Research On Intelligent Prediction And Robust Optimal Control Of Printing And Dyeing Wastewater Treatment System

The improvement of China's national economy in recent years has promoted the development of the printing and dyeing industry.However,incomplete wastewater treatment and slow treatment speed are the key factors leading to environmental pollution.There is indicator data measurement accuracy in the wastewater treatment process.Problems such as low speed and slow speed can easily cause secondary pollution of the water environment.In addition,the internal organization of the wastewater treatment plant is complex and difficult to achieve efficient control.This paper aims to design an intelligent prediction and robust optimization control system for printing and dyeing wastewater treatment to achieve real-time prediction and control optimization of wastewater indicators,and on this basis to develop a central control system applied to wastewater treatment plants to facilitate real-time monitoring and control.This article first introduces the BSM1 model(International Benchmark Simulation Model No.1)proposed by the International Water Quality Association,including the overall framework,key parameters,structure and characteristics of each reaction tank,etc.,and provides simulation experiments for the study of water quality control strategies platform.Secondly,in view of the difficulty of measuring wastewater indicators and the slow speed,this article proposes two intelligent prediction models for wastewater outlet water quality:(1)Based on the improved GWO-SVR prediction model: First,the probabilistic principal component analysis PPCA algorithm is applied to the wastewater treatment tank inlet and outlet water quality index data for feature extraction and dimensionality reduction,and then use the improved gray wolf optimization algorithm GWO to optimize SVR parameters,and establish a water quality index BOD prediction model;(2)CNN-SVR-based prediction model: First,the input easy-to-measure data is constructed in order in the form of a window as the input of the model,secondly,use CNN to extract the feature vector,construct the obtained feature vector in sequence and use it as the input data of SVR,and finally use SVR to predict the index.This article uses the actual wastewater treatment plant process data in the UCI database to perform MATLAB simulation experiments.The results show that the measurement accuracy and speed of the model are high and the model is feasibility.Next,a robust control model of water quality in the wastewater treatment process is proposed: First,a BSM1 model is built on the MATLAB platform,and control simulations under three different weather conditions are carried out.On the basis of the original control,the intelligent algorithm model proposed above is added to optimize control parameters.Through simulation and comparison experiments,the results show that the proposed optimal control model has better stability and robustness.Finally,this article uses tools such as visual studio,combined with frameworks such as ASP.NET MVC,to design a wastewater treatment water quality monitoring and robust control system.The system is divided into two parts:(1)Front-end water treatment automation monitoring system: Realize the functions of user login authentication,real-time display of various data indicators,and historical data curve display;(2)Back-end wastewater treatment central control system: Realize the display of the overall layout,real-time display of key data indicators,manual and automatic control of each control device,display of device parameters,etc.The test results of the system show that the system can meet the needs of the factory to effectively monitor and control,and improve the efficiency of the wastewater treatment plant.