Unmanned Aerial Vehicle Icing Test Flight Path Planning Method Based on a Rapid Prediction of Water Droplet Collection

Abstract Unmanned aerial vehicles (UAVs) are highly vulnerable to in-flight icing in complex meteorological environments, which poses a serious threat to flight safety. To address this challenge, a comprehensive framework integrating meteorological forecasting, rapid prediction of water droplet collection, and icing-tolerant trajectory optimization is proposed. First, a surrogate model combining proper orthogonal decomposition (POD) and Kriging interpolation is constructed based on high-fidelity icing numerical simulations of a NACA0012 airfoil, enabling fast and accurate prediction of water droplet collection. Next, high-resolution meteorological fields (temperature, pressure, liquid water content, and median volume diameter) are forecasted using the Weather Research and Forecasting (WRF) model. By integrating the developed rapid prediction method, the spatiotemporal distribution of icing severity is achieved. Finally, icing severity thresholds derived from FAR are incorporated into a particle swarm optimization (PSO) algorithm as risk constraints to generate icing-tolerant UAV test flight trajectories. Numerical results demonstrate that the proposed method effectively captures the spatiotemporal variability of icing severity and successfully plans optimal trajectories under both moderate and severe icing conditions. The approach significantly enhances the safety and feasibility of UAV natural icing flight tests.