摘要
当前振荡扰动源识别方法由于特征提取能力不足、场景适应性差等问题,扰动源识别准确性偏低。为此设计基于GWO-SVM的升压站强迫共振型低频振荡扰动源识别方法。提取升压站强迫共振型低频振荡信号频率作为信号特征。利用GWO算法优化SVM模型的核参数与惩罚因子,构建优化后的SVM强迫共振型低频振荡扰动源识别模型。在优化后的SVM强迫共振型低频振荡扰动源识别模型中输入提取的升压站强迫共振型低频振荡信号特征,实现强迫共振型低频振荡扰动源识别。测试结果表明,设计方法的OMMD整体高于0.88,NRA整体高于0.87,CSGE整体低于0.02,说明设计方法在振荡扰动源识别中兼具高精度、强鲁棒性与优异泛化能力。
The current methods for identifying oscillation disturbance sources have low accuracy due to insufficient feature extraction capabilities and poor scene adaptability.Design a method for identifying forced resonance low-frequency oscillation disturbance sources in boost stations based on GWO-SVM.Extract the frequency of the forced resonance type low-frequency oscillation signal from the booster station as the signal feature.Using GWO algorithm to optimize the kernel parameters and penalty factors of SVM model,construct an optimized SVM forced resonance low-frequency oscillation disturbance source identification model.In the optimized SVM forced resonance type low-frequency oscillation disturbance source identification model,the extracted characteristics of the forced resonance type low-frequency oscillation signal of the boosting station are input to achieve the identification of the forced resonance type low-frequency oscillation disturbance source.The test results show that the overall OMMD of the design method is higher than 0.88,the overall NRA is higher than 0.87,and the overall CSGE is lower than 0.02,indicating that the design method has high accuracy,strong robustness,and excellent generalization ability in identifying oscillation disturbance sources.
作者
袁勇
陈宏兵
冯海东
YUAN Yong;CHEN Hongbing;FENG Haidong(Jiangsu Jiangnan Power Co.,Ltd.,Nanjing 210028,China;Sichuan Rongtong Sihai Construction Engineering Co.,Ltd.,Nantogn,Jiangshu 226001,China)
出处
《自动化与仪器仪表》
2026年第1期72-76,共5页
Automation & Instrumentation
关键词
GWO-SVM
升压站
信号频率
强迫共振型低频振荡
扰动源识别
GWO-SVM
boosting station
signal frequency
forced resonance type low-frequency oscillation
disturbance source identification
