Parametric Modeling of Multi-Injector Rocket Combustors using Component-based Reduced-Order Modeling Framework

Even with the most advanced computational capabilities, high-fidelity (e.g., large-eddy) simulations of full-scale rocket engines remain far out of reach. In this work, a component-based reduced-order modeling (CBROM) framework is established to enable accurate and efficient parametric modeling of full-scale rocket engines, which are decomposable into different components, including injectors, combustor, and nozzle. These components can be modeled using either reduced-order models (ROMs) or reduced-fidelity full-order models (RF-FOMs). Specifically, ROMs can be trained for individual injectors through component-based training strategies with system-level responses mimicked via external boundary forcing, which only require high-fidelity simulations of a much smaller computational domain, therefore significantly reducing the costs of ROM training. The trained component-based ROMs are then integrated into the CBROM framework to represent identical injectors and are coupled with other models to enable full-scale rocket engine modeling. ROMs are constructed via hyper-reduced model-form preserving least-squares projections with variable transformation (MP-LSVT) formulation~\citep{HuangMPLSVT2022} enhanced with basis and sampling adaptation. The CBROM framework is evaluated based on a multi-injector model rocket combustor configuration exhibiting different self-excited combustion dynamics with geometric variations. The framework is demonstrated to provide accurate parametric predictions of the changes in dynamic behaviors, including the spectra from dynamic mode decomposition (DMD) analysis, time-averaged, and RMS fields of target state variables.