Improved BP Neural Network for Short Circuit Current Calculation in IIDG Distribution Network

With the increasing penetration rate of inverter distributed power sources year by year, traditional short-circuit current calculation methods are no longer applicable. Therefore, this article proposes a short circuit current calculation method for IIDG distribution network based on an improved BP neural network: using the optimal voltage support strategy, the IIDG fault is equivalent to a current source model controlled by the voltage of the grid point. The feature decomposition method is used to improve the sample feature and label selection process of the BP neural network, extract key features and sample labels, and combine common features to obtain the features and sample combinations of the BP neural network. System modeling and network training were conducted in an IEEE33 node system, and compared with traditional short-circuit current calculation methods to verify that the proposed method can effectively solve the short-circuit current calculation problem of distribution networks containing IIDG and has high accuracy.