3. Material modeling for seat structures
An automotive seat is a hybrid structure: stamped high-strength steel recliner plates and tracks, thin-gauge seat-back and cushion pans, welded sub-assemblies, and foam over a wire frame. In an LS-DYNA frontal or rear impact model, the accuracy of the occupant response depends less on the mesh density than on whether the constitutive and failure models reproduce the coupon-level behavior of each of these materials. This section covers the four modeling decisions that dominate seat CAE fidelity: elastic-plastic metal plasticity, strain-rate sensitivity, spot-weld and seam-weld failure, and the coupon-test validation loop that closes the process.
3.1 Elastic-Plastic Metals: Choosing the Right Hardening Law
Seat structural steels are typically modeled with *MAT_PLASTIC_KINEMATIC (MAT_003) or *MAT_024 (*MAT_PIECEWISE_LINEAR_PLASTICITY). The choice is not cosmetic. MAT_003 uses a bilinear isotropic-plus-kinematic hardening formulation controlled by the hardening parameter β (0 = purely kinematic, 1 = purely isotropic). For seat structures subjected to reverse loading — for example, a recliner pawl that loads, unloads, and reloads during a rear impact — kinematic hardening (β ≈ 0 to 0.3) is required to capture the Bauschinger effect. For monotonic collapse of a seat-back frame, isotropic hardening (β = 1) is adequate and cheaper.
MAT_024 accepts a true stress–true plastic strain table, which is the correct way to represent cold-rolled HSLA 340/590 or DP600 sheet. The conversion from engineering to true values must be done before the table is written:
ε_true = ln(1 + ε_eng)
σ_true = σ_eng * (1 + ε_eng)
Example — DP600, 1.5 mm sheet, from tensile coupon:
ε_eng = 0.10, σ_eng = 780 MPa
ε_true = ln(1.10) = 0.0953
σ_true = 780 * 1.10 = 858 MPaTwo practical rules apply. First, the plastic strain column must start at 0.0 with the yield stress, not the proportional limit; LS-DYNA will otherwise produce an artificial elastic offset. Second, the table must be extrapolated to the largest strain the element will experience — typically 0.5–1.0 for seat-back collapse — using a power-law fit (Swift or Voce) rather than a flat plateau, because a flat plateau causes premature numerical unloading and underestimates energy absorption.
ELFORM=2 (Belytschko-Tsay) with at least 5 through-thickness integration points (NIP=5) so that bending-induced plastic strain gradients are resolved. With NIP=2, the through-thickness stress state is too coarse and the pan will appear stiffer than the coupon test predicts.3.2 Strain-Rate Sensitivity
Seat structures see strain rates from 10⁻³ s⁻¹ (quasi-static recliner operation) to 10² s⁻¹ (rear-impact seat-back loading). Mild steel is moderately rate-sensitive; DP600 and TRIP steels are strongly rate-sensitive. Ignoring rate effects can under-predict seat-back strength by 15–30% in a 50 km/h rear impact.
In LS-DYNA, rate sensitivity is added through the Cowper-Symonds model, available in MAT_003 and MAT_024 via the C and P parameters:
σ_y(ε̇) = σ_y0 * [1 + (ε̇ / C)^(1/P)]
Typical values (from coupon tests, illustrative):
Mild steel: C = 40 s⁻¹, P = 5
HSLA 340: C = 800 s⁻¹, P = 6
DP600: C = 4000 s⁻¹, P = 4Alternatively, *MAT_024 supports a rate-dependent table via *DEFINE_TABLE with LCSS curves at discrete rates (e.g., 0.001, 0.1, 1, 10, 100 s⁻¹). This is preferred when the material shows a non-Cowper-Symonds rate dependence, which is common in AHSS. The table must be monotonic in rate and cover the full range the model will encounter; LS-DYNA extrapolates the last curve beyond the highest tabulated rate, which can be non-conservative.
3.3 Spot-Weld and Seam-Weld Failure
Seat frames contain 80–200 spot welds and several continuous seam welds. Modeling every weld as a solid element is impractical; the standard approach is a beam or solid spot-weld with a failure criterion.
For spot welds, *MAT_SPOTWELD (MAT_100) is the workhorse. It combines an elastic-plastic beam formulation with a force-based or stress-based failure surface. The failure criterion is defined by NRR, NRS, NRT (normal, shear, and bending exponents) and the corresponding ultimate forces FN, FS, FT:
(f_n / FN)^NRR + (f_s / FS)^NRS + (f_t / FT)^NRT ≥ 1 → weld fails
Illustrative values for a 5 mm nugget in 1.5 mm DP600:
FN = 8.5 kN, FS = 5.2 kN, FT = 3.1 kN
NRR = 2.0, NRS = 2.0, NRT = 2.0These forces must come from coach-peel, lap-shear, and cross-tension coupon tests, not from handbook values. The weld beam should be modeled with at least 2 elements along its length so that bending deformation is captured; a single-element weld beam is too stiff and will delay failure.
Seam welds are better represented with *MAT_100 in a solid or thick-shell formulation, or with a tied contact plus a stress-based failure criterion. For a continuous seam, the failure criterion should be evaluated along the weld line, and the weld should be allowed to fail progressively rather than all at once. A common error is to use a single force-based criterion for a seam weld that actually fails in peel; the peel stress must be included.
*MAT_100 with default parameters. The default failure forces are generic and will produce either no weld failure or premature failure. Always populate FN, FS, FT from coupon data and verify the weld failure sequence against a physical tear-down of a tested seat.3.4 Coupon-Test Validation Loop
Material cards are only credible if they reproduce the coupon tests used to calibrate them. The validation loop has four steps:
- Build a single-element or small-patch model of the tensile coupon (ASTM E8 geometry, 1.5 mm sheet, 12.5 mm gauge width). Apply displacement control at 1 mm/s for quasi-static and use
*MAT_ADD_EROSIONor element deletion to capture failure. - Compare force–displacement curves between simulation and test. The initial slope should match within 5%; the yield point within 3%; the ultimate force within 5%; and the elongation at failure within 10%.
- Run the same coupon at 0.1, 1, and 10 s⁻¹ to validate the rate-dependent parameters. If the simulation shows no rate effect but the test does, the Cowper-Symonds C value is too high or the rate table is missing.
- Validate the weld model separately with a lap-shear coupon: two 1.5 mm DP600 strips, 5 mm nugget, pulled at 10 mm/min. The simulated peak force and failure displacement must match the test within 10%.
Only after these coupon-level checks pass should the material cards be used in the full seat assembly. A seat model that fails to match a lap-shear coupon will not match a seat-back collapse, regardless of mesh quality.
3.5 Practical Checklist for Seat Material Cards
| Item | What to Verify | Typical Acceptance |
|---|---|---|
| True stress–strain table | Starts at ε_p = 0, monotonic, extrapolated to ε_p ≥ 0.5 | Yield stress within 3% of coupon |
| Hardening law | β selected for loading path (kinematic for reverse loading) | Bauschinger effect reproduced if applicable |
| Rate parameters | C, P or rate table covers 0.001–100 s⁻¹ | Flow stress increase matches coupon at 1 and 10 s⁻¹ |
| Spot-weld forces | FN, FS, FT from lap-shear, coach-peel, cross-tension | Peak force within 10% of coupon |
| Weld beam mesh | ≥ 2 elements along weld length | Bending deformation captured |
| Shell integration | NIP = 5 for thin pans | Bending stiffness matches coupon |
| Failure strain | Element deletion strain from coupon elongation | Elongation at failure within 10% |
The discipline that separates a predictive seat model from a decorative one is the coupon validation loop. Every material card in the seat assembly — the DP600 seat-back frame, the HSLA 340 tracks, the mild-steel recliner pawl, and each weld class — should have a corresponding coupon simulation that matches its test data before it is used in the full model. When a seat sled test later shows a weld tear at a location the model did not predict, the first place to look is not the mesh but the weld failure forces and the rate-dependent flow stress of the adjacent sheet.