Modeling probabilistic micro-scale wind field for risk forecasts of power transmission systems during tropical cyclones

Abstract

Tropical cyclones (TCs) pose significant risks to power transmission systems, causing extensive damage, widespread outages and severe socio-economic impacts. While reliable risk forecasting of these systems during TCs hinges on accurate wind predictions, operational numerical weather prediction (NWP) models struggle to deliver unbiased, high-resolution probabilistic wind-field forecasts necessary for infrastructure risk projections. This study introduces the Probabilistic Micro-Scale Wind-Field model (ProbMicro-WF) designed to enhance real-time hazard modeling for power system risk forecasts during TC evolution. This model improves NWP wind forecast by achieving the following: 1) probabilistic calibration and bias correction for NWP wind forecasts, leveraging historical TC observational data to improve prediction accuracy at high wind speeds; 2) terrain-modified statistical downscaling that translates mesoscale forecasts to micro-scale wind fields, capturing localized wind dynamics critical for tower- and transmission line-specific risk evaluation; and 3) a spatiotemporal stochastic model that preserves wind-field correlation structures, mitigating systemic underestimation of risk variance across geographically dispersed infrastructure during TC evolution. Finally, the ProbMicro-WF model is applied to the power transmission system in Zhejiang Province, China (105,500 km²) during Super Typhoon Lekima in 2019, highlighting its capability to simulate spatially coherent, high-resolution wind fields, enabling robust pre-event mitigation and real-time grid management in TC-prone regions.