Online monitoring of chromium sludge recycling process using recursive MPCA

In this paper, recursive multi-way principal component analysis (MPCA) is applied to the batches of chromium sludge recycling process for the detection of faults without false alarms. The data matrix is being augmented sample wise to reflect the process changes in MPCA. Multivariate statistical process control (MVSPC) techniques; Hotelling's T and squared prediction error; Q, based on PCA have facilitated the analysis of successful and unsuccessful batches. The results have shown that the faults were very quickly identified.