Rubber Hyperelasticity Simulation Testing: Material Model Determination for Finite Element Analysis

Rubber hyperelasticity simulation testing provides the experimentally determined material constants required to build accurate finite element (FE) models of elastomeric components such as seals, bushings, mounts, and tires. By combining uniaxial tension, planar shear, equibiaxial, and volumetric compression test data with rigorous least-squares curve fitting, accredited third-party laboratories deliver validated hyperelastic material models compatible with mainstream FE solvers. This article explains the specimens, test items, methods, and engineering applications involved in hyperelastic material characterization for simulation.

Technical Overview

Rubber and other elastomeric materials exhibit highly nonlinear, large-deformation, nearly incompressible elastic behavior that cannot be described by classical linear elasticity. In finite element simulation, this behavior is represented by hyperelastic constitutive models, in which the stress is derived from a strain energy density function W. The accuracy of any rubber simulation therefore depends directly on how well the coefficients of the chosen model reproduce real material behavior across the strain range of interest.

Rubber hyperelasticity simulation determination is a systematic laboratory process in which a rubber compound is tested under several independent deformation states, and the resulting stress-strain curves are fitted to candidate strain energy potentials. Commonly calibrated models include:

• Neo-Hookean model – a simple two-parameter model suitable only for small to moderate strains
• Mooney-Rivlin model (two, three, five, or nine parameter versions) – widely used for general-purpose sealing and mounting applications
• Yeoh model – a reduced polynomial form that captures the S-shaped stress-strain response of filled rubbers at large strains
• Ogden model – a spectral model offering high flexibility and excellent accuracy for large-strain applications such as tires
• Arruda-Boyce and Gent models – physically motivated models with good behavior extrapolation beyond the fitted strain range
• Van der Waals model – an interaction-based model that balances accuracy with parameter economy

Because each deformation mode probes a different region of the strain energy function, a single uniaxial tension test is mathematically insufficient to uniquely identify a stable material model. Professional determination therefore combines multiple test modes, followed by a material stability check (Drucker stability) to ensure the fitted coefficients produce physically admissible behavior in all deformation states within the simulation strain range.

Testing Samples

Uniaxial Tension Specimens

Standard dumbbell specimens, typically type 1, 2, or 4 according to ASTM D412 or ISO 37, are die-cut from cured sheets of 1 to 3 mm thickness. The dumbbell geometry ensures a homogeneous deformation state within the gauge section, which is essential for extracting true stress-strain data.

Uniaxial Compression Specimens

Cylindrical buttons or discs, commonly 29 mm in diameter and 12.5 mm in height per ASTM D575 or ISO 7743, are tested in compression. To approach a homogeneous stress state, specimens are either bonded to rigid plates or tested with effective frictional lubrication, since surface friction otherwise induces a non-uniform barreling deformation.

Planar Tension (Pure Shear) Specimens

Wide, thin rectangular strips with a high width-to-height ratio (typically greater than 10:1) are gripped along their long edges. Because the width direction resists lateral contraction, the central region approximates a pure shear deformation state, providing the second independent loading mode required for model identification.

Equibiaxial Extension Specimens

Equibiaxial tension can be measured by bubble (bulge) inflation of a clamped circular membrane, by biaxial stretching of a cruciform specimen, or by compression of a thin disc between lubricated platens. Equibiaxial data are particularly important because they probe the strain energy function in a deformation state far from uniaxial, dramatically improving fitted-model reliability.

Volumetric Compression Specimens

Disc or cylindrical specimens are compressed inside a rigid confining cavity to measure the volumetric (hydrostatic) response. Because rubber is nearly incompressible, the bulk modulus is typically two to three orders of magnitude larger than the shear modulus, and dedicated confined compression tests are used to determine the compressibility parameters or Poisson's ratio used in simulation.

Testing Items

A complete hyperelasticity determination program typically delivers the following results:

• Engineering and true stress-strain curves for uniaxial tension, uniaxial compression, planar tension, and equibiaxial tension over the specified strain range
• Calibrated material constants for selected models such as Mooney-Rivlin (C10, C01), Yeoh (C10, C20, C30), Ogden (mu-i and alpha-i), Neo-Hookean, Arruda-Boyce, and Gent coefficients
• Initial shear modulus and bulk modulus, together with the equivalent Poisson's ratio
• Material stability verification results confirming the fitted model satisfies Drucker stability criteria across the target strain range
• Goodness-of-fit metrics (relative error and residual plots) for each candidate model and each deformation mode
• Optional Mullins effect parameters for stress softening in filled rubber compounds
• Optional viscoelastic Prony series coefficients and hyper-viscoelastic coupling parameters for rate-dependent simulations
• A formatted material card ready for import into general-purpose finite element analysis software

Testing Methods

Specimen Preparation and Conditioning

Test plates are molded under controlled cure conditions to ensure representative crosslink density. Specimens are conditioned at standard laboratory atmosphere (23 ± 2 °C, 50 ± 5% relative humidity) for at least 16 hours, and compound hardness, density, and cure state are recorded as supporting metadata.

Uniaxial Tension Testing

Tension tests are performed per ASTM D412 or ISO 37 on a calibrated universal testing machine with a non-contact video extensometer or ink-mark tracking to measure true strain accurately at large extensions. Cyclic pre-loading (mechanical conditioning) is typically applied to remove the Mullins effect and obtain a stable, repeatable response before data acquisition. Tests may be run to a specified maximum strain (for example 50%, 100%, 200%, or 300%) matching the intended simulation use, and at single or multiple temperatures as required.

Compression and Volumetric Testing

Uniaxial compression is conducted per ASTM D575 or ISO 7743 using lubricated or bonded platens, while volumetric compression uses a confined compression fixture to impose a purely hydrostatic deformation. Together, these tests define the compressive branch of the response and the near-incompressibility parameters used in the model.

Planar and Equibiaxial Testing

Planar tension is executed with wide specimens gripped on the long edges, and equibiaxial tension is generated by bubble inflation or biaxial extension rigs. These two modes, combined with uniaxial data, span the principal deformation space of the strain energy function and are the key to eliminating the parameter non-uniqueness that arises from single-mode fitting.

Curve Fitting and Model Calibration

All test curves are converted into a common invariant or principal stretch formulation, and the model coefficients are determined by weighted least-squares regression using either the solver's built-in evaluation capability or a dedicated calibration routine. Multiple candidate models are fitted, and the model offering the best balance of accuracy, stability, and parameter robustness is recommended. A Drucker stability check is then performed to verify that the tangent stiffness remains positive definite over the entire simulation strain range.

Verification and Validation

The calibrated model is verified against the source test data and, where possible, validated against a benchmark component test such as a compression-loaded button, a squeezed O-ring cross-section, or a simple mounted bushing. Agreement within an agreed tolerance (commonly ±10% on load-deflection) confirms that the material card is ready for production simulation work.

Applications

Hyperelastic material determination services support simulation-driven design across many industries, including:

• Automotive: engine mounts, suspension bushings, body seals, weather strips, anti-vibration systems, and air springs
• Oil and gas and fluid power: O-rings, packer seals, gaskets, and downhole sealing elements
• Tire engineering: tread, sidewall, and inner liner material modeling for ride, handling, and durability studies
• Rail and heavy machinery: primary and secondary suspension rubber components and coupler buffers
• Consumer and footwear: midsole and outsole cushioning materials
• Medical devices: balloon catheters, tubing, diaphragms, seals for syringes and infusion systems
• Electronics and appliances: vibration isolators, grommets, and keypads
• Research and development: input data for material model research, durability prediction, and design optimization studies

FAQ

Which hyperelastic model should I choose for my rubber material?

For moderate strains up to roughly 100%, a two-parameter Mooney-Rivlin or Neo-Hookean model may suffice. For carbon-black-filled compounds exhibiting the characteristic S-shaped curve at large strains, the Yeoh model is often the most robust choice. For tire applications and strain ranges approaching 300% or more, the Ogden model with two or three terms typically provides the best accuracy. A determination program that fits several candidate models and compares stability and error metrics removes the guesswork from this selection.

How many deformation modes must be tested?

At minimum, uniaxial tension plus one additional mode (planar shear or equibiaxial) is strongly recommended; fitting a multi-parameter model to uniaxial data alone frequently yields unstable or non-unique coefficients. For high-fidelity work such as seals or tires, the recommended combination is uniaxial tension, planar tension, equibiaxial tension, and volumetric compression.

What strain range should the tests cover?

The tests should cover the full strain range expected in service, plus a safety margin. If the finished part will see 60% compression strain, the material tests should extend at least to that level in the corresponding deformation mode, because extrapolating a hyperelastic fit beyond the fitted range is a common source of simulation error.

What is material (Drucker) stability and why does it matter?

Stability means the model's incremental work is positive for all admissible deformations, so the tangent stiffness remains positive definite. Unstable coefficients may cause convergence failures or non-physical results in finite element analysis even when the fit to test data looks excellent. A professional determination always includes a stability check over the intended strain range.

Should Mullins effect and viscoelasticity be included?

If the component experiences significant cyclic loading, first-time loading after manufacture, or rate-dependent behavior, adding Mullins effect parameters and a Prony series from dynamic or relaxation testing significantly improves realism. For static, monotonic loading cases, a pure hyperelastic model is generally adequate.

How does temperature affect the determination?

Rubber properties are strongly temperature dependent. When the part operates across a temperature range, hyperelastic tests should be repeated at the relevant temperatures so that separate material cards or temperature-interpolated data sets can be generated for the simulation.

What deliverables will I receive?

Typical deliverables include the raw stress-strain data, fitted coefficients for each candidate model with error metrics, stability check results, and a solver-ready material definition file, all documented in a test report issued by an accredited third-party laboratory.