Study On Water Quality Verifying And Monitoring Model Of Urban Water Supply Networks

Water quality has met the drinking water health standards after it left the treatment plant. However, through complex urban drinking water distribution networks,it might deteriorate seariously when reached the users.In order to improve water quality drinking of water distribution networks, to strengthen the running management, it was necessary to study its change in the networks.Establishing water quality model of distribution networks, not only could forecast the water quality, but also provided reliable the oretical basis for the running management of water quality, all of these had vital significance for the water quality security.This paper makes a study of nine indicators of water quality monitoring spots within a network endings in Suzhou, and combining with actual situationtests the ex perimental data with four kinds of the statistical test model, involved 4D,Dixon,Q and Grubbs test model. After analyzing the test results, it could make a scientific trade-off judgment for the prediction of water quality model and water quality,This paper presented the application of two empirical models for simulating and forecasting turbidity within drinking water distribution systems. The first was a multivariate linear regressive model with data analysis tool SPSS; the second was an artificial neural network model With numerical simulation tool MATLAB. The paper had established water quality models with No.1 and No.9 monitoring points, and forecasted the turbidity for these two monitoring points. The results demonstrated the artificial neural network model had better, the former was more accurate than the latter, and the artificial neural network mode could be used to the prediction of practical water supply system. The predict effect shew that it made forecast effect better after the data screening of bizarre datas.Finally, combining with the researched datas of the actual situation, draw a conclusion that establishing the water quality model must be based on a large amount of data, that’s the premise to ensure that accuracy and precision of the model; And then thought it’s necessary to inspect abnormal datas from the group Before building the water quality model. At the end of the paper, it not only analysed the reasons of forecast errors,but also made development prospects for the research of water quality models.