Lightning Risk Assessment Method Considering Environmental Factors Based on Artificial Intelligence Algorithms

With the continuous extension of transmission lines, the risk of lightning strikes on transmission lines has significantly increased, especially in areas with complex terrain. Additionally, as the construction of transmission lines encounters more complex terrain, the influencing factors of lightning strikes have become more diverse. This paper proposes a lightning risk assessment method based on Particle Swarm Optimization and Random Forest (PSO-RF). By utilizing ADTD lightning data and geographical information of the study area, the model evaluates high-risk lightning zones. The results indicate that the PSO-RF model outperforms other machine learning models in terms of accuracy, precision and recall, demonstrating excellent performance on the test set. SHAP value analysis reveals that features such as latitude, longitude, lightning frequency and building height significantly influence lightning risk. This study provides guidance for differentiated lightning protection measures in power system transmission lines.