Fault detection using UKF-based optimized EWMA method in wastewater treatment plant

Abstract

In this work, Unscented Kalman Filter (UKF) based Optimized exponentially weighted moving average (OEWMA) is suggested for fault detection (FD) in a Wastewater Treatment Plant (WWTP). UKF method is suggested, to compute the residual of the true and the estimated data of WWTP, in addition, the Optimized EWMA is utilised to the faults in a simulated WWTP. The FD techniques will be tested using simulated data, which are generated using the simulated COST benchmark BSM1 of wastewater treatment model, provided by the IWA Task Group of Control Strategies. The results of the detection of the UKF-based Optimized EWMA are appraised with two criteria of FD : the missed detection rate (MDR) and the false alarm rate (FAR).