XSEDE Science Successes
Open Science and Industry Collaboration—Addressing New Problems While Improving the Bottom Line
By Scott Gibson
Consumers are happy when products flow nicely, whether the items are tubes of toothpaste or bottles of shampoo, while people in open-science research and private industry, respectively, like workflows that result in better problem solving and higher profits.
A new program from the Extreme Science and Engineering Discovery Environment (XSEDE) called the Industry Challenge brings the scientific and industrial communities together in multidisciplinary collaborative teams and connects them with XSEDE's world-class advanced digital services.
"The Industry Challenge solicits open science proposals that are fundamental to modeling and simulation problems in industry," said David Hudak, director of the program. "By solving these basic problems, we hope industry will advance."
The idea behind the challenge apparently resonated with the intended communities. The Procter & Gamble Company, the world's largest maker of consumer goods, and Rensselaer Polytechnic Institute (RPI), America's oldest technological university, submitted the winning proposals in this year's competition and have assembled teams composed of both industrial and academic researchers.
The brands of Procter & Gamble (P&G) are household names—Braun, Crest, Oral-B, Head & Shoulders, Bounce, Charmin, and Tide, for example. The P&G research team is exploring how to predict the flow properties (rheology) of surfactants, compounds that lower surface tension. The rheology of surfactant formulations defines the dynamics of such things as how shampoo dispenses from the bottle, mounds on the user's hand, and then spreads between the fingers, said Peter Koenig, principal investigator for the P&G-led project. XSEDE resources will enable rheology predictions via molecular simulations and mechanical models.
"The adjustment of flow properties by altering the composition of ingredients is an important part of the design of new consumer products," Koenig said. "Being able to predict the rheology using computer simulations will focus and accelerate the development process."
Koenig explained that the simulations the P&G project participants need to run are too large for the computers at their respective sites, and so they are using Stampede at the Texas Advanced Computing Center and Keeneland at the National Institute for Computational Sciences. Those supercomputers provided by XSEDE enable the project team to perform the requisite demanding novel molecular simulations of surfactant properties that relate to rheology. "Project plans include some large-scale, high-fidelity simulations to establish the approaches in a suitable way for publication in the scientific literature," he said. "Staff from XSEDE will contribute expertise to the research to optimize the speed and robustness of protocols for routine application."
Meanwhile, XSEDE is assisting the RPI team with the use of Stampede as it develops and demonstrates computationally parallel simulation workflows for companies such as Corning, Inc., ITT Gould Pumps, Pliant Energy Systems, and Sikorsky Aircraft.
"Industry is always focused on the bottom line," Hudak said. "I believe that engaging the academic community in general, and XSEDE in particular, must have a demonstrable return on investment [ROI] or else industrial partners will lose interest. I want to find ways to demonstrate ROI for our Industry Challenge projects."
Reflecting the goal of enhancing profitability, the RPI team is delving into projects that will enhance the materials processing, flow control, fluid structure interaction, and design methods of the companies involved in the team collaborations.
"Although the reasons companies are interested in having these simulations developed vary, from improving manufacturing process quality, to designing a more effective product, to reducing the time required to produce time-critical designs, the bottom line for each of the companies is improving their bottom line," said Mark Shephard, principal investigator for the RPI-led project.
Shephard added that the RPI team's study is directed not only at creating improved complex simulation workflows but also at increasing the levels of automation and reliability of the simulations. "As part of this research, specific attention is given to the interoperability of the tools produced so they can be used in the fast and cost-effective construction of simulation workflows that address new industrial simulation needs," he explained.
Hudak believes the various research victories made possible by the collaborations along the way will serve to underscore the overarching virtue of the Industry Challenge program. "Addressing the individual challenges represented by these problems will be significant for their respective domains; however, the larger win will be the demonstration that industry and academic teams can work together to achieve results they could not reach alone," he said.