Control Of The Aeration Process Based On Simplified Models

The activated sludge method used in wastewater treatment processes (WWTPs) is an effective approach of removing organic pollutants in wastewater. It is a main way in the current industrial and urban wastewater treatment processes. Under the conditions of increasingly stringent discharge requirements, the energy-efficient operation of WWTPs faces enormous challenges. As the aeration process is the largest part of energy consumption in WWTPs, therefore, the fine control of the aeration process has a significance impact on the energy-efficient operation of WWTPs.In this paper, on the basis of dissolved oxygen control research development in aeration process, simulations of the traditional control of dissolved oxygen, aeration optimal control strategy, model reduction of the aeration process and dissolved oxygen model predictive control based on the reduced-order model were studied on a benchmark simulation platform (Benchmark Model NO.1, BSM1) of WWTPs, in order to save energy of the aeration process. The main content of this paper include:(1) A simplified model associated with the consumption of the dissolved oxygen was firstly proposed for aeration process. The aeration optimal control based on reduced-order state observer was proposed by constructing a reduced-order state observer which was used to estimate online unpredictable states applying predictable states, based on the simplified model. Compared with PID control strategy of dissolved oxygen recommended by BSM1, aeration energy was saved while the effluent ammonia and chemical oxygen demands were maintained by applying optimal control. The proposed control strategy also improved the effluent quality index.(2) The reaction process is complex, and many parameters are difficult in identification in activated sludge model (ASM1). The aeration process model’s order is high and it is difficult to realize the design of model-based optimization and control. In this paper, for optimization and control, model reduction of the aeration process was studied.Linear state space model of the aeration process was obtained by using the least squares. Comparison between the non-linear model of the ASM1and the state space model characteristics the accuracy of the model. Model reduction of the aeration process was realized by applying balanced reduction method. The number of reduced-order model components and parameters are less than that in ASM1, and simulation shows that the dynamic performance of the aeration process based on the reduced-order model and ASM1is almost the same, which proves the feasibility of the model reduction algorithm.(3) Dissolved oxygen model predictive controller based on the reduced-order model of aeration process was designed. The simulation results show that, dissolved oxygen model predictive controller based on reduced-order model is able to overcome influent disturbance and model mismatch. Compared to PID control, the proposed control strategy reduces the changes of the concentration of dissolved oxygen in the aeration system, to some extent, improves the effluent quality.