Research on risk prediction model for voltage sag based on Elman neural network

Abstract

Voltage sag, as a prominent problem among many power quality problems, has always plagued power grids and users. To reduce the loss caused by the voltage sag and provide a basis for the prevention of the voltage sag, it is important to accurately and efficiently predict the risk of the voltage sag. This paper analyses the related factors of voltage sag risk from multi-directional and multi-dimensional, obtains the data of the influence factors of voltage sag risk in various information systems of the power grid, constructs voltage sag multi-source data, and uses Elman neural network to construct voltage sag risk prediction model. Through the training and learning of multi-source data, the model realizes the voltage sag risk assessment of each node in the power grid. Finally, the feasibility and validity of voltage sag risk prediction model are verified by case study of historical monitoring data in Fujian Province, which provides a method for voltage sag risk prediction.