Abstract: To establish a mechanism for mapping the relationship between interior train vibration and noise response and the state of the line rail surface corrugation, the dynamic time warping algorithm（DTW） is applied to synchronize vibration or noise signals with rail surface corrugation signals. The adaptive wavelet time-frequency analysis（AWTA） method is used to reveal the influence of rail corrugation on interior train vibration and noise in both the time and frequency domains. A mapping relationship between train interior vibration and noise and line rail corrugation state is constructed by the K-means clustering algorithm and custom windowed Pearson coefficients. A new method is proposed to evaluate rail corrugation state based on interior train vibration or noise response. The results show that the main frequencies of train interior noise and vertical floor vibration are primarily caused by rail corrugation. Specifically, the main frequencies of train interior noise are distributed in the range of 400-1 200 Hz, and the main frequencies of vertical floor vibration are distributed in the range of 300-1 000 Hz. The Pearson coefficients of train interior noise and inner and outer rail corrugation of the curve are around 0.82 and 0.62, respectively. Additionally, the Pearson coefficients of train interior floor vertical vibration and inner and outer rail corrugation of the curve are around 0.81 and 0.72, respectively. The rail corrugation state can be evaluated by the Noise Index and Vibration Index. The results are consistent with the measured data in the field. These indices can characterize the wavelength and amplitude of rail corrugation and provide a new method for identifying subway rail corrugation.

Key words: metro, rail corrugation, train interior noise, train interior vibration, correlation analysis