Multivariate statistical modeling and monitoring of SBR wastewater treatment using double moving window MPCA

As effluent criteria become increasingly stringent, advanced monitoring and control strategies for wastewater treatment process (WWTP) attract more and more attention. Multivariate statistical process control has been found wide applications to process performance analysis and monitoring. A simple and straight multivariate statistical model based on double moving window mechanism is used for online monitoring the progress of sequencing batch reactor for wastewater treatment. It replaces an invariant fixed-model monitoring approach with adaptive updating model data within batch-to-batch, which overcomes the problem of changing operation condition and slow time-varying-behavior of industrial processes. It replaces an whole static model with multiple local models along time axis, which copies seamlessly with variable run length and need not estimate any deviations of the ongoing batch from the average trajectories. The case studies demonstrate that the MPCA model using double moving window performs better than a single MPCA model for all the operation time.