1 Scope and motivation

Nonlinear viscoelasticity addresses the mechanical response of materials whose deformation depends on both time-dependent dissipation and nonlinearity in the stress–strain relationship. When the material’s stiffness and internal damping vary with deformation magnitude, loading rate, or loading history, linear viscoelastic theory no longer provides an accurate description.

1.1 Why linear viscoelasticity is insufficient

Linear viscoelastic models assume a proportional relation between stress and strain via a superposition principle, typically leading to responses that can be computed from a linear hereditary kernel. In practice, many soft and complex solids exhibit stiffness that changes with amplitude, relaxation that depends on the strain level, and creep that accelerates or decelerates depending on how the load is applied. These effects break the assumptions behind linear superposition and require nonlinear constitutive descriptions.

1.2 Common materials showing nonlinear behavior

Nonlinear viscoelasticity is widely used to model polymeric solids, polymer gels, elastomers with temperature- and strain-dependent behavior, biological tissues, and other soft matter systems such as hydrogels. Complex formulations—filled polymers, double-network gels, and composite soft materials—also commonly show nonlinear relaxation and hysteresis under realistic loading.

1.3 Typical experimental signatures

Key signatures include strain-dependent relaxation spectra, creep curves that show nonlinear onset, and stress–strain loops whose shape depends on amplitude and waveform. Nonlinear damping often produces cycle-to-cycle changes in dissipated energy. Experimental results may also show pronounced path dependence: response at a given strain or stress can differ depending on prior loading and the time history of the applied deformation.

2 Foundations of nonlinear viscoelastic modeling

Nonlinear viscoelastic modeling aims to represent time-dependent dissipation while allowing constitutive relations to change with deformation state and loading history. A successful framework typically introduces state variables and evolution laws while ensuring physical constraints such as positive dissipation.

2.1 Constitutive descriptions and state variables

Constitutive modeling specifies how stresses are generated from kinematics and internal state. In nonlinear viscoelasticity, internal degrees of freedom track aspects of microstructural rearrangement or structural degradation/strengthening.

2.1.1 Internal variables and structural state

Internal variables represent hidden modes such as relaxing stresses, viscoelastic stretches, damage-like structural measures, or conformational populations. Their values evolve with time and with the mechanical state, enabling the macroscopic stress to depend on both current deformation and past loading.

2.1.2 Strain-history dependence and memory

Memory effects are incorporated through hereditary integrals or evolving internal variables. In hereditary approaches, stress at time \(t\) depends on the entire strain history via a convolution-like structure that may itself become amplitude-dependent. In internal-variable approaches, the “memory” is embedded in the state variables whose trajectories reflect prior deformation.

2.2 Nonlinearity sources

Nonlinearity arises when mechanical response changes with loading amplitude, rate, or the structural state produced by stress and strain.

2.2.1 Amplitude-dependent moduli

Amplitude-dependent stiffness means that the effective modulus at the same time after loading can differ when the imposed strain or stress has a different magnitude. This often produces relaxation that is faster or slower depending on the level of deformation.

2.2.2 Rate dependence beyond linear theory

While linear viscoelasticity already predicts rate sensitivity through time-dependent kernels, nonlinear theory extends this by allowing the material to exhibit different functional forms for response at different rates and loading protocols. Rate effects may also interact with amplitude, yielding distinct behavior when both are varied.

2.2.3 Stress- or strain-induced material changes

Some materials undergo reversible restructuring (e.g., alignment, unfolding/refolding, network rearrangement) or irreversible changes (e.g., softening). Constitutive models may include evolution equations for structural state variables that alter stiffness and dissipation as loading proceeds.

2.3 Thermodynamic consistency

Thermodynamic consistency constrains admissible constitutive equations so the model does not predict unphysical amplification of energy.

2.3.1 Energy dissipation requirements

A key requirement is nonnegative dissipation: the model should not generate energy from the internal processes. This is typically enforced by ensuring that the rate of the free energy decreases by an amount equal to (or exceeding) the dissipative power associated with internal variable evolution.

2.3.2 Frame invariance and objectivity

For materials that may undergo finite deformation, stresses must transform appropriately under changes of observer. Objectivity, often expressed through frame invariance, restricts the choice of strain measures and the form of constitutive relations.

3 Representative constitutive frameworks

Many nonlinear viscoelastic models can be grouped by how they treat history dependence and nonlinearity: via nonlinear hereditary integrals, by nonlinear limits to superposition, through strain- or stress-dependent kernels, or by internal-variable networks and hyperelastic-viscoelastic decompositions.

3.1 Nonlinear hereditary integrals

Hereditary formulations build stress response from integrals over past strain. Nonlinearity is introduced either by expanding around a reference state or by allowing fully nonlinear dependence inside the integrand.

3.1.1 Weakly nonlinear expansions

Weakly nonlinear models treat deviations from a baseline linear response as small perturbations. The stress can be expanded in powers of strain (and sometimes strain rate), leading to higher-order kernels. These models are useful when nonlinear effects are present but not dominant.

3.1.2 Fully nonlinear convolution-type models

Fully nonlinear convolution-type frameworks replace the linear kernel with a function that depends on the loading magnitude and/or state. The resulting integral may not be expressible as a simple superposition of independent contributions, but it can still preserve a structured history dependence.

3.2 Boltzmann-type nonlinear superposition limits

Boltzmann superposition is central to linear viscoelasticity. In nonlinear settings, superposition may hold only approximately, in restricted regimes, or when nonlinearity can be “absorbed” into an effective kernel.

3.2.1 Breakdown of linear superposition

When relaxation depends on strain amplitude, the response to a sum of inputs cannot be obtained by adding responses to each input separately. This failure motivates models that incorporate explicit dependence on amplitude and path, rather than relying on linear superposition.

3.3 Stress- and strain-dependent relaxation functions

Relaxation functions may depend directly on the current strain level, past strain levels, stress level, or the evolving structural state. This approach often yields tractable forms while capturing key nonlinear features.

3.3.1 Strain-dependent time scales

Nonlinearity can be expressed by allowing characteristic relaxation times to change with strain amplitude. A model may predict slower relaxation at one deformation range and faster relaxation at another, producing observed shifts in relaxation curves.

3.3.2 Kernel identification challenges

Because kernels become state-dependent, identification from experiments is harder than in linear viscoelasticity. Different loading protocols can yield different effective kernels, and parameter estimates may become non-unique unless the model structure matches the data sufficiently well.

3.4 Internal-variable (generalized Maxwell/anelastic) models

Generalized Maxwell-type networks use multiple relaxing elements. Nonlinearity is introduced by nonlinear springs and/or nonlinear evolution rules governing the relaxing stresses.

3.4.1 Nonlinear spring–dashpot networks

A common strategy is to replace linear springs and dashpots with nonlinear constitutive elements. The network retains a convenient structure for implementation while allowing stiffness and viscosity to vary with strain level or stress.

3.4.2 Nonlinear evolution laws for internal variables

Internal variables can evolve through differential equations where rates depend on the current deformation measure, internal state, and sometimes temperature. Properly designed evolution laws allow the model to reproduce strain-dependent creep initiation, nonlinear hysteresis, and evolving relaxation behavior.

3.5 Hyperelastic-viscoelastic approaches

For finite deformations, hyperelastic-viscoelastic models combine an elastic free-energy response with time-dependent viscous/viscoelastic mechanisms. Conceptual decompositions often separate the elastic contribution from the dissipative one.

3.5.1 Multiplicative decompositions (conceptual)

In many continuum mechanics settings, deformation may be conceptually split into elastic and inelastic parts. This structure helps define evolving stress-generating measures that remain compatible with objectivity under large strains.

3.5.2 Time-dependent stresses from evolving strain measures

The stress is derived from an evolving strain measure that changes with both deformation and viscous processes. As internal mechanisms evolve, the effective elastic strain and thus the stress response evolve, yielding nonlinear viscoelastic behavior under large deformation histories.

4 Material response and characteristic phenomena

Nonlinear viscoelastic models are motivated by distinctive observed behaviors under common mechanical protocols. The following phenomena serve as benchmarks for model adequacy.

4.1 Strain-dependent stress relaxation

Stress relaxation refers to the decay of stress under imposed strain. In nonlinear materials, the decay rate and residual level can depend strongly on strain amplitude, so relaxation curves measured at different strain levels cannot be collapsed using a single linear kernel.

4.2 Creep with nonlinear onset

Creep is the time-dependent deformation under sustained stress. Nonlinear creep may show accelerated deformation after a threshold time or a change in slope early in the response. Such onset behavior is sensitive to both the stress level and prior loading.

4.3 Hysteresis and loop behavior in cyclic loading

Cyclic loading produces stress–strain loops whose area reflects energy dissipated per cycle. In nonlinear viscoelasticity, loop shape depends on amplitude, frequency, and waveform (e.g., sinusoidal versus triangular), and may change as the material undergoes evolving structural state.

4.4 Path dependence under complex loading

Under multi-step or non-monotonic loading, the response at a given time can depend on the order and timing of imposed stresses/strains. Nonlinear viscoelastic models reproduce this by retaining memory through internal variables or nonlinear hereditary integrals.

4.5 Nonlinear damping and energy loss

Nonlinear damping means the effective damping coefficient is not constant: it changes with deformation amplitude and loading rate. This leads to frequency shifts and amplitude-dependent attenuation in vibration problems.

4.6 Experimental protocol dependence (loading rate, waveform)

Because nonlinear behavior couples amplitude, time, and history, experimental outcomes can vary widely with protocol. A model must therefore be calibrated using tests that cover the relevant loading modes and not merely a single monotonic experiment.

5 Governing equations in continuum mechanics

To use nonlinear viscoelastic constitutive laws in engineering analyses, they must be embedded within continuum mechanics balance laws and kinematic descriptions.

5.1 Kinematics and stress measures

The governing framework begins with kinematics: displacement or deformation measures define strains, stretches, and rates. The choice of stress measure must align with the kinematics and with the intended deformation regime (small versus finite strain).

5.2 Constitutive substitution into balance laws

Balance of linear momentum and, when relevant, angular momentum and mass conservation provide the mechanical field equations. Constitutive equations relate stress to strain measures and internal variables. Substituting these relations yields a closed system for unknown fields and internal states.

5.3 Boundary and initial conditions

Initial conditions determine internal variable states and initial stress/strain configuration. Boundary conditions specify tractions, displacements, or mixed constraints. For nonlinear viscoelasticity, specifying the loading protocol in time is essential because the solution depends on the history.

5.4 Numerical implications for nonlinear viscoelasticity

Nonlinearity introduces challenges for computation: constitutive updates may require iterative procedures, time stepping must resolve relaxation scales, and consistent tangents may be needed for robust convergence. Stability and accuracy depend on both discretization and the chosen integration scheme for internal variable evolution.

6 Parameter identification and model calibration

Model parameters must be inferred from experiments and validated against independent data. Nonlinear viscoelasticity complicates identification due to state-dependent kernels and potential non-uniqueness.

6.1 Data requirements and preprocessing

Calibration typically uses time-resolved stress and strain measurements under multiple protocols, including relaxation, creep, and cyclic tests. Preprocessing includes filtering noise, aligning time references, and ensuring consistent sampling so that derivatives (if required) are computed reliably.

6.1.1 Choosing strain vs stress control tests

Strain-controlled tests more directly reveal relaxation under prescribed deformation, while stress-controlled tests are essential for creep and for identifying nonlinear viscosity. Using only one control mode can leave key parameters weakly constrained.

6.2 Fitting relaxation and creep curves

Relaxation experiments constrain the time-dependent decay under fixed strain, informing relaxation functions or internal variable kinetics. Creep tests constrain deformation growth under fixed stress, providing complementary information on rate and nonlinear onset.

6.3 Parameter estimation strategies

Parameter estimation often uses nonlinear optimization with regularization to reduce sensitivity to noise and to discourage physically implausible parameter sets.

6.3.1 Optimization and regularization

Common approaches include least-squares fitting of stress or strain histories, sometimes with weighting to balance different time regions. Regularization terms can enforce smoothness or constrain parameters to physically meaningful ranges.

6.3.2 Identifiability and sensitivity

Identifiability analysis evaluates whether multiple parameter combinations produce nearly indistinguishable predictions. Sensitivity studies determine which experiments most strongly influence each parameter and help prevent overfitting.

6.4 Validating predictive capability

Validation requires prediction of responses for loading protocols not used in calibration, such as different amplitudes, combined loading sequences, or different frequencies. Agreement in both transient and cyclic behavior supports a model’s generality.

7 Numerical methods and computation

Numerical solution of nonlinear viscoelastic problems requires careful treatment of time integration, constitutive updates, and stability.

7.1 Time integration strategies

Time stepping must resolve rapid relaxation processes while covering long histories. Implicit integration is often used for stiff models, while explicit schemes can be efficient but may require small time steps to maintain stability.

7.2 Handling nonlinear constitutive updates

Constitutive evolution may require solving nonlinear algebraic equations at each time step or updating internal variables via differential equations with appropriate discretization. In many implementations, consistent linearization improves convergence in Newton-type solvers.

7.3 Stability and convergence considerations

Stability depends on model parameters, time step size, and discretization consistency. Convergence tests typically verify mesh/time refinement behavior and confirm that the solution approaches a limiting prediction.

7.4 Multiscale and coarse-grained implementations

Some frameworks attempt to reduce computational cost by deriving macroscopic models from microstructural simulations. Coarse-grained constitutive laws can capture effective viscoelasticity with fewer internal variables, though accuracy depends on the representativeness of the reduced description.

8 Applications in science and engineering

Nonlinear viscoelasticity is used wherever materials show history-, amplitude-, and rate-dependent mechanical behavior.

8.1 Polymer and soft-matter mechanics

In polymer processing and characterization, nonlinear viscoelastic models support prediction of stress development, relaxation during forming, and time-dependent deformation under service loads. Soft-matter systems such as gels and suspensions can show pronounced hysteresis and evolving internal structure.

8.2 Biomedical tissue modeling (mechanical response)

Biological tissues often exhibit nonlinear, time-dependent stress–strain behavior due to complex microstructure and active remodeling. Modeling tools help interpret mechanical tests and may support the design of biomedical devices, such as flexible implants and tissue-mimicking scaffolds.

8.3 Impact, damping, and vibration control

Nonlinear viscoelastic materials are relevant for vibration isolation and energy absorption because their damping changes with amplitude and frequency. Impact problems require capturing both rapid transient response and slower relaxation that can persist after unloading.

8.4 Additive manufacturing and complex polymer networks

Additive manufacturing produces polymer structures with varying crosslink density and local heterogeneity. Nonlinear viscoelastic models can be used to estimate how printed parts relax stress, creep under sustained loads, and respond under cyclic service conditions.

9 Outlook and research directions

Ongoing research focuses on improving model realism, reducing computational expense, and integrating richer data sources for parameter identification.

9.1 Data-driven and hybrid constitutive models

Machine-learning-assisted constitutive laws and hybrid approaches combine physics-based constraints with flexible function approximators. These methods aim to better capture nonlinearities while preserving thermodynamic admissibility and extrapolation reliability.

Linking internal variables to measurable microstructural quantities remains a central challenge. Progress includes multiscale modeling, improved structural state variables, and experimental methods that correlate deformation with microstructural evolution.

9.3 Efficient simulation for large-scale problems

Scalable simulation methods seek to reduce the cost of updating complex internal-variable systems in large finite-element models. Strategies include model reduction, adaptive time stepping, and efficient surrogate models for constitutive predictions.