A Neural Component Analysis Aided Extreme Learning Machine Method for Power System Transient Stability Assessment
Daowei Liu, Gaoshang Zhao, Yang Yandong, Liu Weilin, Jaydar Jingus · 2022 34th Chinese Control and Decision Conference (CCDC) · 2022
In recent years, the transient stability assessment (TSA) of power systems has attracted more and more scholars’ attention. However, the scale of the power system is gradually increasing and the corresponding characteristics are gradually increasing, which makes the accuracy of TSA greatly compromised. Therefore, it is not appropriate to directly use the obtained original data for the assessment. Considering that there are generally nonlinear characteristics in the data of the power systems, this paper adopts a recent neural network dimension reduction method to process the data first, and then performing extreme learning machine (ELM) on reconstructed data to realize TSA. The proposed TSA model can remove redundant and noisy information in the original data, thereby improving the accuracy of evaluation. Finally, the IEEE 10-machine, 39-bus New England simulation system is adopted to demonstrate the validity and superiority of the proposed model.