Induction motor fault detection dataset - Startup currents for mains powered motors.

Experimental data were collected from four identical induction motors. One motor was healthy, while the remaining three had different types of rotor faults: a single broken rotor bar, two broken rotor bars, or a damaged ring.

To increase the moment of inertia, the motors were mechanically coupled in pairs via their shafts. In all experiments, the startup current was measured, but the motors were supplied with different voltages. This resulted in varying startup durations and, consequently, different startup current waveforms.

This variation is important because the objective is to develop a diagnostic procedure that is sensitive to phenomena occurring during the startup process itself, rather than to the specific waveform shape characteristic of a particular motor type.

Measurements were performed using a standard data acquisition card and LEM current transducers. The sampling frequency was kept constant at 1150 Hz in all cases.

Dataset includes:

raw data in "*.dat" files

raw data converted to csv's

processing scripts to clear and analyze the data and filter design information 

processed time domain data

processed frequency domain data

graphs of current waveforms (filtered and not) and their FFTs

The data set was created in the project NCN OPUS "Process Fault Prediction and Detection" (UMO-2021/41/B/ST7/03851)

Data was used in the following papers:

Jarzyna, K., Rad, M., Piątek, P., Baranowski, J. (2023). Bayesian Fault Diagnosis for Induction Motors During Startup in Frequency Domain. In: Pawelczyk, M., Bismor, D., Ogonowski, S., Kacprzyk, J. (eds) Advanced, Contemporary Control. PCC 2023. Lecture Notes in Networks and Systems, vol 709. Springer, Cham. https://doi.org/10.1007/978-3-031-35173-0_2

A. Dudek, K. Jarzyna and J. Baranowski, "Mixture Based Classifier Using Gaussian Processes for Induction Motor Diagnosis," IECON 2024 - 50th Annual Conference of the IEEE Industrial Electronics Society, Chicago, IL, USA, 2024, pp. 1-6, doi: 10.1109/IECON55916.2024.10906040.