A knowledge representation based recognition method on power grid alarm monitoring data

Abstract

The identification of power grid monitoring and alarm information is an essential guarantee for the stable operation of the power grid. Identifying current power grid alarms is based on an expert rule system. This method has the problems of cumbersome rule matching, difficulty in maintaining the rule base, and low matching efficiency. In order to solve the above problems, this paper proposes a method of identifying power grid monitoring and warning information based on knowledge representation. The method first uses grid topology information to construct a knowledge graph. It then uses knowledge representation learning algorithms to represent the power knowledge graph which digitally represents the textual warning information, and finally uses a series of deep learning models to identify ”alarm-event.” This paper uses six algorithms, such as TransE for knowledge representation, combined with three deep learning models such as ConvLSTM for identification. The verification shows that this method has specific ”alarm-event” identification capabilities based on actual power grid monitoring data.