Proton exchange membrane fuel cell system diagnosis based on the multivariate statistical method

Although electric-powered vehicles have developed rapidly in recent years, with significant progress in the lithium power battery industry, the Fuel Cell Electric Vehicle (FCEV) is still a competitive choice for a clean transportation solution, because of its extended driving range, zero emissions, and fast fuel recharging capability. In particular the fuel cell hybrid bus used for city traffic is the FCEV type most likely to be commercialized. Demonstration programs for a fuel cell bus fleet have been operated for a few years in China. It is necessary to develop comprehensive diagnostic tools to increase the reliability of these systems, because fuel cell city buses serve large numbers of passengers using public transportation. This paper presents a diagnostic analysis and implementation study based on the Principal Component Analysis (PCA) method for the fuel cell system. This diagnostic system was successfully implemented for detecting a fuel cell stack sensor network failure in the fuel cell bus fleet at the Shanghai Expo in 2010.