Data. It is the lifeblood of science and engineering. Researchers and engineers use it to model, design, benchmark, evaluate, monitor and compare. But, how easy is it to find, what kinds are needed, how good is it and how accessible is it? These are the questions that a workshop held earlier this week began to address.
The workshop was organized by Steve Freiman and John Rumble, and with support from DOD. Freiman and Rumble (with Lynnette Madsen), you may recall, started bringing the issue of data and databases to the attention of the ceramics community with an article in the ACerS’ Bulletin in March 2011.
Today’s data needs are driven by increased use of modeling and simulation tools in research and development across the full spectrum of engagement: academia, national labs, big industry and start-up innovators.
Recent federal-level initiatives are big drivers of the demand for data, too. The drumbeat of the Materials Genome Initiative since its introduction a year ago is to spur the leveraging computational tools to reduce the time from innovation to marketplace by at least half. Similarly, since its establishment in 2000, the National Nanotechnology Initiative is driving the use of data, informatics and computation for nanotechnology development. (For example, there is a new NNI undertaking, “Nanotechnology Knowledge Infrastructure: Enabling National Leadership in Sustainable Design,” which will develop computational tools in a multidisciplinary and collaborative culture in close concert with MGI.)
The Freiman-Rumble workshop was titled, “E-Ceramics: Prospects and Challenges for Improved Access to Ceramics Property Data.” E-Ceramics refers specifically to electronic access to ceramic materials property data. The goals of the workshop were to assess what data resources are available now, what are the continuing challenges and to ask what the future is for ceramic materials data innovations.
“There’s data out there, it’s just not easy to come by,” Freiman says. He cited as examples of existing databases the ACerS-NIST Phase Equilibria Diagrams, NIST Ceramics Webbook, military handbooks, the International Centre for Diffraction Data, AMPTIAC reports, professional society handbooks and many assorted databases, about which little is known beyond their user bases.
Lew Sloter's talk reviewed DoD's interest in engineered materials to support the agency's mission. Credit: ACerS.
The workshop was underwritten by the DOD, through the program office managed by Lew Sloter, associate director, materials and structures, DOD. In his keynote presentation, “Materials as a Key Defense Strategy,” Sloter said the current workshop grew out of an NNI-related workshop held at Oak Ridge National Laboratory in 2007 on nano-informatics, where several participants (including Sloter, Freiman and Rumble) realized that, in order to gain traction, the topic of data needed to be addressed in a specific context. “We were looking for a materials type that had the necessary complexity to make a good case study,” Sloter said.
About thirty workshop participants heard presentations by representatives from the service labs—Army Research Laboratory, Air Force Research Laboratory and the Office of Naval Research—on their how they access, generate and use data. The perspectives of industry (Boeing, United Technologies Research Center, Kyocera and Du-Co Ceramics), national labs (Sandia, NIST) and academia also were presented.
A few themes emerged, and the issues that attend ceramic property data are indeed complex.
- There are few standard specifications or designations for ceramic materials. Investigators are not necessarily talking about the same material, even when it seems like they are. The oft-cited example is “96 percent” alumina, the properties of which are determined in large part by the unspecified 4 percent.
- Much data is not accessible. It is cloistered into proprietary databases or in protected databases, such as ITAR/EAR databases. And, manufacturers often generate and store their own data on an as-needed basis.
- Lack of standard reference materials and standard testing methods. Also, advances in testing methods themselves can raise questions about the validity of data generated using older methods.
- Support for data generation is declining. Freiman says materials R&D labs are closing, industrial R&D is fragmenting, government support is declining and the overall economic downturn is having an effect. However, new technologies and new materials, such as nanostructured materials, are increasing demands for data.
- Quality: How good is the data and how does the user know? How was it generated; how far back does the data trail go? Matthew Batcher from ARL works on SiC armor tiles and asked in his presentation, “What information is lost or not translated to users as it moves forward?”
- Inconsistent demand. How does fluctuation in demand affect the economics of data acquisition, storage and retrieval? Representatives from large companies said they cannot get enough quality data to run good design FEA models. A researcher from the Air Force Research Laboratory described his lab’s data needs as “sporadic but intense.”
The issues of quality and provenance came up frequently. Without knowing the uncertainty in the data, any use of the data will also have unknown uncertainties. The solution is to attach metadata to the data so that users can evaluate for themselves whether the data is usable.
Kevin Ewsuk of Sandia National Laboratory pointed out that there are tiers of data quality, depending on what the engineering activity is. For example, lower quality data can be used in the development phase when the researcher is looking for trends, however, more exact data is needed for when it comes to applications. Ewsuk also noted that not all applications demand the same quality level of design data. “Think of your cell phone,” he said. “If it quits working, you throw it away and get a new one. That doesn’t work if the same component goes bad on a satellite.”
Freiman challenged the group, “Imagine a world with unlimited access to ceramics data. Now, how can we get there?”
The clear consensus was that one giant, monolithic database would not meet the specific needs of enough users and would not be sustainable. However, with internet tools like Google-type search capability and robust metadata, an organic approach to building and accessing databases is possible and may offer a pathway to assembling a collection of databases. Several ideas were suggested for ways to adapt other internet practices, such as Amazon’s review function, to engage data contributors and users in a dynamic and productive way.
The most challenging questions Freiman posed were “Where do we go from here? Who should lead?”
Several participants suggested The American Ceramic Society is well-positioned to take a leadership, or at least a facilitator, role. Certainly, since its founding, the Society has viewed technical content as core to its mission. Others suggested that the government should provide support, at least in terms of funding.
What do you think? What are your data needs and wants? What role would you expect a society like ACerS to assume?