Computational Models for Nonlinear Aeroelastic Systems

Clear Science Corp. and Duke University propose to develop and demonstrate a new and efficient computational method of modeling nonlinear aeroelastic systems. The method will extend the scope of multi-disciplinary computational tools like NASA Dryden's STARS by augmenting linear eigenmode stability algorithms and coupled time-marching techniques. The objective is a low-dimensional model that accurately reflects nonlinearity in both structure and fluid and that is efficient enough to permit full exploration of parameter space. In Phase I, our team will evaluate two types of model order reduction: proper orthogonal decomposition of the coupled-system variables and the method of harmonic balancing. We will downselect one method based on efficiency, accuracy, and versatility, demonstrate its merit via a prototype problem, and design a comprehensive Phase II plan for model development and testing. The proposed innovation can minimize the risk of failure and maximize flight safety in aircraft like the F-18-AAW and X-43 by accurately and efficiently predicting nonlinear dynamics over a broad range of flight conditions. Integrating the nonlinear model with codes like STARS will augment the capability of quickly determining linear stability with the capability of efficiently analyzing nonlinear behavior like limit cycle oscillations, hysteresis, higher harmonic and sub-harmonic resonances, jump resonances, entrainment, beating, and period doubling.