"QIT-Bearing: A Multi-Sensor Bearing Fault Dataset Under Variable Working Conditions"

"Bearing fault diagnosis under variable operating conditions has long been a research hotspot in the field of mechanical condition monitoring. In this dataset, simulated fault experiments of deep groove ball bearings were conducted under varying rotational speeds and radial loads. The tested bearing model is ER16K, which is widely applied in agricultural machinery. A total of 11 speed levels were set, ranging from 1000 rpm to 3000 rpm with a step of 200 rpm. Meanwhile, 12 radial load conditions were configured, covering loads from 100 g to 200 g at an interval of 10 g, as well as the no-load condition. For each health state (faulty or normal), 132 samples were collected. During the bearing operation, multi-source signals were acquired, including vibration signals in the X, Y, and Z directions of the test bearing, motor vibration signals, vibration signals of the left and right supporting bearings, friction signals at the test bearing, and motor current signals. Eight artificial fault types were simulated, namely, a slight outer race fault, a moderate outer race fault, a slight inner race fault, a moderate inner race fault, a slight rolling element fault, a moderate rolling element fault, a slight compound fault, and a moderate compound fault. Combined with the normal healthy state, this dataset contains nine health categories in total. The proposed dataset can be used to evaluate the performance of bearing fault diagnosis algorithms under variable operating conditions, and effectively compensates for the limitation of insufficient working condition variation in existing public bearing datasets."