基于负荷预测及广义回归神经网络的短路电流超短期预测

Ultra-short term forecasting for short circuit current based on load forecasting and general regression neural network

针对智能电网中实时状态监测和告警需求,提出一种电网短路电流超短期智能预测的方法。通过节点超短期负荷预测进行电网态势外推,采用基于广义回归神经网络的短路电流辨识方法对短期内的全网母线短路电流水平进行扫描,实现短路电流的超短期智能辨识。该方法为智能电网中超短期智能预测提供了一种快速仿真建模(FSM)的新思路,为智能调度辅助决策提供有力的技术支持。通过IEEE30节点系统验证了该方法的可行性与有效性。

On account of the need of real-time monitor and alarm in smart grid, a new approach to ultra-short term intelligent forecasting of short circuit current is proposed. The extrapolation of grid state is conducted by ultra-short term load forecasting of nodes, and a short circuit current identification method based on GRNN is adopted to scan the short circuit current level of whole grid buses in short time, by which the intelligent identification is implemented. The approach provides a new idea of fast simulation modelling （FSM） in ultra-short time intelligent monitor, and a strong technical support for aid in decision （AID） of intelligent dispatch. The feasibility and validity of this approach is verified by the test on IEEE30 system.