Research Of Adaptive Coordinated Protection Based On Multi-agent System

In order to comply with the development trend of current power grid,in the future,China will rely on UHV power grid to construct a solid smart grid.So it makes more rigorous requirements to rely protection.As a guard of power grid,the factor of prohibiting refusing and mistaking to move has been the most basic.Thus,it is no time to delay to design a rely protection which possesses superior adaptability and stability.The ideology of adaptive protection was born under the huge background.As the development of computer technique and artificial intelligence,scholars found that the adaptive protection and artificial intelligence can be highly combined by virtue of the autonomy and study ability in artificial intelligence.The integrate of adaptive protection and artificial intelligence can make up each weakness.Among them,multi-agent technology is the most representative.Aiming at the weakness of traditional rely protection,this paper emphasically applys the multi-agent system in rely protection and designs an adaptive protection system based on multi-agent system,which broadens the domain of rely protection.In the Beginning,the paper introduces the application of multi-agent technology in adaptive rely protection.Adaptive protection can not achieve without the topology analysis of grid.To contrast with the depth first search,using the more reasonable algorithm of breadth first search,a portion of graph theory search to make a topology in power grid without back tracking,and then reducing the frequency of search.The network maximum flow algorithm is introduced soon after and mainly elaborated the Edmonds and Karp algorithm,which belongs to network maximum flow algorithm.Combined with the breadth first search,an electrical connectivity analysis is put forward by Edmonds and Karp algorithm.The program put forward can shorten the time of mathematical operation in real-time environment.Multi-agent system is a hot spot in artificial intelligence.This is paper introduces the mechanism of reinforcement learning among agents and enhances the adaptability.To compare the architecture in distributed and centralized multi-agent system,construcing a wide area protection system based on the layering agent module.And then mainly introducing the coordination layer in the middle,which is the core of the whole system.The coordination layer can accomplish most of the data processing and cooperate with the upper and lower layer,which ensure that the system can make the most correct decision in real-time environment.When the malfunction occurs,thissystem can be the alternative protection to cut the malfunction with the main protection,in order to ensure the safety and stability of grid.Meanwhile,this paper develops the multi-agent system based on JADE in Netbeans integrated development environment,through monitoring platform and the dynamic display of the visual interface,achieving the coordinated protection of grid.In order to reflect the adaptability of protection based on MAS,this paper finally proposes a design of adaptive distance protection based on MAS,which can deal with the issue that traditional distance protection do not well in confronting the transition resistance.This design can eliminate the influences of transition resistance by regulating the protect boundary automatically,reducing the probability of refusing and mistaking to move in protection device.