Attention-LSTM-Based Power Optical Distribution Frame Cable State Hierarchical Pre-Warning Model

Aiming at the problem of the insufficient accuracy of pre-warning decision, this paper proposes a hierarchical pre-warning model based on Attention-LSTM prediction for the power optical distribution frame(ODF) cable. Firstly, the power ODF cable data prediction model is established based on the Attention-LSTM training network, and the power ODF cable loss prediction value is obtained through the SENet attention weighting and LSTM data prediction operation. Then, based on the K-means clustering model, the pre-warning and hierarchical decision of the operation status of power ODF cable is realized by using the power ODF cable loss data and multi-core parameters. Finally, the effectiveness of the proposed model is verified by experiments. The experimental results show that, compared with the comparison method, the prediction accuracy index of the proposed method is improved by 2% on average, which can effectively realize the hierarchical pre-warning, so as to more efficiently complete the operation and maintenance of power ODF cable.