The improved KICA method based on wavelet transform for industrial process monitoring

With the continuous development of large-scale, continuous and complex industries process, the reliability and safety requirements are getting higher and higher of the industrial equipment and control system. Process monitoring has become one of the most important research directions in process automation. In this paper, the ICA-based continuous industrial process system monitoring methods were studied in depth; furthermore, the problem of sorting the independent components in the ICA method is described in detail; at last, make full use of the advantages of independent component analysis, proposed improved fault diagnosis method. The simulation application shows the advantages of the proposed method.