A BP Forecast model for Soil water infiltration Parameters Based on Philip Infiltration Model

Based on the mass measured data of soil infiltration on Loess Plateau,this paper establishes a forecast model based on Philip soil infiltration parameters through using BP neural network.Moreover,the single and integrated errors for 90 min cumulative infiltration volume prediction of the BP neural network model for soil water infiltration steady infiltration rate prediction model and Philip model are respectively discussed.The results show that it is feasible to establish the BP neural network model based on conventional soil physical and chemical parameters,including soil moisture content,volume-weight,clay content,silt particle content and organic matter,and so on,to predict the two parameters,namely soil sorptivety S and steady infiltration rate A,of soil infiltration half experience and half theory model.The relative error between the predicted value and actual value about the parameters and parameters S are 2.174%and 3.080%,respectively.Meanwhile the relative error between the predicted value and actual value about infiltration capacity of 90 min cumulative of Philip model is 2.038%.They are all within the acceptable range.The research results can provide a strong support for practical surface irrigation technical parameters optimization in the worldwide.