Predicting Emissions Across Design Variations of an Aero-Engine Combustor Using FGM and PISO

Abstract

A modeling approach for quickly predicting Carbon Monoxide (CO) and Unburned Hydrocarbon (UHC) emissions is assessed in Simcenter STAR-CCM+ version 2021.3. Both magnitude and trends are evaluated across design variations of an aero-engine test combustor operating at an idle condition. Large Eddy Simulation (LES) is used with the non-iterative Pressure Implicit by Splitting of Operators (PISO) scheme, and the Flamelet Generated Manifold (FGM) combustion model. There are four geometric design variations of the same test combustor where changes are made to the dome effusion cooling, main swirler, and downstream orifices. These four geometries are chosen for this study because they yield distinctly different CO and UHC emissions, thus reducing signal to noise in assessing the predictive capability of the modeling approach. Sensitivity of the results to a key parameter in the liquid fuel spray breakup model is provided. By varying the breakup rate, the prediction of the CO emissions is shown to compare very favorably both in magnitude and trend to the experimentally measured values. The UHC emissions are shown to compare well for three of the four designs. Results are also generated using the Semi-Implicit Method for Pressure Linked Equations (SIMPLE) scheme; these show the same behavior as PISO. However, the results with PISO can be obtained between 2.5 and 3.5X faster than the more traditional approach. Combining the computational efficiency of FGM and PISO allows for fast and accurate predictions of regulated emissions, and therefore down-selection of designs earlier in the design process.