Improved Solution of Nonlinear Reduced-Order Models for Static and Dynamic Response Prediction

Abstract

Nonlinear reduced-order models (NLROMs) have been used recently to predict the response of stiffened shell structures subjected to extreme acoustic loading and aerodynamic heating. The NLROMs are essentially modal models augmented with nonlinear terms to capture the large amplitude, nonlinear geometric effects. Temperature dependent linear stiffness terms can be added to capture thermal effects. Numerical stability issues have risen for quasi-static and dynamic simulations using the NLROMs. Most of these issues have been solved through the use of very high sample rates. However, these stability issues are more severe with thermal models of curved structures with softening effects. In this paper, the nonlinear terms of the reduced-order models will be re-cast in an alternate form that allows for more robust solution techniques. An example problem with a curved panel subjected to a thermal loading shows the benefits of the improved solution procedures.