ANN-Based Model to Predict the Screening Efficiency of EPS Geofoam Filled Trench in Reducing High-Speed Train-Induced Vibration

Trench or wave barrier is an economical and cost effective way to minimize surface vibrations and screen the vibration of sensitive structures from the unwanted shaking. Previous researches mainly focused either on experimental or on analytical work. Owing to the complexities involved in the mathematical formulation, involvement of vast numbers of parameters, and the requirement of computational time to numerical analysis of an effective vibration screening system. The present study aims to explore the use of an artificial neural network to estimate the vibration screening effectiveness of an EPS geofoam filled trench in reducing train-induced vibration. Identifying different key parameters, a MLR model has been developed to know the influence of key parameters governing the vertical vibration screening of EPS geofoam trench wave barriers. It has been seen that the ANN model can effectively and accurately predict the averaged amplitude reduction ratio of the EPS geofoam infilled trench.