Research On Disaster Assessment And Emergency Recovery Method Of Smart Distribution Network

Disasters influenced on power grid frequently in recent years, accompanied by a large amount of the loss of load and the damage of grid infrastructure at each time. It caused serious influence on the industrial production, economic activity and all aspects of people’s life. Under this background, the technology of distribution network coping with disaster becomes the important research topic in the distribution network. The disaster assessment of distribution network and the supply recovery method of responding to the disaster are studied in this thesis. The study is of great significance to not only improve the emergency response capacity which distribution network responds to natural disaster but also have the ability of resisting risk.With the comprehensive consideration of the impact which natural disaster played on power grid, the disaster management visualization system of Liaoning distribution network is developed based on the project which is called research and development of smart distribution network situational awareness and disaster responding technology. The system is a visualization operating platform, which sets computer, network, communication, database, GIS, GPS and other technical means as a whole. It can realize the monitoring of power grid, disaster assessment, early warning, decision-making and so on. The functions of every module are introduced and interfaces of them are showed. It states that this system is an easy operation platform with good interface and more pictures switching function.With in-depth analysis of impact factors which different types of disasters play on distribution network, corresponding grid disaster evaluation index system is established which uses the method of the improved analytic hierarchy process. And on this basis, given that the impact factors have vagueness and uncertainty, disaster assessment method of power grid based on intuitionistic fuzzy rough sets is proposed. It can realize timely and accurate assessment, determine the level of disaster and adjust update along with the development of the situation. So that it can provide scientific basis for emergency decision.In addition, from the perspective of dealing with emergencies and large scope blackout accident which caused by disasters, restoration process is divided into three stages, distribution network’s skeleton channel restoration, local distribution network restoration, regional networked restoration respectively. And aiming at the specific restoration target of every stage, it is proposed that distribution network’s skeleton channel restoration method based on depth first search and local and regional network recovery method based on ADP. Among them the local and regional network recovery method uses RNN to reconstruct the power distribution system dynamic model, so that the unknown distribution network system can be identified. And then it uses BP neural network to training critic network and control network, so that the optimal control strategy can be obtained.Finally, the distribution network power supply recovery system with optimal coordination is built based on MAS architecture. The system adopts the blackboard model cooperation way so that modules can be parallel processing and comprehensive coordination. Besides, the coordinated optimization method which combined ant colony algorithm with the Q learning is put forward. The method establishes the information sharing and enhances the interactivity among agents.