Experimental investigation on statistical characteristics of cell voltage distribution for a PEMFC stack under dynamic driving cycle

Abstract

The proton exchange membrane fuel cell (PEMFC) stack consists of individual cells in series. Its operating life is subjected to performance of the weakest cell because of the short-board effect, thus voltage uniformity during dynamic long-running process is significant to its durability. In this work, based on a 1044 h aging experiment on a 6.55 kW PEMFC stack under dynamic driving cycle, voltage uniformity is analyzed. In a single cycle, voltage uniformity becomes worse with the increase of loading current and there are some local maxima of voltage coefficient variation (Cv) at the moment of loading or unloading step. Cv value at higher current is greater and increases faster with cycles. At the end of experiment, Cv at 135 A is more than 6%. Besides, skewness (Sk) is used to evaluate the skew direction and degree of cell voltage data in a cycle. In most cycles, Sk values at 34.22 A are above 0 and Sk values at 59.70 A and 135 A are below 0. After the Box-Cox transformation, which is used to improve symmetry of data and reduce Sk, the cell voltage data have passed the verifications of normal fitting, probability-probability plot and quantile-quantile plot. Therefore, it is found that cell voltage data tend to obey skewed normal distribution, which is of positive significance for improving voltage uniformity and durability of PEMFC stack.