Stray Flux Signal Analysis for Faults Detection in Induction Motors During Startup Transient By Means Of Statistical Indicators

Electric motors are one of the most used primary drivers for diverse processes involving rotary movement in several fields concerning modern societies. One of the most widespread types of electric motors is Induction Machines (IMs), mainly due to their robustness, easy control, and easy maintenance. Despite their excellent characteristics and robustness, IMs are prone to fail from different constructive parts. One of the most common failures is due to broken rotor bars (BRBs). Although multiple methodologies have been proposed for diagnosing BRBs, these are mainly based on methods that rely on the amplitudes of specific fault-related frequencies, which may yield false diagnoses. More recently, transient methods based on tracking fault-related time-frequency (t-f) patterns involve using t-f methods with a high computational load, and sometimes the resolution may need to be better. This paper proposes a novel non-invasive low-burden method for the diagnosis of BRBs and misaligments. The proposal relies on the computation and amplitude tracking of stray-flux-based statistical indicators from signals captured during the startup transient. The results demonstrate the proposed method's effectiveness, validated on a kinematic chain (composed of a 0.746 KW induction motor, a belt and pulleys transmission system, and an alternator as a load).