N-HiTS Based Tree Barrier Distance Forecasting Method for Enhancing Safety of Electrical Systems

The rapid growth of trees in proximity to overhead electrical lines poses a significant risk to the safe and reliable delivery of electricity services by utility companies. The challenge lies in accurately monitoring and predicting the distance between power lines and surrounding trees, as this distance is crucial for preventing power outages and maintaining system integrity. In this paper, we propose a novel distance forecasting method based on the N-HiTS algorithm, which effectively captures the complex temporal dynamics of tree growth patterns. Specifically, our method leverages hierarchical time series forecasting to enhance the prediction accuracy of tree-barrier distances, integrating historical growth data and environmental factors that influence tree development. Experimental results demonstrate that the N-HiTSbased method outperforms traditional forecasting techniques in terms of accuracy and reliability. The results indicate a significant reduction in forecasting errors, providing utility companies with a valuable tool for proactive vegetation management and improved safety of electrical systems.