This content is not included in your SAE MOBILUS subscription, or you are not logged in.

A Zonal Turbulence Modeling Approach for ICE Flow Simulation

Journal Article
2016-01-0584
ISSN: 1946-3936, e-ISSN: 1946-3944
Published April 05, 2016 by SAE International in United States
A Zonal Turbulence Modeling Approach for ICE Flow Simulation
Sector:
Citation: Krastev, V. and Bella, G., "A Zonal Turbulence Modeling Approach for ICE Flow Simulation," SAE Int. J. Engines 9(3):1425-1436, 2016, https://doi.org/10.4271/2016-01-0584.
Language: English

Abstract:

Turbulence modeling is a key aspect for the accurate simulation of ICE related fluid flow phenomena. RANS-based turbulence closures are still the preferred modeling framework among industrial users, mainly because they are robust, not much demanding in terms of computational resources and capable to extract ensemble-averaged information on a complete engine cycle without the need for multiple cycles simulation. On the other hand, LES-like approaches are gaining popularity in recent years due to their inherent scale-resolving nature, which allows the detailed modeling of unsteady flow features such as cycle-to-cycle variations in a DI engine. An LES requires however a large number of simulated engine cycles to extract reliable flow statistics, which coupled to the higher spatial and temporal resolution compared to RANS still poses some limits to a wider application of such methodology on realistic engine geometries. In this paper a hybrid zonal RANS/LES simulation methodology is proposed, based on a Detached Eddy Simulation (DES) technique previously developed by the authors. The aim is to preserve the turbulence scale-resolving capabilities wherever actually needed, reducing at the same time the overall computational costs compared to standard LES. The resulting zonal model has been implemented into the open-source CFD package OpenFOAM® and then assessed on geometries which reproduce the typical flow conditions at cylinders' intake ports. The presented numerical solutions have been compared with the available experimental benchmarks as well as with previous computational studies from other authors.