Simulated Inter-turn Short Circuit Faults in Wind Turbine Generator Stators: A Benchmark Dataset

While extensive Condition Monitoring and Fault Diagnosis (CMFD) systems are developed based on the Supervisory Control and Data Acquisition (SCADA) system, high-frequency data crucial for monitoring electrical failures is often absent in the standard 10-minute sampling dataset. To address this gap, we have open-sourced a dataset aimed at enhancing the development of CMFD systems in wind turbine generators.

This dataset includes 75 inter-turn short circuit faults (ITSCFs) simulated in the wind turbine (WT) generator stator using MATLAB/Simulink, based on the example: Wind Turbine - MATLAB & Simulink (mathworks.com). In the Simulink WT model, the generator stator is characterized by an equivalent impedance of  R=0.0027 Ohm. The stator winding is partitioned into healthy and faulty segments, with an increasing series of faulty winding ratios p_s:={0.05, 0.1, 0.2, 0.3, 0.5} representing the progression of degradation. Additionally, a decreasing series of short circuit resistances R_s:={1, 0.1, 0.01, 0.001, 0.0001} Ohm models the degradation of the insulation layer. Considering that ITSCFs can occur in three phases, a total of 5x5x3=75 ITCSF scenarios are generated. Including a normal operation scenario without ITSCF, the dataset comprises 76 scenarios in total.

Each scenario is simulated for 80 seconds using the default wind speed profile from the WT example. The phase-to-phase current (in Amperes) and voltage (in Volts) signals for all three phases are recorded at a sampling frequency of 4 kHz.

Both .mat and .csv files encompassing all scenarios are provided, along with metadata. The 76 .mat files, each corresponding to a specific scenario, are compressed into a zip file, where the signals are stored in time-series format, with separate metadata files included.

For those interested in accessing the Simulink model, the Wind Turbine Simulink (ITSCF).zip is available, with a ReadMe.txt file inside for guidance.

We encourage further contributions based on this benchmark WT model. Related materials and updates will be posted at https://github.com/UTD-DOES/WT-ITSCF-Benchmark. Contributions and questions are welcome at: jingyi.yan@utdallas.edu or jiezhang@utdallas.edu