Dissolved oxygen control of BSM1 benchmark using generalized predictive control

This paper presents a generalized predictive control (GPC) to regulate the dissolved oxygen concentration in an activated sludge process. Firstly, an unconstrained GPC controls the dissolved oxygen concentration in the presence of white measurement noise. However, adjusting to influent changes is more important in practice than rejecting white noise. To use the GPC for disturbance rejection, an inequality constraint on the controlled variable augments the control methodology. This constrained GPC for disturbance rejection is the main contribution of this work. Simulations using the Benchmark Simulation Model 1 (BSM1) demonstrate that the proposed method leads to precise setpoint tracking and disturbance rejection.