Principal Component Analysis for Fault Detection and Isolation in a DC-DC Buck Converter

Abstract

The principal component analysis (PCA) is as well as being an eminent technique for dimensionality reduction; a powerful fault detection and isolation (FDI) technique. This reliable FDI has ubiquitously and successfully been used for the monitoring of complex systems. However, in the field of power conversion, PCA has not known the success it has had in the other applications. It has not been used at its full potential as a capable and sufficient diagnostic technique. This paper proposes a method for open-circuit fault detection and isolation (FDI) in DC-DC buck power converters using the PCA technique. In this paper, an efficient algorithm is proposed. The algorithm relies on the squared prediction error (SPE) for fault detection, then the contributions of variables to the SPE statistic are calculated, therefore the fault isolation. The effectiveness of the proposed method is illustrated through diverse simulation results.