An Improved Fault Diagnosis Approach Using LSSVM for Complex Industrial Systems

Fault diagnosis is a challenging topic for complex industrial systems due to the varying environments such find themselves in. In order improve performance of fault diagnosis, this study designs novel approach by using particle swarm optimization (PSO) with wavelet mutation and least square support (LSSVM). The implementation entails following three steps. Firstly, original signals are decomposed through an orthogonal packet decomposition algorithm. Secondly, reconstructed obtain features. Finally, extracted features used as inputs model established in research classification accuracy. This joint method not only solves problem PSO falling easily into local extremum, but also improves effectively. Through experimental verification, optimazation sqaure vector machine ( WMPSO-LSSVM) has maximum recognition efficiency that 12% higher than LSSVM 9% extreme learning (ELM). error corresponding regression under WMPSO-LSSVM algorithm 0.365 less traditional linear model. Therefore, proposed scheme can effectively identify faults occur systems.