Bond-Based Peridynamics

How do materials break? For centuries, classical continuum mechanics, with its smooth functions and differential equations, has struggled to answer this fundamental question. At the tip of a crack, where displacements are discontinuous, its mathematics falters, requiring complex add-ons from fracture mechanics to predict failure. What if there was a better way—a theory that embraces discontinuity as a natural part of its framework?

Enter peridynamics, a revolutionary nonlocal theory that reimagines materials as a collection of points interacting through a network of bonds. By replacing differential equations with integral ones, peridynamics can model the initiation and growth of complex crack patterns without special treatment, allowing failure to emerge naturally from the system's fundamental laws.

This article will guide you through the elegant core of this theory. The first chapter, "Principles and Mechanisms," will unpack the foundational concepts of the horizon, bonds, and how microscopic interactions give rise to macroscopic behavior. Subsequently, "Applications and Interdisciplinary Connections" will explore the profound impact of this perspective, from autonomous fracture simulation and wave dispersion to multiphysics problems like thermal shock and viscoelasticity.

Forget for a moment the smooth, continuous world of classical physics, where materials are like infinitely divisible jellies described by differential equations. Let's try a different picture, a view closer to the atomic reality of things, but with a wonderfully simplifying twist. Imagine that a solid object—a steel beam, a pane of glass, a rubber ball—is not a continuous whole, but a colossal collection of material points, like stars in a galaxy. Each point interacts with its neighbors not just by touching, but through "bonds" that reach across the empty space between them. This is the foundational idea of peridynamics. It’s a return to a kind of atomism, but one where we don’t need to know the messy details of every single atom.

In classical mechanics, a point in a material only "knows" about its immediate, infinitesimal neighbors. All interactions are local. Peridynamics boldly discards this. It proposes that every material point feels forces from all other points within a certain, finite distance. This region of interaction is called the ​​horizon​​, a sphere of radius $\delta$ centered on our point of interest. This parameter, $\delta$, is not just a computational artifact; it is a fundamental length scale built into the physics of the material itself. It defines the extent of ​​nonlocality​​.

This nonlocal character has profound consequences. Unlike a classical model that is "scale-free," a peridynamic material has an intrinsic sense of size. The behavior of the material is an average over the entire state of the horizon, which means that the theory naturally smoothes out or filters features that are much smaller than $\delta$. It has a built-in resistance to the unruly, singular behavior that plagues classical theories at sharp corners and crack tips.

The interaction itself is carried by ​​bonds​​, which are conceptual springs connecting pairs of material points. The state of the material is not described by local strain or stress, but by the collective state of all these bonds.

So, what is the law governing these bonds? Nature often opts for simplicity and elegance. In the simplest form of peridynamics, called ​​bond-based peridynamics​​, the force in each bond depends only on the deformation of that bond alone, independent of its neighbors.

Let's consider a single bond between two points. In its original, undeformed state, it's a vector we can call $\boldsymbol{\xi}$. After the material deforms, the bond becomes a new vector, $\boldsymbol{\xi} + \boldsymbol{\eta}$. The fundamental measure of deformation is the ​​stretch​​, $s$, which is simply the fractional change in the bond's length:

For a simple, linear elastic material, the force in the bond is directly proportional to this stretch. The magnitude of the force is simply $f_{scalar} = c s$, where $c$ is a constant we call the ​​micromodulus​​, representing the stiffness of that particular bond.

But force is a vector. In which direction does it point? For an isotropic material—one that looks and behaves the same in all directions—there is no reason for the force to point anywhere other than straight along the bond itself. If the bond energy depends only on its length (a scalar), then the force, which is the gradient of that energy, must point in the direction of the greatest change in length—along the deformed bond. This is the definition of a ​​central force​​. The pairwise force vector is thus beautifully simple:

where $\mathbf{e}$ is the unit vector along the deformed bond. The micromodulus $c$ can depend on the original bond length $|\boldsymbol{\xi}|$, allowing short bonds to be stiffer or weaker than long ones.

This is a remarkably simple set of rules for the microscopic world. But does it reproduce the familiar, macroscopic world of elasticity we can measure in the lab? This is where the true beauty of the theory reveals itself.

Let's imagine a simple one-dimensional bar under a uniform strain $\epsilon$. In this case, every single bond in the material experiences the exact same stretch, $s = \epsilon$. We can calculate the total strain energy stored in the material by summing up the energy of every bond. The energy of a single bond turns out to be proportional to $s^2$, just like a classical spring. When we integrate this bond energy over the entire horizon of a point, we get the total peridynamic strain energy density, $W_{PD}$.

For our 1D bar, a straightforward calculation shows this energy density is $W_{PD} = \frac{1}{2}(2c\delta)\epsilon^2$. Now, compare this to the classical formula for elastic energy you learned in freshman physics: $W_{classical} = \frac{1}{2}E\epsilon^2$, where $E$ is Young's modulus. They have the identical form! By simply looking at the two expressions, we see that the macroscopic, measurable Young's modulus is nothing more than an emergent property of the microscopic bond stiffness and the nonlocal horizon size:

This is a profound result. It bridges the microscopic model with the macroscopic world. The same procedure can be used to relate the micromodulus $c$ to other elastic constants like the bulk modulus $K$ and, in the limit where the horizon shrinks to zero, to recover the full Cauchy stress tensor of classical elasticity. Classical elasticity is simply the local limit of a more general, nonlocal theory.

If peridynamics was just another way to derive classical elasticity, it would be a mathematical curiosity. Its real power lies in what it can do that classical mechanics cannot: handle failure.

The equations of classical continuum mechanics are differential equations. They rely on the existence of derivatives of the displacement field. But at the face of a crack, the displacement makes a sudden jump. The derivative is undefined; the mathematics breaks down. To model a crack, classical theories must be augmented with separate, complex criteria known as fracture mechanics.

Peridynamics, however, needs no such additions. Its fundamental equation is an integral equation—it sums up forces. It doesn't care if the displacement field is smooth or not. A bond connecting two points on opposite sides of a crack simply sees a large relative displacement. The bond stretches, the force is calculated, and the simulation goes on. There are no derivatives to "blow up".

The mechanism for fracture is as elegant as the theory itself. We simply add one more rule to our bond law: if a bond's stretch $s$ exceeds a certain material-dependent threshold, the ​​critical stretch​​ $s_c$, the bond breaks. It is permanently removed from the system. As the material is loaded, bonds begin to stretch. The most highly strained bonds reach $s_c$ and snap, releasing their stored energy and transferring load to their neighbors. These neighbors stretch further, and a cascade of breaking bonds can lead to the formation of a macroscopic crack.

And here lies another beautiful micro-macro connection. The macroscopic fracture toughness $G_c$, a property that measures the energy required to create a unit area of new crack surface, can be directly related to the microscopic critical stretch. The energy $G_c$ is simply the sum of all the potential energy stored in the bonds that had to be broken to form that surface. This provides a direct physical link between the breaking strength of the microscopic bonds and the toughness of the bulk material.

However, this simple and powerful bond-based model has a fascinating limitation, an "Achilles' heel" that points the way to more advanced theories. When a material is stretched in one direction, it usually contracts in the perpendicular directions. The ratio of this contraction is the Poisson's ratio, $\nu$. Cork has $\nu \approx 0$, while rubber has $\nu \approx 0.5$. These are distinct material properties.

Yet, the simple bond-based model, due to its central-force nature, predicts a fixed Poisson's ratio of $\nu=1/4$ for any isotropic material in 3D (and $\nu=1/3$ in 2D). The simple bonds aren't "smart" enough to provide independent resistance to volume change and shape change.

This limitation was a key motivator for the development of ​​state-based peridynamics​​. In these more advanced models, the force in a bond can depend on the deformation of the entire neighborhood, not just its own stretch. This allows bonds to "talk" to each other and generate non-central forces, breaking the constraint on Poisson's ratio and enabling the modeling of any material type.

Even so, the bond-based model, with its elegant simplicity and profound physical insights, remains a cornerstone of the theory and a powerful tool for understanding the fundamental mechanisms of material failure. Through it, we see how the complex and often violent world of fracture can emerge from the simple, local rules of microscopic bonds.

Now that we have acquainted ourselves with the fundamental rules of our peridynamic world—the neighborhood watch of the horizon, and the non-local chatter between material points—you might be asking a fair question: "What is it all good for?" It is a wonderful question. The true test of any physical theory is not just its internal elegance, but what it can tell us about the world we live in. It turns out that this seemingly simple shift in perspective, from the local "managerial memos" of classical derivatives to the "neighborly conversations" of peridynamic integrals, unlocks a new and profound understanding of phenomena that have long been a headache for scientists and engineers. We have found a new language, one that speaks fluently about the beautiful and violent drama of things breaking.

The star application, the very reason peridynamics was born, is the problem of fracture. In classical mechanics, a crack is a nightmare. Because the theory is built on derivatives—rates of change at a single point—a sharp crack, where the material is discontinuous, creates mathematical absurdities like infinite stresses. To handle it, physicists had to invent clever but cumbersome patches. It’s like trying to describe traffic flow in a city where a giant canyon has suddenly appeared; the original rules of driving are no longer sufficient. You need special rules for the canyon edge. Classical fracture mechanics is full of these special rules: rules for when a crack should grow, in what direction, and whether it should branch. The crack's path must be painstakingly tracked and dictated to the equations.

Peridynamics turns this entire approach on its head. Instead of telling the crack where to go, we just let the simulation run, and the cracks appear all by themselves. They are emergent properties of the system. Imagine the material not as a continuous block, but as a vast, interconnected social network of points, with bonds as the friendships between them. When the material is stretched, some friendships become strained. If a bond is stretched too far—beyond a critical breaking point, $s_c$—it snaps, and that connection is lost forever. The force that bond once carried must now be redistributed among the remaining neighbors. This might, in turn, overload other bonds, causing a cascade of failures.

A crack, then, is simply a region where a significant number of these friendships have been severed. If enough bonds across a surface are broken, the network splits into two separate communities that can no longer talk to each other. A physical gap can open up, and a crack is born. There is no need for infinite stresses or special crack-tip equations. The same integral law of motion that governs the intact material also governs its failure. This is an immense conceptual leap. It allows a computer simulation, for the first time, to predict the wildly complex and beautiful patterns of fracture from first principles—the spiderweb cracks on a car's windshield, the branching path of lightning in a block of acrylic, or the catastrophic shattering of a ceramic plate.

This non-local view does more than just change how things break; it changes how they vibrate. If you pluck a guitar string, classical theory tells us that waves of all frequencies, from the low fundamental tone to the high harmonics, travel along the string at the same speed. But is that entirely true? What if we look really, really closely?

In the peridynamic world, the answer is no. Waves of different wavelengths travel at different speeds, a phenomenon known as dispersion. A calculation for waves in a peridynamic bar shows that the wave frequency $\omega$ is not simply proportional to the wavenumber $k$ (which is inversely related to wavelength). Instead, the relationship is more complex, depending on the size of the horizon $\delta$. Why? Because very short, choppy waves "feel" the discrete, non-local nature of the interactions more acutely than long, lazy waves do. Their propagation is influenced by the details of the neighborhood, not just an averaged-out bulk property.

This isn't just a mathematical quirk! At the atomic scale, materials are non-local, and their vibrational properties (their phonons) do exhibit dispersion. Peridynamics, while still a continuum theory, captures a shadow of this microscopic reality. It acts as a beautiful bridge, linking the familiar world of continuum mechanics to the granular world of atoms without needing to model every single one.

This connection between waves and fracture becomes powerfully clear when we consider impacts. Imagine a brittle coating being struck. A compression wave travels into it, reflects from a surface as a tension wave, and if that tension is strong enough, the material fractures. How fast must the impact be to cause failure? By marrying the peridynamic view of fracture with wave propagation, we can derive a stunningly simple and powerful result. The critical impact velocity, $v_{cr}$, required to initiate fracture can be related to the material's fracture toughness (energy per unit area), $G_c$, its density, $\rho$, and the peridynamic horizon size, $\delta$:

This elegant formula tells us something profound: to make a material more impact-resistant, you can either increase its intrinsic toughness ($G_c$) or... make it lighter ($\rho$)!. It is a perfect demonstration of how a deep theory can crystallize into a beautifully simple and practical piece of engineering wisdom.

The real beauty of a fundamental framework is its extensibility. Once you have the mechanical rules, you can start coupling them with other parts of physics, creating a richer, more powerful simulation. Peridynamics excels at this.

Consider the coupling of heat and mechanics. When an object is heated, its atoms vibrate more vigorously, and it tends to expand. In the peridynamic picture, this means every bond has a "thermal stretch" it wants to have, proportional to the temperature change. If the material is constrained and cannot expand freely, this desired stretch is converted into a real mechanical force. This simple idea allows us to model thermal stresses. A famous and dramatic example is ​​thermal shock​​: drop a hot glass into cold water, and it shatters. Why? The outer surface cools and contracts rapidly, while the interior remains hot and expanded. This mismatch creates immense internal tension. A peridynamic simulation can couple the equation of heat flow with the peridynamic laws of motion to predict the precise moment and location of the first microscopic crack, a feat of great importance for designing everything from engine components to cookware.

The framework is not limited to solids, either. What about materials with "memory," like putty, polymers, or even biological tissue? These are called viscoelastic materials. Their response depends not just on how much you are stretching them now, but on their entire history of being stretched. The peridynamic model accommodates this with remarkable grace. Instead of the force in a bond being a simple function of the current stretch, it becomes a weighted average over the entire history of the stretch, expressed as a hereditary integral. This "fading memory" allows us to model the slow sag of a plastic component over years, the resilient bounce of a rubber ball, or the complex biomechanics of a knee cartilage under load. It opens the door for peridynamics to contribute to soft matter physics, polymer science, and biomechanics.

Of course, using this theory in practice comes with its own set of fascinating challenges, which reveal even more about the nature of a non-local world. For instance, what happens to a material point right at the surface of an object? A point in the bulk is surrounded by neighbors in all directions, but a point on the surface is "lonely"—it has neighbors on one side only. This lack of a complete neighborhood means its response to forces is different from that of its interior cousins. This "surface effect" causes the stress near a boundary to behave differently than classical theory would predict, even in a simple tension test. This is not a bug; it is a feature! It's a real physical effect that becomes important in thin films, coatings, and nanomaterials.

Even something as simple as "holding one end of an object fixed" becomes a richer question. In a non-local theory, you can't just nail down a single point. To be consistent, the boundary condition must also be non-local. The most effective way to do this is to imagine a fictitious layer of "ghost" material just outside the real boundary, and prescribe the motion of this entire ghost layer. The real material points near the boundary can then interact with this ghost layer, feeling the constraint over their entire horizon. It is as if you are holding the object with a non-local "ghost hand," a beautiful reminder that in a peridynamic universe, everything is connected.

From the dramatic shattering of a brittle solid to the subtle whispers of nanoscale waves, from thermal shock to the memory of soft tissues, the peridynamic framework provides a unified and intuitive lens. It shows us that by replacing a local point of view with a more communal one, we can uncover a deeper description of the material world, one that embraces discontinuities and connections with equal elegance.