Bulk Low-Inertia Power Systems Adaptive Fault Type Classification Method Based on Machine Learning and Phasor Measurement Units Data
This research focuses on developing and testing a method for classifying disturbances in power systems using machine learning algorithms and phasor measurement unit (PMU) data. To enhance the speed and accuracy of disturbance classification, we employ a range of ensemble machine learning techniques,...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/13/2/316 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This research focuses on developing and testing a method for classifying disturbances in power systems using machine learning algorithms and phasor measurement unit (PMU) data. To enhance the speed and accuracy of disturbance classification, we employ a range of ensemble machine learning techniques, including Random forest, AdaBoost, Extreme gradient boosting (XGBoost), and LightGBM. The classification method was evaluated using both synthetic data, generated from transient simulations of the IEEE24 test system, and real-world data from actual transient events in power systems. Among the algorithms tested, XGBoost achieved the highest classification accuracy, with 96.8% for synthetic data and 85.2% for physical data. Additionally, this study investigates the impact of data sampling frequency and calculation window size on classification performance. Through numerical experiments, we found that increasing the signal sampling rate beyond 5 kHz and extending the calculation window beyond 5 ms did not significantly improve classification accuracy. |
|---|---|
| ISSN: | 2227-7390 |