Fault detection system of subway sliding plug door based on adaptive EMD method

With the rapid development of urban rail transit, the safety of subway sliding plug doors has become a great concern. To improve the operational reliability of the sliding plug door, we developed a fault detection system based on the adaptive empirical mode decomposition (AEMD). Firstly, we designed a hardware acquisition device and analysis software to collect motor current signal data during the opening and closing of the sliding plug door. Secondly, to address the impact of noise on signal analysis, the AEMD denoising method is proposed. This method employs EMD to obtain intrinsic mode functions (IMFs), and select the appropriate IMF components for reconstruction based on the adaptive threshold of Hausdorff distance, resulting in improved denoising effectiveness. Thirdly, waveform segments of different faults are sliced to reduce the amount of computation and effectively improve recognition accuracy. Meanwhile, this paper utilizes feature selection methods and machine learning techniques to classify the 12 subway sliding plug door faults. It is worth noting that most of these faults have not been extensively studied in previous classification research. The experimental results show that the identification accuracy reaches 98.96% on the practical platform. Moreover, the effectiveness and robustness of our proposed method are further validated through practical tests, ablation experiments, and comparisons with other relevant literature.