Font Size: a A A

Study On Decision-making Technology Of Emergency Repair Strategy For Large-scale Power Failure Based On Reinforcement Learning

Posted on:2020-11-10Degree:MasterType:Thesis
Country:ChinaCandidate:P L DingFull Text:PDF
GTID:2392330572477706Subject:Control engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of Chinese economy,people have higher and higher requirements for power stability.The large-scale blackout influences large area of customers and involves a wide geographical so much so that bring big loss and inconvenience for socioeconomic and people.The establishment of a large-scale blackout emergency repair drill system based on wargame deduction can effectively promote the effectiveness of the power enterprise emergency drill,enhance the scientific,participation and organization of the emergency actual combat,and improve the power enterprise employees'ability to deal with the power emergency.The drill system for emergency repair of large-scale blackout based on wargame deduction can overcome the limitations of traditional drill mode.Based on the traditional wargame deduction,the two sides of the deduction confrontation are transformed into man and nature,and the LBS system is combined to make the emergency command makers closer to the field rescue,so that the emergency command and management personnel can know the emergency repair situation in real time.According to the actual scene,the model is established for each element of wargam,and the data support is provided for the decision-making level.At the decision-making level,the decision-making model of wargame is established by combining the deduction and behavior tree.Through the different combinations of the nodes of the behavior tree,the complex logic function of the strategy deduction and decision-mak:ing is realized,and the behavior tree of large-scale blackout emergency management is constructed.According to the specific scenario,the nodes of the behavior tree are automatically rearranged with reinforcement learning technology,and the behavior tree is optimized to improve the efficiency of emergency management.A set of self-optimizing interactive emergency drill system will be formed to facilitate the organization of emergency drill and exercise the ability of unified linkage and combined peacetime and wartime emergency response.
Keywords/Search Tags:Wargame, Blackout, Emergency repair, Behavior tree, Reinforcement learning
PDF Full Text Request
Related items