Eventos Anais de eventos
ENCIT 2022
19th Brazilian Congress of Thermal Sciences and Engineering
COMPARISON OF RANS, DES AND LES TURBULENCE MODELS TO DETERMINE DISCHARGE COEFFIENTS OF AN ENGINE CYLINDER HEAD
Submission Author:
Jean Fagundez , RS
Co-Authors:
Cristian Douglas Rosa da Silva, Jean Fagundez, Vinícius Roso, Thompson Diordinis Metzka Lanzanova, Mario Martins
Presenter: Jean Fagundez
doi://10.26678/ABCM.ENCIT2022.CIT22-0567
Abstract
In view of recent international agreements on reducing greenhouse gas emissions, some groups advocate that the only solution is the total replacement of internal combustion engines by electric vehicles during the next few years. However, this change is a questionable alternative since the world energy matrix has high levels of non-renewable sources. A sustainable transition from internal combustion engines to hybrid and electrified solutions could be achieved with advanced combustion strategies and renewable fuels utilization. In this sense, the use of computational fluid dynamics (CFD) models is an important ally in the study of microscale phenomena, such as turbulence and its effects on combustion and engine efficiency. This paper intends to show the difference in the modeling and validation of a diesel engine cylinder head in a flow bench using three different turbulence models, from the lowest to the highest complexity in the analysis of turbulence microscales: Reynolds Averaged Navier Stokes (RANS), Detached Eddy Simulation (DES) and Large Eddy Simulation (LES). For this purpose, the inlet and exhaust valves were analyzed with forward and reverse air passes through several lift positions, from 1 to 10 mm, as well as the study of mesh independence for each case study. The discharge coefficients found by CFD modeling, when compared with those determined experimentally by the flow bench, showed that the RANS, DES and LES models reached similar results for valve lift values above 6 mm, but only DES and LES were able to determine air flow values closer to the experimental ones in cases where there was greater air restriction and, consequently, greater turbulence. Thus, for cases where the flow has greater turbulence intensity, the LES model was the most suitable model for predicting the experimental values, while the RANS model had its most suitable application for situations where lower mesh detail was needed or the flow presented lower turbulence intensity.
Keywords
Reynolds averaged Navier Stokes(RANS), Detached Eddy Simulation, Large Eddy Simulation (LES), Discharge Coefficient
