Online Monitoring And Remaining Useful Life Prognostics Of PEMFC

Proton Exchange Membrane Fuel Cell(PEMFC)is the most promising technology for mobile energy storage equipment,stationary power stations and transportation.The main obstacles to the commercialization of PEMFC technology are its durability.Continuous monitoring of PEMFC's degradation processes and accurate prognostics of remaining useful life can be beneficial for PEMFC's performance optimization.Therefore,it's vital to study on state of health assessment and remaining useful life prognostics on PEMFC.This paper reviews the current research status of PEMFC remaining useful life prognostics,introduces the working principle,degradation mechanism and output characteristics of PEMFC,and elaborates the system factors such as hydrothermal balance management and operating conditions that affect the output performance of PEMFC.The selection of degradation indicators is analyzed to provide a theoretical basis for subsequent experimental design,state of health assessment and prognostics of remaining useful life.Firstly,the paper conducts temperature and exhaust experiments on the self-humidifying open cathode PEMFC platform.The effect of spacing and load conditions on output characteristics.Then,Design 900 W vehicle PEMFC design vehicle control circuit,auxiliary circuit and LabVIEW-based online monitoring platform to realize real-time collection of fuel cell operation data,monitoring whether PEMFC has fault or performance degradation,thus ensuring safe and stable operation of vehicle PEMFC.And then,in order to realize the state of health assessment of PEMFC,the raw data is preprocessed,indicators that can reflect the degradation of PEMFC is extracted,the PEMFC health indicator HI is established by combining the voltage balance of the cell,and the health status of the PEMFC is determined by two distance analysis methods.Evaluate and compare the effects of the two algorithms on the actual degradation of PEMFC.Finally,Based on the theory of deep learning algorithm,construct a deep convolutional neural network,use the Python platform to train the model,and combine the previous experimental results,select the current density,the temperature of the inlet and outlet reaction gas,that is,the pressure,the cooling water temperature of the inlet and outlet ports.The input of the neural network considers the output voltage of the stack as the output of the neural network.The results show that the trained network can accurately predict the time node at which the PEMFC reaches the voltage failure threshold,that is,to realize the useful remaining life prognostics of PEMFC,which support the stable operation of PEMFC.