iSGTW is now Science Node Learn more about our evolution

  • Subscribe

VERIFI advances combustion engine research

Combustion engine framework. Image courtesy Argonne National Laboratory.

The automotive industry is just one of many markets reliant on gasoline- or diesel-fueled engines. From heavy-duty trucks and buses, to off-highway vehicles for specialized agriculture and marine markets, the internal combustion engine is an integral part of commerce in a modern world.

According to Doug Longman, manager of engine combustion research at Argonne National Laboratory (Argonne) near Chicago, Illinois, US, ongoing and future changes in requirements and regulations mean engine manufacturers are looking for "tools to shrink design timelines and costs." Some manufacturers have already incorporated combustion engine simulations into the design cycle. However, to keep up with regulatory tailpipe emissions and Corporate Average Fuel Economy (CAFE) standards - as well as customer demands - he says many will need advanced solutions.

At the 2014 Society of Automotive Engineers (SAE) World Congress in Detroit, Michigan, US, Argonne launched the Virtual Engine Research Institute and Fuels Initiative (VERIFI). Argonne mechanical engineer and computational fellow Sibendu Som notes VERIFI is the first environment to offer "high-fidelity, 3D, end-to-end combustion engine simulation and visualization tools capable of simultaneous powertrain and fuel simulation, and uncertainty quantification."


The computational fluid dynamics (CFD) modeling required to accurately simulate an internal combustion engine is intricate and complex, Som explains. Overall, "we need to accurately capture the two-phase flow by solving for liquid fuel spray and its vaporization characteristics." As gasoline or diesel fuel evaporates, it mixes with ambient air to form pockets of ignitable mixtures. Combustion chemistry, says Som, is critical to accounting for aspects such as auto-ignition, how the flame propagates, oxidation of the mixtures, and formation of emissions.

The 50-million peak cell count simulation was performed on the high-performance computing clusters at Argonne National Laboratory, using 512 computational cores for a wall-clock time of more than 2 weeks. Video courtesy Argonne National Laboratory and VERIFI.

"Combustion chemistry becomes even more critical," he notes, "when dealing with alternative fuels" like diesel, bio-diesel, ethanols, and mixtures of each. "At VERIFI," Som adds, "we can develop chemical kinetics models for these different fuel types." Researchers are also working with experimentalists to model and validate one of the most difficult and fundamental problems in classical physics - turbulence (in this case, engine turbulence).

Physics also comes into play when modeling nozzle flow. "Nozzle flow governs the way the fuel spray emerges and mixes with ambient air, and so physics-driven modeling is essential," says Som. No combustion engine is 100% efficient; emissions are the final piece of the equation, and VERIFI researchers are creating advanced emission models for both soot and nitrogen oxides.

Achieving results

In general, many engine simulations are not grid convergent, meaning the computed results change if the CFD mesh size or makeup changes. The ultimate aim is a result that is independent of the mesh. Finer and finer meshes solve more and more of the flow field, but at a certain point - grid convergence - further mesh refinement has no impact on the results.

Grid convergence is most often limited by computational power. However, with access to the MIRA supercomputer and additional HPC clusters, researchers at VERIFI worked with Convergent Science Inc. to scale their code into the hundreds and even thousands of cores. At the rate computational power is changing, "the resources we have available at national labs today," says Longman, "will be available in industry in just a few years."

To provide predictive simulations of engine and fuel performance, combustion models incorporate accurate descriptions of two-phase flows, chemistry, transport phenomena and device geometries. Image courtesy Argonne National Laboratory.

VERIFI scientists and industry partners created the largest grid-convergent, diesel-engine simulation to date, with 50 million cells; it ran for three weeks on 500 cores. Som notes there were a few cores handling 30-40,000 cells, which would eventually run out of memory and crash. "But with improved load balancing," he says, "we were able to distribute the cells evenly and achieve better scaling."

A competitive future

At the Paris Motor Show in 2012, Martin Winterkorn, CEO of Volkswagen AG, professed, "There will be, for a long time to come, no alternative to the internal combustion engine." Yet surprisingly, for the first time in its 64-year history, Motor Trend named Tesla's all-electric Model S sedan the 2013 Car of the Year. The interesting back-to-back dichotomy has yet to play out in long-term markets; however, the US Department of Energy aligns with Winterkorn's statement and estimates that by 2040 around 90% of new cars sold in the US will employ combustion engines.

"Many scientists have been working on the stages of combustion engines individually, in an empirical sense with coarse CFD meshes," notes Som. "But they haven't solved the complete problem; we're looking at the complete cycle, creating models that are extremely refined and predictive in nature." The simulations, says Longman, are intended to replace aspects of the design and test phase iterations including accumulating drive-time and response patterns by building and running test vehicles. "We're pushing the envelope to incorporate all of the expertise we have at Argonne and VERIFI, to lower engine design costs, shorten timeframes, and increase US competitiveness."

- Amber Harmon

Join the conversation

Do you have story ideas or something to contribute?
Let us know!

Copyright © 2015 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.


We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.