Convergent Science simplifies CFD meshing for engine combustion modeling

  • 25-Jun-2010 02:11 EDT

Simulation of temperature contours in a spark-ignition, direct injection engine. (Michigan State University)

For engine combustion modeling, many will agree creating a mesh is a problem. “When we started out as a consulting company over 10 years ago, helping people with meshes was the main aspect of our business,” explains Dr. Daniel Lee, Business Development Manager for Convergent Science. “Not only did it dominate most of the labor time it took to perform an analysis, it also dominates the accuracy.”

For accurate results, a mesh needs both high-quality elements and enough of them in places where important changes are happening to resolve and accurately predict temperatures, pressures, and velocities, according to Lee. “Adapting a mesh is also difficult—few analysts want to change a mesh once it works, even to perform a grid refinement study.”

Enter CONVERGE, the CFD software the engineers at Convergent Science Inc. built to alleviate these issues. It creates—automatically at run time—an orthogonal mesh composed of cubic elements from an STL file input. An STL file is a triangulated approximation of surfaces, a standard output from most CAD systems. After the software generates a mesh, it simulates in-cylinder flow, spray, and combustion using the Finite Volume computational method. “Users no longer create meshes through point-and-click methods,” Lee said.

Merely generating a mesh may not be its biggest attraction. In successive steps, CONVERGE also optionally, locally refines the initial mesh, subdividing cubic elements where gradients are high. This refines the accuracy of the computation only in areas where needed, using a novel cut-cell technique. One can only imagine the complex bookkeeping required to maintain connections between subdivided cells and neighboring larger cells. The cells also remain body-fitted to the triangulated surface provided as input, adjusting to the boundaries even as the mesh is refined.

This is not to say a user does not have control over the mesh. As Lee explains, a user controls four ways how the software constructs the mesh:

• The user defines the base mesh size, using knowledge of the phenomena and factoring in how much computing resource they have.

• Optionally, embedding strategies allows the user to specify areas the solution needs more cells, such as valve seats, spray injectors, or pistons.

• With adaptive mesh refinement (AMR), the user sets the parameters on gradients that will determine when a cell will be subdivided. He/she can choose AMR based on velocity, temperature, or chemical species either singularly or in combination. This feature is optional and can be enabled when the user wishes to increase accuracy yet maintain reasonable run times.

• Finally, the entire base grid size can be scaled as a function of time. Lee reports this useful for diesel engine as well as steady state simulations.

When AMR is used, the user sets a budget for the upper limit of cells so as not to overwhelm computer resources. While some might initially not trust a variable grid, its impact on computing run times can be dramatic. “We did an experiment where we took the smallest cell from a variable grid generated from AMR and created a whole grid to that smallest cell,” explains Lee. “We found that the results were almost identical and the AMR grid ran 28 times faster.”

The full Navier-Stokes equations are solved along with the user’s choice of turbulence model. A chemistry solver, SAGE, was developed by Convergent Science and is used to model kinetically limited phenomena such as flame propagation, ignition, knock, and pollution formation. “SAGE is ideally suited for HCCI analysis,” says Lee.

CONVERGE includes diesel, spark ignition, and homogeneous-charge compression ignition (HCCI) physical models. Users can also create their own combustion models. Advanced submodels for sprays, wall film, turbulence, ignition, combustion, and emissions are included. The code is fully parallelized with automatic domain decomposition for mapping to multiple CPUs.

“Analysis is a critical part of our development process, especially for combustion and aftertreatment system development,” says Dr. Yangbing Zeng, Technical Specialist, Airflow and Combustion Analysis at General Motors Powertrain. Using CAE analysis significantly cuts down the development time and material cost of developing new engine hardware, according to Zeng. He points to the unique mesh approach of CONVERGE as shortening analysis time.

Achates Power, an engine developer, is also a user of CONVERGE that found it speeded development time. “The analysis process accelerated significantly thanks to the elimination of mesh development time, automatic mesh refinement (AMR) during the calculation, and parallel computation. The quality of the results is also impressive,” says Fabien Redon, VP of Performance and Emissions Development at Achates Power.

HTML for Linking to Page
Page URL
Rate It
4.47 Avg. Rating

Read More Articles On

Freudenberg Sealing Technologies has expanded its LESS (Low Emission Sealing Solution) lineup to include new products designed to address challenges associated with powertrain friction, smaller spaces, lighter weight vehicles and growth in the electric mobility vehicle arena.
At least for this moment, all is right with the world as Mercedes-Benz reinstates a new-age version of its iconic inline six-cylinder engine.
Tanktwo, a Finland-based startup company is rethinking the basic battery cell and challenging the fundamental economics and operational assumptions of EVs. The ingenious concept is worth engineers' attention.
Conti’s 48-V system will be standard equipment on both gasoline and diesel versions of the Scenic Hybrid Assist model. It is the first of multiple 48-V production announcements coming over the next few years.

Related Items