Research On Island Division Of Power Cyber-Physical Systems Based On Common Edge Nodes

With the development of power grid becoming more intelligent,it is obvious that information and communication technology is prominently important.The research on power grid has gradually changed from the previous power network to two-layer network containing both physical and information network.Iin order to reduce the occurrence of the phenomenon that the power grid loses monitoring ability due to the large-scale power outage like in North America in 2003 and the blackout in Italy,which lead to the cascading failure caused by the information node,this thesis proposes a method on island division of power Cyber-Physical Systems based on common edge nodes to recover after nodes failure.In the past,the island division of single-layer with DGs only considers about power nodes,information nodes has been ignored.According to this,this paper conducts in deep research.First,the model construction of single-layer power network islanding has been improved.Based on the objective function and the actual operation,the constraints are improved and the simplified islanding model is obtained.Then when the specific algorithm is improved,the fish population algorithm is used to optimize the results.The Kruskal algorithm is first used to generate a radial distribution network,which facilitates depth-first search for the number and forms the initial fish population.Because the fixed step size of the fish swarm algorithm is flawed,the optimization speed of the algorithm is slowed down.This paper proposes an adaptive step size fish swarm algorithm,which effectively solves this problem.According to the switch characteristics in the power grid,the original algorithm is changed to the binary fish swarm algorithm,which is easier to understand and apply,and reduces the coding space and increases the search breadth.After doing the above preparations,the two-dimensional modeling of the power information physical network is carried out,and the cascading failure risk index is proposed.Then,a fault recovery algorithm is proposed using common edge node theory,so that more nodes are classified into islands.The simulation results show that the fault recovery method proposed in this paper improves the load recovery rate and the information node recovery rate also increases accordingly,which greatly improves the ability of regional self-distribution using its own distributed energy when the grid is faulty,providing more time for repairing.At the same time,the information node data is transmitted back to the dispatch center,which also provides an effective way for the staff to perform fault analysis.