Abstract
As engine technologies become increasingly complex and engines are driven to new operating points, understanding transient phenomena is important to ensure reliable engine operation. Unlike Reynolds Averaged Navier-Stokes (RANS) studies that only provide cycle-averaged information, Large Eddy Simulation (LES) studies are capable of simulating cycle-to-cycle dynamics. In this work, a finite difference based structured methodology for LES of IC engines is presented. This structured approach allows for an efficient mesh generation process and provides potential for higher order numerical accuracy. An efficient parallel scalable block decomposition is done to overcome the challenges associated with the low ratio of fluid elements to overall mesh elements. The motion of the valves and piston is handled using a dynamic cell blanking approach and the Arbitrary Lagrangian Eulerian (ALE) method, respectively. Modified three-dimensional Navier-Stokes Characteristic Boundary Conditions (NSCBC) are used in the simulation to prescribe conditions in the manifolds. The accuracy of the simulation framework is validated using various canonical configurations. Flow bench simulations of an axisymmetric configuration and an actual engine geometry are done with the LES methodology. Simulations of the gas exchange in an engine under motored conditions are also performed. Overall, good agreement is obtained with experiments for all the cases. Therefore, this framework can be used for LES of engine simulations. In the future, reactive LES simulations will be performed using this framework.
Highlights
With the ever-rising need for better fuel efficiency and lower emissions, development of improved engine technologies is critical
In a parallel framework based on SIMD such as domain partitioning with the Message Passing Interface (MPI) library, a large ratio of solid cells can lead to severe load imbalance among the processors, resulting in overall reduced computational efficiency
The mesh location is moved in a Lagrangian manner, and the effect of this mesh motion is included in the original Eulerian equation as an additional flux
Summary
With the ever-rising need for better fuel efficiency and lower emissions, development of improved engine technologies is critical. The advantage of LES lies in the direct computation of the large scale flow structures that enables a better description of the turbulence and its interaction with chemistry. If a structured framework is used, higher order finite difference schemes with discrete mass, momentum and energy conservation can be used [6] Use of these schemes ensures that the turbulent energy is not artificially dissipated and small flow structures are retained, leading to a more accurate description of turbulent mixing. A structured domain ensures efficient parallel communication which is essential for scalability to a large number of processors Despite these advantages, the structured framework has not been extensively used for LES of engines due to various challenges. Various validation cases of the framework are presented including flow bench simulations and gas exchange simulations
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
