Prediction for onset of transonic shock buffet using steady thin-layer navier-stokes solver

Abstract

Transonic buffet produces large unsteady aerodynamic forces, which are undesirable for the aircraft performance limiting the cruising speed of transport aircraft and severely degrading the maneuverability of combat aircraft. Consequently, the flight envelope of transonic aircraft is restricted when a certain degree of buffet occurs. Therefore, it is quite important to be able to predict the buffet onset boundary at the earliest possible design stage of an aircraft. A new steady approaching method has been developed to predict theoretically the transonic buffet onset. In this thesis, the flow is assumed to be steady for the buffet onset. The present method involves the analysis of a distinct change in the variation of various aerodynamic parameters calculated from steady Navier-Stokes solver. These distinct changes indicate the onset of transonic buffet. In order to compute the transonic flow over airfoils and wings, the thin-layer form of compressible Navier-Stokes solver expressed in strong conservation-law form is solved using an implicit finite volume method. The numerical algorithm adopted is the upwind Roe’s FDS scheme for calculating inviscid flux. All viscous terms are discretized with second-order central differences. To increase the accuracy of the solution, TVD scheme based on MUSCL type approach with the minmod flux limiter is applied. The DADI scheme is used for the time integration. To treat the turbulent flow, the two-layer algebraic eddy viscosity turbulence model by Baldwin-Lomax is chosen in favor of computational robustness. In this thesis, firstly, the steady Navier-Stokes computations are performed on the NACA0012 conventional airfoil and BGK No.1 supercritical airfoil. Secondly, the steady Navier-Stokes computations are performed on a large swept wing with small aspect ratio. Using these results obtained from steady Navier-Stokes computations, the analysis of the variation of various aerodynamic parameters for predictin...