Groundwater Flow Calculation Model and Neural Network Prediction Based on Groundwater Modeling System

Abstract

Due to its merits of good water quality, pollution-free, and wide range, groundwater has become an important source for the water supply of industry, agriculture and urban domestic living in China. The change trends and laws of groundwater level and flow are important basis for the scientific planning and reasonable management of groundwater. In order to obtain a more accurate 3D groundwater flow model, first, this paper constructed a Groundwater Flow Calculation (GFC) model based on the Groundwater Modeling System (GMS), and analyzed the periodic groundwater recharge, the tridimensional flow space, the complex streamline combination, and the concurrent flow direction; then, the paper gave the preprocessing process of the calculated parameter data of groundwater flow, and built a grayscale-embedded BP neural network to predict the groundwater flow. At last, the paper verified the effectiveness of the model with experimental results. This study provided a reference for the flow prediction of other similar fields and the research on groundwater evaluation.