Developing a clean green driving machine virtual research nets real results

When Dr. Bassem H. Ramadan, associate professor of Mechanical Engineering, goes deep inside an engine, his hands never leave his computer keyboard. Ramadan researches new engine combustion system designs for the Environmental Protection Agency (EPA) using computational fluid dynamics. In other words, he uses three-dimensional computer simulations to "virtually" test new engine designs.

"It's as sophisticated as we can get without actually being inside the engine," said Ramadan. The computer simulations allow Ramadan to look at many different variables like velocity vectors, how air flows inside the cylinder and how different designs of intake systems affect the combustion process in an engine.

The EPA has funded the research to help develop new engine technologies that will reduce emissions and increase fuel efficiency, ultimately reducing U.S. dependency on imported oil. This research is part of a national initiative to dramatically reduce the air pollution and greenhouse emissions from motor vehicles and to improve U.S. auto manufacturers competitiveness in the world market. Some of the designs are conceptual while others are changes to existing engines.

The EPA grants were originally given to Michigan State University (MSU) who subcontracts the research to Kettering. The research project followed Ramadan to Kettering from MSU where he had been a research associate. "When I started in 1998 I was not aware of anyone who was using computer simulations to do engine research at Kettering. I brought the technology here with me from MSU," Ramadan said. He is currently working in his second 2-year grant cycle. The EPA has provided a total of $330,000 for research through 2002.

"What makes these computer simulations valuable is that if you were to do these experimentally (with test engines), you wouldn't be able to get as much information as you would from these simulations. You would have to make thousands of measurements, and it would take a long time to measure at every point in the cylinder under different conditions and at different times, " said Ramadan. Image removed.

Another benefit to the simulations is that you are able to see what is going on during combustion. "In a test engine you need to have a transparent cylinder (may be made of quartz) to be able to visualize what's going on," said Ramadan. Currently it is very difficult to use transparent cylinders, because they may not tolerate the high temperatures and pressures occurring in a combustion engine. "In experimentation, researchers can visualize cold flow, without combustion, which is completely different from what happens during combustion," he added.

Ramadan works with a company called FEV, that builds prototype test engines based on EPA designs. The EPA works with the big three auto manufacturers on design concepts, according to Ramadan. "Then we get a CAD file from FEV that contains a description of the geometry of the proposed engine design," he said.

"We take it and create a "mesh" or finite element model of the geometry. This mesh allows us to perform calculations to determine pressure, density, velocities, chemical species concentrations, turbulent kinetic energy, equivalence ratio, vorticity, and temperature at all the different points in the mesh. It is all based on mathematical models that have been proven to be somewhat accurate," he said.

For Ramadan, the nice thing about working with FEV is that they will eventually build the engine and test it. "Then we are able to compare the numerical results to actual engine test data. We have had some amazing good agreements between the simulations and what the engine tests have shown," he said.

Most discrepancies are found when the actual test doesn't replicate exactly what was simulated. "We may run a simulation that tells us the velocity is 10 meters/sec at a given point in the cylinder, when they measure it they may get 8 or 12 meters/sec. They don't measure exactly what we solve. But the trends are usually similar. The model usually over predicts or under predicts the velocity," Ramadan said.

One pleasant surprise came when Ramadan was working on an engine simulation for direct injection. Direct injection systems basically work like a diesel system, injecting the fuel directly into the cylinder rather than into a port like in a conventional engine. The computer program would abort during certain calculations, and no matter how many times he tried it again, with different numerical solvers, the simulation would abort before completing the cycle.

"I thought maybe there was a bug in the program. I found out later that when they did the testing, the engine did not start. The computer program checks the solution on a regular basis to make sure it does not diverge. When that happens the program aborts. So, in this case the program was aborting because the solution was not physically possible, the physical engine would not start," said Ramadan. "That was a nice correlation. When it happened later, I didn't look at it as being a defect in the program, I knew it was just solving a mathematical problem that was not physically possible. That increased my confidence in the program."

That is why the simulations are so valuable to the EPA, according to Ramadan, because it costs a lot of money to build a test engine. "They were spending a lot of money building engines before we began doing the simulations. Some simulations have shown certain ideas were not going to improve existing designs. Not only the EPA, but most of the larger auto companies are doing simulations before building test models, they are using the simulations as a screening tool, to decide which concept looks promising." he said. Image removed.

Looking at cross sections of the simulation results, Ramadan sees beauty in technology. Printouts of concept engine designs showing concentrations of intake air look like art to him. "It IS art to me," he said.

Ramadan is currently working on developing a Combustion Simulation Design Studio at Kettering to give undergraduate students first-hand experience using computational tools in solving complicated systems, and graduate students an opportunity to design advanced combustion systems.

Johnson Controls Inc., has provided funding toward the development of the lab. "We are also trying to acquire funding from the National Science Foundation for a high-end computer for the lab so we can do these simulations on campus. Currently we have access to an EPA supercomputer in North Carolina via remote connections to do the simulations," Ramadan said.

"We need a high-end computer because these simulations are computationally intensive. We're using three-dimensional models and we do everything under transient conditions, so it is happening in real time. That requires a lot of computational power," he said.

Ramadan currently has four graduate students working with him, two assigned to him directly, one working on the simulations as an independent study project and the fourth, while assigned to another faculty member, is completing a thesis project on a topic related to engine simulations.

Originally from Beirut, Lebanon, Ramadan earned his bachelor's degree in Mechanical Engineering from American University of Beirut, and his master's and Ph.D. in Mechanical Engineering from Michigan State University. He currently lives in Davison, Michigan.

Written by: Dawn Hibbard
(810) 762-9865