Analysis of hydraulic pressure fluctuation in deep geologic formations in Tono area, Japan using artificial neural networks

In this study, the concept and possible uses of artificial neural network (ANN) for mapping hydraulic pressures within boreholes are presented. ANN models were developed using monitored hydraulic pressure data over a certain period representing near-natural conditions to map the fluctuating patterns of hydraulic pressure in deep geologic formations in the Tono area, Japan. The ANN models developed captured the natural effects due to barometric pressure, earth tides, and other factors. It was also observed that geologic structures such as the Tsukiyoshi Fault divided the groundwater flow domains in the area. The formulated models that satisfactorily captured those natural effects were then used to evaluate the effect of pumping over a different period.