Diesel engine fault diagnosis method based on MPDE-SVM and its application

There is much difficulty in fault diagnosis because of lacking of system fault samples. So this paper presents a mixed strategy of combining differential evolution algorithm with local enhanced operator with the optimization of the parameters of support vector machine. For the diesel engine valve clearance fault diagnosis, the measured diesel engine valve vibration signals data after wavelet transform are used as the characteristic values of the diagnostic model. Then the diesel engine valve clearance fault diagnosis model of SVM based on MPDE is established and compared with the BPNN, SVM and IQPSO-RBF models, The results showed that the performance of the SVM optimized by MPDE algorithm is better than the models based on the other three algorithms. The method is practical and effective.