Numerical simulation and experimental validation of soil moisture infiltration patterns under underground porous membranes

In agricultural irrigation engineering, deep leakage is a key factor that significantly reduces the utilization efficiency of irrigation water. Underground installation of porous membranes, as a novel active regulation technology, can effectively reduce deep leakage losses of water in the soil through its physical barrier effect. However, the current understanding of the infiltration patterns of underground porous membranes remains inadequate, limiting the promotion and application of this technology. Therefore, this study integrates a methodology that combines numerical simulations with experimental validations. Using a non-membrane treatment as a control (CK), this study investigated the soil water infiltration of underground porous membranes under various combinations of saturated hydraulic conductivity (K_s), porous membrane diameter (D), burial depth (H), and spacing (S). The results indicated that under the four types of aeolian sandy soil conditions, underground installation of porous membranes had a significant impact on soil infiltration characteristics, exhibiting an infiltration-reducing effect. Upon entering the steady infiltration stage, the minimum reduction in the infiltration rate for the various porous membrane treatments was 2.86 times that of the CK treatment. At a specific irrigation time (t), the steady infiltration rate (i_f) and cumulative infiltration (I) of soil increased with increasing K_s, D, H, and S. There was a strong power function relationship between i_f and the four factors (R²=0.997), with a coefficient of 0.209, and exponents of 1.14, 1.04, 0.48, and 0.30, respectively. Furthermore, based on the Kostiakov infiltration model and comprehensively considering K_s, D, H, S, and t, an estimation model for cumulative infiltration of underground porous membranes was developed. The reliability of the estimation model was assessed using experimental data, with the root mean square error approaching 0 and the Nash-Sutcliffe efficiency coefficient close to 1, indicating the good predictive performance of the model. The findings of this study can provide a scientific basis for the operation and management of underground porous membrane irrigation projects.