A Novel Mutual Information and Partial Least Squares Approach for Quality-Related and Quality-Unrelated Fault Detection

Partial least squares (PLS) and linear regression methods are widely utilized for quality-related fault detection in industrial processes. Standard PLS decomposes the process variables into principal residual parts. However, as part still contains many components unrelated to quality, if these were not removed it could cause false alarms. Besides, although do affect product they have a great impact on safety information about other faults. Removing discarding will lead reduction rate of faults, quality. To overcome drawbacks PLS, novel method, MI-PLS (mutual PLS), is proposed this paper. The algorithm utilizes mutual divide selected components, then uses singular value decomposition (SVD) further decompose quality-unrelated subsequently constructing monitoring statistics. ensure that there no loss can be used detection, component analysis (PCA) model performed obtain its score matrix, which combined with total Finally, method applied numerical example Tennessee Eastman process. has lower computational load more robust performance compared T-PLS PCR.