Study On Data-driven Fault Diagnosis For Proton Exchange Membrane Fuel Cell Stack

As the energy crisis and environmental pollution become more and more serious around the world,fuel cell has become viewed as an important area all over the world.The performance,reliability,environmental adaptability and life span of fuel cell system will directly affect the large-scale commercial promotion of fuel cell.Therefore,it is an important research area that to accurately diagnose,locate the fault location,eliminate the fault in time and ensure the smooth and reliable operation of fuel cells.For proton exchange membrane fuel cell stack,this paper focuses on data-driven fault diagnosis based on the fuel cell experiment platform from the research group.The main contents are as follows:Firstly,several typical fault diagnosis methods are analyzed in this paper.Two kinds of common fault types of proton exchange membrane fuel cell(PEMFC)are analyzed and studied.Based on the data source approach of fault diagnosis of PEMFC,the stack voltage and the single fuel cell voltage of the stack were analyzed and their influence on the overall output performance of the PEMFC was analyzed.Single fuel cell voltage can be used as a reference standard for health management and can be used as an effective data source for fault diagnosis.Compared with the short stack,the single cell voltage is more important to the long stack and the voltage uniformity is a factor that must be taken into consideration.According to the structure of PEMFC stack,based on the fuel cell/super capacitor hybrid 100% low floor tram which was researched jointly by our research group and Tangshan railway vehicle co.,LTD,this paper used principal component analysis(PCA)method for feature extraction.And two proposed fault diagnosis method were used,then fault diagnosis model was established,and in the end the actual test data was used to verify the validity of this method.In this paper,the fault diagnosis model of the typical fault in the power reactor of proton exchange membrane fuel cell is established.The results show that the proposed data-driven method could improve the reliability of PEMFC fault diagnosis.