Quantitative assessment method of typhoon-induced photovoltaic damage and energy production losses: a case study of the 2024 Typhoon Yagi

Climate change has intensified the threat of typhoons to photovoltaic (PV) infrastructure. We present a quantitative assessment method to conduct typhoon-induced PV infrastructure loss assessment. Firstly, we use positive-unlabelled learning random forest (PUL-RF) classifier to generate pre- and post-typhoon PV distribution maps. Then, typhoon consequences are evaluated across three dimensions: physical damage, immediate economic costs, and energy production, providing complete impact assessment for PV infrastructure. This method was applied to assess typhoon-induced PV damage and energy production losses across Hainan Island during Super Typhoon Yagi (2024, maximum wind speeds > 58 m/s). We achieved a classification accuracy (IoU) exceeding 82%, revealing a 3.51% island-wide PV infrastructure damage, more than 5% of PV power plants suffered over 35% damage. Wind scale and PV damage rate showed a linear relationship, with damage prevalence increasing 3% per wind level increment. The direct damages at CNY 253.4 million, with indirect losses totalling 301.5 MW in power output and CNY 603.7 million in revenue (from September 6, 2024 to January 7, 2025). We combine remote sensing, spatial damage and economic modelling to quantify physical damage and indirect economic impacts of typhoons on PV, enabling accurate assessment of disaster-related power disruptions to strengthen renewables' climate resilience.