Monitoring of Wastewater Treatment Plants using Improved Univariate Statistical Strategy

In this paper, we propose an enhanced monitoring of wastewater treatment plant (WWTP) using state estimation-based fault detection strategies. The WWTP state estimation problem is performed using particle filter (PF) technique. The PF has shown good improvement and provides a significant advantage over extended Kalman filter (EKF), unscented Kalman filter (UKF) techniques and can be applied to nonlinear models with non-Gaussian errors. The fault detection phase is achieved using a novel chart called multiscale exponentially weighted moving average chart (MS-EWMA). The new chart allows to optimize the smoothing parameter (λ) and control width L of EWMA chart to deal with the dynamic nature of WWTPs. It enables also to extract accurate deterministic features and decorrelate autocorrelated measurements using dynamical multiscale representation. The fault detection performance is studied using simulated COST wastewater treatment BSM1 model. The BSM1, provided by the IWA Task Group on Benchmarking of Control Strategies, is a simulation platform that allows for creating sensor faults disturbances in a wastewater treatment plant. The developed technique is used to enhance fault detection of the BSM1 system through monitoring some of the key variables involved in this model. The results demonstrate the effectiveness of the proposed PF-based MS-optimized EWMA method over EWMA and Shewhart charts.