Research on Turbine Rotor Fault Diagnosis Based on CPSO-BBO Optimization SVM

Abstract

In order to improve the recognition accuracy and efficiency of the turbine rotor fault diagnosis, a fault diagnosis method based on CPSO-BBO (chaos particle swarm optimization biogeography based optimization) algorithm was proposed to optimize SVM (support vector machine). Firstly, four common faults states of turbine rotors were simulated by ZT-3 rotor test-bench to obtain fault data. Secondly, CEEMD (complementary ensemble empirical mode decomposition) is used to decompose the rotor vibration signal, and the more effective intrinsic mode function (IMF) is screened by combining the variance contribution rate, and the corresponding PE (permutation entropy) is calculated as the fault characteristic value. After that, chaos theory and PSO (particle swarm optimization) algorithm are combined into the theory of BBO (biogeography based optimization) to obtain CPSO-BBO algorithm, which is used to optimize SVM to obtain the optimal parameters of the diagnostic model. Finally, the fault identification is studied by using the acquired fault data.