2. Calibrate MPDB Barrier Face Material and Cell Properties
Why the Barrier Face Needs Its Own Calibration
The MPDB deformable barrier face is not a rigid wall. It is a graded honeycomb block whose force–deflection response is specified by Euro NCAP as a corridor: at any given crush depth the measured barrier force must fall inside a lower and upper bound. If the barrier is too soft, the vehicle over-penetrates and the crash pulse is under-driven; if it is too stiff, the vehicle front structure is loaded unrealistically and occupant restraint timing shifts. Because the barrier is the single component that every one of the 30 cases shares, its calibration error propagates into every subsequent result. This case therefore isolates the barrier: a barrier-only impact simulation is run against a rigid impactor, and the face material model plus cell-level stiffness parameters are tuned until the simulated force–deflection curve sits inside the corridor with margin.
Barrier-Only Test Configuration
The calibration model is deliberately minimal so that the only unknowns are the barrier properties. Build it as follows:
- Impactor: a rigid, non-deformable block with a flat 1000 mm × 1000 mm face, modelled as a rigid body with a prescribed initial velocity. A rigid impactor removes vehicle compliance from the measurement, so the recorded contact force is the barrier reaction alone.
- Barrier face: the full MPDB honeycomb block, fixed on its rear face to a rigid backing plate that represents the trolley front. All rear-face nodes are constrained in all translational degrees of freedom.
- Contact: a surface-to-surface contact between the impactor face and the barrier front face, with a friction coefficient of 0.3 and a penalty stiffness scaled to the honeycomb cell stiffness (start at 1.0 and adjust only if hourglassing or penetration appears).
- Output: contact force in the impact direction and impactor displacement, sampled at 0.1 ms so the corridor comparison is not aliased.
A practical impactor mass of 250 kg at 10 m/s gives roughly 12.5 kJ of kinetic energy, which drives the barrier to a crush depth of about 300–350 mm — enough to cover the full published corridor. Record the force–deflection curve as force versus impactor displacement, not versus time, because the corridor is defined in the displacement domain.
Material Model and Cell-Level Parameters
The honeycomb face is represented with a crushable foam material (for example MAT_CRUSHABLE_FOAM in LS-DYNA or an equivalent *FOAM material in other explicit solvers). The parameters that actually control the corridor are few, and they should be tuned in a fixed order because they interact:
| Parameter | Symbol / keyword | Effect on the force–deflection curve | Typical starting value |
|---|---|---|---|
| Plateau stress | LCID yield stress vs. volumetric strain | Sets the height of the crush plateau; the dominant knob for corridor fit | Derived from cell wall crush strength |
| Densification strain | End point of the stress–strain curve | Controls where force rises steeply at deep crush | 0.7–0.85 volumetric strain |
| Cell wall stiffness | Elastic modulus of the foam | Governs the initial rising slope before the plateau | Scaled from honeycomb relative density |
| Damping / rate factor | DAMP or strain-rate curve | Suppresses spurious oscillation on the plateau | 0.05–0.2 critical |
| Hourglass control | IHQ, QM | Prevents unphysical softening of under-integrated cells | Type 4, coefficient 0.05–0.1 |
Start from the nominal honeycomb crush strength and adjust the plateau stress first: run the barrier-only impact, overlay the simulated curve on the corridor, and scale the plateau stress by the ratio of corridor mid-force to simulated mid-force. Re-run, then adjust densification strain only if the deep-crush tail leaves the band. Leave damping and hourglass control at their defaults unless the plateau shows high-frequency oscillation or the cells visibly collapse into hourglass modes.
Tuning Workflow and Pass Criteria
Run the barrier-only impact, extract the force–deflection curve, and compare it against the corridor at fixed displacement stations (for example every 25 mm from 0 to 350 mm). The pass criterion is that the simulated force lies inside the band at every station with at least 10% margin to both bounds — a curve that touches the boundary is not accepted, because vehicle integration will perturb the response. If the plateau is too high, reduce the plateau stress by the ratio of corridor mid-force to simulated mid-force and re-run; if the initial slope is too steep, reduce the foam elastic modulus; if the deep-crush tail rises too early, increase the densification strain. Iterate in that order, one parameter at a time, and keep a run log of parameter set, mid-force, and margin so the final choice is traceable.
Typical Failure Symptoms
- High-frequency oscillation on the plateau: under-damped foam or insufficient hourglass control — raise damping slightly before touching the plateau stress.
- Force collapses after initial peak: cells are hourglassing; switch to a fully integrated or type-4 hourglass formulation and re-check mesh quality.
- Curve shifted uniformly up or down: plateau stress is wrong; scale it directly rather than adjusting multiple parameters.
- Penetration warnings at the contact: penalty stiffness too low relative to cell stiffness; increase it, but re-verify that the plateau has not stiffened artificially.
Once the barrier-only curve sits inside the corridor with margin, record the final material card and cell parameters as the frozen barrier definition. Every subsequent case in this series should reference that same definition so that differences between cases reflect the vehicle, not the barrier.