Research On Power Supply Recovery Strategy Of Smart Distribution Grid

In recent years,with the transformation of global energy supply towards clean,low-carbon and electrification,the construction of energy Internet has become the future direction of power grid development.As an important part of energy internet,intelligent distribution network is the inevitable trend and goal of the future development of distribution network.Compared with traditional distribution network,intelligent distribution network is based on efficient and reliable communication network,which has the characteristics of self-healing,strong,integrated,compatible and interactive.Power supply recovery is the key link to realize the self-healing performance of intelligent distribution network.It is of great significance to improve the reliability of power supply by reconstructing the form of the network after power failure.Power supply recovery based on heuristic rules and restoring power supply of power loss load according to preset action logic after fault has the advantages of high efficiency and clear physical meaning,which is widely used in practical engineering.Based on the analysis of the similarity of power grid morphology before and after fault,this paper proposes a heuristic power supply restoration method based on transfer affinity.Firstly the concept of relay affinity of load-tie switch is established,and the morphological relationship of power grid before and after lault is constructed.Secondly,the evaluation index and calculation method of relay affinity are studied.Finally,according to the ranking results of fault location and affinity,the relay scheme is quickly generated.This method combines affinity off-line evaluation with online decision-making,which improves the decision-making speed and scheme performance of heuristic method.At present in the research of power supply restoration of distribution network,the unified control model is adopted for different feult situations of power grid,which results in poor adaptability of control strategy to actual fault situations of power grid.By analyzing the relationship between power supply recovery demands and fault situations,this paper proposes a power supply recovery technology considering the impact of fault.Firstly,starting with the loss of power loss and the importance of fault line,the influence degree of distribution network fault is evaluated;secondly,the adaptive matching mechanism of power supply recovery model is established,and the appropriate power supply recovery model is selected according to the fault level to balance the contradiction between solving speed and scheme performance;finally,the multi-objective comprehensive learning particle swarm optimization algorithm is applied to generate the power conversion scheme.This method integrates the relationship among fault level,power supply recovery demand and decision-making algorithm.It can deal with different fault situations flexibly and effectively improve the engineering practicability of recovery scheme.In order to further improve the speed of power supply recovery,referring to machine learning and pattern recognition theory,this paper proposes a fault recovery technology based on self-learning mechanism,which endows the self-learning and self-evolution ability of self-healing system.Firstly,based on the principle of pattern recognition,the matching method of similar fault state in distribution network is studied.The network topology similarity and load distribution similarity are taken as evaluation factors,and the fault similarity evaluation model of distribution network is established.On this basis,the anticipated accidents are simulated according to the results of online risk assessment,and a self-learning database based on simulation samples and historical event processing scheme is constructed.After distribution network failure,fault features are extracted,self-learning database is used for reference,and swift generation of power conversion schemes is achieved.By introducing self-learning mechanism,the self-healing system is upgraded to an intelligent control system with self-learning,self-improvement and continuous evolution ability.In order to achieve the organic coordination of the above three power supply restoration methods and provide an efficient and reliable power supply recovery scheme for power loss load,this paper constructs an intelligent power supply recovery system based on multi-agent system,combining the structure of distribution network and the characteristics of distribution system.Firstly,the system structure model and agent functions are designed,and then the operation mechanism and cooperation strategy of the system are studied.Through the independent decision-making and information exchange of the agents,three kinds of power supply recovery technologies are coordinated and coordinated to complete the power supply recovery task together.The system takes into account the advantages of centralized control and distributed control,and has the characteristics of self-organization,self-adaptation and self-learning.It further strengthens the reliability and accuracy of power supply recovery decision-making function.