eurocent per kilometre on conventional vehicles increases their TCO by 1 eurocent per kilometre. The first BEV to reach the breakeven point with their diesel counterpart is the Kangoo ZE in the LCV segment and the e-worker in the quadricycle segment when the kilometre toll is set at about 1.5 eurocents per kilometre. The breakeven points for the Partner electric and the e-NV200 (battery leased) are located at a respective fee of about 2.5 and 4 eurocents per kilometre. Sensitivity on road taxes Finally, the sensitivity analysis investigated the effects of a reduction of road taxes for BEVs. That advantage is not granted in the Brussels Capital Region but other cities like in London give an exemption to BEVs. The results of the analysis in Table 6 show that the segment of quadricycles is not impacted by this type of measure since they are already exempted from this tax scheme. A reduction of taxes reduces the TCO of battery electric vehicles only in the LCV segment. This effect is however shown to be limited: a full exemption of the road taxes for BEVs would reduce by 0.72 eurocents their TCO. That impact is similar to an additional subsidy of 1,000 euros for the purchase of BEVs.

4.5 Conclusion The results of the total cost of ownership analysis show that electric quadricycles and LCVs can be more financially attractive than their petrol counterparts for freight transport companies in the Brussels-Capital Region but are still more costly than diesel vehicles. They were found to benefit from low operating costs due to their efficient energy consumption, low energy prices and low maintenance costs. But these advantages are offset by the high purchase of the vehicle, the faster depreciation of BEVs and the additional battery costs. The subsidies provided to freight vehicles in the Brussels Capital Region limit however their TCO at a competitive level with conventional vehicles. As a result, electric quadricycles were found to be an affordable solution for transporters to switch from diesel LCVs to battery electric vehicles. These conclusions are however limited to the context of the Brussels-Capital Region. In order to generalize the findings, the authors conducted a sensitivity analysis so more insights are given on the evolution of the competitiveness position of BEVs. First, the influences of different vehicle


- 85 -

utilisations on the TCO were explored. The analysis shows that BEVs are more competitive with conventional vehicles when larger distances are driven per year. The breakeven points of BEVs with diesel vehicles are located between a distance of 16,000 and 25,000 kilometres per year. In a second sensitivity analysis, the evolution of the TCO in function of the years of ownership was investigated. The replacements of the batteries are found to play a critical influence on the costs which results in important variations in the TCO across the years of ownership. The analysis showed that the TCO of battery electric vehicles can be minimized when the ownership period of the vehicle corresponds to the full lifetime of the battery (in transport applications, end life of batteries is reached when its energy capacity is below 80% of its initial capacity). Hence, BEVs should ideally be sold when the battery has to be replaced and the owner is not planning to use the BEV for another battery lifetime. That observation does not however apply on BEVs with a battery leasing model since the battery costs are spread on a monthly basis and risks of battery replacement are supported by the manufacturer. The chapter explored also how the competitive position of BEVs will change with future market conditions. An important expected evolution is the reduction of battery prices. The sensitivity analysis estimated that a reduction of between 25% and 75% in battery prices is required in order for BEVs to be competitive with their diesel counterparts. Residual values might also be captured from used batteries thanks to the development of second hand applications of BEVs batteries. These residual values should be able to recover between 25% and 75% of the initial price in order to reduce the TCO of battery electric vehicles below the TCO of their diesel versions. Finally energy prices were investigated. Diesel prices need to increase between 20% and 55% in order for BEVs to reach the breakeven points with their diesel counterparts. On the other hand, the TCO of battery electric vehicles was found to be sensitive to electricity prices which showed the importance of charging at the best rates. However, given the current market conditions, BEVs require still government support. The current subsidies of the Brussels Capital Region is shown to support effectively their competitiveness with petrol vehicles but should be increased to 10,000 euros in order to support their competitiveness with diesel vehicles. The fiscal system applied on passenger cars could also reduce the TCO of BEVs in city logistics if it was extended to freight vehicles. A deductibility rate of 120% support the competitive position of some BEVs with their diesel counterparts but a rate of 150% is recommended to support every BEVs analysed in this chapter. Access fee to the city was also investigated. An urban toll between 1 euro and 2.5 euros or an urban kilometre tax between 2.5 and 4 eurocents per kilometre on conventional vehicles could support the competitive position of BEVs with their diesel counterparts. Finally, an exemption of road taxes for BEVs was found to have a limited impact on the TCO. A combination of these effects can easily lead to a situation where the BEVs have a lower TCO than diesel vehicles in a near future. Still, these conclusions cannot be extended to the whole segment of LCVs, the heaviest vehicles could not be analysed. Indeed, the comparison of the results with other studies suggested that BEVs compete less with their diesel counterparts in the heavier segments. But new models are coming on the market. New electric LCVs from Smith vehicles and Iveco for example are expected to be marketed soon in the heaviest segment of light commercial vehicles (3.5 tonnes). Future research should therefore analyse the costs of these larger LCVs. Future research should also integrate these results in a more global approach. The


- 86 -

competitive position of a vehicle is indeed not limited to the costs aspects. At a meta level, these results could be used in a fleet management model so operational constraints of range can be integrated for better decision making. At a macro level also, the private interests and public interests of the policies described in this chapter could be evaluated like in Melo, Baptista & Costa (2014). The relevance of such policies could then be discussed by comparing the investments needed for policy support of BEVs and their benefits on the urban environment. Finally, the environmental impact of the vehicles could be integrated in the TCO analysis in order to evaluate their eco-efficiency like in Messagie et al. (2013).


- 87 -

References Browne, M., Allen, J., Nemoto, T. and Visser, J., (2010). Light goods vehicles in urban areas. Procedia - Social and Behavioral Sciences, 2, 5911–5919. Browne, M., Allen, J. and Leonardi, J. (2011). Evaluating the use of an urban consolidation centre and electric vehicles in central London. IATSSResearch, 35, 1-6. Brugel. (2014). Observatoire des prix de l’électricité et du gaz en Région de Bruxelles-Capitale. Crainic, T.G., Ricciardi, N. and Storchi, G. (2004). Advanced freight transportation systems for congested urban areas. Transportation Research Part C, 12, 119–137. CREG (2013). Apercu et évolution des prix de l’électricité et du gaz naturel pour les clients résidentiels et les PME. Brussels. Dablanc, L. (2011). City distribution, a key element of the urban economy: guidelines for practitioners. In Macharis, C. and Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, 13-36. Dablanc, L. and Rakotonarivo, D. (2010). The impacts of logistics sprawl: How does the location of parcel transport terminals affect the energy efficiency of goods’ movements in Paris and what can we do about it? Procedia Social and Behavioral Sciences, 2, 6087–6096. Dablanc, L. (2007). La notion de développement urbain durable appliquée au transport des marchandises. Les Cahiers Scientifiques Du Transport, 51, 97–126. Davis, B. A. and Figliozzi, M. A. (2013). A methodology to evaluate the competitiveness of electric delivery trucks. Transportation Research Part E, 49, 8–23. DELIVER (2012). Deliverable D1.1 - Report on Technology, Market and Urban Logistics Roadmap from 2020 and Beyond. Aachen, Germany. EC (2011). White paper: Roadmap to a Single European Transport Area. Brussels. ECB (2015). Long-term interest rate statistics for EU Member States [WWW Document]. URL http://www.ecb.int/stats/money/long/html/index.en.html (accessed 08.14.15). Electrification Coalition (2010). Electrification Roadmap. Washington. Ellram, L.M. (1995). Total cost of ownership: an analysis approach for purchasing. International Journal of Physical Distribution & Logistics Management. 25 (8), 4–23. European Commission (2002). 2002/24/EC. Brussels. European Commission (2007). 2007/46/EC. Brussels. FEBIAC (2011). Evolution des immatriculations de véhicules utilitaires neufs. Brussels Federal Planning Bureau (2015). Perspectives économiques régionales 2015-2020. Brussels. Feng, W. and Figliozzi, M. (2012). Conventional vs Electric Commercial Vehicle Fleets: A Case Study of Economic and Technological Factors Affecting the Competitiveness of Electric Commercial Vehicles in the USA. Procedia Social and Behavioral Sciences, 39, 702–711.


- 88 -

Feng, W. and Figliozzi, M. (2013). An economic and technological analysis of the key factors affecting the competitiveness of electric commercial vehicles: A case study from the USA market. Transportation Research Part C, 26, 135-145. Fischer, M., Werber, M. and Schwartz, P. V. (2009). Batteries: Higher energy density than gasoline? Energy Policy, 37, 2639–2641. Fisconetplus (2012a). Codes des impôts sur les revenus 1992 - exercice d’imposition 2013 (revenus 2012) [WWW Document]. URL http://ccff02.minfin.fgov.be/KMWeb/document.do?method=view&nav=1&id=bdf4e90dfbb7-4216-a99f-db12efe9a34c&disableHighlightning=true#findHighlighted (accessed 10.20.12). Fisconetplus (2012b). Codes des impôts sur les revenus 1992 - exercice d’imposition 2013 (revenus 2012) [WWW Document]. URL http://ccff02.minfin.fgov.be/KMWeb/document.do?method=view&nav=1&id=4908858aa88b-4c30-832f-4ba4ee0db0ed&disableHighlightning=true#findHighlighted (accessed 10.20.12). GOCA (2015). Age and kilometers driven of Light Commercial Vehicles in Belgium. Brussels Lebeau, K. (2013). Electrifying cars : the economic potential of electric vehicles. Thesis. Vrije Universiteit Brussel, Brussels, Belgium. Lebeau, P., & Macharis, C. (2014). Freight transport in Brussels and its impact on road traffic. Brussels Studies, 80, 1– 14. Lee, D.-Y., Thomas, V.M. and Brown, M.A. (2013). Electric urban delivery trucks: energy use, greenhouse gas emissions, and cost-effectiveness. Environmental Science & Technology, 47, 8022–8030. Messagie, M., Lebeau, K., Coosemans, T., Macharis, C. and Van Mierlo, J. (2013). Environmental and Financial Evaluation of Passenger Vehicle Technologies in Belgium. Sustainability, 5 (12), 5020-5033. Martensson (2005). Volvo’s environmental stratgey for next generation trucks. Presented at BESTUFS Conference, June 2005, Amsterdam. Mearig, T., Coffee, N. and Morgan, M. (1999). Life Cycle Cost Analysis Handbook. State of Alaska, Department of Education & Early Development. Melo, S., Baptista, P. and Costa, Á. (2014). Comparing the Use of Small Sized Electric Vehicles with Diesel Vans on City Logistics. Procedia Social and Behavioral Sciences, 111, 350-359. Moniteur belge, 2009. C- 2009/31231. Moniteur belge, 2013. C- 2013/31464. Nesbitt, K. and Sperling, D. (2001). Fleet purchase behavior: decision processes and implications for new vehicle technologies and fuels. Transportation Research Part C, 9, 297–318. Nykvist, B. and Nilsson, M. (2015). Rapidly falling costs of battery packs for electric vehicles. Nature Climate Change, 5(4), 329-332. Omar, N. (2012). Assessment of Rechargeable Energy Storage Systems for Plug-In Hybrid Electric Vehicles. Thesis. Vrije Universiteit Brussel, Brussels, Belgium. PORTAL (2003). Inner Urban Freight Transport and city logistics. Quak, H. (2011). Urban freight transport: the challenge of sustainability. In Macharis, C. and Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK. SPF Finances (2015). Tarifs de la taxe de circulation camionnettes (jusque 3.500 kg).


- 89 -

Turcksin, L. (2011). Stimulating the purchase of environmentally friendlier cars: a socio-economic evaluation. Thesis. Vrije Universiteit Brussel, Brussels, Belgium. Van Amburg, B. and Pitkanen, W. (2012). Best Fleet Uses , Key Challenges and the Early Business Case for ETrucks : Findings and Recommendations of the E-Truck Task Force. Paper presented at EVS26, May 2012, Los Angeles. Van den Bossche, P., Vergels, F., Van Mierlo, J., Matheys, J. and Van Autenboer, W. (2006). SUBAT: An assessment of sustainable battery technology. Journal of Power Sources, 162 (2), 913–919. Van Vliet, O.P.R., Kruithof, T., Turkenburg, W.C. and Faaij, A.P.C. (2010). Techno-economic comparison of series hybrid, plug-in hybrid, fuel cell and regular cars. Journal of Power Sources, 195, 6570–6585. Van Mierlo, J. and Maggetto, G. (2005). Fuel cell or battery: electric cars are the future. Fuel Cells, 7 (2), 165–173. Zanni, A. M. and Bristow, A. L. (2010). Emissions of CO2 from road freight transport in London: Trends and policies for long run reductions. Energy Policy, 38 (4), 1774–1786.

Chapter 5: Diesel or/and battery electric vehicles: Which mixed technology fleet?

- 90 -

5 Diesel or/and battery electric vehicles: Which mixed technology fleet for city logistics?10 5.1 Introduction A number of trends can be observed in urban freight transport. The current urbanization process generates more freight volumes in cities, transport is increasingly fragmented due to the success of light commercial vehicles and distances are stretching out due to the delocalisation of logistics platforms to the periphery (Dablanc and Rakotonarivo, 2010). Because of these combined effects, vehicle-kilometres of freight vehicles are expected to increase in the future. However, urban freight transport is responsible for negative impacts on the sustainability of cities. These negative impacts can be attributed partly to the intense use of road. Vans and trucks have indeed a worse impact compared to other motor vehicles such as cars and motorcycles (Anderson et al., 2005). Even though road freight transport represents 10 to 15% of vehicle-kilometres in cities (Russo and Comi, 2012; Ségalou et al., 2006), freight vehicles are responsible for around 25% of CO2 emissions, 30% of NOx emissions, 40% of energy consumption and 50% of particles matter (Dablanc, 2011; Schoemaker et al., 2006). Also, noise nuisance caused by freight transport generates around five times more decibels than the circulation noise of private cars during morning rush hour (Ségalou et al., 2006). Recognizing the need for solutions, the European Commission has set the objective of reaching free CO2 city logistics in major urban areas by 2030 (EC, 2011b). 10

This chapter is an update of the following paper: Lebeau, P., De Cauwer, C., Van Mierlo, J., Macharis, C., Verbeke, W., and Coosemans, T. 2015. “Conventional, Hybrid, or Electric Vehicles: Which Technology for an Urban Distribution Centre?” The Scientific World Journal 2015, 11 p.


- 91 -

Research has developed a wide range of logistics concepts, regulations and technologies to fulfil the city logistics carbon free target (Quak, 2011). Among them, battery electric vehicles (BEVs) are considered to be an answer to the negative impacts listed above (Crainic et al., 2004). They have a particularly low environmental impact compared to conventional vehicles (Messagie et al., 2010). Several big companies such as DHL, UPS, DPD and Japan Post have already integrated BEVs in their fleet for last mile deliveries (Schneider et al., 2014). Still, purchase costs and limited battery capacity remain the two most important barriers for BEV adoption (Amburg and Pitkanen, 2012). They both contribute to the paradox of the BEV as depicted in Figure 38. On the one hand, they have to drive a high number of kilometres to be competitive with conventional vehicles. On the other hand, range is limited due to the battery capacity. As a result, BEVs fit in a specific niche. The objective of this chapter is to address these constraints by comparing the use of battery electric vehicles with conventional vehicles in a delivery fleet. Figure 38: The paradox of the battery electric vehicle

Source: own setup

Based on a vehicle routing problem (VRP) that we formulated, we study the case of an urban distribution centre. It will use the (real) case of a distributor which has a depot located in Brussels. The constraint related to the limited range defines the area of possible solutions. The optimal solution is then identified by the composition of the fleet that shows the lowest total cost.

5.2 Literature review The organisation of deliveries tours has generally been investigated through the Vehicle Routing Problem (VRP). It is defined as “the determination of the optimal set of routes to be performed by a fleet of vehicles to serve a given set of customers” (Toth and Vigo, 2002). The optimisation can have different objectives such as the minimisation of traveling time or delivery costs. The VRP can be described as a Travelling Salesman Problem (TSP) where more than one vehicle is used to serve each customer. The original VRP was introduced by Dantzig and Ramser (1959) and then developed in a large variety of more complex versions. Constraints on limited driving range were introduced with the works of Christofides et al. (1981). The VRP was designed such that a maximum cost could not be exceeded by the solution. The maximum costs parameter could be either be replaced by time constraints or distance constraints. This idea was also used by Laporte et al. (1982) where distance travelled by any vehicle could not exceed a defined upper bound. They named it the Distance Constrained Vehicle Routing Problem (DCVRP) and kept improving it in following works such in Laporte et al. (1985) and Laporte et al. (1987). The first attempts to investigate the specific characteristics of BEVs in a VRP was achieved by Gonçalves et al. (2011). They considered a VRP with Pickup and Delivery (VRPPD) and a mix fleet made of BEVs and conventional vehicles. The limited battery capacity was represented by a


- 92 -

time constraint on charging BEVs. The approach enriched the previous work on DCVRP as the distance constraints can be extended against a loss in time due to charging the battery. However, the location of charging spots were not considered in the model, meaning that BEVs could virtually recharge anywhere on the delivery round once the battery was empty. Erdoğan and Miller-Hooks (2012) brought a solution to the weaknesses of Gonçalves et al. (2011) by developing the Green VRP (G-VRP). He considers a network of refuelling stations that alternative fuelled vehicles can use during their delivery tour. They based their mix integer linear program formulation on the VRP with Satellite Facilities (VRPSF) from Bard et al. (1998). They translated the concept of satellites facilities where the cargo of vehicles can be reloaded or unloaded during the route into charging spots where vehicles can be refuelled during the route at specific points in the network. Erdoğan and Miller-Hooks (2012) showed therefore how to consider the location of refuelling in a VRP. Though, the charging time considered by Gonçalves et al. (2011) are missing in the GVRP as it was not designed specifically for BEVs but for alternative fuel vehicles (namely biodiesel, liquid natural gas or CNG vehicles). The contributions of Gonçalves et al. (2011) and Erdoğan and Miller-Hooks (2012) were integrated by Schneider et al. (2012) in their Electric Vehicle Routing Problem with Time Windows (E-VRPTW). Charging locations and charging times are both considered in their model which approaches well the problem of BEVs. In particular, they modelled charging time of BEVs as being a function of the state of charge of the battery. Plus, time windows and vehicle capacity restrictions are also included in the constraints of the E-VRPTW in order to adapt the model to the context of urban freight distribution. At the same time, Conrad and Figliozzi (2011) developed a solution close to Schneider et al. (2012). Based on a Capacitated Problem with Time Windows constraints (CRVRP-TW), they introduced the limited range and charging times in order to get the Recharging Vehicle Routing problem (RVRP). Their main difference is regarding charging locations: Conrad and Figliozzi (2011) consider that charging is possible at some customer locations while the formulation of Schneider et al. (2012) is more flexible as other possible charging locations are possible in the network. So far, these different papers considered VRP with a single type of vehicle. However, electric vehicles are likely to be used in delivery fleets with other kind of vehicles. A well-studied branch of the VRP literature is precisely addressing the problem of heterogeneous fleets in delivery fleets (Lin et al., 2014). Merging the VRP research on electric vehicles with the Fleet Size and Mix Vehicle Routing Problem (FSMVRPTW) is therefore relevant to come with recommendations to logistics decision makers. Van Duin et al. (2013) have been the firsts to develop this idea with their Electrical Vehicle Fleet Size and Mix Vehicle Routing Problem with Time Windows (EVFSMVRPTW). However, they approached the problem from the FSMVRPTW branch without considering the previous work on battery electric vehicles in VRP. As a result, the model was entailed with similar weaknesses than in Gonçalves et al. (2011): they do not consider the locations of charging points. A BEV with a battery swapping system is modelled so that the range of this BEV can be doubled. But the swapping system is not reflected in the constraints. It is in fact reflected in the range parameter of the vehicle which is simply doubled, meaning that the battery of the BEV can be swapped virtually anywhere on the road. Still, the main benefit of Van Duin et al. (2013) is to bring the Fleet Size and mix approach in the discussion of electric vehicle


- 93 -

routing problem. Hiermann et al., (2014) developed that idea further to propose an EFSMVRPTW that considers the decisions regarding the fleet composition and the choice of recharging times and locations. This work can be considered as the state of the art of delivery optimisation with BEVs. Table 7: Overview of literature review

(Christofides et al., 1981) (Laporte et al., 1985) (Gonçalves et al., 2011) (Conrad and Figliozzi, 2011) (Erdoğan and Miller-Hooks, 2012) (Van Duin et al., 2013) (Schneider et al., 2014) (Hiermann et al., 2014)

Charging

Energy consumption model



×

×

×



×

×

×



Fleet Size and Mix

Time windows

×

×

× 

× × 

× 

Range constraints



×



±

(location is not considered)

 



(charging does not depend on the state of charge)

±

× × × ×











×







×

(location is not considered)

Source: Own setup

Table 7 summarises the contributions of the different relevant papers. However, one aspect is forgotten in every paper. Every BEV specific papers assume the range to decrease linearly in function of the distance driven. However, the literature from engineering research recognises that range of BEVs is strongly influenced by other parameters than distance. Hayes, de Oliveira, Vaughan, & Egan (2011) show for example in their paper that the driving range for a specific BEV (Nissan Leaf) can change from 221 km in ideal driving conditions to 99km in bad conditions. In order to facilitate for consumers the comparison between BEVs’ performances, manufacturers have to show the range based on official drive cycles. In the United States, the EPA is used and shows a range of 121km for the Nissan Leaf (Blanco, 2013). In Europe, the NEDC is used and shows a range of 200 km for the Nissan Leaf (Crowe, 2013). Hence, range can change to a large extent depending on the usage of the vehicle. Two current works are integrating vehicle dynamics in a VRP to estimate variable ranges (Afroditi et al., 2014; Goeke and Schneider, 2014). However, more technical knowledge is required to develop an energy model. Auxiliaries for example were not considered even though they represent an important part of the energy consumption. Since Lin et al. (2014) identify precisely the lack of interdisciplinary approach for solving VRP problems, the objective of this chapter is to bring together the developments of the EVFSMVRPTW with real observations conducted on electric vehicles. The model we propose


- 94 -

considers therefore the different aspects shown in Table 7. The fleet size and mix vehicle routing problem considers different vehicle sizes with either electric propulsion or internal combustion engine. They mainly differ by their payload, fixed costs, running costs, energy available in the vehicle and their energy consumption. Charging operations are considered for battery electric vehicle at the depot with fast chargers. Finally, time windows are also considered as they are important to be considered in the context of city distribution. As a result, we call our formulation of the problem a Fleet Size and Mix Vehicle Routing Problem with Time Windows for Electric Vehicles (FSMVRPTW-EV).

5.3 Methodology The parameters influencing the range of BEV The range of an electric vehicle is determined by the amount of energy at disposal in the battery, and the energy consumption of the vehicle. The available total energy in the batteries for vehicles is called the battery capacity and the remaining amount of energy during use is called the State of Charge (SoC) and is expressed in percentage ‘charge’ remaining. The energy required at the wheels to drive a vehicle is determined by the vehicle dynamics. Based on El Baghdadi et al. (2013), we can express a theoretic energy consumption 𝐸𝑖𝑗 at the wheels for a distance 𝑑𝑖𝑗 using the vehicle dynamics described in equation (1). 𝐸𝑖𝑗 =

𝑑𝑏𝑖𝑗 1 [𝑚 . 𝑔. (𝜔. 𝑐𝑜𝑠𝜑 + 𝑠𝑖𝑛𝜑) + 0.0386. (𝜌. 𝜎. 𝜇. 𝑏𝑖𝑗 ²) + 0.2778(𝑚𝑖𝑗 + 𝑚𝑓 ). ] 𝑑𝑖𝑗 3600 𝑖𝑗 𝑑𝑡

(1)

Where 𝐸𝑖𝑗 = 𝑚𝑖𝑗 = 𝑚𝑓 = 𝑔= 𝜔= 𝜑= 𝜌= 𝜎= 𝜇= 𝑏𝑖𝑗 = 𝑑𝑖𝑗 =

Mechanical energy required at the wheels to drive on a distance 𝑑𝑖𝑗 [kWh] Vehicle mass [kg] fictive mass of rolling inertia [kg] Gravitational acceleration (9.81m/s²) Vehicle coefficient of rolling resistance[-] Road gradient angle [°] Air density (1.226kg/m3) Drag coefficient of the vehicle [-] Max. vehicle cross section [m²] Vehicle speed between the point 𝑖 and the point 𝑗 [km/h] Distance driven from point 𝑖 to point 𝑗 [km]

The first term of the formula assesses the rolling resistance due to the work of deformation on wheel from the contact with the road. It also considers the required potential energy for hill climbing. The second term assesses the aerodynamic drag (losses), which are heavily dependent on the shape of the vehicle and the driving speed. Finally, the third term considers the energy required for acceleration. By combining these three factors, we can estimate the theoretical force to move the vehicle. If we consider the distance on which this force is applied, we compute the energy required to move the vehicle. Driving the vehicle is however not the only source of energy consumption. Auxiliaries (AC, heating, …) represent another important part of energy consumption. Additionally, to deliver


- 95 -

energy from the battery to the wheels and auxiliaries, the energy is submitted to a number of conversion stages, each comprising energy losses, as estimated by De Vroey et al. (2013). On the other hand, the amount of available energy is increased by the ability of the electric vehicle to recover (regenerated) part of the kinetic (braking) or potential (hill descend) energy. Figure 39 shows how energy is transmitted from the grid to the wheels and how it is distributed between the drive train and auxiliaries in the final stage. Figure 39: Energy losses from plug to the wheels of battery electric vehicles

Source: De Vroey et al. (2013).

An accurate range model is therefore the combination of an accurate SoC estimation and an accurate energy consumption estimation. As energy consumption varies considerably with changing circumstances, a large impact on the vehicle’s range is expected. Therefore, a first energy consumption model using real-life BEV consumption measurements and the vehicle dynamics can result in a more realistic range estimation of BEVs in the VRP. Data collection and assumptions In order to model the energy consumption of BEVs in the FSMVRPTW-EV, we used real observations of the energy consumption of a Nissan Leaf collected from December 2012 until December 2013. For each trip, duration, distance and date were monitored. Most importantly, the energy consumed and recovered during the trip was registered. Since the car was shared between different drivers, a variety of driving behaviours could also be observed in our data. After filtering the sample, eliminating very short trips (less than 1km) and corrupted data, we kept 838 observations for analysis. Based on these data, we modelled the energy consumed through an ordinary least squares analysis. We considered the theoretical relationships described in the previous section to explain the energy consumption. Hence, before the ordinary least square analysis is conducted, the theoretical energy required by the electric drivetrain was estimated for each trip based on vehicle dynamics. But some information was missing in the description of the trips such as acceleration and road gradient. Also, energy losses when converting electrical energy to mechanical is missing.


- 96 -

This information should be however considered in the model through the error term or the coefficient 𝛽 of the ordinary least square analysis. We also assumed auxiliary consumption to be a function of time. Finally, since temperature affects the consumption of both auxiliary and drivetrain efficiency, an additional parameter is included for correctional purposes. As a result, we explain the energy of a trip according the function described in equation (2). 𝐄𝐧𝐞𝐫𝐠𝐲𝐂𝐨𝐧𝐬𝐮𝐦𝐩𝐭𝐢𝐨𝐧𝐭𝐫𝐢𝐩 = 𝛂 ∗ 𝐃𝐮𝐫𝐚𝐭𝐢𝐨𝐧𝐭𝐫𝐢𝐩 + 𝛃 ∗ 𝐄𝐢𝐣 + 𝛄 𝐓𝐞𝐦𝐩𝐞𝐫𝐚𝐭𝐮𝐫𝐞𝐭𝐫𝐢𝐩 + 𝛆

(2)

The non-linear least square analysis showed that the variable chosen were very significant with a p value of 2.39e-113, 9.42e-235, 5.44e-6 and 7.42e-5 for respectively 𝛼, 𝛽, 𝛾 and 𝜀. The model is also assessed to be excellent with a R² of 0.93. In Figure 40, we can see the distribution of the expected and the observed values. Figure 40 : Observed and expected energy consumption of electric vehicle trips

Source: own setup

As section above described, there is also a positive flow of energy with regenerative braking. In order to model the contribution of the regenerated energy, we developed also a model through an ordinary least squares analysis. We used the function described in equation (3) to explain the regenerative energy in function of duration of the trip, distance and the temperature: 𝐑𝐞𝐠𝐞𝐧𝐚𝐫𝐢𝐭𝐯𝐞𝐄𝐧𝐞𝐫𝐠𝐲𝐭𝐫𝐢𝐩 = 𝛅 ∗ 𝐃𝐢𝐬𝐭𝐚𝐧𝐜𝐞𝐭𝐫𝐢𝐩 + 𝛉 ∗ 𝐓𝐞𝐦𝐩𝐞𝐫𝐚𝐭𝐮𝐫𝐞𝐭𝐫𝐢𝐩 + 𝛑 ∗ 𝐃𝐮𝐫𝐚𝐭𝐢𝐨𝐧𝐭𝐫𝐢𝐩 + 𝛕

(3)

Each variable was assessed to be very significant with a respective p-value of 2.73e-33, 3.59e-8 and 3.97e-45. The constant error term had also a very low p value with 2e-8. Figure 41 shows the distribution of observed and expected value of the regenerated energy. The R² is 0,77.


- 97 -

Figure 41: Observed and expected regenerated energy of electric vehicle trips

Source: own setup

By combining both models in the FSMVRPTW-EV, we can estimate the depletion of the battery in function of the route taken by the BEV. The energy capacity of the battery vehicle is reduced by its specific consumption on each route. Let us note that this specific consumption considers the average driving behaviour from our observations. The FSMVRPTW-EV The formulation of the Fleet Size and Mix Vehicle Routing Problem with Time Windows for Electric Vehicles (FSMVRPTW-EV) is based on the FSMVRPTW of Belfiore and Yoshizaki (2013) and the G-VRP of Erdoğan & Miller-Hooks (2012). It is defined on a complete and directed graph 𝐺 = (𝑉, 𝐴) . 𝑉 denotes a set of vertices with 𝑉 = 𝐶 ∪ {𝑣0 } . 𝐶 is a set of 𝑛 customers with 𝐶 = {𝑣1 , 𝑣2 , … , 𝑣𝑛 } and {𝑣0 } stands for the depot. Then, set 𝐴 represents the set of arcs connecting the vertices of 𝑉, with 𝐴 = {(𝑣𝑖 , 𝑣𝑗 )|𝑣𝑖 , 𝑣𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗}. Each arc (𝑣𝑖 , 𝑣𝑗 ) is associated with a distance 𝑑𝑖𝑗 , a speed 𝑏𝑖𝑗 and a traveling time 𝑡𝑖𝑗 . 𝑃 represents a set of 𝑘 vehicles with 𝑃 = {𝑝1 , 𝑝2 , … , 𝑝𝑘 }. They are either BEVs or conventional vehicles. Hence, they have different properties. They differ according their fixed costs 𝑓𝑘 , their running costs 𝑔𝑘 , their payload 𝑚𝑘 and their volume capacitiy 𝑎𝑘 . They differ also according 𝑘 their energy capacity 𝑧𝑘 and the energy consumption ℎ𝑖𝑗 they spend to travel from 𝑣𝑖 to 𝑣𝑗 . The energy capacity is reduced with 10% of the announced battery capacity to take into account the maximum depth of discharge. The energy consumption is based on the range model presented in the previous section. On the other hand, the vehicles share common characteristics. The driver


- 98 -

cost 𝑐 (€/hour) remains the same across vehicles. Vehicles start and end at the depot 𝑣0 . They travel in the directed graph 𝐺 so that the demand of every customer is fulfilled. Demand is described both in terms of volume with 𝑞𝑛 and in terms of weight with 𝑜𝑛 . Each vertex of 𝐶 is associated with a time window [𝑙𝑖 , 𝑢𝑖 ] and a service time 𝑠𝑖 . Deliveries cannot start before 𝑙𝑖 and after 𝑢𝑖 but can end after 𝑢𝑖 given the service time 𝑠𝑖 . Once the vehicle 𝑘 has come back to the depot, the used vehicle can do additional routes and become vehicle 𝑘′. Hence, 𝑃′ denotes the set of used vehicles. Vehicles in 𝑃′ have a fixed cost of zero since it has already been considered in their first route but they are available later. Let us note that recharging is possible for BEV at the depot only. BEV can fast charge with a power 𝑟 during their loading/unloading operations that we consider set at 50kW. In order to optimise the lifetime performance of the batteries, they can reach a maximum state of charge of 80% of their initial capacity. We assume in our model that a fast charger is always available at the depot. The objective of the FSMVRPTW-EV is to minimise the total costs of fulfilling the demand of 𝑘 customers within their time windows. The binary decision variables 𝑥𝑖𝑗 |𝑘 ∈ 𝑃, 𝑖, 𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗 𝑘 represent the resulting route followed by the vehicles such that 𝑥𝑖𝑗 equals 1 if the arc (𝑖, 𝑗) is

travelled and 0 otherwise. Besides, vertices are associated with additional decision variables: 𝑒𝑖𝑘 shows the available energy of vehicle 𝑘 at customer 𝑖, 𝑎𝑖𝑘 gives the available volume capacity of vehicle 𝑘 at customer 𝑖, 𝑚𝑖𝑘 denotes the available payload of vehicle 𝑘 at customer 𝑖 and 𝑤𝑖𝑘 gives the time of arrival of vehicle 𝑘 at customer 𝑖. Indices and sets 𝑖, 𝑗 = Vertex indices; 𝑉 = Set of all vertices with 𝑉 = 𝐼 ∪ {𝑣0 }; 𝐶 = Set of 𝑛 customers with 𝐶 = {𝑣1 , 𝑣2 , … , 𝑣𝑛 }; 𝐴 = Set of arcs with 𝐴 = {(𝑣𝑖 , 𝑣𝑗 )|𝑣𝑖 , 𝑣𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗}; 𝑃 = Set of 𝑘 vehicles with 𝑃 = {𝑝1 , 𝑝2 , … , 𝑝𝑘 }; Parameters 𝑣𝑛 = 𝑣0 = 𝑞𝑛 = 𝑜𝑛 = 𝑙𝑛 = 𝑢𝑛 = 𝑠𝑛 = 𝑑𝑖𝑗 = 𝑏𝑖𝑗 = 𝑡𝑖𝑗 = 𝑝𝑘 = 𝑎𝑘 = 𝑚𝑘 = 𝑒𝑘 = 𝑔𝑘 = 𝑓𝑘 =

The customer 𝑛 with 𝑣𝑛 ∈ 𝐼; Depot with 𝑣0 ∈ 𝑉; Volume of goods to be delivered at customer 𝑛 (m³); Weight of goods to be delivered at customer 𝑛 (kg); Lower bound of the time windows for customer 𝑛 (h - time); Upper bound of the time windows for customer 𝑛 (h - time); Service time to deliver customer 𝑛 or to load the vehicle if 𝑛=0 (h - duration); Distance from node 𝑖 to node 𝑗 (km); Speed limit between node 𝑖 and node 𝑗 (km/h); Time of travel between node 𝑖 and node 𝑗 (h - duration); The vehicle 𝑘 with 𝑝𝑘 ∈ 𝑃; Volume capacity of the vehicle 𝑘 (m³); Payload of the vehicle 𝑘 (kg); Maximum available energy capacity of the vehicle (kWh);; Running costs of vehicle 𝑘 (€/km); Fixed cost of vehicle 𝑘 (€);


- 99 -

𝑐 = Cost for the driver (€/h); 𝑟 = Charging power at the depot (kW); Variables 𝑤𝑖𝑘 = 𝑒𝑖𝑘 = 𝑘 ℎ𝑖𝑗 = 𝑘 𝑎𝑖 = 𝑚𝑖𝑘 = 𝑥𝑖𝑗 =

Arrival time of vehicle 𝑘 at node 𝑖 (h - time); State of charge of vehicle 𝑘 at node 𝑖 (kWh); Energy consumed by vehicle 𝑘 from node 𝑖 to node 𝑗 (kWh); Goods volume being transported by vehicle 𝑘 at node 𝑖 (m³); Goods weight being transported by vehicle 𝑘 at node 𝑖 (kg); Binary variable to 1 if the arc (𝑖, 𝑗) is travelled, 0 otherwise.

The formulation of the FSMVRPTW-EV can be expressed as the following mixed-integer program: Minimize 𝑘 ∑ ∑ 𝑓𝑘 𝑥0𝑗 + ∑ 𝑘∈𝑃 𝑗∈𝐶

𝑘 𝑘 ∑ ∑ 𝑑𝑖𝑗 𝑥𝑖𝑗 𝑔𝑘 + ∑

𝑖∈𝑉,𝑖≠𝑗 𝑗∈𝑉,𝑖≠𝑗 𝑘∈𝑃

𝑘 ∑ ∑ 𝑡𝑖𝑗 𝑥𝑖𝑗 𝑐

(4)

𝑖∈𝑉,𝑖≠𝑗 𝑗∈𝑉,𝑖≠𝑗 𝑘∈𝑃

Subject to 𝑘 ∑ 𝑥0𝑗 =1

∀𝑘 ∈ 𝑃

(5)

𝑗∈𝐶 𝑘 ∑ 𝑥𝑖𝑗 − ∑ 𝑥𝑗𝑖𝑘 = 0 𝑖∈𝑉,𝑖≠𝑗

∀𝑗 ∈ 𝑉, ∀𝑘 ∈ 𝑃

(6)

𝑖∈𝑉,𝑖≠𝑗

𝑘 𝑎𝑘 ≥ 𝑎𝑗𝑘 − 𝑞𝑗 + 𝑀𝑖𝑗 (1 − 𝑥𝑖𝑗 ) ≥ 𝑎𝑖𝑘 ≥ 0 ∀𝑘 ∈ 𝑃, ∀𝑖, 𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗 𝑘 𝑚𝑘 ≥ 𝑚𝑗𝑘 − 𝑜𝑗 + 𝑀𝑖𝑗 (1 − 𝑥𝑖𝑗 ) ≥ 𝑚𝑖𝑘 ≥ 0 ∀𝑘 ∈ 𝑃, ∀𝑖, 𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗 𝑘 𝑘 𝑒𝑘 ≥ 𝑒𝑖𝑘 − ℎ𝑖𝑗 + 𝑀𝑖𝑗 (1 − 𝑥𝑖𝑗 ) ≥ 𝑒𝑗𝑘 ≥ 0 ∀𝑘 ∈ 𝑃, ∀𝑖, 𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗 𝑘 𝑤𝑖𝑘 + 𝑠𝑖 + 𝑡𝑖𝑗 − 𝑀𝑖𝑗 (1 − 𝑥𝑖𝑗 ) ≤ 𝑤𝑗𝑘 ∀𝑘 ∈ 𝑃, ∀𝑖, 𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗 𝑙𝑖 ≤ 𝑤𝑖𝑘 ≤ 𝑢𝑖 ∀𝑖 ∈ 𝑉, ∀𝑘 ∈ 𝑃 𝑒0𝑘 ≤ (1 − 0.10) ∗ 𝑒𝑘 ∀𝑘 ∈ 𝑃, ∀𝑖 ∈ 𝑉, 𝑒0𝑘 ≤ 𝑠0 ∗ 𝑟 ≤ (0.80 − 0.10) ∗ 𝑒𝑘 ∀𝑘 ∈ 𝑃′, ∀𝑖 ∈ 𝑉, 𝑘 𝑥𝑖𝑗 ∈ {0,1} ∀𝑘 ∈ 𝑃, ∀ 𝑖, 𝑗 ∈ 𝑉, 𝑖 ≠ 𝑗

(7) (8) (9) (10) (11) (12) (13) (14)

Equation (4) represents the objective function. It expresses the total costs associated with a solution of the FSMVRPTW-EV. The first term considers the fixed costs of vehicles leaving the depot. The second term computes the running costs of the vehicles. Finally the last term considers the staff costs to operate the vehicle. Constraints are given by equation (5) to (14). Constraint (5) and (6) guarantee that vehicles start from the depot, visit the customers and come back to the depot. Constraint (6) in particular ensures the conservation of flow by forcing the equality between the number of arrivals and the number of departures at each vertex. Constraints (7) and (8) guarantees that the vehicle capacity is not exceeded in terms of volume and weight. They also track their reduction through the deliveries. Mij is a sufficiently large number. Constraint (9) guarantees that the available energy is always positive and tracks the battery depletion through the route. Constraint (10) sets a


- 100 -

minimum arrival time for vehicle 𝑘 arriving at customer 𝑗. Constraint (11) ensures that customers are visited during their time windows and that vehicles are operated during the opening hours of the depot. Constraints (12) and (13) consider the maximum depletion of the battery. In particular, constraint (13) considers fast charging possibilities at the depot. Finally, constraint (14) ensures the binary integrality. The Algorithm The algorithm we developed to solve the FSMVRPTW-EV is based on the savings heuristic (Clarke and Wright, 1964). First, an initial solution is built with each shop being delivered by one route. Subsequently the solution is iteratively improved by searching for potential savings by merging two or more routes. In order to address the constraints of the FSMVRPTW-EV some adjustments have been made to the original savings heuristic algorithm. When searching for potential savings, the adjusted algorithm limits the computation of potential savings to the fifteen closest shops to the last shop inserted in the route in order to reduce computation times. Moreover, as proposed by Bräysy, et al. (2008), the idea of the insertion-based heuristic is used. When merging an initial route (thus delivering to only one shop) with the route being improved, each possible position of the shop in the route is considered. The position of the shop showing the largest savings is selected and another initial route is investigated for merging. In other words, the shop to be inserted is the one showing the largest savings when located at the best position in that route. Once the route is completed, the algorithm verifies again if remaining savings are possible by reordering a last time the shops in the route. Bräysy et al. (2008) also recommend to use an insertion sequence for customers. The algorithm is more effective when the most critical shops are investigated first (i.e. shops with short time windows). The algorithm starts to build routes around the most difficult shops to insert. If these would be inserted later, once most shops are already included in the constructed routes, there might be no possible insertion left for that shop. The algorithm would then create a single route delivering only to that critical shop which is not cost effective. The insertion sequence is therefore based on a criticality index that considers for each shop the start of the time window, the time window duration and the distance from the depot. Shops that are inserted first are shops with a limited time window, starting early and located closer to the depot. In order to consider the different combinations of vehicles, the algorithm uses a tree. In each branch, one of the vehicles is selected. According the heuristic described above, a maximum of shops are inserted in the route, given the selected vehicle. The resulting node of that branch shows the number of shops that still need to be delivered. The branch is then developed further in sub branches according the number of available vehicles until all the shops are delivered. We have then a possible solution. Still, if the total cost of a branch becomes higher than the total cost of the cheapest completed branch, then the branch is not further explored. Finally, once the tree has investigated the different vehicle combinations, the branch with the lowest cost is considered as the optimal solution.


- 101 -

Application Based on the algorithm we developed, we can explore the paradox of BEVs in a city logistics context. We applied the FSMVRPTW-EV on a real case of the Brussels-Capital Region. It is a distribution firm that operates from an urban depot and distribute 681 shops located in the city and in the periphery early in the morning. The loading of the vehicles at the depot starts from 3am and is assumed to take 45 minutes. The operations face often short time windows as shops prefer to receive goods before first customers arrive. The shops are therefore described in terms of time windows and demand (expressed both in volume and weight). The OD-matrices between each shop are also given in terms of time and distance. They were generated based on the location of the shops and the fastest route given by Google maps in free flow conditions. Finally, we assumed a delivery time of 3 minutes at each shop. Out of these 681 shops, we selected randomly a set of 30 shops in order to keep reasonable computation times in our analyses. We considered also different set of vehicles. They are based on 6 different vehicles that represent the different vehicle types used in urban freight transport. We consider the segments of the small vans (volume of 3m³), large vans (volume of 8.5m³) and trucks (volume of 26m³). For each segment, diesel and electric technologies are represented. Their characteristics are based on the following vehicles. The small van segment uses the electric and diesel versions of the Renault Kangoo Express. The large van segment is represented by the diesel Mercedes Sprinter and the electric Smith Edison. Finally, the truck segment uses two 7.5 tonnes: the diesel Fuso Canter and the electric Smith Newton.

5.4 Results The results of the FSMVRPTW-EV are given in Figure 42. It shows how the different shops are delivered from the depot. In order to deliver the 30 shops at a minimal cost within their time windows, three routes need to be operated with three different vehicles. The green road is achieved by a large diesel van. It delivers 14 shops in 4hours and 9min and drives a distance of 114km. The black road is also achieved by a large diesel van. It delivers 9 shops in 2 hours and 38 minutes and drives a distance of 61 km. Finally, the blue road is operated by a small electric van. It delivers 7 shops in 2hours and 46 minutes. The electric van drives a distance of 80km which consumes about 11kWh. It accounts for 54% of its available energy. In total, the costs to deliver the 30 shops are estimated to account for a minimum daily cost of 426€ and the total distance is 255km.


- 102 -

Figure 42: Results of the FSMVRPTW

If the vehicles available for the last mile in the depot are limited to diesel vehicles only, the organisation of the deliveries does not change. The blue road operated by the small electric van is simply replaced by a small diesel van. However, the costs of the last mile do change. Because the choice of vehicles is more limited, the fleet composition is less flexible. The costs of the optimal fleet increases then to 426€. The blue road meets indeed the lower distance bound and the upper distance bound of the BEV paradox: the distance is sufficiently low for a BEV to operate the road within its range constraints and the distance is sufficiently high for a BEV to offset the high fixed costs with its low variable costs. As a result, the BEV was selected in the optimal mixed technology fleet. Still, the cost reduction it brings is very limited: it improves the diesel only solution by - 0.02%. If the vehicles available for the last mile in the depot are limited to battery electric vehicles, we can observe that the organisation of the deliveries does change. The total distance of the last mile increases indeed from 255km to 278km. Although, the large diesel vans are simply replaced by their electric counterparts, the additional constraints from BEVs impose a more complex organisation of the deliveries which can result in total longer distances. In this case though, we observe that range is not the main constraint but it is rather the more limited payload of BEVs. Because of the additional battery weight compared to the diesel large van, freight needs to be more balanced across the fleet. Some shops that demand a too heavy delivery have to be reassigned in other routes where free capacity is still available. Since that route will be less optimal


- 103 -

in order to meet the constraints of the BEVs, the total distance will be higher. Together with the more difficult competitive position of BEVs, the higher distance contributes to increase the costs of the last mile. The total cost of the optimal solution is 478€ which represents an added cost of 12.36% compared to the mixed technology fleet. Still, we can stress that we do find a feasible solution with a fleet operating only BEVs. The new green road is delivering 15 shops in 4hours and 10 minutes, the new black road is delivering 8 shops in 2 hours and 53 minutes and the new blue road is delivering 7 shops in 2 hours and 49 minutes. The large electric vans operating the new green and the new black roads consumed respectively for 78% and 35% of their available battery capacity and the small electric operating the new blue road consumed for 66% of its available battery capacity.

5.5 Sensitivity analysis In order to evaluate in what kind of environments battery electric vehicles are the most suitable, we conducted a sensitivity analysis based on the instance used in previous section. We refer to it as the baseline scenario and we analyse in this section the modification we brought to this scenario. Time-windows Time windows are often used by authorities to regulate the deliveries within a specific area. But time windows can also depend on the business model of the transport firm. Some transport firms leave the choice of delivery hours to the receivers which can be an added value for them. But since it imposes more constraints to the transporters, this added value comes with additional costs. In the baseline scenario, shops require to be delivered before a given hour. It ranges between 4:30am and 2:10pm with an average around 7:00am. In this sensitivity analysis, we uniformed the time windows in order to better explore their impact on the costs of deliveries. Since the vehicles are assumed to leave the depot at 3h45 the earliest, a time window of 1h will assume that shops have to be delivered before 4h45. The impact of a time window of 2h, 3h and 5h are also explored. The results are depicted in Figure 43. They show that costs of deliveries decrease when time windows become larger. Indeed, with short time windows, the operator needs to use more vehicles in order to satisfy the demand faster: when the time window is 1 hour, the optimal solution uses seven different vehicles to satisfy the demand. Since freight is more fragmented, vehicles are also smaller: six small vans and one large van are used. When the time window is extended of an hour, deliveries can be more consolidated and the number of vehicles can be reduced to four: two large vans and two small vans are selected in the optimal fleet. With one more extra hour, one small van can be saved. Finally, when the time window lasts for five hours, the deliveries can be achieved by one truck and one small van. In most of the cases, these solutions do not involve BEVs. Although operating the routes with BEVs is feasible, freight flows are too fragmented when time windows last for one hour. Distances per vehicle are too low which does not allow BEVs to reach their competitive threshold with diesel vehicles: distances per vehicle are indeed limited to a maximum of 71km per day in this case. But when time windows last for two hours, each vehicle has the time to deliver more shops which results in longer distances per vehicle. We observe then in that instance that one electric small van is selected in the optimal solution. But the savings brought by the mixed


- 104 -

technology fleet compared to the diesel only fleet is very small: it reduces the total cost of the last mile by - 0.02%. These savings are limited because the route that the BEV operates has a distance of 80km. It is just above the lower distance bound of the BEV paradox where they become more competitive than diesel. When time windows are more extended however, diesel becomes again the only type of fuel used: with reduced constraints from time windows, freight can be more consolidated in larger vehicles where diesel is often the most competitive technology. Hence, a mixed technology fleet can contribute to reduce the costs of the last mile when moderate time windows are enforced. On the other hand, diesel remains the most competitive vehicle technology when delivery times are flexible. Figure 43: Impact of time windows on costs of the last mile 700

Last mile costs (€/day)

600 500 400 300 200 100 0 1h Mixed technology fleet

2h

3h Diesel only fleet

5h BEV only fleet

This sensitivity analysis highlights however the potential of extensions of time windows for BEVs. When comparing the last mile costs of a BEV only fleet with a time window of 2hours and the last mile costs of a diesel only fleet with a tighter time window of 1hour, the BEV only fleet becomes much more interesting for transport operators: costs would be reduced by - 19.5% with a BEV only fleet. However, that reduction becomes smaller when the time windows for diesel vehicles get larger. Local authorities have here a powerful regulation that can stimulate city logistics to adopt BEVs. Congestion Efficiency of the urban transport system is often entailed by congestion. In this sensitivity analysis, we increased travel times of the baseline scenario by 150%, 200%, 250% and 300%. Results in Figure 44 show that costs increase with a rising congestion. The effect is in fact quiet similar to time windows: since travel times increase, time windows of shops become more constraining for the routing. As a result, freight flows become also more fragmented with increasing congestion levels. Still, the main difference with time windows is the much more important impact of congestion on last mile costs due to the longer travel times. When congestion increases by 150% compared to the baseline scenario, the number of vehicles remains the same in the mixed technology fleet but one of the two large vans has to be replaced by a truck. Indeed, the increasing time constraints require a new routing of the vehicles to meet


- 105 -

the time windows which reduce the possible consolidation of freight. In the case of a BEV only fleet however, the new organisation of the routes requires an additional vehicle in the fleet: the new fleet is composed of three small vans and one truck. BEVs have indeed a lower payload than diesel vehicles given the extra weight of the battery. Also, the increasing time constraints reduce the possibility to better balance deliveries in the remaining capacity of the fleet. Hence, an extra vehicle is used which increases costs of the BEV only fleet. When congestion doubles, the fleets increase in a similar way across the type of technologies considered: they all need three small vans and two large vans to meet the time windows of the shops. When travel times are further increased by 250% compared to the baseline scenario, the number of vehicles used in the fleet remains the same but one more route needs to be operated: one of the small vans has the time to operate two different roads. Finally, when travel times are tripled, no solutions can be found since at least one shop cannot be delivered within its time windows. Figure 44: Impact of congestion on costs of the last mile 1200


1000 800 600 400 200 0 Baseline

150%

Mixed technology fleet

200% Diesel only fleet

250%

300%

BEV only fleet

Like with time windows, a fragmentation of the flows implies lower distances per vehicles. The vehicles added to the optimal mixed technology fleet because of increased congestion are therefore mostly diesel vehicles. But the electric small van of the baseline scenario remains in the optimal mixed technology fleet across these scenarios. The BEV contributes to reducing the costs of the last mile compared to a diesel only fleet even though savings are here also incremental: they range between a cost reduction of - 0.02% and - 0.30%. Still, that result needs to be interpreted within the limitations of the model. Indeed, the vehicle routing problem considers constant fuel consumptions from diesel vehicles. In congestion however, their consumption is higher. Hence, this sensitivity analysis might underestimate the costs of diesel vehicles and thereby overestimate the number of diesel vehicles in the optimal mixed technology fleet. The lower distance bound of the BEV paradox might be too high. This sensitivity analysis highlights another powerful regulation that local authorities could use to stimulate the adoption of BEVs in fleets of city logistics operators. The authorised use of priority lanes for electric vans and trucks could indeed be more convincing than extension of time


- 106 -

windows. Since operators can use less vehicles and save time when they face lower levels of congestion, such a regulation would reduce the last mile costs of BEV fleets by - 21.4% compared to a diesel only fleet that faces travel times that are 150% more important. Shop density The concentration of the shops to be delivered around the urban consolidation centre can also influence the costs of deliveries with decreasing distances and therefore decreasing travel times. Figure 45 shows indeed that a concentration of the shops in the Brussels-Capital Region is less costly than the baseline scenario where shops are sometimes located also in the periphery (outside the Brussels-Capital Region) as shown in Figure 42. On the other hand, the baseline scenario is less costly than a scenario where all the shops are located in the periphery, far from the urban distribution centre. Figure 45: Impact of shops density on costs of the last mile 600


500 400

300 200 100 0 Urban Mixed technology fleet

Baseline Diesel only fleet

Periphery BEV only fleet

In the urban scenario, time to drive between each customer is smaller. As a result, two vehicles are sufficient to deliver the customers within their time windows. Since the fleet is reduced, larger vehicles are however required: a truck is replacing both the small van and one of the two large vans of the baseline scenario. As a result, the optimal mixed technology fleet does not include a BEV and have the same costs than the diesel only fleet: their last mile costs 274€ per day. In the segment of larger vehicles, BEV competes indeed less with conventional vehicles than in the segment of smaller vehicles. The cost of the last mile operated by a BEV only fleet increases therefore to 299€ per day which represents an added cost of 9%. In the periphery scenario, the moderate distances offer a more interesting environment for BEVs: distances should be large enough to fully benefit of the lower running costs of BEVs but low enough to be compatible with range constraints. Still, the fleet does not change compared to the baseline scenario. It involves two large diesel vans and one small electric van that drive a total distance of 374km. If the fleet is limited to diesel or BEVs, the organisation of the route does not change and the total distance remains the same. But in the electric scenario, the fleet needs to be adapted. One of the two large vans is replaced by an electric truck. That replacement is however not achieved because of a too limited payload. The truck is preferred because it has a larger


- 107 -

battery which allows to cover the 171km required by the first route (it has a battery of 80kWh while the large electric van has a battery of 40kWh). That long route requires 91% of the available energy capacity of the electric truck. The replacement of a large van by a truck for the electric solution contributes to increasing the total cost of the last mile. It is entailed by an added cost of 7.83% compared to the mixed technology fleet. We should stress here a second limitation of the model. The energy model we used could not indeed estimate the influence of acceleration on the energy consumption. Instead, the ordinary least squares analysis considered that effect within the error term. However, these energy consumption data we analysed where collected on a Nissan Leaf. Since acceleration on larger vehicles demands more energy, the range from larger vehicles might be overestimated. Still, regenerative braking should reduce that effect. Larger vehicles can also recover more energy when decelerating. Nevertheless, this choice for larger vehicles because of limited battery capacity on smaller vehicles stresses the importance for manufacturers to offer different size of batteries to the customers. This way, the operator can drive longer routes without purchasing a larger vehicle. Or conversely, he can operate shorter routes with a battery electric vehicle that has a smaller battery. The financial constraint can then be reduced. Demand We tested also the influence of the demand on distribution costs from an urban consolidation centre. The type of freight to be delivered can indeed change to a large extent given the diversity of receivers in a city. We modified the weight and the volume of the goods demanded by each shop according to the following percentage of the baseline scenario: 25%, 50% and 200%. The results are pictured in Figure 46. Figure 46: Impact of the demand on costs of the last mile 700


600 500 400 300 200 100 0 25% Mixed technology fleet

50%


200% BEV only fleet

The results show that a change in the size and weight of freight has a much more limited impact on the costs of the last mile than the effects presented in the previous sensitivity analyses. Indeed, the number of vehicles remains the same across the scenarios because deliveries are still constrained by the different time windows. Three vehicles still achieve the distribution. But the sizes of the vehicles do change with the demand. In the 25% scenario, only small vans are used.


- 108 -

In the 50% scenario, the fleet is made of a large van and two small vans. In the baseline scenario (100% scenario), two large vans are used with one small van. And finally, one truck is used with a large van and a small van when the demand has doubled (200% scenario). In terms of technology, we observe a preference for BEVs in scenarios where the weight and size of freight are lower than in the baseline scenario. The optimal mixed technology fleet is indeed made of two BEVs and one diesel vehicles. Also, BEV only fleets have a lower cost than diesel only fleet. Since BEVs compete better with diesel vehicles in segments with small vans, that result could be understandable in such scenarios where optimal fleets involve small vans. However, it is not the unique reason. In the scenario where the demand is reduced to 25%, the total distance of the BEV only fleet is 260km while the total distance of the diesel only fleet increases to 307km. However, a diesel vehicle should be able to operate a route that is achieved by a BEV. The lower distance driven by the electric fleet compared to the diesel fleet shows therefore that the route solution found for the diesel fleet is suboptimal. There should be some unexplored solution by our algorithm. Indeed if the routes of the electric fleet are kept but diesel vehicles are operating them, we can find a total cost reduced to 413€ per day for the demand scenario at 25% (439€ found by the algorithm). Still, the full BEV solution remains less costly than the diesel only fleet solution that show a total cost of 412€ per day. We find the same possible improvement for the scenario where the size and weight of deliveries decrease by 50% compared to the baseline scenario. In that case however, the diesel only fleet become less costly than the BEV only fleet. It is reduced to 414€ per day (442€ found by the algorithm) while the total cost of the BEV only fleet is 434€. These differences can be explained by the use of heuristics which aim at finding an optimal solution within reasonable computation time. The additional constraints of BEVs can sometimes change the way the algorithm is navigating within the space of solutions. In this case, the constraints of the BEVs have led the algorithm to find more optimal routes than with a diesel fleet. Cost of the vehicles Finally, we tested the impact of a change in vehicles costs on their selection in the optimal mixed technology fleet. In some cities indeed, local authorities can give incentives to BEVs through subsidies or fiscal deduction. They can also increase taxes on fuel or implement distance tolls on conventional vehicles. Costs of the vehicles might also change in the future with less costly batteries for BEVs or more important fuel costs for conventional vehicles. We tested therefore the increased variable costs of conventional vehicles together with the decreased of fixed costs from BEVs at different levels. Figure 47 shows the impact on the last mile costs for an increase of diesel’s variable costs by 10%, 20%, 30% and 50% together with a decreasing of BEVs’ fixed costs by 10%, 20%, 30% and 50%. As expected, the last mile cost gap between diesel fleets and electric fleets are reducing with an increased variation in vehicle costs. The added costs of a BEV only fleet compared to a diesel only fleet is reduced from 12.34% in the baseline scenario to 5.41%, 2.86%, 0.62% and -1.58% in the following scenarios. The reduction of the last mile cost gap stimulates also more operators to integrate BEVs in their diesel fleets. The cost reduction of a mixed technology fleet compared to a diesel fleet is in the respective order -0.02%, -0.46%, -1.04%, -1.38% and -3.04%. However,


- 109 -

these costs changes of the vehicles do not influence much the number of BEVs selected in the mixed technology fleet. The sensitivity analysis shows that fixed costs of BEVs must be reduced by 50% and variable costs of diesel must increase by 50% in order to influence the mixed technology fleet. In that case, the mixed technology fleet is composed of one large diesel van, one large electric van and one small electric van. Compared to the baseline scenario, the large electric van replaces the large diesel van but keeps operating the same route: it delivers 9 shops in two hours and 38minutes and drives 61km. Figure 47: Impact of the costs of the fleet on costs of the last mile 700


600 500 400 300 200 100 0 Baseline

10%

Mixed technology fleet


30%

50%

BEV only fleet

5.6 Conclusions Battery electric vehicles can reach a competitive threshold with diesel vehicles when the distances driven are large enough to fully benefit of their lower running costs and offset their higher purchase costs (financial constraint of BEVs). However, battery electric vehicles are also limited by their range constraints and their sometimes lower payloads (operational constraints). The competitive position of BEVs must therefore consider also these aspects in order to evaluate their potential in city logistics. Hence, we developed in this chapter a fleet size and mix vehicle routing problem with time windows that was adapted to BEVs. In particular, we integrated an energy consumption model in order to consider the variable range of BEVs depending on speed, distance, weight, temperature and aerodynamics of the vehicles. We used then the model to analyse the most cost effective fleet of an urban distribution centre located in the city of Brussels. Based on a subset of their customers, we conducted a sensitivity analysis in order to evaluate in what kind of city logistics environments BEVs suit the best. The results showed that BEVs can indeed contribute to a cost reduction of the last mile in city logistics. But their applications are limited. Diesel remains the technology that provides most often the lowest last mile cost. But in some logistics environment, BEVs are selected in the optimal mixed technology fleet. The model showed that when time windows are narrow, when congestion increases, when shops are not concentrated around the depot and when freight demand involves small deliveries, BEVs should be used as a complement to a diesel fleet in order


- 110 -

to minimize the last mile costs. In those environments, freight flows are typically fragmented and the optimal fleet requires more but smaller vans. Since BEVs are mostly competitive with diesel vehicles in the small van segment, BEVs are more often selected in the optimal mixed technology fleet that operates a fragmented distribution. Still that fragmentation should not be too extreme because distance per vehicle need to be high enough for a BEV to be competitive with diesel. The financial constraint of BEVs (the lower distance bound of the BEV paradox) is therefore a frequent reason for not selecting a BEV in the optimal mixed technology fleet. And when a BEV is selected, the cost reduction it provides is incremental which might not stimulate transport operators to change. On the other hand, the operational constraints of BEVs (range constraints and more limited payload) are often compatible with the routes of the diesel fleet. Still, compared to fleets with diesel vehicles only, fleets with only BEVs result sometimes in solutions with a longer total distance (because an extra vehicle needs to be used or because deliveries need to be reassigned in less optimal routes) or in solutions with larger vehicles (because of more limited payload or limited battery size on smaller vehicles). These adaptations lead finally in additional costs which occur especially in fleets that are fully electric. BEVs have therefore a narrow distance window where they can be profitable, located between their lower and upper distance bounds. Still, manufacturers could make that distance window more flexible by leaving the choice of the battery size to their customers. With smaller batteries, that distance window could be moved to lower distances. Conversely, with larger batteries, routes with longer distances could become profitable with BEVs. Besides the development that manufacturers can bring to stimulate BEV adoption, local authorities have also an important role in that context given the lower impact of BEVs on the urban environment. The sensitivity analysis showed that integrating BEVs in the fleet of an urban distribution centre can reduce the total costs of the last mile by up to - 3% when running costs of diesel increase by 50% and fixed costs of BEVs decrease by 50%. But the sensitivity analysis showed also the important potential of specific advantage granted to BEVs such as time windows extensions or use of bus lanes. In those cases, costs reduction of BEV only fleets can be increased to about - 20%. The results have however to acknowledge the limitations of the model. We observed indeed that some better solutions could be found. The heuristics that we used in our algorithm explain that the solutions might be less optimal than the exact solution of our mixed integer linear program. Hence, future research should improve the performance of the algorithm used to solve the fleet size and mix vehicle routing problem with time windows for electric vehicles. Future research should also consider the variable consumption of diesel vehicles and develop a more detailed energy model for larger freight vehicles. The energy model should also be able to consider more extreme temperatures in order to estimate the influence of that factor on the limited range. Finally, the potential of other environmentally friendly vehicles should be analysed within that framework.


- 111 -

References Afroditi, A., Boile, M., Theofanis, S., Sdoukopoulos, E., Margaritis, D., 2014. Electric Vehicle Routing Problem with Industry Constraints: Trends and Insights for Future Research. Transp. Res. Procedia 3, 452–459. Amburg, B. Van, Pitkanen, W., 2012. Best Fleet Uses , Key Challenges and the Early Business Case for E-Trucks : Findings and Recommendations of the E-Truck Task Force. In: EVS26. Los Angeles, California, May 6-9, pp. 1–12. Anderson, S., Allen, J., Browne, M., 2005. Urban logistics––how can it meet policy makers’ sustainability objectives? J. Transp. Geogr. 13, 71–81. Bard, J.F., Huang, L., Dror, M., Jaillet, P., 1998. A branch and cut algorithm for the VRP with satellite facilities. IIE Trans. 30, 821–834. Belfiore, P., Yoshizaki, H.T.Y., 2013. Heuristic methods for the fleet size and mix vehicle routing problem with time windows and split deliveries. Comput. Ind. Eng. 64, 589–601. Blanco, S., 2013. 2013 Nissan Leaf gets 75-mile range (actually 84) in new EPA test [WWW Document]. Autobloggreen. URL http://green.autoblog.com/2013/02/21/2013-nissan-leaf-revealed-gets-75-mile-rangeactually-84-in-n/ (accessed 5.20.14). Bräysy, O., Dullaert, W., Hasle, G., Mester, D., Gendreau, M., 2008. An Effective Multi-restart Deterministic Annealing Metaheuristic for the Fleet Size and Mix Vehicle Routing Problem with Time Windows. Transp. Sci. 42, 371–386. Christofides, N., Mingozzi, A., Toth, P., 1981. Exact algorithms for the vehicle routing problem, based on spanning tree and shortest path relaxations. Math. Program. 20, 255–282. Clarke, G., Wright, J.W., 1964. Scheduling of vehicles from a central depot to a number of delivery points. Oper. Res. 12, 568–581. Conrad, R.G., Figliozzi, M.A., 2011. The Recharging Vehicle Routing Problem. In: Doolen, T., Van Aken, E. (Eds.), Proceedings of the 2011 Industrial Engineering Research Conference. Crainic, T.G., Ricciardi, N., Storchi, G., 2004. Advanced freight transportation systems for congested urban areas. Transp. Res. Part C Emerg. Technol. 12, 119–137. Crowe, P., 2013. European-Specific Nissan Leaf To Be Unveiled In Geneva [WWW Document]. Hybrid Cars. URL http://www.hybridcars.com/european-specific-nissan-leaf-to-be-unveiled-in-geneva/ (accessed 5.20.14). Dablanc, L., 2011. City distribution, a key element of the urban economy: guidelines for practitioners. In: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 13–36. Dablanc, L., Rakotonarivo, D., 2010. The impacts of logistics sprawl: How does the location of parcel transport terminals affect the energy efficiency of goods’ movements in Paris and what can we do about it? Procedia Soc. Behav. Sci. 2, 6087–6096. Dantzig, G.B., Ramser, J.H., 1959. The truck dispatching problem. Manage. Sci. 6, 80–91. De Vroey, L., Jahn, R., El Baghdadi, M., Van Mierlo, J., 2013. Plug-to-wheel energy balance - Results of a two years experience behind the wheel of electric vehicles. In: EVS27. Barcelona, pp. 1–5. EC, 2011. White paper: Roadmap to a Single European Transport Area. Brussels.


- 112 -

El Baghdadi, M., De Vroey, L., Coosemans, T., Van Mierlo, J., Jahn, R., 2013. Electric Vehicle Performance and Consumption Evaluation. In: EVS27. Barcelona, pp. 1–8. Erdoğan, S., Miller-Hooks, E., 2012. A Green Vehicle Routing Problem. Transp. Res. Part E Logist. Transp. Rev. 48, 100–114. Goeke, D., Schneider, M., 2014. Routing a Mixed Fleet of Electric and Conventional Vehicles. Gonçalves, F., Cardoso, S.R., Relevas, S., Povoa, A.P.F.D.B., 2011. Optimization of a distribution network using electric vehicles: a VRP problem. Lisboa. Hayes, J.G., de Oliveira, R.P.R., Vaughan, S., Egan, M.G., 2011. Simplified electric vehicle power train models and range estimation. IEEE Veh. Power Propuls. Conf. 1–5. Hiermann, G., Puchinger, J., Hartl, R.F., 2014. The Electric Fleet Size and Mix Vehicle Routing Problem with Time Windows and Recharging Stations. Submitt. to Transp. Sci. Laporte, G., Nobert, Y., Desrochers, M., 1982. Two exact algorithms for the distance constrained vehicle routing problem. Cah. du GERAD G-82-05. Laporte, G., Nobert, Y., Desrochers, M., 1985. Optimal Routing under Capacity and Distance Restrictions. Oper. Res. 33, 1050–1073. Laporte, G., Nobert, Y., Taillefer, S., 1987. A Branch-and-Bound algorithm for the asymmetrical distanceconstrained vehicle routing problem. Math. Model. 9, 857–868. Lebeau, P., Macharis, C., Van Mierlo, J., Lebeau, K., 2013. Electric vehicles for logistics: a total cost of ownership analysis. In: HESSE, Et al. (Eds.), Proceedings of the BIVEC-GIBET Transport Research Days 2013. Walferdange, Luxemburg-City, pp. 307–318. Lin, C., Choy, K.L., Ho, G.T.S., Chung, S.H., Lam, H.Y., 2014. Survey of Green Vehicle Routing Problem: Past and future trends. Transp. Res. Procedia 3, 452–459. Messagie, M., Boureima, F., Matheys, J., Sergeant, N., Timmermans, J.-M., Macharis, C., Van Mierlo, J., 2010. Environmental performance of a battery electric vehicle : a descriptive Life Cycle Assessment approach. World Electr. Veh. J. 4, 782–786. Quak, H., 2011. Urban freight transport: the challenge of sustainability. In: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 37–55. Russo, F., Comi, A., 2012. City Characteristics and Urban Goods Movements: A Way to Environmental Transportation System in a Sustainable City. Procedia - Soc. Behav. Sci. 39, 61–73. Schneider, M., Stenger, A., Goeke, D., 2014. The Electric Vehicle Routing Problem with Time Windows and Recharging Stations. Transp. Sci. 48, 500–520. Schoemaker, J., Allen, J., Huschebeck, M., Monigl, J., 2006. Quantification of Urban Freight Transport Effects I. Ségalou, E., Routhier, J., de Rham, C., Albergel, A., 2006. Mise en place d’une méthodologie pour un bilan environnemental physique du transport de marchandises en ville. Lyon. Toth, P., Vigo, D., 2002. Models, relaxations and exact approaches for the capacitated vehicle routing problem. Discret. Appl. Math. 123, 487–512.

Van Duin, J.H.R., Tavasszy, L.A., Quak, H.J., 2013. Towards E ( lectric ) - urban freight : first promising steps in the electric vehicle revolution 1–19.

PART III

Chapter 6: Strategic Scenarios for the Brussels-Capital Region

- 114 -

6 Strategic Scenarios for Sustainable Urban Distribution in the BrusselsCapital Region Using Urban Consolidation Centres11 6.1 Introduction The Brussels-Capital Region (BCR) is increasingly aware of the negative impacts of freight vehicles on the urban welfare. Vans and trucks were indeed estimated to be responsible of a third of fine particles generated by the transport sector and of a fourth of CO2 emissions (Lebeau and Macharis, 2014). They account also for 14% of the vehicles on the road. Given the important contribution of these vehicles on mobility, air quality and climate change, different type of solutions have been planned by the authorities in their strategic plan for urban freight transport in the BCR (Bruxelles Mobilité, 2013). Among others, the Urban Consolidation Centre (UCC) receive a specific attention in that plan. It has been one of the most investigated solutions in the Region. Several research papers (Hubert et al., 2008; Debauche and Duchateau 1998; Van Mierlo et al., 2004) have indeed addressed the implementation of a UCC in Brussels. Recent years have also seen some concrete steps, with the setting-up of several small UCCs in the region, run by private companies. Still, regional authorities require more insights in the possible future 11

This chapter is based on the following paper: Janjevic, M., Lebeau, P., Ndiaye, A.B., Macharis, C., Van Mierlo, J. and Nsamzinshuti, A. (2016). “Strategic Scenarios for Sustainable Urban Distribution in the Brussels-Capital Region Using Urban Consolidation Centres”. Transport Research Procedia 12, 598-612.


- 115 -

developments. In order to take decisions, they must address a large number of questions such as: (1) how many consolidation centres are necessary to service the Region? (2) what is the optimal size and location of these centres? (3) what part of the freight demand are they likely to capture? (4) is their operation possible without public subsidies? (5) what is the impact of other accompanying measures on their development? (6) what are the environmental impacts of such schemes? (7) what type of vehicles are best suited? (8) how will the urban stakeholders respond to different policy options? The multiplication of options leaves the decision-makers in front of a large number of uncertainties and dimensions of analysis. Hence, the chapter has the objective of giving a method to tackle this complexity. A scenario-planning methodology is described and applied on the case of the BCR in order to select a limited number of scenario themes that can be evaluated. The different quantitative models that support that methodology are also presented.

6.2 Methodology The methodology that was used for constructing strategic scenarios is scenario planning and comes from the field of strategic planning. Scenario planning is a systematic methodology method for designing possible futures that simplifies the avalanche of data into a limited number of possible states (Schoemaker, 1995). It differs from other planning methods as contingency planning which examines only one uncertainty and the sensitivity analysis that examines the effect of a change in one single variable (Schoemaker, 1995). According to Chermack et al. (2001), the two academic approaches most often cited in the field of scenario planning are those by van der Heijden (2011) and Schoemaker (1995)). Although different scenario planning approaches differ in their details, they share certain characteristic process steps (Wulf et al., 2010): (1) Definition of the scope; (2) Perception analysis; (3) Trend and uncertainty analysis; (4) Scenario building; (5) Strategy definition and (6) Monitoring. Scenario planning is an iterative methodology, but for better reader comprehension, this article will be structured around the subsequent steps. The article focuses on the first four steps of the process that lead to the elaboration of strategic scenarios. In the fifth step (“strategy definition”), public stakeholders can use results of the scenario building and evaluation to define and communicate a proper strategy in terms of public policy required to promote the shift towards the optimal scenario, while private stakeholders can use these results in order to anticipate market developments and adapt their business strategy. The last step of the scenario building methodology (“monitoring”) corresponds to the development of performance measurement system with specific KPIs that allow overseeing the implementation of different strategies. We will now address each of the first four steps of the scenario building methodology and explain their application to the particular case of scenarios for sustainable urban distribution in the BCR using UCCs. Definition of scope The first step of the scenario-planning methodology is the definition of the scope that sets the foundation for the analysis and strategy definition phases by specifying important characteristics for the scenario planning project such as the time frame, the commodity type and the market segment, geographic area or scope of analysis (Wulf et al., 2010; Schoemaker, 1995; Van der Heijden, 2005).


- 116 -

We will delimit the scope of scenarios using the following elements: (1) which types of UCCs are to be considered for the strategic scenarios? (2) which market segments are to be targeted by the UCCs in the BCR? (3) which is the geographical coverage of UCC services? (4) what is the time frame of the implementation of the strategic scenarios? Type of Urban Consolidation Centers Allen et al. (2012) describe three generic types: (a) UCCs serving all or part of an urban area (b) UCCs serving large sites with a single landlord (c) Construction project UCCs. Scenarios presented in this chapter focus on the first type of UCCs, i.e. UCCs serving all or part of an urban area since the two latter types of UCCs should be established on a case-to-case basis. Market segments UCCs are likely to be better suited to some types of goods and vehicle movements than others (Browne et al., 2005). In order to define relevant market segments, we will discuss the product/commodity type and the supply chains that are best suited to the UCC concept. The product/commodity type has been found to be an important explanatory variable in almost all the freight transportation choices (Holguin-Veras, 2002; Holguín-Veras et al., 2008). First of all, the product/commodity type determines the way in which the cargo is handled (HolguinVeras, 2002; Holguín-Veras et al., 2008). Several authors (Boudouin, 2006; Browne et al., 2005; Panero et al., 2011; Van Duin, 2009) find that UCCs are not suitable and cannot handle all types of products. Indeed, it is difficult for a single centre to be able to handle the wide range of goods moving in and out of an urban area, for example due to different handling and storage requirements (Browne et al., 2005). For this reason, most of the existing UCCs have started by distributing only parcels, as this doesn’t require specialized handling - although sometimes additional types of goods are handled by the UCC, either when it has proved successful, either in order to increase its volumes and reduce operating deficits (Panero et al., 2011). Secondly, the product/commodity type can be seen as a proxy for the market segment and the behaviour of actors and as such will have a major importance in the carriers’ decision to use the UCC (Holguin-Veras, 2002; Holguín-Veras et al., 2008). The empirical evidence with regards to the type of products that could benefit from the UCCs is however scattered, such as observed by Danielis et al. (2010), in a literature review that presents several studies on the subject. A detailed study performed for the city of Mestre, Italy (near Venice) evaluating the possibility to adopt a UCC scheme similar to Padua, concludes that they are likely to be clothing, specialised retails and dry food (Danielis et al., 2010). A survey among 118 retailers in Broadmead Bristol in the VIVALDI project found that the most suitable goods for the UCCs were of medium size, nonperishable, and not of high value (Hapgood, 2005). Marcucci and Danielis (2008) find that clothing and other specialised goods other than food are most likely to accept to use the UCC, while Ho.Re.Ca is more unlikely to use it. With regards to these elements, Danielis et al. (2010) conclude that pharmaceutical products and fresh food will not make use of a UCC, while clothing and footwear and Ho.Re.Ca, especially when supplied via own-account, might accept to use a UCC. Another study performed in Manhattan and Brooklin by Holguín-Veras et al. (2008) shows that food carriers (with exception of large food carriers), chemical carriers, household good carriers, textile carriers and plastic carriers are inclined towards using the joint delivery services. Finally, Panero et al. (2011) summarize some generic limitations to the type of goods that can be consolidated in UCCs: (1) perishable goods (with some exceptions such as Stockholm


- 117 -

UCC which can handle cold foods, Heathrow airport UCC which has chilled and frozen facilities and one of the two UCCs in Siena (Italy) which is specialized in food products); (2) heavy and bulky goods which are not suitable for certain types of vehicles used by the UCCs and (3) highvalue products whose transhipment is often prohibited by insurance companies. With regards to these elements from the literature, the strategic scenarios for urban distribution in the BCR will therefore focus on goods that do not require special holding conditions, are of medium-value and are transported in form of parcels or pallets. With regards to the type of supply chains, Routhier et al., (2001) characterize the supply chains of urban goods movements with regards to the activity sector (small retail, wholesale, handcraft and services, supermarkets, industry, transport and warehouses, offices and agriculture) and the type of transport management mode (own account shipper, own account receiver or hired transport). With regards to the type of activity sector, Browne et al. (2005) find that the biggest potential beneficiaries of the UCC are retailers who are not part of supply chains in which deliveries are already highly consolidated at distribution centres, and/or are receiving full vehicle loads. Major supermarkets and similar outlets who operate their own stock consolidation centres are not envisaged as being beneficiaries of UCCs since they essentially transfer full vehicle loads (Browne et al., 2005). A similar conclusion can be taken for the industry, transport and warehouses and agriculture activity sectors. On the other hand, handcraft and services and offices can be suitable goods to be handled by a UCC since they mostly receive parcel deliveries. Browne et al. (2011) present a micro-consolidation centre that is focused on delivering office equipment and Janjevic et al. (2013) confirm that office supplies and equipment are a relevant sector for microconsolidation initiatives. A similar conclusion is made by Belouannas (2014) that concludes that retail, offices and crafts and services are all relevant for the use of a UCC in Saint-Etienne (France). The wholesale sector is not typically considered to be a potential beneficiary since wholesalers tend to provide themselves consolidation services (Browne et al., 2005). However, the acid test will be whether or not the final deliveries are so geographically focused that a high vehicle utilisation for a specific urban centre can be achieved (Browne et al., 2005). Moreover, there seem to be some empirical evidence of wholesalers being users of UCCs such as in Aachen (Browne et al., 2005) and in Brussels (information gathered during an interview with regional transport authorities). With regard to these elements, the following activity sectors will be considered for the strategic scenarios in the BCR: small retail, offices, handcraft and services and wholesale. With regards to the type of management, Browne et al. (2005) conducted an extensive review of 67 UCC schemes in the world and mention UCCs that are used by carriers (e.g. Tenjin, Kassel, Basel, Tokyo, La Rochelle, Monaco, etc.) or shippers (e.g. Munich, Basel, Zurich, etc.). However, there are no examples of UCCs capturing goods movements from receivers transporting their goods on own account. These actors might indeed more difficult because transport is not part of their core business and they might therefore less perceive the added value of a UCC. For this reason, strategic scenarios in the BCR will focus on the two following transport management modes: own account shipper and third party transport. Geographical coverage of UCC services Due to the specific form of the BCR with urban sprawl outside the administrative borders of the Region, it is not sufficient to consider goods movements (delivery or pick-up) that take place


- 118 -

within the administrative boundaries of the BCR. We will therefore consider the BCR and the 13 municipalities that are adjacent to the Brussels ring road. Time frame of the implementation The local authorities, in their Strategic Plan for Urban Freight Transport in the BCR (Bruxelles Mobilité, 2013) have established goals for a more sustainable distribution with aims in decreasing 10%, 20% and 30% of vehicle movements by 2020, 2030 and 2050 respectively. The design of the scenarios is therefore made in a prospective setting. We selected a time frame to 2030 for the possible implementation of the scenarios. Perception analysis This step of the scenario planning methodologies aims in identifying major stakeholders participating in the scenario project and in analysing their perceptions, roles, interests, and power positions (Wulf et al., 2010; Schoemaker, 1995). For strategic scenarios in the BCR, we can identify two main groups of stakeholders: private stakeholders (carriers, shippers and receivers) and public stakeholders (administrators). We can highlight several important elements with regards to the perceptions of private stakeholders and public stakeholders that will have a major impact on the way we construct the strategic scenarios. The first issue is the financial viability of the UCC. As noted by Verlinde et al. (2012), many pilots and test cases show that many of the UCCs are granted only a short life because the cost of an additional transhipment prevents them of being cost-effective. Moreover, even if the introduction of the UCC produces net benefits for the overall transport chain, the fact that the benefits for different users are more difficult to quantify and allocate between the different stakeholders than the costs reinforces the status-quo and has probably been a factor inhibiting the development of UCCs in the past (Browne et al., 2007). In fact, Allen et al., (2012) mention that one of the major barriers concerned with making UCCs financially sustainable is the extent to which the various participants (carriers, receivers and local authorities) are willing and able to meet the financial costs of the UCC in return for the benefits that they receive. With regards to the financial intervention of the public stakeholders, we can highlight a certain paradox: on the one hand, the general consensus shows that UCCs must be financially viable in their own right in the medium-to long-term and that subsidies are not a desirable solution (Browne et al., 2007), but on the other hand there is no evidence that any truly UCC selffinancing schemes yet exist (Browne et al., 2005; Van Duin, 2009). Still, the authorities of the BCR are not willing to support structurally a UCC. They prefer to stimulate the development of self-sustaining privately run UCCs. Some support for the setting-up of the UCC could be offered (e.g. by renting public space at a decreased cost), but no long-term subsidies are intended. On the private side, the allocation of costs and benefits between supply chain participants leads us to a wider issue of power relations between the supply chain actors. Traditionally, UCCs have been focused on carriers (Van Rooijen and Quak, 2010) but only a few of these initiatives were realised in practice (Quak, 2008). However benefits of the consolidation centres for carriers have been demonstrated (Quak, 2008). Moreover, carriers acknowledge indeed the advantage of delivering a site rather than multiple points (Boudouin, 2006). According to Holguín-Veras and


- 119 -

Sánchez-Díaz (2015), two independent surveys (Regan and Golob, 2005; Holguín-Veras et al., 2008) have estimated that carrier’s willingness to participate in UCC is in the range of 16%-18%, which is significantly larger than the observed participation. Holguín-Veras et al. (2015) show that the ability of carriers to change behaviour is in fact constrained by the other participants in the supply chain and that shippers and receivers have a great deal of power to specify how the transportation is to be done. Receivers are unaffected by a UCC scheme (Holguín-Veras and Sánchez-Díaz, 2015) and will not proactively push for it as they are currently served in conditions that already meet their needs, even if they are aware of the problems associated with it (Boudouin, 2006). On the other hand, shippers could be negatively impacted by participation because of factors such as the lost of the control of their deliveries and loss of the direct interface between suppliers and customers (Holguín-Veras and Sánchez-Díaz, 2015; Browne et al., 2005). The fact that the agents that yield the most benefits from the scheme are the ones with the least power can therefore be an important hindrance to the development of UCCs. These elements from the literature lead us to the following conclusions for the strategic scenarios in the BCR: (1) UCCs are only feasible if they present net benefits for the transport chain as a whole; (2) no long-term subsidies are currently available for their operation; (3) the mechanism of the allocation of costs and benefits between the private supply chain actors is essential for their adoption; (4) regardless the possible financial benefits, other considerations must be taken into account in the implementation of the UCC scheme. With regards to the first element, strategic scenarios for the BCR will only consider transport chains for which the UCC(s) decrease(s) the total cost of transport. This is in fact considered as a prerequisite to their adoption. The second element signifies that only financially self-sustaining schemes shall be considered for the strategic scenarios. Finally, the last two elements reaffirm the need to consider stakeholder acceptance of the proposed scenarios beyond the financial measures. For this reasons, once that they are constructed and evaluated in a quantitative manner, strategic scenarios must be evaluated based on a consultation of the local stakeholders in order to assess their acceptance. Trend and uncertainty analysis All major approaches to scenario planning include an analysis of the most important trends and uncertain elements (Wulf et al., 2010). This process stage is sometimes conducted in two distinct steps, as in Schoemaker’s phases ‘Identify basic Trends’ and ‘Identify Key Uncertainties’ (Schoemaker, 1995), or combined into one ‘Data Analysis’ step (van der Heijden, 2005). We will now present some of the most important trends and uncertainties that were identified for strategic scenarios for urban distribution using consolidation centres in Brussels. Physical structure of the urban freight distribution Regarding the physical structure of the urban freight distribution, based on the Strategic Plan for Urban Freight Transport in the BCR (Bruxelles Mobilité, 2013), we can identify the following types of UCC configurations: (1) A centralized structure where a large UCC serves the entire region; (2) A hierarchical structure where a city distribution centre and several satellites (microhubs) are distributed throughout the Region; (3) A distributed system with interconnected UCCs that serve a particular geographic area or deliver specific product categories; (4) A distributed system with independent UCCs that all service the entire urban area. To our knowledge, there are


- 120 -

no examples of practical implementations of a hierarchical (or a two-echelon) structure or of the distributed system with interconnected UCCs. This could be explained by the additional step required by these configurations in the transport chain thus resulting in higher costs of deliveries. The strategic scenarios for urban distribution in the BCR will therefore consider two types of structure: a centralized structure and a distributed structure with independent UCCs. This means that in case of a distributed structure, we will take into account the market competition between different UCCs. Regarding the number of UCCs required to serve the region, Panero et al. (2011) highlight a certain paradox: on the one hand, it has been noted that enabling a single consolidation centre for the distribution in a large urban area is unlikely to be attractive for many freight flows due to the degree of diversion required from normal route; on the other hand, there are examples of UCCs that failed because the number of customers served was insufficient to reach the break-even volume, like the UCC in Leiden which originally targeted only the city centre. The service area was then expanded to the whole city and later to the surrounding cities in order to attract more freight flows (Panero et al., 2011; Van Duin, 2009). Therefore, the number of UCCs needs to be a compromise between the need to offer multiple accessible locations in order to capture freight flows and the need to reduce the total number of UCCs in order to ensure sufficient market share for each centre in particular. In order to construct our strategic scenarios, we will consider up to 5 UCCs servicing the BCR and determine the number of UCCs that fits best these two requirements. Possible locations of UCC(s) Browne et al. (2005) and Van Duin (2009) highlight the importance of the location for the success of the UCC. The important considerations are: (1) their proximity to the area served, especially if the UCC scheme involves the usage of vehicles with a limited range (i.e. electric vehicles); (2) the upstream accessibility (proximity of major roads) and the downstream accessibility and (3) an environment that does not create neighbourhood problems and that is secure. In order to establish a list of possible locations for UCC(s) in Brussels, we have identified a list of 26 zones in Brussels and its periphery that meet these criteria. For some zones, based on the information from the planning documents (such as Bruxelles Mobilité (2013), Région de Bruxelles-Capitale (2014), Région de Bruxelles-Capitale (2002)), it was possible to identify a specific site that could be used for the setting-up of a UCC. For other zones, we have used the geographical epicentre of the zone. Considering that all planning documents mention the establishment of a logistical pole at the site of Schaerbeek Formation (a location in the north of the Region with tri-modal access), a UCC at this location will be considered in all strategic scenarios. This leaves us with 25 additional locations that could be considered for additional UCCs. Fleet of UCC vehicles The fleet of UCC vehicles can vary according to their size and vehicle technology. The decision regarding the size of vehicles is extremely important as it highly impacts the vehicle-trips and the congestion in the urban area. Strategic scenarios will therefore consider both vans and trucks. Regarding the vehicle technology, there are a number of possibilities: diesel, hybrid, CNG, electric vehicles, etc. Currently, diesel vehicles are the most commonly used in city distribution in


- 121 -

Brussels (Lebeau and Macharis, 2014). However, the local authorities have objectives of reaching free CO2 logistics by 2050 (Bruxelles Mobilité, 2013). In that context, battery electric vehicles can be considered as the most efficient technology to reach that objective. Browne et al. (2011) have indeed shown the potential that a combination between battery electric vehicles and a UCC has on reducing CO2 emissions. Besides, Lebeau et al. (2015) have demonstrated the feasibility and the economic relevance of introducing battery electric vehicles in urban distribution in Brussels. Hence, scenarios for urban distribution will consider the two following technologies: diesel and battery electric vehicles. Operation hours of the UCC Boudouin (2006) mentions that a UCC should be open during the entire time of the day when the goods are treated. The same author mentions the following hourly activity of a UCC: first receptions from 7AM and the distribution of goods until 6PM. In their strategic plan for urban freight transport in the BCR (Bruxelles Mobilité, 2013), regional transportation authorities advocate the usage of the night and off-hours deliveries in the BCR. In that sense, off-hour deliveries for delivering goods to the UCC can be considered as a possibility in the strategic scenarios. Regulatory and market framework We can highlight two main strategies used by public authorities to promote UCCs: regulatory strategy and market-based strategy. Regulatory instruments can differ according to the level of public intervention (Panero et al., 2011; Ville et al., 2013): (1) obligation to use a UCC by granting a special status (i.e. the permission to deliver an area) to a single market player, the UCC operator; (2) obligation to use a UCC by granting a special status (i.e. the permission to deliver an area) to a series of market players through a license system for example (Ville et al., 2013) and (3) inducing vehicles to use the UCC by reducing the accessibility of the area served by the UCC (Ville et al., 2013). Some European cities, including La Rochelle in France, Monte Carlo in Monaco, or Vicenza in Italy prefer to implement strong municipal regulations to stimulate the use of the UCC. They consider indeed that a UCC is the only way to ensure a successful rationalisation of urban goods distribution. On the other hand, other European cities do not consider the rationalization of freight delivery to be a municipal responsibility; the UCC should result of market mechanisms and should not require regulations to stimulate it. Regarding the market-based measures, we can distinguish between two major types of measures: (1) measures targeting carriers (freight pricing) and (2) measures focusing on changing the behaviour of the receivers of goods (mobility credits). De Palma and Lindsey (2011) categorize freight or congestion-pricing schemes according to several dimensions: (a) the type of scheme (facility-based schemes, cordon schemes, zonal schemes or distance-based schemes); (b) the degree to which pricing schemes vary over time (flat, depending on the time-of-day, day of week or season according to a predetermined schedule or responsive and function of prevailing traffic conditions); (c) the differentiation of the scheme with regards to the vehicle characteristics or driving behaviour (differentiation according to the vehicle type, vehicle technology, number of axles, weight of the vehicle or speed of the vehicle). In recent years, London (2003), Stockholm (2006), Durham (2002), Milano (2008), Rome (2001), Valletta (2007), Oslo and Trondheim have


- 122 -

all introduced different forms of charging or permit systems to combat congestion and/or environmental problems (Börjesson et al., 2012; Transmodal, 2012). However, despite the growing success of these schemes, road-pricing appear to have a limited impact on the freight transport demand management. Whereas there is ample theoretical support and empirical evidence that show that road pricing is an effective transportation demand management technique in the case of automobile transportation, this is not so clear in the case of freight transportation (Holguín-Veras, 2010; Quak and van Duin, 2010). Still, authorities of the BCR are currently implementing a distance-based freight-pricing scheme planned to come into action in 2016. Under its current form, the distance-based freight pricing concerns only heavy goods vehicles (HGVs). However, it is possible to imagine the extrapolation of this scheme to light commercial vehicles (LCVs). The strategic scenarios for the implementation of UCCs in the BCR will therefore consider only the distance-based freight pricing since this is the most probable measure but the type of vehicles targeted by this measure will vary across the scenarios. Pricing of the UCC services With regards to the pricing of the UCC services, we can rely on several sources. Boudouin (2006) reports that an average price of the UCC services is 3 euros/parcel. This order of magnitude is confirmed by Courivault (2004) that reports a price of 3,80 euros/parcel for the UCC in La Rochelle and a price of 4,90€/100kg for the UCC in Monaco. Data from Brussels gathered during interviews confirms a price of 2,5-5€/parcel and of 10-20€/pallet. In a feasibility study of freight consolidation centre in South London, Lewis et al. (2007) use the price hypothesis of £2-5 for an individual parcel and £5-10 for a pallet (2,9-7,1€ per parcel and 7,1-14,3€ per pallet according to the exchange rate in July 2015). This is also in line with the pricing of other urban logistics services reported by Chiron-Augereau (2009). We will therefore use three cost hypothesis for the pricing of the UCC services ranging from 2,5€ to 5€ per parcel and from 10€ to 20€ per pallet. Scenario building The scenario building phase is the core element of the traditional approaches to scenario planning where the previously identified key uncertainties are converted into distinct scenarios that describe different future states of the world (Wulf et al., 2010). In order to build the scenarios, we will proceed in three steps inspired by Schoemaker (1995): (1) Developing initial scenario themes and learning scenarios, (2) Defining research needs and develop quantitative models, (3) Choosing scenarios with the most realistic narrative. Initial scenario themes and development of learning scenarios This phase consists in combining the main ingredients for scenario construction. Based on the trends and uncertainties identified in the previous section of the chapter, the following options have been selected for the building of the scenarios: (1) Physical structure of urban distribution: centralized structure with 1 UCC situated at Schaerbeek Formation or distributed structure with up to 5 independent UCCs (1 UCC situated at Schaerbeek Formation and 1 to 4 additional UCCs situated in other areas); (2) Possible location of UCC: in all scenarios 1 UCC shall be situated in Schaerbeek Formation site; in case of a distributed structure, 25 additional locations for setting-up 1 to 4 UCCs have been identified;


- 123 -

(3) Fleet of UCC vehicles: scenarios will investigate 2 possibilities with regards to the type of UCC vehicles (diesel vehicles or electric vehicles) and 2 possibilities with regards to the size of the UCC vehicles (trucks or vans); (4) Operation hours of the UCC: strategic scenarios will investigate the possibility of allowing or not allowing night distribution to the UCC by formulating two hypothesis with regards to it operation hours; (5) Regulatory and market framework: the scenarios will investigate a UCC implementation with no freight pricing, with a distance based freight pricing limited to HGV and a distance based freight pricing extended to LCVs, considering the environmental performance of the vehicles; (6) Pricing of the UCC services: for the development of the scenarios, three cost hypothesis will be tested: low-cost (2,5 euros/parcel; 10 euros/pallet), medium-cost (3,75 euros/parcel; 15 euros/pallet), high-cost (5 euros/parcel; 20 euros/pallet). Define research needs and develop quantitative models The number of possibilities across the different dimensions identified in the previous step leads to a very large number of scenarios that could be considered. In order to reduce the number of possible scenarios, we will develop a series of quantitative models to evaluate each scenario for consistency and plausibility and to quantify the consequences of various scenarios. For evaluating the consistency and plausibility of scenarios, we will look at their operational feasibility (e.g. ensuring that scenarios are feasible in terms of necessary infrastructure, ensuring that scenarios are capturing a sufficient amount of freight flows). For quantifying the consequences of each scenario, we will look at two main elements: the financial feasibility (e.g. ensuring that scenarios are financially sustainable without public subsidies) and environmental impact of each scenario. In order to respond to the aforementioned research needs, several quantitative models have been developed. The overall integration of the models is shown on Figure 48. We will now provide a brief description of each model. Freight routes characteristics: in order to quantify the consequences of the scenarios, we need first to have an overview of the current situation. We have used several primary sources of data. The first one is an origin destination matrix of urban goods movements in the BCR and the municipalities located in the periphery of the city. The data were generated within the European project LaMiLo (see http://www.lamiloproject.eu/) using the FRETURB software (Routhier and Toilier, 2007) for modelling urban goods movements. A second source came from Google Maps API that gave the travel distances and travel times. Finally, a third source came from Gerardin et al. (2000) that provided information about the stop duration of freight vehicles. Following some data adjustments, the combination of these two data sources has allowed us to characterize urban freight routes in terms of origins and primary destinations, average number of stops within a route, average route travel distances and average route times. This characterization has been done for each type of vehicle, each type of management and each route size (direct trip or 6 categories of round sizes according to the number of stops in a round). The location-allocation model: the model uses the data on freight routes characteristics as input. It computes the optimal structure (number, location and amount of captured movements and trips) of the network of UCCs servicing the BCR. The location-allocation model is a bi-level


- 124 -

model that simulates two levels of decision: the decision of the planner who aims in selecting the locations of the UCCs that capture the highest amount of freight movements (i.e. the locations that are most likely to reach the break-even volume) and the decision of each particular vehicle that chooses to perform a direct trip or to use a specific UCC based on the total cost. The model allocates a specific freight flow to a specific UCC if the cost is lower than a direct trip or a passage though any other UCC. The model considers three different pricing scenarios for parcels and pallets, three types of distance-based road fees and the possibility of having off-hours deliveries to the UCC(s). The compilation of results provides the optimal configurations of UCC(s) in terms of number, size and location that capture the highest amount of vehicle trips from 15 302 possible location patterns of UCC(s) (corresponding to a maximum of 5 UCCs from 26 candidate sites with 1 fixed site). The model also provides information for all scenarios about the number of freight movements, number of trips, vehicles*km, total duration of deliveries and total cost of deliveries. The vehicle routing model: the data about the optimal location of the UCC(s) as well as the number of captured deliveries was used then as input to a vehicle routing model developed by Lebeau et al. (2015). It computed the last mile operations of the UCC considering the different constraints of urban distribution. First, the time windows of the different shops are considered so that the shops are delivered between 10am and 6pm. But when nigh distribution is considered in the scenario, the time windows are extended from 8am until 6pm. A second constraint is the capacity of the vehicles. We considered two types of vehicle: a van with a volume of 7,5m³ and a truck with a volume of 21,2m³. Finally, a third constraint considered the more limited range of battery electric vehicles compared to diesel vehicles. The external cost model: Based on the one hand on the kilometres driven by the vehicle to the UCC provided by the location-allocation model and on the other hand on the kilometres driven from the UCC provided by the vehicle routing model, the external cost model estimates for each scenario the following environmental impact: climate change, air pollution and noise emissions. In order to do this, the external cost model considers how the kilometres are driven. First, it differentiates the vehicles according to their size categories (vans, truck and trailers/semi-trailers). Heavier vehicles produce more global, local and noise emissions and contribute more to congestion. Then, it takes into account the vehicle technologies (diesel and electric). For diesel trucks, we considered an EURO V vehicle while battery electric vehicles were considered to use electricity produced from renewable energy sources. Finally, it distinguishes deliveries that are made at night and during the day. External costs of noise are indeed more important at night. The external cost model was constructed based on the report of the DG MOVE (2014). The UCC financial model: finally, the UCC financial model was used to test the financial viability of each scenario. The revenues of the UCC were established based on the number of freight movements captured by each UCC (coming from the location-allocation model) and the cost hypothesis used in each model. The following cost categories were considered: (1) infrastructure costs were estimated based on the UCC surface (calculated based on the UCC throughput, i.e. number of parcels and pallets) and the UCC location (with varying price per square meter); (2) vehicle costs were estimated based on the number of vehicles tours from the vehicle routing model and the cost data used by Lebeau et al. (2013); (3) equipment costs were


- 125 -

estimated based on the UCC throughput; (4) human resources costs were estimated based on the number of tours (drivers) and on the UCC throughput (warehouse staff), (5) overhead costs. It is to be noted that the hypothesis used in the financial model suppose optimal operations of the UCCs. Figure 48: Overall modeling structure

Freight routes characteristics: - Type of vehicle - Type of management - Size of tour - Duration of stops - Average distance - Geographical concentration

Cost of UCC services scenarios: - Price/pallet - Price/parcel Physical structure and location patterns of UCCs - Number of UCCs - Location of UCCs

Location-allocation model

Optimal structure of urban distribution - Number of UCCs - Location of UCCs - Size of UCCs - Captured movements

Accompanying measures: - Road tolls - Night distribution

Vehicle routing model

Vehicle fleet scenarios: - Type (diesel, EV) - Size (vans, trucks)

UCC vehicles operational data: - Number of tours - Number of deliveries - Vehicles*km

Financial viability

UCC financial model

Operational viability

Environmental impact

External cost model

The quantitative models have allowed limiting the number of possible scenarios due the following considerations: (1) The marginal impact of adding one additional UCC decreases with the number of UCCs. A 2nd UCC enables capturing 27,6% more vehicle-routes that the first one whereas a 5th UCC allows capturing 2,9% more vehicle-routes compared to four UCCs. We have therefore reduced the number of possible UCC(s) to 1, 2 and 4. Moreover, the precise location of a UCC has very little impact on the captured routes and movements as UCC locations in neighbouring zones tend to have the same performance. The scenarios shall therefore be presented in terms of UCC zones rather than precise locations. (2) The results of the location-allocation is used to estimate the necessary surface for setting-up a UCC at each particular location with regards to the flows captured, allowing eliminating some scenarios with regards to their operational viability (e.g. a scenario with one single UCC and distance based pricing for all vehicles would require dimensions exceeding the possible space) (3) The low-cost scenario for pricing of the UCC services has consistently produced negative gross margins whereas the high cost scenario for the pricing of the UCC has not been able to capture a sufficient number of deliveries. Only the medium cost scenario shall be considered for the decision scenarios. In fact, public subsidies are possible in the start-up phase, but all scenarios must eventually reach financial sustainability.


- 126 -

(4) The setting-up of 4 UCCs is only possible with strong regulations and accompanying measures such as night distribution to the UCC as other alternatives do not capture sufficient freight flows. (5) Under the hypothesis that each UCC serves the entire region and with regards to the distances to be travelled by UCC vehicles, vans are appropriate for the scenario with 4 UCCs only. Choosing scenarios with the most realistic narrative The quantitative models have allowed reducing significantly the number of scenarios, but still resulting in a large number of possibilities. In order to choose the final decision scenarios, we will choose among the possible scenarios those who have the strongest narrative. In fact as scenario is more than a random combination of elements: each scenario tells a story of how various elements might interact under certain conditions (Schoemaker, 1995). The scenarios that have been finally selected must provide storylines that are clearly distinguishable. With this in mind, the following scenarios have been selected: (1) Scenario 1: one UCC situated at Schaerbeek Formation, freight road pricing for HGVs, regular trucks and medium price of UCC services. Night distribution is not allowed to the UCC. This scenario positions as an extrapolation of the political option that is currently chosen by regional authorities. (2) Scenario 2: Two UCCs (one at Schaerbeek Formation and one at the north of Brussels), night distribution to the UCCs and no freight road pricing, trucks and medium price of UCC services. This scenario is an option to which the freight carriers are expected to be most favourable, as it does not introduce any additional costs and still offers the possibility of using a consolidation centre if more cost-effective. Moreover, the night distribution to the UCC allows decreasing even more the cost of transport to the UCC. Since no additional pressure is put on the carriers to use the UCC services, two UCCs are expected to suffice to cover the demand. In order to test the effect of the introduction of the electric-vehicles, two sub-scenarios are defined: scenario 2a using the diesel trucks and scenario 2b using electric trucks (3) Scenario 3: Four UCCs, night distribution allowed to the UCCs, freight road pricing based for all vehicles based on their environmental performance, vans and medium price of UCC services. This scenario presents the extreme point of possible futures and the most radical change in urban distribution as it combines heavy regulation, large number of UCCs and night distribution. It is the scenario with the best results in terms of environmental performance but with the heaviest constraints for the carriers. In order to test the effect of the introduction of the electric-vehicles, two sub-scenarios are defined: scenario 3a using the regular vans and scenario 3b using electric vans. Figure 49 shows the optimal UCC locations for the three scenarios. These locations will be used to evaluate the scenarios – however, the precise location of a UCC has very little impact on the evaluation.


- 127 -

Figure 49: Optimal locations of the UCCs for the Scenario 1, 2 and 3 respectively

UCC location

City centre

Zone within the ring road

Zone outside the ring road

6.3 Results Based on the tools we developed, we quantified the impact of the scenarios on the urban freight flows in Brussels as depicted in Table 8. The evaluation of the selected scenarios details: (1) the impact on transport costs for the carriers (the cost variation is expressed for all carriers regardless of the fact that they use the UCC(s)); (2) the impact on congestion from all carriers in the region (to include the different impact on congestion of different vehicle categories, the indicator did not consider vehicles arriving at the depot before 6am and expressed the vehicles*km in terms of “personal vehicle equivalent” which implies a multiplication by 1,5 for vans, 2 for trucks and 2,5 for trailers and semi-trailers as considered by Routhier et al. (2001)); (3) the impact on external costs of local and global emissions from all carriers in the region; (4) the impact on external costs of noise emissions from all carriers in the region; and (5) the gross margin of the UCC(s) which is expressed for all the UCC(s) in each specific scenario. Table 8: Evaluation of decision scenarios with regards to business as usual

Scenario 1

Scenario 2a

Scenario 2b

Scenario 3a

Scenario 3b

Cost of transport

+6,3%

-1,3%

-1,3%

13,4%

13,4%

Congestion

-2,2%

-4,1%

-4,1%

-7,2%

-7,1%

Emissions

-3,0%

-2,9%

-3,7%

-5,8%

-8,1%

Noise

-2,2%

5,6%

4,7%

3,9%

2,0%

Gross margin UCC(s)

8,3%

15,3%

3,2%

11,9%

7,8%

Cost of transport is being reduced in scenario 2a and 2b when no toll is introduced. This benefit is coming from the optimisation of the goods flows achieved by the services of the UCCs. However, in scenario 1, costs increase because of the introduction of the toll on heavy goods


- 128 -

vehicles. Costs increase further in scenario 3a and 3b when the toll is extended to the light commercial vehicles. The second indicator shows that congestion is reduced with the introduction of additional UCCs. The increasing number of UCCs offers a better geographical coverage on the city and captures therefore more flows. We can highlight the minimal influence of battery electric vehicles on congestion: the scenario with battery electric vehicles (3b) reduces less congestion than the scenario with diesel vehicles (3a). Their limited range imposes additional constraints which lead to, sometimes, more delivery turns and thus more vehicle kilometres. Still, that difference is very low since no difference can be observed between scenario 2a (with diesel vehicle) and scenario 2b (without battery electric vehicle). The benefits of battery electric vehicles appear better in the impact of the scenarios on emissions. Table 8 shows that they improve significantly the environmental performance of their similar diesel scenarios. The reduction of vehicle kilometres due to an optimisation of freight flows by the UCC contributes also to a better environmental performance. Interestingly, that is not the case between the scenario 1 and scenario 2a. The toll in scenario 1 was indeed incentivising more heavy goods vehicles to use the UCC than in the scenario 2 where no toll is considered. As a result, the UCCs in scenario 2 capture a larger share of light commercial vehicles than heavy goods vehicles compared to the scenario 1 which explain a lower environmental efficiency of scenario 2 despite a better result on the criteria congestion. The impact of the scenarios on external costs of noise shows more negative results. The night distribution allowed to the UCC impacts significantly noise levels in the city. The reduction of noise because of less freight vehicles in the traffic (through a rationalisation of freight flows) cannot compensate the impact of a few freight vehicles driving at night. The marginal impact of a freight vehicle is indeed much higher during the night and when the density of the traffic is low than during the day and when the density of the traffic is high. Still, we can note the positive influence the battery electric vehicles can bring to reduce noise levels in the city. Regarding the gross margin, we can see that while using the same type of vehicles (diesel trucks), the scenario 2a presents a higher profitability than the scenario 1: the addition of a second UCC allows shortening the total distances and times travelled by UCC vehicles, thus reducing the related operational costs. The scenario 3a (4UCCs and diesel vans) results however in lower profitability. In fact, although vans have lower kilometric and hourly costs than trucks, their small capacity results in lower efficiency in terms of volume transported per euro of operating cost. We can observe also the influence of battery electric vehicles on the gross margin by comparing scenarios 2b with 2a and comparing scenarios 3a with 3b. Battery electric vehicles produce in each scenario a lower margin compared to their similar scenarios with diesel vehicles (2a and 3a). In fact, battery electric vehicles can improve the profitability of a fleet if it is used in combination with conventional vehicles as observed by Lebeau et al. (2015). The negative impact of electric vehicles on the gross margin is lower for the scenario 3b than for the scenario 2b for two reasons: (1) scenario 3b uses smaller electric vehicles that are more competitive than larger electric vehicles (as demonstrated by Lebeau et al., 2013); (2) electric vehicles are not affected by the distance-based freight pricing in the scenario 3b since in this case, freight-pricing is based on the environmental performance of the vehicles and not their weight.


- 129 -

6.4 Conclusion This chapter uses a combination of qualitative and quantitative approach in order to construct scenarios for the urban city distribution in the Brussels-Capital Region using UCCs. The large number of options regarding the possible scenario components calls for a use of a formal methodology in order to be able to select a limited number of scenario themes. These themes are then detailed though a series of quantitative models in order to reach a limited number of decision scenarios. This chapter shows that the combination of these two approaches can be an interesting option for enhancing decision-making under uncertainty and that scenario planning methodology which is usually applied in corporate setting can in fact bring interesting results in the fields of public decision making in urban freight transport. It is however important to leave different possible alternatives open in order to evaluate the support of the stakeholders. They are indeed recognised to be critical to guarantee the success of projects in urban freight transport. As previously discussed in this chapter, the non-monetary benefits and the power relations between supply chain actors have a major impact on the acceptance of UCC schemes. Future research is therefore needed to further evaluate the scenarios based on a consultation of the local stakeholders. This will lead to the identification of the scenario with the highest acceptance among the stakeholders.


- 130 -

References Allen, J., Browne, M., Woodburn, A., Leonardi, J., 2012. The Role of Urban Consolidation Centres in Sustainable Freight Transport. Transport Reviews 32, 473–490. Belouannas, F., 2014. MEDUSE: Marchandises et Etude de la Distribution Urbaine à St-Etienne «tous acteurs de la logistique urbaine». Börjesson, M., Eliasson, J., Hugosson, M.B., Brundell-Freij, K., 2012. The Stockholm congestion charges—5 years on. Effects, acceptability and lessons learnt. Transport Policy 20, 1–12. Boudouin, D. 2006. Les espaces logistiques urbains (guide méthodologique). Paris: La Documentation française Transports, recherche, innovation guide Browne, M., Sweet, M., Woodburn, A., Allen, J., 2005. Urban Freight Consolidation Centres Final Report. Browne, M., Woodburn, A., Allen, J., 2007. Evaluating the potential for urban consolidation centres. European Transport / Trasporti Europei 35 Browne, M., Allen, J., Leonardi, J., 2011. Evaluating the use of an urban consolidation centre and electric vehicles in central London. IATSS Research 35, 1–6. Bruxelles Mobilité, 2013. Plan stratégique pour le transport de marchandises en région de Bruxelles-Capitale. Brussels. Chermack, T., Lynham, S., Ruona, W., 2001. A review of scenario planning literature. Futures Research Quarterly 17. Courivault, N., 2004. Les livraisons/enlèvements en centre-ville: quels problèmes? Quelles solutions? Université Lumière - Lyon 2 Chiron-Augereau, V., 2009. Du transport de marchandises en ville à la logistique urbaine, quels rôles pour un opérateur de transports publics urbains?: l’exemple de la RATP. Université Paris-Est. Danielis, R., Rotaris, L., Marcucci, E., 2010. Urban freight policies and distribution channels. European Transport / Trasporti Europei 46, 114–146 Debauche, W., Duchateau, H., 1998. Urban freight transport strategy in Brussels. Proceedings of Seminar B Held at AET European Transport Conference, Loughborough University, England, Volume P421. De Palma, A., Lindsey, R., 2011. Traffic congestion pricing methodologies and technologies. Transportation Research Part C: Emerging Technologies 19 (6), 1377–1399. DG MOVE (2014. Update of Handbook on external costs of Transport. report MOVE/D3/2011. Gerardin, B., Patier, D., Routhier, J.-L., Segalou, E., 2000. Diagnostic du transport de marchandises dans une agglomération - Document technique n°1. Hapgood, T. , 2005. The Brisol VIVALDI Project experience, in: 1st BESTUFS II Workshop Thematic focus: Approaches to Urban Consolidation: concepts and experiences , London. Holguín-Veras, J., 2010. The truth, the myths and the possible in freight road pricing in congested urban areas. Procedia-Social and Behavioral Sciences, The Sixth International Conference on City Logistics, 2, 6366–6377 Holguin-Veras, J., 2002. Revealed preference analysis of commercial vehicle choice process. Journal of transportation engineering 128, 336–346.


- 131 -

Holguín-Veras, J., Silas, M., Polimeni, J., Taniguchi, E., Thomson, R., 2008. An investigation on the attitudinal factors determining participation in cooperative multi-carrier delivery systems. Innovations in city logistics IV. Nova Science Publishers 55–68. Holguín-Veras, J., Aros-Vera, F., Browne, M., 2015. Agent interactions and the response of supply chains to pricing and incentives. Economics of Transportation (in press). Holguín-Veras, J., Sánchez-Díaz, I., 2015. Freight Demand Management and the Potential of Receiver-Led Consolidation programs. Transportation Research Part A: Policy and Practice (in press) Hubert, M., Dobruszkes, F., Macharis, C., 2008. La mobilité à, de, vers et autour de Bruxelles. Brussels Studies, 1–14. Janjevic, M., Kaminsky, P., Ballé Ndiaye, A., 2013. Downscaling the consolidation of goods – state of the art and transferability of micro-consolidation initiatives. European Transport / Trasporti Europei 54 Lebeau, P., De Cauwer, C., Van Mierlo, J., Macharis, C., Verbeke, W., Coosemans, T., 2015. Conventional, Hybrid, or Electric Vehicles: Which Technology for an Urban Distribution Centre? The Scientific World Journal., in press) Lebeau, P., Macharis, C., 2014. Le transport de marchandises à Bruxelles : quels impacts sur la circulation automobile ? Brussels Studies, 80, 1–14. Lebeau, P., Macharis, C., Van Mierlo, J., Lebeau, K., 2013. Electric vehicles for logistics: a total cost of ownership analysis. Proceedings of BIVEC-GIBET Transport Research Days, 307–318. Lewis, A., Lagrange, A., Patterson, D., Gallop, N., 2007. South London Freight Consolidation Centre Feasibility Study - Final Report. Marcucci, E., Danielis, R., 2008. The potential demand for a urban freight consolidation centre. Transportation 35, 269–284. Panero, M.A., Shin, H.-S., Lopez, D.P., 2011. Urban distribution centres–A Means to reducing freight vehicle miles traveled. New York University Région de Bruxelles-Capitale, 2014. Plan Régional de Développement. Moniteur belge 15. Région de Bruxelles-Capitale, 2014. Projet de Plan Régional de Développement Durable. Retrieved from http://www.prdd.be/. Quak, H.J., (2008) Sustainability of urban freight transport: Retail distribution and local regulations in cities. Rotterdam: Erasmus Research Institute of Management (ERIM) Regan, A.C., Golob, T.F., 2005. Trucking industry demand for urban shared use freight terminals. Transportation 32, 23–36. Routhier, J.-L., Segalou, E., Durand, S., 2001. Mesurer l’impact du transport de marchandises en ville: le modèle de simulation FRETURB (V.1). Programme national marchandises en ville. Routhier, J.-L., Toilier, F., 2007. FRETURB V3, A Policy Oriented Software Tool for Modelling Urban Goods Movement. Presented at the 11th World Conference on Transport Research. Quak, H., van Duin, J.H.R., 2010. The influence of road pricing on physical distribution in urban areas. Procedia Soc. Behav. Sci. 2, 6141–6153. Schoemaker, P.J., 1995. Scenario planning: a tool for strategic thinking. Sloan Management Review. 36, 25–25.


- 132 -

Transmodal, M.D.S., 2012. DG MOVE European Commission: Study on Urban Freight Transport. Final report. Chester, UK: MDS Transmodal Limited. Triantafyllou, M.K., Cherrett, T.J., Browne, M., 2014. Urban Freight Consolidation Centers. A Case Study in the UK Retail Sector. Proceedings of Transportation Research Board 93rd Annual Meeting. Van Duin, R., 2009. To be or not to be, a typical City Distribution Centre question. Research on success and failures in ten European CDC-cases. Proceedings of Bijdragen vervoerslogistieke werkdagen, 123–145. Van Mierlo, J., Van Den Bossche, P., Maggetto, G., 2004. Integrated modelling of the urban development mobility and air pollution analysis in the Brussels-Capital region: Policy measures based on environmentally friendly vehicle technologies. Final report of a Prospective Research for Brussels. Van Rooijen, T., Quak, H., 2010. Local impacts of a new urban consolidation centre–the case of Binnenstadservice. Procedia-Social and Behavioral Sciences, 5967–5979. Van der Heijden, K., 2011. Scenarios: the art of strategic conversation. John Wiley & Sons Verlinde, S., Macharis, C., Witlox, F., 2012. How to consolidate urban flows of goods without setting up an urban consolidation centre? Procedia-Social and Behavioral Sciences 39, 687–701. Ville, S., Gonzalez-Feliu, J., Dablanc, L., 2013. The limits of public policy intervention in urban logistics: Lessons from Vicenza (Italy). European Planning Studies, 21, 1528–1541. Wulf, T., Meissner, P., Stubner, S., 2010. A scenario-based approach to strategic planning–integrating planning and process perspective of strategy. Leipzig Graduate School of Management

Chapter 7: Operationalising a bottom-up approach through a MAMCA

- 133 -

7 Operationalising a bottom-up approach through a multi-actor multicriteria analysis: an application to city logistics 12 7.1 Introduction City logistics is one of the most polluting segments of transport. Although it accounts for 10 to 15% of city’s traffic, it is responsible for one fourth of CO2 emissions, one third of NOx emissions and half of the particulate matter generated by the transport sector in large cities (Dablanc, 2011). However, authorities want it to become one of the cleanest segments in the future. The European Commission for example has set the objective of reaching CO2 free city logistics by 2030 in major urban areas (EC, 2011). As a result, research projects have developed a wide range of solutions to improve the sustainability of the logistics sector in cities which are often called “best practices” (Bestfact, 2013). They typically include an integrated set of instruments aspects such as regulations, economic measures, physical infrastructure, urban planning, new technologies and educational and informational measures (Quak, 2011). The challenge for local authorities is therefore to choose the right set of measures to implement. Best practices are however often context-dependant (Janjevic et al., 2013; Patier and Browne, 2010). Their success depends indeed on the support that they receive by the local business groups, 12

This chapter is based on the following paper: Lebeau, P., Macharis, C., Van Mierlo, J. and Janjevic, M. 2015. “Implementing an urban consolidation centre: involving stakeholders in a bottom-up approach”. Transportation Research PART A (submitted).


- 134 -

the local residents, the freight carriers and their organizations (Dablanc, 2011). Holguín-Veras et al. (2015) stress also the importance of considering the stakeholders: they find that, in order to have an effective policy that is likely to change the behaviour of the overall supply chain, all participants must agree on a common operational strategy. Hence, authorities should choose the set of measures that receives the best support from the stakeholders (Stathopoulos et al., 2012). However, we identify a lack of methods to involve stakeholders in a bottom up approach for developing integrated strategies for sustainable city logistics. We have observed several approaches but a formal methodology that considers the interests of stakeholders when selecting a solution is missing. In that context, the multi-actor multi-criteria analysis (MAMCA) can contribute to a better involvement of stakeholders. This methodology has already been used in the context of city logistics to evaluate different solutions based on the interest of stakeholders (Macharis, et al. 2014). This chapter aims therefore at demonstrating how integrating the MAMCA within the consultation process of stakeholders can be used to structure a bottom up approach in city logistics and guide stakeholders towards a consensus.

7.2 Stakeholders involvement in city logistics The importance of stakeholder involvement in policy designs has been discussed in a number of areas, ranging from urban development projects to solutions addressing air pollution (Soste et al., 2015; Yang, 2014). In city logistics also, their importance has often been highlighted as an enabler for various urban freight policies (Awasthi et al., 2016; Awasthi et al., 2011; Browne et al., 2005; Nordtømme et al., 2015; Lindholm and Browne, 2013; Holguín-Veras et al., 2014). The development of urban freight transport requires indeed considerable interaction between the public and private sector (Lindholm and Browne, 2013). Crainic et al. (2004) advocate therefore for a better collaboration and innovative partnerships between stakeholders in order to ensure a successful implementation of sustainable freight strategies. However, until recently there has been little involvement of private companies in the transport planning process (Lindholm and Browne, 2013). The focus of authorities was limited to a regulatory perspective (Vieira et al., 2015). Experience has shown that this lack of consideration for stakeholders has often led to ineffective efforts, unintended effects and antagonism between those making the rules - public authorities - and those trying to adapt to the rules - transport operators - (Holguín-Veras et al., 2014; Lindholm and Browne, 2013). For example, time-window schemes can in some cases lead to decreased transport efficiency, additional cost as well as increased pollution and congestion (Quak and De Koster, 2006). Ville et al. (2012) and van Duin et al. (2010) described also the significant opposition from the private sector following the introduction of restrictive regulations intended to support urban consolidation centers operations in Vicenza (Italy) and in Leiden (Netherlands) which led to a series of judicial proceeding in the first case and contributed to a project failure in the second case. Besides actors performing the deliveries, additional stakeholders should also be considered as well as their interactions between each other (Holguín-Veras et al., 2014). For example, time-of-day pricing schemes in New-York and London were expected to influence trip timing and reduce the pressure of freight traffic in peak hours. However it only resulted in the absorption of toll increase by carriers because of their inability to influence the receivers’ behaviour (Holguín-Veras, 2011; Holguín-Veras et al., 2006). The implementation of off-hour delivery programs in Manhattan was only possible by tailoring policies that targeted specifically both carriers and receivers of freight (Holguín-Veras, 2008).


- 135 -

Successful urban freight policies need therefore to account for the different stakeholders of city logistics. They are however not limited to the economic sphere: other stakeholders from the urban sphere can also be relevant (Routhier, 2002). Citizens in particular can become critical stakeholders in city logistics. The project of the Brussels International Logistics Centre met for example an important opposition from the local inhabitants, which led to the failure of the project (Mazy, 2014; Van Waes, 2011). In this framework, the involvement of all relevant stakeholders in the policy making process is increasingly recognised by authorities to be a critical factor of success for the implementation of new measures in city logistics (European Commission, 2012). A proper and honest stakeholder engagement enables to exchange information and knowledge in order to confirm the problems and the necessity of public intervention as well as to provide the public sector with a thorough idea of the constraints and expectations of the various stakeholders (Holguín-Veras et al., 2014; Lindholm and Browne, 2013). It allows public decision-makers to change or modify a course of action, fostering a collaborative environment and deploying implementation paths that can ensure a more successful transition towards more sustainable practices (Holguín-Veras et al., 2014). Furthermore, a participatory planning process improves the commitment to the project being discussed (Soste et al., 2015). Given the benefits of a bottom-up approach in city logistics, several approaches have been developed to involve stakeholders. The most generic approach to involve stakeholders is the establishment of platforms organised by the public sector where problems and measures concerning urban goods transport are discussed with the private sector. They are commonly referred as “Forums”, “Freight networks” or “Industry Advisory Groups” (OECD, 2003; Quak et al., 2015). They have been mostly used at a local level in countries like Belgium, the Czech Republic, Germany and the United States (European Commission, 2012; OECD, 2003). But consultation forums were also observed at a regional or national level such as in the Netherlands (“Platform Stedelijke Distributie”). Some hybrid forms exist also where national consultation forums are based on a network of municipalities. We can note in Denmark the “Forum for city logistics” and in France, the GART - Groupement des autorités responsables de transport (OECD, 2003). In these platforms, the interests of the different stakeholders are represented and solutions can be debated in order to reach a consensus or an agreement between stakeholders (Muñuzuri et al., 2005). An interesting output of these forums has been the development of Charters, which ensure that the solutions discussed in forums are respected through the commitment of the stakeholders to the charter. This document becomes then a materialisation of the stakeholders’ involvement. That approach was especially used in France. Based on a consultation of the different stakeholders of the logistics sector, the mayor of Paris published the first “charter of good practice for transport and goods deliveries” in 2006 (Mairie de Paris, 2006). It was then updated in 2013 (Mairie de Paris, 2013). The charter is based on key principles shared by the stakeholders. These principles are then operationalised in projects and policies proposed by the charter. An operational committee ensures afterwards that these propositions are discussed regularly with the signatories of the charter (around 47 members for the Charter in Paris). The city of Toulouse implemented a similar charter in 2012 (Toulouse, 2012). Some charters can also be more specific. The charter of CERTIBRUIT for example is addressing specifically the off-hour deliveries. By gathering the different stakeholders, this charter developed a list of important measures that


- 136 -

guarantee the quietness of the neighbourhood. The respect of these guidelines is valorised by a label (CERTIBRUIT, 2015). A similar approach to the charter is the development of strategic plans based on stakeholder concertation. The freight transport plan of the Brussels-Capital Region is in that context an interesting example (Bruxelles mobilité, 2013). In order to improve the sustainability of the logistics sector, the regional authorities developed a strategic plan by involving the different stakeholders from the beginning of that project. First meetings with professional of the industry have allowed the authorities to define the priorities in the sector and solutions were discussed in several workshops open to the different stakeholders of city logistics. The sector reacted positively to these workshops with an attendance of around 40 actors. That consultation could create a large support among stakeholders for the new freight plan of the Brussels-Capital Region. But perhaps the most elaborated approach to involve stakeholders in the development of solutions in city logistics has been the Freight Quality Partnerships (FQPs). It provides a framework to include the industry, authorities and representatives of different interest groups in the development of solutions regarding freight transport (Browne et al., 2007). It typically starts with a diagnostic phase where a survey is conducted among the different stakeholder groups in order to identify the main issues that they perceive regarding freight transport. Based on that diagnostic, the solutions that each stakeholder group can develop are identified. The measures that authorities can implement are also discussed. Several examples of FQPs that have been implemented in the United Kingdom are given by the Department for Transport (2003a). These experiences have shown that dialogue was improved between the different stakeholder groups and resulted in more efficient operations in freight transport (Allen et al., 2010). However, Browne et al. (2007) recognise also that this approach might be more difficult to use at a regional or national level. Engagement of stakeholders is indeed already limited at a local level. Nevertheless, authorities at higher levels can support these local initiatives by providing guidance and advice (OECD, 2003). That strategy was adopted by the government of the United Kingdom. They supported the scheme with several publications such as the guidelines on how to set up a FQP (Department for Transport, 2003b). As a result, the FQPs have been growing and are now used outside the UK with similar partnerships in Sweden, Holland and France for example (Ballantyne et al., 2013; Lindholm, 2014). Literature also provides some evidence of additional approaches and frameworks for managing the stakeholder consultation processes. Zunder et al. (2014) describe a Design and Monitoring Framework (or DMF) methodology that was applied to engage stakeholders in 3 city regions in Europe (Como in Italy, Berlin in Germany and Newcastle in the UK) within the SMARTFUSION project. Iwan (2014) proposed also an adaptive approach including eleven steps where stakeholders are identified and integrated in the development of the good practice. Finally, Quak et al. (2015) presented the concept of Living Lab as an action-driven form of stakeholder consultation.

7.3 Relevance of MAMCA for a bottom up approach in strategy formulation The different approaches that we have identified to involve stakeholders in city logistics have shown to improve dialogue and stimulate collaboration in the sector. They offer a space where stakeholders can debate and discuss new solutions for city logistics. However, that space should


- 137 -

be well organised; the discussion should be well structured in order to guarantee a significant contribution. Indeed, Lindholm (2014) highlights that these different consultations have the risk to become a “talking shop” that does not produce outcomes or progresses for city logistics. We identify therefore a need to improve that critical step of the consultation process of stakeholders. We identify a need to better integrate the different stakeholders groups in the discussion. In that context, the Multi-Actor Multi-Criteria Analysis (MAMCA) represents a relevant methodology that can structure the consultation process of stakeholders. An important benefit of that methodology is to focus the consultation process around the priorities of stakeholders in a first stage and discuss the solutions only in a next stage, based on these priorities. The objectives for each stakeholder group in city logistics are identified and their importance is assessed directly by the members of the group. These objectives are used as evaluation criteria for the different possible scenarios that can improve the sustainability of the sector. And the importance of these objectives is used to weight the performance of the scenarios on the evaluation criteria of each stakeholder group. As a result, the MAMCA produce a visualisation of the support that the different scenarios receive from each stakeholder group. This visualisation represents another key benefit of the MAMCA. It can indeed guide the discussion between the different stakeholders groups towards a consensus and avoid the debate to be focused on conflicting scenarios. The MAMCA can be considered as an extension of a multi-criteria analysis. Multi-criteria decision-aid (MCDA) can be used by public decision-makers to overcome the shortcomings of traditional decision-support tools used in economics (e.g. cost-benefit or cost-effectiveness analysis) and can support complex decision problems in areas such sustainability policies (Gamper and Turcanu, 2007; Munda et al., 1998) which require trade-offs between sociopolitical, environmental, ecological, and economic factors (Kiker et al., 2005; Lahdelma et al., 2014). Stakeholder management can be incorporated in the MCDA approaches and the need for public participation has been more and more recognised in an MCDA framework (Munda, 2006) with the development of techniques such as participatory multi-criteria evaluation (Banville et al., 1998) or social multi-criteria evaluation (Munda, 2004). In multiple criteria approach, the aim is to build several attributes/criteria considered to be representative of the various elements which the decision-makers consider in a decision-making process (Bouyssou, 1990). The particularity of MAMCA is the explicit use of stakeholders’ objectives as evaluation criteria. This way, their interests lie at the core of the methodology. Also, the visualisation of the results produces a multi-actor view where the results are not aggregated across stakeholders but are instead shown per stakeholder group. The objective of the MAMCA is indeed to identify the different position of stakeholders regarding the scenarios in order to find the solution that is the most supported by the different stakeholders. This methodology has been developed by Macharis (2007) and was mainly applied in transport evaluation projects (Macharis et al., 2009). It has also been used recently as an evaluation framework for comparing different European projects in the field of city logistics (STRAIGHTSOL, 2015).

7.4 Application to a stakeholder consultation in Brussels In this section, we describe the different steps of the MAMCA shown in Figure 50 and we illustrate them with an application of the methodology in a workshop conducted in Brussels with the different stakeholders of city logistics. The event was hosted at the freight division of the “regional commission of mobility”, a platform managed by the transport ministry of the Brussels-


- 138 -

Capital Region. That platform is dedicated to the consultation of stakeholders in the logistics sector in Brussels and was used, among others, for the development of the regional freight plan (Bruxelles mobilité, 2013). Figure 50: The seven steps of the MAMCA methodology (Source: Macharis, 2007)

Step 1: Define alternatives The Brussels-Capital Region has adopted in 2013 a new freight transport plan that has the ambition of rationalising freight flows and reducing their emissions (Bruxelles mobilité, 2013). One of the key measures proposed in this document is the establishment of a logistical pole at a tri-modal site Schaerbeek Formation as well as the establishment of a network of urban consolidation centres (UCCs) servicing the Region. UCCs are indeed a popular measure to improve the sustainability of city logistics: Allen et al. (2012) have referenced 114 implementation cases. However, their implementation was not always successful, mainly because of their inability to attract sufficient participation (Van Rooijen and Quak, 2010; Holguín-Veras and Sánchez-Díaz, 2015). A significant body of literature discusses the factors that influence the success of the UCC schemes, which include elements such as characteristics of the service area, regulatory framework, location of the UCC or the availability of subsidies (Quak and Tavasszy, 2011; Panero et al., 2011; Browne et al., 2005; Van Rooijen and Quak, 2010). As a result, the implementation of UCC(s) is often not a unique measure but is integrated and supported by a larger freight strategy. Several additional measures that are being considered in Brussels could therefore be integrated with the implementation of a UCC. Lebeau and Macharis (2014) show indeed that a road pricing scheme is expected to be implemented in April 2016 for heavy goods vehicles; night distribution is also one of the measures that receive the most important interest in the logistics sector in Belgium; finally, the purchase of electric vehicles is promoted by authorities. As a result, Janjevic et al. (2015) have reduced the wide range of possible combinations of measures for Brussels to a set of five possible scenarios that were considered relevant by the local authorities. We use therefore these scenarios as alternatives to be discussed with the stakeholders. They are represented in Figure 51.


- 139 -

Scenario 1: One UCC is introduced in the North of the Brussels-Capital Region. A distancebased road pricing scheme is applied on heavy goods vehicles. In this framework, the UCC is an alternative for transporter to avoid the higher cost of transport in urban areas. Diesel trucks are operating the last mile from the UCC. Scenario 2a: Two UCCs are introduced, one in the North, the other in the South of the Region. No road pricing scheme is applied. However, night distribution is allowed for the vehicles delivering the goods to the UCC, stimulating its use by carriers. Diesel trucks are operating the last mile from the UCC. Scenario 2b: Two UCCs are introduced, one in the North, the other in the South of the Region. No road pricing scheme is applied. However, night distribution is allowed for the vehicles delivering the goods to the UCC, stimulating its use by carriers. Battery electric trucks are operating the last mile from the UCC. Scenario 3a: Four UCCs are introduced, in the North, the South, the West and East of the Region. A distance-based road pricing scheme is applied on both heavy goods vehicles and light commercial vehicles, except for battery electric vehicles. The four UCCs offer viable alternatives for all transport operators to avoid the higher cost of transport in urban areas. Moreover, night distribution is allowed for the vehicles delivering the goods to the UCC, stimulating its use. Diesel vans are operating the last mile from the UCC. Scenario 3b: Four UCCs are introduced, in the North, the South, the West and East of the Region. A distance-based road pricing scheme is applied on both heavy goods vehicles and light commercial vehicles, except for battery electric vehicles. The four UCCs offer viable alternatives for all transport operators to avoid the higher cost of transport in urban areas. Moreover, night distribution is allowed for the vehicles delivering the goods to the UCC, stimulating its use. Battery electric vans are operating the last mile from the UCC.

Figure 51: Scenarios for the Brussels-Capital Region UCC location

Scenario 1 Road tax on HGV Night distribution not allowed Diesel truck

City centre

Zone within the ring road

Scenario 2a Scenario 2b No road tax Night distribution allowed Diesel truck Electric truck

Zone outside the ring road

Scenario 3a Scenario 3b Eco road tax on all Night distribution allowed Diesel van Electric van

In order to compare the scenarios with the current situation, we included also a Business As Usual (BAU) scenario. It does not consider any UCC in Brussels, night distribution is not allowed and no road pricing scheme is implemented.


- 140 -

Step 2: Stakeholders analysis In the second step of the MAMCA, stakeholders that are concerned with the scenarios described in step 1 are identified. However, Binsbergen & Visser (2001) stress the difficulty for local authorities of identifying the relevant stakeholders in city logistics. Hence, we first performed a literature review in order to identify a preliminary list of stakeholders. As Table 9 shows, the local authorities, citizens, receivers, shippers and freight carriers are the most popular stakeholders’ groups that are referred to in city logistics. But there are small differences across authors. Sometimes two groups of stakeholders are assimilated. Behrends (2011) for example merged shippers and receivers together as he considers their main interests to be similar (high accessibility between them). Also, Ballantyne et al. (2013) and Bjerkan et al. (2014) merged receivers and customers together as they consider them to be both the end destination of the last mile. But other stakeholders are sometimes considered. For example, Tamagawa et al. (2010) include the motorway operators in their multi agent model. Also, Witkowski and KibaJaniak (2014) add public transport operators in a survey of stakeholder’s perception on different policies that address city logistics. Though, these kinds of actors are considered by Ballantyne, Lindholm, & Whiteing (2013) to have only an indirect impact on city logistics. In their framework where relationships between stakeholders is explained, other stakeholders with an indirect effect on city logistics are identified such as vehicle manufacturers, trade associations, land owners, etc. But the main actors influencing directly the system remain the ones identified in Table 9, which are shippers, freight transport operators, authorities and customers (receivers and citizens). Macharis et al. (2014) have confirmed these five stakeholders to be the main groups in city logistics based on practical inputs of a European project dedicated to city distribution (STRAIGHTSOL, 2015). Step 3: Define evaluation criteria and weights with stakeholders Once the stakeholders have been identified in step 2, each stakeholder group is characterised in step 3 by their objectives and the importance that they associate to these objectives. Based on a deeper review of the literature presented in this section, we identified the main objectives of each stakeholder group. These objectives were then presented to the stakeholders of the logistics sector in Brussels in a small survey sent with the registration for the workshop. They were asked to select the stakeholder group they belong to (with an option “other” available) and comment the objectives we proposed for their stakeholder group based on our literature review. That survey allowed us to validate the stakeholder groups and include a few additional objectives for the authorities (consistency with urban planning), the citizens (safety), the transporters (congestion) and the receivers (congestion). The final list of objectives is shown in Table 10. During the workshop, the importance of the objectives was evaluated by each stakeholder group. The 32 participants were divided in five tables according to the group they belong to. By discussing together, each group could assess the weight they give to their different objectives based on a pairwise comparison method managed by the MAMCA software. The results of the pairwise comparison are also given in Table 10.


- 141 -

Table 9: Identification of stakeholders in city logistics. Local authorities Citizens Receivers Forwarders Administrators Residents Retailers Shippers Policy makers Customers (Ogden, 1992) (Tseng et al., 2005) (Muñuzuri et al., 2005) (Taniguchi et al., 2007) (Quak, 2008) (Tamagawa et al., 2010) (Behrends, 2011) (Stathopoulos et al., 2012) (Taniguchi et al., 2012) (Ballantyne et al., 2013) (Macharis et al., 2014) (Witkowski and Kiba-Janiak, 2014) (Bjerkan et al., 2014) TOTAL

             12

 

  

  

       



   

 8

8

10

Carriers Logistics service providers Transport operators Freight carriers              13

Others



 2

Logistics service providers Logistics service providers (LSPs) provide the transport services demanded by the shippers and/or the receivers. They are usually carriers but they can also be third party logistics operators that organize transport and subcontract the carriage of the goods (Stathopoulos et al., 2011). The drivers can also be included since they receive some autonomy in routing (Friedrich et al., 2003). The LSPs can be contracted by the shipper, by the receiver or by both if the shipper and the receiver agree on the transfer of risks at one place during the transportation. In any case, the LSPs are the designers of the transportation system (Behrends, 2011). They shape it by minimizing the cost of transportation and maximizing sales (Bjerkan et al., 2014). Therefore they organize transport as efficiently as possible by optimizing load capacity under the constraints of providing the quality of transport service required by their customers (PORTAL, 2003). LSPs also have to consider the constraints imposed by local authorities (Behrends, 2011). The main challenge of the carrier relies therefore in the planning of pick-up and delivery, select the shortest vehicle routing and minimizing operational costs (Bjerkan et al., 2014). But they also consider their environmental impact and the satisfaction of their employees in their decision (Macharis et al., 2012).


- 142 -

Table 10: Objectives of the stakeholder’s groups and their importance Actors

Receivers

Shippers

Logistics service providers

Citizens

Authorities

Objective

Weight

Transport costs Level of service One shot deliveries Congestion Transport emissions Transport costs Level of service Transport emissions Profitability Transport emissions Congestion Satisfaction of LSP employees Level of service Optimized load late Safety Noise level Space occupancy Transport emissions Congestion Cost for UCC support Business climate Congestion Consistency with urban planning Citizens support Noise level Transport emissions

11,6% 20,1% 26,3% 37,5% 4,6% 75,2% 19,7% 5,1% 38,7% 10,9% 13,3% 15,3% 11,4% 10,5% 30,7% 12,2% 13,5% 39,8% 3,8% 4,4% 13,1% 13,7% 3,3% 6,1% 20,3% 39,1%

Shippers Shippers can be manufacturers, wholesalers or retailers. They send the goods from the warehouses they operate. The goods are then delivered by the LSPs to the receivers. The main objective of shippers is to satisfy their customer’s needs, the receivers. In order to do this, they keep improving their product (Ballantyne et al., 2013). Besides product’s price and quality, shippers can also improve the accessibility of their product by organizing the transport for the receivers account (Dablanc, 2011). Shippers are therefore most of time the customers of LSPs (Melo and Costa, 2011). It simplifies the order for the receiver, contributing to the level of service of the shippers. Also, the cost of the transport service can be minimized as shippers have usually more bargaining power than fragmented receivers. This direct contact with LSPs make shippers a critical stakeholder as they determine the transport demand in terms of shipment size, frequency, lead-time, delivery precision and flexibility (Behrends, 2011). They can act on these operational aspects to minimize stock and optimize their infrastructure costs. But these operational aspects are sometimes directly designed by the shippers as they organise sometimes transport themselves through own-account transport (Ballantyne et al., 2013). Finally, Macharis et al., (2014) indicates that shippers dedicate also some attention to their environmental impact. Receivers Receivers can be either the final consumer in the urban supply chain or an intermediate player such as a retailer. If receivers are organizing the transport, they prefer a transport service with


- 143 -

low costs (Behrends, 2011; Macharis et al., 2012). But they usually do not deal with transport issues as shippers are often organizing it (Dablanc, 2011). Hence, they do not care much about the way goods are transported, as long as the delivery requirements are fulfilled (Melo and Costa, 2011). Still they can value the way the last mile is operated as they want to be located in an attractive urban environment (Ballantyne et al., 2013; Macharis et al., 2012). The environmental impact of deliveries is therefore one of their concern. But perhaps one of their most important concerns is the high level of service that they get from their deliveries. In case of regular orders, receivers expect a high frequency of deliveries in order to keep their stock at a minimum which are particularly expensive in urban areas (Anderson et al., 2005). In case of occasional orders, they expect a short delivery time (Stathopoulos et al., 2011). Also, they prefer a secure, safe, on-time and high reliable door-to-door service (Melo and Costa, 2011), preferably when there are not many customers in order to handle the deliveries and reduce the risk of theft or complaining customers (Macharis et al., 2012). In that context, the degree of deliveries grouping can be valorised by receivers. If all deliveries arrive in one shot, it will burden less the core business of the receiver. Citizens The population is increasingly moving in urban areas. As inhabitants, citizens are very much concerned about their living environment (Macharis et al., 2014). They value an urban environment that is not affected by trucks and vans traffic that are responsible of accidents, noise, emissions, visual intrusion and congestion (Macharis et al., 2012). They want quietness and clean air, in particular in their neighbourhood (Taniguchi et al., 2007). Authorities The authorities are elected by the citizens to design and manage the city. In the context of city logistics, their role is to finance the public infrastructure (e.g. roads, rail trucks, etc.) and to subsidize some services of public interest. They are also responsible for the land planning and the regulations: weight or size restrictions, time windows, low environmental zone or urban toll. Through the use of these tools, the main interests of authorities are about raising the overall socio-economic welfare and controlling the externalities of transport operations (Janic, 2006). This role is difficult because they have to match two conflicting interests. On the one hand, the government wants its city to be pleasant and attractive for the citizens with a reduced impact of freight transport on congestion, noise and emissions (Witkowski and Kiba-Janiak, 2014). On the other hand, the government wants the city to be well accessible for freight transport so it becomes an attractive location for shops, offices or restaurants (Taniguchi et al., 2007). This trade-off can be sometimes difficult knowing that the benefits of city logistics are less recognized by the voters than their negative impacts. This might justify some politically hard decisions when the situation makes necessary to implement effective – but vote loosing – measures (Begg and Gray, 2004). Still, authorities prefer solutions that are easily enforced and supported by the citizens (Macharis et al., 2012). Knowing that their resources are also limited, they also prefer to leverage the current infrastructure so measures are less costly. Step 4: Evaluation of the scenarios based on stakeholders criteria The objectives of the stakeholders are used in step 4 as evaluation criteria for the scenarios. In order to estimate the impact of the scenarios on these criteria, indicators are selected and described in this section and summarized in Table 11. We present also the tools that were used to


- 144 -

evaluate these indicators. Table 12 gives the impact of the scenarios on the criteria compared to the business as usual scenario. The authors achieved these evaluations before the workshop, once objectives of stakeholders were evaluated. The criterion “transport costs” uses as indicator the total cost paid by all transporters servicing the region: for those who choose to deliver on their own, it consists of the hourly and kilometric costs of transportation; for those who chose to use the UCC, it consists of the cost of transport to the UCC as well as the out-of-pocket expenses for the UCC services. This indicator was assessed with the location-allocation model described by Janjevic et al., (2015). The same indicator is used to assess the “profitability” criterion of the logistics service providers because we assumed that costs and benefits are equally distributed among supply chain actors in the case of an introduction of a UCC. Indeed, there is no clear evidence about the way these are reallocated between the shippers, receivers and the logistics service providers. Since an improvement in the cost and the profitability parameters of all actors leads to improving the general competiveness of the local economy, the same indicator is used to evaluate the authorities’ criterion “business climate”. It is to be noted that although the three aforementioned criteria essentially correspond to the same indicator (i.e. the total cost paid by all transporters servicing the region), they correspond to three different points of view (i.e. of receivers and shippers, carriers and authorities respectively). For the sake of keeping the vocabulary used by each stakeholder group, distinct criteria names were kept. Let us note also that the criterion “cost for UCC support” was always set to zero because the Region is not willing to support structurally a UCC or a network of UCCs. Janjevic et al. (2015) considered already that position of authorities and developed therefore scenarios with UCC(s) that show a positive business model without subsidies. The criteria “congestion”, “transport emissions”, “noise level” and “safety” all rely on the vehicle kilometres. The vehicle kilometres achieved by the transporters are estimated by the locationallocation model described in Janjevic et al. (2015) and the vehicle kilometres achieved by the UCC operator are estimated by the vehicle routing model developed by Lebeau et al. (2015). In order to estimate these different criteria, the vehicle kilometres are then coupled with different ratios. For evaluating the criterion “congestion”, the equivalent vehicle passenger assumptions are used on the vehicle kilometres to consider the more important contribution of large vehicles in the traffic compared to smaller vehicles (Routhier, 2002). Also, the vehicle kilometres driven before 6am are not considered since their marginal impact on congestion can be considered to be null. For the criterion “transport emissions”, the external costs of the CO2 emissions, PM emissions, NOx emissions, SO2 emissions and NMVOC emissions are estimated per type of vehicles in an urban environment based on the figures of the DG MOVE (2014). Likewise for the “noise level” criterion, the external cost of noise considered, in an urban context, the difference between vehicle types and the difference between day and night transport (DG MOVE, 2014). Also the “safety” criterion used as indicator the vehicle kilometres coupled with the estimations from the OGP (2010) on the number of serious injuries per kilometres in an urban context per type of vehicles.


- 145 -

Table 11: Summary of indicators and evaluation used for estimating the criteria

Quantitative indicators Transport costs Profitability Business climate Cost for UCC support Congestion

Transport emissions

Noise level

Safety

Space occupancy One shot deliveries Optimized load late Level of service Qualitative indicators Citizens support Fit with urban planning Satisfaction of LSP employees

Indicator

Evaluation method and references

Total cost of transport for deliveries Total cost of transport for deliveries Total cost of transport for deliveries n/a Total vehicle*km* (adjusted to vehicle size)

Location-allocation model (Janjevic et al., 2015) Location-allocation model (Janjevic et al., 2015) Location-allocation model (Janjevic et al., 2015) n/a Location-allocation model (Janjevic et al., 2015), Vehicle routing model (Lebeau et. al, 2015) and space occupancy factors (Routhier, 2002). Location-allocation model (Janjevic et al., 2015), Vehicle routing model (Lebeau et. al, 2015) and external cost model (DG MOVE, 2014)

External cost of CO2 emissions, PM emissions, NOx emissions, SO2 emissions and NMVOC emissions according External cost of noise

Location-allocation model (Janjevic et al., 2015), Vehicle routing model (Lebeau et. al, 2015) and external cost model (DG MOVE, 2014) Number of serious injuries Location-allocation model (Janjevic et al., 2015), Vehicle routing model (Lebeau et. al, 2015) and number of serious injuries model (OGP, 2010) Total time during the deliveries in the Location-allocation model (Janjevic et al., 2015) urban area Probability of bundling deliveries from Location-allocation model (Janjevic et al., 2015) different suppliers Average load rate of the vehicles Location-allocation model (Janjevic et al., 2015) and leaving their depot Vehicle routing model (Lebeau et. al, 2015) Frequency of service Vehicle routing model (Lebeau et. al, 2015) Scale Scale Scale

Estimation in consultation with regional authorities Estimation in consultation with regional authorities Estimation based on literature review

The criterion “space occupancy” is estimated based on the total time spent by the vehicles in the urban area for performing the deliveries. This time is estimated based on the locationallocation model described by Janjevic et al. (2015) which includes hypotheses about the duration of deliveries according to the vehicle category and the number of stops within a tour, based on results from Gerardin et al. (2000). The criterion “optimized load rate” is estimated based on the average of the load rate of the vehicles leaving their depot weighted by the number of kilometres driven by that vehicle. For the transporters that choose to deliver on their own, average load rates for different vehicle categories have been established based the assumptions of the location-allocation model described in Janjevic et al. (2015). For the UCC vehicles, the average load rate was estimated by the vehicle routing model developed by Lebeau et al. (2015). The criterion “one shot delivery” reflects the possibility of bundling several shipments from different suppliers during the same delivery and depends directly on the probability of two or more random shipments transiting through the same UCC. The location-allocation model


- 146 -

described in Janjevic et al. (2015) allows to calculate the probability of two random suppliers using the same UCC. Hence, the probability is zero when no UCC is implemented and the probability is reduced when more UCCs are introduced. The criterion “level of service” encompasses freight transport service attributes such as reliability of delivery, service frequency or absence of losses (Beuthe and Bouffioux, 2008). Elements such as reliability and absence of losses are difficult to estimate given the prospective nature of our scenarios, however, service frequency can be estimated based on the vehicle routing model developed by Lebeau et al. (2015). Moreover, experiences from UCC implementations show that the fact that receivers can decide on the delivery time is seen as an important benefit (Blom and van Nunen, 2009). Our scenarios consider therefore that the level of service is the best in scenarios 3a and 3b because vans operate four delivery turns per day which will give more chances of matching better the preferred timing of receivers. The chances are however lower in scenarios 1, 2a and 2b because the number of delivery turns is reduced to two per day. Trucks have indeed longer routes and more addresses to deliver given their higher payload. Finally, the business as usual considers only one delivery per day.

Table 12 : Evaluation of the scenarios compared to the business as usual scenario Scenario 1 Scenario 2a Scenario 2b Scenario 3a Scenario 3b Transport costs Profitability Business climate Cost for UCC support Congestion Transport emissions Noise level Safety Space occupancy One shot deliveries Optimized load late Level of service Citizens support Fit with urban planning Satisfaction of LSP employees

+6,3% +6,3% +6,3% +0,0% -2,2% -3,0% -2,2% -2,2% -5,3% +0,8% +5,6% ++ + + +

-1,3% -1,3% -1,3% +0,0% -4,1% -2,9% +5,6% -1,9% -6,6% +0,6% +5,9% ++ + ++ ++

-1,3% -1,3% -1,3% +0,0% -4,1% -3,7% +4,7% -2,0% -6,6% +0,6% +5,9% ++ ++ +++ ++

+13,4% +13,4% +13,4% +0,0% -7,2% -5,8% +3,9% -5,0% -13,1% +0,5% +8,8% ++++ + ++ ++

+13,4% +13,4% +13,4% +0,0% -7,1% -8,1% +2,0% -5,0% -13,1% +0,5% +8,8% ++++ ++ +++ ++

As Table 12 shows, some additional criteria were evaluated qualitatively. The criterion of authorities “citizen’s support” is evaluated on a scale from 0 to 2: if nothing is done, citizens’ support receives a score of 0, if at least one UCC is introduced, citizen’s support receives a score of 1 and if battery electric vehicles are combined with the UCC, it receives a score of 2. The criterion “consistency with urban planning” measures the extent in which a specific scenario follows the policy formulated in the regional urban freight strategy and planning documents (Bruxelles mobilité, 2013; PRDD/GPDO, 2013). In consultation with the regional authorities, the following scores were set: if nothing is done, the score is 0, if one UCC is


- 147 -

introduced, the score is 1, if multiple UCC are introduced the score is 2 and if battery electric vehicles are introduced with the UCCs, the score is 3. Finally, the criterion “satisfaction of LSP employees” is evaluated on a scale from 0 to 2: if no UCCs are offering last mile services, the criterion receives a score of 0, if last mile services are available to avoid the congested area, the score is 1 and if drivers have the possibility to deliver the UCC earlier in the morning, outside peak hours, the score is 2. This is in line with findings from literature that indicate that stress levels for delivery drivers are lower in more peripheral areas than in the downtown (Kawamura et al., 2014) and during the off-hour deliveries (HolguinVeras, 2014). Step 5: Analyses Since the criteria of the stakeholders group were validated and evaluated before the workshop, the results of the MAMCA could directly be presented in the workshop. Once the weights were collected from the stakeholders group, the MAMCA software could compute the overall evaluation scores of the different scenarios for each stakeholder group based on Promethee. Step 6: Results The results of the MAMCA are summarised in the multi-actor view depicted in Figure 52. This figure was presented in the workshop to the different stakeholder groups as a result of their internal discussion where they assessed the importance of their objectives. The multi-actor view can then guide the discussion between the different stakeholder groups towards a consensus. The visualisation helps identifying the support of each stakeholder for the different scenarios. A higher evaluation score means a scenario that fulfils better the objectives of one stakeholder group. For example, we can see that receivers, citizens and authorities show the highest support for the scenario 3b. That scenario is however the worst alternative for the shippers. It is also not the preferred one by the LSPs. That scenario is therefore not guaranteed to be successful. The main reason for the lack of support by LSPs and shippers is found in the introduction of a distance-based road pricing scheme that leads to increased transport costs. The step 3 assessed that the LSPs and in particular the shippers are very sensitive to that criterion. This scenario illustrates well that a solution for one actor can lead to new problems for another actor as explained by Browne and Allen (1999). An alternative approach to analyse Figure 52 is to observe the patterns of the scenarios across stakeholders. We can notice that the scenario 2a and scenario 2b share a similar curve shape with the business as usual (BAU). But scores of scenario 2a and 2b are higher than the scores of the business as usual. Shifting from the business as usual to scenario 2a and 2b improves therefore the situation of every stakeholder group. Hence, we can identify the scenario 2b to be an interesting compromise for all stakeholders.


- 148 -

Figure 52: Multi-actor view of stakeholder’s preferences for the scenarios

However, the transport sector is undergoing important changes. The three Regions of Belgium (thus also the Brussels-Capital Region) are currently implementing a distance-based road pricing scheme on heavy goods vehicles. They receive strong opposition from LSPs but little space is left for discussion. According to the current plans, the new regulation should be active in April 2016. Besides, a UCC has been launched in September 2014 and is currently growing. These market changes are best represented by scenario 1. Figure 52 can therefore show the impact that these evolutions have on the stakeholders. Compared to the business as usual, the scenario 1 is particularly supported by the authorities, the citizens and the receivers. However, shippers prefer the business as usual given their sensitivity to transport costs. Still, it is interesting to see that LSPs support scenario 1 although a distance-based road pricing scheme is included. The difference between LSPs and the shippers is coming from the positive effects of the scenario 1 on the logistics operations that are more valued by the LSPs than by the shippers. Indeed, Table 12 showed that congestion decreased, the level of service improved and load rate also. As a result, the downsides of scenario 1 are compensated by their benefits according LSPs. Figure 53: Uni-actor view of LSPs evaluations for the scenarios


- 149 -

When the pattern of scenario 1 is compared with the pattern of the business as usual, we can see that the shapes of the two curves are almost reversed, which demonstrates the importance of the expected changes brought by these new measures in the logistics sector. Such modification in the baseline scenario could therefore be a game changer with regards to the acceptance of subsequent measures: assuming that the transport sector will adopt that new structure, scenario 3b might receive a better support than in the current context. The pattern of scenario 3b is indeed similar to scenario 1 and its evaluation scores are in general higher than scenario 1. Shifting from scenario 1 to scenario 3b improves therefore the situation for every stakeholder, except for the shippers. The sensitivity of the shippers on transport costs explain again that barrier for shifting from scenario 1 to scenario 3b. The impact of the road pricing scheme is indeed higher since it is extended to the vans and not limited anymore to the heavy goods vehicles as in scenario 1. We can also notice that transporters prefer scenario 3b and 3a despite the increased transport costs. That position of the LSPs can again be explained by the improvement of the logistics operations that compensates the increased transport costs. In order to shift from scenario 1 to scenario 3b, it is therefore essential to convince the shippers of their logistics impact on the city. Since the position of LSPs is sometimes ambiguous, we can push the multi-actor multi-criteria analysis further by investigating the preferences of that particular actor group in the different scenarios. Figure 53 shows the uni-actor view of the LSPs where are displayed the importance that this group associate to their objectives as well as the performance of each scenario on these objectives. The final score – used in the multi actor view – depicts for each scenario the total performance of their criteria weighted by their importance. Hence, the uni-actor view reveals the underlying reasons of the stakeholder position shown in the multi-actor view. Figure 53 shows for example why the business as usual is the least preferred scenario for LSPs. The current strong opposition of LSPs against the road pricing scheme would suggest that business as usual is preferred over scenario 1. Indeed, their main objective, profitability, is seriously affected by the scenario 1. But Figure 53 shows also that LSPs consider several other criteria that have a minor influence in their evaluation but together, they represent a weight of more than 60% in the total evaluation. Since scenario 1 improves each one of these criteria, the overall score of scenario 1 is higher than the overall score of the business as usual. That difference is pushed to the extreme with scenario 3a and 3b where profitability is even more affected but other criteria are much more improved which result in a better overall score. Nevertheless, we observe that the most preferred scenario is scenario 2a and 2b as it gives the most profitable scenario for LSP and improves reasonably the other criteria compared to the BAU scenario. It can be considered as the scenarios with the best trade-offs between the different objectives of the LSP. Finally, we can stress the contribution of the battery electric vehicles on the acceptance of the different scenarios. By observing in Figure 52 the difference between scenario 2a and 2b and the difference between the scenario 3a and 3b, we can see that citizens and in particular the authorities value very positively the introduction of battery electric vehicles. Hence, authorities should be pro-active in stimulating the introduction of battery electric vehicles as they value it the most. The UCC operator is indeed supporting alone the costs of these vehicles in the scenarios, which reduce its gross margin (Janjevic et al., 2015). Authorities should therefore give to the UCC operator incentives to use battery electric vehicles. However, Figure 52 shows also that LSPs, receivers and shippers do not have a significant preference for a scenario with or without battery


- 150 -

electric vehicles. Although battery electric vehicles bring an important reduction on transport emissions and noise according Table 12, the overall score barely change between a scenario with or without electric vehicles. These criteria receive indeed only a limited attention among these stakeholders as depicted in Table 10. Nevertheless, we can see that the different stakeholders of city logistics support positively an electrification of city logistics. That result deserves to be stressed given the often conflicting views across stakeholders.

7.5 Discussion Importance of the bottom up approach The MAMCA was previously used especially as an evaluation tool to support the decision-making process and choose, among a set of scenarios, the one that matches the best the objectives of stakeholders. Within city logistics, the MAMCA was used as evaluation framework for the research project STRAIGHTSOL. Among others, it evaluated the introduction of a UCC in L’Hospitalet de Llobregat in Barcelona (STRAIGHTSOL, 2014). Although there is apparent similarity with the evaluation in Brussels (both MAMCA evaluates the introduction of UCC(s)), there are some notable differences in the stakeholder groups and choice of criteria. For example, the biggest preoccupations of shippers in Brussels are the cost (75.2%) and service quality (19.7%), which seems natural when considering policies that include road pricing and UCCs. In Barcelona, no road pricing was considered and negotiated prices were not impacted, resulting in a most predominant importance for the level of service (75.1%) and the omission of the cost criterion for the shippers. We observe therefore that the exact stakeholder preoccupations are extremely context-dependant, which demonstrates the need to include stakeholders from the very beginning of the policy-building process rather than using a pre-established list of criteria. The differences in stakeholder preoccupations, however small they might be, may in fact result in very different options with regards to policy design and support instruments. Contribution of MAMCA in the consultation process The goal of the chapter is however not primarily focused on showing the importance of including stakeholders in the development of freight strategies in cities. The objective was first to demonstrate the contribution of the MAMCA in improving the consultation process of stakeholders. It does not aim at replacing the previous approaches we have identified in the beginning of the chapter but rather at complementing them. A MAMCA could for example be integrated within a Freight Quality Partnership. The application of the methodology to a workshop in Brussels has demonstrated the benefits of the method especially in structuring the discussion and guiding the stakeholders towards a consensus. The consultation process was mainly organised around three key steps in the MAMCA methodology. The first step (1) was a preparation phase before the workshop took place. We validated the different stakeholder groups and their most important objectives though a small survey that they received when registering for the workshop. Based on that input, the authors could evaluate before the workshop the different scenarios according their impacts on the objectives of the stakeholders. The two following steps were then organised within the workshop. Stakeholders were first divided in tables according the stakeholder group they belong to. The workshop was introduced


- 151 -

by presenting the different scenarios and the objectives that each stakeholder group had validated in the survey. The second step (2) of the consultation process followed that introduction and consisted in a table discussion per stakeholder group where the members were asked to assess together the weights of their objectives. That step allowed the workshop to focus the discussion first on the priorities of stakeholders regarding the different scenarios proposed. The MAMCA software could assist this discussion with a pairwise comparison and use these results to compute directly the multi-actor view thanks to the evaluations of the scenarios achieved before the workshop. The third step (3) of the consultation process could therefore address the different strategies based on the results of the MAMCA. Because the results show the performance of the different scenarios based on the objectives of stakeholders, the multi-actor view translate the degree of support of each stakeholder group for the different scenarios. The discussion could therefore be guided around the scenarios that show the largest consensus. We could focus the debate around the scenario 2b and 3b that received the most interesting support from stakeholders and avoid conflicting discussions around the scenarios that match the interest of a limited number of stakeholder groups. Towards implementation? Once the results of the MAMCA have been discussed, there is however one last step in the MAMCA methodology that we did not develop in the application to Brussels. According, Figure 50 there is still the step 7 which aims at adapting the best scenarios for their final implementation or iterating the method with improved scenarios. The assessment of the most important objectives of stakeholders is in that context particularly valuable. The scenarios can be better adapted by considering both the discussion between stakeholders and their priorities. For example, if considering the most environmental configuration (scenario 3b), the results have shown that the biggest opposition to that scenario is expected to come from the shippers: they are indeed particularly sensitive to delivery costs which are increased because of the introduction of the road freight pricing. It is therefore necessary to provide some incentives in order to guarantee their commitment to the new scheme. One such incentive could be changing the regulations on the off-hour deliveries that are currently not allowed in Brussels combined with the exemption of the road pricing during the off-peak hours. This would bring down costs of deliveries for the shippers and increase their adherence to the scheme. However, receivers are not always disposed to receive off-hour deliveries (Verlinde, 2015). Authorities can therefore use funds from the toll collection in order to provide to receivers financial incentives to accept off-hour deliveries. An example of such a self-supported freight demand management system that uses a small toll surcharge to generate an incentive budget that, in turn, is used to induce receivers of goods in congested urban areas to accept unassisted offhour deliveries is described by Holguín-Veras and Aros-Vera (2014). Funds from the toll collection could also be used to give incentives to operate electric vehicles in city logistics. Janjevic et al. (2015) showed indeed that the business model of the UCC operator is affected in scenarios that consider the use of a fully electrified fleet instead of a diesel fleet. Given the global positive support of stakeholders for electric vehicles, the toll could be used to induce a shift from diesel towards battery electric vehicles among logistics service providers. The first iteration of MAMCA can therefore be used to design a more optimal set of scenarios, which can in turn be evaluated and presented to stakeholders through an iterative process. May


- 152 -

(2009) highlights that the policy-making process can be presented as an iteration starting with the definition of objectives and ending with the evaluation of solutions. The application of the MAMCA in Brussels shows that this methodology can be used from the very beginning of this process. Apart from allowing a design of more optimal scenarios from a mathematical point of view, the iteration of the MAMCA creates also a higher commitment to the policy schemes by clearly demonstrating the inclusion of the different viewpoints of stakeholders in the policy building process. Limitations of the approach There are however some precautions to the use of this methodology. A first precaution regards the difficulties of establishing homogenous and representative stakeholder groups. On the one hand, stakeholders can often play numerous roles (e.g. shipper and carriers, citizen and receiver) and on the other hand, there can be significant differences within each stakeholder groups (e.g. all shippers are not equal and their preoccupations might vary across the sub-groups). The selection of stakeholder groups might therefore change across applications. Hence, we stress the importance of validating the groups with the stakeholders. A second precaution regards the evaluation of the scenarios. The selection of criteria is left to stakeholders so that they represent their objectives. That approach can however lead to an establishment of imprecise or overlapping criteria. Also, it can lead to difficulties in choosing appropriate indicators for their criteria as well as their evaluation methodologies. In some cases (e.g. cost or congestion impact), the choice is more straightforward and we can rely on quantitative methods and models. However, in other cases, in particular those concerning qualitative criteria (e.g. quality of service, fit with urban planning, satisfaction of LSP employees), the choices of indicators and evaluation methodologies are particularly challenging. This is particularly true in the case of ex-ante analysis where these qualitative criteria cannot be preassessed by the stakeholders themselves. Nevertheless, leaving the choice of criteria to each stakeholder group ensures a higher degree of commitment to the results generated by the MAMCA and involves better the stakeholders in a bottom-up approach.

7.6 Conclusions The objective of the chapter is to contribute to the development of a bottom-up approach where stakeholders are involved in the design of more sustainable strategies for city logistics. The authors presented the MAMCA as a method which structure the consultation process of stakeholders. This approach was tested in a workshop in Brussels and the chapter uses that application as illustration of the methodology. Finally, the lessons learned from that application were discussed. The MAMCA was found to bring several important benefits to the consultation process of stakeholders. First, the methodology helps in identifying the priorities of the stakeholder groups regarding the different scenarios considered. The stakeholders are indeed asked to assess the importance of their objectives with the members of their stakeholder group. These weights are then used to compute the results of the MAMCA and show the position of each stakeholder group regarding the different scenarios. The multi-actor view is therefore a second important benefit of the MAMCA. That visualisation of the results can open the discussion to all stakeholders and guide the debate towards the most supported strategy. Finally, the inputs


- 153 -

collected by the MAMCA through the consultation process provide a good basis to decide on the implementation of one strategy or on a further adaptation of the scenarios based on the priorities of the stakeholders.


- 154 -

References Allen, B., Browne, M., Piotrowska, M., Woodburn, A.G., (2010), Freight Quality Partnerships in the UK - an analysis of their work and achievements, Green Logistics. Allen, J., Browne, M., Woodburn, A., Leonardi, J., (2012), The Role of Urban Consolidation Centres in Sustainable Freight Transport. Transp. Rev. 32, 473–490. doi:10.1080/01441647.2012.688074 Anderson, S., Allen, J., Browne, M., (2005), Urban logistics––how can it meet policy makers’ sustainability objectives? Journal of Transport Geography, 13, pp. 71–81. Awasthi, A., Adetiloye, T., Crainic, T.G., (2016), Collaboration partner selection for city logistics planning under municipal freight regulations. Appl. Math. Model. 40, 510–525. Awasthi, A., Chauhan, S.S., Goyal, S.K., (2011), A multi-criteria decision making approach for location planning for urban distribution centers under uncertainty. Math. Comput. Model. 53, 98–109. Ballantyne, E.E.F., Lindholm, M., Whiteing, A., (2013), A comparative study of urban freight transport planning: addressing stakeholder needs. Journal of Transport Geography, 32, pp. 93–101. Banville, C., Landry, M., Martel, J.-M., Boulaire, C., 1998. A stakeholder approach to MCDA. Systems Research and Behavioral Science 15, 15–32. Begg, D., Gray, D., (2004), Policy instruments for achieving sustainable transport. Issues on Environmental Science and Technology, 20, pp. 65–80. Behrends, S. (2011), Urban freight transport sustainability: The interaction of urban freight and intermodal transport. Chalmers University of Technology, Gothenburg. Bestfact, (2013), Best Practice Handbook 1, Deliverable 2.2. Beuthe, M., Bouffioux, C., (2008), Analysing qualitative attributes of freight transport from stated orders of preference experiment. J. Transp. Econ. Policy 105–128. Binsbergen, A. Van, Visser, J. (2001), Innovation Steps Towards Efficient Goods Distribution Systems for Urban Areas. Delft University. Bjerkan, K.Y., Sund, A.B., Nordtømme, M.E., (2014), Stakeholder responses to measures green and efficient urban freight. Research in Transportation Business & Management. 11, pp. 32–42. Blom, F.M., van Nunen, I.J., (2009), Determining the value of Binnenstadservice. nl for their customers. Erasmus Universiteit. Bouyssou, D., 1990. Building criteria: a prerequisite for MCDA, in: Readings in Multiple Criteria Decision Aid. Springer, pp. 58–80. Browne, M., Allen, J., (1999), “The impact of sustainability policies on urban freight transport and logistics system”, In: Meermans, H., Van de Voorde, E., Winkelmans, W. (eds.), World Transport Research: Selected Proceedings of the 8th World Conference on Transport Research, pp. 505–518. Browne, M., Piotrowska, M., Woodburn, A., Allen, J., (2007), Literature Review WM9 : Part I - Urban Freight Transport, Transport. London. Browne, M., Sweet, M., Woodburn, A., Allen, J., (2005), Urban freight consolidation centres final report. Bruxelles mobilité, (2013), Plan stratégique pour le transport de marchandises en Région de Bruxelles-Capitale.


- 155 -

CERTIBRUIT, (2015), Charte livraisons de nuit respectueuses des riverains et de l’environnement. Crainic, T.G., Ricciardi, N., Storchi, G., (2004), Advanced freight transportation systems for congested urban areas. Transp. Res. Part C Emerg. Technol. 12, 119–137. doi:10.1016/j.trc.2004.07.002 Dablanc, L., (2011) “City distribution, a key element of the urban economy: guidelines for practitioners”, In: Macharis, C., Melo, S. (eds.), City Distribution and Urban Freight Transport : Multiple Perspectives, Edward Elgar, Cheltenham, UK, pp. 13–36. Department for Transport, 2003a. Freight Quality Partnerships Case Studies. Department for Transport, 2003b. A guide on how to set up and run Freight Quality Partnerships. DG MOVE, (2014), Update of the Handbook on External Costs of Transport. EC, (2011), White paper: Roadmap to a Single European Transport Area. Brussels. European Commission, (2012), DG MOVE: Study on Urban Freight Transport. Friedrich, M., Haupt, T., Noekel, K., (2003) “Freight modelling: data issues, survey methods, demand and network models”, In: 10th International Conference on Travel Behavior Research. Lucerne. Gamper, C.D., Turcanu, C., 2007. On the governmental use of multi-criteria analysis. Ecological Economics, Special Section: Ecological-economic modelling for designing and evaluating biodiversity conservation policiesEE Modelling Special Section 62, 298–307. doi:10.1016/j.ecolecon.2007.01.010 Gerardin, B., Patier, D., Routhier, J.-L., Segalou, E., (2000), Diagnostic du transport de marchandises dans une agglomération - Document technique n°1. Gonzalez-feliu, J., (2011), Costs and benefits of logistics pooling for urban freight distribution : scenario simulation and assessment for strategic decision support 1–24. Hesse, M., (1995), Urban space and logistics: on the road to sustainability? World Transport Policy and Practice, 1, pp. 39–45. Holguín-Veras, J., (2008), Necessary conditions for off-hour deliveries and the effectiveness of urban freight road pricing and alternative financial policies in competitive markets. Transp. Res. Part Policy Pract. 42, 392–413. doi:10.1016/j.tra.2007.10.008 Holguín-Veras, J., (2011), Urban delivery industry response to cordon pricing, time–distance pricing, and carrier– receiver policies in competitive markets. Transp. Res. Part Policy Pract. 45, 802–824. Holguin-Veras, J., (2014), The Off-Hours Delivery Project in New York City. Holguín-Veras, J., Aros-Vera, F., (2014), Self-supported freight demand management: pricing and incentives. EURO J. Transp. Logist. 1–24. doi:10.1007/s13676-013-0041-1 Holguín-Veras, J., Aros-Vera, F., Browne, M., (2015), Agent interactions and the response of supply chains to pricing and incentives. Economics of Transportation. (in press) Holguín-Veras, J., Sánchez-Díaz, I., (2015), Freight Demand Management and the Potential of Receiver-Led Consolidation programs. Transp. Res. Part Policy Pract. doi:10.1016/j.tra.2015.06.013 Holguín-Veras, J., Wang, C., Browne, M., Hodge, S.D., Wojtowicz, J., (2014), The New York City Off-hour Delivery Project: Lessons for City Logistics. Procedia-Soc. Behav. Sci. 125, 36–48.


- 156 -

Holguín-Veras, J., Wang, Q., Xu, N., Ozbay, K., Cetin, M., Polimeni, J., (2006), The impacts of time of day pricing on the behavior of freight carriers in a congested urban area: Implications to road pricing. Transp. Res. Part Policy Pract. 40, 744–766. doi:10.1016/j.tra.2005.12.011 Iwan, S., (2014), Adaptative Approach to Implementing Good Practices to Support Environmentally Friendly Urban Freight Transport Management. Procedia - Soc. Behav. Sci. 151, 70–86. doi:10.1016/j.sbspro.2014.10.009 Janic, M., (2006), Sustainable transport in the European Union: a review of the past research and future ideas. Transport Reviews, 26, pp. 81–104. Janjevic, M., Kaminsky, P., Ballé Ndiaye, A., (2013), Downscaling the consolidation of goods – state of the art and transferability of micro-consolidation initiatives. Janjevic, M., Lebeau, P., Ndiaye, A.B., Macharis, C., Van Mierlo, J., (2015), Strategic Scenarios for Sustainable Urban Distribution in the Brussels-Capital Region Using Urban Consolidation Centres. Procedia - Social and Behavioral Sciences, Ninth International Conference on City Logistics 17-19 June 2015, Tenerife, Spain. Kawamura, K., Sriraj, P.S., Surat, H.R., Menninger, M., (2014), Analysis of Built Environment Features and their Effects on Freight Activities. Procedia - Soc. Behav. Sci., Eighth International Conference on City Logistics 17-19 June 2013, Bali, Indonesia 125, 28–35. doi:10.1016/j.sbspro.2014.01.1453 Kiker, G.A., Bridges, T.S., Varghese, A., Seager, T.P., Linkov, I., 2005. Application of multicriteria decision analysis in environmental decision making. Integr Environ Assess Manag 1, 95–108. doi:10.1897/IEAM_2004a-015.1 Lahdelma, R., Salminen, P., Hokkanen, J., 2014. Using Multicriteria Methods in Environmental Planning and Management. Environmental Management 26, 595–605. doi:10.1007/s002670010118 Lebeau, P., Cauwer, C. De, Van Mierlo, J., Macharis, C., Verbeke, W., Coosemans, T., (2015), Conventional, hybrid or electric vehicles: what technology for an urban distribution centre? The Scientific World Journal, Vol. 2015, 11 pages. Lebeau, P., Macharis, C., (2014), Le transport de marchandises à Bruxelles : quels impacts sur la circulation automobile ? Brussels Studies, 80, pp. 1–14. Lindholm, M., (2014), Successes and Failings of an Urban Freight Quality Partnership – The Story of the Gothenburg Local Freight Network. Procedia Soc. Behav. Sci. 125, 125–135. doi:10.1016/j.sbspro.2014.01.1461 Lindholm, M., Browne, M., (2013), Local authority cooperation with urban freight stakeholders: A comparison of partnership approaches. Eur. J. Transp. Infrastruct. Res. 13, 20–38. Macharis, C., (2007) “Multi-criteria analysis as a tool to include stakeholders in project evaluation: The MAMCA method”, In: Haezendonck, E. (ed.), Transport Project Evaluation, Edward Elgar Publishing. Macharis, C., de Witte, A., Ampe, J., (2009), The multi-actor, multi-criteria analysis methodology (MAMCA) for the evaluation of transport projects: Theory and practice. Journal of Advanced Transportation, 43, pp. 183–202. Macharis, C., Milan, L., Verlinde, S., (2012), Straightsol Deliverable D3.2. Stakeholders, criteria and weights. Macharis, C., Milan, L., Verlinde, S., (2014), A stakeholder-based multicriteria evaluation framework for city distribution. Research in Transportation Business & Management, 11, pp. 75–84. Mairie de Paris, (2006), Charte de bonnes pratiques des transports et des livraisons de marchandises dans Paris. Mairie de Paris, (2013), Charte en faveur d’une logistique urbaine durable.


- 157 -

Mazy, K., (2014), Villes et ports fluviaux: le projet comme dispositif de reconnexion? Regards croisés sur Bruxelles et Lille. Université des Sciences et Technologie de Lille-Lille I; Université Libre de Bruxelles. May, A., 2009. Improving decision-making for sustainable urban transport: an introduction to the DISTILLATE research programme. EJTIR 3. Melo, S., Costa, A., (2011) “Definition of a set of indicators to evaluate the performance of urban goods distribution initiatives”, In: Macharis, C., Melo, S. (eds.), City Distribution and Urban Freight Transport : Multiple Perspectives, Edward Elgar, Cheltenham, UK, pp. 120–147. Munda, G., Nijkamp, P., Rietveld, P., 1998. Environmental decision making: a comparison between cost-benefit analysis and multicriteria decision aid, in: Sustainable Development: Concepts, Rationalities and Strategies. Springer, pp. 227–238. Munda, G., 2004. Social multi-criteria evaluation: Methodological foundations and operational consequences. European journal of operational research 158, 662–677. Munda, G., 2006. Social multi-criteria evaluation for urban sustainability policies. Land Use Policy, Resolving Environmental Conflicts:Combining Participation and Muli-Criteria Analysis 23, 86–94. doi:10.1016/j.landusepol.2004.08.012 Muñuzuri, J., Larrañeta, J., Onieva, L., Cortés, P., (2005), Solutions applicable by local administrations for urban logistics improvement. Cities 22, pp. 15–28. Nordtømme, M.E., Bjerkan, K.Y., Sund, A.B., (2015), Barriers to urban freight policy implementation: The case of urban consolidation center in Oslo. Transp. Policy 44, 179–186. doi:10.1016/j.tranpol.2015.08.005 OECD, (2003), Delivering the Goods: 21st century challenges to urban goods transport. Ogden, K.W. (1992) Urban Goods Movement: A Guide to Policy and Planning. University Press, Cambridge, United Kingdom. OGP, (2010), Risk Assessment Data Directoty - Land Transport accident statistics. Panero, M.A., Shin, H.-S., Lopez, D.P., (2011), Urban Distribution Centers - a Means to Reducing Freight Vehicles Miles Travelled. Patier, D., Browne, M., (2010), A methodology for the evaluation of urban logistics innovations. Procedia - Soc. Behav. Sci., The Sixth International Conference on City Logistics 2, 6229–6241. doi:10.1016/j.sbspro.2010.04.033 PORTAL, (2003), Inner Urban Freight Transport and city logistics. PRDD/GPDO, 2013. Projet de Plan Régional de Développement Durable. Quak, H. (2008) Sustainability of Urban Freight Transport - Retail Distribution and Local Regulations in Cities. Erasmus Universiteit Rotterdam. Quak, H., (2011), Urban freight transport: the challenge of sustainability, in: Macharis, C., Melo, S. (Eds.), City Distribution and Urban Freight Transport : Multiple Perspectives. Edward Elgar, Cheltenham, UK, pp. 37–55. Quak, H., Lindholm, M., Tavasszy, L., Browne, M., (2015), From freight partnerships to city logistics living labs – Giving meaning to the elusive concept of living labs. Presented at the 9th International Conference on City Logistics. Quak, H., Tavasszy, L., (2011), Customized solutions for sustainable city logistics: The viability of urban freight consolidation centres, in: Transitions towards Sustainable Mobility. Springer, pp. 213–233.


- 158 -

Quak, H.J., De Koster, M.B.M., (2006), Urban distribution: The impacts of different governmental time-window schemes. ERIM Rep. Ser. Ref. No ERS-2006-053-LIS. Routhier, J.L., (2002), Du transport de marchandises en ville à la logistique urbaine. Soste, L., Wang, Q.J., Robertson, D., Chaffe, R., Handley, S., Wei, Y., (2015), Engendering stakeholder ownership in scenario planning. Technol. Forecast. Soc. Change 91, 250–263. Stathopoulos, A., Valeri, E., Marcucci, E., (2011), “Urban freight policy innovation for Rome ’ s LTZ : a stakeholder perspective”, In: Macharis, C., Melo, S. (eds.), City Distribution and Urban Freight Transport : Multiple Perspectives, Edward Elgar, Cheltenham, UK, pp. 75–100. Stathopoulos, A., Valeri, E., Marcucci, E., (2012), Stakeholder reactions to urban freight policy innovation. Journal of Transport Geography, 22, pp. 34–45. STRAIGHTSOL, (2014), Deliverable 5.4. Final evaluation of all STRAIGHTSOL city distribution concepts by use of the MAMCA. STRAIGHTSOL, (2015), Straightsol - Final publishable report. Tamagawa, D., Taniguchi, E., Yamada, T., (2010), Evaluating city logistics measures using a multi-agent model. Procedia – Social and Behavioral Sciences, 2, pp. 6002–6012. Taniguchi, E., Okamoto, M., Yamada, T., (2007), Multi-agent modelling for evaluating dynamic vehicle routing and scheduling systems. Journal of the Eastern Asia Society for Transportation Studies, 7, pp. 933–948. Taniguchi, E., Thompson, R.G., Yamada, T., (2012), Emerging Techniques for Enhancing the Practical Application of City Logistics Models. Procedia - Social and Behavioral Sciences, 39, pp. 3–18. Toulouse, M. de, (2012), Charte livraisons centre-ville. Tseng, Y., Taylor, M.A.P., Yue, W.L., (2005), The role of transportation in logistics chain. Proceedings of the Eastern Asia Society for Transportation Studies, 5, pp. 1657–1672. van Duin, J.H.R., Quak, H., Muñuzuri, J., (2010), New challenges for urban consolidation centres: A case study in The Hague. Procedia - Soc. Behav. Sci. 2, 6177–6188. doi:10.1016/j.sbspro.2010.04.029 Van Rooijen, T., Quak, H., (2010), Local impacts of a new urban consolidation centre–the case of Binnenstadservice. nl. Procedia-Soc. Behav. Sci. 2, 5967–5979. Van Waes, N., (2011), Les discours environnementaux produits par les acteurs portuaires et logistiques urbains : une analyse critique du cas bruxellois ou Le BILC : les discours à l’épreuve des faits. Université Libre de Bruxelles. Verlinde, S. (2015). Promising but challenging urban freight transport solutions. Vrije Universiteit Brussel. Vieira, J.G.V., Fransoo, J.C., Carvalho, C.D., (2015), Freight distribution in megacities: Perspectives of shippers, logistics service providers and carriers. J. Transp. Geogr. 46, 46–54. doi:10.1016/j.jtrangeo.2015.05.007 Ville, S., Gonzalez-Feliu, J., Dablanc, L., (2012), The limits of public policy intervention in urban logistics: Lessons from Vicenza (Italy). Eur. Plan. Stud. 1–14. Witkowski, J., Kiba-Janiak, M., (2014), The Role of Local Governments in the Development of City Logistics, Procedia - Social and Behavioral Sciences, 125, pp. 373–385. Yang, R.J., (2014), An investigation of stakeholder analysis in urban development projects: Empirical or rationalistic perspectives. Int. J. Proj. Manag. 32, 838–849.


- 159 -

Zunder, T.H., Aditjandra, P.T., Schoemaker, J.T., Vaghi, C., Laparidou, K., Österle, I., others, (2014), Engaging city stakeholders to achieve efficient and environmentally friendly urban freight movements, in: Transport Research Arena (TRA) 5th Conference: Transport Solutions from Research to Deployment.

Conclusions

- 160 -

Conclusions Battery electric vehicle: a solution for sustainable city logistics? The transport sector is facing important challenges in European cities. It contributes to climate change through greenhouse gas emissions, it affects urban welfare through noise and exhaust gas emissions, and it impacts energy security given the substantial fossil fuel consumption of the sector. These negative impacts are reinforced by congestion, which is common in cities areas. Hence, solutions need to be developed in order to drive urban transport systems towards a more sustainable equilibrium. In that context, city logistics appears to be a key sector to improve. It is recognised to be responsible of an important share of the total transport emissions and forecasts indicate that this impact will increase in the future. To keep cities attractive, local authorities are therefore devoting greater attention to this segment of the transport sector. Additionally, the European Commission established in its white paper on transport the ambitious objective of achieving CO2-free logistics in major urban areas by 2030. As a result, an increasing number of research projects combining academia, companies and the public sectors are developing solutions to improve the sustainability of city logistics. Among the various proposed solutions, battery electric vehicles (BEVs) are often considered promising for the last mile. Their environmental performance can contribute to a reduction in CO2 emissions, exhaust gas emissions and traffic noise. Additionally, BEVs reduce the transport sector’s dependence on fossil fuels. However, there are barriers to BEV adoption that limit the diffusion of that technology in the automotive sector. Nevertheless, captive fleets are regarded as potential early adopters. The objective of the thesis was therefore to explore the feasibility of introducing BEVs in urban supply chains. To achieve that objective, the following three research questions were investigated:   

RQ1: What is the potential of battery electric vehicles in city logistics? RQ2: What are the factors that can stimulate a shift from diesel to battery electric vehicles? RQ3: What is the level of support from stakeholders for the electrification of city logistics?

To address these general research questions, the thesis investigated the different aspects of BEVs through six chapters, divided into three parts. The findings of these chapters are used to answer

Conclusions

- 161 -

the research questions in the next section of this conclusion. Then, the contributions of these findings regarding the state of the art are highlighted. Finally, the limitations of the thesis are stressed and the resulting opportunities for future research are identified.

Findings RQ1: What is the potential of battery electric vehicles in city logistics? The introduction of the thesis showed that the market penetration of BEVs is highly dependent on the country under investigation. Because Brussels is used as case study throughout the thesis, the context of the research was first introduced in chapter 2. Still, the analyses in the following chapters take care of generalising the results to other urban environments. The Brussels-Capital Region is presented by merging the different figures available for the logistics sector. The results are in line with observations in other European cities. In 2012, vans and trucks accounted for 14% of traffic on average, with a concentration between 6:00 and 8:00 am, which further increases congestion during morning peak hours. However, the most important negative impact of city logistics is on emissions. Logistics accounts for one fourth of CO2 emissions and approximately one third of local pollutants generated by the transport sector in the city. Furthermore, forecasts indicate that city logistics will have a larger impact in the future: volumes to be transported are growing with the urban population, distances should increase because of the delocalisation of logistics platforms from the city, and LCVs (Light Commercial Vehicles, also more commonly called “vans”) are increasingly used because of a fragmentation of volumes. Given the expected growth of city logistics, the Brussels-Capital Region has developed a freight plan to alleviate these negative effects on the urban environment. That plan considers BEVs part of the solution. The negative impacts of city logistics are indeed largely explained by the dominant share of trucks and vans using diesel technology. Electrifying city logistics could therefore contribute to reducing emissions of the transport sector in Brussels. As a result, the regional authorities provide a maximum subsidy of €5,000 to companies purchasing a freight electric vehicle. However, that subsidy is found in chapter 3 to stimulate insufficiently the purchase of BEVs by freight operators. Further support is required to overcome the barriers to the adoption of these vehicles. That was concluded from a survey conducted among freight transport operators located in the Brussels-Capital Region. Their attitudes regarding the advantages and disadvantages of BEVs were first assessed in that third chapter. The results reveal that the operational constraints of BEVs are the first concerns of transporters, before their financial constraints. The limited range and limited public charging infrastructure are assessed to be the most important disadvantages. High purchase costs are found to be the third most critical disadvantage. This ranking can be confirmed through an assessment of the advantages of BEVs: the ability to charge at the depot is considered the most important advantage, while low operational cost is ranked second. These results therefore indicate that operational constraints of BEVs should first be compatible with the logistics environment before being profitable. The environmental performance of BEVs was also assessed: it is the third most critical advantage, after charging at the depot and low operating costs.

Conclusions

- 162 -

Still, positive attitudes regarding BEVs can have sometimes a more limited influence on the final choice decision. A choice-based conjoint (CBC) experiment was therefore included in chapter 3 in addition to questions eliciting respondents’ attitudes. This methodology explores the trade-offs that the respondents make in a choice situation to estimate the preference structure of the population. The CBC revealed that the attribute contributing the most to the adoption of BEVs is their low operating costs. However, that advantage does not compensate for the negative impact of high purchase costs on the choice behaviour. Nevertheless, the most important utility gap between a diesel and a BEV is produced by the range attribute. The results of the CBC confirm the findings from respondents’ attitudes: operational constraints represent the most important challenge, and costs are second. However, the results of the CBC differ from the importance expressed by the respondents’ attitudes regarding the environmental performance of BEVs. That attribute was indeed found in the CBC to have a much more limited impact on the choice behaviour. The attitude–action gap identified by Mairesse et al. (2012) among private households was therefore stressed in chapter 3 to be also at stake among freight operators. Hence, the potential of BEVs is found to be limited in city logistics because their benefits are insufficiently valued by transporters to overcome their barriers. However, the positive attitude of freight operators regarding BEVs can still support the electrification of city logistics. It should facilitate the introduction of policies that stimulate the adoption of BEVs. Indeed, 85% of freight operators were found to agree with the idea that the authorities should support the introduction of BEVs in city logistics. Among the different type of policies, they are particularly supportive of the use of financial incentives such as subsidies for the purchase of electric freight vehicles, exemptions for BEVs from an urban toll at the entrance of the city or fiscal deductions for BEVs. However, the most supported measure is an exemption for BEVs from a kilometre tax. Perhaps the intense debate surrounding such a measure in the Brussels-Capital Region influenced that result. Nevertheless, the results of the CBC reveal indeed that increasing the operating costs of conventional vehicles would particularly influence the choice of transporters towards BEVs. A combination of the current €5,000 subsidy with an exemption for BEVs from a kilometre tax of €0.15/km would effectively influence the average choice behaviour towards BEVs rather than diesel vehicles in the small van category. Financial incentives can therefore compensate for the operational and financial constraints of BEVs. Those conclusions are however valid only if fast charging is available at the depot. Developing charging infrastructure therefore appears to be another important measure to undertake. Installing fast chargers was considered the second most important measure that the authorities should implement, according to transporters. The support of freight transporters is however more mixed regarding measures that regulate vehicle access. In particular, measures that restrict access of some urban areas to polluting vehicles are only little supported by freight carriers. That result might reflect the willingness of transport operators to keep the city accessible for their diesel vehicles. In that category of measures, most promising policy support is therefore rather in granting more access to BEVs through extended time windows or access to bus lanes. An authorization to deliver at night for BEVs is in that context particularly interesting. Indeed, night distribution is another solution considered in the freight plan of the Brussels-Capital Region. Because noise of deliveries is an important constraint for the development of night deliveries, BEVs can represent a technological

Conclusions

- 163 -

solution given their more silent motor (Cavar et al., 2011). Such an extension of time windows for BEVs could then reduce the pressure of freight vehicles on congestion which is especially important in the morning peak hours as identified in chapter 2. RQ2: What are the factors that can stimulate a shift from diesel to battery electric vehicles? Chapter 3 showed that the potential of BEVs in city logistics is mostly limited by their financial and operational constraints. Given their importance, chapter 4 and chapter 5 investigated these barriers in greater depth to identify the factors that can reduce these barriers and stimulate the electrification of city logistics. Chapter 4 was focused on the financial constraints of BEVs. Costs were identified in chapter 3 to be both an advantage and a disadvantage for BEVs compared to conventional vehicles. On the one hand, BEVs benefit from low operating costs; on the other hand, they entail substantial purchase costs. Given the differences in cost structure compared to conventional vehicles, a comprehensive analysis is required to assess the competitiveness of BEVs. Hence, a total cost of ownership (TCO) model was developed to explore the financial constraints of BEVs in the LCV segment. The results indicated that BEVs were more competitive than petrol vehicles thanks to the subsidy of maximum €5,000 provided by the Brussels-Capital Region. However, they remain more costly than diesel vehicles. That result explains the dominant share of diesel vehicles in city logistics and the current limited potential of BEVs in Brussels identified in the first research question. However, these observations depend on the assumptions that were defined based on the average characteristics of the Brussels-Capital Region. To generalise the results, a sensitivity analysis was performed in chapter 4 to relax these assumptions. That analysis was also able to identify how factors influencing the TCO should change to make the BEVs that were analysed more competitive than diesel vehicles. A first category of factors is related to their usage. Light commercial vehicles drive on average almost 15,000 kilometres per year in Belgium, but if they reach a minimum distance of 18,000 kilometres per year, BEVs begin to become competitive with their diesel counterparts. After that threshold, the low operational costs of BEVs offset their high purchase costs. The number of years of ownership should also be optimised by operating a BEV until its battery reaches its end life in transport applications. Ideally, a BEV should be sold before the battery has to be replaced (which depends on the battery technology and its usage), if the owner is not planning to use its BEV for another battery lifetime. A second category of factors is related to market dynamics. In that context, energy prices represent a critical factor that support or entail the TCO of BEVs. The sensitivity analysis revealed that a minimum increase of 20% in diesel prices from July 2015 (€1.28/l) would be necessary for the first BEVs to become competitive with diesel vans. Conversely, an increase in electricity prices might increase the competitive gap between electric and diesel vehicles. The costs of BEVs should therefore be minimised by charging the vehicle at the best rates of the day. Another important market change can come from the development of batteries. Because batteries are responsible for an important share of the TCO of BEVs, their development has the potential to change the competitive position of BEVs relative to diesel vehicles. At present, no residual values on batteries are considered. However, at the end of their transport application, 80% of the battery’s initial energy capacity remains, which leaves the potential for second-hand

Conclusions

- 164 -

applications. The sensitivity analysis estimated that the residual value should be able to recover a minimum of 25% of the initial battery price for electric vans to start becoming competitive with their diesel counterparts. Battery prices are also expected to decline in the near future. That development should impact the TCO of BEVs to the same extent: a reduction in battery prices by a minimum of 25% is required for the first BEVs to become competitive with their diesel counterparts. A combination of these various effects might therefore change the competitive position of BEVs in the near future. Manufacturers have a particularly important role in encouraging that change, in particular in valuing the second-hand applications of batteries. The final category of factors that can influence the competitiveness of BEVs relates to governmental support. Financial incentives were found to be the type of measures that are regarded by transporters as the most important to implement. The sensitivity analysis of chapter 4 indicated that subsidies should be increased to €10,000 in order for BEVs to yield a similar TCO to that of diesel vans. The fiscal deductibility of BEVs is an alternative in this regard. The fiscal system applied in Belgium on passenger cars for BEVs could be extended to city logistics. The deductibility rate of 120% supports however the adoption of a limited number of BEVs. A deductibility rate of 150% is instead recommended to support all the BEVs that were analysed in the segment of light commercial vehicles. An urban toll of between €1 and €2.5 per day or an urban kilometre tax of between €0.025 and €0.04 per kilometre on conventional vans were also found to support the competitiveness of BEVs with respect to their diesel counterparts. The competitive position of BEVs was however considered in chapter 4 only from a financial perspective. Indeed, the results of chapter 3 indicated that other aspects influence the choice of a vehicle such as the vehicle’s capacity, the limited range and the charging time of BEVs. Hence, the financial constraints of BEVs were investigated together with their operational constraints in chapter 5. To integrate these two aspects, a fleet size and mix vehicle routing problem with time windows was developed for battery electric vehicles. That optimisation problem searches for the optimal set of vehicles required to deliver a set of shops at a minimal cost. For a BEV to be selected in the optimal fleet, the distance of the route should be sufficiently large to reach the competitive thresholds with diesel vehicles identified in chapter 4 (lower distance bound). The distance of the last mile should however be sufficiently low to respect the limited range (upper distance bound). The profitable applications of BEVs are therefore limited by these distances bounds. Given these limitations, BEVs were observed in chapter 5 to contribute to a cost reduction of the last mile mostly when they are used as a complement to a diesel fleet. They were also observed to be operated on the routes that have a distance greater than 80 km per day and require small vans. Through a sensitivity analysis, those routes were mostly observed in distribution schemes with fragmented freight flows. Fragmented flows are identified in environments where time windows are narrow, congestion is important and delivery density is low. They are also identified in distribution involving small deliveries. The gap between the lower distance bound (financial constraint) and the upper distance bound (operational constraint), where BEVs are economically feasible, could however become more flexible. By allowing trade-offs between battery price and limited range, transporters could select the optimal battery capacity based on the specific needs of their logistics activity. To increase the

Conclusions

- 165 -

number of profitable applications of BEVs, manufacturers should therefore offer various options of battery capacities to transporters. The gap between the lower distance bound and the upper distance bound could also be widened through financial incentives. The financial constraint would then be relaxed in favour of BEVs. However, financial incentives were found in chapter 5 to have a limited impact on the total costs of the last mile. Granting specific advantages to BEVs such as time window extensions or the use of bus lanes is considered to be the most powerful policies to reduce the last mile costs of electrified fleets. Indeed, time windows and congestion were assessed to have a critical impact on the last mile costs. Relaxing these constraints specifically for BEVs was estimated to reduce last mile costs by approximately 20%. They allow operators to consolidate more freight flows and thereby decrease the number of vehicles in the fleet and the number of drivers in the staff. Such policies were however estimated in chapter 3 to be less frequently considered by transporters as a measure stimulating the adoption of BEVs. Transporters might therefore underestimate the potential of such measures. RQ3: What is the level of support from stakeholders for the electrification of city logistics? The previous research questions focused on the challenges of introducing BEVs into the fleets of transport operators. However, city logistics is a system that also involves stakeholders that are impacted by transport operations. Previous projects in city logistics have shown that the support of these stakeholders is sometimes critical for driving change in the sector. Hence, that last research question is also important for generating insights into the feasibility of electrifying city logistics. The impact of introducing BEVs was first assessed in chapter 6. Because chapter 2 showed that implementing urban consolidation centres (UCCs) was another important measure considered by the Brussels-Capital Region to improve the sustainability of city logistics, the contribution of BEVs was assessed in the context of a UCC. Five scenarios introducing UCCs in Brussels were developed, including two pairs of scenarios differing only by the type of fleet that the UCC operates for the last mile in order to measure the net contribution of BEVs. The first pair compared a fleet of conventional trucks with a fleet of battery electric trucks operating the last mile of the UCCs. The context is defined by two UCCs delivering the city, one located in the north and the other one in the south. Night distribution is also permitted, but only towards the UCCs. Compared to the business as usual, the result of the scenario with conventional trucks reveals a reduction of -2.9% of external costs related to global and local emissions generated by logistics in Brussels. When replacing the fleet of the UCC in that scenario by electric trucks, external costs are further reduced to -3.7%. Hence, BEVs contribute to 22% of these reductions, the rest being due to the consolidation of the flows. The impact of the scenarios on the noise levels of city logistics was also evaluated. Because night distribution is allowed towards the UCC in that scenario, external costs related to noise of city logistics increase by 5.6% relative to the business as usual. But the introduction of electric trucks in the last-mile limits that increase to 4.7%. The electrification of the fleet of the UCCs thus contributes to a 1% reduction in the external costs related to the noise of city logistics in Brussels. That effect is limited given the dense traffic in the city during the day.

Conclusions

- 166 -

The second pair of scenarios compared a fleet of conventional vans with a fleet of battery electric vans operating in UCCs in Brussels. In these two scenarios, the city is served by four UCCs located in the north, south, east and west. Night distribution is still allowed towards the UCCs. Additionally, a road-pricing scheme is introduced on every freight vehicle except on BEVs. Because of the better geographical coverage and the road-pricing scheme, the UCCs capture more freight flows in these scenarios, which result in more important impacts on city logistics. UCCs operating battery electric vans were found to reduce the external costs related to the global and local emissions of city logistics by -8.1%. This reduction would be limited to -5.8% if conventional vans were used. BEVs therefore contribute to 28% of the reduction in these external costs, the rest being due to the consolidation of the flows. BEVs reduce also the external costs related to noise by 2% compared to the scenario where conventional vans would be used in the fleet of UCCs. Again, the noise reduction from BEVs is less valorised because that benefit is diluted in the dense traffic during the day. The evaluation of chapter 6 shows therefore that the electrification of city logistics reduces especially the negative impact of freight vehicles on the emissions and to a lower extent their negative impact on the noise that they generate in the city. However, it does not show the degree of support that the stakeholders of city logistics have for these scenarios. Their support for an electrification of city logistics was therefore explored in chapter 7 through a multi-actor multicriteria analysis. Five main stakeholder groups were identified: receivers, shippers, logistics service providers, citizens and authorities. These stakeholders were invited to a workshop in which the different scenarios were presented. To evaluate their support for these scenarios, they were asked to report their objectives that would be impacted by these scenarios as well as the importance that they assign to these objectives. The results of the multi-actor multi-criteria analysis revealed that citizens and authorities represent the stakeholder groups with the highest support for the introduction of BEVs in city logistics. They both assign a high importance to noise and transport emissions. However, an electrification of city logistics contributes further to the objectives of authorities as stated in their strategic plans for urban development. As a result, authorities represent the stakeholder group that support the most the introduction of BEVs in city logistics. That result supports the idea that authorities should lead the transition towards electrified city logistics. Receivers, shippers and logistics service providers do not exhibit however a significant preference for the technology used in the last mile. Although they consider transport emissions in evaluating the scenarios, they devote little attention to this issue. As a result, they do not show an important support for the electrification of city logistics. However, the analysis has shown also that they are not opposed to an electrification of city logistics. Their support, even if limited, remains positive. That result is stressed given the previous solutions that have been developed in city logistics. Their feasibility was often limited by the conflicting views of stakeholders: some solutions were indeed improving the situations of one stakeholder group while it was affecting the objectives of another stakeholder group. Hence, this last chapter shows that the electrification of city logistics is feasible because it is supported by the different stakeholders of city logistics. The main challenge is therefore to convince freight transport operators to adopt BEVs in their fleets.

Conclusions

- 167 -

Contributions and implications of findings The findings collected across the different chapters of this PhD contributed to the research in city logistics in several ways. They extended the state of the art. But some chapters also nuanced some previous findings reported in the literature. Chapter 2 has contributed to a diagnostic of the logistics sector in Brussels. It consolidated the different data and documents addressing freight transport in the Brussels-Capital Region in order to give a global picture of the sector. That diagnostic was important for introducing the thesis and for developing solutions adapted to the city in the next chapters. But the contribution of chapter 2 was especially important within the Brussels-Capital Region. The publication of this chapter improved the awareness about the logistics sector and its negative impacts in Brussels (Strale et al., 2015). In chapter 3, the thesis has investigated the vehicle choice behaviour of freight transport operators. That chapter represents an original contribution in understanding the limited adoption of BEVs by freight transport operators. It is indeed the first application of a conjoint analysis among freight transport operators in city logistics. The identification of an action-gap confirmed the relevance of assessing the choice behaviour and not only the attitudes of freight transport operators. Hence, the conjoint analysis mostly contributed to the state of the art by estimating the trade-offs that freight transport operators make between the different attributes of a vehicle. Based on that survey, the chapter could also contribute to the possible policies that can stimulate a shift from conventional to battery electric vehicles. The findings of chapter 3 could also nuance what can be found in the literature. Limited range is sometimes presented as a less important barrier in city logistics because vehicles are assigned to regular, short delivery routes. However, companies in city logistics generally operate small fleets, a context in which vehicles should be able to achieve a wider set of missions and consequently be more flexible. BEVs would then be less adapted for such fleets because of their additional constraints. The strong negative influence that the CBC estimated for the limited range of BEVs might therefore be explained by those small operators who represent the majority of the sector. Nevertheless, the CBC revealed a greater heterogeneity in the preferences for low ranges, which indicates that some transporters are less affected by the operational constraints of BEVs. These findings support the hypothesis that potential adopters can be found in city logistics, especially among transporters operating large fleets. Another common idea in the literature presents long charging times and the lack of charging infrastructure as less problematic in city logistics because vehicles can charge at the depot. The results of the CBC, however, revealed that charging at the depot without fast chargers is the worst attribute that a BEV can have. That result was explained by the large share of transport companies that do not park their vehicles in a depot but on streets, where charging stations are scarce. That result also explains why the lack of charging infrastructure was considered the most critical disadvantage of BEVs, even before limited range. It also explains that charging at the depot was assessed to be the most important advantage that a BEV can have. Charging at the depot can indeed solve three important barriers. First, it reduces the problem of limited charging infrastructure. Second, it can extend the range of BEVs by charging the battery each time the vehicle returns to the depot. Third, long charging times become less problematic because

Conclusions

- 168 -

charging can occur during loading and unloading operations at the depot. Early adopters might therefore also be found among transporters that own a depot. Given that the main barriers of battery electric vehicles were identified in chapter 3 to be their operational and financial constraints, the two following chapters contributed to the state of the art by investigating further the conditions where these barriers can be solved or at least reduced. The chapter 4 mainly contributed by extending the TCO analysis to the light commercial vehicle segment. Previous TCO analyses were indeed limited to the segment of heavy goods vehicles although city logistics is dominated by light commercial vehicles as shown in chapter 2. Besides assessing the current competitive position of BEVs, that chapter contributed also to the assessment of the costs factors that influence their TCO compared to their diesel and petrol versions. The sensitivity analysis could identify the different competitive thresholds where battery electric vehicles become the most competitive alternative. The economic constraints investigated in chapter 4 were integrated in the chapter 5 in a vehicle routing problem to consider also the impact of the operational constraints of BEVs on the costs of the last mile. The main contribution of that chapter was the development of a fleet size and mixed vehicle routing problem with time windows (FSMVRTPW) adapted to battery electric vehicles. Because distance is not the only factor influencing the range of BEVs, the model extended previous vehicle routing problems by including an energy consumption model for BEVs that considers the influence of speed, temperature and weight on range limitations. This energy consumption model was then included in the algorithm that was developed to solve the vehicle routing problem. Based on that algorithm, the chapter 5 contributed to the field by identifying the conditions when battery electric vehicles can reduce the costs of the last mile. Among the different findings summarised in previous section, BEVs were mostly observed to be profitable in freight flows that are fragmented and in routes that have a moderate distance. These findings are particularly interesting for BEVs given the current trends of city logistics. Indeed, it was found in chapter 2 that freight flows are increasingly fragmented and that logistics platforms are being delocalised to the periphery of cities. BEVs can therefore be considered an a technology which suits particularly well the current trends of city logistics. In particular, the rise of e-commerce can be identified as an opportunity for the development of BEVs in city logistics. A growing number of profitable applications of BEVs should therefore be expected. The two last chapters contributed to the evaluation of the net impact of an electrification of city logistics. The introduction stressed indeed an important research gap there since the impacts of electric vehicles are often evaluated in city logistics together with a new organisation of deliveries. Based on scenarios developed in chapter 6, the net contribution of battery electric vehicles could be analysed on criteria of sustainability such as transport costs, profitability of the operator, emissions and noise. The results showed that the benefit of an electrification of the fleet was especially in the reduction of the environmental impact of city logistics. BEVs were estimated to be responsible for approximately 25% of the reduced external costs related to local and global emissions of the UCC scheme. That contribution seems however more limited than what the evaluation of Chronopost indicated. In the context of a UCC also, BEVs were reported to be responsible for 2/3 of CO2 emissions reductions while the benefits of consolidation were limited to 1/3 of such reductions (C-LIEGE, 2012; SUGAR, 2011). The difference might be explained by the type of BEVs used by Chronopost. They used electric quadricycles which have a limited

Conclusions

- 169 -

payload. As noted in the introduction, these small vehicles limit the potential for the consolidation of freight flows. Hence, the evaluation of this project underestimated the benefits of the UCC and overestimated therefore the contribution of BEVs to the environmental performance of the scheme. Given the potential of consolidation in improving the sustainability of logistics, the thesis argues that freight flows should be first optimised and then electrified. And not the inverse. The most environmentally friendly vehicle is indeed the one that can be Avoided, as noted by Macharis et al. (2014) in the 4As of sustainable city distribution. BEVs are interesting to reduce the impacts of freight flows that cannot be avoided. Finally, the chapter 7 contributed to the field by estimating the support of the different stakeholders in city logistics for an electrification of the sector through a multi-actor multi-criteria analysis. The results have highlighted the important interest of authorities. But the results showed also the interesting support of citizens. Their positive attitude shows that they might also have a role in stimulating the electrification of city logistics. In addition to encouraging the authorities to move in that direction through their votes, they can also stimulate the introduction of BEVs through their purchasing behaviour. The rise of e-commerce in that context could also provide a greater power to the consumers in how their deliveries are operated. Leaving them the option to have their products delivered with a BEV could be a powerful strategy to stimulate the development of transporters providing sustainable services. Because green deliveries would be valued, the higher costs of an all-electric fleet could be covered by a higher price for the service. This behaviour can also be expected in the B2B as mentioned by Taefi et al. (2013). Some companies are indeed committed to a reduction of their environmental footprint within their programmes of corporate social responsibility. Deliveries with BEVs could contribute in that context to their objectives. These examples highlight therefore the potential of BEVs in segments where a higher transport price can be charged because their benefits are more valorised.

Limitations and future research The thesis has brought several new contributions to the city logistics domain. But the thesis must also acknowledge the limitations of the analysis given the constraints of the field. One problem that all researchers encounter in city logistics is the lack of data. Considerable efforts were made to overcome that problem. Available data sources were identified and used to their full potential, as in chapter 2. However, most of the efforts had to be directed on collecting new data from the sector: in chapter 3, transport operators were surveyed; in chapter 4, the different manufacturers were contacted to identify the various costs components of their vehicles; in chapter 5, the data from an operator was used as a case study; in chapter 6, a collaboration with the European project LAMILO allowed the estimation of freight flows in Brussels; and in chapter 7, a workshop was organised with stakeholders to collect the importance they attribute to their objectives. Thanks to these new data collection efforts achieved within the interdisciplinary MOBI research group, the thesis was able to draw the conclusions described above. The precisions of the conclusions are however limited by the amount of data that could be collected. That constraint might be the most important in chapter 3, where the choice preferences of transporters were estimated for BEVs. The sample size limited the conclusions to the average choice behaviour of transporters. However, there is a variety of segments in city logistics. Browne et al. (2010) showed that the patterns of freight flows can differ considerably according the type of supply chain involved. Because heterogeneity was also observed in the

Conclusions

- 170 -

preferences of transporters, the conclusions suggested several characteristics of transporters that could reveal potential adopters. However, these hypotheses could not be verified based on the choice-based conjoint analysis. The number of choice tasks that was possible to collect was insufficient to conduct a segment analysis on the sample. Future research should therefore investigate the potential of BEVs in some specific segments of city logistics. In addition to identifying segments of early adopters, future research should also bring better tools that can assess the economic feasibility of electrifying the fleet of potential adopters based on their logistics requirements. The thesis started to develop that approach with the fleet size and mix vehicle routing problem with time windows. That model faces however some limitations that should be addressed by future research. In particular, the algorithm should be improved to compute more optimal and faster solutions. More data should also be collected to improve the energy consumption model of BEVs, especially for heavy goods vehicles and for more extreme temperatures. Such a tool could then provide multiple recommendations for the fleet management of transporters at a strategic, tactical and operational level as indicated by Pelletier et al. (2014). One of the important findings stressed that transporters would be highly reliant on fast charging. But other types of charging methods might also facilitate the use of BEVs in city logistics. Future research should therefore investigate further the impact of the different charging methods on the operations of transporters and on their choice behaviour. Their impacts on the battery lifetime should also be further investigated given the important influence of the battery replacements on the TCO of BEVs. That research area might be particularly interesting in city logistics given the diversity of batteries that would be used. They should indeed be adapted to different vehicle sizes and to different logistics environments. The economics of the charging infrastructure deserve also a greater attention. That aspect was indeed not considered in the thesis while it can be considerable, especially in the case of fast charging. An interesting extension of the thesis could also investigate the potential for retrofitting. Converting formerly conventional vehicles into BEVs might be an efficient solution to reduce the high purchase costs barrier and reduce the environmental impact of old freight vehicles. When observing the different experiences of using BEVs in city logistics, converted vehicles are indeed a frequently used option (E-Mobility NSR, 2013). But perhaps the most important research area regarding costs reduction of BEVs is related to their batteries. Research & Development has definitively an important role in improving the energy density of the batteries and reducing their costs. The interdisciplinary MOBI research group is in that field an important contributor to the state of the art thanks to its battery innovation centre. But further research should also address management issues of batteries. The potential second hand application of batteries is a first research area. Capturing their residual value was indeed identified in chapter 4 to strongly improve the competitiveness of BEVs. Further research should also address the potential of right sizing batteries for transporters. Reducing battery capacities could indeed reduce battery costs and therefore improve the total cost of ownership of BEVs. Conversely, increasing battery capacities can convince additional operators that require higher ranges. In that respect, further research on electrification of city

Conclusions

- 171 -

logistics should also investigate the potential of plug-in hybrid electric vehicles and range extended electric vehicles. Finally, one last future research avenue can be identified by taking a bit of distance. The findings developed in this thesis could indeed be brought to the macro-level. The question of efficiency should particularly address the policies that were recommended across the different chapters. Future research should therefore investigate the extent to which the proposed policies are correcting the market failures in the transport sector. The costs of these measures might indeed exceed the benefits of the electrification of city logistics. Hence, future research should identify the optimal level of market intervention from local authorities. Such an analysis would be helpful to define the extent to which the authorities should finance the electrification of city logistics.

Conclusions

- 172 -

References Browne, M., Allen, J., Steele, S., Cherrett, T., McLeod, F., 2010. Analysing the results of UK urban freight studies. Procedia - Soc. Behav. Sci. 2, 5956–5966. Cavar, I., Kavran, Z., Jolic, N., 2011. Intelligent transportation system and night delivery schemes for city logistics. Comput. Technol. Appl. 2, 782–787. C-LIEGE, 2012. C-LIEGE: Clean Last Mile Transport and Logistics Management. E-Mobility NSR, 2013. Comparative Analysis of European Examples of Schemes for Freight Electric Vehicles Compilation report. Macharis, C., Melo, S., Woxenius, J., Van Lier, T., 2014. Sustainable Logistics. Emerald, Bingley, UK. Mairesse, O., Macharis, C., Lebeau, K., Turcksin, L., 2012. Understanding the attitude-action gap: Functional integration of environmental aspects in car purchase intentions. Psicologica 33, 547–574. Pelletier, S., Jabali, O., Laporte, G., 2014. Goods Distribution with Electric Vehicles: Review and Research Perspectives. Strale, M., Lebeau, P., Wayens, B., Hubert, M., Macharis, C., 2015. Cahiers de l’Observatoire de la mobilité. Cah. l’observatoire la mobilité la Région Bruxelles-Capitale 4, 111. SUGAR, 2011. Sustainable Urban Goods Logistics Achieved by Regional and Local Policies, City. Bologna, Italy. Taefi, T.T., Fink, A., Kreutzfeldt, J., Held, T., 2013. On the profitability of electric vehicles in urban freight transport, in: Proceedings of the European Operations Management Association. Dublin, pp. 1–11.

Annexes

- 173 -

Annexes Annex 1: Publications ARTICLES IN SCIENTIFIC JOURNALS WITH AN INTERNATIONAL REFEREE SYSTEM JANJEVIC, M., LEBEAU, P., NDIAYE, A.B., MACHARIS, C., VAN MIERLO, J. and NSAMZINSHUTI, A. (2016). “Strategic Scenarios for Sustainable Urban Distribution in the Brussels-Capital Region Using Urban Consolidation Centres”. Transport Research Procedia 12, 598-612. LEBEAU, P., MACHARIS, C., VAN MIERLO, J. and LEBEAU, K,. 2015. “Electrifying light commercial vehicles for city logistics? A total cost of ownership analysis”. The European Journal of Transport and Infrastructure Research 15 (4), 551-569. LEBEAU, P., DE CAUWER, C., VAN MIERLO, J., MACHARIS, C., VERBEKE, W. and T. COOSEMANS, 2015. “Conventional, Hybrid, or Electric Vehicles: Which Technology for an Urban Distribution Centre?”. The Scientific World Journal, Vol. 2015, Article ID 302867, 11 pages, DOI:10.1155/2015/302867. LEBEAU, K., VAN MIERLO, J., LEBEAU, P., MAIRESSE, O. and C. MACHARIS, 2013. “Consumer attitudes towards battery electric vehicles: a large scale survey”. International Journal of Electric and Hybrid Vehicles, Vol. 5(1), pp. 28-41 (0.18: SCI 2013). LEBEAU, K., VAN MIERLO, J., LEBEAU, P., MAIRESSE, O. and C. MACHARIS, 2012. “The market potential for plug-in hybrid and battery electric vehicles in Flanders: a choice-based conjoint analysis”. Transportation Research Part D, Vol. 17, Issue 8, pp. 592-597, DOI: 10.1016/j.trd.2012.07.004 (1.291: JCR 2012).

ARTICLES IN SCIENTIFIC JOURNALS WITH A NATIONAL REFEREE SYSTEM LEBEAU, P. and C. MACHARIS, 2014. “Freight transport in Brussels and its impact on road traffic” (in NL, FR en EN). Brussels Studies, Number 80, October 2014.

Annexes

- 174 -

ARTICLES/CONTRIBUTIONS IN SCIENTIFIC MONOGRAPHS WITH AN INTERNATIONAL REFEREE SYSTEM MOMMENS, K., LEBEAU, P. and C. MACHARIS, 2014. “A modal shift of palletized fast moving consumer goods to the inland waterways: a viable solution for the Brussels-Capital Region?”. Urban Transport XX, C.A. Brebbia (ed.), WIT Press, pp. 359-371.

ARTICLES/CONTRIBUTIONS IN NATIONAL REFEREE SYSTEM

SCIENTIFIC

MONOGRAPHS

WITH

A

LEBEAU, P. and C. MACHARIS, 2014. “City Distribution in Brussels”. In: Macharis, C., Dobruszkes, F. and M. Hubert (Eds.), Mobilité et logistique à Bruxelles. Urban Notebooks/Stadsschriften/Cahiers urbains. Brussels Hoofdstedelijk Gewest (Brussel Mobiliteit en InnovIris), VUBPRESS, Brussel, pp. 163-189. LEBEAU, P., VERLINDE, S. and C. MACHARIS, 2013. “Urban Freight Transport: description and future evolutions”. In: Macharis, C. and J. Van Mierlo (Eds.), Sustainable Mobility and Logistics, VUBPress, Brussels, pp. 203-239.

PAPERS AT INTERNATIONAL PUBLISHED IN PROCEEDINGS

CONGRESSES/SYMPOSIA

INTEGRALLY

LEBEAU, P., MACHARIS, C., VAN MIERLO, J. and M. JANJEVIC, 2015. “Implementing an urban consolidation centre: involving stakeholders in a bottom-up approach”. URban freight and BEhavior change conference, October 1-2, 2015, Rome. MACHARIS, C., KIN, B. and P. LEBEAU, 2015. “Multi Actor Multi Criteria Analysis as a tool to involve stakeholders in the city distribution context”. URban freight and BEhavior change conference, October 1-2, 2015, Rome. LEBEAU, P., MACHARIS, C. and J. VAN MIERLO, 2015. “The choice of battery electric vehicles for urban logistics: a conjoint based choice analysis”. EVS28 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium, 3-6/05/2015, KINTEX, South-Korea. LEBEAU, P., VAN MIERLO, J., MACHARIS, C. and W. VERBEKE, 2014. “Developing a fleet size and mix vehicle routing problem with time windows for electric vehicles”. European Electric Vehicle Congress (EEVC), 3-5/12/2014, Brussels, Belgium. MACHARIS, C., LEBEAU, P., VAN MIERLO, J. and K. LEBEAU, 2013. “Electric versus conventional vehicles for logistics: A total cost of ownership”. EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium, 17-20/11/2013, Barcelona, Spain. LEBEAU, P., VAN MIERLO, J., MACHARIS, C. and G. MAES, 2013. “Implementing electric vehicles in urban distribution: a discrete event simulation”. EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium, 17-20/11/2013, Barcelona, Spain.

Annexes

- 175 -

LEBEAU, P., VAN MIERLO, J., MACHARIS, C. and K. LEBEAU, 2013. “The electric vehicle as a viable solution for urban freight transport? A total cost of ownership analysis”. Selected Proceedings of 13th World Conference on Transport Research, 15-18/07/2013, Rio de Janeiro, ISBN: 978-85-285-0232-9. LEBEAU, P., MACHARIS, C. and J. VAN MIERLO., 2013. “Electric vehicles for logistics: a total cost of ownership analysis”. In: Hesse, M. et al. (eds.) Proceedings of the BIVEC-GIBET Transport Research Days 2013, 30-31/05/2013, Walferdange, Luxemburg-City (Luxemburg), pp. 307-318. LEBEAU, P., VAN MIERLO, J., MACHARIS, C. and K. LEBEAU, 2012. “Commercial electric vehicles in logistics networks: a profitable niche market?”. European Electric Vehicle Congress (EEVC), 20-22/11/2012, Brussels, Belgium. LEBEAU, K., VAN MIERLO, J., LEBEAU, P., MAIRESSE, O. and C. MACHARIS. 2012. “A choice-based conjoint analysis on the market potential of PHEVs and BEVs in Flanders”. EVS26: The 26th World Battery, Hybrid and Fuel Cell Electric Vehicle Symposium & Exhibition, 69/05/2012, Los Angeles, California, United States of America. LEBEAU, K., VAN MIERLO, J., LEBEAU, P., MAIRESSE, O. and C. MACHARIS, 2011. “The market potential for (plug-in) electric vehicles in Flanders: a choice-based conjoint analysis”. European Electric Vehicle Congress (EEVC), 26-28/10/2011, Brussels, Belgium.

COMMUNICATIONS AT INTERNATIONAL CONGRESSES/SYMPOSIA NOT PUBLISHED OR ONLY AVAILABLE AS AN ABSTRACT LEBEAU, P., VERLINDE, S., MACHARIS, C. and J. VAN MIERLO, 2015. “How Authorities can support Urban Consolidation Centres? A Review of the Best Practices”. City logistics and freight transport workshop (meeting of NECTAR Cluster 3), 16-17/04/2015, Algarves, Portugal. LEBEAU, P., MACHARIS, C., VAN MIERLO, J. and K. LEBEAU, 2015. "Electric vehicles in city logistics: A classification of the most popular applications". NECTAR 2013 International conference, Sao Miguel Island, Azores, 16-18/06/2013.

Annexes

- 176 -

Annex 2: City logistics projects involving battery electric vehicles BE BE

Brussels Brussels

Start up year 2011 2013

Courier

GE

Hamburg, Berlin

2010

Courier Courier Courier

GE GE GE

2011 2013 2009

DHL

Courier

GE

UPS

Courier

GE

Abby Couriers Correos UPS TNT express Loomis Danmark A/S KLS Grafisk Hus STEF-TFE Deret Gefco Meyer&Meyer Hermes Versand Service

Courier Courrier Courrier Courrier Distribution Distribution Distribution Distribution Distribution Distribution

UK ES NL UK DK DK FR FR FR GE

Stuttgart Karlsruhe Hamburg Berlin, Bonn, Dusseldorf, Hamburg, Rosenheim Hamburg, Frankfurt, Dusseldorf, Cologne, Munich, Nurnberg London Various Amsterdam Scotland Copenhagen, Aarhus Copenhaguen Lyon Various cities Paris Berlin

Distribution

GE

Various

1995

Allkopi

Distribution

NO

TGM GreenWay Nappy ever after SpeedyHire Brewers Melrose & Morgan To Door

Distribution Distribution Distribution Distribution Distribution Distribution Home deliveries

SE SK UK UK UK UK DK

Erosiki

Home deliveries

ES

Seur Joey's pizza delivery Peter Appel Sainsbury's Tesco Borgerstein Institute Hospital Aviapartner SEAS-NVE Tre-For A/S Effenberger Bakery Manus

Home deliveries Home deliveries Home deliveries Home deliveries Home deliveries

ES GE NL UK UK

Oslo, Bergen, Stavanger, Kristiansand Gothenburg Bratislava London London London London Copenhagen Bilbao, San Sebastian, Vitoria/Spain Valencia, Madrid Hamburg Amsterdam London London

Internal transport

BE

Antwerp

2011

Internal transport Internal transport Internal transport Internal transport Maintenance

BE DK DK GE BE

Zaventem Haslev and Svinninge Kolding Hamburg Brussels, Antwerp

2012 2010 2010 2012 2008

Operator Ecopostale TNT Hermes Logistik Gruppe (HLG) DPD UPS City Express

Logistics framework Courier Courier

Country City

2010 2010 2009 2008 2012 2011 2011 2009 2003 2011

2009 2013 2013 2005 2007 2008 2010 2011 2011 2012 2012 2005 2007

Annexes Municipalities (41) in Wallonia The municipality of Frederiksberg The municipality of Copenhagen Technische Unie Enterprise Mouchel Goupil Antverpia City of Lommel Post Danmark Finland Post Corporation La poste Norway Post Posten Norge CityDepot CUDE Chronopost (chronocity) Geodis (Distripolis) Transport Genty (ELCIDIS) La petite Reine The green link DPD Eco-logis Life CEMD M.E.R.Ci. (Mobilità Ecologica Risorsa per la Città) City logistics company Interporto (cityporto) Merci in Centro Peeters Vervoercentrale L.A.J. Duncker BV (020Stadsdistributie.nl) Cornelissen (cityshopper) Delta stadsdistributie Hoek Transport (cargohopper) Van de Bogerd Stadsleveransen Gnewt Cargo Office depot UPS DHL Newcastle University

- 177 -

Maintenance

BE

Various Walloon cities

2011

Maintenance

DK

Frederiksberg

2004

Maintenance

DK

Copenhagen

2009

Maintenance Maintenance Maintenance Maintenance Postal services

NL UK BE BE DK

Amsterdam London Antwerp Lommel Bornholm

2011 2007 2007 2011 2011

Postal services

FI

Turku, Kajaani, Rovaniemi

2002

Postal services Postal services Postal services UCC UCC

FR NO NO BE ES

Various Bergen, Oslo Various Hasselt Malaga

2011 2011 2004

UCC

FR

Paris

2004

UCC

FR

Paris

2011

UCC

FR

La Rochelle

2001

UCC UCC UCC UCC UCC

FR FR GE IT IT

Paris Paris Nurnberg Brescia Lucca

2001 2009 2000 2012 2005

UCC

IT

Genoa

2003

UCC UCC UCC

IT IT IT

Siena Padova Como

1999 2013 2014

UCC

NL

Amsterdam

2009

UCC

NL

Amsterdam

2009

UCC

NL

Nijmegen

2010

UCC

NL

Zutphen

2009

UCC

NL

Utrecht

2010

UCC UCC UCC UCC UCC UCC UCC

NL SE UK UK UK UK UK

Leiden Gothenburg London London London Bristol Newcastle

1997 2012 2009 2009 2009 2004 -

Annexes VanGansewinkel The municipality of Rodovre VanGansewinkel TIL Citycargo Aad de Wit Verhuizingen Go-Ahead Seymour Green

- 178 Waste

BE

Mechelen

2011

Waste

DK

Rodovre

2009

Waste Others Others

NL IT NL

Rotterdam Reggio Emilia Amsterdam

2009 2003 2007

Others

NL

Amsterdam

2011

Others Others

UK UK

London London

2008 -