MOTOR FAULT DETECTION DATA

Induction motors play a crucial role in industrial applications, but their operation is often compromised by various mechanical and electrical faults. This paper presents a new dataset for comprehensive fault diagnosis of three-phase induction motors, using synchronized multi-sensor data collection. The dataset includes real-time measurements of vibration, voltage, and current collected from a 0.2 kW squirrel cage induction motor. Fault scenarios such as phase removal and mechanical misalignments were simulated to capture a wide range of motor behaviors. Data were collected using high-resolution sensors, with the vibration ,voltage and current sampled at 50kHz. The dataset is organized into tem distinct CSV files, covering different operational scenarios, providing a comprehensive resource for researchers aiming to develop or test fault detection algorithms. The dataset was used to train a RandomForest classifier for fault detection, achieving an accuracy of 99.82%. This demonstrates the effectiveness of the dataset for developing machine learning models aimed at real-time fault diagnosis and predictive maintenance. Unlike existing datasets, this collection provides synchronized data across multiple sensor types, enabling cross-analysis of electrical and mechanical faults. The dataset is publicly available, offering a valuable tool for advancing research in motor fault diagnosis and predictive maintenance.

感应电动机在工业应用中发挥着至关重要的作用，但其运行常受各类机械与电气故障的影响而出现阻滞。本研究针对三相感应电动机（three-phase induction motor）的综合故障诊断任务，构建了一套采用同步多传感器数据采集方案的新型数据集。该数据集采集自一台0.2kW笼型感应电动机（squirrel cage induction motor），包含振动、电压与电流的实时测量数据。研究人员模拟了断相、机械不对中等多种故障场景，以覆盖感应电动机的各类运行特性。数据采集采用高分辨率传感器，振动、电压及电流的采样率均为50kHz。该数据集包含10个独立的CSV文件，覆盖不同的运行场景，可为致力于开发或测试故障检测算法的研究人员提供一套全面的研究资源。研究人员利用该数据集训练了随机森林（RandomForest）分类器用于故障检测，最终实现了99.82%的分类准确率。这充分证明了该数据集在开发面向实时故障诊断与预测性维护的机器学习模型方面的有效性。与现有数据集不同，本数据集可提供多传感器类型的同步采集数据，支持对电气与机械故障的交叉分析。该数据集已公开，可为感应电动机故障诊断与预测性维护领域的研究推进提供极具价值的研究工具。