Identification of Low-Order Unsteady Aerodynamic Models for Store Release in Transonic Flow
This work investigates two low-order modelling strategies, namely dynamic derivatives and impulse response functions (IRFs), for predicting the quasi-steady/unsteady aerodynamic loads on a store during release into a freestream and, ultimately, more efficient store certification. Computational fluid dynamics (CFD) simulations using an Euler formulation and overset-grid methodology are performed to generate training and validation data. Dynamic derivatives are identified through sinusoidal perturbations at varying amplitudes, frequencies, and mean angles of attack, while IRFs are identified using LASSO regression from step and multi-sine perturbations. The resultant models are evaluated against full-order CFD simulations, considering both forced vibration and store-release trajectories. The results show that dynamic derivatives provide strong performance for small-amplitude, low-angle-of-attack conditions but exhibit phase lead and reduced accuracy for nonlinear regimes. IRF-based models, particularly those identified from step responses, capture both continuous and discontinuous inputs with higher accuracy. Overall, the study demonstrates that while each modelling approach has its own advantages, the IRF-based methods provide a more robust linearised representation of store loads across a wider range of dynamic behaviors.
