Prediction of pore pressure change due to the shaft excavations using artificial intelligent techniques

Takeuchi, Shinji; Chinone, Tatsuya*; Mebruck, N.*; Watanabe, Kunio*

Tunnel or shaft excavation in fractured rock mass brings out the pore pressure change around the site. The effect of the construction is spatially different due to the existence of high permeable fractures around the site. The precise prediction with real time of pore pressure change is very important to maintain the groundwater in good condition. In this paper, pore pressure measured by 50 sensors which was installed in 5 boreholes around two shafts that are now under construction for Mizunami underground research laboratory project of JAEA, was analysed by using Genetic Algorithm (GA) and Artificial Neural Network (ANN). It was found that the pore pressure change in this area could be divided into four patterns. The pore pressure fluctuation at an arbitrary point was precisely predicted on real time with pore pressure data obtained at other sections when the fluctuation patterns of those data are identical.