Application of Neural Network Model to Evaluate Hydro-Geological Parameters

The study proposes to apply the artificial intelligence combining neural network control system and groundwater flow control equations to investigate the hydro-geological structure and hydraulic parameters in the area of Chou-Shui alluvial fan. We set up the closed-loop control system to simulate the variations of groundwater level in the such area. The close-loop control system collects the feedback data generated from the neural network model and conveys them to the sensor plant (groundwater flow control equation) in the control system. It is found that when the control system responses, the minimum error (within 5.1 × 10^-4 to 9.2 × 10^-3 in test and predict stage) of both the neural network model and the sensor plant is occur the results also indicate optimum hydraulic parameters that the innovative method works better than just applied neural network model and MODFLOW software to simulate the groundwater flow.