Multi-Operational Condition Broken Rotor Bar Fault Detection in Induction Motors Using Accelerometers and 2D Maps

Abstract

Three-phase induction motors (TIMs) play a crucial role in the industrial sector, powering 90 % of machines and applications. They are favored for their simplicity, versatility, reliability, and durability. However, at the same time, these motors present failures of various types, such as broken rotor bars, which represent on average 10 % of the defects. These failures can lead to significant financial losses, forcing companies to perform preventive maintenance, often unnecessarily disassembling motors to prevent unexpected breakdowns. In this sense, researchers explore non-invasive diagnostic methods such as acoustic sensors and accelerometers, facilitating early detection of faults without disassembly. However, the detection of low-magnitude damage can be masked by noise due to the different operating conditions of the machine. Some of recent research also uses expensive sensors and overly difficult methodologies, which motivated the following proposal: a low-cost accelerometer and simplified signal processing to detect minor damages. Therefore, this study focuses on the application of accelerometers to detect a minor rotor failure in TIMs by analyzing vibration patterns. The paper proposes a method that involves accelerometer sensors, root mean square (RMS), and Skewness signal processing. Experiments were carried out to test motor performance under different voltage and loading conditions that could mask the fault signal and hinder fault classification. The results demonstrate the effectiveness of using accelerometers for fault detection, providing a viable, noiseresistant, and cost-effective option for industrial applications.