Kespry and NVIDIA demonstrate deep learning for commercial autonomous drones; NVIDIA Jetson TX1
How a tire company is doing its part to recycle and reuse; Michelin’s TREC

ORNL, industry partners using high-performance computing to develop new high-temperature aluminum alloys for engines

The Department of Energy’s Oak Ridge National Laboratory, FCA US LLC, and Nemak, a specialist in the production of high complex aluminum components for the automotive industry such as cylinder heads and engine blocks, are partnering to create lightweight powertrain materials that will help the auto industry meet the mandated target of 54.5 mpg (4.3 l/100 km) by 2025. Using high-performance computing, ORNL researchers are modeling the atomic structure of new alloys to select the best candidates for physical experimentation.

The ORNL-led project is part of a new initiative from DOE’s Vehicle Technologies Office to develop such new high-performance alloys. Ford, General Motors and FCA US are collaborating with national laboratories, universities and the casting industry to develop an affordable, 300 ˚C-capable high-strength cast aluminum alloy.

This target means engineering a material that is 25% stronger than current alloys and durable at temperatures 50 degrees Celsius higher—a necessity for next-generation engines—while keeping costs low for automotive manufacturers and consumers.

Automakers need powertrain materials that are not only lighter but also low cost and able to withstand the elevated temperatures and pressures in high-efficiency turbocharged engines. A typical materials development cycle takes 10 to 20 years.

The aggressive goals of these projects compress about half a century of typical materials development into a four-year project.

—DOE program manager Jerry Gibbs

The researchers from ORNL, FCA US and Nemak are using integrated computational materials engineering (ICME) to speed the development of new high-temperature aluminum alloys for automotive cylinder heads. ICME enables researchers to tailor new alloys at the atomic level to achieve desired properties such as strength and ease of manufacturability.

In a presentation at the 2010 US DOE BES (Basic Energy Sciences) Workshop on CMS, University of Michigan Professor John Allison (then also at Ford as Senior Technical Leader – Research and Advanced Engineering), noted that the materials domain is a different class of computational problem.

Materials response and behavior involve a multitude of physical phenomena with no single overarching modeling approach, he observed. Capturing the essence of a material requires integration of a wide range of modeling approaches dealing with separate and often competing mechanisms and a wider range of length and time scales. The integration of knowledge domains, he said, is the key to ICME.

Aluminum has been in mass scale production for more than a century, but current cast aluminum alloys cannot withstand the temperatures required by new advanced combustion regimes. Our goal is to take high-temperature cast aluminum where it has never been.

—Amit Shyam, ORNL principal investigator

ORNL is breaking new ground by scaling ICME to run on DOE’s Titan supercomputer. (Earlier post.) Using Titan’s speed and parallel processing power, ORNL researchers can predictively model new alloys and select only the best candidates for further experimentation. This predictive capability significantly reduces the time, energy, and resources devoted to casting trial alloys.

Using approximately 100,000 cores simultaneously on Titan, we can increase the speed and scale of our first-principles quantum mechanics calculations by at least an order of magnitude.

—Dongwon Shin, ORNL researcher

Before the shift to Titan, Shin was using a Linux cluster with approximately 300 cores to create atomistic simulations of single elements diffusing to intermetallic precipitates within the alloy. Now researchers can achieve larger scale simulations on Titan that are much closer to real world scenarios.

The team is also verifying the computational models through atomic scale imaging and analytical chemistry measurements. ORNL’s scanning transmission electron microscopy and atom probe tomography allow researchers to identify and examine the location and chemistry of each atom in the alloy matrix, precipitates, and the interfaces between them.

ORNL and collaborators are creating a database that captures their aluminum alloy materials discoveries. This materials genome approach will help guide efforts to improve ICME capabilities and accelerate the development of new high-performance materials.

The Oak Ridge Leadership Computing Facility, a DOE Office of Science User Facility, approved six million hours on Titan for the ORNL alloy development project. The research uses microscopy resources at ORNL’s Center for Nanophase Materials Science, a DOE Office of Science User Facility. The alloy development research is funded by the Propulsion Materials Program in the Vehicle Technologies Office of DOE’s Office of Energy Efficiency and Renewable Energy.

Resources

Comments

HarveyD

Lighter ICE, drive trains, frames, wheels, cabs etc is a strong near future possibility to cut total vehicle weight, reduce fuel consumption, GHG etc.

More could have been done many years ago.

Henry Gibson

Forget more efficient engines, first use Artemis digital control hydraulics and that of NOAX for transmissions. Where are all the gas pressure bladder tanks to save braking energy. ..HG..

Henry Gibson

What about Riccardo flywheels and magnetic gears perhaps? ..HG..

The comments to this entry are closed.