A multifidelity framework oriented toward efficient modeling of combustion dynamics that integrates a reduced-order model (ROM) for the combustion response into the Euler equations is proposed. The ROM is developed from computational fluid dynamics (CFD) simulations of a combusting flow that is periodically forced at the boundaries of a reduced domain. Galerkin’s method is used to reduce the high-order partial differential equations to a low-order ordinary differential equation system via proper orthogonal decomposition eigenbases generated from a reduced-domain dataset. Evaluations of the framework are performed based on parametric studies of a simplified test problem for a model combustor showing distinguishable combustion instability behavior. Two-way information transfer between the ROM and Euler solutions is accomplished by interface matching at the boundaries of the reduced domain. It is shown that accurate predictions require the use of multiple ROMs to account for both upstream- and downstream-traveling perturbations. Characteristic boundary conditions are required at both reduced-domain boundaries to minimize wave reflections and preclude generic unstable responses in the multifidelity model predictions. Comparisons with CFD solutions show the multifidelity model is capable of capturing overall instability trends.
Multifidelity Framework for Modeling Combustion Dynamics
AIAA Journal ; 57 , 5 ; 2055-2068
2019-01-30
14 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
MULTIFIDELITY MODELING FRAMEWORK FOR BAYESIAN-BASED MULTIDISCIPLINARY AIRCRAFT DESIGN OPTIMIZATION
British Library Conference Proceedings | 2012
|