GA based fault parameter identification for hybrid system with unknown mode changes

This paper presents a method for the identification of the fault parameters of hybrid systems with unknown mode changes after fault occurring. The identification method utilizes genetic algorithm (GA) to identify fault parameters and unknown mode changes simultaneously based on global analytical redundancy relation (GARR). Fault parameters and mode change time of all switches are encoded into one chromosome as potential solution of the identification process. The GARR is adopted as the performance index of GA search. With the fault parameter values identified by the proposed method, we can tell the healthy status of monitored hybrid system. Experiment results show the efficiency of the proposed method.