3. LS-DYNA composite material models overview

Selecting a constitutive model is the single most consequential decision in a composite crash simulation. The material card determines which stiffness terms are tracked, which failure criteria are evaluated, how damage evolves, and — critically — how much calibration data you must supply. LS-DYNA offers several families of composite models, each built around a different physical assumption. This section maps the four most common families used in vehicle structural work, then gives practical selection criteria for real components.

The Four Model Families and Their Physical Assumptions

MAT_054/055 (MAT_ENHANCED_COMPOSITE_DAMAGE) is a thin-shell, plane-stress model for unidirectional and woven lamina. It uses a Chang–Chang style failure surface with separate tensile and compressive strengths in the fiber and matrix directions, plus an optional nonlinear shear stress–strain curve. MAT_054 requires a defined failure strain and deletes elements once the criterion is met; MAT_055 is the same formulation but supports a damage-based softening option and a solid element form. Both are inexpensive and robust, which is why they remain common for large body panels where the dominant concern is first-ply failure and ply drop-off, not progressive crush.

MAT_058 (MAT_LAMINATED_COMPOSITE_FABRIC) is the workhorse for crash-grade composite structures. It is a continuum damage mechanics (CDM) model: each of the four in-plane modes — fiber tension, fiber compression, matrix tension, matrix compression — has its own damage variable that evolves from 0 to 1. Stiffness is degraded progressively rather than by instant element deletion, so the model captures the softening tail that governs energy absorption. MAT_058 needs a full set of strengths (XT, XC, YT, YC, S) and, for accurate results, fracture energies (GFT, GFC, GMT, GMC) that set the softening slope. It is available for shells and solids and is the default choice for CFRP/GFRP crash tubes, rocker sections, and floor structures.

MAT_161/162 (MAT_COMPOSITE_MSC / MAT_COMPOSITE_DIRECT) is a ply-by-ply model that tracks individual lamina through the thickness. MAT_161 uses a stress-based failure surface with a defined damage evolution law; MAT_162 is the "direct" variant where you supply the full 3D orthotropic stiffness and strength data directly. Both support solid elements and are the right tool when through-thickness stress matters — thick laminates, bolted joints, T-joints, and delamination-prone regions. The cost is higher: you must mesh each ply or use a layered solid formulation, and you must supply interlaminar properties if delamination is modeled.

MAT_261 (MAT_LAMINATED_FRACTURE_DAIMLER_PINHOLE) is a more recent CDM model developed for automotive crash. It uses a physically motivated fracture plane approach (Puck-style) for matrix failure and a separate fiber failure criterion, with a single set of fracture energies controlling the softening. It is available for shells and solids, handles both unidirectional and woven materials, and is generally more accurate than MAT_058 for mixed-mode matrix-dominated failure — at the cost of more calibration effort and a steeper learning curve.

Selection Criteria for Vehicle Structural Components

The choice is driven by three questions: what is the dominant failure mode, how thick is the laminate, and what calibration data do you actually have?

A practical rule: start with MAT_058 for any component whose job is to absorb energy. Move to MAT_161/162 only when through-thickness effects are demonstrably important — for example, when a bolted joint fails by delamination rather than net-section. Move to MAT_261 when MAT_058 under-predicts matrix-dominated failure under combined loading, which typically shows up as a too-early or too-late trigger in off-axis coupon validation.

Calibration Data Requirements

Every model above needs at minimum: elastic constants (E1, E2, G12, ν12), tensile and compressive strengths in fiber and matrix directions, and in-plane shear strength. MAT_058 and MAT_261 additionally need fracture energies. A common mistake is to supply strengths but leave fracture energies at defaults — this produces a brittle, deletion-driven response that absorbs far too little energy. For a typical carbon/epoxy crash tube, reasonable starting values are GFT ≈ 90 kJ/m², GFC ≈ 40 kJ/m², GMT ≈ 0.6 kJ/m², GMC ≈ 2.0 kJ/m²; these must be tuned against a coupon or sub-element test.

Worked Example: Selecting a Model for a CFRP Rocker Section

Consider a side-impact rocker made from a 4 mm quasi-isotropic CFRP laminate, loaded in three-point bending. The design intent is progressive crush with controlled energy absorption. Inputs: ply properties from the datasheet (E1 = 135 GPa, E2 = 9 GPa, G12 = 5 GPa, ν12 = 0.3, XT = 2200 MPa, XC = 1200 MPa, YT = 60 MPa, YC = 200 MPa, S = 90 MPa). Because the laminate is thin (4 mm, 8 plies) and the dominant mode is fiber compression followed by matrix crushing, MAT_058 with shell elements is the correct starting point. Mesh the rocker with 5 mm shell elements, assign MAT_058 with the properties above, and set GFT/GFC/GMT/GMC from the datasheet or a starting estimate. Run a three-point bend simulation and compare the force–displacement curve against a physical test. If the peak force is within 10% but the post-peak energy is low, increase GFC and GMC; if the trigger is too early, reduce the matrix tensile strength or increase GMT.

Practical Selection Checklist

• Identify the dominant failure mode first. Fiber-dominated crush → MAT_058 or MAT_261. Matrix-dominated or mixed-mode → MAT_261. Through-thickness or joint failure → MAT_161/162. First-ply only → MAT_054/055.
• Match element type to model. MAT_054/055 and MAT_058 are shell-friendly; MAT_161/162 and MAT_261 are solid-friendly. Using a solid-only model on shells forces approximations.
• Check your calibration data. If you only have datasheet strengths, MAT_054/055 or MAT_058 with estimated fracture energies is realistic. If you have coupon-level fracture data, MAT_261 or MAT_161/162 is justified.
• Validate on a sub-element before full vehicle. A three-point bend or axial crush coupon is the minimum. Compare peak force and post-peak energy, not just peak force.
• Document your choice and its assumptions. The model card, element type, mesh size, and fracture energies should be recorded so results are reproducible.

The next section moves from model selection to the mesh and element formulation decisions that determine whether the chosen model can actually deliver accurate results.