(362g) Enforcing Elemental Mass and Energy Balances for Reduced Order Models Generated From CFD Simulations

Development of economically feasible gasification and carbon capture, utilization and storage (CCUS) technologies requires a variety of software tools to optimize the designs of not only the key devices involved (e., g., gasifier, CO₂ adsorber) but also the entire power generation system.  High-fidelity models such as Computational Fluid Dynamics (CFD) models are capable of accurately simulating the detailed flow dynamics, heat transfer, and chemistry inside the key devices.  However, the integration of CFD models within steady-state process simulators, and subsequent optimization of the integrated system, still presents significant challenges due to the scale differences in both time and length, as well the high computational cost.  A reduced order model (ROM) generated from a high-fidelity model can serve as a bridge between the models of different scales.  While high-fidelity models are built upon the principles of mass, momentum, and energy conservations, ROMs are usually developed based on regression-type equations and hence their predictions may violate the mass and energy conservation laws.  A high-fidelity model may also have the mass and energy balance problem if it is not tightly converged.  Conservations of mass and energy are important when a ROM is integrated to a flowsheet for the process simulation of the entire chemical or power generation system, especially when recycle streams are connected to the modeled device.  As a part of the Carbon Capture Simulation Initiative (CCSI) project supported by the U.S. Department of Energy, we developed a software framework for generating ROMs from CFD simulations and integrating them with Process Modeling Environments (PMEs) for system-wide optimization.  This paper presents a method to correct the results of a high-fidelity model or a ROM such that the elemental mass and energy are conserved perfectly.  Correction factors for the flow rates of individual species in the product streams are solved using a minimization algorithm based on Lagrangian multiplier method.  Enthalpies of product streams are also modified to enforce the energy balance.  The approach is illustrated for two ROMs, one based on a CFD model of an entrained-flow gasifier and the other based on the CFD model of a multiphase CO₂ adsorber.