Evaluation And Countermeasure Research Of Black And Odorous Water In River Based On BP Neural Network

Because of the accelerated urbanization process,industrial development makes most urban rivers seriously polluted.According to statistics,about 80% of urban rivers in China have black and odor problems in varying degrees.With the development of economy and the increase of population,the discharge pressure of industrial wastewater and domestic wastewater also increases rapidly.However,due to the imperfection of river embankment construction,the rivers that bear the discharge of wastewater are seriously polluted by non-point sources,resulting in poor fluidity of water body and intensified eutrophication.Urban rivers lose their landscape functions and have a very bad negative impact on the surrounding living environment.In recent years,the number of enterprises stationed in a city's Development Zone has increased,which has increased the discharge pressure of rivers.This not only aggravates the pollution level in the middle and lower reaches of the urban area,but also restricts the green sustainable development of the city.With the development of information engineering system for non-linear water quality assessment,it is a scientific and feasible method to construct the evaluation model of black and odorous water body by using artificial neural network.Taking a city river as the research background,based on the full investigation of natural environment and social economy in the study area,combined with the annual routine monitoring data of the river,the main pollution factors of the river and the water quality categories of the river are analyzed and determined.BP neural network is established by using MATLAB software.The evaluation model of black and odorous water body based on the network is validated by the measured data.The output of the model shows the actual situation of black and odorous water body in the study area intuitively.According to the principle of "one river,one policy" for the treatment of black and odorous water body,and combined with the results of black and odorous assessment,the pollution control countermeasures for each river were formulated.The main contents and achievements of this paper are as follows:(1)Analysis of main water quality parameters of rivers: monitoring section covers the whole development zone along the inland river system,and the distribution of upper,middle and lower reaches is balanced.Monitoring points are set up in Ligong River,Pengbai River,Xiebai River,Yubai River and Huangbai River,totaling 12 rivers.The single factor index and pollution load rate of eight conventional water quality indexes,DO,COD,BOD5,ammonia nitrogen,total phosphorus,fluoride,sulphide and cyanide,were calculated.Five factors which have great influence on the river were determined,namely DO,COD,BOD5,ammonia nitrogen and total phosphorus.Through the calculation of comprehensive water quality index method,the research factors and river water quality categories are obtained.The calculation results show that the water quality of the studied rivers is in V standard or inferior V standard.It is necessary to further classify the river water quality into black and odor grades.(2)Establishment and validation of BP neural network water quality evaluation model: Based on the current monitoring data and field investigation,the classification standard of black and odorous water body is established.The BP neural network is established by using the function of MATLAB toolbox.The number of nodes in input layer,hidden layer and output layer is determined according to the evaluation needs,so that the output range of the model is in the I-V level.Based on the established classification standard of black and odorous water body,the training samples of the network are expanded.One part is taken as the learning samples of the network,the other part is taken as the testing samples of the network.After continuous debugging and self-learning of the network,a mature learning network is obtained,and the test results meet the actual requirements.(3)Application of BP neural network water quality evaluation model: When the network model is mature,the known measured data are substituted into the network,and the evaluation results of 12 monitoring sections are obtained.According to the output of the model and the field investigation,the annual proportion of each river's black-odor grade and the annual variation of the black-odor grade are analyzed,and the corresponding causes are given.After mastering the pollution sources and general situation of the rivers,according to the principle of "one river,one policy" for the treatment of black odor,the corresponding countermeasures for each river are given.