Modeling And Control Strategy Of Dissolved Oxygen Concentration In Sewage Treatment Process

With the advancement of industrialization and urbanization in our country,the situation of water resources in our country is more severe,showing the characteristics of water shortage,rising water consumption and serious water pollution.Water pollution also affects people's drinking water safety and human health.Sewage treatment is imminent.In the process of wastewater treatment,the concentration of dissolved oxygen in the aeration tank is a very important control parameter,high and low will have a great impact on the effect of sewage treatment.It can intuitively and quickly reflect the operation of the whole system,and also directly affect the effect of sewage treatment and energy consumption of the system.In this paper,the characteristics of influent flow fluctuation,water quality variation,non-linearity,strong coupling,large lag and high precision of DO concentration control in activated sludge process are studied.An improved differential evolution based BP fuzzy neural network control method is proposed to control DO concentration in activated sludge process with variable parameters.In AS,the DO concentration of activated sludge process wastewater treatment system is controlled.Based on the simplified M1 model,the simulation results show that the control method can adjust the membership function online,optimize the control rules,and apply it to the activated sludge sewage treatment system,which can quickly and accurately meet the requirements of dissolved oxygen and has strong robustness.Based on the above,the main research contents of this paper are as follows:Firstly on the basis of in-depth analysis of the mechanism of activated sludge sewage treatment,the process model is deduced based on ASM1 model,and the system model is deduced on the basis of reasonable simplification of the process;the controllability and stability of the system model are analyzed;on this basis,the state feedback gain matrix is obtained by Ackerman calculation method,which improves the stability of the system.Continuous research provides a stable and reasonable system model.Secondly the fuzzy neural network(FNN),BP algorithm,difference algorithm and improved difference algorithm are summarized.Aiming at the fact that BP algorithm commonly used in fuzzy neural network is easy to fall into local extremum,a hybrid algorithm of improved differential evolution(IDE)and BP is proposed.The operation andadvantages of the algorithm are analyzed and applied to the parameter optimization of fuzzy neural network.Based on the test function,two algorithms are tested.The results show that the convergence speed of IDEBP algorithm is obviously faster than that of DE algorithm,and its generalization ability is also improved.It is superior to DE algorithm and its convergence accuracy is higher than DE algorithm.IDEBP algorithm is applied to the simulation of second-order lag system to prove the effectiveness of the fuzzy neural network based on IDEBP algorithm.It also proves that the optimization performance of IDEBP algorithm is better than that of BP algorithm.Lastly according to the above method,the DO concentration controller is designed.Aiming at the four parts of fuzzy control,a four-layer fuzzy neural network based on the simplification of standard neural network model is designed,and a set of optimal or optimal weights are obtained by using IDE off-line training BP.It can provide a better initial value for online learning of fuzzy neural network.The variable rate BP algorithm with momentum term is added as the online learning algorithm of fuzzy neural network controller,and the online learning algorithm is built.The system simulation platform controls DO concentration.Through comparative analysis of simulation experiments,it is concluded that the performance of the designed controller can meet the actual control requirements,and the robustness and adaptive ability of the controller are better verified on the international benchmark simulation platform.