Performance Analysis for Directional Residual Based Fault Isolation

Authors:	Hu Yongtong, George Mason University, United States Gertler Janos, George Mason University, United States
Topic:	6.4 Safeprocess
Session:	Signal Based Fault Detection and Isolation
Keywords:	Fault detection and diagnosis; Fault isolation; Directional residuals; Statistical testing; Generalized likelihood ratio test

Abstract

Generalized likelihood ratio test (GLRT) for directional residual based fault isolation is studied in this paper. The conditional probability of correct fault isolation is used to measure the performance of the test. The explicit form of this probability is derived with the help of two GLRT invariant properties and the standard fault isolation subspace. The effect of linear transformation is studied as well.