恶劣气候诱导的冰冻、风灾、雷害、洪涝、泥石流等自然灾害已成为威胁电网安全的主要因素。针对物联网环境下电网灾害预警及防御系统面临的决策模型动态优化、决策信息人机交互两方面的问题，本课题拟基于电力物联网获取的全景数据，研究多尺度的“灾害-电网”交互影响机理，提出基于数据-事件驱动的电网灾变过程动态演化模型；结合动态演化预测结果及在线监测数据，研究应急响应决策在线多目标优化算法，解决模型的动态匹配问题，提高决策过程的快速和精确性；建立基于情景感知的应急响应协同防御人机交互模型，实现运维业务链与信息链的双向互动，提高决策过程的效率和可靠性。通过云南电网已建成的输变电设备物联网系统的应用案例对课题研究成果进行验证，最终为实现电网灾害预警及防御系统具备实时、智能、闭环的决策能力提供理论及技术支持。

The natural disasters, including freezing, windstorm, lightning damage, dense fog,floods, landslides, have become the key element threatening the security of power grid. Disater early warning and cooperative defense for power transmission line based on internet of thing are faced with updating decision-making models for emergency and human-machine interacting. This project intends to study the dynamical evolution process of multi-scale "disasters - grid" evolution model, and presents an analysis method of the time series of catastrophic threats and a dynamic prediction model of cascading failures based on the Data - Event Driven methods. Combined with emergency response decision-line optimization algorithm and the predicted result of cascading failures model, online monitoring data improve emergency response in real-time decision-making process and solve the problem of dynamic decision model to match. This paper also establishes the disaster-based context-aware collaborative model of human-computer interaction defense to realize intelligent push information of disaster prevention decision and enhance the efficiency and reliability of emergency decision process.The methods of decision-making of Disaster prevention will be presented to optimize de-icing plants and can be corrected by feedback results, which are helpful for icing disaster prediction and preventing.Final research results are including 8-10 pieces of paper in important journals or conferences, and 2 items of originate patent.