3. Spot weld failure mechanisms

A spot weld is not a single material point. It is a small, rapidly solidified casting (the nugget) surrounded by a heat-affected zone (HAZ) and bounded by two sheet interfaces. Because the geometry, microstructure, and stress state all change within a few millimetres, the joint can fail along several distinct paths. For crash and durability simulation, the critical question is not simply "does the weld break?" but "which mechanism governs, and how does that mechanism depend on loading rate?" This section defines the three canonical failure modes, explains the mechanics that select between them, and shows how to translate that understanding into LS-DYNA joint models.

3.1 The Three Canonical Failure Modes

Interfacial fracture occurs when the crack propagates through the nugget–sheet boundary, roughly in the plane of the faying surface. The weld separates with little or no material pulled from either sheet. In a coupon test this produces a clean, button-free fracture surface. Interfacial failure is favoured by small nugget diameter, thin sheets, high hardness (martensitic) nuggets, and shear-dominated loading. It is generally the least desirable mode because it provides almost no energy absorption and is sensitive to weld quality.

Pull-out (plug) failure occurs when the crack leaves the interface and propagates through the HAZ around the nugget, tearing a plug of material out of one sheet. The fracture surface shows a distinct button. Pull-out requires the nugget to be stronger than the surrounding HAZ, so it is favoured by larger nuggets, thicker sheets, softer microstructures, and tensile or peel-dominated loading. Pull-out absorbs significantly more energy and is the preferred mode in crash-critical joints.

Base metal failure occurs when the sheet itself fails away from the weld, either by necking and tearing or by gross plastic collapse. This mode dominates when the weld is over-matched relative to the sheet (a very strong, large nugget in a thin or low-strength sheet), or when the specimen geometry concentrates strain outside the weld zone. In simulation terms, base metal failure means the joint model should not fail at all — the surrounding shell elements must capture the failure instead.

3.2 What Selects the Mechanism

The governing mode is set by a competition between the load path and the relative strengths of the three regions. Three variables dominate:

• Loading direction. Pure shear (lap-shear coupon) drives the crack along the interface and promotes interfacial fracture. Peel or cross-tension loading opens the interface and promotes pull-out. Real automotive joints see mixed mode, so the mode is a function of the local force ratio.
• Nugget diameter to sheet thickness ratio (d/t). Small d/t concentrates shear on the interface; large d/t shifts the critical location into the HAZ and favours pull-out. A common rule of thumb is that d/t below roughly 4√t (with t in mm) risks interfacial failure.
• Strength mismatch. A hard, martensitic nugget with a soft HAZ promotes interfacial fracture because the interface is the weak link. A soft nugget with a strong HAZ promotes pull-out.

Loading rate adds a fourth dimension. Under dynamic loading, strain-rate sensitivity differs between the nugget, HAZ, and base metal. HAZ material is often more rate-sensitive than the nugget, so at high rates the HAZ strengthens relative to the interface, shifting the balance toward interfacial fracture. This is why a joint that pulls out in a quasi-static test can fail interfacially in a crash — a critical point for model calibration.

3.3 Static vs. Dynamic Response

Under quasi-static loading, the joint response is governed by plastic deformation and ductile tearing. The force–displacement curve shows an initial elastic rise, a yield plateau, and a gradual drop as the crack advances. Interfacial fracture produces a sharp, brittle drop; pull-out produces a longer, more gradual softening. These shapes are the primary calibration targets for a joint model.

Under dynamic loading, inertia and strain-rate effects change both the peak force and the failure mode. The peak force increases, and the post-peak softening becomes steeper because crack propagation is faster and less stable. In LS-DYNA, this is captured by rate-dependent failure criteria. A common approach is to define the failure strain or failure force as a function of effective strain rate:

*MAT_ADD_EROSION
$#   mid        ex        mx        mn        mx        mn        mx        mn
         1       0.0       0.0       0.0       0.0       0.0       0.0       0.0
$#  sigp      sigvm      epsp      epsf      epst      epss      epss      epss
       0.0       0.0       0.0      0.25       0.0       0.0       0.0       0.0
$#  epsf      epsf      epsf      epsf      epsf      epsf      epsf      epsf
      0.25      0.30      0.35      0.40      0.45      0.50      0.55      0.60

Here the failure strain epsf is tabulated against strain rate. The values shown are illustrative: a quasi-static failure strain of 0.25 rising to 0.60 at high rate. The exact curve must come from coupon tests at multiple rates. The key point is that the same weld model can switch from pull-out to interfacial fracture simply by changing the rate-dependent failure threshold.

3.4 Mapping Mechanisms to LS-DYNA Joint Models

LS-DYNA offers several spot weld representations, and the choice depends on which mechanism you need to capture:

For MAT_100, the failure surface is defined by normal and shear force resultants. The model fails when the resultant exceeds a user-defined envelope. To distinguish interfacial from pull-out, you can define separate envelopes for normal and shear loading, or use a mixed-mode criterion. A practical calibration workflow is:

1. Run lap-shear and cross-tension coupon tests at quasi-static and dynamic rates.
2. Extract peak force and displacement at failure for each mode.
3. Fit the MAT_100 failure envelope to the shear (lap-shear) and normal (cross-tension) data.
4. Validate on a mixed-mode specimen, such as a coach-peel or KS-II test.
5. Check that the model predicts the correct failure mode, not just the correct peak force.

3.5 Practical Example: Rate-Dependent Mode Shift

Consider a 1.0 mm mild steel sheet with a 5.0 mm nugget (d/t = 5.0). Quasi-static lap-shear testing shows pull-out failure at 4.2 kN with 2.5 mm displacement. At a loading rate of 1 m/s, the same joint fails interfacially at 5.8 kN with only 0.8 mm displacement. The peak force rises, but the energy absorption drops sharply.

To reproduce this in LS-DYNA, the MAT_100 shear failure force is defined as a function of strain rate. At quasi-static rate, the shear failure force is set to 4.2 kN. At 1 m/s, it is set to 5.8 kN, but the failure displacement is reduced to 0.8 mm. The model then predicts the correct mode shift. If the failure displacement is not reduced, the model will over-predict energy absorption and under-predict the severity of the crash event.

3.6 Summary

Spot weld failure is a competition between three mechanisms: interfacial fracture, pull-out, and base metal failure. The winner is determined by loading direction, nugget geometry, strength mismatch, and loading rate. In LS-DYNA, the joint model must be calibrated not only to the correct peak force but also to the correct failure mode and energy absorption. The next section builds on this foundation by introducing the specific material cards and failure criteria used to represent each mechanism in a full vehicle model.