Modeling And Optimal Control Of Wastewater Treatment Process Based On Extreme Learning Machine

Sewage treatment plants generally have the problems of low treatment efficiency and high operation cost.Therefore,this thesis studies an energy-saving and consumption-reducing sewage treatment control system based on the actual project of a sewage treatment plant in Liaoning,aiming to improve the sewage treatment efficiency and reduce the system operation cost.The sewage treatment process has the characteristics of nonlinearity,strong coupling and large time delay.The influence of the change of influent water on the effluent quality will not be shown until a long time later,resulting in the control system not being able to obtain the response of the effluent quality in time,thus leading to untimely control and low treatment efficiency.In order to get the feedback of effluent quality in time,this thesis put forward a prediction model of Chemical Oxygen Demand(COD)of effluent based on Extreme Learning Machine(ELM)with the relevant parameters of influent as auxiliary variables and COD of effluent as leading variables.At the same time,the improved PSO algorithm(IPSO)was used to optimize ELM parameters,finally an IPSO-ELM prediction model for COD of effluent is established.Simulation results show that IPSO-ELM model has better prediction accuracy,convergence speed and anti-interference ability than PSO-ELM and ELM model,and has better prediction performance.Based on the prediction of effluent quality,this thesis proposes an energy-saving optimal control strategy to further reduce the system operation cost.Firstly,the biochemical treatment process of sewage is modeled,dissolved oxygen(DO)concentration and excess sludge discharge(Qn)which have a great impact on effluent quality and operation cost are selected as optimization control variables,effluent quality is taken as constraint condition and operation cost is taken as performance index,and IPSO-ELM model is continuously used to predict effluent quality,and IPSO algorithm is used to optimize DO and Qn on the premise of ensuring effluent quality up to standard.In order to improve the optimization effect,the 1-day time is divided into n segments on average and optimized segment by segment.The simulation results show that the combination of IPSO algorithm and IPSO-ELM prediction model,as well as the subsection optimization method,improves the sewage treatment efficiency and reduces the operation cost.Through simulation and comparison,IPSO algorithm is superior to PSO algorithm in energy saving and optimization performance.Finally,according to the sewage treatment process and control requirements,a set of energy-saving optimal control system is designed with WinCC as the upper computer monitoring software and S7-300 series PLC as the lower computer,combined with IPSO-ELM prediction model and the optimal control strategy based on IPSO algorithm.The optimal control system uses MATLAB to realize the operation of algorithm,realizes the data communication between MATLAB and WinCC through OPC technology,takes the optimization value of MATLAB as the set value of the field controller,and then realizes the energy-saving optimization control of the sewage biochemical treatment process.