Expected-Mode Augmentation Based Variable Structure Multiple-Model Approach for Multiple Faults Detection and Isolation

摘要

In this paper, we proposed a new approach for multiple faults detection and isolation (FDI). This FDI method combines the variable-structure interacting multiple-model (VSIMM) method with expected mode augmentation (EMA). In this approach, the model set is composed of a basic model set and an adaptive EMA set. The EMA set is generated from the basic models based on their predicted probabilities. This variable structure method reduces the size of the model set which is very large in fixed structure. It can detect and isolate single or multiple simultaneous faults. The simulation results show that the proposed method is very effective in detecting faults of various types and estimate unknown fault parameters.