Strain Prediction for High-Speed Rail Canopies in Cold Regions Based on LSTM Models

Abstract

With the rapid development of high-speed rail (HSR) in China, the platform canopies of HSR stations have become crucial structures for ensuring operational safety and providing sheltered waiting areas for passengers. Temperature variations, being the primary factor affecting structural strain, lead to internal temperature responses that significantly impact the health of these structures. Modern Structural Health Monitoring (SHM) systems collect structural response data to evaluate health status and detect anomalies in real time. With the advancement of data-driven models, machine learning, particularly deep learning, is increasingly applied in civil engineering. This study employs Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) networks to handle time series data, establishing a health monitoring and early warning system for HSR station canopies. The results demonstrate that deep learning models effectively capture the complex relationship between temperature and strain, enhancing the accuracy of strain variation predictions. This provides strong support for the safe operation of HSR station canopies.