Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town)
Network constraints for competing renewable power plants in a resource rich area Jarrad Wright PrEng MScEng MIEEE MSAIEE
Senior Consultant Energy Exemplar (Africa) Johannesburg, South Africa
[email protected]
Abstract—Significant growth in installed capacity of solar PV and wind generation has taken place globally with the largest contributors being the USA, Europe (mostly Germany) and China. Projections for solar PV and wind generation capacity seem to indicate that this growth will continue into the future. The lack of electrical networks in resource rich areas to integrate these technologies is well known. Considering this, a model R to assess long run tariff has been developed in PLEXOS competition between solar PV and wind power projects in a particularly resource rich area with varying levels of network constraints. The model includes functionality for individual technology specific developers (solar Photovoltaic (PV) or wind only) as well developers who own a mixed portfolio of projects (solar PV and wind). Network details to simulate maximum capacity exports from the various projects are also modelled. Hourly solar irradiation and wind speeds along with associated power curves define the expected power output from the projects on an hourly basis. The projects compete via a tariff based on technical and financial parameters defined for each with a cost recovery algorithm utilised. The effects on individual project and portfolio tariffs for a range of network constrained scenarios are presented. Applications could include inter alia network expansion approaches (by network operators and project developers) as well as tariff bidding strategies by developers in renewable energy auctions like the REIPPPP in South Africa .
Figure 1.
Worldwide installed wind capacity
plants in an area with considerable available solar and wind resources. The focus is on solar PV and wind It is structured with this introductory section outlining wind and solar PV worldwide as well as in Africa followed by a brief overview of network constraints as they relate to renewables. This is followed by a brief outline of the model developed and mathematical formulation that is used to analyse the effect that network constraints have on the required tariff for a renewables plants (including those in a portfolio). Results and analysis are then presented with a concluding section thereafter.
I. N OMENCLATURE PV Photovoltaic IEA International Energy Agency REIPPPP Renewable Energy Independent Power Producer Procurement Programme SAURAN Southern African Universities Radiometric Network WASA Wind Atlas of South Africa FOM Fixed Operations and Maintenance II. I NTRODUCTION Significant growth in installed capacity of solar PV and wind generation has taken place globally with the largest contributors being the USA, Europe (mostly Germany) and China. Projections for solar PV and wind seem to indicate that this growth will continue. The lack of networks in resource rich areas to connect these technologies is well known. This paper attempts to assess the effect of network constraints on the behaviour of competing renewable power
A. Wind and Solar PV worldwide The installed capacity of wind power worldwide has grown significantly since 2000 as shown in Figure 1 [1]. This growth has been near exponential and resulted in 318.5 GW of installed wind capacity worldwide by the end of 2013 with dominant regions including North America, Europe and Asia. This means that 23 GW of wind capacity has been installed every year on average since 2000 (an average growth of 25% per year). Based on their high renewables scenario, the International Energy Agency (IEA) predicts installed wind capacity of up to 1 600 GW by 2030 (>4x 2013 levels) and 2 700 GW by 2050 (>7x 2013 levels) [2]. Significant growth is expected from Asia, Europe and USA. Worldwide installed capacity of solar PV is shown in Figure 2 [3]. This growth is even more dramatic when compared to the growth trends in installed wind capacity worldwide. Average annual growth of 148% since 2000 has resulted in
Page 1 of 7
Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town)
Figure 2.
Worldwide installed solar PV capacity
138.8 GW of installed capacity worldwide in 2013 with an average of 10.5 GW installed annually since 2000 growing installed capacity by 108x since 2000. Solar PV installed capacity is dominated by Europe (mostly Germany), Asia Pacific (including China) and the Americas. The IEA predicts solar PV capacity to grow even faster than wind capacity into the future with predictions of up to 1 721 GW by 2030 (>12x 2013 levels) and 4 670 GW by 2050 (>34x 2013 levels). Most of this growth is expected to come from China (and other Asian countries), USA, India and the Middle East with Africa growing from a small base in 2013. This growth in installed capacity means growth rates of 200% per year will be required along with the installation of 93 GW annually on average until 2030. Up to 2050, 225% growth per year will be required along with the installation of 123 GW annually on average until 2050. B. Wind and Solar PV in Africa Although solar PV and wind generation capacity have grown significantly globally, Africa has not yet featured in this growth. Africa had only 1.2 GW of installed wind capacity by 2013 [4] including South Africa, Morocco, Ethiopia, Egypt, Cape Verde and Tunisia with negligible solar PV capacity by 2013. In recent years, wind and solar PV deployments have taken place in South Africa under the Renewable Energy Independent Power Producer Procurement Programme (REIPPPP) [5], in Kenya under the Lake Turkana Wind Power Project [6] as well as in Rwanda with the development of the Agahozo Shalom Youth Village solar PV project [7] to name a few. The expected solar PV and wind installed capacity in South Africa as a result of the REIPPPP is summarised in Table I. These are geographically shown in Figure 3 [8]. Based on information from [8], 1049 MW of solar PV and 706 MW of wind capacity were operational by the second quarter of 2015. This is a significant increase in installed capacity for Africa (albeit form a very low base). There are significant solar and wind resources in Africa [9– 11] and as a result many solar PV and wind projects are being planned for development. Prospective renewable resources in Africa cover all regions as shown stylistically in Figure 4 [9] and includes solar and wind. Although African solar PV and wind capacity is quite small at the moment, the latest developments in South Africa as well as expected developments in other African nations based on the available solar and wind resources should bode well
Figure 3. South African REIPPPP preferred bidder locations by technology
to grow the solar PV and wind power generation industry in Africa. Table I E XPECTED SOLAR PV AND WIND IN S OUTH A FRICA (REIPPPP)
Round Round Round Round Total
1 2 3 4
Solar PV (MW)
Wind (MW)
710 349 435 415
641 559 787 676
1 909
2 663
C. Network constraints with renewables The issue of network constraints as a result of the deployment of renewables (predominantly solar PV and wind) is well known around the world in Asia [12], United Kingdom [13], USA [14], Europe [15] and South America [16]. The location of major load centers and conventional generation fuel sources have determined the planning of networks in the past. Due to the deployment of non-conventional generation sources like solar PV and wind there is currently and will be a need to reassess the required network infrastructure required to get the required power to major load centers. In particular, areas where significant solar and wind resources are present will require localised network investment and thus possible constraints if a number of developers wish to develop solar PV and/or wind power plants in those locations. The network costs included in integrating plants could range from shallow network costs (only costs to integrate from the plant to the nearest point in the network) to deep network costs (additional costs imposed including expansion of networks in other areas not directly related to the plant) [17]. In South Africa in particular, following the integration of Round 1 and Round 2 REIPPPP projects, network constraints appeared at a transmission level as well as in localised areas where
Page 2 of 7
Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town)
Figure 4.
Stylised geographical view of renewables potential in Africa [9]
prospective renewables plants are being developed [18, 19]. There has been a requirement to reprioritise projects and adjust timing of network expansions to fit within the budget available at the national transmission provider[18]. This paper will focus on localised network constraints (not transmission level backbone expansion) as they relate to competing solar PV and wind power plants in a resource rich area. III. T HE M ODEL
AND
F ORMULATION
The model is built using PLEXOS R Integrated Energy R Model (developed by Energy Exemplar) [20]. PLEXOS is a best in class software tool used around the world for energy market analysis and optimisation.
the prospective projects need to compete only on tariff in a fixed quantity single round auction. The hourly profiles of solar irradiation and wind speeds are based on data from Wind Atlas of South Africa (WASA) [21] and Southern African Universities Radiometric Network (SAURAN) [22]. In addition to adding a degree of variability to the expected solar irradiation and wind speed (20% hourly standard deviation), the data is converted via look-up tables to actual power outputs per technology. Solar irradiation and wind speed variables for each project have an assumed correlation of 50% with each other. Example power curves for solar PV and wind are shown in Figure 6 and Figure 5 respectively where data points for one year are plotted.
B. Network A. Generation The model considers the behaviour of competing wind and solar PV plants with technical and financial properties as defined in Table II. These parameters can be adjusted by the user if desired and simulations re-run to assess the effect on tariffs required. Two individual solar PV projects (solarPV1 and SolarPV2 ) along with two individual wind projects (Wind1 and Wind2 ) are considered along with a solar PV and wind project owned by one developer (SolarPVH and WindH ). This is to simulate the possibility of portfolio optimisation and economies of scale available to a developer with both technologies in their prospective portfolio. It is assumed that
The network is modelled in a number of scenarios as shown in Table III. The scenarios are chosen to simulate the effect of no network constraints, all projects having network constraints, self-interest network expansion by each project developer and collaborative network expansion. When the network is constrained, a representative value of half of the project capacities is chosen i.e. 50 MW. The costs for network expansion are assumed to be as shown in Table IV. Similar to the technical and financial parameters defined for the solar PV and wind projects, these parameters can be adjusted by the user if desired, new scenarios defined and simulations re-run to assess the effect on tariffs.
Page 3 of 7
Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town) Table II P LANT PROPERTIES
Max Capacity Fixed O&M (USD/kW/yr) Build Cost (USD/kW) Technical/Economic Life (yrs) Debt:Equity Cost of Debt (%) Return on equity (%)
SolarPV1
Solar PV2
Solar PVH a
Wind1
Wind2
Winda H
100 50 2 730 25 0.70 7.50 25.0
100 50 2 600 25 0.70 7.50 25.0
100 50 2 600 25 0.70 6.75 22.5
100 60 2 100 25 0.70 7.50 25.0
100 60 2 000 25 0.70 7.50 25.0
100 60 2 000 25 0.70 6.75 22.5
a Developed by one developer
Table III N ETWORK SCENARIOS SolarPV2 (MW)
SolarPVH (MW)
Wind1 (MW)
Wind2 (MW)
WindH (MW)
100 50 50 100 50 100
100 50 50 100 50 100
100 50 100 50 50 100
100 50 50 50 100 100
100 50 50 50 100 100
100 50 100 50 50 100
100
100
90
90
80
80
70
70
Plant output (MW)
Plant output (MW)
No constraints Network constrained Network expansion (portfolio only) Network expansion (solar only) Network expansion (wind only) Network expansion (collaborative)
SolarPV1 (MW)
60
50
40
60
50
40
30
30
20
20
10
10
0
0 0
2
4
6
8
10
12
14
16
18
0
20
Figure 5.
Figure 6.
Wind project power output versus wind speed
C. Formulation In order to assess the effect on the overall tariff of the prospective projects, a cost recovery algorithm is utilised. In this case, the long term capacity expansion module is utilised with a large load connecting the projects via network elements (as previously described) to force the building of the projects and supply their power to the load across the modelled network. Annualised build costs are automatically calculated (as required to pay debt and return to shareholders). In addition, annual Fixed Operations and Maintenance (FOM) costs previously defined are added to annualised build costs which then defines the required cost to be recovered annually. The tariff required will then be these costs divided by the generation in the year.
200
400
600
800
1000
1200
1400
1600
Solar irradiation (W/m 2 )
Wind speed (m/s)
Solar project power output versus solar irradiation
Of course, when there are network constraints, the required tariff will be higher (as a result of undispatched generation capacity from the project). It is also important to note that the cost recovery is performed across portfolios of projects (and not necessarily individual projects). This has been simulated via assigning SolarPVH and WindH to one developer while all other solar PV and wind projects are assumed to be developed by individual developers. IV. R ESULTS AND A NALYSIS For the scenario with no network constraints, the power generation from all of the projects (once operational) for a representative one week period, are shown in Figure 7. The power generation for the developer with the portfolio is shown
Page 4 of 7
Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town) Table IV N ETWORK EXPANSION COSTS ( TO DEVELOPER )
100 90
Unitised cost (USD/kW)
11.26 10.63 10.63 35.82
56.30 106.30 106.30 11.94
No constraints Network constrained Network expansion (portfolio only) Network expansion (solar only) Network expansion (wind only) Network expansion (collaborative)
80 70
Power (MW)
Cost USD-mln
60 50 40 30 20 10
200
0 0
20
40
60
80
Solar (gen.) 1
160 140
Plant output (MW)
100
120
140
160
Hour
180
Solar (avail. gen.) 1
Wind (gen.) 1
Wind (avail. gen.) 1
Figure 8. Generation from projects for a week (network expansion for wind only)
120 100 80
a 41% decrease in solarPVH tariff and 29% increase in WindH tariff when compared to individual solar and wind projects. This makes the SolarPVH project extremely well placed to compete with other solar PV projects but places the WindH project in a difficult position as it would have a higher tariff compared to other competing wind projects in the area.
60 40 20 0 0
20
40
60
80
100
120
140
160
Hour Solar 1
Figure 7.
Solar 2
Wind 1
Wind 2
Portfolio (Solar H +Wind H )
Generation from projects for a week (no network constraints)
along with the individual solar PV and wind projects. The hourly randomisation imposed on the solar irradiation and wind speeds are clearly seen where power generation differs across solar and wind projects. To demonstrate the effect of network constraints, the power generation and available power generation for the solarPV1 and Wind1 projects for a representative one week period are shown in Figure 8 (for the self-interest wind network expansion scenario). As can be seen, the solarPV1 project is curtailed as a result of insufficient network capacity while the Wind1 project can export all available power generation. For all of the network scenarios, the tariff required for each of the projects is shown in Figure 9. Of interest is that the change in build cost between solar PV and wind projects is offset by the change in cost of capital (cost of debt and return on equity) to result in a very similar required tariff (see Table II). A. No constraints If one were to consider the scenario where the network company were to pay for and build the required network (unconstrained scenario), the tariff required is the lowest possible tariff i.e. 253 USD/MWh for solar PV projects and 100 USD/MWh for wind projects. The portfolio projects owned by one developer recover their tariff across the portfolio. This allows for the developer to offer a lower tariff for their solarPVH plant (149 USD/MWh) but means an increased required tariff for their WindH plant (129 USD/MWh). This is
B. Network constrained With the constrained network scenario, the solar PV projects’ required tariff increased by 20.6% (from 253 USD/MWh to 305 USD/MWh) and wind projects by 58% (from 100 USD/MWh to 158 USD/MWh). This is as a result of the constrained network (with only 50 MW being available compared to the 100 MW maximum capacity of the projects). The developer portfolio requires an increase in tariff of 35% (from 149 USD/MWh to 202 USD/MWh) and 6% (from 129 USD/MWh to 138 USD/MWh) for the SolarPVH and WindH projects respectively. It is interesting to note that even with a constrained network, the SolarH project is still cheaper than the other competing solar PV projects (when they have no network constraints). The WindH project becomes slightly less competitive but not substantially. C. Network expansion (self-interest) The decrease in tariff as a result of a developer investing in network expansion that is in their self-interest (compared to a constrained network) is: i Solar: 42 USD/MWh ii Wind: 51 USD/MWh iii Portfolio: SolarPVH : 32 USD/MWh WindH : 7 USD/MWh The benefit of network expansion is clear with a decrease in tariff of 14% for solar PV projects and 32% for wind projects while portfolio projects decrease by 16% and 5% for solar and wind projects respectively.
Page 5 of 7
Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town) 350
Tariff required (USD/MWh)
300
250
200
150
100
50
0 NoConstr Solar
Figure 9.
Constr 1
Solar
Portfolio
Solar
Solar
2
H
Wind
Collaborative 1
Wind
2
Wind Wind
H
Tariff required for projects under various network scenarios
It should be noted though that there would be an increase in tariff required for the projects as a result of the network expansion (compared to when the network company performs and pays for the network expansion). These will be: i Solar: 10 USD/MWh ii Wind: 8 USD/MWh iii Portfolio: SolarPVH : WindH :
21 USD/MWh 2 USD/MWh
As can be seen, self-interest network expansion does of course add a small amount to the tariff when compared to the case where the network operator performs the network expansion. Increases of 4% for solar PV, 7% for wind projects are noted while for the portfolio projects increases of 12% for solar and 2% for wind are noted.
D. Network expansion (collaborative) Continuing from the previous section, the decrease in tariff as a result of investment in network expansion collaboratively compared to network expansion in their self-interest is: i Solar: 8 USD/MWh ii Wind: 4 USD/MWh iii Portfolio: SolarPVH : WindH :
V. C ONCLUSIONS Considering the significant growth in installed capacity of solar PV and wind capacity around the world as well as growth projections thereof, there will be a need for expanded networks in resource rich areas to connect these technologies to load centers. The model developed allowed for projects to compete on tariffs, based on technical and financial parameters defined. A cost recovery algorithm was run to find the tariffs required by the projects (individual and portfolio bidding has been included). When there were no network constraints, the lowest tariff per project was possible (as expected). A developer owning a portfolio of projects was able to bid lower on a particular technology project as they could recover costs through their other prospective project. Once a constrained network was assumed, tariffs increased but it was interesting to note that a project within a portfolio still had a lower tariff than an equivalent technology project (even if they were unconstrained by their network). Network expansion in the self-interest of a particular developer is never better than collaborative network development but only a marginal decrease in required tariff is noted. Applications could include inter alia network expansion approaches (by network operators and project developers) as well as tariff bidding strategies by developers in renewable energy auctions like the REIPPPP in South Africa.
3 USD/MWh 2 USD/MWh
R EFERENCES
The benefits of investing in network expansion collaboratively is marginal (1-4% decrease compared to self-interest network expansion). However, if the possibility of network expansion is available in a collaborative manner it will be beneficial (to all projects).
[1] The Wind Power, “World and continents capacity,” 2015. [Online]. Available: http://www.thewindpower.net/ statistics world en.php [2] International Energy Agency (IEA), “Technology Roadmap: Wind energy,” France, Tech. Rep., 2013.
Page 6 of 7
Power-GEN/DistribuTECH Africa, 15-17 July 2015 (Cape Town) [3] European Photovolatic Industry Association (EPIA), “Global Market Outlook: For Photovoltaics 2014-2018,” Tech. Rep., 2014. [Online]. Available: http://www.epia.org/index.php?eID=tx nawsecuredl&u=0&file=/uploads/tx epiapublications/ 44 epia gmo report ver 17 mr.pdf&t=1425826730& hash=ddb55baff6b2f9c4dcac72643f9094731a86e757 [4] Global Wind Energy Council (GWEC), “Global Wind Report 2013,” 2013. [Online]. Available: http://www.gwec.net/publications/ global-wind-report-2/global-wind-report-2013/ [5] Department of Energy, “Renewable Energy Independent Power Producer Procurement Programme,” 2015. [Online]. Available: http://www.ipprenewables.co.za/ [6] Lake Turkana Wind Power, “Lake Turkana Wind Power Project: Overview,” 2012. [Online]. Available: http:// www.ltwp.co.ke/the-project/overview [7] Gigawatt Global, “Scatec Solar and Norfund begin construction on East Africas First UtilityScale Solar Plant,” 2015. [Online]. Available: http://energiyaglobal. com/east-africas-first-utility-scale-solar-field-launched/ [8] The Energy Blog, “Renewable Energy Project Sites South Africa,” 2015. [Online]. Available: http://www. energy.org.za/knowledge-tools/map-of-sites [9] International Renewable Energy Agency (IRENA), “Africa’s Renewable Future: The Path to Sustainable Growth,” Tech. Rep., 2013. [Online]. Available: http://www.irena.org/menu/index.aspx?mnu=Subcat& PriMenuID=36&CatID=141&SubcatID=276 [10] ——, “West African Power Pool: Planning and Prospects for Renewable Energy,” Tech. Rep., 2013. [Online]. Available: http://www.irena.org/DocumentDownloads/ Publications/GA ECOWAS Solar Web.pdf [11] ——, “Southern African Power Pool: Planning and Prospects for Renewable Energy,” Tech. Rep., 2013. [Online]. Available: http://www.irena.org/ DocumentDownloads/Publications/SAPP.pdf [12] M. Ni and Z. Yang, “By Leaps and Bounds: Lessons Learned from Renewable Energy Growth in China,” IEE Power and Energy Magazine, 2012. [13] K. Neuhoff, J. Cust, and K. Keats, “Implications of intermittency and transmission constraints for renewables deployment,” 2007. [14] T. Gentile, D. Elizondo, and S. Ray, “Strategic Midwest Area Renewable Transmission (SMARTransmission) Study,” Quanta Technology, Tech. Rep., 2010. [15] CIGRE WG C6.08, “Grid Integration of Wind Generation,” CIGRE, Tech. Rep., 2011. [16] H. Rudnick, L. A. Barroso, D. Llarens, D. Watts, and R. Ferreira, “Flexible Connections: Solutions and Challenges for the Integration of Renewables in South America,” IEEE Power and Energy Magazine, 2012. [17] D. Jakus, J. Krstulovic, and J. Vasilj, “Algorithm for optimal wind power plant capacity allocation in areas with limited transmission capacity,” International transaction on Electrical Energy Systems, vol. 24, pp. 1505–1520, 2014. [18] Eskom Transmission Group, “Transmission development
[19]
[20] [21] [22]
Page 7 of 7
plan 2015-2024,” Eskom Holdings SOC Limited, Tech. Rep., 2014. Antoinette Slabbert, “Unconnected renewable energy costs Eskom R2m a month,” 2014. [Online]. Available: http://www.moneyweb.co.za/uncategorized/ unconnected-renewable-energy-costs-eskom-r2m-a-mon/ Energy Exemplar, “PLEXOS Integrated Energy Model,” 2015. [Online]. Available: http://energyexemplar.com/ CSIR, “Wind Atlas of South Africa (WASA),” 2015. [Online]. Available: http://www.wasa.csir.co.za/ Stellenbosch University and and University of KwaZuluNatal, “Southern African Universities Radiometric Network (SAURAN),” 2015. [Online]. Available: http:// www.sauran.net/
