Condition assessment of high-speed railway track structure based on sparse Bayesian extreme learning machine and Bayesian hypothesis testing

ABSTRACT

Aiming at condition assessment of ballastless high-speed railway track structures, in this study, a probability prediction model of the structural static responses under temperature loads based on sparse Bayesian extreme learning machine (SBELM) is constructed. Utilizing the probabilistic predictions of the structural static responses, a Bayesian hypothesis testing-based condition assessment method for track structures is proposed. This method is employed for long-term monitoring data analysis of a high-speed railway track structure with a small-radius curve. Implicit mappings between the temperature loads and the structural static responses are obtained by training a SBELM model, and reliable predictions of the subsequent structural static responses based on the monitored temperature data are yielded. Subsequently, the probabilistic predictions of the structural responses are compared with measured data by Bayesian hypothesis testing for effective condition assessment. The illustrative application validates that the proposed method can realize the condition assessment of high-speed railway track structures effectively.

KEYWORDS:

• high-speed railway
• structural health monitoring
• structural condition assessment
• sparse bayesian learning
• extreme learning machine
• bayesian hypothesis testing

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Major Technology R&D projects of China Railway Construction Co., Ltd [2021-A03]; Young Elite Scientists Sponsorship Program by CAST [2021QNRC001]; the National Natural Science Foundation of China [52078174]. .