Product quality and operation safety are important aspects of industrial processes, particularly those with large numbers of correlated process variables. Principal component analysis(PCA) has been widely used in multivariate process monitoring for its ability to reduce process dimensions. PCA and other statistical techniques, however, have difficulties in differentiating faults with similar time-domain process characteristics. A wavelet-based time-frequency approach is developed in this paper to improve PCA-based methods by extending the time-domain process features into time-frequency information. Subsequently, a similarity measure is presented to compare process features for on-line process monitoring and fault diagnosis. Simulation results show that the proposed multivariate time-frequency process feature is effective in both fault detection and diagnosis, illustrating the potentials for real-world application.

A PLS-based multirate dynamic modeling is proposed for quality prediction at a faster rate for multivariable processes with di6erent sampling rates between the process and quality variables. Depending on the nature of the process and quality variables, two model representations are proposed, one-block form for the process with closed-loop quality control, and two-block form for the process without the closed-loop quality control. The applications of the proposed models are compared with respect to two processes, a three-stage absorber and the TE process. The ability of the proposed model to predict quality at a faster rate is demonstrated with the complex TE process.

This paper extends the stage-based sub-PCA modeling method originally proposed by the authors to the monitoring of batch processes with durations of uneven lengths. Two models for each stage are developed, one for the stage division and the other for process monitoring. The purposes of the stage division are two-fold, to enhance process understanding and to provide stage-division information necessary for the development of PCA monitoring models. With the proposed method, batch processes with durations of uneven lengths can be effectively monitored for fault detection and diagnosis.