Fault diagnosis approach based on probabilistic neural network and wavelet analysis

A fault diagnosis method based on probabilistic neural network and Harr wavelet (HWPNN) to Tennessee Eastman (TE) process was presented. Noises and outliers in the data were firstly eliminated by Harr wavelet, and then the denoised data were used in probabilistic neural network to diagnose the faults. To validate the performance and effectiveness of the proposed scheme, the HWPNN was applied to diagnose the faults in TE process, and the classification accuracies of the classifiers were compared. The results showed that significant improvement in diagnosis accuracy was achieved by using HWPNN. HWPNN is better than PNN in classification ability and fault diagnosis accuracy.