Distributed Fault Diagnosis for Electro-hydraulic Suspension by Model Decomposition

This paper presents a distributed fault diagnosis approach for electro-hydraulic suspension system based on model decomposition technique. As a complex nonlinear system, the electro-hydraulic suspension involves various components and traditionally centralized diagnosis method will suffer from some problems such as poor scalability, high computation cost and so on. To solve above problems, the distributed diagnosis approach is adopted in this paper. Firstly, the diagnostic hybrid bond graph model of electro-hydraulic suspension system is decomposed into several submodels by model decomposition technique. Then, the distributed analytical redundancy relations are derived from each submodel for fault detection, and the fault signature matrix is constructed to isolate possible faults set (PFS). After that, a Cuckoo Search algorithm cooperated with differential evolution is proposed to determine the true faults from the PFS. Finally, simulation results validate the effectiveness of the proposed distributed method.