Study On Intellectualized Irrigation System Of Urban Green Land

Water is the source of life, and human essential basis of survival and social development. Water shortages, and the inadequate supply water, have become a global problem.Partcularly China's water situation is severe, less than 1/4 of the per capita water resources in the world, lines in No.109 in the world, are in more than 400 cities (aucncy 670 cities) in varying degrees of water shortage, including 108 serious water shortages. Green system is not only to beautify the city, to improve the urban quality, or to improve the urban environment and so on, but also to offer an open place for residents. With the improvement of living standards, the Government is more and more emphasised on the urban green construction, urban green coverage rate will be increased in the future substantially, and urban green areas will be rapidly increased. However, this increase of green areas will inevitably lead to the increasing of the amount of green irrigation water substantially. For the development of irrigation adapt to the needs of urban development, it is the most effective way to advance water-saving irrigation. Precise intelligent precision irrigation would replace the current widespread extensive irrigation.This thesis includs two parts. Firstly, the author established a GA-BP network model to predict of evapotranspiration ETo reference crop, and to analysis prediction results. Then, the formula was utilized to caculate the moist layer of soil moisture combined with canopy temperature to establish the fuzzy decision system for irrigation and date meeting appropriate crop irrigation purposes timely.Establishment the network prediction model,it needs to determine the network structure of BP neural network, the author conducted random combination meteorological factors that influences the evapotranspiration reference crop as the network's input, and used the reference crop evapotranspiration ETO as output, and determined the hidden nodes on the basis of principles of gold-segmentation feedforward neural network hidden nodes of optimization algorithm.Network structure was determined by using genetic algorithms (GA) optimization of BP neural network connection weights, and used the meteorological data from Beijing(year 2001,2002) and PM ETO to train network model, established GA-BP network prediction model. Using the trained GA-BP model to predicte 2003 the reference transpiration for Beijing, and then the prediction rusults were compared with the PM ETO. The analysis shows that the GA-BP prediction model would quickly obtain a high precision reference transpiration prediction by inputing the four factors built, day ordinal, mean air temperature, wind speed and sunshine duration. If weather datas are incomplete, this study would be some guiding significanct to guid the prediction reference evapotranspiration.In this paper, a fuzzy decision system was set up, wet floor plans average water content and canopy temperature as input data, irrigation coefficient as output data, conducted fuzzy decision-making on irrigation coefficient, author introducted restricted coefficients(α1,α2)to prevent rain and irrigation water over hours in the next day, used the formula of irrigation to predict the amount and, forecast date of the irrigation was predicted on the basis of the soil moisture content, using this system to predict irrigation amount and date would keep the soil moisture content in the range of crops healthy growth.