Prediction model of soil water infiltration under moistube irrigation

As a new type of underground water-saving irrigation method, moistube irrigation has great potential for promotion and development. However, the multi-factor interaction influence and water infiltration prediction model of soil water infiltration under moistube irrigation need to be improved. This paper aims to obtain soil water infiltration data for different bulk densities (1.2, 1.3 and 1.4 g/cm³) and textures (loam, sandy loam and clay loam) under different pressure heads (1, 1.5 and 2 m) through an indoor moistube irrigation soil water infiltration test. By analysing multiple factors affecting soil water infiltration under moistube irrigation, the calculation method was determined with the initial soil moisture content, pressure head, bulk density and texture as the input variables and the Kostiakov model parameters as the output variables. Finally, the method of optimizing the back propagation (BP) neural network by a genetic algorithm (GA) was used to establish the Kostiakov prediction model of soil moisture in moistube irrigation, and the model was verified with analytics. The results showed that the Kostiakov prediction model based on GA-BP had high accuracy and good consistency. The research results provide more practical proofs and perfect theoretical supplements for the study of soil water infiltration under moistube irrigation.