Lifetime Prediction of Underground Cable Based on GA-BP Neural Network

As China’s urbanization progresses at an accelerated pace, cross-linked polyethylene (XLPE) cables are playing an increasingly important role as the electrical energy transmission. However, issues of aging and faults due to prolonged operation frequently arise, making it a critical challenge to ensure the safety and reliability of cables in the maintenance of power equipment. Therefore, this paper conducts research on the lifespan prediction of XLPE cables based on deep learning. The study utilizes a publicly available underground cable detection dataset and employs a Genetic Algorithm (GA) to optimize the Back Propagation (BP) neural network for predicting the cable’s lifespan. correlation analysis and lifespan distribution analysis are applied. Important factors influencing lifespan are selected and preprocessed. The GA-BP algorithm is employed to establish an underground cable lifespan prediction model, introducing three evaluation indicators—Mean Squared Error (MSE), Mean Absolute Error (MAE), and Coefficient of Determination (R
²
)—to assess the model’s predictive performance. The results are compared with the predictions of a BP neural network. The findings indicate that the GA-BP outperforms the BP in all evaluation indicators, confirming that the improved model demonstrates better predictive capabilities. This provides valuable insights for subsequent applications in cable lifespan prediction.