4. Material and property assignment for dummy components

A crash dummy is not a single homogeneous body. It is an assembly of foams, rubbers, plastics, and metals, each chosen because its mechanical response matches a specific region of the human body. In PAM-CRASH, the accuracy of an occupant injury prediction depends far more on how these materials are assigned than on mesh density alone. A head skin modeled with a linear elastic solid will produce a Head Injury Criterion (HIC) that is essentially meaningless, no matter how fine the elements are. This section explains how to assign viscoelastic foams, rubber joints, and strain-rate-sensitive metals to dummy components, and how to verify that the assigned behavior is physically plausible before running a full sled or vehicle model.

Why Dummy Materials Are Not Interchangeable

Dummy components fall into three broad mechanical families, and each demands a different constitutive treatment:

• Viscoelastic foams — head skin, chest flesh, abdomen inserts, and pelvis padding. These are highly compressible, dissipate energy, and show strong hysteresis and rate dependence.
• Rubber joints — neck bushings, lumbar spine connections, and some shoulder and hip articulations. These are nearly incompressible, hyperelastic at large strain, and viscoelastic at high loading rates.
• Metallic segments — ribs, femur, tibia, and the metallic spine elements. These are elastic-plastic and, critically, their yield stress rises with strain rate.

Assigning the wrong family is the most common modeling error. A foam modeled as incompressible rubber will lock up and generate unrealistically high forces; a rib modeled as rate-independent steel will under-predict fracture risk in high-speed impacts.

Viscoelastic Foams

For dummy foams, PAM-CRASH offers several material laws. The most practical for head skin and chest flesh is a crushable foam with viscoelastic relaxation, often implemented through a foam material card combined with a viscoelastic overlay, or through a dedicated low-density foam law. The key inputs are:

• Density — typically 50–200 kg/m³ for dummy flesh simulants.
• Loading/unloading curve — a nominal stress versus volumetric strain table, usually derived from uniaxial compression tests at quasi-static rate.
• Relaxation or damping parameters — a shear modulus ratio and decay constant, or a Prony series if the law supports it.
• Tension cutoff — foams carry almost no tension; set a small tensile limit to avoid negative pressure artifacts.

A representative example for head skin might use a density of 120 kg/m³, a plateau stress near 0.35 MPa at 30% volumetric strain, and a relaxation time constant of roughly 1 ms. These values are illustrative and must be replaced with data from the actual dummy certification or material test report.

Rubber Joints

Rubber bushings and neck joints are best represented with a hyperelastic model, commonly a Mooney–Rivlin or Ogden formulation, optionally combined with viscoelastic damping. Because rubber is nearly incompressible, the bulk modulus should be set high (typically 1000–2000 times the shear modulus) while the shear behavior is defined by the hyperelastic constants.

For a neck bushing, a practical workflow is:

1. Obtain uniaxial tension and compression data for the rubber compound.
2. Fit Mooney–Rivlin constants C10 and C01 using a curve-fitting tool.
3. Assign the fitted constants to the rubber material card.
4. Add a small viscoelastic damping term if the joint shows rate-dependent stiffness in dynamic tests.

If only a rotational stiffness is known, an alternative is to model the joint as a rigid connection with a rotational spring element, which is often sufficient for the neck and lumbar joints in a full dummy model.

Strain-Rate-Sensitive Metals

Ribs, femurs, and tibias are metallic and exhibit clear strain-rate sensitivity. In PAM-CRASH, this is handled with an elastic-plastic material law that includes a rate-dependent yield function, such as a Cowper–Symonds or Johnson–Cook formulation. The essential inputs are:

The Cowper–Symonds model scales the yield stress as σ_y = σ_0 [1 + (ε̇/C)^(1/p)], where σ_0 is the static yield stress and ε̇ is the plastic strain rate. For a rib loaded at 5 m/s, this can raise the effective yield stress by 20–40% compared with quasi-static loading, which directly affects fracture timing and chest deflection.

Assignment Workflow and Verification

The assignment process should follow a consistent sequence. First, partition the dummy mesh into component groups by material family. Second, assign the appropriate material card to each group. Third, run a single-element or small coupon simulation to verify the stress–strain response against the reference data. Fourth, run a component-level test, such as a head drop or chest impact, and compare the force–deflection curve with the dummy certification data.

Typical failure symptoms and their causes:

• Foam locks up too early — the volumetric strain table is too steep or the density is too high.
• Rubber joint is too stiff — the hyperelastic constants were fitted to the wrong test mode, or the bulk modulus is too low causing volumetric locking.
• Rib fractures too early — rate sensitivity is missing or the hardening curve is too low.
• Rib never fractures — the failure strain or plastic strain limit is set too high, or the element size is too coarse to capture localization.

The SVG below summarizes the mapping from dummy component to material family and the key parameters that must be verified for each.

Once the materials are assigned and verified at the component level, the dummy is ready for full assembly. The next step is to confirm that the assembled dummy reproduces the certification corridors before it is placed in a vehicle environment.