Application Of Data Mining Technology In DMA Distribution Losses System

Water is the source of life. As the lifeblood of city, urban water supply pipe network has a very important mission to bring the water to every corner of the city. However, along with the rapid development of social economy, water supply pipe network distribution losses delay less than expected. On the one hand, people has poor awareness of saving water; on the other hand, urban water supply pipe network exists a lot of leakage.At the same time, there are all kinds of water leakage phenomenon. It is not only a waste of valuable water resources, but also a waste of manpower and material resources. Therefore, as a major event related to the people's livelihood,reducing the pipe network distribution losses is a urgent problems to be solved in water supply industry.Water supply pipe network partition management technology (DMA) for leakage control and reduce water distribution losses is of great significance, at present DMA technology has been recognized by water supply industry as a effective pipeline leakage control means.DMA technology has been widely used in major cities at home and abroad, companies and tap water can collecte a large number of water data of users. Based on a large number of real water data, mining the rules and laws of using water of people from the surface of water by statistical techniques and data mining technology for processing and mining, generating rarious water report and push it to the leadership of the enterprise so as to provide auxiliary decision support.By analysing and mining the rules and laws of using water of people in each independent district using minimum flow analysis, regression analysis and cluster analysis methods based on the water data of people in part of the small independent metering district of Fuzhou City, thus provides a new perspective to solve the marketing problem.The main contents and contributions of this paper are as follows:(1)Summary and analysis of current situation and existing problems: This paper introduces the background and significance of the subject, and describes the development status and existing problems of DMA technology at home and abroad, focusing on the regression analysis in data mining technology, clustering analysis method.(2)The accurate calculation of minimum night flow:The error for the calculation of the traditional minimum night flow, related research literature, the calculation method of normal distribution theory based on minimum night flow, the more accurate calculating minimum night flow, can help judge leakage partition residential pipeline measurement.(3)Water consumption prediction:According to the water company need for water consumption prediction and analysis of data of historical water found with the general trend of the water changes, using regression model to carry on the analysis, prediction model was established to predict the area of the short-term water consumption.(4)Cluster analysis and evaluation of water behavior in different regions:On the independent district metering area, from the water consumption, production and marketing difference, minimum flow and other dimensions of the cluster analysis, found that the law of water use, and objective assessment of the water.(5)Design and implementation of DMA system:According to the characteristics of partition measurement, design and deployment of DMA production and marketing system, including the system architecture, the overall functional design, database design, the realization of the main function module.