Objective Rates in Continuum Mechanics

In the study of how materials deform—a field known as continuum mechanics—a fundamental principle governs our mathematical descriptions: the laws of physics must not depend on the observer. When materials undergo large rotations, like a steel beam twisting or a polymer melt flowing, a simple description of stress change becomes contaminated by the observer's viewpoint. This creates a significant knowledge gap, where naive models predict physically impossible phenomena, such as a material generating stress simply by spinning. This article addresses this problem head-on by delving into the concept of ​​objective rates​​.

This exploration will guide you through the core theory and practical implications of ensuring our material models are objective. You will learn:

• ​​Principles and Mechanisms:​​ We will first uncover why the standard time derivative fails and how objective rates are constructed to correct for material spin. This section introduces key models like the Jaumann and Green-Naghdi rates and reveals the surprising and unphysical consequences that can arise from choosing the wrong one.
• ​​Applications and Interdisciplinary Connections:​​ Next, we will bridge theory and practice by examining how these concepts are crucial in modern computational engineering. We will discuss the impact of objective rates on simulation accuracy, algorithmic efficiency in the Finite Element Method, and the development of robust models for plasticity, rheology, and material failure.

Imagine you are a sculptor working with a lump of clay on a spinning potter's wheel. The clay deforms under your hands, but it’s also spinning. How would you describe the true rate of change of stress inside the clay? The stress is changing because you are squeezing it, but its orientation is also changing simply because it's rotating. If you describe the change from your stationary viewpoint, you mix up these two effects. If a friend were spinning on the wheel alongside the clay, they would see the same squeezing but no rotation. Surely, a fundamental law of the clay's material nature shouldn't depend on whether you or your spinning friend are the ones describing it. This is the heart of a deep principle in physics: ​​material frame indifference​​, or ​​objectivity​​. Our description of a material's intrinsic behavior must be independent of the observer's own rigid motion.

Let's think like a physicist. We want to write a law that connects how a material deforms to how its internal stress changes. A simple guess for a rate-type law might be: "the rate of change of stress is proportional to the rate of deformation." Mathematically, we might write $\dot{\boldsymbol{\sigma}} = \mathbb{C}:\boldsymbol{D}$, where $\dot{\boldsymbol{\sigma}}$ is the simple time derivative of the Cauchy stress tensor $\boldsymbol{\sigma}$, $\boldsymbol{D}$ is the rate-of-deformation tensor (which measures stretching), and $\mathbb{C}$ is a tensor of elastic constants.

This looks plausible, but it hides a fatal flaw. Consider a block of steel that is already stressed, but is now simply spinning rigidly in space, like a tossed brick. Since it is not deforming, its rate-of-deformation tensor $\boldsymbol{D}$ is zero. Our naive law would predict $\dot{\boldsymbol{\sigma}} = \boldsymbol{0}$. This means the components of the stress tensor, measured in a fixed laboratory frame, should not be changing. But this is absurd! As the block rotates, the direction of its internal forces rotates with it. The components of the stress tensor are changing in our fixed frame. Our proposed "law" has failed spectacularly; it predicts that a rigid rotation generates or alters stress, a phantom effect that depends entirely on our viewpoint. The simple material time derivative, $\dot{\boldsymbol{\sigma}}$, is not ​​objective​​; it is tainted by the observer's motion.

The principle of objectivity demands that for a pure rigid rotation where no actual deformation occurs ($\boldsymbol{D}=\boldsymbol{0}$), our constitutive law must predict no change in the intrinsic stress state. The failure of the simple time derivative means we need a more sophisticated way to measure the rate of change of stress.

The solution is as elegant as it is intuitive: we must measure the stress rate in a reference frame that co-rotates with the material. We need a special kind of derivative that automatically subtracts the part of the stress change that is merely due to the material's spinning. This special derivative is called an ​​objective stress rate​​, and there are many types, but they all share a common structure.

A corotational objective rate, let's call it $\overset{\diamond}{\boldsymbol{\sigma}}$, is generally defined as:

Here, $\boldsymbol{\Omega}$ is a skew-symmetric tensor representing the chosen ​​spin​​ of the co-rotating frame. The term $\boldsymbol{\Omega}\boldsymbol{\sigma} - \boldsymbol{\sigma}\boldsymbol{\Omega}$ is a commutator that precisely cancels out the non-objective part of $\dot{\boldsymbol{\sigma}}$ arising from rotation. Now, our hypoelastic law becomes $\overset{\diamond}{\boldsymbol{\sigma}} = \mathbb{C}:\boldsymbol{D}$. In our spinning block example, $\boldsymbol{D}=\boldsymbol{0}$, so the law correctly predicts $\overset{\diamond}{\boldsymbol{\sigma}}=\boldsymbol{0}$. This doesn't mean $\dot{\boldsymbol{\sigma}}=\boldsymbol{0}$, but rather that $\dot{\boldsymbol{\sigma}} = \boldsymbol{\Omega}\boldsymbol{\sigma} - \boldsymbol{\sigma}\boldsymbol{\Omega}$, which is exactly the equation describing a tensor whose components are changing solely because its reference frame is rotating with spin $\boldsymbol{\Omega}$! We have successfully separated deformation-induced stress from a simple change in orientation.

Interestingly, this rotational correction has a beautiful property. The trace of a tensor represents its mean value; for stress, this relates to pressure. The trace of the commutator term, $\mathrm{tr}(\boldsymbol{\Omega}\boldsymbol{\sigma} - \boldsymbol{\sigma}\boldsymbol{\Omega})$, is always zero because of the cyclic property of the trace ($\mathrm{tr}(AB) = \mathrm{tr}(BA)$). This means our rotational correction, regardless of the spin $\boldsymbol{\Omega}$ we choose, never affects the rate of change of the mean stress, a physically satisfying result.

Here the plot thickens. We said we need a frame that "rotates with the material," but what does that mean precisely? At a point in a deforming body, different material lines rotate at different rates. This ambiguity has led to a "zoo" of different objective rates, each distinguished by its choice for the spin tensor $\boldsymbol{\Omega}$.

• ​​The Jaumann Rate:​​ This is perhaps the most straightforward choice. It defines the spin $\boldsymbol{\Omega}$ as the ​​spin tensor​​ $\boldsymbol{W}$, which is the skew-symmetric part of the velocity gradient $\nabla \boldsymbol{v}$. You can think of $\boldsymbol{W}$ as the rotation rate of an infinitesimally small pinwheel embedded in the material. It represents the average angular velocity of the material at that point.

• ​​The Green-Naghdi Rate:​​ A more sophisticated approach arises from the ​​polar decomposition​​ of the deformation gradient, $\boldsymbol{F}=\boldsymbol{R}\boldsymbol{U}$. This mathematical theorem states that any deformation can be decomposed into a pure stretch ($\boldsymbol{U}$) followed by a pure rigid rotation ($\boldsymbol{R}$). The Green-Naghdi rate uses the spin of this "true" material rotation, $\boldsymbol{\Omega} = \dot{\boldsymbol{R}}\boldsymbol{R}^{\top}$. It attempts to track the rotation of the material element as a whole, decoupled from its internal stretching.

Other rates, like the ​​Truesdell rate​​, exist as well, each with its own physical interpretation and mathematical properties. The Jaumann and Green-Naghdi rates are generally not the same because the spin of the infinitesimal pinwheel ($\boldsymbol{W}$) is not the same as the spin of the material element as a whole ($\dot{\boldsymbol{R}}\boldsymbol{R}^{\top}$), except in special cases like pure rigid-body motion.

Are these different rates just academic curiosities? Far from it. The choice of objective rate can have dramatic and physically significant consequences, especially when rotations are large.

The classic test case is ​​simple shear​​, the motion you create when you slide a deck of cards. Let's imagine performing a computational experiment on a simple elastic material using a hypoelastic law with different objective rates.

If we use the ​​Jaumann rate​​, we get a bizarre result. As we monotonically increase the shear, the predicted shear stress does not increase monotonically. Instead, it rises, reaches a peak, and then begins to decrease, producing spurious oscillations! This is completely unphysical; our model suggests the material gets weaker and then stronger again as we continue to shear it in the same direction. This pathological behavior stems from the fact that the Jaumann rate model is not ​​integrable​​; it does not correspond to a material with a well-defined elastic potential energy. Objectivity alone is not enough to guarantee a physically sensible elastic model.

In contrast, if we run the same simulation with the ​​Green-Naghdi rate​​, the response is much more reasonable. The shear stress typically increases monotonically, avoiding the unphysical oscillations of the Jaumann rate. Because its spin is tied to the more fundamental polar rotation $\boldsymbol{R}$, it performs better in this large-rotation scenario.

This issue is even more critical in simulations of plasticity. A naive numerical implementation of a perfectly plastic material (one that should not get any stronger after it starts to yield) can exhibit ​​artificial hardening​​ when using a poorly chosen rate or integration scheme. The accumulation of small errors in integrating the rotation step by step can make the material appear to strengthen, a purely numerical artifact that can lead to completely wrong engineering predictions.

The challenges and paradoxes associated with hypoelasticity and the zoo of objective rates have guided the development of modern computational mechanics. While historically crucial, these rate-based formulations are often superseded today by more robust frameworks.

Many modern approaches are founded on ​​hyperelasticity​​, where the stress is derived directly from a stored energy function, guaranteeing integrability from the outset. In plasticity, the ​​multiplicative decomposition​​ of deformation ($\boldsymbol{F} = \boldsymbol{F}^e \boldsymbol{F}^p$) provides a powerful kinematic foundation that elegantly separates elastic deformation, plastic flow, and rigid-body rotation. Formulations based on ​​logarithmic strains​​ and their corresponding objective rates are also known to be free from many of the artifacts that plague simpler models. These advanced theories correctly handle the interplay of stretching and spinning, ensuring that a simulated block of steel behaves like a real block of steel, no matter how it tumbles and turns. The journey through the zoo of objective rates, with all its pitfalls, was a necessary one, teaching us the profound importance of getting rotation right.

Now that we have grappled with the principles of objective rates, we might find ourselves asking, "What is it all for?" It is a fine thing to construct an elegant mathematical framework, but does nature really care about our choice of derivatives? The answer, it turns out, is a resounding yes. The journey to understand and apply objective rates is a wonderful illustration of how a seemingly abstract physical principle—that the laws of nature cannot depend on the observer—becomes a powerful and indispensable tool in science and engineering. This principle guides us in describing the behavior of everything from the steel in a skyscraper and the rubber in a tire, to the living tissues in our own bodies.

Why do we need objective rates at all? Let us begin with a simple, yet profound, observation. Imagine a block of metal being bent and twisted. Its internal resistance to deformation—its stress—will change. A simple material time derivative, $\dot{\boldsymbol{\sigma}}$, seems like the natural way to describe this change. However, as we have seen, this simple derivative is a bit naive. If the block is simply spun around without any change in shape, $\dot{\boldsymbol{\sigma}}$ will be non-zero! It gets confused between real changes in stress due to deformation and apparent changes due to the rotation of our viewpoint. The principle of material frame indifference tells us this cannot be right; a mere rotation shouldn't generate stress.

This is the very heart of the problem. For small, gentle deformations and rotations, this spurious rotational effect is tiny—a second-order nuisance we can often ignore. But in the world of finite deformations, where things can rotate significantly, this "nuisance" becomes a first-order disaster, leading to completely wrong predictions. We are thus forced to invent a "smarter" derivative, an objective rate, that knows how to subtract the purely rotational effects.

This leads us to a fascinating fork in the road of material modeling. It turns out that not all materials require us to go through the trouble of defining and using these special rates. The necessity depends entirely on the nature of the material's "memory."

For a vast class of materials, like metals undergoing plastic deformation or polymers exhibiting viscous flow, their current state of stress depends on their entire history of deformation. Their memory is encoded in the path they took. For these materials, we have no choice but to describe their behavior with a rate-type constitutive law, relating the rate of stress change to the rate of deformation. In this world, using an objective rate is mandatory. This is not only true for the stress $\boldsymbol{\sigma}$, but for any internal memory variables as well. For instance, in more advanced models of plasticity that account for the Bauschinger effect, a "backstress" tensor $\boldsymbol{\alpha}$ is introduced to track the center of the yield surface. Just like stress, this backstress is a tensor quantity that rotates with the material, and its evolution law must also employ an objective rate to remain physically meaningful under large rotations. The same logic applies beautifully to the field of rheology, which studies the flow of materials like polymers and biological fluids. Complex models such as the generalized Maxwell model, which describes viscoelastic behavior, are now routinely formulated for large strains using these very principles, ensuring that predictions for things like polymer melts in an extruder or biological tissues under load are physically consistent.

But there is another, more elegant path for a different class of materials. Think of a perfectly elastic material, like a spring or a rubber band, which has no memory of its history. Its stress depends only on its current state of deformation, not on how it got there. For these so-called hyperelastic materials, we can define a potential energy function—the strain energy $W$. The beauty of this approach is that we can build objectivity in from the very beginning. Instead of defining the energy as a function of the non-objective deformation gradient $\boldsymbol{F}$, we define it as a function of a purely objective measure of strain, such as the right Cauchy-Green tensor $\boldsymbol{C} = \boldsymbol{F}^{\top}\boldsymbol{F}$, which cleverly "forgets" any rigid rotation. Once we have $W(\boldsymbol{C})$, the stress at any configuration is found simply by taking a derivative of this potential. There is no need to integrate a rate over time, and thus no need for an objective stress rate!. This potential-based approach is the gold standard for its elegance and inherent consistency and is the foundation for the most advanced models of material behavior, including those that couple elasticity with other phenomena like material damage.

The distinction between these modeling approaches is not just an academic's delight; it has profound consequences in the world of computational engineering. The breathtaking simulations you see of car crashes, jet engine turbines, or artificial heart valves all rely on solving the equations of continuum mechanics on a computer, most often using the Finite Element Method (FEM). In these simulations, the computer must update the stress in every tiny piece of the model, step by step, as it deforms. This is where our choice of objective rate comes to life.

In a typical nonlinear simulation, the computer solves a massive system of equations at each time step using a method akin to Newton-Raphson. The efficiency of this solver depends critically on a quantity known as the consistent tangent modulus (or tangent stiffness matrix). This matrix tells the solver how the internal forces will change in response to a small change in displacement. For the solver to converge quickly and robustly, it is highly desirable for this matrix to be symmetric.

Herein lies the rub. If one builds a model using a simple hypoelastic law with, say, the Jaumann objective rate, the resulting consistent tangent matrix is generally not symmetric. The very terms that we add to make the stress rate objective come back to haunt us in the linearization, introducing non-symmetric components. We can see this explicitly if we perform the calculation: the spin of the chosen rotational frame, which differentiates one objective rate from another, appears directly in the off-diagonal terms of the final tangent matrix, spoiling its symmetry.

This has motivated a decades-long quest in the computational mechanics community for better formulations. One path is to use clever "corotational" techniques, which try to perform the stress update in a local reference frame that rotates with the material, effectively filtering out the troublesome rotations. Another, more modern approach is to abandon simple hypoelastic laws altogether and build elastoplastic models on a hyperelastic foundation, for instance, using logarithmic strain measures. These formulations are not only more accurate but are designed to produce a symmetric tangent modulus by construction, fulfilling the desires of both the physicist for energetic consistency and the computational scientist for algorithmic efficiency.

What happens if we make the wrong choice? The consequences can be startlingly unphysical. Consider one of the most famous test cases in the field: a block of material subjected to large-amplitude cyclic simple shear, where it is sheared back and forth symmetrically.

Common sense suggests what should happen. For an elastic material, it should deform and return to its original stress-free state after each cycle, with no net energy spent. For an elastoplastic material with isotropic hardening, it should shake down into a stable, symmetric hysteresis loop centered around zero stress.

Yet, if we simulate this with a hypoelastic model based on the Zaremba-Jaumann rate, something bizarre occurs. The model can predict that after a closed elastic cycle, the material has somehow generated energy from nothing! Worse, in the plastic case, the stress response becomes asymmetric. Even though the strain is cycling symmetrically around zero, the stress loop starts to "walk" or "ratchet" up cycle after cycle, predicting a buildup of mean stress that simply doesn't happen in reality. This is a spectacular failure of the model, a ghost in the machine born from an imperfect treatment of rotation.

In contrast, a corotational formulation based on the Green-Naghdi rate (which follows the true material rotation) or a hyperelastic-based plastic model behaves perfectly. It correctly predicts zero net work in an elastic cycle and a stable, non-ratcheting response in the plastic case. This simple example is a powerful reminder that our abstract mathematical choices have direct, observable consequences and can be the difference between a predictive simulation and computational fiction.

The principles of objectivity reach into the most challenging and critical areas of materials science, such as predicting when and how materials fail. To model the process of ductile fracture, for example, we must account for the nucleation and growth of microscopic voids within the material, a phenomenon we can represent with a scalar "damage" variable, $D$.

How should we build a theory for this complex, coupled process of elastic-plastic deformation and accumulating damage, valid for the large strains that precede failure? The most robust and successful theories do exactly what we have discussed. They begin with a hyperelastic potential, $\psi$, that depends on objective measures of elastic strain and the scalar damage variable. This immediately guarantees that the model is frame-indifferent and thermodynamically sound. By deriving all driving forces—for stress, for plastic flow, for damage growth—from this single potential, we create a unified and consistent framework. It is a beautiful synthesis, where the abstract requirement of frame indifference, first recognized over a century ago, provides the essential foundation for building practical tools to ensure the safety and reliability of the world around us.