Fuzzy Logic Control of a Fuel Cell/Battery/ Ultra-capacitor Hybrid Vehicular Power System M. C. Kisacikoglua, b, Student Member, IEEE, M. Uzunoglua, c, Member, IEEE, M. S. Alama, Senior Member, IEEE a
Department of Computer and Electrical Engineering, University of South Alabama, Mobile, AL 36688, USA Department of Computer and Electrical Engineering, University of Tennessee, Knoxville, TN 37996, USA c Department of Electrical Engineering, Yildiz Technical University, Istanbul 34349, Turkiye
[email protected],
[email protected],
[email protected]
b
Abstract- Fuel cell (FC) hybrid electric power systems appear to be very promising in satisfying high energy and high power requirements for vehicular applications. The power management of the hybrid system is very important for dynamic response under various load conditions. In this study, an FC and an energy storage system (ESS) supply power demand using a DC/DC boost converter and a dual source bidirectional DC/DC converter, respectively. We focus on a fuzzy logic control (FLC) algorithm integrated into the power conditioning unit (PCU) for FC/Battery/Ultra-capacitor (UC) hybrid vehicular power systems. The control strategy is capable of managing power flow from FC system, battery and UC bank and keeps the DC bus voltage around its nominal value by supplying propulsion power and recuperating braking energy. Simulation results obtained using MATLAB® & Simulink®, SimPowerSystems®, Fuzzy Logic Toolbox and ADVISOR® are presented to verify the effectiveness of the proposed control algorithm.
I.
INTRODUCTION
Increasing concerns of global and local pollution, depletion of fossil fuels, and higher gas prices motivated ambitious plans for new vehicle types with alternative energy sources. FC vehicles are expected to be the next generation vehicle type since they promise very clean operation and show higher energy efficiency than conventional vehicles [1]. Fuel cells are electrochemical devices that convert chemical energy of a reaction directly into electrical energy. Among the various types of fuel cells, Proton Exchange Membrane Fuel Cell (PEMFC) has drawn the most attention due to its simplicity, viability, quick start up, higher power density and operation on lower temperatures [2], [3]. Therefore, PEMFC is a serious candidate for automotive applications [2]. However, a sole FC system may not be sufficient to satisfy the load demands, especially during start-up and transient events. Moreover, an FC system alone would have to supply all power demand thus increasing the size and cost of the FC system. As a result, downsizing FC system and hybridizing it with an energy storage system decreases system cost, improves performance, and provides capturing regenerative braking energy. In this study a dual energy storage system, comprising of batteries and ultra-capacitors, is proposed. An energy storage system with a battery-ultra-capacitor combination provides higher power density with increased energy storage capability. Therefore, with a dual energy storage system, the capability of
the peak shaving of the vehicle power profile and recovering regeneration energy will be increased. Another significant contribution to effective operation is the design and control of the PCU integrated into the hybrid system. There are several PCU topologies that can be used in hybrid systems. The location of ESS in a hybrid system and the connection way to the DC bus distinguishes the topology of the PCU. Among these topologies, parallel connection of the FC system and ESS to the DC bus via two different DC/DC converters may be chosen for better system utilization [4], [5]. In this paper, a DC/DC boost converter and a dual source bidirectional DC/DC converter [6] are employed for the power control of FC/Battery/UC hybrid system. This topology boosts output voltages of the hybrid sources to higher values and enables the control of the DC bus voltage for the proper operation of the drive inverter. Also, the proposed FLC algorithm manages power flow of the system. In the control system, vehicle load is separated into three components; steady-state, intermediate and transient which are supplied by FC, battery and UC respectively. Moreover, DC bus voltage is kept inside a tolerable region. II. SYSTEM DESCRIPTION A. Design and Dynamic Modeling of PEMFC The PEMFC model used in this paper is realized in MATLAB and Simulink. This model has been built using the relationship between output voltage and partial pressure of hydrogen, oxygen and water. Fig. 1 shows the detailed model of the PEMFC, which is then embedded into the SimPowerSystems of MATLAB as a controlled voltage source and integrated into the overall system. The FC system model parameters used to obtain this model are as follows: Constants to simulate the activation over voltage in PEMFC system [A-1] and [V] Nernst instantaneous voltage [V] Standard no load voltage [V] Faraday’s constant [C kmol-1] Hydrogen valve molar constant [kmol (atm s)-1] Water valve molar constant [kmol (atm s)-1] Oxygen valve molar constant [kmol (atm s)-1] Modeling constant [kmol (s A)-1]
Number of series fuel cells in the stack Number of fuel cell modules Hydrogen partial pressure [atm] Water partial pressure [atm] Oxygen partial pressure [atm] Input molar flow of Hydrogen [kmol s-1] Hydrogen input flow [kmol s-1] Hydrogen output flow [kmol s-1] Hydrogen flow that reacts [kmol s-1] Universal gas constant [(1 atm) (kmol K)-1] FC internal resistance[Ω] The Hydrogen-Oxygen flow ratio Absolute temperature [K] DC output voltage of FC system [V] Utilization rate Hydrogen time constant [s] Oxygen time constant [s] Water time constant [s] Activation over voltage [V] Ohmic over voltage [V]
(2) where is the terminal voltage [V]; and is the maximum voltage [V] of the UC bank. The UC bank model has been implemented in MATLAB and SimPowerSystems for this study. D. Drive Cycle The Urban Dynamometer Driving Schedule (UDDS) [9] is selected as a reference cycle to model, design and simulate the proposed vehicle power system. Fig. 3 shows the power demand of vehicle according to standard UDDS as a function of the time. The power demand required to meet the vehicle speed is obtained from the ADVISOR analysis tool and the load model is realized with respect to this power profile using MATLAB and SimPowerSystems environment.
The FC system consumes hydrogen according to the power demand, and hydrogen is obtained from a high pressure storage tank. To control the hydrogen flow rate according to the FC power output, a feedback control strategy is used. The FC system output current is taken back to the input to adjust the load change and produce the DC output voltage. Detailed description about the modeling of the FC system can be found in [3]. B. Design and Dynamic Modeling of Battery The battery model in the Fig. 2a is used for this study. An internal ohmic resistance ( ) can be considered constant for steady state operation [7]. The state of charge of the battery can be expressed as [8]:
(1) where is the initial charge, is the rated ampere-hour (Ah) value and is the current of the battery. The battery current can either be negative or positive, with respect to the mode of operation, e.g. charging or discharging. The battery model has been implemented in MATLAB and SimPowerSystems for this study. C. Design and Dynamic Modeling of Ultra-capacitor Bank The classical equivalent circuit of the UC unit is showed in Fig. 2b. The model consists of a capacitance (C), an equivalent series resistance (ESR) representing the charging and discharging resistance, and an equivalent parallel resistance (EPR) representing the self-discharging losses. The EPR models leakage effects and affects only the long term energy storage performance of the UC. The SOC of the UC can be expressed as:
Fig. 1. Dynamic model of the FC system.
Ro Vbat
a)
b) Fig. 2. Dynamic models of the ESS components a) The battery model b) The UC model.
Fig. 3. Power demand according to UDDS for an FC powered vehicle.
E. Load Sharing and Sizing of the Power Sources In order to determine the appropriate sizes of the FC system and the ESS, principle of the vehicle load sharing between them should be defined first. FC system is sized to give only the base load or cruising power of the vehicle, and the peak power for up-hill or accelerated driving is provided by the ESS. According to the control algorithm, FC system delivers power to the load, averaged with a filter time constant of 20s [10] which downsized the FC system to 30kW. This power profile also ensures that the membrane is not subjected to sharp peak loads, thus lifetime of the FC system increases. Secondly, ESS size is calculated based on the following roles: Power-assist during the entire drive cycle to achieve peak shaving and improve the transient response capability of the hybrid system. Capturing regenerative braking energy. The individual sizing requirements for each of the objectives are given in Table I. The values are calculated for the entire UDDS. TABLE I INDIVIDUAL REQUIREMENTS FOR THE ESS OBJECTIVES Objective
ESS Peak Power
ESS Energy
Peak Shaving (no regeneration)
27.4 kW
Transient response
No assistance needed for UDDS
Regenerative Braking
21 kW
100 Wh
44 Wh
First objective gives the power and energy values of the ESS to assist the FC system for the entire UDDS. The highest demand is 57.4kW; therefore peak power of the ESS is selected as 27.4 kW. The power demands exceeding the FC system power have a cumulative energy content of 100 Wh. Also, because FC system can increase its power output from 10% of rated power to 90% of rated power in 2s, FC system doesn’t need assistance for the transient events for the UDDS [11]. The first objective is the greatest of the three cases and determines the ESS requirements as shown in the Table I. Since the FC system is capable of supplying 30 kW power, ESS can level the peaks of power with a power rating of 27.5 kW and an energy rating of 100 Wh. The ESS energy requirement is divided between the battery and the UC so that 75% of this energy will be supplied by the battery and the remaining energy (25%) will be given by UC. According to the proposed control algorithm, SOC of the UC can change ∆50% with a center point of 75%. If the SOC of UC is allowed to decrease to 50% from its operating value (75%), the usable energy is calculated as:
(3) Rearranging (3) yields:
(4) Therefore, the UC should be rated to deliver 80 Wh of energy. Proposed control system maintains SOC of the battery around 50% with a delta SOC of 25%. Hence, only 25% of the total capacity is used which implies that the total energy capacity of the battery is: (5) F. Power Conditioning Unit Utilizing three different power sources simultaneously requires a power converter interface to effectively control the power flow. This power interface also allows sharing of the power transmitted from FC system and ESS to the drive train based on the rules defined in the fuzzy logic controller. A DC/DC boost converter couples FC system to drive train and a dual source bidirectional DC/DC converter transmits power from/to ESS. The input/output voltage relationship for the converters is: (6) where is the load voltage [V]; is the input voltage of the source [V]; and is the duty ratio. The minimum and maximum values for the input voltages of the hybrid power sources and the corresponding duty ratios in order to get 400 V DC load voltage, are given in the below table. TABLE II VOLTAGE VALUES OF THE HYBRID SYSTEM Max.-Min. Min.-Max. Power Source Input Voltage (V) Duty Ratio FC System 140 – 120 0.65 – 0.70 Battery 165 – 145 0.59 – 0.64 UC 150 - 75 0.63– 0.81
General system configuration is shown in Fig. 4. The motorinverter system is considered as a “black box”, and its power requirements are known from the UDDS. This power profile also includes the accessory loads which is 700 W. G. Fuzzy Logic Control Main requirements of the control algorithm are satisfaction of the power demand and management of the power flow in accordance with efficient operation of the different power
Fig. 4. Hybrid power system with dual energy storage
sources. Coupled with this, the control system should utilize the FC system, battery and UC bank to match the vehicle load profile with consideration of different features of the power sources. Since stable operation of the FC system is vital for efficiency, lifetime, and cost, the FC system should deliver the base load power without responding to peak power demand. Therefore, the UDDS power profile is averaged with a filter time constant of 20s and compared with the FC system power to form the error signal which is then sent to FLC. In the FLC, steady state power plus ESS recharging power becomes the net power command for FC fed DC/DC converter. Hence, FC system is responsible for delivering the base load power and supervising the SOC of the battery and UC. While FC system satisfies the base load conditions of the load profile, the battery assists the FC system to increase the total energy capacity of the hybrid system. The power profile of the battery is at intermediate level. Therefore, it gives the required energy with responding to load power slower than the UC bank to prevent power cycling losses. Therefore, the UDDS power profile is averaged with a filter time constant of 10s and compared with the battery power to get the error signal. The FLC computes the net power command for the battery with observing the SOC of the UC. Another purpose of the control system is to adopt the UC bank in stabilizing the DC bus voltage. This enables the remaining power to be transferred to/from the UC bank and improves the overall driving performance. Consequently, FC system delivers the steady state load power, battery transfers intermediate power and UC bank controls the DC bus voltage in an acceptable interval, and assists for the fast current response with its high power density. III. SIMULATION RESULTS Simulation results are obtained by developing a detailed model using MATLAB, Simulink and SimPowerSystems using the mathematical and electrical models of the system described
earlier. Simulations are conducted using a computer having 3 GB of RAM. Switching frequency of the power converters should be high, and therefore sample time could not exceed 45μs for reliable operation. Moreover, since memory is limited, total simulation time was decreased to 350s as illustrated in Fig. 5. Nevertheless, the highest power demand and the highest power regeneration made by the vehicle are included in that time interval. Fig. 6 shows the power profile transmitted by the FC system to the drive train. In normal conditions, the FC system transmits power averaged with a 20s time constant and it does not respond to power peaks unless the ESS is in need of charge. When the ESS needs recharge power, the FLC adjusts the duty ratio for the FC-fed DC/DC converter in order to keep SOC of the battery and the UC in an acceptable region as shown in Fig. 7 and Fig. 8. While the FC delivers steady state power, the battery supplies intermediate loads as in Fig. 9 with monitoring SOC of the UC. Fig. 9 shows that battery delivers power with a more energy content, and its SOC has a lower peak to peak value compared to SOC of the UC shown in Fig. 8. On the other hand, during sudden changes in the load, the UC bank supplies the transient power with alacrity. Due to its fast response, the UC fulfils the transient power demand and increases the hybrid system power density as illustrated in Fig. 10. Moreover, the UC bank keeps the DC load voltage within the allowable limits as shown in Fig. 11. It shows that the DC bus voltage stays a stable interval ( %10 for 400V nominal DC bus voltage). Therefore, the hybrid system with dual energy storage is capable of delivering power without fully relying upon the FC system which reduces the size and cost of the FC system.
Fig. 5. Zoomed version of the load power demand
Fig. 9. Battery power
Fig. 6. FC system power
Fig. 10. UC power
Fig. 7. State of charge of the battery
Fig. 11. DC bus voltage
IV. CONCLUSION
Fig. 8. State of charge of the UC
An FC/Battery/UC vehicular hybrid power system is simulated using fuzzy logic control strategy. The hybrid vehicular power system with the proposed control algorithm is used to meet the demand under the load conditions of the UDDS cycle. The proposed load sharing principle demands power as stable as possible from the FC System and uses ESS in both increasing the overall power capability of the hybrid system and regenerating the excess power during decelerations of the vehicle. When the vehicle power is smaller than the FC system power, UC first recovers the energy. If it cannot take all the energy, battery accepts the leftover.
The load power is shared so that FC System delivers base load power, battery supplies the intermediate power and UC gives the instant power. Therefore, FC system is downsized; ESS successfully achieved peak shaving, fuel consumption is decreased, total braking energy is regenerated. ACKNOWLEDGEMENT This work was supported in part by the U.S. Department of Energy under Grant DE-FG02-05CH11295. REFERENCES K.-H. Hauer, “Analysis tool for fuel cell vehicle hardware and software (controls) with an application to fuel economy comparisons of alternative system designs,” Ph.D. dissertation, Dept. Transportation Technology and Policy, Univ. California, Davis, 2001. [2] F. Barbir, “PEM fuel cells: Theory and Practice”, Burlington, MA: Elsevier Academic Press, 2005, pp. 1-10, 342-344. [3] M. Uzunoglu and M. S. Alam,” Dynamic Modeling, Design and Simulation of a PEM Fuel Cell/Ultracapacitor Hybrid System for Vehicular Applications”, Energy Conversion & Management, vol.48, no.5, pp. 1544-1553, May 2007. [4] P. Dietrich, F. Buchi, A. Tsukada, M. Bärtschi, R. Kötz, G. G. Scherer, P. Rodatz, O. Garcia, M. Ruge, M. Wollenberg, P. Lück, A. Wiartalla, C. Schönfelder, A. Schneuwly and P. Barrade. “Hy.Power - a technology platform combining a fuel cell system and a supercapacitor” in Handbook of Fuel Cells – Fundemantals, Technology and Application, W. Vielstich, Arnold Lamm and Hubert A. Gasteiger, Eds. New York: John Wiley&Sons Inc., 2003 vol. 4, pp. 1184-1198. [5] D.-K. Choi, B.-K. Lee, S.-W. Choi, C.-Y. Won, D.-W. Yoo. “A novel power conversion circuit for cost-effective battery-fuel cell hybrid systems”, J. Power Sources, vol. 152, pp. 245-255, December 2005. [6] M. Marchesoni, and C. Vacca, . New DC-DC Converter for Energy Storage System Interfacing in Fuel Cell Hybrid Electric Vehicles., IEEE Tran. Power Electron. , vol. 22, no. 1, pp. 301-308, January 2007. [7] J. Moreno, M. E. Ortuzar, and J. W. Dixon. “Energy-Management System for a Hybrid Electric Vehicle, Using Ultracapacitors and Neural Networks”, IEEE Tran. Ind. Electron., vol. 53, no. 2, pp.614-623, April 2006. [8] J. Lee, J. Jo, S. Choi, and S.-B. Han. “A 10-kW SOFC low-voltage battery hybrid power conditioning system for residential use”, IEEE Tran. Energy Convers., vol.21, no.2, pp. 575-585, June 2006. [9] EPA Urban Dynamometer Driving Schedule (UDDS) [online]. Available: http://www.epa.gov/otaq/emisslab/methods/uddsdds.gif, accessed August 2007. [10] M. Zolot. “Dual-source energy storage−control and performance advantages in advanced vehicles”, EVS-20, November 2003. [11] T. Markel, M. Zolot, K. B. Wipke, and A. A. Pesaran. “Energy storage system requirements for hybrid fuel cell vehicles”, Advanced Automative Battery Conference, Nice, France, June 2003. [1]
