A Real-Time Multi-Source Meteorological Data Integration Framework for Advanced Lightning Risk Detection and Protection in Wind Power Plants

Abstract

The rapid expansion of wind power plants (WPPs) has increased their exposure to lightning-related risks, which are further intensified by climate variability and rising atmospheric instability. Lightning strikes remain one of the most frequent and damaging hazards, leading to turbine blade degradation, system failures, and significant economic losses. This study aims to develop an integrated, real-time framework for lightning risk detection and protection by combining multiple meteorological data sources. The proposed system incorporates satellite observations, ground-based Internet of Things sensors, remote sensing technologies such as light detection and ranging and radar, and historical lightning datasets to provide a comprehensive atmospheric assessment. Data are processed through a unified architecture using embedded computing platforms and analyzed using machine learning techniques, including logistic regression and gradient boosting, to classify lightning types and generate a composite risk index. The system enables automated alerts and protective responses, such as turbine shutdowns, when risk thresholds are exceeded. Experimental results demonstrate that the framework achieves high predictive accuracy with a response latency of less than three seconds, allowing timely identification of cloud-to-ground, intra-cloud, and upward lightning events. The modular, cost-effective design supports scalable deployment across varying wind farm capacities and operational contexts. The findings indicate that integrating heterogeneous meteorological data significantly enhances the reliability and responsiveness of lightning protection systems compared to conventional single-source approaches. This study provides a practical, adaptable solution to improve the safety, resilience, and operational efficiency of wind energy infrastructure in increasingly volatile climatic conditions.