On-line Process Monitoring based on Wavelet-ICA Methodology

In this paper, a new process monitoring methodology is presented to detect fault occurrence. The proposed methodology incorporates a wavelet de-noising approach based on the fast wavelet transform (FWT) to extract the embodied fault dynamics from the noisy measured data. A level dependent soft thresholding technique using Daubechies 3 with three levels of decomposition is utilized. An appropriate sliding window scheme is presented to enable on-line implementation of wavelet denoising filtering. An ICA statistical monitoring technique is employed to detect fault. To enhance ICA monitoring capability, a new statistic measure is developed to cater for monitoring the excluded part which has not been captured by the main dominant part. An approach based on cumulative percent variance (CPV) is presented to mechanize the selection of dominant independent components in the presented monitoring methodology. The effectiveness of the proposed wavelet-ICA approach will be demonstrated by applying on the Tennessee Eastman challenge process plant.