基于SOM算法的城市给水管网水质评价市政工程专业论文.docxVIP

Abstract Water distribution network as a ‘carrier’ to provide users with excellent quality of water, it finishes the important task of transporting water which corresponds to the national water quality to various types of users. In the transmission process, the physical, chemical and biological act happened leads to water quality indicators change a lot and not correspond to the standard. The analysis of water quality in water distribution network is an important step. In this thesis, significant water quality indicators reflected the actual status in the water distribution network had selected. Its changes in laws and predictive models discussed. Because domestic research on the above indicators was not yet mature enough, quoted internationally authorized the three drinking water standards for reference, to grade water quality using residual chlorine, trihalomethanes, water-age as indicators in Harbin, respectively. Due to single indicator grade limited, the study attempts to build Self- Organizing Feature Map, as SOM, clustering evaluation on water quality. Human intervention to cluster evaluation is less than the previous gradation evaluation, so it can get more objective results. Actual observation through the operation of water distribution network with computer aided, inputting water quality value 24 h, compiled the cluster evaluation process using MATLAB. With examples of the network to trial, the water quality of nodes was divided into 8 categories and pipes was divided into 7 categories. From Class 1 to Class 7 or 8, the water quality gradually deteriorated. This cluster method was applied in the distribution network in Harbin, and the water quality of nodes and pipes was divided into 8 categories, respectively. The quality belonged to 8 type was worst, unable to meet water quality standards. Compared with the single-evaluation indicators, SOM method can take into account water quality monitoring data of single indictor and the overall data to analysis, mo