Nanson's formula

In the study of how materials stretch, twist, and flow—the domain of continuum mechanics—the deformation gradient tensor ($\mathbf{F}$) is a central character. It masterfully describes how infinitesimal lines are transformed and, through its determinant, how volume elements expand or contract. However, a critical piece of the puzzle remains: what happens to surfaces? While we have clear rules for lines and volumes, the transformation of an oriented area is a more subtle and non-intuitive affair. This article addresses this fundamental question by exploring Nanson's formula, the elegant equation that governs how areas change in size and orientation during deformation. The following chapters will first unpack the "Principles and Mechanisms," deriving the formula and revealing its surprising consequences for geometry. Subsequently, the "Applications and Interdisciplinary Connections" chapter will demonstrate how this seemingly abstract rule is a cornerstone of modern engineering, essential for linking stress measures and building accurate computational simulations.

Imagine you are watching a piece of dough being kneaded. It stretches, it folds, it gets squashed. If you had painted a tiny line on it, that line would now be longer and pointing in a new direction. If you had somehow injected a tiny bubble of air inside, that bubble would now be a different size, and likely a different shape. Describing this complex dance of deformation is the heart of continuum mechanics.

We already know from our introduction that the motion of a material is a story of points moving from a reference, or "before," configuration $\mathbf{X}$ to a current, or "after," configuration $\mathbf{x}$. The local instructions for this transformation are encoded in a remarkable mathematical object called the ​​deformation gradient​​, denoted by $\mathbf{F}$. In essence, $\mathbf{F}$ tells any infinitesimal arrow $d\mathbf{X}$ in the original body where it ends up, both in length and orientation, in the deformed body: $d\mathbf{x} = \mathbf{F} d\mathbf{X}$.

This single tensor holds profound secrets. Its determinant, a single number we call the ​​Jacobian​​ and write as $J = \det(\mathbf{F})$, tells us how the volume changes. If you take an infinitesimal volume element $dV$ in the "before" state, its new volume $dv$ in the "after" state is simply $dv = J \, dV$. If $J \gt 1$, the material has expanded locally; if $J \lt 1$, it has been compressed. Physicists and engineers generally insist that $J$ must be positive, because a negative $J$ would mean the material has turned itself inside-out, a mathematical possibility we don't observe in the physical world.

But this leaves us with a fascinating puzzle. We have a rule for lines and a rule for volumes. What about surfaces?

A surface is more subtle than a line or a volume. It has not only a size—its area—but also an orientation, which we can describe with a normal vector pointing away from it. How does an infinitesimal area element, say a tiny patch with area $dA$ and normal vector $\mathbf{N}$ in the reference body, transform into its new state with area $da$ and normal $\mathbf{n}$?

You might guess that the new area is just the old area scaled by the volume ratio, $da = J \, dA$. Or perhaps the new normal vector is just the old one pushed forward by the deformation, $\mathbf{n} = \mathbf{F} \mathbf{N}$. Both of these intuitive guesses turn out to be wrong, and the truth is far more beautiful and interesting.

Let's uncover this truth with a bit of reasoning. Imagine a very flat, coin-shaped volume element in the reference configuration. Its volume $dV$ can be thought of as the area of its face, $d\mathbf{A}$ (a vector whose magnitude is the area $dA$ and whose direction is the normal $\mathbf{N}$), multiplied by its infinitesimal thickness, represented by a vector $d\mathbf{X}$. The volume is the dot product: $dV = d\mathbf{X} \cdot d\mathbf{A}$.

Now, let's see what happens after deformation. The little coin is now a warped, stretched version of its former self. Its new volume is $dv$, its face is $d\mathbf{a}$, and its thickness has become $d\mathbf{x}$. The new volume is $dv = d\mathbf{x} \cdot d\mathbf{a}$. We can now use the rules we already know:

1. $dv = J \, dV = J(d\mathbf{X} \cdot d\mathbf{A})$
2. $d\mathbf{x} = \mathbf{F} d\mathbf{X}$

Substituting these into our expression for the new volume gives:

A wonderful property of tensors allows us to move $\mathbf{F}$ to the other side of the dot product, as long as we take its transpose: $( \mathbf{A} \mathbf{u} ) \cdot \mathbf{v} = \mathbf{u} \cdot (\mathbf{A}^T \mathbf{v})$. Applying this, we get:

This equation must hold true for any choice of the tiny thickness vector $d\mathbf{X}$. The only way for this to be universally true is if the parts multiplying $d\mathbf{X}$ on both sides are identical. And so, we arrive at a profound conclusion:

By simply rearranging this equation to solve for the new area vector $d\mathbf{a}$, we find the celebrated relationship known as ​​Nanson's formula​​:

Here, $\mathbf{F}^{-T}$ is shorthand for the inverse of the transpose of $\mathbf{F}$. This is the master formula that governs how oriented areas transform.

Let's take a moment to appreciate what this equation is telling us. The new area vector $d\mathbf{a}$ depends on three things: the original area vector $d\mathbf{A}$, the volume change factor $J$, and this strange character, $\mathbf{F}^{-T}$, which accounts for the change in shape and orientation. Notice that the new normal vector $\mathbf{n}$ is not parallel to the pushed-forward vector $\mathbf{F}\mathbf{N}$, but rather to $\mathbf{F}^{-T}\mathbf{N}$. This is a crucial, non-intuitive insight.

The most striking consequence is that area change is ​​anisotropic​​—it depends on the orientation of the surface. A simple, concrete example makes this astonishingly clear. Imagine a block of material that is stretched to three times its length in the $\mathbf{e}_1$ direction, while being compressed to 0.6 times its size in the $\mathbf{e}_2$ and $\mathbf{e}_3$ directions. The deformation gradient would be a diagonal matrix with these values: $\mathbf{F} = \mathrm{diag}(3, 0.6, 0.6)$.

First, what happens to the volume? The Jacobian is $J = \det(\mathbf{F}) = 3 \times 0.6 \times 0.6 = 1.08$. The volume has increased by 8%!.

Now, let's use Nanson's formula to see what happens to a surface that was originally facing the direction of the stretch, with a normal vector $\mathbf{N} = \mathbf{e}_1$. The ratio of the new area to the old area, $da/dA$, can be found by taking the magnitude of Nanson's formula: $da/dA = J |\mathbf{F}^{-T} \mathbf{N}|$. For our diagonal $\mathbf{F}$, its inverse transpose is simply $\mathbf{F}^{-T} = \mathrm{diag}(1/3, 1/0.6, 1/0.6)$. Plugging in $\mathbf{N} = \mathbf{e}_1 = (1, 0, 0)^T$, we find that $\mathbf{F}^{-T}\mathbf{e}_1$ is the vector $(1/3, 0, 0)^T$, whose magnitude is $1/3$. The area change factor is therefore:

This is remarkable! Even though the total volume of the material expanded, this particular surface has shrunk to only 36% of its original area. Meanwhile, a surface facing the $\mathbf{e}_2$ direction would have its area multiplied by a factor of $1.08 \times (1/0.6) = 1.8$, meaning it grows by 80%. Area change is not a simple, uniform affair; it is a rich, directional phenomenon, beautifully captured by the anisotropy embedded in Nanson's formula. We can use this formula to directly calculate the deformed area vector for any given deformation.

The beauty of the formula is that the area change can be conceptually split. Part of the change comes from the overall swelling or shrinking of the volume, and the other part comes from the pure change in shape (shear), which stretches the area in some directions while squashing it in others.

This might seem like a purely geometric curiosity, but Nanson's formula is a cornerstone of mechanics because it connects geometry to the real-world concept of ​​forces​​. In engineering and physics, we are often concerned with ​​stress​​, which is force per unit area.

The problem is, which area? Do we mean the area in the original, undeformed shape, or the area in the final, deformed shape? These are different! Nanson's formula provides the crucial link. The "true" physical stress in the deformed body is the ​​Cauchy stress​​, $\boldsymbol{\sigma}$, which is the force $d\mathbf{f}$ per current area $d\mathbf{a}$. However, for calculations, it is often more convenient to work with the ​​First Piola-Kirchhoff stress​​, $\mathbf{P}$, which relates the same force $d\mathbf{f}$ back to the original area $d\mathbf{A}$.

The force is the same physical entity, regardless of our description:

If we substitute Nanson's formula, $d\mathbf{a} = J \mathbf{F}^{-T} d\mathbf{A}$, into this relationship, a direct connection between the two stress tensors magically appears:

This expression is fundamental to virtually all modern engineering simulations, from designing airplane wings to modeling biological tissues.

Furthermore, Nanson's formula is part of a deep and unified mathematical structure. It connects the deformation gradient $\mathbf{F}$ not only to stress but also to measures of strain, like the ​​Cauchy-Green deformation tensors​​ ($\mathbf{C} = \mathbf{F}^T \mathbf{F}$ and $\mathbf{B} = \mathbf{F} \mathbf{F}^T$). These tensors quantify how lengths and angles have changed. For example, knowing the deformation, one can express the original area's magnitude purely in terms of the deformed area and the tensor $\mathbf{B}$. Conversely, if we have special information—for instance, if we know that a certain surface's area is perfectly preserved during deformation—we can use Nanson's formula to deduce specific properties of the strain tensor $\mathbf{C}$.

Nanson's formula is far more than a dry equation. It is a portal that connects the elegant geometry of deformation to the physical reality of forces and material response, revealing the surprising and anisotropic ways in which surfaces stretch, shrink, and reorient themselves in a deforming world.

After our exploration of the principles behind Nanson's formula, you might be left with a sense of its geometric elegance. But is it merely a mathematical curiosity? Something for the connoisseurs of abstract kinematics? Far from it. As we are about to see, this single, beautiful relationship is a master key that unlocks doors across the vast landscape of mechanics, engineering, and computational science. It acts as a universal translator, a robust bridge between the world as it is and the world as it was, allowing us to build models that are not only powerful but physically true.

Imagine you are trying to describe the internal forces within a squashed rubber ball. You could measure the force acting on a tiny patch of the squashed ball and divide it by that patch's current, deformed area. This gives you what we call the ​​Cauchy stress​​ ($\boldsymbol{\sigma}$), a true measure of force intensity in the here-and-now. But what if it's easier to do your bookkeeping on the ball's original, perfectly spherical shape? You might instead relate the same force to the area of the patch before it was squashed. This gives you a different quantity, the ​​first Piola-Kirchhoff stress​​ ($\mathbf{P}$), a sort of hybrid measure that connects the original geometry to the current forces.

Are these two types of stress just different accounting methods? How are they related? The answer lies in a beautiful piece of physical reasoning. The actual incremental force, $d\mathbf{f}$, acting on that tiny patch of material is a physical reality; it cannot depend on which coordinate system or area definition we choose to use. The force on the current patch, which we can write as the current traction $\mathbf{t}$ times the current area $da$, must be identical to the force described in the reference frame, the nominal traction $\mathbf{t}_0$ times the original area $dA$.

$\mathbf{t} da = \mathbf{t}_0 dA$

This simple statement of force balance is where Nanson's formula enters the stage with breathtaking effect. We replace the oriented area element on the left, $\mathbf{n} da$, with its referential equivalent given by Nanson's formula, $\mathbf{n} da = J \mathbf{F}^{-T} \mathbf{N} dA$. By substituting the definitions of traction ($\mathbf{t} = \boldsymbol{\sigma}\mathbf{n}$ and $\mathbf{t}_0 = \mathbf{P}\mathbf{N}$) and applying this transformation, we discover that what seemed like two different descriptions of stress are in fact inseparable. Because the relationship must hold for any surface patch we choose, the tensors themselves must be related by the magnificent formula:

$\mathbf{P} = J \boldsymbol{\sigma} \mathbf{F}^{-T}$

This is the Rosetta Stone of continuum mechanics. It allows us to translate perfectly between the language of the current, deformed world (Cauchy stress) and the language of the original, undeformed world (Piola-Kirchhoff stress). This isn't just an exercise in algebra; it's the conceptual link that ensures our physical models are consistent. The power of this translation extends even further; it is the very tool that allows us to transform the local differential equation for momentum balance, relating the divergence of the Cauchy stress in the current coordinates to the divergence of the Piola-Kirchhoff stress in the reference coordinates. Nanson's formula ensures that the laws of physics are preserved across these different mathematical descriptions.

This ability to translate between configurations is not just a theoretical nicety; it is essential for solving real-world engineering problems. Consider the challenge of analyzing an inflatable structure, like a weather balloon or a car tire. The force driving the inflation is a pressure, which acts perpendicularly on the balloon's current, stretched surface. However, to design the balloon, we want to perform our calculations on its simple, uninflated shape. How do we correctly represent the pressure acting on a complex, deformed geometry as a force on our simple, original model?

You can't just apply the same pressure value. When the balloon inflates, any given patch of its surface not only stretches—changing its area—but also reorients itself in space. Nanson's formula is the precise tool to account for both effects. By using it, we can calculate the equivalent traction (force per unit original area) that we must apply to our reference model to produce the same physical effect as the pressure on the final, deformed shape.

A beautiful example of this is the bending of a plate. Imagine bending a flat rectangular plate into a curve. The outer surface must stretch more than the inner surface. If a uniform traction is applied to the deformed outer surface, what is the equivalent load on the original flat plate? Nanson's formula tells us that the area of the top surface stretches by a factor that depends on the plate's thickness and the curvature of the bend. This means the equivalent force on the original, flat reference shape is stronger than what you might naively expect, and this strengthening effect is more pronounced for a tighter bend. Nanson's formula automatically captures this intuitive physical phenomenon, translating a simple load in the deformed world into a more complex, position-dependent load in the reference world, where our calculations are often easier to perform.

The role of Nanson's formula becomes even more profound when we venture into the world of computational simulation. Modern engineering relies on the Finite Element Method (FEM), a numerical technique for solving the complex equations of mechanics. When dealing with large deformations—like a car crash or the deployment of a stent—engineers use one of two main frameworks: the Total Lagrangian (TL) or the Updated Lagrangian (UL) formulation. Nanson's formula is the quiet, indispensable engine at the heart of both.

In a ​​Total Lagrangian​​ formulation, the simulation always refers back to the object's original, undeformed geometry. This is computationally convenient because the "grid" never changes. The challenge, as we've seen, is that real-world forces like pressure are defined on the current, deformed shape. To include these forces in the simulation, they must be mathematically "pulled back" and converted into an equivalent load on the reference grid. The explicit recipe for doing this, a cornerstone of every nonlinear FEM code, is derived directly from Nanson's formula. It computes the work-equivalent referential traction vector, often called a "follower force" because its equivalent representation on the reference body changes as the body deforms. This very dependence on the deformation is a source of what is called ​​boundary nonlinearity.​​

In an ​​Updated Lagrangian​​ formulation, the simulation advances in small steps, and the reference configuration is continuously updated to be the configuration from the previous step. Here, Nanson's formula plays a different but equally critical role. At the end of each incremental step, the code must know the new geometry of the boundary—the new areas and new normal vectors—to apply forces for the next step. Nanson's formula, applied incrementally, provides the exact rule for updating the oriented area vectors from one configuration to the next. It is the kinematic engine that drives the evolution of the model's geometry through time.

Finally, Nanson’s formula serves a subtle but vital role as a quality-control inspector, helping us diagnose errors in our computational models. In the world of FEM, a smooth, curved surface is always approximated by a mosaic of small, flat patches, or "facets." This is a necessary simplification, a "geometry crime" we commit to make the problem tractable for a computer. But does it introduce errors?

Consider a single flat facet on the original mesh. After deformation, its vertices are mapped to new positions, defining a new facet in the current configuration. We can now describe the oriented area of this new facet in two different ways:

1. Directly, by calculating the area and normal from the coordinates of the new, deformed facet.
2. Indirectly, by taking the original facet's area and normal and "pushing it forward" using the discrete version of Nanson's formula.

In the perfect world of continuum mathematics, these two descriptions would be identical. In the discrete world of our computer model, they are not! The difference between them is a direct measure of the error introduced by our geometric approximation. This "Nanson's relation residual" tells us how faithfully our discrete model is capturing the true, continuous kinematics of the deformation. If this error doesn't shrink as we refine our mesh, it signals a fundamental problem in our simulation's ability to represent reality. This provides a powerful, physics-based diagnostic for developers and analysts to verify the consistency and accuracy of their complex simulations.

From a piece of pure geometry to a translator of stress, a practical tool for engineers, a core algorithm in computational solvers, and a final check on the integrity of our simulations, Nanson's formula stands as a testament to the profound unity of mathematics and the physical world. It is one of the unseen architects of modern mechanics, working silently in the background to ensure that our models of reality are not only powerful, but true.