上海交通大学学报 ›› 2021, Vol. 55 ›› Issue (S2): 22-30.doi: 10.16183/j.cnki.jsjtu.2021.S2.004

• • 上一篇    下一篇

台风气象环境电网设备风险量化预警及其N-m故障处置预案在线生成方法

周毅1, 秦康平1, 孙近文1, 范栋琦1(), 郑义明2   

  1. 1.国家电网有限公司华东分部, 上海 200120
    2.国电南瑞科技股份有限公司, 南京 211106
  • 收稿日期:2021-10-21 出版日期:2021-12-28 发布日期:2022-01-24
  • 通讯作者: 范栋琦 E-mail:dqfan@foxmail.com
  • 作者简介:周 毅(1982-),男,上海市人,高级工程师,从事电网调度、电力系统自动化研究.
  • 基金资助:
    国家电网有限公司华东分部科技项目(SGHD0000DKJS100235)

Real-Time Risk Evaluation Method of Power System Equipment and N-m Fault Contingency Plan Generation Under Typhoon Meteorological Environment

ZHOU Yi1, QIN Kangping1, SUN Jinwen1, FAN Dongqi1(), ZHENG Yiming2   

  1. 1. East Branch of State Grid Corporation of China, Shanghai 200120, China
    2. Nari Technology Co., Ltd., Nanjing 211106, China
  • Received:2021-10-21 Online:2021-12-28 Published:2022-01-24
  • Contact: FAN Dongqi E-mail:dqfan@foxmail.com

摘要:

为提升台风等灾害性气候下电网运行的风险预警预控能力,将气象环境数据、地理环境、输变电设备状态与电网实时状态结合,通过机器学习技术提取电网历史故障特征,将数字化输电通道的气象风险概率量化研判,实现台风气象环境下电网高风险故障的超前量化预警,并根据当时的电网状态实时生成当前和未来态N-m故障处置预案,显著增强故障预想的及时性与针对性.该方法克服了气象风险模型需要不断根据实际情况修正的困难,考虑了高风险概率设备故障后电网中的各类安全自动装置的响应情况和需要采取的控制措施.实际应用表明, 该方法可为电网调度运行风险预控提供重要决策支持.

关键词: 气象风险, 机器学习, 量化预警, 预案在线生成, 在线安全分析

Abstract:

In order to improve the power grid operation risk early warning and pre-control capabilities in typhoon and other disastrous climates, the historical fault features of power grid are extracted by utilizing the machine learning technology based on the meteorological environment data, geographic environment, and the status of power transmission and transformation equipment. The meteorological risk probability of digital transmission channel is quantified to realize the advanced quantitative warning of high-risk faults in the power grid under typhoon weather environment. The proposed method can genarate the current and future N-m fault handling plans based on the current power grid status and significantly enhance the timeliness and pertinence of fault prediction. This method overcomes the difficulty that the meteorological risk model needs to be continuously revised according to the actual situation, and takes into account the response of various safety automatic devices in the power grid after a high-risk equipment failure and the control measures that need to be taken. The practical application shows that this method can provide significant decision supports for risk pre-control in the power grid dispatching operation.

Key words: meteorological risk, machine learning, quantitative warning, fault contingency plan generation, online dynamic security analysis

中图分类号: