Skip to main content


Scaling-up Predictive Engine Simulations Through High Performance Computing

Company Name: Convergent Science
Program Office: Vehicles
Location: Madison, WI
Email: Peter Kelly Senecal, Co-Owner & Vice President;
Award Amount: $92,000
Project Term: 12 months
Project Status: Active
Participating Lab(s): Argonne National Laboratory


At the heart of every internal combustion engine, complexity reigns supreme. Valves and pistons lunge up and down at 25 meters per second, pressure spikes in an instant, and sprays of fuel spread throughout the maelstrom in incredibly intricate patterns. That complexity is daunting for anyone trying to improve engine design.

With increasingly stringent emission regulations and a growing demand for better fuel economy, engine manufacturers and energy companies are pursuing novel technologies that simultaneously improve efficiency and reduce emissions. In order to create market-ready products that incorporate these complex new technologies, engineers need a fundamental understanding of the in-cylinder combustion processes and how these processes are affected by engine geometry, spray characteristics, and fuel chemistry.

Computational fluid dynamics (CFD) simulations can illuminate these fundamental processes, but both the accuracy and speed of CFD simulations must be improved. Higher accuracy is required so that engineers can use CFD for predictive simulations rather than simply to confirm experimental results, while reduced computational time is required to meet the increasingly tight design schedules of automotive companies.

Convergent Science is working to reduce the simulation time needed to achieve acceptably accurate results. Through the SBV pilot, the company will have access to high performance computing (HPC) tools at Argonne National Laboratory, which will allow it to test and refine its code. The work can play an important role in product development by reducing the cost of design and optimization studies for new and existing products.


Convergent Science's software is a state-of-the-art CFD code that eliminates all user meshing time by automatically generating the mesh – a grid that defines a geometric area to be examined - at startup. By contrast, other codes require time-consuming manual mesh generation or require specific templates or scripts for their automated meshing features. In addition to automatic runtime mesh generation, the company's code uses its unique adaptive mesh refinement technology to automatically adjust the mesh throughout the simulation. This development helps users optimize the accuracy and speed tradeoffs inherent in CFD simulation. In addition to its significant meshing advantages, the code will allow the company to optimize the spray and combustion models through the lab's HPC capabilities. Finally, the company's software offers a robust detailed chemistry solver that presents a unique combination of multizone modeling, dynamic mechanism reduction, and simulation of an unlimited number of species.



Automotive suppliers and manufacturers are looking for new ways to improve engine efficiency and these companies are leading employers in the advanced manufacturing field. Further, when consumers spend less money on gasoline, they spend more money bolstering local economies.


Gasoline engines convert about 20 percent of the fuel they burn into energy. Increasing engine efficiency reduces gasoline use and helps avoid heat-trapping emissions that cause climate change as well as conventional air pollutants.


The U.S. automotive fleet consumes 385 million barrels of gasoline a day, on average, much of which is imported through global oil markets. Improving fuel economy is the most effective way for the country to reduce its oil dependence.