Predicting Mechanical State of High-Speed Railway Elevated Station Track System Using a Hybrid Prediction Model

Abstract

Elevated station track system is one of the most vulnerable parts in high-speed railway and prone to various defects during long-term service. The structural mechanical state will be deteriorated with the occurrence of defects, which will finally threaten the operation. Therefore, it is essential to monitor and accurately predict the structural mechanical state of elevated station track system. However, the existing prediction methods cannot achieve an accurate prediction of the structural mechanical state of the elevated station track system. Aiming at the problem, a hybrid model integrating wavelet transform, convolutional neural network, and long-term memory was proposed, which has the best performance compared with state-of-art methods and can be expanded to the state prediction of civil infrastructures. The prediction method can pre-evaluate the structural state, guide timely maintenance, and contribute to the safety of the high-speed railway.