Investigation of the environment around the deep of Shahid Baqdrat underpass in Kerman due to changes in the mechanical properties of soil and water pumping and its prediction in artificial neural network

Due to construction of high-rise buildings and construction of underpass intersections and also due acceleration of construction of high-rise buildings in most metropolitan areas, deep and shallow excavations always bring advantages and disadvantages for engineers. During the construction of underpasses that are located below the groundwater level, no problem can disrupt the project development process as much as the presence of water problem ،due to this problem in the Baghodrat martyr underpass project of Kerman city to reduce groundwater ،different methods were studied. Eventually Valid scientific documents, the use of pumping wells and guidance to the intended location for drainage and stabilization of the well by nailing were confirmed. In this research, first numerical modeling Baghodrat martyr underpass of Kerman city using finite element method using Plaxis2D software and by changing the mechanical properties soil including internal friction angle, adhesion and permeability of soil as well as outlet pumping discharge, the maximum vertical subsidence The edge of the pit was calculated and the necessary analysis was performed in this regard. The results showed that increasing the angle internal friction and adhesion and soil permeability coefficient reduced sedimentation and increasing pumping discharge increased soil sedimentation. Then, using multilayer prosthetic artificial neural network in MATLAB software and Plexis software data as network entrance, the values of ambient subsidence around the pit in other cases have been predicted. Finally, an artificial neural network with a correlation coefficient of 0.982 is obtained, which indicates the accuracy and power of accurate prediction of real values.