Bearing fault detection in induction motors based on density of maxima analysis

This paper presents a novel approach for early detection of bearing faults in induction motors based on chaotic behavior analysis. Two comprehensive case studies are detailed. The first study delves into an analysis utilizing an internal database containing one hundred tests for each operational condition, ensuring enough statistical information for effective fault detection. Three distinct fault types (ball damage, outer race damage, and rusted bearing) were induced to simulate real-world scenarios. The second case study focuses on vibration signals sourced from Case Western Reserve University, specifically investigating ball and inner race damage for both loaded and unloaded conditions. Both current and vibration signals were analyzed employing chaos-based density of maxima calculations. This methodology entails quantifying the peaks and evaluating the density of maxima within the motor signals for correlation length calculation, providing valuable chaotic insights for fault detection with results that yield 100% of accuracy.