基于参数优化的VMD-SVD和LSTM的输电杆塔倾斜状态识别  

Transmission Tower Tilt State Recognition Based on Parameter Optimization of VMD-SVD and LSTM

作　　者：赵隆[1,2] 温冠儒 刘志成 袁鹏 董新胜 ZHAO Long;WEN Guanru;LIU Zhicheng;YUAN Peng;DONG Xinsheng(School of Electronics and Information,Xi'an Polytechnic University,Xi'an 710048,China;Xi'an Qinchuang Electric Co.,Ltd.,Xi'an 710000,China;State Grid Xinjiang Electric Power Research Institute,Urumqi 830063,China)

机构地区：[1]西安工程大学电子信息学院,陕西西安710048 [2]西安勤创电气有限责任公司,陕西西安710000 [3]国网新疆电力有限公司电力科学研究院,新疆乌鲁木齐830063

出　　处：《中国电力》2023年第12期217-226,237,共11页Electric Power

基　　金：陕西省重点研发计划资助项目(2023-YBGY-069)。

摘　　要：针对输电杆塔结构状态信息提取难度大、精度低等问题,提出了一种基于北方苍鹰算法优化的变分模态分解(northern goshawk optimized variational mode decomposition,NGO-VMD)与长短期记忆(long shortterm memory,LSTM)神经网络的输电杆塔倾斜状态识别方案。通过北方苍鹰优化算法解决了变分模态分解参数难确定的问题,并且证明其分解的各阶本征模态分量(intrinsic mode function,IMF)可以有效提取出杆塔结构的模态信息。为了使信息特征更为明显,对IMF分量进行奇异值分解(singular value decomposition,SVD),发现各阶分量的奇异值在杆塔不同状态下有较为明显的区别。最后引入LSTM神经网络进行特征分类,形成故障诊断模型。依托某110kV猫头塔对模型进行试验验证,结果表明:所提方法对杆塔倾斜状态的识别准确率为96.68%,与其他方法相比,具有效率更高、稳定性更强、更加精准的优势。To address the problems of high difficulty and poor accuracy in extracting the structural state information of transmission towers,a transmission tower tilt state recognition solution is proposed based on the northern goshawk optimized variational mode decomposition(NGO-VMD)and long short-term memory(LSTM)neural network.The problem to determine the VMD parameters is solved by NGO,and it is demonstrated that the decomposed intrinsic mode function(IMF)components of each order can effectively extract the modal information of the tower structure.In order to make the information features more obvious,the singular value decomposition(SVD)of IMF components is performed,and it is found that the singular values of each order component have more obvious differences in different states of the tower.Finally,the LSTM neural network is introduced for feature classification to form a fault diagnosis model.A 110 kV cathead-type tower is used to verify the proposed model,and the results show that the proposed method can achieve an accuracy of 96.68%in identification of tower tilting state.Compared with other methods,this solution has the advantages of higher efficiency,stronger stability and better accuracy.This work is supported by Key Research and Development Program of Shaanxi Province(No.2023-YBGY-069).

关 键 词：杆塔倾斜 状态识别 北方苍鹰算法优化 自适应变分模态分解 奇异值分解 长短期记忆神经网络 

分 类 号：TM75[电气工程—电力系统及自动化] TP18[自动化与计算机技术—控制理论与控制工程]