Reserch Of Interval Predictor Models For Key Effluent Water Quality Parameter Of Waetewater Treatment

At present,the problem of water environment in China is becoming more and more prominent,and wastewater treatment has become an important measure to protect the environment.It is difficult to realize the real-time monitoring of the key water quality parameters of the effluent.Fortunately,soft sensors with the advantages consist of low cost,timeliness and high accuracy have been introduced into the process of wastewater treatment and provide an effective way for real-time monitoring of effluent parameters.In the light of the inherent complexity of the mechanism of wastewater treatment process providing with high nonlinearity and uncertainty,the result of traditional method building the soft sensor based on the biochemical reaction mechanism is unsatisfactory.So the neural network has been widely used in soft sensor modeling of wastewater treatment by virtue of its excellent approximation ability.This paper presents a new soft-sensing method,which can predict the upper and lower bounds of the effluent during the wastewater treatment system.The main research work of this paper includes the following points:(1)Establishment of soft sensor model for key water quality parameters of wastewater treatment;Firstly,design the structure of soft sensing model according to the general steps,and the initial selection of the secondary variables that related to the estimated variables is done.The resulting variables are the original secondary variable.And then the collected data is preprocessed,including the elimination of abnormal data,data normalization,and reduced dimension of original secondary variables.The principal component analysis method is used to reduce the dimensionality of the secondary variables of the soft measurement model.And they are used as the input of the RBF(Radial Basis Function)neural network and the key water quality parameter to be measured is used as the output of the RBF neural network to establish the soft measurement model.(2)Fundamental calculation of the confidence interval for key water quality parameters of wastewater treatment;Considering the bounded modeling error,linearin-parameters set membership identification algorithm is used to obtain a description of the uncertain set of the output weights after the determination of centers and width of the RBFNN by K-means.Applying this soft measurement model to the wastewater treatment system can predict the upper and lower bounds of the effluent.In addition,in order to reduce the conservativeness of the results obtained by a single model,a multimodel strategy is adopted here.So a plurality of models are established and the measurement results obtained by these models are merged.And then analyze and explain the use of confidence intervals.Finally,experiments were conducted when BOD(Biochemical Oxygen Demand)and TP(Total phosphorus)are the model output parameters.The experimental results show that the method is feasible and effective comparing with the experiment that the neural network weights obtained by least square method(3)The design of interval forecasting system for key water quality parameters of wastewater treatment;The interval forecasting system for key water quality parameters of wastewater treatment is designed in this paper,which mainly includes a user registration module,sample data management module,neural network model selection,neural network model training and prediction module and water quality parameter knowledge.In the process of system design,SQL Server 2008 database is used to store user information and auxiliary variables,and C# and MATLAB mixed programming techniques are used to implement the MATLAB program in the interface,so as to realize the display of key parameters of the effluent and the prediction results.Through the information transmission among various modules such as user management module,data processing,neural network model training and prediction,the output of the predicted value of the key water quality parameters of the water outlet can be realized and displayed,and the purpose of the visual interface of the soft measurement system can be achieved.The system is exemplified by BOD and TP and can be extended to other key water quality parameters,and other functions can be added if we need.