Logarithmic Strain

How do we accurately describe the stretch of a rubber band or the forging of a metal part? While the question seems simple, conventional measures of strain break down when deformations become large or occur in multiple steps. This gap in understanding limits our ability to predict the behavior of materials under extreme conditions. This article introduces logarithmic strain, a powerful and elegant concept that provides the "true" measure of deformation. In the chapters that follow, we will explore its fundamental principles and then journey through its diverse applications. The first chapter, "Principles and Mechanisms," will deconstruct the idea of strain, revealing why the logarithm is the key to correctly accumulating deformation and how this concept extends into three dimensions using the language of tensors. Following this, the chapter on "Applications and Interdisciplinary Connections" will demonstrate how logarithmic strain is an indispensable tool in fields ranging from materials science and plasticity to geophysics and computational mechanics, providing a unified language to describe the way things bend, flow, and break.

Imagine stretching a rubber band. It gets longer. How much has it stretched? This simple question, which seems to have an obvious answer, opens a door to one of the most elegant concepts in the mechanics of materials. The way we choose to answer it determines our ability to describe the world of large, complex deformations, from the yielding of steel to the flow of glaciers.

Let's start with a simple metal bar of initial length $L_0$. We pull on it until its length becomes $L$. The most intuitive way to quantify the stretch is to calculate the change in length, $L - L_0$, and divide it by the original length, $L_0$. This gives us what engineers call the ​​engineering strain​​:

If we define the "stretch ratio" as $\lambda = L/L_0$, then the engineering strain is simply $\lambda - 1$. If the bar doubles in length, $\lambda=2$, and the engineering strain is $1$ (or 100%). This seems perfectly straightforward.

But let's perform a thought experiment. Suppose we first stretch the bar by 50%, so its new length is $L_1 = 1.5 L_0$. The engineering strain is $0.5$. Now, let's take this newly stretched bar and stretch it again by another 50% of its current length. The final length will be $L_2 = 1.5 L_1 = 1.5 \times (1.5 L_0) = 2.25 L_0$. The total engineering strain, measured from the very beginning, is $L_2/L_0 - 1 = 2.25 - 1 = 1.25$.

Here's the catch: we performed two "50% strain" operations, but the total strain is not $0.5 + 0.5 = 1.0$. It's $1.25$. The engineering strain is not additive. This is because at each step, our reference for "100%" length changes. This might seem like a minor inconvenience, but for materials that undergo enormous deformations, or for processes that occur in many sequential steps (like metal forging), this lack of additivity becomes a serious conceptual problem.

There must be a better way, a more "truthful" way to measure strain. The great minds of mechanics proposed this: instead of looking at the total change from the beginning, let's consider the deformation as a process of continuous, infinitesimal changes. At any given moment when the bar has length $l$, we stretch it by a tiny amount $dl$. The "instantaneous" strain is $dl/l$. To find the total "true" strain, we simply add up all these tiny fractional changes by integrating from the starting length $L_0$ to the final length $L$:

This quantity, $e_{\text{true}} = \ln(\lambda)$, is the ​​true strain​​, or, as we will call it, the ​​logarithmic strain​​.

Now let's revisit our two-step stretching experiment. The first 50% stretch corresponds to a stretch ratio $\lambda_1 = 1.5$. The logarithmic strain is $\ln(1.5)$. The second 50% stretch corresponds to a stretch ratio of $\lambda_2 = 1.5$ relative to the intermediate state. The logarithmic strain for this second step is also $\ln(1.5)$.

The total stretch ratio from start to finish was $\lambda_{\text{total}} = 2.25$. The total logarithmic strain is $\ln(2.25)$. But wait! Since the total stretch is the product of the individual stretches, $\lambda_{\text{total}} = \lambda_1 \lambda_2$, the beauty of the logarithm function reveals itself:

The logarithmic strains add up perfectly! This additive property is not just a mathematical convenience; it reflects a deeper truth about the nature of deformation as a cumulative process. This is the first hint of the power and elegance of the logarithmic strain measure. It is the correct way to accumulate finite stretches.

The real world, of course, is in three dimensions. A block of clay being squashed doesn't just stretch in one direction; it bulges in others, it shears, and it twists. To describe this, we need a more powerful mathematical tool: the ​​deformation gradient tensor​​, denoted by the matrix $\boldsymbol{F}$. You can think of $\boldsymbol{F}$ as a "master map" that tells you how any tiny vector in the undeformed body is transformed into a new vector in the deformed body.

A crucial insight, formalized in what is known as the ​​polar decomposition​​, is that any deformation can be uniquely broken down into two distinct actions: a pure stretch followed by a pure rigid-body rotation. Mathematically, we write this as:

Here, $\boldsymbol{U}$ is the ​​right stretch tensor​​, a symmetric matrix that captures all the stretching and shearing—the actual change in shape of the material. $\boldsymbol{R}$ is a ​​rotation tensor​​, which describes how the stretched shape is then rotated in space without any further change in shape. Since strain is fundamentally about the change in shape and size, it must be hidden entirely within $\boldsymbol{U}$. The rotation $\boldsymbol{R}$ has nothing to do with strain. All true measures of finite strain must, in some way, be functions of $\boldsymbol{U}$ alone, making them independent of any subsequent rigid rotation.

With this beautiful separation of concerns, we can now generalize our 1D logarithmic strain to 3D. The ​​Hencky strain tensor​​ (or logarithmic strain tensor) is simply defined as the logarithm of the stretch tensor:

What on earth does it mean to take the logarithm of a matrix? The secret lies in looking at the deformation from the right perspective. For any symmetric stretch tensor $\boldsymbol{U}$, we can always find a special set of three perpendicular axes—the ​​principal directions​​—along which the deformation is a pure stretch, with no shearing. The amount of stretch along these principal directions are the ​​principal stretches​​, $\lambda_1, \lambda_2, \lambda_3$. These are the eigenvalues of the matrix $\boldsymbol{U}$.

Taking the logarithm of the tensor $\boldsymbol{U}$ is then fantastically simple: we just take the logarithm of each of its principal stretches! The principal directions of the strain tensor $\boldsymbol{H}$ are the same as for the stretch tensor $\boldsymbol{U}$, and its principal values are simply $\ln(\lambda_1), \ln(\lambda_2),$ and $\ln(\lambda_3)$. So, the seemingly abstract tensor $\boldsymbol{H} = \ln \boldsymbol{U}$ is just a neat and tidy package for the three 1D logarithmic strains happening along the three principal axes of the deformation. Our 3D definition perfectly collapses back to our intuitive 1D picture.

This definition can also be written in terms of the ​​right Cauchy-Green tensor​​ $\boldsymbol{C} = \boldsymbol{F}^T\boldsymbol{F} = \boldsymbol{U}^2$. Since $\ln(\boldsymbol{U}^2) = 2 \ln \boldsymbol{U}$, we often see the Hencky strain defined as $\boldsymbol{H} = \frac{1}{2}\ln \boldsymbol{C}$.

This tensorial formulation of logarithmic strain possesses some remarkably elegant properties that make it a favorite of physicists and engineers.

One of the most beautiful is its connection to volume change. The volume of a small piece of material changes by a factor of $J = \det \boldsymbol{F}$. It turns out that the trace of the Hencky strain tensor (the sum of its diagonal elements, which is also the sum of its principal values) is exactly the logarithm of the volume change ratio:

This provides a wonderfully clean way to separate a deformation into a part that changes volume and a part that only changes shape (an ​​isochoric​​, or volume-preserving, deformation). For any deformation that preserves volume, such as the plastic flow of metals, we have $J=1$, which immediately implies that $\mathrm{tr}(\boldsymbol{H}) = 0$. This leads to an exact additive split of the strain into a volumetric part and a shape-changing (deviatoric) part, a feature that is immensely powerful in building theories of material behavior.

Furthermore, the additivity we cherished in the 1D case extends to 3D, with one crucial condition. If we perform a sequence of stretches, the total logarithmic strain tensor is the sum of the individual strain tensors if and only if the principal directions of stretch are the same for every step. We call such deformations ​​coaxial​​. This condition is required because, in the language of matrices, the logarithm of a product is the sum of the logarithms ($\ln(\boldsymbol{U}_2 \boldsymbol{U}_1) = \ln \boldsymbol{U}_2 + \ln \boldsymbol{U}_1$) only if the matrices commute ($\boldsymbol{U}_1 \boldsymbol{U}_2 = \boldsymbol{U}_2 \boldsymbol{U}_1$), a condition met by coaxial stretch tensors.

While there are many ways to define a finite strain tensor—such as the Green-Lagrange strain $\boldsymbol{E} = \frac{1}{2}(\boldsymbol{C}-\boldsymbol{I})$—the logarithmic strain holds a special place. Its additive nature for coaxial stretches makes it the natural language for theories of ​​plasticity​​, where materials flow and deform in a long sequence of incremental steps.

Moreover, in the world of computational mechanics, algorithms based on logarithmic strain are known for their exceptional robustness and accuracy, especially in problems involving large rotations. They correctly predict the material's response without producing "phantom" stresses that can plague simpler models, because they are built around the clean separation of stretch and rotation that we saw in the polar decomposition.

From a simple question about a rubber band, we have journeyed through a landscape of mathematics to find a quantity, the logarithmic strain, that is not just a definition, but a deep and unified principle. It gracefully handles the accumulation of deformation, elegantly separates shape from size, and provides the foundation for some of the most powerful theories and computational tools we have to understand our physical world.

We have journeyed through the abstract world of tensors and logarithms to define a new way of thinking about strain. You might be wondering, was it worth the effort? Is this "logarithmic strain" just a mathematical curiosity, a clever trick for the initiated, or does it tell us something profoundly true about the physical world? The answer, you will be delighted to find, is the latter. Like a key that unexpectedly opens not one but many doors, the concept of logarithmic strain unlocks a deeper understanding across a vast landscape of science and engineering. It reveals a surprising harmony in the way things bend, flow, and break.

Let’s now walk through some of these doors and see for ourselves how this idea is not just useful, but in many cases, indispensable.

Imagine you are watching a rubber band stretch in extreme slow motion. You analyze the movie frame by frame. In each tiny interval of time, the band gets a little bit longer. If you wanted to find the total amount it stretched, what would you do? Your first instinct might be to calculate the small percentage stretch in each frame and simply add them all up. But as we saw with the pitfalls of engineering strain, this will give you the wrong answer! Because the "base length" is constantly changing, the percentages don't add up correctly over large changes.

Here, logarithmic strain comes to the rescue with its most elegant and powerful property: it is additive. If you calculate the logarithmic strain for each small step of the deformation and add them all up, you get the exact total logarithmic strain from the beginning to the end. It’s the only strain measure that behaves this way for large, sequential deformations.

Think of it like compound interest. Engineering strain is like simple interest, always calculated on your initial deposit. Logarithmic strain is like continuously compounded interest, where each increment of growth is based on the current total. It is, in this sense, the "true" measure of accumulation.

This isn't just an academic point. Modern experimental techniques like Digital Image Correlation (DIC) rely on this very principle. DIC can track thousands of points on a deforming surface, providing a full map of the deformation from one video frame to the next. To get the total strain over a long experiment, scientists must use logarithmic strain; otherwise, their tally of the deformation would be fundamentally incorrect.

This additive nature also means that the rate of logarithmic strain is precisely the rate of deformation that a physicist would measure in the laboratory. This beautiful identity allows engineers in fields like rheology—the study of flow—to program a device to apply a complex, time-varying stretch rate to a material like a polymer melt and know that the total accumulated strain is simply the integral of that rate over time.

Deformation is not just about getting longer or shorter; it's about changing volume and shape. Once again, logarithmic strain provides a beautifully clean way to distinguish between these two effects.

Consider a block of geophysical material being compressed deep within the Earth's crust. It will be squished, reducing its volume, but it will also be distorted, changing its shape. Logarithmic strain allows us to separate these two actions perfectly. The sum of the principal logarithmic strains—the strains along the three perpendicular axes of maximum stretch—gives the exact logarithmic change in volume. The strain to change volume is simply $\epsilon_{\text{volumetric}} = \epsilon_1 + \epsilon_2 + \epsilon_3$. It's a wonderfully simple and exact relationship.

What's left over after we account for the volume change? The pure change in shape. This is what physicists call the deviatoric strain. The power of this concept is most clearly seen in a classic example: simple shear. Imagine shearing a deck of cards. The top card slides over the bottom one, but the total volume of the deck doesn't change. When we calculate the logarithmic strain tensor for this motion, we find a remarkable result: its trace (the sum of its diagonal elements) is exactly zero. This means the volumetric strain is zero, and the entire deformation is purely a change in shape. The mathematics perfectly mirrors the physical reality.

This ability to decompose deformation isn't limited to the principal axes. What if we are interested in a material with an internal structure, like the grain in a piece of wood or the fibers in a high-tech composite? We can use the machinery of continuum mechanics to calculate the logarithmic strain along any arbitrary fiber direction within the material, giving us a precise measure of how that specific structural element is being stretched or compressed. This is essential for designing materials that are strong in specific directions.

Perhaps the most profound applications of logarithmic strain are found in the field of materials science, where it has become the natural language for describing how materials behave under extreme conditions.

When you bend a paperclip and it stays bent, it has undergone plastic deformation. These deformations are often large and irreversible. To create theories that predict this behavior, we need a robust way to quantify the strain. Here, logarithmic strain is the undisputed champion. Engineers combine the components of the logarithmic strain tensor into a single scalar value called the von Mises equivalent strain. This value represents the total amount of distortion the material has experienced, irrespective of direction. By plotting the stress in a material against this equivalent logarithmic strain, we can obtain the material's unique "fingerprint" or constitutive law. This stress-strain curve, properly measured using logarithmic strain, is the fundamental input for computer simulations that predict everything from how a car's frame will crumple in a crash to how a building will respond to an earthquake. The framework is even flexible enough to be adapted for materials like soils and rocks, whose behavior depends on the ambient pressure.

The story doesn't end with metals. Consider "smart materials" like shape memory alloys (SMAs). These materials can be deformed and then, upon heating, magically return to their original shape. This effect is due to a phase transformation in the crystal structure. This transformation involves a large, built-in strain, which is best characterized as a logarithmic transformation strain. Understanding this strain is key to designing devices from self-deploying space antennas to medical stents that expand inside an artery.

Or think of soft matter, like a polymer melt being pulled into a thin fiber. As the fiber is stretched, its diameter shrinks. Incompressibility dictates a precise relationship between the axial stretch and the radial shrinkage. When expressed in terms of Hencky strain, this relationship takes on a beautifully simple exponential form: the diameter decreases exponentially with half the axial strain. This fundamental result is used daily in the rheology of polymers, foods, and other complex fluids.

In our modern world, much of engineering and science is done inside a computer. How do we translate these physical ideas into robust algorithms? This is where the true genius of logarithmic strain reveals itself. The equations for large plastic deformation are notoriously complex. However, by adopting a framework based on logarithmic elastic strain, computational mechanicians have found that the numerical algorithm for solving these complex problems—the so-called "return mapping"—becomes formally identical to the much simpler algorithm used for infinitesimal strains. It is a stunning result. The right choice of mathematical language tames the wild complexity of finite plasticity, making our simulations more elegant, efficient, and robust.

But with great power comes the need for great understanding. Is logarithmic strain a panacea? No theory is. The simplest and most elegant plasticity models built on Hencky strain are, at their heart, non-linear elastic models. This means they are path-independent: the stress depends only on the final strain, not the history of how it got there. Real materials, however, often have memory. If you bend a metal bar one way, it often becomes easier to bend it back the other way—a phenomenon known as the Bauschinger effect. A simple Hencky deformation theory, by its very nature, cannot capture this path-dependent behavior.

This limitation is not a failure but a lesson. It tells us precisely where the boundary of our simple model lies and illuminates the path forward to more sophisticated "flow theories" that incorporate history. A good theory not only explains a great deal but also clearly defines the borders of its own knowledge.

From tracking the slow creep of rock to simulating the high-speed crash of a vehicle, from pulling taffy to designing self-healing materials, the logarithmic strain provides a unified and powerful perspective. It is a testament to the fact that sometimes, the most abstract-seeming mathematical ideas turn out to be the most practical, revealing the deep and elegant structure of the physical world.