Irrigation Forecasting Model Of Irrigation Region Based On Artificial Neural Network

High level's management of irrigation water is the important tache of saving-water, advancing agricultural outputs of irrigation region and exerting folly the benefit of irrigation project. Implementing planed-using-water is the core of irrigation water management, the plan of using-water is the foundation of instructing planed-using-water. The former plan of using-water of irrigation region, which is obtained by historical datum, is different with the actual status, and is difficult to effectively instruct how to use water. The dynamic plan of using-water bases on the real-time irrigation forecasting, and includes dynamic getting-water, dynamic distributing-water and dynamic irrigating-water. It is short-term plan that can make the best of real-time information. So the dynamic plan of using-water comparatively accords with actual facts, and it's applied value is high.The model of optimal irrigation schedule, which bases on the water production function and adopts the means of dynamic programming, can solve the problem how to reasonably distribute the water supply to all development stage, and can make the limited water resource create maximal value; the model of real-time irrigation forecasting can effectively utilize the weather information of the day, and can make irrigation forecasting in good time. Synthetically using the two-model to instruct the irrigation of irrigation region can achieve the aim of saving-water and high yield, and has important realistic meaning.The text gains the sensitivity index of all development stage according to the testing datum, gains the sensitivity index of every ten days using the measure of Artificial Neural Network, establishes the water production function taking ten days as a phase; using the historical rainfall datum, the text establishes the grey Markov model, which can forecasts the rainfall of next year; based on the former water production function and the rainfall-forecasting model, the text uses the measure of dynamic programming to establish optimal irrigation schedule model that can forecast the plan of using-water when the irrigation region is short of water, and compiles friendly interface between the person and the computer; the text takes the datum of crop water requirements and it's influence factor of ZaoHe irrigation region about 17 years as educating-datum, establishes the neural network model of crop water requirements, the model forecasts crop water requirements using the short-term weather forecast information; the text establishes the real-time irrigation forecast model using the water balance equation. The text takes the ZaoHe irrigation region as example, checks up the models, the effect is ideal.