Multivariate nonlinear statistical process control of a sequencing batch reactor

Abstract

This research describes the application of a multivariate statistical process control method to a pilot-scale sequencing batch reactor (SBR) using a hatchwise nonlinear monitoring technique for a denoising effect. Three-way batch data of normal batches are unfolded batch-wise and then a kernel principal component analysis (KPCA) is applied to capture the nonlinear dynamics within normal batch processes. The developed monitoring method was successfully applied to an 80-l sequencing batch reactor (SBR) for biological wastewater treatment, which is characterized by a variety of nonstationary and nonlinear characteristics. In the multivariate analysis and batch-wise monitoring, the developed nonlinear monitoring method can effectively capture the nonlinear relations within the batch process data and clearly showed the power of nonlinear process monitoring and denoising performance in comparison with linear methods.