Transient Stability Assessment of Power Systems Based on Slow Feature Analysis

The classify technology to transient stability assessment (TSA) has been studied and applied as an effective method for real-time transient monitoring. However, Large power systems usually have the high dimensionality of variables, which raises the curse of dimensionality and requires an extra dimensionality reduction technique to make the TSA methods appropriate in practice. This paper firstly employs a prevailing data-driven method, slow feature analysis (SFA), to assess transient stability, and shows its suitability for TSA. SFA-based TSA can be a classifier also an online monitoring tool, which simultaneously keeps high performance in handling the curse of dimensionality and outperforms in computational complexity. The algorithm of SFA is conceptually explained, and the IEEE 10-machine, 39-bus New England case is utilized to verify the validity and effectiveness of SFA-based TAS.