Experimental Vibration Dataset for Rolling Bearing Outer Race Fault Diagnosis under Variable Load Conditions
Description
This dataset provides vibration signals from rolling element bearings, specifically the UC204 insert ball bearing (GBR), collected under healthy and faulty conditions for research in fault diagnosis and condition monitoring. Experimental Setup: Hardware: Measurements were captured using an ADXL210 accelerometer mounted near the bearing housing. Data Acquisition: Signals were processed via a National Instruments USB-6212 (16-bit resolution) interfaced with LabVIEW. Operational Parameters: The shaft speed was constant at ~1500 rpm (25 Hz). Three load levels were tested: 0.25 hp, 0.4 hp, and 0.6 hp. Fault Conditions: Artificial outer race defects were introduced as linear grooves with four severity levels (lengths): 0.18 mm, 0.36 mm, 0.54 mm, and 0.72 mm. Data Structure: Sampling Rate: 3200 Hz. Signal Length: 32,000 samples (~10 seconds) per file. Quantity: 10 signals per operating condition. Format: Plain text (.txt) time-series files, organized into folders by condition and load. This dataset is suitable for vibration analysis, signal processing, and machine learning applications in predictive maintenance.
Files
Steps to reproduce
1. Experimental Setup: Install the UC204 bearing into the housing unit of the laboratory test rig. Mount the ADXL210 accelerometer near the bearing housing to capture radial vibrations. Connect the sensor to the NI USB-6212 DAQ board and initialize the LabVIEW interface. 2. System Startup and Calibration: Power on the system via the electrical panel. Set the shaft rotational speed to 1500 rpm (25 Hz) using the frequency inverter. Verify speed stability using the encoder readings. 3. Data Acquisition Protocol: Set the resistive load bank to the first level (0.25 hp). Perform signal acquisition: record 10 signals at a sampling frequency of 3200 Hz (32,000 samples per signal). Repeat the acquisition for the remaining load levels (0.4 hp and 0.6 hp). 4. Fault Induction and Comparison: Replace the healthy bearing with those containing artificial outer race defects (grooves of 0.18 mm, 0.36 mm, 0.54 mm, and 0.72 mm). For each damaged bearing, repeat the full acquisition cycle (Steps 2 and 3) under all load conditions to ensure data consistency. 5. Data Storage: Export the raw data from LabVIEW as plain text (.txt) files for post-processing and analysis.
Institutions
- Universidade Regional do CaririCeará, Crato
- Universidade Federal da ParaíbaParaíba, João Pessoa