| A key factor for Proton Exchange Membrane Fuel Cell(PEMFC)to be applied in large scale in vehicles,which is start-up under zero degree,hasn’t been solved yet.The start-up performance of fuel cell stacks under zero degree is largely influenced by the initial water content in stack before start-up.Gas purge during shutting down of the stack can effectively change the initial water content in stack before start-up,thus promoting the start-up performance of fuel cell stacks under zero degree.Therefore,it’s very necessary to study gas purge process during shutting down of the stack.As nitrogen is hard to be carried in actual vehicles,and purging the anode with hydrogen consumes a lot of hydrogen fuel,also water is always produced and accumulated at the cathode,this paper is going to focus on cathodic gas purge process.This paper is going to study the issue from three aspects,which are experimental study,modeling and optimal control strategies.In order to study gas purge process during shutting down of the stack,firstly,we built an experimental device in the environmental cabin.A lot of cathodic gas purge experiments have been conduct to study the influences of different purge factors on purge effectiveness and purge energy consumption.Because water content in fuel cell stack is hard to be measured directly,the integration method using the values of temperature and relative humidity of the purge gas to calculate water removal rate is used to characterize water content change in fuel cell stack in this paper.The experimental results show that two key factors which have decisive influences on purge effectiveness are the temperature of the stack and flow rate of purge gas.According to the experimental results,a qualitatively purge strategy can be promoted that gas purge process should be done before the temperature of the stack has been decreased largely and flow rate of purge gas should be large at the beginning of purge process while flow rate of purge gas should be small at the end of the purge process.Secondly,based on previous studies gas purge process during shutting down of the stack is divided into three stages,which are through-plane water transfer stage,in-plane water transfer stage and membrane water transfer stage.A mechanism model of gas purge process during shutting down of the stack has been built based on previous studies,which has been added by high frequency impedance model of the membrane and purge energy consumption model in this paper.The accuracy of the model has been verified by the experimental results.Besides,the influences of some factors on purge effectiveness which are hard to study through experiments,such as inlet relative humidity of purse gas and initial water content of the stack before gas purge process,are studied based on the model we have built.The simulation results show that inlet relative humidity of purse gas mainly influences the water content that can be left in the stack finally,and initial water content of the stack before gas purge process mainly influences the total purge time of the whole gas purge process.Thirdly,three different optimal control strategies with the least purge energy consumption under the limit of purge time have been proposed,which are constant flow rate control strategy,variable flow rate global optimization control strategy based on dynamic programming algorithm and rule-based three-stages real-time control strategy.Constant flow rate control strategy is based on the maps which are obtained from the model,which is simple and it can realize least energy consumption while the flow rate is a constant during the whole purge process under the limit of maximum purge time.The variable flow rate global optimization control strategy is based on dynamic programming algorithm,which has least energy consumption of all purge strategies,while it takes much time to run the program,so it can’t be used for real-time control.The rule-based threestages control strategy is induced from the results of dynamic programming algorithm,whose energy consumption is nearly the same with the control strategy based on dynamic programming algorithm and it’s easy to be used for real-time control. |