1 Constitutive relations in continuum theory
1.1 Purpose and role in closing governing equations
In continuum modeling, balance laws express conservation of quantities such as mass, momentum, and energy, but they do not specify how the material’s internal response develops under applied influences. Constitutive relations supply this missing information by prescribing a functional link between measurable driving variables (for example, displacement gradients or temperature) and internal variables that characterize response (for example, stress or heat flux). Without such relations, the system of governing equations is typically underdetermined.
1.2 Relationship to balance laws
Balance laws provide the evolution and constraints of fields within space and time, while constitutive relations determine the constitutive fields that enter these laws. For instance, in solid mechanics the momentum balance involves stress, and the constitutive relation provides stress as a function of deformation measures. In thermodynamics, energy balance includes heat flux and entropy production; constitutive prescriptions specify heat conduction and possible coupling between heat transfer and other processes.
1.3 Local vs. nonlocal and instantaneous vs. history-dependent forms
Constitutive models are commonly classified by how they depend on the state of the material. “Local” models relate response at a point to state variables at that same point. “Nonlocal” models include dependence on neighboring regions or introduce gradient terms, reflecting effects such as size dependence or long-range interactions. Similarly, “instantaneous” relations depend only on present state (and sometimes present rates), whereas “history-dependent” relations incorporate past deformation or past thermodynamic states through time-dependent kernels, internal variables, or hereditary integrals. These choices strongly affect both predicted behavior and computational requirements.
2 Variables and state descriptions
2.1 Kinematic quantities (strain, deformation, strain rate)
The kinematics of motion introduce measures of deformation that quantify how material elements change shape and size. Common choices include displacement gradients, strain tensors, and strain rates. For large deformations, nonlinear strain measures and objective kinematic descriptions are used so that predicted stresses do not depend on rigid-body motion.
2.2 Field quantities (stress, electric displacement, magnetic induction)
Field variables represent response within the continuum. In mechanics, stress characterizes internal forces per area. In electromagnetism within materials, electric displacement and magnetic induction describe how polarization and magnetization alter the fields. These response quantities are linked to driving variables such as displacement fields, electric fields, and magnetic flux density through material-specific constitutive laws.
2.3 Thermodynamic variables (temperature, entropy, internal energy)
Thermodynamic state is described by quantities like temperature, entropy, and internal energy. Constitutive relations often ensure consistency with the second law by connecting heat flux, entropy production, and internal state evolution. Temperature dependence of elastic moduli or viscosity is likewise introduced through thermodynamic parameters or empirical correlations.
2.4 Material state and constitutive assumptions
A “material state” is an abstract description sufficient to determine the response within the chosen model class. Depending on modeling assumptions, state may include only current deformation and temperature, or it may also include internal variables representing microstructural effects (for example, damage variables, viscoelastic strains, or plastic hardening measures). These assumptions determine whether the model is memoryless, rate-dependent, or path-dependent.
3 Linear constitutive models
3.1 Hooke’s law and elastic moduli
For small strains, elastic response is often modeled linearly: stress is proportional to strain through elastic moduli. In isotropic solids this yields two independent parameters (e.g., Young’s modulus and Poisson’s ratio), while anisotropic materials require a tensor of elastic constants. Linear elasticity is foundational for many engineering analyses and provides a baseline for more complex nonlinear theories.
3.2 Newtonian viscosity and viscous stress
In simple fluid mechanics, Newtonian viscosity relates viscous stress to the strain rate tensor. The proportionality constant is the dynamic viscosity, possibly varying with temperature. This assumption produces linear stress–rate behavior and leads to classical results such as linear momentum diffusion in certain flow regimes.
3.3 Linear thermoelasticity
Thermoelastic models extend linear elasticity by incorporating temperature variations. Stress depends not only on mechanical strain but also on temperature changes, often through thermal expansion coefficients. Heat conduction may be coupled to the mechanical fields depending on whether the model includes thermoelastic coupling in the energy and entropy balances.
3.4 Onsager reciprocity and linear transport
Near equilibrium, transport processes such as heat conduction and diffusion can be modeled with linear relations between fluxes and thermodynamic forces. Onsager reciprocity constrains the coefficients of these linear laws under standard symmetry and time-reversal assumptions, implying that certain cross-effects (for example, coupling between different fluxes) are pairwise consistent.
4 Nonlinear and rate-dependent constitutive relations
4.1 Nonlinear elasticity and hyperelasticity
Nonlinear elastic theories are needed when strains are not small or when material response exhibits strong stiffening, softening, or non-symmetric stress–strain behavior under large deformations. Hyperelasticity models the stress as derived from a strain-energy density function, ensuring a conservative elastic structure and supporting robust treatment of large strain kinematics. Nonlinear elasticity also accommodates pressure-dependent behavior in some materials.
4.2 Viscoelasticity and time-dependent stress-strain behavior
Viscoelastic models represent materials that exhibit both elastic and viscous effects, producing stress that depends on deformation history. Two common viewpoints are “spring-dashpot” element models and formulations using integral constitutive laws or internal variables.
4.2.1 Maxwell models
The Maxwell model combines a spring and a dashpot in series, capturing stress relaxation: under a sudden strain step, stress decays over time. This structure is useful as an idealization for polymers and other media where long-time response becomes more fluid-like.
4.2.2 Kelvin–Voigt models
The Kelvin–Voigt model places a spring and dashpot in parallel, capturing creep: under a sudden stress step, strain increases over time. It represents materials that resist immediate deformation but gradually yield while still retaining some elastic character at long times.
4.2.3 Boltzmann superposition and hereditary integrals
Many linear viscoelastic behaviors can be represented by Boltzmann superposition, where current response is obtained from an integral over past strain or stress weighted by relaxation or retardation functions. Hereditary integral formulations express memory through convolution-type kernels and are widely used in analytical and numerical studies when linearity holds.
4.3 Plasticity and yield criteria (general modeling framework)
Plasticity models permanent deformation once stress reaches a yield threshold. A yield function defines the onset of plastic flow, while flow rules determine how plastic strain evolves after yielding. Hardening laws describe how the yield surface evolves with accumulated plastic strain, capturing material strengthening or softening. Rate-independent plasticity is common in engineering solids, though rate-dependent extensions exist for viscous-plastic materials.
4.3.1 Stress-strain curves and hardening concepts
Stress–strain curves summarize macroscopic response under controlled loading. Hardening concepts include isotropic hardening (expanding yield surface), kinematic hardening (translating yield surface), and more complex combined mechanisms. These ideas translate into constitutive evolution equations and influence predicted phenomena like elastic recovery, cyclic stabilization, and the Bauschinger effect in appropriate frameworks.
4.4 Rheological models for complex fluids
Complex fluids such as polymers, suspensions, and gels often show nonlinear viscosity, shear-thinning or shear-thickening, and time-dependent relaxation. Rheological constitutive models—beyond simple Newtonian behavior—describe how stress depends on deformation history and deformation rate. Common approaches include generalized Newtonian models for shear-dependent viscosity and more elaborate viscoelastic or microstructure-based theories.
5 Anisotropy, heterogeneity, and internal structure
5.1 Anisotropic constitutive behavior
Anisotropic materials have directional dependence of response, arising from crystal structure, fiber reinforcement, or manufacturing-induced alignment. Constitutive relations must therefore use tensorial forms rather than scalar moduli. Anisotropy can affect wave propagation, yield behavior, and stiffness in ways that cannot be captured by isotropic assumptions.
5.2 Heterogeneous materials and effective properties
Heterogeneous media vary in space at microscopic or mesoscopic scales. Continuum models often replace the detailed microstructure by effective properties that approximate overall behavior. Effective stiffness, conductivity, or permeability may depend on volume fractions and geometry of phases, and must be determined by homogenization methods, mixture rules, or calibrated experiments.
5.3 Microstructure-informed modeling (conceptual connection)
Microstructure-informed constitutive modeling aims to link macroscopic behavior to mechanisms at smaller scales, such as polymer chain orientation, fiber network deformation, or particle interactions. While fully resolving microscopic physics is often impractical, internal variables and evolution laws are used to represent microstructural states, enabling more predictive modeling across regimes of loading and temperature.
5.4 Objectivity and frame invariance requirements
Constitutive relations must be consistent with the principle that rigid-body motions should not produce spurious stresses or altered material response. “Objectivity” or “frame invariance” imposes constraints on how strains, rates, and stress measures transform under changes of observer. This requirement influences the selection of strain measures and stress-rate formulations, especially in finite deformation settings.
6 Governing principles and constraints
6.1 Thermodynamic admissibility
A constitutive model is often required to be thermodynamically admissible, meaning it does not violate fundamental thermodynamic laws. In practice, this typically constrains allowable forms of constitutive functions and internal variable evolution so that the model respects relationships among energy, entropy, and irreversible processes.
6.2 Energy conservation and dissipation inequalities
The second law is frequently expressed via an entropy inequality or, equivalently, a dissipation inequality. Constitutive relations must ensure that the rate of dissipation is nonnegative under all admissible processes. This requirement eliminates physically inconsistent constitutive choices and guides the formulation of irreversible mechanisms such as viscosity, plastic flow, damage, and diffusion.
6.3 Symmetry considerations (material and response symmetry)
Material symmetry limits how constitutive tensors can be structured. For example, in linear elasticity, symmetry reduces the number of independent elastic constants. Response symmetries can also be used to restrict constitutive equations for transport and coupling terms, ensuring compatibility with invariances of the underlying microstructure or with measured behavior.
6.4 Invariance under coordinate transformations
Constitutive laws must remain consistent under changes of coordinate systems. This includes rotational invariance in spatial descriptions and proper transformation properties of tensor fields. Ensuring invariance avoids coordinate artifacts and helps guarantee that numerical implementations produce physically meaningful results regardless of mesh orientation or local basis choices.
7 Constitutive relations across physical domains
7.1 Stress–strain in mechanics of solids
Solid mechanics constitutive relations connect stresses to deformation measures and their rates, possibly including temperature and history effects. The modeling range spans linear elasticity, viscoelasticity, hyperelasticity, plasticity, and combinations thereof (for example, viscoelastic–plastic models). The selected form determines predicted phenomena such as stress relaxation, permanent set, and nonlinear load–displacement curves.
7.2 Constitutive laws in fluid mechanics and rheology
Fluid constitutive behavior typically relates stress to velocity gradients and possibly to concentration or temperature. Beyond Newtonian viscosity, non-Newtonian rheology introduces shear-dependent or time-dependent effects. In multiphase contexts, constitutive relations may also couple with interfacial dynamics and with constitutive transport of additives or solutes.
7.3 Electromagnetic constitutive relations in media
In materials, electromagnetic constitutive relations link electric field to polarization effects and magnetic field to magnetization, commonly expressed through permittivity and permeability, potentially as functions of frequency, temperature, or field magnitude. In more advanced settings, constitutive laws include anisotropy, magnetoelectric coupling, and dispersive behavior requiring time-domain or frequency-domain formulations.
7.4 Heat conduction and thermotransport laws
Heat conduction constitutive relations connect heat flux to temperature gradients, often using Fourier’s law in the simplest case. More general thermotransport theories introduce non-Fourier effects, temperature-dependent conductivity, and coupled transport phenomena such as thermoelectric or diffusion-driven heat flux. These laws determine how thermal fields evolve and how mechanical or electrical processes can contribute to heating.
8 Modeling, parameter identification, and calibration
8.1 Experimental measurements for model fitting
Model parameters are usually inferred from controlled experiments that produce measurable stress, deformation, temperature, or field response under known loading histories. For elastic models, standard tests include tension, compression, or shear. For viscoelasticity and plasticity, relaxation tests, creep tests, cyclic loading, and hardening characterization provide the information needed to fit time-dependent or path-dependent constitutive functions.
8.2 Inverse problems and parameter estimation
Parameter identification is often posed as an inverse problem: given experimental observations, determine constitutive parameters that best reproduce the data. This may involve nonlinear least squares, Bayesian inference, or regularized optimization to address non-uniqueness and noise. Model calibration frequently requires careful selection of objective functions and constraints to ensure physical plausibility.
8.3 Scaling, nondimensionalization, and model selection
Scaling arguments and nondimensionalization clarify which terms dominate under specific operating conditions. By comparing characteristic times, lengths, stresses, and rates, one can determine whether a linear model suffices or whether nonlinear, rate-dependent, or history-dependent effects are necessary. Model selection balances predictive accuracy with simplicity and computational cost.
8.4 Sensitivity analysis and uncertainty quantification
Uncertainty quantification assesses how measurement noise and parameter uncertainty propagate into predictions. Sensitivity analysis identifies which parameters most strongly influence output quantities such as stress peaks, relaxation times, or effective stiffness. These analyses guide experimental design and improve confidence in model applicability across regimes.
9 Computational implementation
9.1 Constitutive updates in numerical methods
In finite element and finite volume methods, the constitutive relation must be evaluated repeatedly at integration points. For rate-independent and rate-dependent models, the computation typically involves determining stress (or other response variables) from current strain measures and possibly solving for internal variable updates. Efficient algorithms are essential because constitutive evaluations dominate runtime for nonlinear problems.
9.2 Time stepping and stability considerations
When constitutive laws depend on time derivatives or history, time discretization choices affect stability and accuracy. Viscous regularization, implicit versus explicit schemes, and consistent linearization are common considerations. For viscoelastic and viscoplastic behavior, poorly chosen step sizes can produce numerical oscillations or inaccurate relaxation dynamics.
9.3 Finite element integration of constitutive laws
Finite element implementation requires mapping kinematic measures to integration points, evaluating constitutive updates, and providing consistent tangent operators for nonlinear solvers. Viscoelastic and plastic constitutive algorithms may require local iterative procedures, such as return-mapping schemes for plasticity, or convolution updates for hereditary integrals. Ensuring algorithmic consistency helps maintain convergence of global Newton iterations.
9.4 Validation and verification of simulations
Verification confirms that the numerical implementation solves the governing equations correctly given the constitutive model, often through method-of-manufactured-solutions or mesh/time refinement studies. Validation checks that model predictions match experimental data within acceptable uncertainty. Together, these steps establish credibility for using constitutive models in predictive simulations.
10 Extensions and modern directions
10.1 Nonlocal and gradient-enhanced constitutive theories
Nonlocal and gradient-enhanced approaches include additional length scales through dependence on gradients of strain, phase fields, or other state variables. These theories can reduce mesh dependence and capture size effects observed in microstructured materials. They often involve higher-order governing equations or additional boundary conditions.
10.2 Data-driven constitutive modeling (general overview)
Data-driven methods aim to learn the mapping from inputs (deformation, temperature, fields) to outputs (stress, heat flux) or to infer internal variable evolution. Approaches range from surrogate models and regression to hybrid physics–based models that enforce constraints such as invariance or thermodynamic admissibility. Data-driven models require careful training data coverage and robust extrapolation checks.
10.3 Multiphysics constitutive coupling (mechanical–thermal–electrical)
Coupled constitutive theories allow interactions among mechanical, thermal, and electrical processes. Examples include thermo-mechanical expansion, electro-mechanical deformation, and electrically driven heating affecting viscosity or phase behavior. Proper coupling requires consistent treatment of energy exchange and of how each field influences the others through shared internal variables or cross-coefficients.
10.4 Machine-learning-assisted material response modeling
Machine learning can accelerate parameter fitting, identify hidden relationships in experimental data, or build reduced-order surrogate models for constitutive response under varying conditions. When combined with mechanistic constraints and uncertainty quantification, these methods can provide faster predictions while retaining interpretability. Implementation still requires validation against experiments and careful attention to generalization beyond the training set.