Cylindrical Roller Bearing Vibration Dataset with Multiple Defect Orientations and Widths under Varying Speeds and Loads
Description
Cylindrical Roller Bearing Vibration Dataset with Multiple Defect Orientations and Widths under Varying Speeds and Loadsunder Varying Speeds and Loads contains single-channel vibration signals acquired from cylindrical roller bearings under controlled laboratory conditions. The dataset covers five bearing conditions: healthy condition (N), inner-race fault (IR), outer-race fault (OR), rolling-element fault (B), and outer-race–rolling-element compound fault (ORB). Data were collected under two load conditions, 0HP (no-load) and 1HP (loaded), and six rotational frequencies of 5, 7.5, 10, 12.5, 15, and 20 Hz. Artificial localized groove-like defects were introduced with controlled defect orientations and widths. Inner-race and outer-race defects include orientations of 30°, 45°, and 90°, while the rolling-element defect orientation was fixed at 90°. Defect widths of 0.2, 0.4, and 0.6 mm were considered, with a fixed defect depth of 1.0 mm. The compound-fault data combine outer-race and rolling-element defects with matched defect widths. The dataset comprises 372 experimental conditions and 1,488 raw vibration records, with four repeated recordings for each condition. Each record contains 1,000,000 samples acquired at 20 kHz, corresponding to a recording duration of 50 s. The dataset is provided together with metadata describing the operating conditions, fault type, defect geometry, and trial information for each record. It can support bearing fault diagnosis, signal-processing studies, cross-speed and cross-load evaluation, defect-geometry analysis, compound-fault identification, and machine-learning or deep-learning research.
Files
Institutions
- Wuhan University of Science and TechnologyHubei, Wuhan
Categories
Funders
- Jiangsu Province Qinglan Project Outstanding Teaching Team ProjectGrant ID: [2024] No. 14