基本信息:
- 专利标题： 一种基于机器学习的电网全局延时态势感知方法
- 专利标题（英）：Overall and delayed situation awareness method for power grid based on machine learning
- 申请号：CN201611160493.0 申请日：2016-12-15
- 公开(公告)号：CN106779215A 公开(公告)日：2017-05-31
- 发明人： 饶玮 , 蒋静 , 胡斌 , 裘洪彬 , 赵兵兵 , 曹军威 , 明阳阳 , 陈建会
- 申请人： 全球能源互联网研究院 , 清华大学 , 国家电网公司
- 申请人地址： 北京市昌平区未来科技城北区国网智能电网研究院院内; ;
- 专利权人： 全球能源互联网研究院,清华大学,国家电网公司
- 当前专利权人： 全球能源互联网研究院,清华大学,国家电网公司
- 当前专利权人地址： 北京市昌平区未来科技城北区国网智能电网研究院院内; ;
- 代理机构： 北京安博达知识产权代理有限公司
- 代理人： 徐国文
- 主分类号： G06Q10/04
- IPC分类号： G06Q10/04 ; G06Q50/06 ; G06N99/00
The invention relates to an overall and delayed situation awareness method for a power grid based on machine learning. The method comprises the following steps: establishing a sample matrix through measurement values of sampled nodes; creating stability flag values of the samples according to voltage values of the samples in the sample matrix; compressing the dimensionality of the sample matrix, and training classifiers through the sample matrix that has undergone the dimensionality compression; and predicting a power grid state stability probability at a next moment through the classifiers. The method provided by the invention has the advantages that the power grid voltage stability is predicted through the machine learning algorithm, the nodes are further combined to provide a power grid situation awareness evaluation result, and a feature selection period is staggered from a prediction time node during training of each node classifier to establish a delayed prediction method, so that better restoration effects for a complex system are achieved.
公开/授权文献:
- CN106779215B 一种基于机器学习的电网全局延时态势感知方法 公开/授权日：2021-12-03