Tilting rotor aerodynamic performance prediction based on vortex particle method coupled with large eddy simulation

Tiltrotors are crucial in a low-altitude economy, combining the advantages of fixed-wing aircraft and rotorcraft. However, traditional numerical simulation methods are not accurate enough when simulating the transition state of a tiltrotor, especially under large tilting angles. This paper introduces a novel meshless computational framework that integrates a reformulated vortex-particle method (rVPM) with large-eddy simulation (LES) and a machine-learning-based airfoil model to predict tiltrotor performance across all flight phases. To address rVPM’s accuracy degradation at extreme tilting angles, a hybrid aerodynamic model is developed: neural networks trained on high-fidelity pre-stall data are combined with Truong’s analytical deep-stall formulation, seamlessly blended via a logistic-based confidence metric. Validation against wind-tunnel measurements of the TUD_F29 rotor at tilt angles of 0°, 30°, 60°, and 90° demonstrates thrust and power predictions within 5% of experimental data. A chord-distribution sensitivity analysis further reveals the method’s capability to resolve enhanced solidity effects, including increased thrust and induced power losses with larger chord lengths. By unifying a stable, meshless LES solver with data-driven airfoil modeling, this framework offers an efficient, high-fidelity tool for designing and optimizing next-generation electric vertical take-off and landing tiltrotor designs.