A new process called double-sided incremental forming holds promise for low-volume parts
When a legacy automotive or aircraft part breaks and needs a replacement, manufacturers currently have no good options. To be prepared, they could keep expensive dies for each part indefinitely. But the large dies require a lot of storage space. They could also create new dies as needs arise. But designing and building a die can take 15 weeks and cost hundreds of thousands of dollars.
Academics, automakers and others are looking at an emerging technology known as double-sided incremental forming (DSIF). Alternate names are dual-sided dieless forming and Ford Freeform Fabrication Technology/F3T. No matter the name, it involves two robots working simultaneously on one sheet of metal—from the bottom side and the top side.
DSIF shows promise for Batch Size One, prototyping, legacy parts and any low-volume production, said Jian Cao, associate VP for research at Northwestern University’s McCormick School of Engineering.
In the automotive sector, she said, DSIF, compared with traditional processes, could:
- Reduce design-to-production time to less than one week from six to 15 weeks;
- Require only a small and simple facility size compared with a larger, more complex facility;
- Require no tooling storage, and
- Offer broad material choices compared with limited choices in additive.
“The fender for a car is not that costly but if you go to a body shop and try to get a fender not currently in production, you have to pull out the old die, stamp it, then make the fender,” said Cao, who is also a mechanical engineering professor. “Even with state-of-the-art equipment, it can take six to 12 weeks to make the die, do the calibration, test it and get the part.”
The DSIF process can produce geometries without the cost of multi-million-dollar, large die presses and matched die sets, said Alan Taub, a University of Michigan professor and tech advisor for LIFT, a Manufacturing USA institute.
“You’re avoiding the cost and time associated with producing those dies,” he said. “I can produce geometries from the sheet without multi-million-dollar die presses and matched die sets that cost hundreds of thousands of dollars. It takes longer to form each part because you’re doing it incrementally. If you’re going to make thousands of units a year, you should do it with conventional processes.”
How DSIF works
Cao compared DSIF to holding a clay figure in your hands, using your fingers on each side to deform the doll to a different shape. By working two sides at the same time, achieving great geometric accuracy is possible—at least when robots are doing the work.
“Imagine a generic tool, two-sided with one on top of the sheet and the other on the bottom of the sheet surface,” she said. “We are doing CNC deformation, not taking any material away but deforming the sheet to a different position. The top and the bottom work collectively. Gradually, the tool forms the material into the final geometry.”
In a factory, an automated, two-sided computer numerical control (CNC) robotic tool could deform the top and bottom of a sheet of metal at the same time, Cao said. Although some processes employ single-sided incremental forming, working on only one side of the metal at a time means the tool can only poke at as opposed to being able to truly shape the metal.
“If you work only on the top, the geometry you generate is only in that direction,” she said. “There’s little control in terms of the boundary. One can deform the top side, push the metal down, and then flip the part and push it again. The previously deformed pieces can be pushed to new positions.
“However, there’s little control in terms of the boundary and geometric accuracy.”
Another advantage of DSIF is the metal can stretch more without failure, Cao said.
“It’s the so-called noodle theory,” she added. “If you take a spaghetti noodle and pull on both sides, you can break it pretty quickly. But if you manipulate and stretch if little by little, incrementally, you can deform it more, make it much longer.”
Where DSIF makes sense
DSIF will not replace traditional stamping and die tech, which uses dies to form and shape metal and is well-suited to make large-volume sheet metal parts with a typical cycle time of less than a minute, Cao said. But because DSIF eliminates the need, expense, storage and design/build requirements of using dies, the process shows promise for low-volume production of legacy, prototyping and personalized products.
As of now, the cost benefit for DSIF is at a volume of 300 to 500 parts per year, Taub said. The goal is to make it cost-effective at 1,000 parts per year.
In the automotive sector, low-volume derivatives and legacy parts for classic/antique vehicles are vi-able use cases, he said.
“Handling legacy parts is a major issue for automakers,” Taub said. “Dies required for one car can be in the hundreds. Normally, automakers plan for 15 years of use. But they have to keep those dies around for spare parts for people who keep their cars longer. Think of all the storage costs for legacy parts.”
For aircraft manufacturing, where production is often in the hundreds per year, “you’re right in the sweet spot” of annual production, Taub said. Thousands of airplanes and cars remain in the skies and on the roads after active production has halted. For example, Boeing built 744 B-52s from 1952 to 1962 and about 58 remain in service.
A close cousin
The process is related to single-sided incremental forming, which involves working on only one side of the metal at a time. LIFT is working only on single-sided incremental forming, said Chief Technology Officer Hadrian Rori.
DSIF does show promise, especially in some applications, he said.
In the double-sided process, one robot can act as a pseudo-die for the other, helping to increase geometric accuracy, Rori said. Having two forming tools also allows an operator to form concave and convex geometries in the same part. With the doubled-sided process, users can define a set distance between the tools, allowing for control of material thinning.
Automakers studying DSIF
So far, DSIF remains in the research stage. Nissan and Ford (www.tinyurl.com/DSIFford), have been investigating it for years.
Cedric Xia, now at Apple, led a team at Ford to experiment with DSIF. Its focus was on cutting the time to prototype parts.
Prototyping often takes six to eight weeks. Xia and his team at Ford came up with F3T, a technology that reduced the time to an average of one week.
Later, Xia and Cao led researchers and engineers from Ford, Northwestern and MIT and further developed the technology on an industrial scale in a U.S. Department of Energy project. “We demonstrated we could scale it up,” he said. “On the commercial side, we need system integrators who can fabricate hardware and develop, maintain and integrate the software.”
Nissan Japan is researching dual-sided dieless forming, potentially to use to make replacement parts for discontinued models, the company said.
Challenges remain, particularly geometric accuracy.
A typical design tolerance for large-volume production to industry standards is less than 2 millimeters of a large panel, Cao said. In some cases, the machines, tool holders or robots are not still enough because the tool deforming the metal is itself experiencing a force and the resulting displacement causes shape deviation that puts the part out of compliance with industry standards.
Sheet metal itself has an elastic response to the deformation and bounces back. “Springback is not unique to incremental forming,” Cao said. “It exists in room-temperature metal-stamping operations. People have been spending a lot of time trying to understand springback and compensate” for it.
Using sensors, researchers have made some progress predicting and mitigating springback.
“Based on real-time sensing, when we do the forming, we can predict springback and vary the tool path in situ to compensate,” she said.
Finally, researchers and manufacturers must be able to validate the ultimate performance of the part, based on the deformation, Taub said.
“The way these parts are formed, the parts undergo a different deformation strain, which can result in a different microstructure,” he said. “You have to validate the mechanical performance.”