| 引用本文: | 但扬清,刘文颖,朱艳伟,蔡万通,王维洲,梁 琛.含大规模风电集中接入的电网自组织临界态辨识[J].电力自动化设备,2016,36(5): |
| DAN Yangqing,LIU Wenying,ZHU Yanwei,CAI Wantong,WANG Weizhou,LIANG Chen.Self-organized critical state identification of power grid with centralized integration of large-scale wind power[J].Electric Power Automation Equipment,2016,36(5): |
|
| 摘要: |
| 基于学习向量量化（LVQ）神经网络法，提出了一种含大规模风电集中接入的电网自组织临界态辨识方法。该方法以加权潮流熵、网络拓扑熵和风电波动熵等熵值物理指标为主要输入对象，以停电数据的幂律尾曲线拟合方法生成的数据作为训练样本，采用LVQ1和LVQ2算法对创建的LVQ神经网络进行训练，然后利用经过训练后的网络模型进行含大规模风电集中接入的电网自组织临界态辨识。该方法建立了物理指标与自组织临界态之间的直接联系，避免了采用传统辨识方法多次仿真和较多主观干预的问题。实例仿真结果表明，所提方法能够正确地对电网运行状态进行辨识。 |
| 关键词: 集中式大规模风电 连锁故障 自组织临界态辨识 LVQ神经网络 物理指标 幂律尾曲线 风电 熵 神经网络 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金资助项目（51377053）；国家科技支撑计划项目（SQ2015BA0502239） |
|
| Self-organized critical state identification of power grid with centralized integration of large-scale wind power |
|
DAN Yangqing1, LIU Wenying2, ZHU Yanwei3, CAI Wantong2, WANG Weizhou4, LIANG Chen4
|
|
1.State Grid Zhejiang Electric Power Company Economic and Technology Research Institute，Hangzhou 310007，China;2.School of Electrical & Electronic Engineering，North China Electric Power University，Beijing 102206，China;3.State Grid Zhejiang Electric Power Company Ningbo Power Supply Company，Ningbo 315010，China;4.State Grid Gansu Electric Power Company Electric Power Research Institute，Lanzhou 730050，China
|
| Abstract: |
| A method of self-organized critical state identification based on the LVQ（Learning Vector Quantization） neural network is proposed for the power grid with centralized integration of large-scale wind power，which takes the entropy indicators as the main input objects，including weighted power flow entropy，network topology entropy，wind power fluctuation entropy，etc.，applies the data generated by the power-law tail curve fitting based on the outage data as the training samples and adopts the LVQ1 and LVQ2 algorithms to train the created LVQ neural network，which is then used to identify the self-organized critical state of power grid with centralized integration of large-scale wind power. A direct connection is established between the physical indicator and the self-organized critical state by this method，avoiding the repeated simulation and subjective intervention existed in the traditional identification methods. Simulative results show that the proposed method can correctly identify the operating states of power grid. |
| Key words: centralized large-scale wind power cascading failure self-organized critical state identification LVQ neural network physical indicators power-law tail curve wind power entropy neural networks |
