| Working with college faculty and graduate students allows the company to try valuable experiments that can result in promising tools such as the robotic probe. |
A robotic welder that does not know the relationship between where it is and where its work is, produces 100 percent junk. But if the robot can figure out where it and its work are in relation to each other, the robot can make fast, quality welds.
A.O. Smith's automotive division in Milwaukee has a long history of pioneering automation in the assembly of automobiles. The company also has a long history of working with the college.
In its most recent partnership, graduate student Brandon Tarr is expanding on the work of COE alumnus Chris Smith, who now works for A.O. Smith. Smith earned his graduate degree by developing a robotic sensor under the direction of Mechanical Engineering Professor Neil A. Duffie. Duffie also earned his master's in mechanical engineering by working on a project at A.O. Smith under the direction of COE Dean John G. Bollinger. Bollinger and Professor Howard Harrison's work with A.O. Smith in the 1960s brings the story full circle.
In 1963, Bollinger was developing a radar measuring system for use in the construction industry. The University-Industry Relations program put A.O. Smith in contact with Bollinger. "A.O. Smith wanted to know if I could use my system to pinpoint the location in space of two plates coming together," Bollinger said. "They were interested in automating their welding process."
Bollinger successfully demonstrated that he could indeed use his system to find the joint, but thought the radar system would be too expensive for this kind of application. He suggested that A.O. Smith fund research into a seam tracker that could traverse the seam and relay information to an automated weld head.
A.O. Smith agreed and through phased development and years of work, Bollinger, Harrison and A.O Smith developed a water-cooled copper probe which traveled along the seam just ahead of and beneath the angle of hot spewing weld debris. The probe was controlled in five dimensions (x, y, z, pitch and yaw) using relay logic. It was the first five-axis robotic welder.
A.O. Smith then hired two of Bollinger's graduate students to build the robot for the production line. The robot successfully welded car frames until the industry switched to unibody designs in the 1970s.
A.O. Smith technician Roger Bratberg (left) discusses a robotic sensing device under refinement by graduate student Brandon Tarr (center) and Mechanical Engineering Professor Neil A. Duffie.
|
A.O. Smith recently returned to Bollinger's work when it ran into problems programming a modern, computer-controlled robotic welder. Again, the goal was to build a sensor that could locate the position of a seam to be welded on an automobile undercarriage, but this time A.O. Smith needed to program a robot to compensate for variations in a seam. Today, computer-controlled laser systems can do the job, but cost about $50,000 each. Retired A.O. Smith manager John Hinrichs, who worked with Bollinger on the relay logic system years before, remembered the technology and called Bollinger to see if elements of that system could help solve the problem. Bollinger contacted Neil A. Duffie who put graduate student Chris Smith on the case. Smith was charged with building a new sensor for less than $5,000.
"About that time, Asea Brown Boveri Inc. came out with a new robot including a more powerful programming language," said Duffie. "Chris Smith was the first to learn that language and implement a mathematically complex program on the robot."
The team's sensor plugs into the arm of the welding robot. It is a two-dimensional stylus that probes a seam by applying a voltage and measuring the potential in relation to position. The farther away from the seam the stylus is, the higher the voltage. The computer program analyzes the information and then plots the path for the welder. The robot must gather information, process it and plot the weld's course in a matter of milliseconds. The weld must be accurate to within one-half millimeter. Mechanical engineering graduate student Brandon Tarr is now working with Chris Smith to expand the abilities of the sensor to include a third dimension.
"It has truly been a cooperative project," said Smith. I acquired real-world experience while getting my master's, and then got a job to boot."
As a full-time engineer for A.O. Smith, Chris Smith says the relationship is a good deal for both the company and the college. Working with college faculty and graduate students allows the company to try valuable experiments that can result in promising tools such as the robotic probe. The college provides experienced researchers and graduate students that A.O. Smith could not afford to hire for one- time experimental projects. In addition, the company has the opportunity to evaluate potential employees while graduate students gain valuable experience.
"It really comes down to people," said Duffie. "When you work with a company on an operational problem, the more you can be there and work on a personal level, the more success you will have and the more the company will want to continue to develop the relationship. That's certainly been the case with A.O. Smith."
--By Jim Beal--
| For further information, please contact: |
Neil A. Duffie, 608/262-9457
duffie@engr.wisc.edu
© Copyright 1997 The Board of Regents of the University of Wisconsin System
Content by perspective@engr.wisc.edu
Markup by webmaster@engr.wisc.edu
Date last modified: Wednesday, 19-Mar-1997 12:00:00 CST
Date created: 19-Mar-1997
