| 基于支持向量机和长短期记忆网络的暂态功角稳定预测方法 |
| |
| 引用本文： | 刘俐,李勇,曹一家,汤吉鸿,朱军飞,杨丹,王炜宇.基于支持向量机和长短期记忆网络的暂态功角稳定预测方法[J].电力自动化设备,2020,40(2):129-136. |
| |
| 作者姓名： | 刘俐 李勇 曹一家 汤吉鸿 朱军飞 杨丹 王炜宇 |
| |
| 作者单位： | 湖南大学 电气与信息工程学院，湖南 长沙 410082,湖南大学 电气与信息工程学院，湖南 长沙 410082,湖南大学 电气与信息工程学院，湖南 长沙 410082,国网湖南省电力有限公司，湖南 长沙 410007,国网湖南省电力有限公司，湖南 长沙 410007,国网湖南省电力有限公司，湖南 长沙 410007,湖南大学 电气与信息工程学院，湖南 长沙 410082 |
| |
| 基金项目： | 国家自然科学基金资助项目(51520105011)；国网湖南省电力有限公司科技项目(5216A5170012) |
| |
| 摘 要： | 为实现暂态功角稳定性及功角轨迹的预测,提出一种支持向量机(SVM)与长短期记忆(LSTM)网络相结合的预测方法。根据系统动态特性构造暂态特征变量,采用SVM训练暂态稳定性分类器,对暂态稳定进行初步评估;利用LSTM网络对分类器评估的失稳样本进行发电机功角轨迹预测,提前发现失稳机组,减少误判样本数。通过IEEE 10机39节点系统产生训练样本并对所提方法进行测试,结果验证了所提方法的快速性和精确性。
|
| 关 键 词： | 暂态功角稳定预测 支持向量机 循环神经网络 长短期记忆网络 功角轨迹预测 |
| 收稿时间： | 2019/4/4 0:00:00 |
| 修稿时间： | 2019/11/15 0:00:00 |
Transient rotor angle stability prediction method based on SVM and LSTM network |
| |
| LIU Li,LI Yong,CAO Yiji,TANG Jihong,ZHU Junfei,YANG Dan and WANG Weiyu.Transient rotor angle stability prediction method based on SVM and LSTM network[J].Electric Power Automation Equipment,2020,40(2):129-136. |
| |
| Authors: | LIU Li LI Yong CAO Yiji TANG Jihong ZHU Junfei YANG Dan and WANG Weiyu |
| |
| Affiliation: | College of Electrical and Information Engineering, Hunan University, Changsha 410082, China,College of Electrical and Information Engineering, Hunan University, Changsha 410082, China,College of Electrical and Information Engineering, Hunan University, Changsha 410082, China,State Grid Hunan Electric Power Company Limited, Changsha 410007, China,State Grid Hunan Electric Power Company Limited, Changsha 410007, China,State Grid Hunan Electric Power Company Limited, Changsha 410007, China and College of Electrical and Information Engineering, Hunan University, Changsha 410082, China |
| |
| Abstract: | In order to realize the prediction of transient rotor angle stability and rotor angle trajectory, a prediction method with the combination of SVM(Support Vector Machine) and LSTM(Long Short-Term Memory) network is proposed. The transient characteristic variables are constructed according to system dynamic features, and SVM is adopted to train the transient stability classifier for preliminary assessment of transient stability. LSTM network is used to predict the generator rotor angle trajectory of instability samples assessed by the classifier for discovering the instability generators in advance and reducing the number of misjudged samples. The training samples are generated by IEEE 10-generator 39-bus system and the proposed method is tested, and the results verify the quickness and accuracy of the proposed method. |
| |
| Keywords: | transient rotor angle stability prediction support vector machines recurrent neural network long short-term memory network rotor angle trajectory prediction |
| 本文献已被 CNKI 维普 等数据库收录！ |
| 点击此处可从《电力自动化设备》浏览原始摘要信息 |
|
点击此处可从《电力自动化设备》下载全文 |
|
