Multimodal Bearing Failure Diagnosis Dataset Under Various Startup Modes of an Asynchronous Electric Motor

A dataset for failure diagnosis of bearings is proposed taking into account different startup modes of the asynchronous motor. Five starting modes were considered. These are the direct-on-line (DOL) starting, variable frequency drive (VFD) starting with periods of 5, 10, 15, and 30 seconds. The experiments were carried out in a BALTECH WS-3060 experimental setup with double-row bearing type 11206. We implemented eleven different scenarios. They include a normal bearing condition, a broken ball, a bearing outer ring failure, a broken bearing retainer, a combination of inner ring and retainer failures, lack of lubrication in a bearing, an unbalanced working wheel, and combinations of the unbalanced wheel with listed bearing failures. All failures are manually implanted before experiments. The dataset has the following parameters: sampling periods with account of the start period in various start modes, 12 vibration signals , values of temperature, rotation frequency, and acoustic data. Sampling frequency is 19 kHz.