Sheet Metal Forming Simulation
Professor, I'm starting to work on automotive stamping simulation. I've seen terms like "springback," "wrinkling," "FLD" thrown around everywhere. Can you give me a mental map of what sheet metal forming simulation actually covers?
Sure — let me paint the picture. Sheet metal forming starts with a flat blank, usually 0.5–3 mm thick steel or aluminum. You press it between a punch and die, and through plastic deformation, it takes the shape of the tool cavity. Sounds simple, but the physics are anything but. Different regions of the blank experience very different stress states simultaneously: the flange is in biaxial compression and might wrinkle; the draw wall is in uniaxial tension and might fracture; the punch nose region is in biaxial tension and thins out. Understanding where each failure mode is active — that's what forming simulation gives you.
What's the single biggest challenge in sheet metal forming simulation? I've heard springback mentioned a lot.
Springback is arguably the number one headache. After you form a part and release it from the tool, it elastically recovers — the part geometry shifts from what the die geometry intended. For a high-strength steel hood panel, you might have designed the die with a 15° angle, but after springback you measure 17°. Now your part doesn't fit the assembly. The challenge is that springback is sensitive to almost everything: material elastic modulus, yield stress, friction coefficient, blank holder force, and the strain path history through the thickness. Predicting it within 1–2° is the current industry benchmark, and even that requires careful material model calibration.
You mentioned the die geometry is like a rigid body — does that mean I model the die as a rigid surface in FEM?
Almost always, yes. The die, punch, and blank holder are orders of magnitude stiffer than the thin sheet, and their elastic deformation is negligible. Modeling them as rigid bodies saves enormous computation time — you only need to discretize their surfaces with contact facets, not a volumetric mesh. The deformable blank is meshed with shell elements, typically about 50,000–200,000 elements for a car body panel. All the plastic deformation, thinning, and springback happen in the sheet mesh.
What's the Forming Limit Diagram? I keep seeing it in AutoForm results.
The FLD is the most practical tool in forming engineering. It's a chart with major strain on the vertical axis and minor strain on the horizontal axis. You plot where every element in your simulation lands — each element's strain state is a dot on the diagram. There's a forming limit curve (FLC) that separates safe strains (below the curve) from fracture (above). There's also a wrinkling limit in the lower-left compression region. When you look at the color-coded FLD plot from AutoForm, green means safe, orange means marginal risk, and red means the part will fracture or wrinkle in production. It's an immediate visual check of formability.
I've also seen "Hill 48" and "Yld2000" mentioned as yield criteria. What's the difference, and which should I use?
Both are anisotropic yield criteria for rolled sheet metal — they account for the fact that a rolled sheet has different properties in the rolling direction, transverse direction, and thickness direction. Hill 48 is simpler: it's a quadratic yield function with parameters calibrated from the r-values (Lankford coefficients) measured at 0°, 45°, and 90° to the rolling direction. It's well-established, supported everywhere, and adequate for mild and high-strength steels. Yld2000-2d (Barlat 2000) is a higher-order criterion that also uses biaxial test data to calibrate additional parameters. It gives significantly better predictions for aluminum alloys and advanced high-strength steels where Hill 48 shows noticeable deviation. In practice: use Hill 48 for standard steel forming, upgrade to Yld2000 for aluminum or AHSS if your software supports it and you have the biaxial test data.
For the FEM setup, why are shell elements used instead of solid elements? I thought shell elements were less accurate.
For thin sheet metal — typical thickness-to-span ratio of 1:100 or more — shell elements are actually more accurate than solid elements at the same computational cost, not less. A solid element mesh for a 1 mm thick panel would need at least 5–7 elements through the thickness to capture the bending stress gradient accurately, multiplying the element count by 5–7 times. The Belytschko-Tsay (BT) shell element used in LS-DYNA is particularly efficient: it's a 4-node quadrilateral with 5 integration points through the thickness, reduced integration in the plane to avoid shear locking, and hourglassing control. For springback-critical work, the 7 integration-point version is recommended to better capture the through-thickness stress gradient that drives springback.
So the standard workflow is: run forming with explicit, then run springback with implicit?
That's the industry-standard two-step workflow, yes. Step one: forming simulation using explicit time integration — handles the fast contact dynamics robustly. Step two: springback — you release the tooling boundary conditions and let the part spring back under its own residual stresses, using implicit quasi-static analysis. The explicit springback approach is also used but requires many more time steps to damp out dynamic oscillations. Some tools like AutoForm handle both steps internally. In LS-DYNA you'd use LS-DYNA/Explicit for forming and then switch to LS-DYNA/Implicit for springback, mapping the stress state between the two analyses.
Sheet Metal Forming Processes
Sheet metal forming encompasses a broad family of manufacturing processes that plastically deform flat metal blanks (sheets) into complex three-dimensional shapes without removing material. Key processes include:
- Stamping (pressing): A punch presses a blank into a die cavity. Most common in automotive body panel production. A single press stroke produces one part; progressive dies form multiple features in sequence.
- Deep drawing: Forming a flat blank into a cup or box shape by drawing it over a punch. The blank flange is held under pressure and pulled inward. Used for beverage cans, fuel tanks, sink bowls.
- Bending: Simple angular deformation along a straight or curved line. Springback is the primary concern; overbending is the standard compensation strategy.
- Roll forming: Continuous bending of a strip through a series of roll pairs, gradually forming a constant cross-section profile. Used for automotive door frames, roofing sections, structural tubes.
- Hydroforming: High-pressure fluid replaces the punch, pressing the blank or tube into the die. Enables complex shapes and reduced springback compared to die stamping.
- Stretch forming: A blank is gripped at its edges and stretched over a form die under tension. Used in aerospace for fuselage skin panels with compound curvature.
Key Forming Phenomena
Plastic Deformation and Thinning
Sheet metal forming operates well beyond the elastic limit. The material yields and plastically flows — permanently taking the die shape — while conserving volume (no density change during plastic deformation). Local thinning occurs wherever the material is stretched biaxially or uniaxially. Excessive thinning — typically defined as a thinning ratio $t/t_0 < 0.75$ (25% thinning) — indicates risk of fracture at that location. Draw ratio optimization and blank shape optimization aim to keep thinning within acceptable limits.
Springback
Springback is the elastic recovery that occurs when the tooling force is removed. It is the primary source of dimensional error in formed parts. The springback ratio for bending is often expressed as:
$$\frac{\Delta\kappa}{\kappa} = \frac{2\sigma_y}{E \cdot t \cdot \kappa}$$
where $\kappa = 1/\rho$ is the forming curvature, $\sigma_y$ is the yield stress, $E$ is Young's modulus, and $t$ is the sheet thickness. This shows why high-strength steels (large $\sigma_y/E$ ratio) spring back much more than mild steel — a direct consequence of the wider elastic range. Advanced high-strength steels (AHSS) with $\sigma_y > 700$ MPa can spring back 5–10× more than equivalent mild steel.
Wrinkling
Wrinkling is a compressive instability of the sheet in regions where the in-plane compressive stress exceeds the wrinkling critical stress. The flange of a deep-drawn cup is especially prone to wrinkling. Blank holder force (BHF) prevents wrinkling by applying a through-thickness compressive stress that stiffens the sheet against buckling. Too low BHF → wrinkles form; too high BHF → material can't flow inward and fractures at the punch radius.
Fracture
Fracture occurs when the local strains exceed the material's forming limit, as captured by the FLD. In simulation, the distance from each element's strain state to the FLC (the "FLD safety margin") is the key output metric. Typical engineering targets: >10% safety margin in production, >5% in prototype tryout.
Earing
Earing is the formation of uneven cup heights (ears) around the rim of a deep-drawn cup, caused by planar anisotropy in the sheet. The number and position of ears is controlled by the crystallographic texture of the sheet: high $\Delta r$ (planar anisotropy) produces 4-eared cups; isotropic material produces no ears. Earing is predicted by anisotropic yield criteria and affects material utilization.
Forming Limit Diagram (FLD)
The FLD is the fundamental tool for forming feasibility assessment. It plots principal strains ($\varepsilon_1$ major, $\varepsilon_2$ minor) with a Forming Limit Curve (FLC) separating safe and fractured states.
| Region | Strain State | Process Mode | Risk |
|---|---|---|---|
| Right side, above FLC | $\varepsilon_1 > 0$, $\varepsilon_2 > 0$ | Biaxial stretching | Fracture |
| Left side, above FLC | $\varepsilon_1 > 0$, $\varepsilon_2 < 0$ | Plane strain / draw | Fracture |
| Lower left region | $\varepsilon_2 \ll 0$ | Compression | Wrinkling |
| Safe zone (below FLC) | Any combination | — | Safe forming |
FLCs are measured experimentally using the Nakajima or Marciniak test. They depend on sheet thickness, strain rate, and temperature. The Keeler-Goodwin formula provides an estimate of the minimum FLC value (at plane strain): $\varepsilon_1^0 \approx (0.233 + t/2.54) \times n/0.21$ for mild steel.
Yield Criteria for Sheet Metal
Hill 1948 (Hill 48)
The Hill 48 quadratic anisotropic yield criterion generalizes von Mises for orthotropic materials. For a sheet with principal axes aligned with the rolling (RD), transverse (TD), and normal (ND) directions:
$$F(\sigma_{22}-\sigma_{33})^2 + G(\sigma_{33}-\sigma_{11})^2 + H(\sigma_{11}-\sigma_{22})^2 + 2L\sigma_{23}^2 + 2M\sigma_{31}^2 + 2N\sigma_{12}^2 = 1$$
Parameters $F$, $G$, $H$, $N$ are calibrated from r-values measured at 0°, 45°, 90° to the rolling direction:
$$r_0 = H/G, \quad r_{90} = H/F, \quad r_{45} = \frac{N}{F+G} - \frac{1}{2}$$
Normal anisotropy $\bar{r} = (r_0 + 2r_{45} + r_{90})/4$ — high $\bar{r}$ means the sheet resists thinning, improving deep drawability.
Yld2000-2d (Barlat 2000)
A non-quadratic yield function for in-plane loading of orthotropic sheets, using two linear transformations of the stress tensor to produce a more flexible yield surface shape:
$$\Phi = |X_1' - X_2'|^a + |2X_2'' + X_1''|^a + |2X_1'' + X_2''|^a = 2\bar\sigma^a$$
Eight coefficients calibrated from 7 mechanical tests (3 r-values, 3 yield stresses at 0°/45°/90°, 1 biaxial test). Significantly better for aluminum alloys and AHSS where Hill 48 underpredicts the equibiaxial yield stress.
FEM Setup for Forming Simulation
Element Selection: Belytschko-Tsay Shell
The Belytschko-Tsay (BT) shell is the workhorse element for sheet forming in LS-DYNA and most forming-specific codes:
- 4-node quadrilateral, one integration point in-plane (reduced integration), 5–7 integration points through thickness (Gauss or Lobatto)
- Hourglass control required to prevent zero-energy spurious deformation modes from reduced integration
- Consistent with small elastic strains but large rigid body rotations — appropriate for thin sheet forming
- For springback accuracy: use 7 integration points through thickness (compared to 5) to better represent the bending stress gradient that drives elastic recovery
Contact Modeling
Sheet-to-tool contact uses surface-to-surface contact with Coulomb friction $\tau = \mu p$. Typical friction coefficient $\mu = 0.10–0.15$ for lubricated steel-to-steel. Die and punch modeled as analytical rigid surfaces (no mesh deformation) — much faster than deformable tool models. Blank holder force applied as a uniform pressure on the blank holder rigid surface.
Solver Strategy
- Forming phase: Explicit time integration. Mass scaling used to artificially increase the critical time step (ensure kinetic energy < 10% of internal energy to confirm quasi-static validity).
- Springback phase: Implicit quasi-static analysis. Release contact constraints, apply gravity. Solve with Newton-Raphson iteration to convergence.
- Trimming (optional): Remove flange material after forming; re-run springback to capture the effect of trimming on residual stress redistribution.
Software Comparison
| Software | Solver Type | Key Strength | Industry Adoption |
|---|---|---|---|
| AutoForm | Implicit incremental + one-step | Purpose-built UI, die compensation, fast springback workflow | Dominant in automotive OEM |
| LS-DYNA | Explicit + Implicit | Most material models, research-grade accuracy, crash integration | Widely used, general-purpose |
| PAM-STAMP | Explicit | Multi-stage progressive die, good draw bead modeling | European automotive |
| Abaqus/Explicit | Explicit + Standard | Seamless explicit-implicit switch, user subroutine flexibility | Aerospace, research |
| eta/DYNAFORM | LS-DYNA based | Turnkey forming pre/postprocessor for LS-DYNA | Stamping shops |