A comprehensive observer-based fault isolation method with application to a hydraulic power train

This paper concerns observer-based fault isolation. As a new contribution it combines structural diagnosability analysis with observer-based residual generation. Structural analysis techniques are used to find diagnosable subsystems within the system model from each of which an observer can be built. The resulting bank of observers provides a vector of residuals. Evaluation of this residual vector is realized by structural methods which makes fault isolation possible. The result is a comprehensive way to build a fault diagnoser for a class of nonlinear systems. The proposed diagnosis method is evaluated at an automotive application, which is a hydraulic power train.