Multi-step ahead fault prediction method based on PCA and EMD

In recent years, fault prediction method, which means forecast process fault in an early time based on the current condition of the system, has attracted more and more attention by companies and scientists. However, it still has many problems in this area, especially for its application in industrial process. In the present work, a multi-step ahead fault prediction method combining principle component analysis, empirical mode decomposition and extreme learning machine are developed to realize early prediction of fault. The application of the presented method is illustrated with respect to simulated data collected from the Tennessee Eastman process. The experimental results demonstrate the effectiveness of the proposed method.