ENERGY EFFICIENCY OF MULTIPORT POWER CONVERTERS USED IN PLUGIN/V2G FUEL CELL VEHICLES
Nicu Bizon University of Pitesti, Romania E-mail:
[email protected],
[email protected] Tel +40 348 453 201, +40 722-895624, Fax +40 348 453 200 Address: 110040 Pitesti, 1 Targu din Vale, Arges, Romania
Abstract. In this paper is presented an analysis of energy efficiency for the Multiport Power Converters (MPCs) used in Plug-in Fuel Cell Vehicles (PFCVs). A generic MPC architecture for PFCVs is proposed, which is analyzed for different operating modes of MPC in relation with PFCV operating regimes and the plug-in feature. The basic MPC architecture is described in relation with the PFCV operating regimes. Two MPC architectures are derived from the basic MPC architecture: (1) the MPC1 architecture, which is the MPC architecture without reverse power flow during regenerative braking process, and (2) the MPC2 architecture - MPC architecture without charging mode of Energy Storage System (ESS) from the FC system. Taking in account the imposed window for the ESS state-of-charge, the MPC can be connected to Plug-in Charging Stations (PCS) to exchange power with the Electric Power (EP) system, which will include renewable Distributed Generation (DG) systems. The Energy Management Unit (EMU) of MPC can communicate with the EP system to determine the moments that match the energy demand of plug-in vehicle with the supply availability of the EP system, stabilizing the EP system. The MPC features regarding its energy efficiency were shown by analytical computing performed
1
and appropriate simulations presented in relation with the ESS that can be charged (discharged) from (to) the home/DG/EP system. Keywords: multiport power interface, energy efficiency, fuel cell hybrid power source, fuel cell vehicle, plug-in feature, state-of-charge window.
1. Introduction Electric Vehicle (EV) is a technology that promises to drastically reduce emissions associated with road transport. In the last decade the technology has been supported by different manufacturers and specialists as a key element in reducing CO2 emissions (as well as emissions of pollutants and noise) of cars and light commercial vehicles. But at the same time, EV technology is still far from being projected as necessary, emphasizing too many uncertainties regarding the issues to be addressed, such as [1]:
The battery technology (energy capacity in relation to vehicle range and road range, fast charging, durability, availability and environmental impacts of used materials).
Well–to-wheel impacts on emissions.
Interaction with the DG system.
Cost of large scale introduction.
On the other hand, battery-powered EVs technology has some advantages over conventional Internal Combustion Engine (ICE) vehicles, such as high-energy efficiency and zero environmental pollution. However, the performance is far less competitive than ICE vehicles, due to the much lower energy density of the batteries than that of gasoline. Consequently, the Hybrid Electric Vehicle (HEV) that uses two power sources has the advantages of both technologies – the
2
ICE and EV technology, and overcome their disadvantages. So, the HEV technology is promoted in meantime by the main companies that design and produce cars. As it is known, HEV combines an ICE with an on-board rechargeable ESS to achieve better fuel economy than a conventional vehicle, without having a road range limitation as an electric vehicle. A number of HEVs are in current production and now are available for purchase, such as the hybrids models from Toyota (Lexus), Honda, Chevrolet, Ford, Mercedes Benz, and so on. Interaction of vehicle with the DG system is a relatively new concept defined by the plug-in features (referring mainly to the integration of an on-board charger or using of an external charger). Plug-in vehicles can be classified into different categories such as EV, HEV, Plug-in Hybrid Electric Vehicles (PHEV), and plug-in fuel cell vehicles (PFCV) [2]. Update information about EVs and a concise classification of EVs are given in [3], as below:
Full Electric Vehicles (FEVs) that have an electric motor and no ICE or Fuel Cell (FC) system.
PHEVs that have both an ICE and an electric motor, and a battery that can be charged from the renewable PCS/DG system.
Electric Vehicles with a Range Extender (EREVs) that have one or more electric motors and an ICE or a FCS that can be used to charge the ESS, and thus extend the vehicle’s road range. The ESS of an EREV can be charged from renewable PCS/DG system, too.
A FCV is an EV with a FC system operating as range extender. Even if the FCVs are totally different from the conventional ICE powered vehicles and ICE-based hybrid drive trains, however the main car’s manufactures already announced their FCVs, with or without plug-in facilities. They have supported ongoing research into the development of FC technology for use in FCVs
3
and other applications. Hydrogen production, storage and distribution are the biggest challenges, so the FCVs still have a long way before entering the market. The purpose of this paper is to demonstrate that there are many design options for MPC architecture of a PFCV. The typical MPC architecture for a PFCV is presented in Figure 1, where FC system usually powers (near to its Maximum Power Point (MPP)) the DC bus via a DC-DC power converter named Power Interface 1 (PI 1). The inverter system is usually of bidirectional type, so the PI 2 must be of bidirectional type, too. The PIs 1&2 could be integrated in one PI of multi-inputs type using different integrated power topologies [4, 5]: bi-buck, bi-boost, or hybrid integrated topologies. During the regenerative braking process, the power flows from load to ESS via the inverter diodes and the PI2 operating in buck mode [5]. Because the bidirectional type is more expensive than unidirectional type, some FCVs architectures use the unidirectional boost converter type for PI 2 [4]. Consequently, for charging the ESS used in the MPC architecture based on unidirectional inverter is necessary to use two additionally PIs, which are power converters of DC-DC and AC-DC unidirectional type (named as PI 3 and PI 4, respectively). If the inverter system (PI 5) and the PI 2 are of bidirectional type, then a series conection of these PIs (working in reverse mode during regenerative braking process) could be modeled by the PI 4, too. Consequently, the architecture shown in Figure 1 is a generic MPC architecture that permits to study the energy efficiency of whole MPC, having as energy sources the FC system and the ESS, and as output(s) the AC electrical machine(s). The concept of MPC (or Multi-port Power Electronic Interface) is commonly adopted to process the renewable power from multiple sources and loads [6, 7, 8], having the following main features: (1) maximum energy harvesting from renewable sources, (2) optimal management of
4
energy from multiple sources, (3) optimal ESS management, and (4) adaptive energy management system for the best performance. The MPC represents a particular case of energy hub concept that is considered as a unit where multiple energy carriers can be converted, conditioned, and stored, representing an interface between different energy sources and loads [9, 10, 11].
Figure 1. Plug-in Fuel Cell Vehicle – generic MPC architecture The MPC creates an interface between loads, renewable sources, and storage elements to efficiently provide and recover power. Consequently, the MPP tracking guarantees optimal energy harvesting from energy sources that have a power characteristic with a maximum at its MPP. The proposed architecture is a flexible MPC topology, generalizing the most used MPC topologies in automotive applications such as the series and parallel MPC topologies, which means the use of an ESS of low voltage (LV) and high voltage (HV) type, respectively [12, 13, 14]. In [12] it is
5
shown that this MPP architecture is more efficient than both series and parallel MPC topologies when the ESS is operated to have final State-Of-Charge (SOC) equal with the initial SOC. Here, is analyzed the case when the final ESS SOC is different from the initial SOC. This case may be of interest in operating the PFCV. Also, for the architectures that are derived from the basic MPC proposed, in this paper will be shown that each could work more efficiently under certain conditions that will be specified by analytical calculation performed. Consequently, this basic MPC architecture helps to boost the energy efficiency performance as well as the flexibility in operating the PFCV to increase the driving range and fuel economy. Also, the mass of MPC could be reduced by using the modularization and integration techniques for MPCs [6, 15], but this is off topic of this paper. Finally,it can be noted that the MPC concept represents both integrated PIs and their appropriate control, and also the fact that they operate efficiently under an energy management strategies implemented in the Energy Management Unit (EMU). A brief comparison of energy management approaches applied for Polymer Electrolyte Membrane (or Proton Exchange Membrane - PEM) FC systems in the literature is made in [16]. The energy management of multiple sources and loads is a complex and difficult task that need to be solved by the specialists using different strategies based on (1) intelligent concepts [17, 18], (2) local and global optimization approaches [19, 20], (3) frequency decoupling techniques [21, 22], (4) linear and nonlinear controllers used in multiple control loops [23] etc. The ECU design takes a more important and detailed place in the literature recently, especially based on fuzzy logic and nonlinear optimization [16, 24]. This design issue is mainly related to ECU features of easy adaptation to more complex MPC architectures, computational efficiency, driving rules integration, capacity to modeling uncertainties, etc [25]. Real-time optimization based on equivalent consumption minimization strategy may be a good choice for online implementation of
6
ECU [20]. Finally, all proposed ECU strategies compare the MPC energy efficiency in term of equivalent hydrogen cost. From the point of view of power flows balance in hybrid MPCs, the energy is balanced by the DC bus voltage regulation, usually using an ESS [23]. This study focuses on this kind of MPC architecture powered by a FC system and equipped with an ESS that usually has two energy storage devices: battery and ultracapacitor (UC). The UCs stack, as a high dynamic and high power density device, operates for supplying power to regulate the DC bus voltage on load transitions. The batteries stack, as a high energy density device, operates for supplying energy to satisfy the power flows balance on DC bus and to keep charged the UC stack, which could be connected to the ESS bus via a bidirectional PI for obtaining high performance in operating the ESS [26]. The FC system, as a slowest dynamic source in PFCV, operates to provide the average load power and to keep charged the ESS stack that also can be charged from the regenerative braking power flow [21, 23]. Therefore, there are three voltage control loops: DC bus voltage regulated by the ESS, ESS voltage regulated by the FC system, and UC voltage regulated by the battery stack (see Figure 1, where the last loop is not shown; a cheap ESS topology of passive parallel type was shown [26]). The FC system combined with an ESS can operate via MPC at best performance for FCVs propulsion and other applications [27, 28]. If the ESS can fulfill the transient power demand fluctuations, the FC system can be downsized to fit the average power demand in the load following strategy, without facing to the load peaks [27]. The dynamic performances of a FCV operating in all above mentioned regimes, with and without the assistance of the ESS, are systematically investigated in [28, 29].
7
The MPC architecture and energy management in such hybrid power sources play a significant role in their design while considering the (1) performance, (2) efficiency, and (3) reliability issues. As it was mentioned above, different MPCs have been investigated by researchers in their effort to design better PFCVs. The trade-off between the complexity of the energy management strategies and the system performance may provide different results for different types of PFCVs that operating in the specified regimes under a given driving cycle. The EMU controls the power flows in different PFCV operating modes such as [13, 29]:
Charge-Sustaining (CS) Mode - operating mode in which the ESS SOC may fluctuate, but on average is maintained at a certain level while driving.
Charge-Depleting (CD) Mode - operating mode in which the average of ESS SOC decreases while driving.
Charge-Increasing (CI) Mode: operating mode in which the average of ESS SOC increases while driving.
The CI mode is a specific operating mode only for EREV (in particular for FCV) in which the ESS is charged (PESS < 0) while driving, so it is different by charging mode when the plug-in vehicle is connected to a PCS. In Figure 1, PESS represents the average power exchanged by the ESS in a load or drive cycle:
If PESS > 0, then this power represents the net power discharged by the ESS, and thus final SOC is lower than initial SOC. In this case the PFCV operates in CD mode and, considering that FC system already operates at full power, the PFCV must be connected to renewable PCS/DG system as soon after the SOC level reaches the low value imposed.
8
If PESS < 0, then this power represents the net power from the FC system and regenerative load that charged the ESS, and thus the final SOC is higher than the initial SOC. In this case the PFCV operates in CI mode and must be connected to the renewable DG system as soon after the SOC level reaches the high value imposed. In this way the plug-in vehicles could operate as grid stabilizers.
The key issue related to grid integration of plug-in vehicles will be matching the power demand with supply availability from the renewable DG system. A connection to the DG system allows more opportunities for all plug-in vehicles. The plug-in vehicles can not only charge their ESS, but also discharge the available ESS energy into the DG grid. In these operating regimes the plugin vehicles can support the DG grid and this concept is named as vehicle-to-grid (V2G). However, plug-in vehicles should not be confused with the vehicle-to-grid (V2G) concept, in which electric drive vehicles can provide value to electric utilities through peak saving and auxiliary services. V2G defines system architecture in which plug-in vehicles, such as EVs, PHEVs and PFCVs, communicate with the DG system to sell demand response services by either delivering electricity into the grid or by controlling their charging rate. Such of operating regimes for PFCV will be considered in energy efficiency analysis performed in this paper. The EMU controls the PIs to operate the ESS in charge, discharge, or sustaining mode in order to increase the energy efficiency of the whole MPC. SOC regulation, ESS life, power dynamic of load, power capacity of FC system near to its MPP, and available or required power to or from grid, respectively, should be the main factors in designing the PFCV EMU. The energy efficiency of the whole MPC shown in Figure 1 will be analyzed in this perspective of ESS operating modes. The ESS SOC must always be in the SOC window imposed for a drive cycle (between two consecutive connections to the PCSs). The analyze performed will show that
9
energy efficiency of whole MPC depends by SOC window width, energy efficiency of each PI, regenerative braking efficiency and duration of this occurrence, kFC fraction of FC power used to charge ESS, and load power profile (that depend by road profile). Anyway, to implement PCS system it is necessary to upgrade the distribution system at level of [30]:
Wide Area Monitoring Systems.
Two way distribution grids.
Two-way high speed communications.
Renewable Energy Sources integration.
Furthermore, plug-in vehicles can also be designed to provide power for stationary power applications such as back up power to a home. Usually, a charger is a bidirectional two-stage topology consisting in a DC-AC inverter and a DC-DC converter that uses a high frequency transformer for isolation and/or a HV conversion ratio. The EMU can be designed to control the range extender system (ICE or FC) to keep the ESS SOC within a specified SOC window. The EMU receives operation commands from the driver (signals from the accelerator pedal and brake pedal) and feedback from the drive train (for example, the speed signal) and other components (Figure 1), and then takes decision to use a proper operation mode and the appropriate levels for power flows to satisfy the power balance. The EMU design is a complex issue and is out of the goal of this paper that presented the EMU strategies only at conceptual level. In first part of this paper (introduction and next two sections) was briefly presented the state of the art for EVs and HEVs, and especially for FCV, with or without the plug-in feature integrated, for an easy understanding of the last sections where it is performing an energy efficiency analysis for
10
the FCV MPC. So, next section shortly presents the common electrical motors and power converters usually used in automotive applications. The features of Advanced Power Interface Modules (APIMs), which are scalable to meet different power requirements, are presented in section 3 in relation with input sources connected to APIM DC-ports and grid support, respectively. In the last sections are presented the analysis of three MPC architectures in term of energy efficiency related to ESS power flow and MPC parameters’ values. Finally, some conclusions are given related to MPC energy efficiency and PFCV capacity to stabilize the grid by charging/discharging the ESS under the EMU control. 2. Electrical motors and APIMs used in current HEVs and EVs The main types of electric traction motors adopted for HEVs include the induction motor (IM), the switched reluctance motor (SRM), the permanent magnet (PM) Synchronous Motor (SM) motor (PMSM), the brushless DC motor (BLDC), and the new motors such as PM brushless DC (PMBLDC) [13]. These motors have many advantages when compared with other motor types, such as: efficiency, cost, size and weight, life span and maintainability. Of course, the current EVs must use an electric motor that can also act as a generator to recharge the ESS when the brakes are applied. During regenerative braking, the motor acts as a generator, providing power back to the ESS while slowing down the vehicle. If the vehicle must be stopped quickly or if the ESS is near to full charge, then friction brakes are used. Usually, both regenerative braking and mechanical braking modes appear into a single foot pedal, in following order: the foot pedal controls the regenerative braking in the first part and then will control the mechanical braking for stop safely and quickly. However, the braking requirements for all cars will have to be enhanced in next years to meet new safety requirements.
11
Compared to IM, the PM motors have higher efficiency due to the elimination of magnetizing current and copper loss in the rotor. Also, high performance torque control is easier to be obtained with PM motors, in particular, with PMBLDC motors. Recent advancements in permanent magnetic materials and motor design have made the PM motor a great candidate for traction drives in EV and HEV applications [31]. Also, the BLDC motor requires a rather complex control to switch from motoring mode to regenerative braking. So, a number of different ways to implement regenerative braking in a BLDC motor were developed. On the other hand, the IMs and SRMs have the highest reliability. According to the comparative study presented in [32], which are based on the following major requirements of HEVs electric propulsion, the IM seems to be the most widely adapted candidate for the electric propulsion of HEVs, but new motors have been designed in recent years for HEV applications. The size of the electric motor for standard HEV applications is expected to vary from about 10 kW to as high as 100 kW, so the inverter system must meet these designing requests. The basic inverter topologies, which are powered from the HV DC bus used in HEV or EREVs architecture or directly from ESS bus in FEV architecture, are shown in Figure 2: The first (a) is a traditional PWM inverter that operates in rectified mode during the regenerative braking. The second (b) is a DC-AC PI (also named inverter) plus a bidirectional DC-DC PI. The third (c) is a Z-source inverter (ZSI), with the unidirectional variant shown in plot (d).
12
Figure 2a. Traditional PWM inverter
Figure 2b. Inverter plus a bidirectional DC-DC PI
Figure 2c. Bidirectional Z-source inverter
Figure 2d. Unidirectional Z-source inverter Figure2. Inverter topologies Usually a bidirectional DC-DC PI is necessary in front of DC-AC PI. This operates in boost mode during motoring regime and will operate in buck mode during the regenerative braking regime. As a result, the bidirectional power converters are desirable. As it was mentioned, the generic architecture shown in Figure 1 can model both regimes in study of FCVs energy efficiency for standard drive cycles. The voltage on ESS bus has a large variation (larger as 50%, depending of
13
the battery type used) [13]. The voltage on DC bus must be stabilized to level required for inverter system used. If the PI 2 will not be used, the inverter system must to be oversized to handle this large ESS voltage range and appropriate value of current at full power load. At full power load, double the rated current will appear at 50% of rated ESS voltage. Consequently, the cost of the inverter system will increase. So, the bidirectional PI 2 is necessary to minimize the stress of the inverter, even if this will increase the whole system cost and control complexity, and thus the system reliability decreases [33]. ZSI topology shown in Figure 2c is a new PI topology that has interesting features such as buck-boost characteristics and single stage conversion, which overcome the above mentioned problems for automotive applications [34]. For example, ZSI topology can produce any desired AC output voltage, even one greater than the input DC voltage, making the ZSI fed adjustable speed drive systems competitive for FC automotive applications [35, 36]. A bidirectional ZSI with nine-switches can replace the conventional HEV inverter with a bidirectional DC-DC PI and two inverters, which are used for power transfer among a battery, an electric motor and an electric generator [37]. The resonant-phase leg inverters could be used instead of hard-switching inverter to increase the energy efficiency, and new control techniques could reduce the ripple on DC bus [38, 39]. Taking in account the above cited references, the energy efficiency for inverter system may be considered in range of 0.8 to 0.95. PIs must be modular and scalable because this will simplify the design, leading to increased use [15]. Consequently, a higher production volume will permit lower PI costs. Also, further improvements in the PI system package, PI standardization and interoperability among PIs and systems are necessary in APIMs. APIMs that are scalable to meet different power requirements, with modular design, lower cost, and improved reliability, will allow designers to use a hubenergy-block approach with software and wiring, defining the specific functions of each identical
14
block. As it was mentioned above, combination of multiple energy sources can be done using a MPC, thus, the following should be provided for the design of APIMs [13]:
Each APIM should have at least one DC-port and one AC-port.
All APIM ports must be bidirectional and should be able to operate in a buck or boost mode.
APIM device’s rating is limited by manufacturer datasheets, so paralleling of APIMs is necessary for high power applications.
The wiring will be based on the load power required.
Filters, transformers, and other external circuit designs will be dependent of the operational power implemented.
These APIMs features will improve the overall cost and durability of automotive power systems and renewable PCS systems. The concept of Integrated Power Electronic Modules is already implemented on the market. This PI combines power electronic devices, dc-link filter capacitors, current and temperature sensors, gate drivers, heat sinks, and optional Digital Signal Processing controllers into a single optimized power module. 3. DC input sources for APIM ports The DC input sources for APIMs shown in Figure 1 are the FC system and ESS. 3.1. Fuel Cell System Compared with the ICE, FC system has the advantages of high energy efficiency and much lower emissions. Because the response of the fuel cell is much slower than electric motor traction response, especially when this power changes abruptly, the use of ESS is obligatory. This is the case of FCV automotive power systems, but also of renewable DG systems based on FC system, especially when the FC system is used for supplying power in the islanded mode to some local
15
loads (V2G or plug-in vehicle, for example). The ESS employs high-power and high-energy density devices, such as UCs and batteries. Hybridization of the FC system with an ESS is an effective technology to overcome the disadvantages of the FC-alone-powered vehicles. The PEMFC stack represents one of the most used solutions as main energy source in automotive applications because of its small size, the ease of construction, a fast start-up and low operating temperature. Unfortunately, its relatively short life cycle is still an impediment to FCV commercialization [13, 28]. As it is known, the inverter current ripple is the main factor responsible for low performance regarding the PEMFC energy efficiency [40, 41] and PEMFC life cycle [42]. Also, it is known that low frequency spectral components of the PEMFC current ripple affect the PEMFC life cycle. Consequently, the controller of PI 2 must to have a control loop for ripple mitigation, for example as in parallel active filtering. Obviously, this control loop will force the PI 2 to supply from ESS the sharp load power profile, too [4]. The PEMFC stack (PFC) and ESS stack (PESS) assure the power flows on DC bus via the PI 1 (Pout1) and PI 2 (Pout2), respectively. The power balance on DC bus is PDC=Pout1+Pout2, where the load power flow, PDC, is the input DC power required by the inverter system. The power flows management is performed by the controller of Controlled Voltage Source (CVS) (having the optional feature of power spreading) and the controller of Controlled Current Source (CCS) (that have included the mitigation feature of current ripple and peaks of inverter power). Consequently, the FC system operates near to MPP and the ESS operates inside of SOC window [27, 43]. The MPP current (IMPP) can be tracked in an adaptive feedback loop by injecting the probing current or by controlling the FC fuel flow using FC current as control variable. 3.2. Energy Storage System
16
The on-board ESS is charged from FC and load by a power fraction, kFCPFC and kLPL, respectively, and provides peak power to the motor controller during vehicle acceleration. Consequently, the ESS is one of the most important components in the PFCVs, so a number of requirements, such as specific energy, specific power, efficiency, maintenance requirement, management, cost, environmental adaptation and friendliness, and safety, must be fulfilled. There is a wide variety of batteries’ technologies from standard flooded and enhanced flooded to next Absorbed Glass Mat (AGM) and lithium-ion polymer batteries, which all are claimed to be “the best choice” for automotive applications [13, 44]. As it is known, most current HEVs utilize NiMH batteries [44], but the most likely alternative battery technologies for use in PHEVs, EREVs and FCVs are Li-ion and ACM batteries that have the advantage of higher energy density and charge-discharge efficiency [13]. The new semi-solid flow batteries described in [45] and nanotechnologies in development will overcome the limitation of energy density, providing a possible 10-fold improvement in this performance over present liquid flow-batteries, and a lowercost manufacturing than conventional lithium-ion batteries [44]. The SOC of the ESS should be kept inversely proportional to the vehicle’s speed [13]:
At low speeds it is probable that vehicle will accelerate in the near time, so the ESS should have a high SOC to power that acceleration.
At high speeds, it is probable that vehicle will be brake soon, so the ESS should have a low SOC to have capacity to accept as much of the regenerative braking energy as possible.
Consequently, the durability and lifetime of the battery under SOC deep cycling are key measures of battery performance. Automotive batteries require a cycle life of at least 5,000 charge –
17
discharge cycles and a calendar lifetime of over ten years [44]. To overcome this situation, the hybridization of ESS by using UCs stack seems to be now the best way to obtain high performance for ESS. Several ESS topologies (passive, semi active and active battery–UC hybrid) have been proposed and tested for using in PFCV [13, 26]. In general, the UC stack voltage rating is lower than ESS voltage, so a bidirectional PI is required to boost this UC voltage to ESS voltage required on UC semi-active ESS topology. An optimal operating window in SOC middle (around of 0.7 SOC) is usually chosen for ESS in order to obtain a high efficiency and flexibility of MPC operation in all regimes, including the safe operation of PFCV in braking regime [13]. Also, in can be noted that UCs stack has a very limited energy change range because of the limited voltage variation in passive ESS topology, which (because is cheap and simple) is by far the most common topology employed in commercial products [26]. Expanding (by using an active or semi-active ESS topology) of the usable SOC window will substantially reduce incremental cost, and thus should be emphasized in PFCV design [16, 46]. For example, a UC voltage window of 50% from its rated voltage will allow about 75% utilization of the overall energy [46]. However, until the PFCV will be plugged to PCS, EMU must decide how best will use both sources of energy (FC and ESS) in a CD, CI or CS operating manner. Specifically, the energy content of the vehicle battery must be sufficient to operate the vehicle through the full distance of the anticipated use-restricted zone or distance to next PCS. The FCV will use the stored electrical energy during initially CD operation. After exhausting part of energy obtained during charging, the ECU will switch to CS operation if the next PCS is far away or to CI mode if next PCS has communicated its energy demand availability to stabilize the EP system. The best EMU energy
18
strategy and the optimum ESS requirements depend by the PFCV architecture used under various load changes [47], as it will be presented below, too. 4. Basic MPC architecture In Figure 1 was proposed a generic architecture for APIMs used in MPC of PFCV. For simplicity, the power converters included in the MPC architecture were named power interfaces (PIs). The whole MPC has as inputs the FC and ESS bus and as outputs the AC electrical machines, depending of number of well-mounted motors. The simplified diagram of MPC for a PFCV using one AC electrical machine is presented in Figure 3 (without plug-in feature shown).
Figure 3. Plug-in Fuel Cell Vehicle – simplified architecture of basic MPC The average output power of the FC system and ESS under a load or drive cycle is PFC and PESS. The PFC is always positive, but the sign of PESS depends by state-of-charge (SOC) regulation performed by EMU to ESS system:
19
PESS > 0 represents the net power discharged by the ESS in a drive cycle.
PESS < 0 represents the net power charged by the ESS in a drive cycle.
PESS = 0 means that final SOC is equal with initial SOC in a drive cycle.
The FC system can operate under EMU control in switching mode or can supply continuously the ESS bus. From FC system via PI3 will result a controlled charging power flow to ESS that is defined by k FC ratio. If the current SOC becames lower than the minimum level accepted for SOC or if a nearest connected PCS grid communicates that is some energy is required to stabilize the grid, then FC system can operate at full power. In the last case, the FCV will operate in CI mode before connection to that PCS in order to increase the current SOC near to maximum level accepted for ESS SOC. Also, excess generation can be used to charge the ESS during periods of low demand. Then, the stored energy can be used to provide energy into grid (via PCSs on route) during periods of high demand in EP system. The ESS is charged from the regenerative power flow, too. The regenerative power flow represents in average a kL ratio from the load power, PL. Called the regenerative braking factor, kL is the percentage of the total braking energy that can be regenerated by the electric motor. The regenerative braking factor depends by design and control of the braking system and is a function of the applied braking strength [13]. In the motor regime of AC electrical machine, the power flows will circulate from FC and ESS system to AC motor via PIs, as are indicated in figure 3. The reverse power flow, kLPL, will circulate from the AC electrical machine (operating as generator during regenerative braking process) to the ESS system via PI 5 and PI 2 connected in series (that is formally represented by PI 4). This representation is more flexible for analysis of energy efficiency performed in next sections of this paper. The
20
overall MPC performance regarding energy efficiency can be potentially optimized by a proper configuration of MPC and applying a specific EMU strategy for FCV on different road conditions. The MPC architecture proposed in Figure 3 contains the following PIs:
PI 1 is a unidirectional DC-DC power converter, having the energy efficiency coefficient
denoted by 1. The FC system powers the DC bus via PI 1.
PI 2 - unidirectional DC-DC power converter, having the energy efficiency coefficient
denoted by 2. The ESS system powers the DC bus via PI2. If PI 2 is of bidirectional type and operates in reverse mode, then energy efficiency coefficient of reverse power flow during regenerative braking process is noted with 2R.
PI 3 - unidirectional DC-DC power converter, having the energy efficiency coefficient
denoted by 3. The FC system powers with kFCPFC the ESS bus via PI 3 by a ratio from the rated output power of the FC system, PFC. The remaining power, (1-kFC)PFC, is delivered to DC bus via PI 1.
PI 5 –unidirectional inverter system (DC-AC power converter), having the energy
efficiency coefficient denoted by 5. The inverter system powers the AC electrical machine with PL from the DC bus via PI 5. If PI 5 is of bidirectional type and operates in reverse mode, then energy efficiency coefficient is notes with 5R and is applied to reverse power flow during regenerative braking process, kLPL. The net power delivery to AC electrical machine is (1-kL)PL.
PI 4 - unidirectional AC-DC power converter, having the energy efficiency coefficient
denoted by 4. Because PI 2 and PI 5 are considered of unidirectional type, PI 4 represents the
21
series connection of PI 2 and PI 5 operating in reverse mode, so 4=2R5R. Also note that plug-in feature of PI 4 is not shown in figure 3. Consequently, the basic architecture shown in Figure 3 permits to study the energy efficiency of whole MPC, having as inputs the FC and ESS bus and as output the AC electrical machine. This systemic analysis will be performed in next sections considering as parameter the kESS ratio defined as kESS = PESS/ PL, which will generalize the results from [12], where kESS=0 is considered. It is noted that energy efficiency coefficient of each PI varies with operating conditions, thus it is defined as average value. 5. Energy efficiency of basic MPC architecture (MPCb) The power flows balance on DC bus is: PDC=Pout1+Pout2
(1)
where the power flow on DC bus, PDC, is DC input power required by the inverter system to power the load. The power balance on ESS bus is: PESS+3kFCPFC+4kLPL=Pout2/2
(2)
where the used notations have been mentioned above. The energy efficiency for basic MPC architecture is given by relation: MPCb=(1– kL)PL/(PFC+PESS)
(3)
By simple mathematical manipulations of the first two relations, relation (3) became: MPCb=(1– kL)A / Bb
(4)
A= 1 + kFC(2 3 – 1),
(5)
where:
22
B=1/5 – 2 4 kL – kESS [2 - 1 – kFC (2 3 – 1)], Bb B k
ESS
k ESSb
=1/5 – 2 4 kL – kESSb [2 – 1 – kFC (2 3 – 1)]
6. Energy efficiency of MPC architecture without regenerative power flow (MPC1) In this case, the energy efficiency for MPC is given by relation (4) considering kL=0: MPC1=A1 / B1
(6)
where: A1 A k Bb1 B k B1 Bb1
ESS
kL 0
k ESS 1
L
0
= 1 + kFC(2 3 – 1),
=1/5 – 2 4 kL – kESS1 [2 – 1 – kFC (2 3 – 1)]
(7)
=1/5– kESS1 [2 – 1 – kFC (2 3 – 1)] = Bb1 + 2 4 kL
If both MPC architectures (MPCb and MPC1) exchange the same ESS power in a drive cycle, which is then available (necessary) to be discharged (charged) to (from) a PCS, then kESSb= kESS1= kESS(b1), and it is important to know which ones operate more efficiently.
MPCb MPC1
(8)
By equivalent mathematical manipulations of the relations (4) and (6), relation (8) is equivalent with relation (9):
k ESS (b1) CFC CL
(9)
where: CFC 2 1 k FC ( 23 1 )
(10)
CL 1 / 5 24 0
Therefore, because always CL>0 (for example see Figure 4a for kFC=0.1), the sign of kESS(b1) depends only by the sign of CFC (for example see Figure 4b for 4=0.95). For a finite value of kESS
23
is necessary to have CFC0 by controlling kFC value. So both MPC architectures could have the same efficiency for kESS given by relation (11): k ESS (b1) CL / C FC
(11)
For example see the CL/CFC surface shown in Figure 4c for kFC=0.1, 3=0.95, 4=0.95, and 5=0.9. Analyzing Figure 4c it can be observed that ESS could operate on a drive cycle in CD mode (with PESS > 0, resulting a final SOC lower than initial SOC) or in CI mode (with PESS < 0, resulting a final SOC higher than initial SOC). Consequently, the PFCV ESS must be charged or discharged from or to the next renewable PCS system, respectively.
Figure 4a. CL(2, 5) for 4=0.95
Figure 4b. CFC(1, 2) for kFC=0.1 and 3=0.95
24
Figure 4c. CL/CFC(1, 2) for kFC=0.1, 3=0.95, 4=0.95, and 5=0.9 If value of kESS parameter is the same for both MPC architectures (but different to kESS(b1) value), then energy efficiencies of considered MPC architectures are different, MPCb and MPC1, respectively (see Figure 5).
Figure 5. MPCb(1, 2) and MPC1(1, 2) for kFC=0.1, kL=0.1, 3=0.95, 4=0.95, and 5=0.9 From relations (8) and (9) it can be concluded that: k ESS CL / CFC MPCb MPC1
(12)
25
If different values of kESS parameter are used in operation of MPCb and MPC1 architectures (noted with kESSb and kESS1, respectively), then MPC architectures could have the same efficiency (see Figure 5): MPCb MPC1 k ESSb k L CL / CFC (1 k L )k ESS1
(13)
The relation (13) for kFC=0.1, kL=0.1, 3=0.95, 4=0.95, 5=0.9 and different kESS1 values (negative and positive) is shown in Figure 6. Analyzing Figure 6 it can be observed that same efficiency could be obtained for MPC1 ESS and MPCb ESS operating in CD mode or in CI mode. High values for kESSb are obtained when CFC=0, but this case will be avoided by control of kFC.
Figure 6. kESSb(1, 2) for kFC=0.1, kL=0.1, 3=0.95, 4=0.95, 5=0.9 and different kESS1 values 7. Energy efficiency of MPC architecture without charging power flow from the FC system (MPC2) In this case, the energy efficiency for MPC is given by relation (4) considering kFC=0: MPC2=(1 – kL)A2 / B2
(14)
where:
26
A2 A k Bb 2 B k B2 Bb 2
k FC 0
ESS
k ESS 2
FC
0
= 1,
= 1/5 – 2 4 kL – kESS2 [2 – 1 – kFC (2 3 – 1)]
(15)
=1/5 – 2 4 kL – kESS2 (2 – 1)= Bb2 – kESS2 kFC (2 3 – 1)
If both MPC architectures (MPCb and MPC2) exchange the same ESS power (positive or negative), then kESSb= kESS2= kESS(b2), and it is also important to know which ones operate more efficiently:
MPCb MPC 2
(16)
Relation (16) is equivalent with relation (17) by equivalent mathematical manipulations of the relations (4) and (14):
CkL sign(2 3 1 ) kESS (b2) 2 sign(2 3 1 )
(17)
CkL 1 / 5 24 k L 0
(18)
where:
Of course, if 2 3= 1, then both MPC architectures considered are equivalent in terms of energy efficiency. This is clearly shown in Figure 3: the kFCPFC power flow reaches on the DC bus via PI2 and PI3 in series and via PI1 for MPCb and MPC2 architecture, respectively. Therefore, if 23 1, then both MPC architectures have the same efficiency for kESS value given by relation: k ESS (b 2) CkL / 2 >0
(19)
Because always CkL>0 (for example see Figure 7 for kL=0.1), one can notice that same efficiency could be obtained if ESS operates in CD mode.
27
Figure 7. CkL(2, 5) for 4=0.95 and kL=0.1 It is obvious that for a given value of common kESS parameter, which is different by kESS(b2) value, the energy efficiencies of considered MPC architectures have different values, MPCb and MPC2, respectively. From relations (17) and (19) it can be concluded that:
CkL / 2 kESS 1 23 0 MPCb MPC2
(20)
If different values for kESS parameter are used in the operation of MPCb and MPC2 architectures (noted with kESSb and kESS2, respectively), then MPC architectures could have the same efficiency: MPCb MPC 2 k ESSb CkL / 1 [CFC / k FC ( 2 1 ) / 1 ] k ESS 2
(21)
The relation (21) for kFC=0.1, kL=0.1, 3=0.95, 4=0.95, 5=0.9 and different kESS2 values (negative and positive) is shown in Figure 8. Analyzing Figure 8 it can be observed that same efficiency could be obtained for MPC2 ESS operating in CD mode or in CI mode and MPCb ESS operating in CI mode. From relations (12) and (20) it can be noticed that MPC1 and MPC2 architectures must be compared in terms of energy efficiency, too. This analysis will be performed in the next section.
28
Figure 8. kESSb(1, 2) for kFC=0.1, kL=0.1, 3=0.95, 4=0.95, 5=0.9 and different kESS2 values 8. Energy efficiency of MPC1 vs. MPC2 architectures Taking in account relations (6) and (14), the considered MPC architectures (MPC1 and MPC2) could have the same efficiency for a common value of ESS power (positive and negative), kESS1= kESS2= kESS(12). This means: MPC1 k
ESS (12 )
MPC 2
k ESS (12 )
(22)
By simple mathematical manipulations of the relation (22) is obtained: kESS(12) 1 24 kL (1 kL )/{(1/5 24 kL 1 2 ) [kFC (23 1 ) 1kL ] 1kFC (1 kL )}
(23)
By simulations (see Figure 9, where kESS12 was limited to range [-20, 20]), one can notice that ESS could operate with PESS > 0 (CD mode) or PESS < 0 (CI mode).
29
Figure 9. kESS12(1, 2) for kL=0.1, 3=0.95, 4=0.95, 5=0.9 and different kFC In some particular cases, this can be put in evidence by analytical calculus, too, as below. If kFC= kL will be considered in relation (23), then this relation will become: 2 kESS (12)a 1 24 (1 kL ) /[23 / 5 1 kL (1 234 ) (1 2 )23 ]
(23a)
If 23=1, then relation (23a) will become simpler: k ESS (12)b 2 4 (1 k L ) / [1 / 5 1 k L (1 2 4 ) (1 2 )]
(23b)
Also, if 2=1 is considered in relation (23a), this will become simpler: k ESS (12) c 2 4 (1 k L ) / [3 / 5 1 k L (1 234 )]
(23c)
In all particular cases it may be noted that sign of kESS(12) can be positive or negative, depending on the MPC parameters’ values. Consequently, the PFCV ESS must be charged or discharged from or to renewable PCS system, respectively. Obviously, the MPC architectures have different efficiencies, MPC1 and MPC2, for values of kESS parameter that are different to kESS(12) value. If different values of kESS parameter are used in operating of MPC1 and MPC2 architectures (noted with kESS1 and kESS2, respectively), then MPC architectures could have the same efficiency:
30
k {[ k ( ) 1 k L ] [1 / 5 2 4 k L 1 2 ] 1 2 } MPC1 MPC 2 ESS1 FC 2 3 1 k ESS 2 [1 k FC (1 k L ) 1 2 ] 1 2 4 k L (1 k L )
(24)
Relation (24) is represented in Figure 10 and 11 for different values of kFC and kL parameters, respectively. The values of energy efficiency coefficients for PIs were chosen to obtain negative and positive sign for expressions (1 - 2 3) and (1 - 2), respectively. The used parameters are: 1=0.85 (respectively 1=0.95), 2=0.9, 3=0.95, 4=0.94, and 5=0.9.
a) 1=0.85
b) 1=0.95 Figure 10. kESS2 vs. kESS1 for different kFC and kL=0.1, 2=0.9, 3=0.95, 4=0.94, 5=0.9
31
a) 1=0.85
b) 1=0.95 Figure 11. kESS2 vs. kESS1 for different kL and kFC=0.2, 2=0.9, 3=0.95, 4=0.94, 5=0.9 Analyzing Figures 10 and 11 it can be observed that same efficiency could be obtained for MPC1 ESS and MPC2 ESS operating in CD mode or in CI mode. This means high flexibility in choosing the EMU strategy to control the MPC of FCV. The energy efficiency of whole MPC depends by the energy efficiency coefficients of all PIs. This is shown in Figure 12, where relation (24) is represented as a (1, 2) surface for two values of kESS1 parameter (- 0.5 and 0.5, respectively). The rest of used parameters are: 3=0.95, 4=0.94, 5=0.9, kFC=0.1, and kL=0.1.
32
Figure 12. kESS2(1, 2) for kFC=0.1, kL=0.1, 3=0.95, 4=0.94, 5=0.9 and different kESS1 values High values for kESS2 are obtained when 2 / 1 k FC (1 k L ) 1 1
(25)
but this case will be avoided by appropriate control of kFC. In Figure 12 the kESS2 was limited to range [-20, 20]. 9. Discussion Looking to relations (9), (17), (23) and appropriate figures to these relations, it can be noted that is difficult to say which MPC architecture is more energy efficiently for a specific operating mode of PFCV. Depending by the energy efficiency coefficients of all PIs, and kF or kL ratio, each of above MPC architectures could represent an energy efficient solution, but in terms of need and flexibility for PFCV the basic MPC could be the best choice. Consequently, the conditions of energy efficiency for MPCb in different operating mode will be discussed below considering the common value of kESS parameter. The mathematical relations (8) and (9) indicate that variations in the PIs efficiencies and kFC ratio modify the ESS limit power for which MPCb architecture is more efficient than MPC1 architecture:
33
MPCb MPC1 k ESS CFC CL 0 k ESS [ 2 1 k FC ( 23 1 )] 1 / 5 2 4 0
(26)
Because the parameter CFC may can be positive or negative (see Figure 4b), the following equivalent mathematical relations can be written: { 23 1 } {k FC (2 1 ) / (23 1 )} {k ESS CL / CFC 0} MPCb MPC1
(27a)
{ 23 1 } {k FC ( 2 1 ) / ( 23 1 )} {k ESS CL / CFC 0} MPCb MPC1
(27b)
{ 23 1 } {k FC (2 1 ) / (23 1 )} {k ESS CL / CFC 0} MPCb MPC1
(27c)
{ 23 1 } {k FC (2 1 ) / (23 1 )} {k ESS CL / CFC 0} MPCb MPC1
(27d)
Also, because is obviously that: (2 1 ) (23 1 ) 1 3
(28a)
{ 2 1 } { 23 1 } { 23 1 }
(28b)
{ 2 1 } { 23 1 } {2 1 }
(28c)
{ 2 1 } { 23 1 } 3 1
(28d)
{k FC 0} {k FC 1}
(28e)
and it is impossible to have:
the relations (27) can be rewritten in a concise equivalent form (after canceling of impossible cases) as: {[ 23 1 ] [ 2 1 23 ]} {k ESS CL / CFC 0} MPCb MPC1 , k FC [0,1]
(29a)
{ 2 1 } {0 (1 2 ) / (1 23 ) k FC 1} {k ESS CL / CFC 0} MPCb MPC1
(29b)
{ 2 1 } {0 k FC (1 2 ) / (1 23 ) 1} {k ESS CL / CFC 0} MPCb MPC1
(29c)
The power flows to load from FC system and ESS via PI1 and PI2. Consequently, the relation between energy efficiencies of these PIs decides the above cases mentioned in relations (29). It can be noted that ESS can work in all operating modes (CS, CD and CI).
34
The mathematical relations (16) and (17) indicate that variations in the PIs efficiencies and kL ratio modify the ESS limit power for which MPCb architecture is more efficient than MPC2 architecture: { 23 1 } {0 k L 1/ 2 45 } {k ESS CkL / 2 0} MPCb MPC 2
(30a)
{ 23 1 } {0 k L 1 / 2 45 } {k ESS CkL / 2 } MPCb MPC 2
(30b)
The relations (30) can be rewritten in a concise equivalent form as: {[ 23 1 ] [k ESS CkL / 2 ]} {[23 1 ] [k ESS CkL / 2 ]} MPCb MPC 2 , k L [0,1]
(31)
The load is powered from FC system via PI 1 and partially via PI 2 and PI 3 in series, so the relation between energy efficiencies of these PIs decides the above cases mentioned in relations (31). It can be noted that ESS can works in all operating modes (CS, CD and CI), too. So, basic MPC architecture may operate efficiently for all kL values of regenerative braking factor in range, if proper EMU control of kFC ratio is performed in all ESS operating modes.
9. Conclusion In this paper a systemic analysis of energy efficiency for an innovative MPC architecture that could be used in PFCVs was presented. Two MPC architectures (MPC1 architecture - obtained from MPC architecture without reverse power flow during regenerative braking process, and MPC2 architecture - obtained from MPC architecture without ESS charging from the FC system) were derived for PFCVs. The PFCV equipped with one of these MPCs is analyzed in relation with potential connection to PCS in order to exchange energy. The analytical computing for energy efficiency of all MPC architectures is performed when the ESS has final SOC different to initial SOC between two plug-in connections. For MPC1 and MPC2 architectures it was shown that each
35
can operate efficiently under certain conditions, which have been specified. Also, in discussion made above it was concluded that under certain conditions the basic MPC architecture is more efficient than both MPC architectures derived from its architecture, and obviously is more flexible for an appropriate energy management of PFCV under EMU control. The ESS of PFCV could be charged between two plug-in to PCSs on route from FC system and load under EMU control of kFC and kL ratios, respectively. The EMU set these moments for a communicant vehicle and then notify the driver in due time. Monitoring of ESS SOC for these vehicles will require common standards and protocols for data exchange, which should be a priority area for policy making in relation with future developing of EP system to upcoming smart grids that will include renewable PCSs. Improvements in the area of energy supply, infrastructure and communication are underway in all countries. The impact of PFCVs on the electricity sector depends on (1) the magnitude of the market penetration, (2) timing of charging, (3) energy management, (4) structure of the power sector, and, finally but not least, (5) availability of FC energy sources for automotive vehicle [1, 44, 48]. The PFCVs can contribute to a higher potential for intermittent use of the renewable energy sources and also can serve to buffer short term and potentially even long-term imbalances between electricity supply and demand. From these conclusions and above discussions, a number of questions arise for further exploration in next research: (1) What is the suitable type of energy management strategy for EMU of PFCV in relation with upcoming smart grids? (2) What are the optimum requirements for hybrid ESSs (batteries and ultracapacitors) used in various hybrid power sources of PFCV configurations? (3) Can be quantified the impact of plug-in feature on battery life? (4) Can be quantified the PFCV
36
impact on renewable intermittent potential? (5) Can be evaluated the PFCV consumption characteristics under a given EMU strategy? For high flexibility of PFCV connected to upcoming smart grids, this paper proposes for PFCVs an efficient MPC architecture and an alternative operating concept, named charge-increasing mode, which can be used to sell the power back to the grid (to stabilize the grid). Overall, in this paper is shown that MPC efficiency could be increased by EMU control during both chargedepleting (CD) and – charge-increasing (CI) operation modes of ESS. These multi-operating modes that make PFCVs attractive also make it difficult to estimate their in-use fuel economy. The PFCV consumption characteristics in the designated operating modes under EMU strategy and the utility factor must be evaluated as a function of both CD and CI distances, because in the rest of the drive cycle the ESS of PFCV will operate in CS mode. Further, the complications surrounding management of three operating modes make far from a trivial task to design an EMU strategy in order to more closely approximate real-world operation. In this paper was demonstrated for a given MPC that PFCVs can work in all operating mode (CS, CD and CI) at high energy efficiency by controlling one of the kF and kL ratios or both ratios, and proposes the basic MPC as a competitive topology for PFCV architecture from both efficient and flexible energy management point of view.
Acknowledgement. This work was partially supported by research project PC no. 72209 (#3733/1.10.2008) and no. 2-2292 (#0369/7.11.2011), under PNCD II - Partnerships frame.
References
37
[1] Hacker F, Harthan R, Matthes F, Zimmer W. Environmental impacts and impact on the electricity market of a large scale introduction of electric cars in Europe - Critical Review of Literature, ETC/ACC Technical Paper 2009; 4: 56-90. [2] http://www.pluginamerica.org/vehicles [accessed on January 5, 2011]. [3] http://evworld.com [accessed on January 5, 2011]. [4] Bizon N. A new Topology of Fuel Cell Hybrid Power Source for Efficient Operation and High Reliability. J Power Sour 2011; 196(6): 3260–70. [5] Qian Z, Al-Atrash H. Modeling and Control of Three-Port DC/DCConverter Interface for Satellite Applications, . IEEE Trans Power Electron 2010; 25(3): 637-49. [6] Jiang W, Fahimi B. Multiport Power Electronic Interface—Concept, Modeling, and Design, IEEE Trans Power Electron 2011; 26(7): 1890-1900. [7] McDonough M, Shamsi P, Fahimi B. Application of multi-port power electronic interface for contactless transfer of energy in automotive applications, Vehicle Power Propul Conf 2009: 936-40. [8] Geidl M, Andersson G. Optimal Power Flow of Multiple Energy Carriers. IEEE Trans Power Electron ; 22(1): 145 – 155. [9] Geidl M, Koeppel G, Favre-Perrod P, Kloeckl B, Andersson G, Frohlich K. Energy hubs for the future. IEEE Power Energy Mag 2007; 5(1): 24–30. [10] Krause T; Andersson, G; Frohlich K, Vaccaro, A. Multiple-Energy Carriers: Modeling of Production, Delivery, and Consumption. Proceedings of the IEEE 2011; 99(1):15 – 27. [11] del Real AJ, Arce A, Bordons C. Optimization strategy for element sizing in hybrid power systems. J Power Sour 2009; 193(1): 315–321.
38
[12] Ramos-Paja CA, Romero A, Giral R, Calvente J, Martinez-Salamero L. Mathematical analysis of hybrid topologies efficiency for PEM fuel cell power systems design, Electrical Power and Energy Systems 2010; 32: 1049–61. [13] Mehrdad E, Yimin G, Ali E. Modern Electric. Hybrid Electric, and Fuel Cell Vehicles. 2nd ed. New York: CRC Press; 2010. [14] Bayindir KC, Gozukucuk MA, Teke A. A comprehensive overview of hybrid electric vehicle: Powertrain configurations, powertrain control techniques and electronic control units. Energy Conversion and Management 2011; 52(2): 1305–13. [15] Chakraborty S, Kramer B, Kroposki B. A review of power electronics interfaces for distributed energy systems towards achieving low-cost modular design, Renewable and Sustainable Energy Reviews 2009; 13(9): 2323–35. [16] Erdinc O, Uzunoglu M. Recent trends in PEM fuel cell-powered hybrid systems: Investigation of application areas, design architectures and energy management approaches, Renewable and Sustainable Energy Reviews 2010; 14: 2874–84. [17] Gao D, Jin Z, Lu Q. Energy management strategy based on fuzzy logic for a fuel cell hybrid bus. J Power Sources 2008; 185(1): 311–7. [18] O. Erdinc, B. Vural, M. Uzunoglu.A wavelet-fuzzy logic based energy management strategy for a fuel cell/battery/ultra-capacitor hybrid vehicular power system J Power Sources 2009; 194(1): 369-380. [19] Rodatz P, Paganelli G, Sciarretta A, Guzzella L. Optimal power management of an experimental fuel cell/supercapacitor-powered hybrid vehicle. Control Eng Pract 2005; 13(1): 41–53.
39
[20] Feroldi D, Serra M, Riera J. Energy management strategies based on efficiency map for fuel cell hybrid vehicles. J Power Sources 2009; 190(2): 387–401. [21] Erdinc O, Vural B, Uzunoglu M, Ates Y. Modeling and analysis of an FC/UC hybrid vehicular power system using a wavelet-fuzzy logic based load sharing and control algorithm. Int J Hydrogen Energy 2009; 34(12): 5223–33. [22] Liu G, Zhanga J, Sun Y. High frequency decoupling strategy for the PEM fuel cell hybrid system. Int J Hydrogen Energy 2008; 33(21): 6253–61. [23] Thounthong P, Rael S, Davat B. Energy management of fuel cell/battery/supercapacitor hybrid power source for vehicle applications. J Power Sources 2009; 193(1): 376-85. [24] Wang Y, Chen KS, Mishler J, Cho SC, Adroher XC. A review of polymer electrolyte membrane fuel cells: Technology, applications, and needs on fundamental research. Appl Energy 2011; 88(4): 981–1007. [25] Doucette RT, McCulloch MD. Modeling the prospects of plug-in hybrid electric vehicles to reduce CO2 emissions. Appl Energy 2011; 88(7); 2315–23. [26] Kuperman A., Aharon I. Battery–ultracapacitor hybrids for pulsed current loads: A review, Renewable and Sustainable Energy Reviews 2011; 15 981–92. [27] Tang Y, Juan W, Pan M, Wan Z. Experimental investigation on the dynamic performance of a hybrid PEM fuel cell/battery system for lightweight electric vehicle application, Appl Energy 2011; 88(1): 68-76 [28] Barelli :L, Bidini G, Ottaviano A. Dynamic analysis of PEMFC-based CHP systems for domestic application. Applied Energy 2012; 91(1): 13–28. [29] Thounthong P, Rael S, Davat B. Control strategy of fuel cell/supercapacitors hybrid power sources for electric vehicle. J Power Sources 2006; 158(1): 806-814.
40
[30] Markel T, Kuss M, Denholm P. Communication and Control of Electric Drive Vehicles Supporting Renewables, Vehicle Power Propul Conf 2009: 27-44. Available in: http://www.nrel.gov/vehiclesandfuels/vsa/pdfs/46224.pdf [accessed on January 5, 2011]. [31] Singh B, Singh S. State of the Art on Permanent Magnet Brushless DC Motor Drives, J Power Electron 2009; 9(1): 1-17. [32] Zeraoulia M, Benbouzid MEH, Diallo D. Electric motor drive selection issues for HEV propulsion systems: a comparative study, IEEE Trans Vehicular Tech 2006; 55(6): 1756-64. [33] Shen M, Peng FZ. Converter systems for hybrid electric vehicles. Electrical Machines Syst Conf 2007: 2004-10. [34] Peng FZ, Yuvan X, Fang X, Qian Z. Z-source inverter for motor drives. IEEE Trans Power Electron 2005; 20(4): 857-63. [35] Thangaprakashy S, Krishnan A. Current mode integrated control technique for Z-source inverter fed induction motor drives. J Power Electron 2010; 10(3): 285-92. [36] Haiping X, Peng FZ, Chen L, Wen X. Analysis and design of bi-directional Z-source inverter for electrical vehicles. APEC Conf 2008: 1252-57. [37] Dehghan SM, Mohamadian M, Yazdian A. Hybrid electric vehicle based on bidirectional Zsource nine-switch inverter, IEEE Trans Vehicular Tech 2010; 59(6): 2641-53. [38] Yamanaka M, Koizum H. A bi-directional Z-source inverter for electric vehicles. Power Electron Drive Syst Conf 2009: 574-8. [39] Ding X, Qian Z, Yang S, Peng FZ. A new feedforward compensation to reject DC-link voltage ripple in bi-directional Z-source inverter ASD system. APEC Conf 2008: 1809-13.
41
[40] Fontes G, Turpin C, Astier S, Meynard TA. Interactions between fuel cells and power converters: Influence of current harmonics on a fuel cell stack, IEEE Trans Power Electron 2007; 22(2): 670-678. [41] Thounthong P, Davat B, Raël S, Sethakul P. An Overview of Power Converters for a Clean Energy Conversion Technology, IEEE Ind Electron Magaz 2009; 3(1): 32-46. [42] Wang Y, Chen KS, Mishler J, Cho SC, Adroher XC. A review of polymer electrolyte membrane fuel cells: technology, applications, and needs on fundamental research. Appl Energy 2011; 88(4): 981–07. [43] Bizon N. Nonlinear control of fuel cell hybrid power sources: Part I –Voltage control, Appl Energy 2011; 88(7): 2559-74. [44] Duleep G, Van Essen H, Kampman B, Grünig M. Impacts of Electric Vehicles - Assessment of electric vehicle and battery technology, Delft, CE Delft 2 (2011) 9-29. [45] Duduta M, Ho B, Wood VC, Limthongkul P, Brunini VE, Carter WC, Chiang Y-M. SemiSolid Lithium Rechargeable Flow Battery, Advanced Energy Materials Adv. Energy Mater. 2011; XX: 1–6, [46] Grüning M, Witte M, Boteler B, Kantamaneni R, Gabel E, Bennink D et all. Impacts of Electric Vehicles - Assessment of the future electricity sector, Delft, CE Delft 3 (2011) 35-70. [47] Tang Y, Yuan W, Pan M, Li Z, Chen G, Li Y. Experimental investigation of dynamic performance and transient responses of a kW-class PEM fuel cell stack under various load changes. Appl Energy 2010; 87(4): 1410–17 [48] Wagner G-O. Real-time diagnosis and maintenance for automotive system, J Electronics and Computer Science 2011; 2: 51-6, http://fecc_bulletin.upit.ro/PublicSite/Main.aspx?cid=14 [accessed on January 5, 2012].
42
