Safety Analysis of High-Speed Trains under Poor Track Conditions Based on Entropy Change Theory

Abstract The smoothness of the track structure plays a critical role in ensuring the safety of high-speed trains (HSTs). However, the long-wavelength characteristics of the track structure and the timeliness limitations of existing methods pose challenges to HST safety detection. This study proposes a real-time evaluation method for HST safety based on the dynamic response entropy change rate. First, the typical wavelength characteristics of random track irregularities are extracted using wavelet theory, thereby generating a sample library of track irregularities with varying degrees of deterioration. Subsequently, a vehicle–track coupling dynamic model that considers track deterioration states is established to investigate the evolution law of the Shannon entropy of the vehicle system’s dynamic response as a function of track deterioration. The correlation between the dynamic response entropy and the running state of the train is also revealed. The results show that track deterioration causes the vehicle acceleration response to deviate from a normal distribution. The Shannon entropy of the wheelset acceleration is significantly correlated with the derailment coefficient, whereas the Shannon entropy of the train body acceleration is less sensitive to track deterioration. Furthermore, the Shannon entropy of the dynamic response increases with the length of the line section. Using the change rate of the acceleration Shannon entropy effectively eliminates the influence of line-section length. Setting the change rate of the Shannon entropy of wheelset lateral acceleration to 0.6 as the safety threshold results in a misjudgment rate of less than 3% for derailment coefficient exceedance. In summary, this study provides a theoretical basis for real-time monitoring and early warning of train running safety on deteriorated track sections.