摘要
Abstract
To improve the accuracy of OPGW cable remaining life prediction,multi-factor meteorological data from Guangzhou between 2011 and 2023 are utilized to investigate cable degradation under temperature fluctuations and wind loads.A time series of cable length variation is constructed based on thermal expansion and contraction and mechanical load models,and an Autoformer-based remaining life prediction method is developed.Lagrange interpolation is applied to repair anomalous values,and Min-Max normalization is used for feature scaling.Comparison with LSTM,Bi-LSTM and Bi-LSTM+Attention models shows that Autoformer achieves optimal performance across RMSE,MAE,and MSE metrics,demonstrating a superior capability in capturing long-term variation trends of cable remaining length.The findings indicate that the proposed method enhances prediction accuracy and provides data-driven support for OPGW operation,maintenance,and replacement decision-making.关键词
OPGW/寿命预测/Autoformer/时间序列预测Key words
OPGW/remaining life prediction/Autoformer/time series forecasting分类
信息技术与安全科学
