Rapid Computational Aerodynamics for Aircraft Gust Response Analysis

Abstract

Computational engineering methods play a more and more important role in building aircraft that move people and goods. Particular in high-speed civil air transport increased usage of higher fidelity simulation tools are expected to enable greener designs with a reduced environmental footprint. The challenge of including computational fluid dynamics in aircraft loads and aeroelasticity is addressed herein. During the aircraft design and certification process a tremendous number of dynamic responses to atmospheric turbulence need to be analysed. Current industrial loads computations are based on corrected linear potential flow methods which offer fast predictions but suffer several drawbacks once aerodynamic non-linearities occur. Instead, aerodynamic loads offered by computational fluid dynamics are highly accurate also at these non-linear conditions. However, computational cost necessary for performing time-marching simulations makes these methods prohibitive for unsteady loads in an industrial context. This work addresses how to efficiently introduce computational fluid dynamics based aerodynamics during gust loads analysis. It is shown that using frequency domain methods in conjunction with reduced order modelling techniques based on modal decomposition and projection offer accurate models which can be analysed at low cost. The three requirements of such an industrial gust loads process are, first, the need for high accuracy, secondly, a significant reduction of runtime compared to unsteady full order time-marching simulations, and thirdly, the ability to automatise the generation and solution process of the reduced model as well as the design and certification process. Therefore, the linearised frequency domain method is extended towards gust responses by altering the right-hand side forcing term. An aerodynamic reduced order model is constructed by computing a modal basis using proper orthogonal decomposition and projecting the linearised equations afterwards. Finally, a coupled aeroelastic model is obtained by combining the aerodynamics model with eigenmodes of the coupled Jacobian matrix for the structural vibration and projecting the coupled linearised equations. The final small sized aeroelastic model enables the inclusion of highly accurate loads during time-critical gust loads analysis and provides the opportunity to introduce these loads in a wider multidisciplinary context. Thus it is a substantial step towards establishing computational fluid dynamics for unsteady aeroelastic analysis.