Construction of health status recognition and prediction model for power communication equipment based on TFIDF-COS

When power communication equipment malfunctions, the stability and safety of the power grid are compromised. This is due to the health status of the equipment. The safety and stability of the power grid will be impacted if dispatchers take too long to identify the fault’s origin and kind, which will disrupt the power communication system’s regular operations. In order to solve the problem of poor health status of power communication system due to the inability to timely determine and deal with the faults of power communication equipment, the study proposes the construction of health status recognition of power communication equipment with prediction model based on term frequency-inverse document frequency and cosine similarity. The model firstly extracted the fault information of power communication equipment and builds the fault knowledge graph. Secondly, the study identified and built a prediction model for the health status of power communication equipment based on term frequency-inverse document frequency and cosine similarity model. The outcomes revealed that the training model had the highest accuracy and the lowest loss rate when the learning rate was set to 1 × 10⁻⁵. When the iterations was set to 70, the training and test sets had the highest accuracy and the lowest loss rate. When the model utilized in the study was compared to other models with varying numbers of samples in the dataset, it performed well in terms of runtime and fault diagnosis accuracy. The model developed by the study improves the accuracy of fault extraction and recognition and can better ensure the normal operation of power communication equipment.