Evaluation Method of Low-Carbon Communities Distribution Facilities Based on BiTCN-SE Model

This paper presents a method to evaluate the condition of power distribution facilities in low-carbon communities. It starts by preprocessing operational data using the moving average line and SMOTE method to enhance the dataset's diversity. The main fault types and their risks are then classified, and an index system is built using the AHP-entropy weight method for more accurate evaluation. Finally, the BiTCN-SE model, incorporating a bidirectional spatio-temporal convolutional network and SE attention mechanism, is applied to improve the accuracy of evaluating the operational state of distribution facilities, exceeding 94 %. This ensures the safe and stable operation of these facilities.