3. Fabric mechanical characterization

Airbag fabric is not a simple isotropic sheet. It is a woven, coated or uncoated textile whose mechanical response depends on yarn direction, weave architecture, coating, and loading path. Before any seam or deployment model can be trusted, the fabric constitutive behavior must be characterized with test methods that reproduce the relevant deformation modes. This section covers the four core characterization families — tensile, tear, biaxial, and shear — and explains how each test maps to a material parameter used in crash-safety simulation.

Why Directional Testing Matters

Airbag textiles are typically woven from high-tenacity nylon 6.6 or polyester yarns. Because warp and weft yarns interlace, the fabric exhibits strong anisotropy: stiffness and strength along the warp differ from the weft, and both differ from the 45° bias direction. Coating (silicone, neoprene, or polyurethane) further changes friction between yarns and alters shear and tear behavior. A single uniaxial test therefore cannot describe the material. The characterization program must sample multiple orientations and loading modes, then fit a constitutive model — commonly an orthotropic hyperelastic or a fabric-specific model such as the one used in LS-DYNA MAT_FABRIC — to the combined data set.

Tensile Testing

Tensile tests establish the in-plane stiffness and strength of the fabric along principal yarn directions. The standard reference is ASTM D5035 (strip method) or ISO 13934-1, adapted for high-modulus airbag textiles. Specimens are typically 50 mm wide and 200 mm long, cut along warp (0°), weft (90°), and bias (45°). Grip pressure must be controlled to avoid yarn slippage without crushing the weave; pneumatic grips at roughly 0.4–0.6 MPa are common.

A typical test sequence for a 470 dtex nylon 6.6 fabric might use a 100 kN load frame with a 10 mm/min crosshead rate. The output is a force–displacement curve converted to engineering stress–strain using the initial cross-section. Key extracted parameters include:

• Initial modulus (the slope of the near-linear region, often 0.5–2 % strain)
• Stress at 5 % and 10 % strain, used as stiffness anchors for simulation
• Ultimate tensile strength (UTS) and elongation at break
• Secant modulus at specified strain levels for model calibration

Because yarns crimp and straighten before load transfer, the curve is nonlinear and often shows a toe region. Reporting only UTS hides this behavior; the full curve should be digitized and used for fitting. A practical check: if the warp and weft curves differ by more than about 15 % in initial modulus, the weave is unbalanced and the model must retain separate directional properties.

Tear Testing

Tear resistance governs how a fabric behaves once a crack or slit initiates — critical for airbag deployment where stress concentrations form at seams and vents. The two common methods are the trapezoidal tear (ASTM D5587) and the tongue tear (ASTM D2261). For airbag textiles, the trapezoidal method is preferred because it produces a stable, reproducible tear path.

In a trapezoidal tear test, a specimen is cut with a small slit, mounted in a trapezoidal template, and pulled. The peak and average tear forces are recorded. A typical coated nylon might show an average tear force of 80–150 N, while an uncoated fabric of similar areal weight may tear at 40–90 N because yarns can slide and bunch. The ratio of tear strength to tensile strength is a useful indicator of coating adhesion and weave tightness. Low tear resistance relative to tensile strength signals a fabric prone to catastrophic rip propagation during deployment.

Biaxial Testing

Uniaxial tests constrain the fabric laterally and do not capture the coupling between warp and weft. During airbag inflation, the membrane is loaded in two directions simultaneously, and the Poisson-like interaction between yarn systems strongly affects the inflated shape and stress distribution. Biaxial testing addresses this.

The most common configuration is the cruciform specimen loaded on four independent axes. A central 100 mm × 100 mm gauge region is monitored with optical strain measurement (digital image correlation or video extensometry) because contact extensometers would reinforce the specimen. The test applies a controlled ratio of warp-to-weft stress or strain — for example 1:1, 2:1, and 1:2 — and records the resulting strain field.

Output data are typically presented as stress–strain curves for each loading ratio. From these, the fabric's biaxial modulus and the degree of yarn coupling are extracted. A practical calibration target: at a 1:1 load ratio and 5 % strain, the measured warp and weft stresses should be within 10 % of each other for a balanced fabric. If they diverge, the constitutive model needs separate directional hardening parameters. Biaxial data are also used to validate the simulation's inflated geometry — a model that matches uniaxial data but ignores biaxial coupling will over-predict or under-predict the cushion volume.

Shear Testing

Shear characterization captures the resistance of the weave to angular distortion — the mechanism by which a fabric drapes and conforms. The picture-frame (or bias-extension) test is standard. A square specimen is clamped in a hinged frame and pulled diagonally, forcing the yarns to rotate. The shear angle is measured optically, and the shear force is recorded.

For airbag fabrics, the shear response is highly nonlinear: initial resistance is low as yarns rotate, then rises sharply as the weave jams (locking angle). Coated fabrics typically show a higher initial shear stiffness and a lower locking angle than uncoated ones because the coating restricts yarn mobility. Typical locking angles for nylon airbag fabrics fall between 25° and 40°. The shear data feed into the simulation's fabric model to predict wrinkle formation and seam-line distortion during folding and deployment.

Test Matrix and Data Reduction Workflow

A complete characterization campaign for one fabric grade typically requires the following specimen set. The table below summarizes the minimum recommended matrix for a simulation-grade dataset.

*Biaxial testing for textiles does not have a single universally adopted ISO standard; many laboratories use internal procedures derived from ISO 16810 or ASTM D6856. The exact procedure should be documented and reported with the data.

After testing, the raw curves are filtered, the initial gauge length and cross-section are verified, and the data are fitted to the chosen constitutive model. A common workflow is:

1. Fit tensile data first to establish directional moduli and hardening.
2. Use biaxial data to calibrate coupling terms and validate the fit under combined loading.
3. Use tear data to set failure or damage-initiation criteria.
4. Use shear data to calibrate the fabric's angular resistance and locking behavior.

Interpreting the Combined Dataset

The four test families are not independent. A fabric with high tensile strength but low tear resistance will fail at seams before the base material reaches its limit. A fabric with high shear stiffness may resist wrinkling but concentrate stress at fold lines. The engineer's task is to read the combined dataset as a mechanical fingerprint. For example, if biaxial testing shows strong warp–weft coupling but shear testing shows a low locking angle, the fabric will likely form tight wrinkles during folding, and the simulation must include a damage model at fold intersections. If tear resistance is low and tensile elongation is high, the fabric will absorb energy by stretching but may propagate a slit rapidly once initiated — a scenario that demands careful seam design and reinforcement at attachment points.

In practice, the characterization data feed directly into the material card of the simulation model. The quality of the crash-safety prediction — seam integrity, cushion shape, and occupant interaction — depends on how faithfully these four test types capture the real fabric behavior. Skipping biaxial or shear testing because tensile data are available is a common and costly shortcut.

The workflow above shows how the four test families converge on a single calibrated material model. Each test type contributes a distinct set of parameters, and the validation feedback loop ensures that the fitted model reproduces the biaxial and shear behavior, not just the uniaxial response. In the next stage of the course, these calibrated properties will be applied to seam-level models where the interaction between fabric deformation and stitch line failure determines the overall crash-safety performance.