Source Scheme Design of 220kV Power Grid in Vision Based on Back Propagation Neural Network

Short-circuit capacity of external system, operation mode of main transformers in 500kV substations, and capacity of local power plants dominate the source program of 220kV power grid. Since such factors and their combination are variable and flexible, and the network parameters of the power grid in vision are uncertain, it is not easy to design reasonable source schemes. In order to improve the guidance of long-term grid planning and reduce the calculation burden of planning engineers, this paper proposes a vision 220kV power grid source design method determining the number of transformers in parallel of substation, based on Back Propagation (BP) neural network. In accordance with typical network structures of 220kV power grid, the parameters of equipment such as generators and transformers are chosen as the input variables of the BP neural network, and the short-circuit current of the 220kV busbar of transformers of substation as the output variable. The Cuckoo Search Algorithm (CSA) is employed to optimize BP neural network parameters to improve model accuracy. After trained by historical data, the BP neural network is tried to facilitate substation source scheme design of a future 220kV power grid with cases analysis. The results indicate the effectiveness and practicability of the proposed method.