Adaptive-high-gain observers with an application to Wastewater Treatment plants

In the context of state estimation, most of bioprocessing systems show in general three key features: 1. The presence of a few reliable on-line measurements 2. These measurements are moreover very noisy 3. Due to the natural environment they are subject to large disturbances (meteorological conditions in our case). This paper applied to a small single basin Wastewater Treatment Plants (WWTPs) the adaptive high-gain observer developed in [5]. This observer takes advantages of both the Extended Kalman filter (EKF) in the presence of noisy measurements, and the high-gain Extended Kalman filter (HG-EKF) when facing large magnitude variation in the influent concentrations. Simulations were carried out with the Activated Sludge Model No.1 (ASM1). Estimations were performed with 5-dimensional dynamical model based on ASM1. Estimation results of the three observers: EKF, HG-EKF and the adaptive High-gain observer were compared with the obtained simulation data from the ASM1 model to highlight the efficiency of the adaptive observer.