Fusion of Model-based and Data-based Fault Diagnosis Approaches

Abstract

This paper presents a generic approach that combines model-based and data-based methods for fault detection and diagnosis. A proposed generic representation framework is used to express the different diagnosis results and to merge them without taking into account neither their internal characteristics, nor the nature of their outputs. Within this framework, the generic approach is performed in two steps. The first step consists in operating several diagnosis methods using the system measurements. The second step deals with the fusion of various methods results. This diagnosis approach is evaluated and tested on an anti-lock braking system. Simulations show that methods combination and results fusion make our diagnosis approach more efficient.