Stamping Out Springback

By Mark Broadworth

Springback is an enormous challenge in sheet-metal stamping. It iswidely accepted by most people involved in tool and die engineeringthat FEA, while being accurate for predicting formability issues, isnot an accurate tool for predicting springback. Atlas Tool, however, isproving that with a sound engineering process, clean CAD data, anexpert design and build team, and ESI’s PAM-STAMP 2G it can accuratelycompensate for the effects of springback for high-strength anddual-phase steel for even the most aggressive product geometry.

Ultra-high strength and dual-phase steels makes the stamping processmore challenging than with other materials since they tend to springback three to four times more than low-carbon sheet steel. Dual-phasesteel combines two of the many crystalline structures that steel formsinto, ferrite and martensite, to achieve greater strength thantraditional high-strength steels.

> >Figure 1: Cross-section of virtual die during forming simulation.

Dual-phase steel provides yieldstrengths between 550 megapascals (MPa) and 1000 MPa, whileconventional high-strength steels fall within the range of 210 to 550MPa. The result is improved crashworthiness and weight reduction.However, dual-strength steel provides some significant formabilitychallenges. The material tends to work-harden to the point that itcannot be reformed in subsequent operations. Dual-phase steels alsogenerate higher levels of springback, which can create dimensional partquality problems.

Several years ago, Atlas Tool made the decision to upgrade tosimulation software that was capable of incrementally simulating theforming process with as many steps as needed to provide a much higherlevel of accuracy. We evaluated several different software packages andselected PAM-STAMP 2G because we found it to be both more accurate thanthe other programs and easier to use.

< < Figure 2: An early springback prediction created with PAM-STAMP 2G.

Because we drive our design process using PAM-STAMP 2G, we areconfident we can go to hard tool directly from “virtual-tryout” knowingwe will be better than 95% fit to gauge, eliminating the need forprototype-tryout tooling (see Figure 1, above). This includeschallenges associated with laser-welded blanks, patch-welded blanks,and ultra-high strength steel including dual-phase steel.

In the beginning of the process, Atlas Tool engineers outline apreliminary die process and a draw-die development is drawn in CADsoftware from UGS. Care is taken to ensure that the binder and addendasurfaces are tangent to one another and the binder surfaces are fully"developable.” These steps help ensure better manufacturability whenmachining and building the tools. This also makes certain that thetools closely represent the FEA model.

Modeling physical drawbeads in CAD is one of the keys to a successfulspringback prediction. But unless physical drawbeads are modeled andused in FEA, springback prediction will not be accurate because thework-hardening that exists from the bending and unbending of materialthrough the drawbeads will not be present (see Figure 2, above).

After all the formability issues are solved using PAM-STAMP 2G, thetask of compensating the CAD model for springback begins. PAM-STAMP2G’s fully automated springback compensation module is a key componentin this process. The module provides a mapping of cloud-points to usewhen compensating the CAD data. These points are used inside of thethinkcompensator module, a product of think3 Inc. The thinkcompensatormanipulates the surface data using the two sets of cloud points fromPAM-STAMP 2G to provide an accurate surface model to use to build thetools.

Communication between all departments is a vital consideration in asuccessful process. The same data that was used in the die simulationis used when cutting the tooling. High-speed machining techniques alongwith accurate CNC machines and highly skilled operators are also anintegral part of the equation.

> > Figure 3:  The trimmed A-pillar sitting on draw-punch.

The tools are cut very accurately using the most sophisticated andaccurate machining methods to ensure that very little handwork will benecessary and that the tools will be a very close representation of thestamping simulation. Once the tools are in the press, the pressoperator works with the simulation department to make sure all theboundary conditions are the same as in the FEA. The binder tonnageinformation is passed on from PAM-STAMP 2G as well as the initial blankoutline and the amount of blank draw-in during forming. Duplicating theresults of the FEA is critical in a successful process.

This process was used in the engineering of an “A-pillar” for a majorautomobile manufacturer. A-pillars are the sheet-metal columns betweenthe windshield and front windows of an automobile. Strong A-pillars arevital to the safety of the passengers in a vehicle because they preventa vehicle’s roof from caving in during a rollover accident. A-pillarsare becoming more difficult to form because automobile OEMs areincreasingly specifying that they be built from high-strength anddual-phase steel.

This particular A-pillar exhibited very aggressive geometry with a3.5-in. draw depth, which is considered a fairly “deep” draw for a partof this nature (see Figure 3, above). The fact that it required alaser-welded blank consisting of both 600DT dual-phase steel and 50KSIhigh-strength steel also made the job challenging.

Early springback analysis using PAM-STAMP 2G showed about 18mm of twistthroughout the part after forming and trimming operations werecompleted (see Figure 2, above). Engineers at Atlas determined early onthat the draw-die would need to be compensated in order to produce adimensionally acceptable part. The compensation process using PAM-STAMP2G and think3 was then put to the test.

< < Figure 4:  First part on gauge dimensions.

The operating parameters that were optimized during the simulation wereclosely followed in producing the initial parts for the A-pillar—andthe results exceeded expectations. After compensating for about 18mm oftwist through the center axis of the part, the first part to gaugeshowed the A-pillar was accurate to within 1mm to the net surfaces onthe gauge.

In three weeks of simulation work, Atlas developed a design that thesimulation predicted would meet all of the customer’s requirements.

Speaking of this development process, Mark R. Schmidt, Atlas Tool’spresident, praised the usefulness of the ESI tool. “The use of anadvanced incremental simulation tool enables us to overcome theformability challenges posed by ]high-strength steels and dual-phasesteels] and meet our customers’ requirements in as little time aspossible,” he said. “We believe our expertise with PAM-STAMP 2G is asignificant competitive advantage.”

Mark Broadworth works for Atlas Tool as a tool engineer. Send yourcomments about this article through e-mail by clicking here. Pleasereference “Springback, November 2006” in your message.

Atlas Tool, founded in 1962, specializes in producing stamping dies andprototype parts for the automobile industry. The company buildsprototype parts to suit customer objectives, from simple handmade partsto completely tooled prototype programs that verify manufacturingprocesses.

Atlas has a wide variety of conventional and high-speed machiningequipment capable of operating at feed rates up to 600 inches perminute. Among its assets are 40 stamping presses, 38 CNC machines, 10CMMs, 46 CAD/CAM/CAM seats, and a crane capacity of 70 tons (64 metrictons). Experienced program managers are assigned at the start of atooling program and are involved in every step of a project.Solid-modeling technology is used for die design and CNC programming.—MB

Atlas Tool, Inc.
Roseville, MI

ESI North America
Bloomfield Hills, MI

think3, Inc.
Cincinnati, OH

UGS Corp.
Plano, TX