Diagnosis of PEMFC by using data-driven parity space strategy

In this paper, a data-driven strategy is proposed for PEMFC (polymer electrolyte membrane fuel cell) diagnosis. In the strategy, parity space is directly identified from normal process data without modeling. With the identified parity space, a group residuals can be generated and evaluated to achieve fault detection. In addition, a multi-class SVM (support vector machine) is adopted to realize fault isolation. Experiments of a 40-cell stack are dedicated to highlight the approach.