Vibration sensor data acquisition and analysis for metro equipment

This paper adopts the method of artificial simulation equipment abnormal, collected the electrical equipment abnormal data, and puts forward the analysis of equipment abnormal data, the model can provide the premise for real-time automatic detection of anomalies, the method is proposed, can provide the electrical equipment abnormal perception data set, algorithm, model, greatly reduce the frequent maintenance frequency, provide technical solutions for automation operations and remote operations. This paper provides a major dataset of metro device vibration sensors to build an efficient machine learning-based analytical model that can detect device anomalies. The data set includes collected more perceptual signals, including speed, acceleration, frequency and other 15 variables. the data is from August 2022 to August 2023

本研究采用人工模拟电气设备异常工况的方法，采集了电气设备异常数据并开展设备异常数据分析相关研究。所构建模型可为异常实时自动检测提供前提支撑；本研究提出的方法可提供电气设备异常感知数据集、算法与模型，大幅降低设备的高频运维频次，为自动化作业与远程运维提供技术解决方案。本研究还提供了一套面向地铁设备振动传感器的大规模数据集，用于构建高效的基于机器学习的设备异常检测分析模型。该数据集涵盖多类感知信号，包含速度、加速度、频率等共15项变量，数据采集时段为2022年8月至2023年8月。