Fault classification method of alarm information based on TextCNN

Abstract:
With the continuous development of smart grid informatization, the alarm information data generated by the power grid continues to increase. How to use these text data to ensure the stability of the power grid is the main goal of the current power grid. This paper proposes an alarm information-oriented fault classification method based on TextCNN, which realizes the end-to-end discrimination of power grid fault types and promotes the development of power grid intelligence and stability. Firstly, the text of the alarm information is vectorized through the Word2vec model; then the TextCNN model is built to construct a power grid fault classification model based on TextCNN to judge the type of power grid faulty equipment and whether it has protection and whether it is a complex fault; finally, it is verified through the test set. The method has good expressivity in fault classification.