Objective In response to issues such as reduced passenger ride comfort and increased difficulty in detecting mechanical faults caused by excessive noise during URT (urban rail transit) train operation, the causes of train operation noise and mitigation measures based on a multi-feature comprehensive analysis are studied.

Method Taking Kunming Metro Line 4 as a case study, an analysis method for identifying the causes of train operation noise is proposed to determine the key noise-influencing factors. Considering the interactive effects of parameters such as train operating speed and line characteristics on noise, a multi-feature model for train operation noise is constructed based on the random forest regression model and SHAP (SHapley Additive exPlanations) interpretation method, and the multi-feature factors are comprehensively analyzed.

Result & Conclusion  Train operating speed is the most significant factor affecting passenger cabin noise, with a relative importance of 41.32%. The relative importance ranking of factors influencing cabin noise is as follows: train operating speed > gradient > curve radius. Under low-speed conditions, the marginal contribution of train speed to noise is relatively small. When the line curve radius is small, the impact of train speed on noise is magnified. When the train is running on an uphill gradient, the operating speed has a stronger positive contribution to the noise level.