摘要
为了保证电力系统在风荷载作用下安全稳定运营,建立准确的输电线舞动模型面临巨大挑战。本研究以复杂线路中的覆冰四分裂导线为研究对象,采用数据驱动稀疏识别算法重建舞动模型。首先,推导了覆冰四分裂导线舞动的理论模型;然后,基于非线性动力学稀疏识别(SINDy)算法重建舞动模型;最后,通过具体算例来研究不同噪声幅值下舞动模型的重建精度。结果表明:在无噪声测量数据下,SINDy算法准确地恢复了舞动模型的结构及参数;通过对舞动模型进行灵敏度分析,发现随着噪声幅值的增大,三次项的系数均值误差较一次项大;在噪声幅值为0.1以内时,重建模型位移时程的决定系数在0.93以上,其中噪声幅值为0.05时,决定系数高达0.97。此外,该算法仅需少量数据即可达到很高的模型重建精度。研究成果对验证防舞设计及制定防舞计划具有重要作用。
To ensure the safe and stable operation of the power system under the action of wind loads,it is a huge challenge to establish an accurate transmission line galloping model.In this paper,taking the iced quad bundle conductors in complex lines as the research object,a data-driven sparse identification algorithm is used to reconstruct the galloping model.First,the theoretical galloping model of the iced quad bundle conductors is derived.Then,the galloping model is reconstructed based on the sparse identification of nonlinear dynamics(SINDy)algorithm.Last,the reconstruction accuracy of the galloping model under different noise amplitudes is studied through specific examples.The results show that under the noise-free measurement data,the SINDy algorithm can accurately restore the structure and parameters of the galloping model.By sensitivity analysis of the galloping model,with the increase of the noise amplitude,the coefficient mean error of the cubic term is larger than the coefficient mean error of the first term.When the noise amplitude is within 0.1,the determination coefficient of the displacement time history of the reconstructed model is above 0.93.Meanwhile,when the noise amplitude is 0.05,the determination coefficient is as high as 0.97.Furthermore,the algorithm achieves high model reconstruction accuracy with only a small amount of data.The research results play an important role in verifying anti-galloping design and formulating anti-galloping plans.
作者
刘小会
陈黎兵
杨军
赵建坤
安东
LIU Xiaohui;CHEN Libing;YANG Jun;ZHAO Jiankun;AN Dong(Inner Mongolia Power(Group)Co.,Ltd.,Inner Mongolia Power Research Institute Branch,010020 Hohhot,China;Inner Mongolia Enterprise Key Laboratory of High Voltage and Insulation Technology,010020 Hohhot,China;School of Civil Engineering,Chongqing Jiaotong University,400074 Chongqing,China;State Key Laboratory of Bridge and Tunnel Engineering in Mountain Areas,Chongqing Jiaotong University,400074 Chongqing,China)
出处
《应用力学学报》
北大核心
2026年第1期68-78,共11页
Chinese Journal of Applied Mechanics
基金
国家自然科学基金资助项目(No.51308570)
重庆市自然科学基金资助项目(No.cstc2021jcyj-msxmX0166)。
关键词
覆冰导线
数据驱动
稀疏识别
舞动方程
四分裂
iced conductor
data-driven
sparse identification
galloping equation
quad bundle
