Improving oxygen concentration control in activated sludge process with estimation of respiration and scheduling control

The paper presents an efficient and simple model-based method of on-line estimation of respiration from the current values of dissolved oxygen concentration and airflow signals. Fast estimation of respiration is important because respiration appears as the most important disturbance signal in the control of dissolved oxygen concentration, and could therefore be used for improving control efficiency. A parameter scheduling PI (proportional and integral) control scheme is proposed, where the estimated respiration signal is used for scheduling of the PI controller parameters, by using local linearization of the process dynamics. The performance of the scheduling controller is compared to a conventional PI controller in simulation using the COST simulation benchmark, and experimentally on an activated sludge process pilot plant.