Performance Degradation Analysis of Railway Vehicle Door System Based on Density Peak Clustering

This paper proposed an intelligent analysis with Density Peak Clustering (DPC) to deal with low accuracy in the performance degradation analysis of Railway Vehicle Door System (RVDS). First, the features were evaluated and selected with the Elastic Network (EN). Subsequently, an analysis model that can select degraded data from the door operation data was established by DPC flexibly. The performance degradation analysis of RVDS proposed in this paper has been verified by experiments with data collected from the bench testing door system of Hangzhou Metro Line 4, and the results can be used to validate high accuracy and practical value of the model.