Research On Intelligent Optimal Control System For Deep Nitrogen Removal In Reclaimed Water Plant

In recent years,the country has put forward higher requirements for the total nitrogen discharge limit of wastewater treatment,and it is difficult to balance the operation cost of reclaimed water plants with the problem of exceeding the effluent quality standard.Therefore,the study of intelligent and optimal control theory for deep nitrogen removal in reclaimed water plants has extremely far-reaching implications.The paper adopts the activated sludge biochemical treatment process as the research object,fully considering the characteristics of non-linear,strong coupling and large lag of wastewater treatment in reclaimed water plants,and realizes the research of intelligent optimization control system of modeling-optimization-tracking control for the whole process.(a)Research on feedforward neural network water quality prediction method based on inflection-free adaptive learning rateThe process of wastewater treatment in reclaimed water plants is complex,with many factors affecting denitrification and strong coupling links between variables,and the analysis of the internal mechanism of the process is not yet clear,so it is difficult to find analytical expressions or mechanistic models describing the reaction process from multidimensional and massive data.In this paper,based on feedforward neural network,the adaptive learning rate algorithm is used to make the weights at the inflection points of the neural network pass quickly and solve the problem that the neural network is easy to fall into the local optimum and the convergence rate is slow.The results show that the model effectively avoids the local optimum and the convergence efficiency is significantly improved.(b)Research on multi-objective optimization problem based on improved non-dominated ranking genetic algorithmThe wastewater treatment system has time-varying and non-linear characteristics,and has two conflicting optimization objectives of reducing energy consumption and improving water quality,which is a typical multi-objective optimization problem.In this paper,we analyze the non-dominated ranking genetic algorithm in the evolutionary process,design a custom dynamic adjustment mechanism of variation rate and crossover rate,introduce an individual information exchange mechanism,increase the equilibrium candidate particles,determine the comprehensive satisfaction index based on Topsis algorithm instead of congestion distance,and propose an improved IMNSGA2 algorithm to balance the local search ability and global exploration ability of the algorithm.The results show that the IMNSGA2 algorithm can obtain better population diversity and equilibrium for multi-objective optimization problems.(c)Research on intelligent decision control for wastewater treatmentThe objective of energy saving and consumption reduction is to suppress the concentration of ammonia and total nitrogen from exceeding the standard.This paper first uses NI-BP algorithm to achieve the prediction of ammonia nitrogen and total nitrogen in the effluent;combined with the prediction model using the improved IMNSGA2 multi-objective optimization algorithm to optimize the setting values of dissolved oxygen and nitrate nitrogen concentration in the second partition and the fifth partition;using the proportional integral controller to track and control the optimized setting values,and according to the prediction model prediction,over control the water quality that will exceed the standard.The results show that the intelligent decision control has perfect control effect in suppressing the exceeded water quality and achieves the goal of energy saving and consumption reduction compared with other suppression algorithms.