Fast Prediction Method Of Environmental Vibration Induced By Metro Trains Based On Database

With the expansion of urban scale and the rapid development of rail transit,the environmental vibration sensitive areas are gradually increasing,and the vibration requirements are increasingly stringent.If the scope and requirements of vibration reduction sensitive areas are quickly screened in the planning stage of rail transit,the lines can avoid passing through the vibration sensitive areas as far as possible,for some extremely vibration sensitive areas,the lines can avoid as far as possible,which can not only provide the basis for the selection of rail transit lines,but also provide a certain basis for the design of vibration reduction measures in the future.However,it also faces the requirements of wide range,high efficiency and accurate prediction results.At present,the commonly used method is empirical formula method,which has some problems such as low prediction accuracy,unable to give more accurate reference value for some sensitive targets with high requirements in specific frequency band.Therefore,it is necessary to find a fast and accurate method to predict regional environmental vibration.The purpose of this paper is to establish a rapid prediction method of rail transit environmental vibration based on database.Based on Sobol’s global sensitivity analysis method and weighted K-means clustering algorithm,an algorithm for finding the optimal solution of transfer function is designed to achieve fast and accurate regional vibration prediction relying on transfer function database,train vibration load database and tunnel structure influence correction function database.The main research contents and achievements of this paper are as follows:（1）The database of environmental vibration prediction of rail transit is designed by MATLAB.The three sub databases,such as single and multi-layer soil transfer function database,train vibration load database and tunnel structure correction function database are completed.The source data of each sub database is calculated by expanding.（2）Sobol’s global sensitivity analysis method is used to analyze the influence of seven parameters on the transfer function.The global sensitivity curve of the law is obtained in frequency domain,which provides a basis for the design of transfer function optimization algorithm.The results show that the shear wave velocity,compression wave velocity,the change of the depthand the horizontal distanceof the vibration source have great influence on the transfer function.（3）The weighted K-means clustering algorithm is introduced to design the optimization program of transfer function.According to the parameters of the predicted target point,the transfer function database source data is called to calculate the optimal transfer function solution of the target point.Taking correlation coefficient and relative error as evaluation indexes,the accuracy of transfer function optimization algorithm is evaluated,and the results show that the algorithm has high accuracy.（4）Based on the above research results,the program of MATLAB is used to standardize the prediction flow of the rapid prediction method of rail transit environmental vibration based on database.The GUI module based on MATLAB is used to design the graphical user interface of human-computer interaction.（5）In this paper,the ground vibration response of Niwa-xiju section of Beijing Metro Line 10 is tested on site,and the prediction method established in this paper is used for prediction.The prediction results are compared with the measured results,which verifies the rationality of the rapid prediction method of rail transit environmental vibration based on database.