Back analysis of displacement of deep foundation pit based on genetic algorithm

Abstract

At present, the deformation prediction during the construction of foundation pit is still in the exploratory stage. The more common prediction methods are to calculate the deformation based on numerical simulation software, or to find the law between the excavation conditions and the deformation according to the field measured data and use it for prediction of subsequent operating conditions. The deformation calculation accuracy of the numerical simulation software depends on the value of the model parameters, but the geotechnical parameters themselves are difficult to determine, the excavation process of the foundation pit is more complicated, and the excavation process of deep foundation pits is more complicated.The difficulty in determining the parameters can be efficiently overcome by doing an inverse analysis of the rock and soil characteristics based on field monitoring data (deformation, stress, etc.).According to the parameters obtained from the inverse analysis, the subsequent working conditions are simulated and calculated, and the deformation and stress of the foundation pit in the subsequent working conditions can be predicted.When comparing the excavation displacement inversion curve to the actual deformation curve of the soil layer parameters after inversion, the change trend of the two is essentially the same, and the relative error of the inversion is small. This paper uses the genetic algorithm to perform the inverse analysis of the engineering monitoring displacement. Experiments have proved the effectiveness of genetic algorithm prediction.