LEARN MORE

Zones



GM's Drive For Flexibility

As it works to put out more and better cars and trucks—more economically—GM is seriously pursuing a manufacturing strategy in which robots play a significant part.

 Marty Linn maintains that flexibility is key today at General Motors' manufacturing operations. Consequently, engineers at the vehicle manufacturer are finding or developing the ways and means to increase the organization's ability to handle variations within the plants in short order. Linn knows more than a little something about this activity, as he is the principle engineer of Robotics within GM's Controls, Conveyors, Robotics & Welding Group (CCRW). Prior to taking this position at the start of 2004, he spent two years as a Total Integration engineer for Advanced Technology for the Body-in-White Group. And for about five years before that, he was the principal engineer for Robotics, Welding & C-Flex within CCRW. One of the things that is consistent about all of these positions is that they are focused on increasing flexibility. Which is one of the things that General Motors is aggressively pursuing. In fact, Linn says, "We want to be the most flexible auto manufacturer in the world." The plan is not to simply have assembly plants being capable of building variations off of a single product architecture, but of actually building multiple architectures in a single plant. In order to achieve this capability, it means the utilization of robots.

Robots Abound. Linn says, "We use robots in every North American assembly plant." The roll call of the robots is pretty much what's expected. The big hitter: spot welding in the body shop. There is, he says, some arc welding, too. And there are material handling and material dispensing. The paint shops are heavily roboticized. And the metal components that are produced for assembly by the GM Metal Fabricating Div. have likely undergone press-to-press transfer with robots. "Robots are specified almost as commodities in every area of our assembly plants." Linn adds that there are numerous applications of the "bread-and-butter spot welding robot," but goes on to note, "In the last four years I've seen more and more applications of things like robots on 7th-axis slides for more capability." They're doing more "cooperative" operations, as when, say, one robot holds a part and two others weld it.

The objective, one that Linn mentions frequently, is to "eliminate style-specific machinery and equipment." In other words, hard tools and fixtures, long the standard in automotive assembly operations, are giving way to robots. In fact, Linn goes so far as to say, "We believe we're considerably ahead of pretty much anyone in the world with the use of flexible robotic fixtures and tooling." And he specifically points to the C-Flex robot. C-Flex is a programmable body shop tooling system that permits multiple body panels—including floor pans, deck lids, hoods, and engine compartments—from different models to be welded with the same set of programmable tools and robots. The benefits that are achieved by doing this are both impressive and varied. For one thing, the programmable approach can save $100 million in a body shop when introducing new models—and given that GM is in the process of rolling out more new or heavily modified cars and trucks than at any time in its history—in 2004, 20 models are being launched, and 15 of 30 North American plants are involved in program launches—that savings adds up fast. C-Flex saves space, too, reducing requirements in the body shop by as much as 150,000 ft2. Programmability saves time, as well. GM has installed C-Flex in a number of its assembly plants.

The GM-developed C-Flex system
The GM-developed C-Flex system is saving a tremendous amount of money in the body shops as flexible automation is used in place of product-specific tooling.

There are still places where they have yet to install robots. For example, he notes that chassis marriage is a place where there are "style-specific, big, expensive fixtures. We'd like to roboticize that application, increase our flexibility and eliminate the style-specific capital investment that's dependent on one model."

Multi-program Investment. Historically, one of the inhibitors to the implementation of truly flexible equipment is that investments for tooling and equipment were tied to specific programs. Although there would be the opportunity to use the flexible equipment for follow-on programs, if the goal is to control spending on program A, what's the incentive to possibly pay a little more for flexibility even though it would be of benefit to program B? Linn admits that while they have struggled with that, he points out, "We've gotten past that with robots for some time now." Linn explains, "The way our financial people allocate funds for programs, you're looking at making an investment for multiple programs—amortizing those costs over multiple programs."

Reuse is now part of the equation. Of course, Linn acknowledges of engineers, "We like to use the latest and greatest, newest and best." But the flexible equipment is often serving the purposes of permitting quick changeovers and getting the job done. So while there is extensive reuse of robots, that's not always the case. "As we roll programs over, we still buy new robots and equipment. We have some applications where the robots have been beat up because of the payload and duty cycle. So we have to refurbish or replace them. What we're seeing in some of the new programs is that we're doing a lot of reuse, but we're going out and buying new robots to augment what we're reusing."

Linn says that they've standardized on robot suppliers including Fanuc, ABB, KUKA, and Comau. "We have those vendors' products in our plants around the world." But this is not to say they aren't interested in new technology developments that may come from other equipment vendors: "With any technology—whether it is a vendor's product or new technical concept—we have a validation process in place. We go through and do the business case and engineering analysis. If it all works out and it is the right thing to do, we validate it. We'll purchase the equipment and try it out in plants before we actually standardize on it." He goes on to say that just because they aren't using something doesn't mean they aren't interested in something that can provide a competitive advantage.

Making It With Math. Linn says, "We do off-line programming in 99.9% of our applications. It works very well. The things that mess us up from an integration perspective are the things that aren't modeled well—hoses and cables and things that flop around are not modeled well in the off-line packages. We end up going in and touching up points and making sure we don't hook cables on clamps." Yet even this is being minimized as robots are being fitted with integrated dress packages so that there is less (or no) flopping. Off-line programming, of course, is part and parcel of GM's on-going efforts to have math-driven product and process development. Specific to manufacturing, Linn says, "We want to take everything in the manufacturing process and program it off line, verify and certify that everything runs, and then be able to download it onto the shop floor. The things that screw us up now are the things that don't model well."

It's interesting to note that one thing that's not screwing them up are robots themselves. Once derided for their glitches (real or imagined), Linn states, "The reliability of robots is astounding. The robot is one of the most reliable pieces of equipment that we have in any part of the plant." Which, given their comparative sophistication, capability, and flexibility is an impressive accomplishment, indeed.

Robotic System for Faster Aluminum Welding
Tim Nacey, Industrial Group manager, Panasonic Factory Automation (Elgin, IL), notes that the number of aluminum welding applications in automotive has taken an up tick during the past couple of years. Which is certainly of interest to Panasonic because they've had a particular interest in that for the past 20 years. Recently, they've developed what's called the MIG Force push-pull write feed system that is specifically designed to facilitate aluminum MIG welding—as in providing the means to make lap welds at 5 m/min and fillet laps at 2 m/min. The system works with the company's G-series robots. What's important about those robots being matched with this system is the fact that they are equipped with a 64-bit RISC controller, so that the robot controller not only controls the movements of the robot itself, but also the wire feed system and the wave form programs (as distinct from conventional arc welding systems, wherein there tends to be separation between these functions). Nacey explains that in order for there to be fast, good, arc welding, it's necessary to have coordinated control of the wire feed speed, pulse waveform, and robot movement.

The wire feed system has a planetary roller arrangement that straightens the wire so that as it moves to the joint it isn't curved, and therefore more accurate. Importantly, this feeding system is coordinated with a buckling detection circuit, which is particularly important when small-diameter wire is used. This setup helps assure reliable arc starting, which is particularly important when there is stitch welding performed (i.e., when the arc is struck and stopped repeatedly). If there is a change in the motor load that indicates the potential of wire buckling, the controller shuts down the wire feed drive motor before buckling occurs.

Because of the nature of aluminum, welding can be tricky because of the transfer of heat from the weld to the aluminum being welded. One consequence of this can be a change in the weld joint. The Panasonic system accommodates this via its waveform control by adjusting the amount of amps without reducing the welding speed.

Nacey points out that while some people are looking at hybrid laser systems for aluminum welding, the Panasonic system provides 80% of the speed at one-tenth the price.