Buffet Onset Prediction With the Linearized Frequency-Domain Method

Transonic buffet is a flow instability associated with shock-wave oscillation and boundary-layer separation that can cause unsteady aerodynamic loads, structural fatigue, and degraded aircraft performance. This paper presents a method to predict buffet onset using a linearized frequency-domain computational fluid dynamics approach. A reduced state-space aerodynamic model is constructed from frequency-domain data using the Loewner framework, and buffet stability is determined from its eigenvalues. Buffet onset is identified when the real part of the dominant fluid eigenvalue changes sign and becomes positive. To represent buffet conditions, the realization is allowed to retain physically meaningful unstable fluid poles, while a residue-based criterion is used to distinguish them from spurious poles introduced by interpolation. In combination with a projection of the aerodynamic outputs based on proper orthogonal decomposition (POD), the reduced model also provides the corresponding eigenvectors, enabling reconstruction of their spatial structure in physical space in addition to onset prediction. The approach is demonstrated for the OAT15A airfoil and the NASA Common Research Model. Results agree well with reference global stability analysis and show that the combined Loewner–POD framework robustly extracts the dominant buffet-driving fluid modes. The method provides an efficient reduced-order alternative to direct global stability analysis and can be extended toward combined flutter and buffet prediction.