Multi-Model Predictive Control Strategies for an Activated Sludge Model

This paper investigates the use of Generalized Predictive Control (GPC) on an Activated Sludge Reactor. The reduced bio-reactor activated sludge ASM1 model, which describes the biological degradation of an activate sludge reactor, is designed based on several simplifications, as a Takagi Sugeno fuzzy model (TS). The TS model structure is based on a set of linear sub models, covering the process input-output space, interpolated by a nonlinear weighting function. In the case of the ASM1 model, as specified in this paper, the linear sub models turn out to be non minimal phase, and therefore the system needs to be decoupled prior to design the control formulation. The classical Multi-Input Multi-Output (MIMO) GPC formulation is then modified to integrate the TS formulation as the controller internal model. The simulation results show the effectiveness of the proposed GPC controller compared to benchmark PID in terms of error and response dynamics.