TOP STORY

Center Researchers Use Genetic Algorithms to Design Phononic Bandgap Structures
By Diane Kukich

Research being carried out under the ARL Composite Materials Technology (CMT) program may lead to devices that can protect sensitive electronic equipment from damaging vibration.  Dan Weile, Assistant Professor of Electrical and Computer Engineering, is collaborating with George Gazonas of ARL in a project that exploits phononic bandgaps to filter out deleterious vibration frequencies.

Bandgaps are basically are ranges of frequency where waves cannot propagate in materials.  “Whether we’re talking about photonic or phononic bandgaps, what we’re doing is exploiting the properties of waves,” Weile says.  “The periodic arrangement of a material will determine whether a wave travels through it or is blocked.  With photonic bandgaps, we’re referring to light waves; with phononic bandgaps, it’s sound waves.” 

Energy distribution of an acoustic wave in a periodic
lead silicone acoustic structure

The challenge with phononic bandgaps is actually the opposite of that encountered with light.  “Light waves are very short, so very tiny devices are required to capture the bandgap effect,” he explains.  “With sound, the waves are huge, so we need a way to exploit this phenomenon in a smaller environment.”

One way to do this is to choose materials that are very different from each other to take advantage of the impedance mismatch at the boundary between them.

There is a good news/bad news story in this effort to capitalize on material properties to take advantage of the phononic bandgap.  Weile explains that working with the photonic bandgap is an electromagnetic problem, with two types of variable properties:  permittivity, which describes electrical properties, and permeability, which describes magnetic properties.  Since most materials are not magnetic, only the permittivity varies.

However, the two concomitant terms for working with the phononic bandgap, which is a mechanical problem, are density and modulus.  Since all materials have different values for those properties, there is much more freedom.  It is therefore much more difficult to determine an optimal design, but, at the same time, more freedom has the potential to lead to better designs.

Enter Charles Darwin.  Weile is using a design methodology based on Darwinian theory to select the best design for the job.  “The basic idea behind the use of genetic algorithms is that solutions to a problem are encoded into a ‘chromosome,’” he says.  “The solutions that don’t work are then killed off, while the ones that work are retained and bred with each other.  It’s natural selection applied to design.”

The problem that Weile is attempting to solve is determining how to control the frequency so that a small structure can be induced to resonate at low frequency.  “One way to do this,” he says, “is by making ‘weird’ shapes.”

Unfortunately, genetic algorithms (GA) are not very efficient at geometric design, so Weile and his team have turned to a variation called genetic programming (GP).  The former works with a fixed list of options, while the latter is much more general. 

“If you wanted to use GAs to make a pencil,” he explains, “you would provide the program with the information that the device uses a piece of wood, a piece of lead, and some paint.  The GA would then arrive at a design with the lead and wood each of a certain diameter, the pencil of a certain length, and yellow paint applied to the outside.  But it would never have an eraser.”

“With GP,” he continues, “the system might suggest that a piece of rubber affixed to the end of the pencil could serve as an eraser, as long as the user had suggested that ‘erasability’ was a desired function of a pencil.  What we have with GP is a new way of doing genetic design.”

Band structure designed by the team shouwing
the existence of a bandgap

Preliminary results have demonstrated some success with this approach.  The next step will be to provide other researchers at CCM with some designs for fabrication and testing.

Weile admits that he’s an electrical engineer working on a mechanical problem, but that hasn’t daunted him.  “I’ve done a lot of work with genetic algorithms and with waves,” he says, “but we will need some help with expansion of the concept into a product and with further code development.  This project is exciting because it’s giving me a chance to move from the computational side of the problem to optimization.”

Although his computational work capitalizes on the most advanced computing capabilities available to engineers, Weile acknowledges the important contributions of a nineteenth century Scottish physicist and an English naturalist from the same era. 

“I owe a lot of what I’m doing in my work to James Maxwell and Charles Darwin,” he says. 

To give credit where it’s due, Weile has posted pictures of the two bearded scientists on the door of his lab in DuPont Hall

OTHER NEWS

spacebeam1

CCM Participation Continues in FAA Center of Excellence

Another win was added to CCM’s decades-long track record of success as a Center of Excellence when the Federal Aviation Administration recently decided to renew its program on VARTM processing of aerospace structures for a third year.  The program is part of the FAA’s Joint Advanced Materials and Structures (JAMS) Center of Excellence, which focuses on research, engineering, and development of safe and reliable uses of advanced materials and composites in aircraft.

CCM is involved as a member of the Center of Excellence for Composites and Advanced Materials (CECAM) team,  which is led by the Wichita State University. The team also includes Northwestern, Purdue, Tuskegee Universities and the University of California at Los Angeles. Dirk Heider, Assistant Director for Technology at CCM and Associate Professor in the Department of Electrical and Computer Engineering, is the UD PI on the program

"More primary aerospace structures—for example, the Boeing 787 pressure bulkhead and the Airbus A380 flap tracks—are being fabricated using liquid molding processes today,” says Heider. “In the past, these parts were made via autoclaving of prepregs.  What we’re trying to do is gain a better fundamental understanding of the process so that we can reduce variability of the VARTM process to the point where it’s equivalent to that of autoclave processing.”

The work focuses on three primary technologies: (1) membrane-based VARTM developed by EADS, (2) the Controlled Atmospheric Pressure Resin Infusion (CAPRI) developed by Boeing and (3) optimization of the standard VARTM process for new resin systems developed by Hexcel and Cytec.

Heider explains the first area focuses on a novel membrane, which “breathes”.  It shows promise for reducing the voids that may occur in VARTM process.  “The membrane entraps the resin while allowing volatiles to escape,” says Heider. 

In effect, this supplemental layer provides an areal vacuum over the entire surface and enables uniform compaction, decreasing the thickness gradient and eliminating the dry spots characteristic of manufacturing setups that use “point” vents.

One challenge to the application of this innovative technology, developed by German-based EADS, is the resin and the membrane pore size must be compatible with each other for it to work.  Solange Amoroux, a Ph.D. candidate in Materials Science, is working to develop a model that will ultimately enable users to tailor membrane selection in the VARTM process to the specific resin being used.

The second project focuses on the patented process developed by Boeing for its next generation of aircraft.  This process will be characterized to identify its capabilities and sources of variability.

The third area focuses on two new resins—Cytec’s Cycom 977-20 and Hexcel’s RTM6—developed for liquid molding processes and being used in new aerospace applications. 

“We’re working to optimize the VARTM process, develop a structural property database, and an optimization protocol available to all fabricators,” Heider explains. 

“One of the best things about this program is that we’ve had lots of industrial participation during the past two or three years,” he continues.  “With FAA’s requirement of a 1:1 match, we’ve received strong financial support from Cytec, Gore and other partners.  Overall, we have 16 companies interested in the membrane technology.”

“The CCM VARTM project was the result of an FAA interest in providing standardized control for raw materials and processing that was historically individual company proprietary data.” explains Curtis Davies, FAA Program Manager of the JAMS Center of Excellence.  “This work extends our original work on prepreg materials into this emerging technology to provide direction for proper control and results in increased safety for aviation structures. The experience CCM possesses in processing composites combined with other members analytic and testing capabilities has made JAMS a unique and valuable resource for our research, engineering and development program.”

CCM Associate Director Suresh Advani also points out that the FAA program complements the work done earlier under another Center of Excellence program at CCM.  “Beginning in 1997, our ONR Advanced Materials Intelligent Processing Center focused on developing and demonstrating intelligent RTM and VARTM processing to improve quality and reduce variability and cost,” he says.  “The simulations we developed under the Navy program are now being integrated into the FAA program in an effort to improve dimensional tolerances and overall quality.”

Response to the work at the annual JAMS CoE Technical Meeting in Seattle was excellent, and the program was renewed based on successful progress to date.

VARTM has the potential to reduce cost with equivalent
repeatability compared to autoclave processing.

CCM Offers Composite Materials Workshop

On October 10 - 11, 2006, CCM will host a workshop entitled “Introduction to Vacuum-Assisted Resin Transfer Molding (VARTM)”, at the Composites Manufacturing Science Laboratory
Registration - 8:00 a.m.
Class times - 8:30 a.m.-4:30 p.m.
Registration deadline: Sept. 26, 2006

Course Description
The Vacuum-Assisted Resin Transfer Molding (VARTM) process is a Liquid Molding Process used to fabricate large-scale composite structures that can have substantial laminate thicknesses. The closed-molding approach dramatically reduces emissions compared to open molding while allowing the use of low-cost tooling.  The course will introduce engineers and technicians to the fundamentals of the process with hands-on lab classes as well as theoretical lectures. The lectures include a review of the important process parameters, an introduction to simple analytical tools to predict the flow behavior, and a practical demonstration of the Liquid Injection Molding Simulation (LIMS) software, a finite element package capable of predicting the flow in complex geometries.  A focus of the class is also the practical aspects of the VARTM process. The class will characterize the permeability of several reinforcement and distribution media materials and fabricate several specimens from flat panels to a complex geometry part.

Click here to register and here to learn more.

CONSORTIUM

CCM would like welcome Kop-Flex Emerson Power Transmission, Hanover, MD, to our University-Industry consortium. We also with to thank our current members for their ongoing support, and for continuing to participate in consortium activities. To learn more about our Industry-University consortium, please visit http://www.ccm.udel.edu/Consortium/members.html .

NEW PUBLICATIONS

Conference Proceedings

Su, D., M. H. Santare, and G. A.Gazonas, “Numerical Modeling of Wave Propagation in Anisotropically Microcracked Media,” Shock Compression of Condensed Matter-2005, Proceedings of the 14th APS Conference on Shock Compression of Condensed Matter, Baltimore-2005, eds. Furnish, M. D., Elert, M., Russell, T. P., and White, C. T., American Institute of Physics, p. 359-62, 2006.

Invited Talks/Presentations

Vinson, J. R., invited lecture, “Composite Materials: Their Future and Creative Uses,” presented to the thirty-five students in the senior design class in Electrical and Systems Engineering at the University of Pennsylvania, Wednesday, September 13, 2006

 

Home | Newsletter Archives | Current Newsletter | News Archives | Current Events