Comparative study of inversion-based and observer-based approaches for fault diagnosis in DC-DC converters

Abstract

Recently, model-based approaches have been proposed for fault diagnosis as an emerging alternative to traditional techniques. Particularly, the inversion-and observer-based were proposed for fault diagnosis as successful model-based approaches. However, the research's results related to these approaches are reduced and generally applied to a specific converter or control type. Therefore, the inversion-and observer-based approaches performance presented in the literature does not permit to compare both techniques in order to conclude which one is more suitable for fault diagnosis. In this context, this paper presents a comparative study of inversion-and observer based approaches for fault diagnosis in a DC-DC boost converter. More specifically, after modeling both the power converter and the fault detection and isolation methods, it were inserted faults on the transistor and capacitor (degradation) in order to validate the performance of the proposed approaches. The results demonstrate that these are efficient alternatives to the conventional techniques that are usually used for fault diagnosis in power electronics systems. Both inverse-and observer-based approaches showed similarities and effectiveness in detecting and isolating faults on the studied DC-DC boost converter.