Intelligent Prediction on the Vibration of Piled Buildings Induced by Subway Trains

The construction of subway lines inevitably passes through or adjacent to the buildings. This study presents an analytical model of the piled building near the subway lines and investigates the effects of soil and structural parameters on the piled building vibration. Then, a model for predicting the vibration of piled buildings near the subway based on the machine learning method is constructed, and an intelligent prediction method for obtaining the vibration of piled buildings with soil surface vibration response is developed. The results demonstrate that the vibration of different floors in the low-frequency region is almost independent of soil and structural parameters, and the vibration acceleration level has an amplification effect in the high-frequency region. The GA-BP model provides an effective description of the vibration properties of the piled buildings, especially for the [Formula: see text]-vibration level. The piled building vibration caused by the subway operations can be obtained by constructing an intelligent model with the ground response near the building as an input layer. The accuracy of pile foundation vibration prediction improves as the number of ground surface vibration points increases, offering the potential to simplify the analysis of train-induced vibrations in piled buildings.