A D V A N C E D M A T E R I A L S & P R O C E S S E S | F E B R U A R Y / M A R C H 2 0 1 7
1 0
TESTING | CHARACTERIZATION
APPLYING MACHINE
LEARNING TO
NANOMATERIALS
Researchers at the DOE’s Argonne
National Laboratory (ANL), Lemont, Ill.,
for the first time used machine learning
to predict the physical, chemical, and
mechanical properties of nanomateri-
als—and found it to be more accurate
than traditional approaches. The team
created the first atomic-level model
that predicts the thermal properties of
stanene—a 2D graphene-like material
made of tin atoms potentially useful
for thermal management in certain
nanoscale devices. Badri Narayanan,
postdoctoral researcher, explains, “We
input data obtained from experimental
or expensive theory-based calculations,
and then ask the machine, ‘Can you
give me a model that describes all of
these properties? Can we optimize the
structure, induce defects, or tailor the
material to get specific desired proper-
ties?’” Machine learning can be applied
to a range of materials, and unlike tra-
ditional approaches, it can accurately
capture bond formation and break-
ing events. The efficiency of the new
method is unprecedented—until now,
atomic-scale materials models took
years to develop, and researchers relied
largely on their own intuition to identify
parameters. With machine learning, the
need for human input is reduced and
development time is shortened to a few
months.
anl.gov.
DIGITAL INFRASTRUCTURE
PROJECT EXPLORES
ADVANCED STRUCTURAL
MATERIALS
An international group of organi-
zations—including Oak Ridge National
Laboratory, the Electric Power Research
Institute, the European Commission
Joint Research Center for Energy and
Transport, and ASM International—is
working together on a joint project to
build a digital infrastructure for Mech-
anical Testing Data (calledMeTeDa). The
project consists of developing standard
data formats for uniaxial creep, fatigue,
creep crack growth, and creep-fatigue
crack growth test data. The team is cur-
rently working with data generated by
the nuclear energy industry, but would
like to broaden the scope of its efforts to
include such data from other industries
as well.
MeTeDa is an outgrowth of the DOE
International Nuclear Energy Research
Initiative (I-NERI), established in 2001
to foster international cooperation on
nuclear energy and its use. Critical to
I-NERI’s mission is the development
of data standards for materials used in
The Argonne research teamwho pioneered the use of machine learning tools in 2Dmaterial
modeling. In the background is a 2Dmodel of stanene, which is softer andmuch more rippled
than its cousins graphene and silicene. Courtesy of ANL.
BRIEFS
The
American Association for Lab-
oratory Accreditation
granted the
metallurgical laboratory of AM man-
ufacturer
Sintavia LLC,
Davie, Fla.,
ISO 17025, the highest recognized
quality standard in the world for
calibration and testing. Previously,
AM manufacturers offering this level
of testing had to use independent
laboratories for powder and material
validation. Now, Sintavia’s in-house
accredited laboratory will allow
faster analysis and more complete
process security, according to com-
pany sources
. sintavia.com.Bodycote,
UK, announces that its Chesterfield hot isostatic pressing (HIP) location
earned, for the second time, the highest level of Nadcap accreditation following the
most recent Nadcap audit.
bodycote.com.The University of Alabama in Huntsville
opened a new laboratory for predictive
failure testing of electrical insulators used by utilities, component manufacturers,
aerospace firms, and NASA. The Power Systems Insulation Laboratory performs
real-time testing and computer modeling of insulating materials under conditions
such as temperature variation, mechanical stress, electrical stress, dirt buildup,
corrosion, moisture, and inherent natural decay.
uah.edu.