Supervised learning-based automatic fault classification method for power communication systems

With the expansion of the power grid scale and the improvement of the degree of intelligence, the demand for fault diagnosis of power communication systems is becoming increasingly urgent. However, existing fault diagnosis methods have many limitations, such as dependence on the experience of operation and maintenance personnel, inefficiency, and lack of systematic and standardised restoration measures. To address these issues, this paper proposes a cascade model based on TF-IDF, random forest and multilayer perceptron (MLP) for automatic classification of power communication system faults. Firstly, the TF-IDF algorithm is used to extract features from the fault text and transform the text into a vector representation; secondly, the Random Forest model is used to classify the fault text to achieve the differentiation of the four fault categories; finally, the classification results are further optimised through the introduction of the MLP layer to enhance the correction capability of the model. The experimental results show that the model performs well in the fault classification task with an accuracy rate of 91.1%, which is significantly better than the traditional method. In addition, this paper constructs the fault knowledge graph of electric power communication system by taking the classification results as entity nodes and relationship edges, which not only improves the construction efficiency of the fault knowledge graph, but also provides strong support for fault diagnosis and root cause localisation.