Page 48 - MetalForming September 2014

Tooling by Design
By Peter Ulintz
Understanding Formability Results
Process-modeling software makes science-based manufacturing knowledge readily available early in the product-design and process- quotation phases. Modeling techniques routinely employed for metal stamping often are referred to as metalforming simulation or virtual stamping. Design- ers select from two basic types of solvers (methods) for process model- ing: one-step solvers and incremental solvers, each having unique advan- tages and disadvantages.
One-step solvers, generally used during product development, process planning and quoting, primarily find use for assessing manufacturing feasi- bility. Incremental codes are used as final validation for completely defined product and process designs.
Capability and Limitations
Estimating engineers can take advantage of blank-prediction soft- ware, a derivative of the one-step method that provides users with a fast and accurate method for developing blank shapes, blank nesting and cost estimating. These solvers can capture sheetmetal stretching and compres- sion that normally occurs during the forming process, but which may not easily be accounted for using classical
Peter Ulintz has worked in the metal stamping and tool and die indus- tries since 1978. He has been employed with the Anchor Manufacturing Group in Cleveland, OH, since 1989. His back- ground includes tool and die making, tool engi- neering, process engi-
neering, engineering management and product development. Peter speaks regularly at PMA semi- nars and conferences. He is also vice president of the North American Deep Drawing Research Group. Peter Ulintz
pete.ulintz@toolingbydesign.com www.toolingbydesign.com
length-of-line measure- ments or unfolding soft- ware.
46 MetalForming/September 2014
www.metalformingmagazine.com
The primary disadvan-
tage of a one-step solver is
that it works only for sin-
gle-step forming opera-
tions. Due to simplifying
assumptions, one-step
codes compromise accu-
racy for the sake of time.
But the advantage is that
engineers can evaluate
numerous what-if scenar-
ios, in a matter of min-
utes, to help identify Fig. 1—Puckering in a safe zone important process param-
eters. Because of their speed and min-
imal input data, one-step codes rou-
tinely find use for evaluating product
designs and for establishing feasible
processing methods when little or no
process data is available.
Unlike one-step codes, incremental codes are not restricted to a single form- ing operation. Modeling all of the process steps allows strains generated in previous forming operations to be car- ried over to subsequent operations. This is important because strain histo- ry plays an important role in formabil- ity accuracy and springback prediction.
The advantage of incremental solvers is that they can be used to con- duct a series of virtual die tryouts using specific blank shapes and material properties with production-intent tool- ing geometry. In addition to formabil- ity assessment, designers use incre- mental solvers to conduct trim-line optimization, springback analysis and process-sensitivity studies.
On the other hand, incremental solvers inherently are more sophisti- cated, requiring longer learning curves, more preprocessing steps and much longer computing time compared to one-step methods. Depending on
Interpreting the Results
Understanding the capabilities and limitations of simulation software is necessary in order to accurately inter- pret formability results.
Blank-prediction software relies on simplifications for preprocessing and solving in order to minimize the required amount of input data. The software quickly provides useful blank- prediction results, but sacrifices forma- bility accuracy for ease of use and cost- effectiveness.
One-step codes are fast and easy to use. All they require from the user are final product geometry, usually in 3D CAD format, and minimal material properties, but more process data can be input compared to blank predic- tion software. The software takes the final product geometry and forces it into a flat blank. Resulting strains are calculated based on how much the material has moved, and then are mapped back onto the original product geometry.
If a one-step solver predicts exces-
process complexity and available com- puting power, this can take as long as a day or two, or as little as a few hours.