3. Vehicle body-in-white (BIW) modeling for roof crush

Roof crush simulation in LS-DYNA begins long before the first keyword is written. The fidelity of the body-in-white (BIW) finite element model determines whether the predicted force–displacement curve, roof intrusion, and occupant survival space are physically meaningful or merely numerical artifacts. This chapter walks through the complete BIW modeling chain for roof crush analysis: preparing CAD geometry, extracting mid-surfaces from thin-walled stampings, generating a crush-appropriate mesh, and assigning rate-sensitive material cards for both steel and aluminum structures.

3.1 Geometry Preparation and Cleanup

Roof crush loads are carried primarily by the A-pillars, B-pillars, roof rails, roof bows, and the roof panel itself. These components are stamped from thin-gauge sheet metal, typically 0.7–2.0 mm thick, and are joined by spot welds, structural adhesives, or laser welds. Before any meshing can begin, the CAD assembly must be cleaned and simplified.

Start by isolating the load-bearing body structure. Remove closures (doors, hood, trunk lid), interior trim, seats, glass, and powertrain components unless they are explicitly part of the crush load path. In LS-DYNA roof crush models, the doors are often removed but their latch and hinge attachments are represented by rigid or constrained nodes to preserve the structural boundary. Glass is typically omitted, though the windshield and backlight can be modeled as shell elements with a brittle failure material if their contribution to roof stiffness is under investigation.

Next, repair geometric defects that would otherwise propagate into the mesh: sliver surfaces, gaps between adjacent panels, overlapping faces, and non-manifold edges. Use the CAD cleanup tools in ANSA, HyperMesh, or ANSYS SpaceClaim to stitch coincident edges with a tolerance of 0.01–0.05 mm. For spot-welded flanges, ensure that the mating surfaces are offset by the sum of the two sheet thicknesses so that mid-surfacing produces parallel, properly spaced shells.

3.2 Mid-Surfacing Thin-Walled Stampings

Because roof crush components are thin-walled, shell elements are the correct idealization. Mid-surfacing extracts the neutral surface of each sheet metal part and assigns the original thickness as a shell section property. The workflow is:

1. Select the solid or surface representation of the stamped part.
2. Run the mid-surface extraction tool with a thickness detection tolerance (typically 5–10% of the nominal thickness).
3. Verify that the extracted surface lies at the mid-plane and that the thickness value matches the drawing.
4. Trim and extend the mid-surface to the theoretical intersection with adjacent panels so that the shell mesh connects cleanly at flanges and corners.

For aluminum extrusions used in some roof rail designs, the mid-surface approach still applies, but the extrusion may have variable wall thickness. In that case, assign a separate shell section for each thickness region rather than averaging.

3.3 Mesh Generation Strategy for Roof Crush

Roof crush is a quasi-static large-deformation problem. The mesh must be fine enough to capture local buckling and folding of the pillars, yet coarse enough to keep the explicit time step manageable. A target element size of 5–8 mm on the pillars and roof rails is a common starting point. The roof panel can be coarser, 10–15 mm, because it deforms in a more distributed manner.

Use quadrilateral-dominant meshing with a maximum aspect ratio of 3:1 and a Jacobian above 0.6. Triangular elements should be limited to transition regions and should not exceed 5% of the total element count. Around spot welds, refine the mesh locally to 2–3 mm so that the weld nugget and its surrounding heat-affected zone are represented by at least two element rings.

The following table summarizes recommended mesh parameters for the principal roof crush components:

After meshing, check for free edges, duplicate nodes, and initial penetrations. A penetration of more than 0.5 mm between adjacent shells will cause an initial contact force spike and may trigger premature buckling. Use the model checker in HyperMesh or LS-DYNA's own *INITIAL_PENETRATION diagnostics to resolve these before running the simulation.

3.4 Material Assignment for Steel and Aluminum

Roof crush involves strain rates ranging from 0.1 to 100 s⁻¹ in the pillars. Therefore, rate-sensitive material models are mandatory. For steel, the widely used LS-DYNA card is *MAT_PIECEWISE_LINEAR_PLASTICITY (*MAT_024) with a defined LCSS curve for true stress versus effective plastic strain and a LCSR curve for rate scaling. A typical high-strength steel (DP780) might use:

*MAT_PIECEWISE_LINEAR_PLASTICITY
$ MID  RO  E       PR    SIGY   ETAN   FAIL  TDEL
  1001 7.85e-9 210000 0.30  420.0  1200.0  0.0  0.0
$ LCSS   LCSR
  2001   3001
$ VP
  0.0

For aluminum, *MAT_PLASTIC_KINEMATIC (*MAT_003) or *MAT_024 with an appropriate rate curve works well. Aluminum 6061-T6 has a yield strength near 275 MPa and exhibits moderate rate sensitivity. The Cowper–Symonds parameters C and P can be used in *MAT_003 to capture rate effects:

*MAT_PLASTIC_KINEMATIC
$ MID  RO      E       PR    SIGY   ETAN   BETA
  2001 2.70e-9 70000   0.33  275.0  800.0  0.0
$ SRC  SRP
  40.0 5.0

Thickness must be assigned per shell section using *SECTION_SHELL with the ELFORM=2 (Belytschko–Tsay) formulation, which is robust for crash analysis. For aluminum extrusions with variable wall thickness, create multiple sections and assign them to the corresponding element sets.

3.5 Joining and Boundary Representation

Spot welds are represented by *CONSTRAINED_SPOTWELD or by solid hexahedron elements when the weld nugget's stiffness matters. For roof crush, a common practice is to use a single beam element with *MAT_SPOTWELD and a failure criterion based on normal and shear force. The weld spacing should match the CAD weld map, typically 20–40 mm along flanges.

Structural adhesive is modeled with *MAT_COHESIVE_MIXED_MODE or simplified as a tied contact with a defined failure stress. For a first-pass roof crush model, adhesive can be represented by shared nodes or *TIED_SHELL_EDGE_TO_SURFACE if the adhesive is not the focus of the study.

3.6 Model Verification Before Crush Simulation

Before applying the roof crush platen, run a modal analysis or a simple static load case to verify the model's stiffness. A common check is to apply a 1 kN vertical load at the center of the roof rail and compare the deflection against a hand calculation or a known reference. If the deflection is more than 20% off, investigate mesh quality, material properties, or missing connections.

The SVG below illustrates the complete BIW modeling workflow for roof crush, from CAD cleanup through material assignment, with the key decision points and verification gates.

3.7 Practical Notes on Steel vs. Aluminum BIW

Steel BIW structures are more forgiving in roof crush modeling because their higher ductility allows larger plastic strains before failure. The mesh can be slightly coarser, and the material card is well-characterized. Aluminum, by contrast, has lower ductility and a sharper transition to failure. For aluminum pillars, consider using *MAT_024 with a failure strain defined via *DEFINE_CURVE or switch to *MAT_ADD_EROSION to capture cracking. The mesh around aluminum extrusions should be finer, 3–5 mm, to resolve local buckling modes accurately.

When mixing steel and aluminum in the same BIW (e.g., aluminum roof panel on a steel body), pay attention to the thermal and mechanical mismatch at joints. Adhesive bonding is common here, and the cohesive zone parameters must be calibrated against lap-shear test data. A typical adhesive strength of 20–30 MPa with a fracture energy of 1–3 N/mm is a reasonable starting point for simulation.

With the BIW model prepared and verified, the next step is to define the roof crush loading platen, boundary conditions, and output requests. The quality of this modeling phase directly determines whether the simulation can predict the force–displacement response and roof intrusion with confidence.