This article highlights about getting more out of information technology, from online services to digital databases, in order to make plants more flexible, to improve product development, or to let people at far-flung sites cooperate more closely. There are industry-led groups namely the Consortium for Advanced Manufacturing-International, which has an office in Bedford, TX, and Intelligent Manufacturing Systems, headquartered in Tokyo. These organizations address concerns ranging from budgeting to the best practices for designing a plant floor. The Consortium for Advanced Manufacturing and Intelligent Manufacturing Systems support a research initiative called Next Generation Manufacturing Systems (NIST). NIST also operates the Manufacturing Engineering Laboratory, where a variety of projects are being run. In another part of NIST—the physics lab, to be specific—Marc Desrosiers, a research chemist, has led an effort to cut the time and cost of certifying the calibration of irradiation equipment, which can be used to cure materials and coatings, or to kill bacteria on products ranging from medical devices to hamburger.
The people who work to push manufacturing to new heights have clear goals for the future. They want to develop processes that will turn out goods more accurately and with less expenditure of time, effort, and materials. What’s more, they say they don’t want to settle for two out of three: They want to make products faster and cheaper, and turn them out better, too.
In short, they have the same concerns those guys had in the old days, when they sat around the fire chipping arrow points.
But now, instead of flint versus obsidian, the discussion often revolves around getting more out of information technology, from online services to digital databases, in order to make plants more flexible, to improve product development, or to let people at far-flung sites cooperate more closely.
There are industry-led groups like the Consortium for Advanced Manufacturing-International, which has an office in Bedford, Texas, and Intelligent Manufacturing Systems, headquartered in Tokyo.
These organizations address concerns ranging from budgeting to the best practices for designing a plant floor. The Consortium for Advanced Manufacturing and Intelligent Manufacturing Systems support a research initiative called Next Generation Manufacturing Systems.
Governments, of course, back efforts to advance the manufacturing technology of their countries. Intelligent Manufacturing Systems, although it is largely sponsored by industrial companies and is an international effort, grew out of a meeting led by Japanese government officials.
In the United States, the National Institute of Standards and Technology sees manufacturing—industry in general—as its turf. NIST, for instance, last May hosted a meeting in which various researchers representing Next Generation Manufacturing Systems discussed some of their findings.
NIST also operates the Manufacturing Engineering Laboratory, where a variety of projects are being run.
As David Stieren, strategic relations manager for the lab, sees it, “NIST is the agency that exists to aid U.S. industry.”
Stieren’s job is to oversee a number of programs in the works. One is a project to offer an online service that will markedly simplify the certification of irradiation equipment. Another is a drive to standardize the documentation of machine tool data, a movement with implications that could influence a range of manufacturing operations.
Work on the machine tool project began in the mid- 1990s, according to Alkan Donmez, NIST’s group leader for machine tool metrology. It grew out of talks with Caterpillar, Boeing, and other manufacturers about problems in reducing time from concept to production.
At present, the project involves a consortium of 10 companies, Donmez said.
NIST’s leader for the machine tool project is Hans Soons, who is also known as program manager for metrology and smart sensor systems for manufacturing equipment. As he described it, his job deals with “the health and accuracy of machines.”
According to Soons, when manufacturers periodically test their machine tools for accuracy, they use various methods and equipment from different manufacturers. A problem arises, he said, when machine owners want to compare data from different tests. To do so becomes troublesome because each maker of test equipment has its own format for storing information.
Meanwhile, owners and operators of machines are creating comprehensive databases of their test information, which they see as offering a number of potential uses. But they say the incompatible file formats are a source of frustration. They must translate the files of the different brands of test equipment into a common format to make them comparable and available throughout the organization.
The test information is relevant to the activities of departments ranging from maintenance, which has to schedule servicing with the least disruption to the factory, to budgeting, which has to predict how much to set aside to repair or replace machines.
Translation takes a lot of time and money. Some of the file formats resist the process, and there is always a danger that errors of translation may creep into the record.
Divided By Common Languages
What’s more, each manufacturer is translating its data files into its own common language, and no two may be alike.
James Katter, team leader for advanced production technologies at the Caterpillar Technical Center in Mossville, Ill., is one of several representatives of the private sector who are working with NIST to promote industry standards for documents relating to tools.
Katter pointed out that, although a company may have a database of translated files for internal use, two companies can’t exchange the information.
For instance, a part specifier and a would-be supplier might communicate about a job more quickly and clearly if they could share information on the performance of tools.
Katter cited another disadvantage to relying on the translation method: Each new supplier of test devices requires a new translator.
According to participants front NIST and industry, compiling databases of machine tool test results is complicated not only by differences in software, but also by content. Testing systems don’t all save the same information in the same place, and sometimes use different terms.
According to an executive involved in the NIST machine tool project, “We are trying to come up with a common language for machine test data.” Data files would follow a pattern, particularly in structure and description, regardless of the vendor of the test equipment.
Databases could let machine shops track the change in the accuracy of a machine over time, or in a group of machines, or compare the behavior of one machine with another. If test data were standardized, the facts could be gathered from different kinds and brands of testing equipment and readily grouped for comparison.
Soons and others in the project say the information may also one day be combined with machine programming technology to run simulations that will predict the performance of individual machines in producing specific parts.
The machine tool project has spawned an ASME technical committee, which Soons chairs. The committee, B5/TC 56, has submitted a draft of a standard that spells out the information to be generated and saved in a test data file, and suggests XML, an Internet code, as the markup language for the standardized data files.
Browsing in History
According to Soons, it may be possible one day to upload test results on an ordinary browser to compare with previous data for patterns of performance. If the user of the tool judges something isn’t right, the information can be relayed to the maker of the machine for analysis.
Suppliers in the future may share the information with customers to prove that they can make parts to a required degree of precision.
Another potential use of a performance database is for virtual machine tooling. This technology, which has only reached an early stage of development, combines a computer-aided manufacturing file with test information about an individual machine.
VulcanCraft, a software developer in Carrboro, N.C., has a prototype program that will run a simulation incorporating a tool’s anomalies— perhaps the misalignment of the spindle or a guide out of true—to predict how accurately the machine can render a design, a visual representation of the tool’s accuracy.
Don Esterling, president of VulcanCraft, said he has run simulations using the software, but it is “sitting on the shelf.”
The main interest at present is the standardization of files, he said, and added that the software program would take “another year’s worth of work, at least,” before it would be ready for market.
Esterling said he is waiting for additional data and more indications of interest from machine tool operators before he invests more time in the simulation program.
Katter suggested that standardization might lead to automatic compensation by machine tools for their errors.
Machine tools contain error compensation tables, but in different formats. Compensation for inaccuracy is now done manually.
“A standard format will allow makers to come up with more robust mechanisms for compensation,” Katter said.
One executive said the consortium members hope that, once a standard has been devised, the vendors may adopt it painlessly, by writing it into the normal upgrades of their software.
Another Movement on Tap
Katter said there is another movement afoot, to standardize not only test data, but also machine tool spec sheets and add that information, too, to electronic databases.
Suppliers’ descriptions of machines now contain a variety of information, Katter said, but they are inconsistent. They don’t all have the same information, and as in the test data files, it is not presented in the same manner.
The committee is trying to spell out the information needed not just by the process planners and the management of the machine shop, but also by design, accounting, maintenance, and other departments.
Meanwhile, in another part of NIST—the physics lab, to be specific—Marc Desrosiers, a research chemist, has led an effort to cut the time and cost of certifying the calibration of irradiation equipment, which can be used to cure materials and coatings, or to kill bacteria on products ranging from medical devices to hamburger.
Operators of radiation sources, whether they use cobalt or electron beams, have to make periodic checks of the energy levels of their devices. Currently, they receive dosimeters, which are put through the workaday radiation process. Instead of being sterilized, the dosimeters record the level of radiation they have received. The dosimeters are mailed to NIST, where technicians analyze the results and issue a certificate of calibration.
The process uses conventional mail, so it takes days. NIST charges $2,100 to $2,200 for the first three dosimeter readings, or “dose points,” Desrosiers said, and $400 for each additional dose point in the same calibration test.
Desrosiers said he has been working with an instrument company to develop a reader that will work with a PC at the operator’s site and log into a NIST server.
Dosimeters will not have to leave the irradiation company’s hands. They will be inserted one at a time into the reader, which will transmit the information to NIST’s server. The server will run 24 hours a day, and will issue a provisional certificate of calibration automatically.
According to Desrosiers, once the subscriber has logged in over the Internet and requested a calibration, NIST’s server will take control of the PC and reader to render them inaccessible to the user. According to Desrosiers, this provision is intended to rule out the possibility of tampering with calibration results. Other procedures are designed to keep tabs on the performance of the dosimeter reader.
Test results can be sent to NIST at any time, and during work hours, technicians can confirm the results and issue a permanent certificate of calibration.
Desrosiers said the new service will be by subscription, and NIST will charge according to server time, not by the number of individual readings, as it does now. He estimates it could bring down the cost to as little as $20 a dose point.
NIST demonstrated the idea using a reader and PC set up at a conference in San Diego. “We connected to the NIST server and performed measurements,” he said in an e-mail. “We are now waiting for these readers to be produced commercially. As soon as companies have them, we will begin testing our service.”