4. Mesh generation strategies for seat structures
Mesh generation is where the seat model's predictive fidelity is won or lost. A seat frame is a hybrid structure: thin stamped steel rails and recliner plates, thick foam blocks, cast recliner mechanisms, and dozens of welded or bolted joints. No single element type or mesh density serves all of these. This chapter establishes a disciplined strategy for choosing shell versus solid elements, setting mesh density, enforcing quality criteria, and representing welds and bolts so that the LS-DYNA/PAM-CRASH model behaves like the physical seat under crash loading.
Shell versus solid element selection
The governing rule is the through-thickness stress state. Use shell elements (typically 4-node Belytschko-Tsay or fully integrated ELFORM=16) when the structure is thin-walled and bending-dominated, and through-thickness stress gradients are negligible. Seat tracks, recliner arms, backrest frames, and mounting brackets are stamped from 1.0–2.5 mm sheet steel and are ideal shell candidates. Shells are computationally cheap, and with five through-thickness integration points they capture bending and membrane behavior accurately.
Switch to solid elements when the component is genuinely three-dimensional: cast recliner housings, foam cushions, head restraint posts with large diameter, and any region where contact pressure or plastic hinge formation through the thickness matters. Solid elements are also mandatory where a shell mid-surface cannot represent the geometry — for example, a thick rubber isolator or a polyurethane foam block. A practical compromise for thick sheet (above roughly 4 mm) is layered shells, but for castings use hexahedral or tetrahedral solids with at least two elements through the thickness.
Transition regions between shells and solids are a common source of error. Do not connect a shell edge directly to a solid face with shared nodes unless the shell mid-surface is offset to the solid's mid-plane. Instead, use a tied contact or a transition element, and verify that the offset is applied consistently. In PAM-CRASH, shell offsets are defined per property; a missing offset produces an artificial membrane stiffness that can shift intrusion results by 10–20 mm in a full frontal sled test.
Mesh density and target element size
Mesh density should be driven by the deformation wavelength you need to resolve, not by a uniform global size. For seat structures, a workable baseline is:
- Primary load paths (seat tracks, recliner arms, belt anchorages): 5–8 mm shell edge length, with at least three elements across any flange width.
- Secondary structure (backrest frame, head restraint support): 8–12 mm.
- Foam and trim: 10–15 mm solid elements, refined to 8 mm near the occupant contact surface.
- Local refinement zones: 2–4 mm around weld nuggets, bolt holes, and plastic hinge locations.
A useful convergence check is to refine the critical region by a factor of two and confirm that peak plastic strain and intrusion change by less than 5%. If the result moves more than that, the coarse mesh was under-resolved. Avoid excessive global refinement: doubling the element count in a seat model roughly quadruples explicit solver cost, and the smallest element governs the stable time step.
Quality criteria
Poor element quality causes negative volumes, spurious hourglassing, and premature termination. Apply the following acceptance thresholds before submitting a job:
| Metric | Shell target | Solid target |
|---|---|---|
| Jacobian / aspect ratio | < 5 | < 5 |
| Warpage | < 15° | n/a |
| Skew angle | < 60° | < 60° |
| Quad interior angle | 45°–135° | 45°–135° |
| Tria percentage | < 5% | n/a |
| Minimum edge length | ≥ 2 mm | ≥ 2 mm |
Triangular shells are acceptable only in transition zones; they are overly stiff in bending and should not populate load-bearing flanges. For solids, check that no element has a negative Jacobian and that the minimum edge length is consistent with the intended time step.
Weld and bolt representations
Spot welds in a stamped seat frame are the most numerous connections and the most frequent cause of model instability. Three representations are common:
- Shared nodes / rigid links: fastest, but over-stiff and unable to fail. Use only for non-critical welds.
- Beam or solid nugget with failure: a single beam element (or a small cluster of solid elements) connecting the two sheets, with a plastic strain or force-based failure criterion. This is the recommended approach for crash-critical welds because it can rupture and release the joint.
- Spot-weld contact: tied contact with a failure threshold; convenient when weld positions are dense and irregular.
For a typical 6 mm weld nugget, model a beam of equivalent diameter with a failure plastic strain around 0.15–0.25 for mild steel, calibrated against a lap-shear or coach-peel coupon. Bolted joints should be represented by a rigid spider or beam at the bolt axis plus a contact definition between the clamped plates; do not simply merge the nodes, as this removes the joint's rotational compliance and can over-predict seat back stiffness.
Practical workflow and verification
A reproducible meshing workflow for a seat frame proceeds as follows. First, import the CAD and mid-surface the thin-walled parts; assign shell properties with correct thickness and offset. Second, mesh the primary load path at 5–8 mm and refine weld and bolt locations to 2–4 mm. Third, mesh cast and foam components as solids with at least two elements through the thickness. Fourth, define welds as beam or solid nuggets with calibrated failure strain, and bolts as rigid spiders plus contact. Fifth, run the quality checks in the table above and fix any element exceeding the thresholds. Finally, run a short (5–10 ms) verification simulation and inspect the energy balance: hourglass energy should stay below 5% of internal energy, and the added-mass percentage should remain under 1%.
Typical failure symptoms and their causes: negative volume errors usually indicate a distorted element near a weld or a missing offset; excessive hourglass energy points to under-integrated shells with insufficient refinement; and premature weld rupture in the verification run suggests the failure strain was calibrated on the wrong coupon or the nugget diameter is too small. Each of these is diagnosable from the model's quality report and energy summary before the full crash run is attempted.