Intelligent Prediction Of Total Phosphorus In Effluent Based On Fuzzy Neural Network

The phosphorus is one of the important pollutants in water bodies.It can be used as an important indicator to evaluate the quality of water bodies.The excessive content of phosphorus in water will lead to the eutrophication of water bodies and the rapid propagation of algae and other plankton.It can also result in the reduction of dissolved oxygen in water bodiesand make the water quality worse.An important measure to control the eutrophication of water bodies is to treat the phosphorus-rich wastewater.Moreover,it has to strictly prevent the discharge of total phosphorus(TP)from the effluent during the wastewater treatment process.Therefore,in order to reduce the influence bring from the excessive total phosphorus,monitoring the effluent TP in the wastewater treatment is important.It is difficult to achieve online detection of the effluent TP due to the cumbersome operation,low measurement accuracy,long lag time and high maintenance cost of the instrument in the wastewater treatment plant.An effective online detection method,data-driven method has been widely used to predict the effluent TP and other water quality.However,it may affect the prediction effect of the data-driven method due to the detection data,which are lost,noisy,and have different temporal and spatial scales in the wastewater treatment process.Therefore,in order to utilize the data of different time scales and different spatial scales effectively to complete the real-time detection of effluent TP in wastewater treatment process,a spatio-temporal data fusion model based on fuzzy neural network(FNN)is proposed.The model model can solve the problem of data fusion in different spatial and temporal scales.In addition,the on-line prediction of effluent TP is achieve based on this method.The research work in this study is mainly divided into the following points:(1)A spatial-temporal data fusion model based on FNN is proposed: In order to solve the problem that the date during the wastewater treatment process can't be utilized effectively because of the multi-temporal scale characteristics,a spatial-temporal data fusion model based on FNN is proposed.Firstly,the model can express the mapping relationship between the target domain and the source domain that the data from different time scales and spatial scales.Secondly,a spatio-temporal data fusion transfer learning algorithm is proposed,which can extract the data information of different source domains and realize data fusion at different spatio-temporal scales.Finally,the proposed data fusion method is verified by typical experiments.The results show that the proposed method can achieve multi-temporal data fusion and have better performance.(2)An intelligent prediction model of effluent TP based on FNN is proposed: In order to predict the TP of effluent in wastewater treatment process in real time and accurately.Firstly,Analyzing the reaction mechanism of effluent TP to determine its related variables.Then,principal component analysis(PCA)was used to select the variables related to effluent TP as input and effluent TP as output.Then,an intelligent prediction model of effluent TP based on FNN was established.Finally,the validity ofthe model is validated by using data samples collected from actual wastewater treatment process.The experimental results show that the intelligent prediction model can the intelligent prediction model of effluent TP is effective and can predict the effluent TP with high accuracy.(3)An intelligent prediction system for TP of effluent is developed: It mainly including user management module,user registration and login module,total phosphorus intelligent prediction system homepage module,intelligent prediction method module,Spatio-temporal data fusion model based on FNN module and second-order anaerobic-anoxic-aerobic(A2/O)process module module.Firstly,completing the system login and registration interface design by using Visual Studio2010 software and the database of SQL Server 2008 in the process of developing and implementing the system.Secondly,designing the space-time data model module based on FNN by combining the mixed programming technology of C# language and MATLAB,Finally,each module is integrated through information transmission,so that it can achieve the purpose of visualization and complete the development of intelligent prediction system for TP of effluent.