Study On Raw Water System Optimization

Water, which is the source of life, is one of the most active and extensive ingredents in the ecological system. In China, per capita water resources is only 2,220 m3/a, it ranks110th place in the world. China is on the list of the world's 12 water poor countries. On the other hand, 80% of waters, 45% of the underground water, and more than 90% of urban water are polluted in our country. Meanwhile, such as Beijiang River in Guangdong Province water pollution incidents occurred in recent years, are not only a serious threat to the urban water supply, but also have a significant impact on people's health and the safety of drinking water. Therefore, a reasonable application of the valuable water resources is very important, and the raw water allocation system is required a higher level.On the basis of briefly exposition about water demand forecast methods'development and analysing water demand forecast methods'classifycation and in influence factors,the paper details water demand forecast methods in common use, and establishes the combined model with optimum weight based on GM (1,1) model and BP neural network model, and researches its application in regional total water demand forecast, the precision test shows that the model is reasonable and feasible.The paper summarizes and analyses the largescale systems theory and the multi-objective question theory, details existing water resources optimization disposition model and water plants layout optimization model aiming at the year operating cost minimum, establishes the raw water system optimization overall model which integrates water resources optimization and water works layout optimization, and adopts the genetic algorithm to solve model.The raw water system planning of Dapeng byland is as an application example aiming at all-around benefit maximum. According to the statistical data of Dapeng byland,the paper with the survey of nature, economy, society and status quo of water resources utilization forecasts the regional total water demand the GM(1,1)-BP neural network model in 2010 (32,150,000m3/a), 2020(68,810,000m3/a), and forecasts various departments'water demand by water used standards based on the population and using ground nature; Definites the research area's water supply source, user constitution and water used relations, optimized goal and constraint condition, model parameter and so on, uses the Matlab optimization toolbox to solve the model; To be the allocations of 2010,2020, the results show that the departments'water demand basically satisfy;Comparison between two water plant layout projects, the paper chooses a better one.