Evaluation of the Genericity of an Adaptive Optimal Control Approach to Optimize Membrane Filtration Systems

This study explores the application and robustness of an adaptive optimal control (AOC) strategy to optimize the operation of membrane filtration systems. The proposed control is based on a constant flux model where fouling is primarily due to cake layer formation. The algorithm dynamically finds the optimal ratio between the filtration (F) and backwash (BW) time ratio in response to system disturbances, thereby adapting the operational state of the membrane in order to optimize its performance in terms of energy consumption. The strategy was successfully applied to both microfiltration (MF) and ultrafiltration (UF) systems and quantitatively demonstrated its effectiveness in reducing energy consumption and controlling fouling. It proved robust against model uncertainties and demonstrated real-time adaptability even under varying and realistic disturbance conditions. The implementation of this control strategy facilitated real-time adaptation of the filtration/backwash (F/BW) ratio in response to dynamic system disturbances. The result underlines that the control behavior is predominantly driven by fluctuations in mixed liquor suspended solids (MLSSs). Compared to conventional fixed-time modes, the AOC led to significant energy savings, ranging from 7% to 30%, and membrane lifespan extension, mainly through more efficient permeate pump usage.