电工技术学报  2021, Vol. 36 Issue (1): 179-190    DOI: 10.19595/j.cnki.1000-6753.tces.191703
电力系统 |
采用深度迁移学习定位含直驱风机次同步振荡源机组的方法
陈剑1, 杜文娟2, 王海风1,2
1.新能源电力系统国家重点实验室(华北电力大学) 北京 102206;
2.四川大学电气工程学院 成都 610065
A Method of Locating the Power System Subsynchronous Oscillation Source Unit with Grid-Connected PMSG Using Deep Transfer Learning
Chen Jian1, Du Wenjuan2, Wang Haifeng1,2
1. State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources North China Electric Power University Beijing 102206 China;
2. College of Electrical Engineering Sichuan University Chengdu 610065 China
全文: PDF (1839 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 随着新能源电力电子器件的广泛接入,电力系统次同步振荡问题的诱发机理越来越复杂。为了能够及时定位到诱发次同步振荡的机组并采取措施,基于深度迁移学习提出了一种次同步振荡源定位的方法。该方法首先依据开环模式谐振理论构建仿真系统,并在仿真系统中获取训练数据样本;其次,运用卷积神经网络(CNN)进行振荡源特征提取并建立训练定位模型;最后,通过迁移学习将训练模型迁移到实际系统,以实现定位模型的应用。为验证所提方法的有效性,设计了含直驱风机并网的电力系统的仿真系统测试算例。结果表明,该方法相比于传统的特征值分析方法,具有定位准确率高、在线应用方便等优势。该方法能够在较短时间内给出判别结果,为实现振荡源的在线识别奠定了基础。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
陈剑
杜文娟
王海风
关键词 开环模式谐振次同步振荡直驱风机振荡源定位机器学习    
Abstract:With increasing penetration of new energy power electronic devices, the induced mechanism of the subsynchronous oscillation problem of power systems has become more complicated. In order to locate the unit that induces subsynchronous oscillation in time, a method for locating sub-synchronous oscillation sources based on deep transfer learning is proposed. The method first builds a simulation system based on the open-loop mode resonance theory, and obtains training data samples in the simulation system. Second, it uses convolutional neural networks(CNN) to extract the characteristics of the oscillation source and establish a training localization model. Finally, the training model is transferred to the actual situation through transfer learning, and to realize the application of the location model. In order to verify the effectiveness of the proposed method, this paper designs a simulation system example of a power system with direct-drive generators connected to the grid. The test results show that the proposed method has the advantages of high location accuracy and convenient online application, compared with the traditional eigenvalue analysis method. Besides, this method can give the identification result in a short time, which lay a foundation for realizing the online identification of the oscillation source.
Key wordsOpen loop mode resonance    subsynchronous oscillation (SSO)    permanent magnet synchronous generator (PMSG)    oscillation source localization    machine learning   
收稿日期: 2019-12-06     
PACS: TM712  
基金资助:中央高校基本科研基金资助项目(YJ201654)
通讯作者: 杜文娟,女,1979年生,教授,研究方向为电力系统稳定性理论与控制,柔性输电、大规模新能源接入电力系统的分析与控制。E-mail:ddwenjuan@qq.com   
作者简介: 陈剑,女,1992年生,博士研究生,研究方向为机器学习在电力系统中的应用、电力系统稳定性分析。E-mail:sdqzjane316@gmail.com
引用本文:   
陈剑, 杜文娟, 王海风. 采用深度迁移学习定位含直驱风机次同步振荡源机组的方法[J]. 电工技术学报, 2021, 36(1): 179-190. Chen Jian, Du Wenjuan, Wang Haifeng. A Method of Locating the Power System Subsynchronous Oscillation Source Unit with Grid-Connected PMSG Using Deep Transfer Learning. Transactions of China Electrotechnical Society, 2021, 36(1): 179-190.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.191703          https://dgjsxb.ces-transaction.com/CN/Y2021/V36/I1/179