Fault Detection Based on Deep Learning

In modern industry, bearings play a crucial role as important components of mechanical equipment. The failure and damage of bearings have significant impacts on production. Therefore, real-time detection and accurate identification of bearing faults are essential for improving industrial production efficiency and reducing industrial accident rates. This study is based on bearing data from Case Western Reserve University, and it utilizes Continuous Wavelet Transform(CWT) to transform time-series vibration data into time-frequency images. It combines CWT with the RegNet algorithm to detect bearing faults. The experimental results demonstrate that CWT-RegNet exhibits excellent performance in bearing fault detection.