Internal combustion engines have been identified as the source of one-third of ozone-depleting greenhouse gases and pollution. Vehicular accidents are a leading cause of death and injury in the U.S. ORNL researchers are addressing these problems with computational studies of structural integrity, combustion, and driver behavior.

Lighter vehicles use less fuel and generate less pollution. Sounds simple enough... but what are the effects on safety and performance, and what special manufacturing problems result?

ORNL researchers are answering those questions through use of advanced computational simulations -- accelerating the development and introduction of new auto body materials (such as ultralight steel, lightweight aluminum, and composites).

• Assessment of material related effects on vehicle design and performance.
• Evaluation and prediction of effects of the advanced manufacturing and materials processing techniques in realistic automotive conditions.

Geometry models are obtained by disassembling the vehicle, scanning the shapes, and measuring the mass and inertia of each component.

The Finite Element (FE) model is derived from the geometric model by discretizing. This separation of geometrical and computational FE representations allows greater flexibility in applying modifications and constraints.

Models are validated by comparing simulation results with an actual controlled crash.

Today's typical, state-of-the-art full vehicle models have around 100,000 finite elements and use many simplifying assumptions. A simulation requires several days of computational time on high-performance multi-processor workstations, but can be cut to one day with use of massively parallel computers.

Realistic large scale simulations would require detailed geometry representation, complex material and failure models integrated with scientific tools for visualizing and interpreting the computed data -- both on a scale not manageable on today's supercomputers. A new generation of computer architectures that surpass teraflops computations by at least an order of magnitude is needed.

• DOE Lightweight Materials Program
• Automotive Composites Consortium
• National Highway Traffic Safety Administration
• American Iron and Steel Institute

One way to reduce the pollution generated by vehicles is to improve the efficiency of catalytic after treatments. Inside a catalytic converter, combustion products are passed through an array of thin channels which have catalytic particles imbedded on their porous inner surfaces. A greater understanding of the fluid dynamics, surface chemical reactions, porous diffusion, and heat transfer involved in the catalytic process is needed in order to improve performance.

ORNL researchers are developing a dynamic simulation model for automotive catalytic converters. Massive computational resources are required to address all the factors involved in handling transients:

• Chemical kinetic mechanisms for surface reactions
• Flow through porous media including mass transfer to and from the bulk flow
• Conductive heat transfer between the channels through solids

Using computers to design vehicles can help make them safer and less polluting -- so putting more computers IN the vehicles will make them safer to drive...or will it?