The Construction And Application Study Of Water Quality Of Moshui River Based On Bp Neural Network

Water quality assessment and simulation prediction of river were foundation jobs for water environmental management. And it was the basement for solving serious problems of water environment. Water quality assessment can make scientific evaluation for water environmental quality. It also can supply a scientific base on environmental standards establishment, environmental planning and water pollution control. According to current water environment quality situation, contaminant transportation and transformation of the area and discharge situation of wastewater, water quality simulation prediction can forecast the changing tendency of the water quality. As the basement of water pollution control, both of them provided evidence for studying water quality development regular, diagnosing the problems of water pollution and estimating carrying capacity of water environment.On the base of pre-study, water quality model was constructed in this paper. In project of South-North Water Transfer Line Nansi Lake Comprehensive Improvement of Ecological Restoration and Technology Development of National Eleventh Five-Year Key Project of Science and Technology (Item Number: 2006BAC10B03), by using Matlab software, remote sensing software and Surfer software, our group made comprehensive assessment, eutrophication assessment and spatial distribution of pollutants for Nansi Lake .All of these made the basement for water quality model of Moshui River. Based on the pre-study, some expanding study was carried out and the original water quality model was optimization, which not only improved the model accuracy, but also enhanced its stability.As the most important river for receiving industrial wastewater in the northern area of Jiaozhou Bay in Qingdao, Shandong province, this paper taked Moshui River as the area for studying. Starting from the viewpoint of system theories and complex theories, combining with water quality situation of northern area of Jiaozhou Bay and programming of new and high-tech industry regions, water quality assessment and simulation prediction model had been established by using the BP neural network in Matlab software. This paper made synthesis assessment according to the monitoring data of quality from 2001 to 2007 of the river, and the temporal-spatial changes and main pollutant consistence of the estuary had been predicted. According to the results of water quality assessment and simulation prediction, reasons for pollution of water quality had been found and water environmental capacity had been estimated. Based on the study of pollution control and treatment advanced technologies of river inside and outside, some advices had been given for the study area, and provided the basis for water environment system sustainable development. The main contents of this research were as follows:1) Water environment quality status and socio-economic situation of Moshui River had been studied. By using Matlab and Surfer8.0 software, temporal and spatial variation of major pollutants trends and the spatial distribution of major pollutants map of estuary in Jiaozhou Bay had been drawed. That intuitively described water environment quality status and the spatial distribution of pollutants fo the area.2) Water quality assessment model was an uncertainties theory model, whichwas based on BP neural network algorithm in smart pattern recognition . The model included the input layer, hidden layer and output layer. In order to enhance adapt and improve network accuracy and training efficiency, a series of optimal design of the model had been done. According to results of water quality assessment, the monitoring data of water quality from 2001 to 2007 of Yanxia Bridge, Zhanjia Xicheng and Shi Bridge of Moshui River met the class III, IV and V criteria specified separately in Surface Water Environmental Quality Standards. Water quality of estuary was poor and inferior V criteria that will affect water quality of coastal areas.3) Based on the BP neural network, water quality simulation prediction model of Moshui River was built up. Model of network architecture was divided into model training, model validation and application. As the training samples, water quality monitoring datas from 2001～2006 of the river were brought into the network for training. Then, datas of 2007 were brought for testing the accuracy of the network. Surface Water Environmental Quality Standards were brought into the network for predicting concentration of the major pollutants in the estuary. According to results of water quality simulation prediction, under the circumstances of all functional areas meeting the standard of the standard, water quality of estuary were poor and inferior V criteria. the main reasons were the production wastewater and domestic wastewater directly discharged and the self-purification capacity of water environment substantially reduced. This would result in economic development along the coast to some extent hampered.In general, according to results of water quality assessment and simulation prediction, the water environment quality in the estuary of Moshui River is the important part of water environment pollution control. The contravention between water environment protection and economic development can be solved by using the countermeasure of pollution treatment and control, combining with the economic development of the drainage area.