Drucker's Postulate

In engineering and materials science, the predictability of a material's behavior under load is not just a desirable quality—it is a fundamental requirement for safety and design. While we intuitively understand that stable materials should resist deformation and dissipate energy during permanent changes, formalizing this concept into a rigorous mathematical framework is a profound challenge. This challenge lies at the heart of plasticity theory: how do we create reliable rules that govern a material's response once it passes its elastic limit?

This article addresses this knowledge gap by exploring Drucker's postulate, a cornerstone theory of material stability. It provides the "rules of the game" for stable plastic deformation. In the chapters that follow, we will unpack this powerful concept. First, we will examine the core "Principles and Mechanisms" of the postulate, revealing how a simple inequality about work leads to the elegant geometric requirement of yield surface convexity and predicts the onset of material failure. Following that, in "Applications and Interdisciplinary Connections," we will see how this principle is applied in the real world to ensure uniqueness in engineering calculations, diagnose instabilities in geophysics, and provide a critical sanity check for modern computational simulations.

Imagine you are building something—a bridge, an airplane, a skyscraper. You choose your materials, steel, aluminum, concrete, with care. You expect them to be reliable. When you push on a steel beam, it pushes back. If you bend it so far that it stays bent, you understand that you’ve had to do work to make that happen. The beam didn't spontaneously bend itself and hand you back energy. This fundamental intuition—that stable matter resists, and that permanent deformation costs energy—is the soul of what we are about to explore. It seems obvious, but transforming this simple physical idea into a rigorous, predictive mathematical theory is a journey of profound discovery. This journey was charted in large part by the engineer Daniel C. Drucker, and his "rules of the game" for material stability are now known as ​​Drucker's postulates​​.

At its heart, plastic deformation—the permanent, irreversible change in a material's shape—is a messy business. Atoms slide past one another, dislocations tangle, and heat is generated. It is a ​​dissipative​​ process, much like friction. The second law of thermodynamics already tells us that the total energy dissipated must be non-negative. But Drucker proposed a stricter, more powerful condition focused squarely on the mechanical work.

He suggested we consider the work done by the stress on the plastic part of the deformation alone. His first, and most fundamental, postulate can be stated simply: for any process that causes plastic flow, the incremental work done by the stress $\boldsymbol{\sigma}$ on the resulting increment of plastic strain $\delta\boldsymbol{\varepsilon}^{p}$ must be non-negative.

This little inequality is more demanding than the second law of thermodynamics. The second law only requires the total dissipation (which includes energy stored in microscopic hardening mechanisms) to be non-negative. Drucker's postulate insists that the plastic work term itself cannot be negative. A material cannot use its internal stored energy to deform plastically against the direction of the applied stress.

But Drucker didn't stop there. He also considered how a material changes while it is yielding. Suppose a material is right on the cusp of yielding. We give it a tiny nudge of extra stress, $d\boldsymbol{\sigma}$, which causes a tiny bit more plastic strain, $d\boldsymbol{\varepsilon}^{p}$. Drucker’s second postulate, often called the ​​hardening postulate​​ or the condition of ​​stability in the small​​, requires that the work done by this stress increment on the resulting plastic strain increment must also be non-negative.

This ensures the material provides stable resistance to further deformation. A positive value means the material is ​​hardening​​—it's getting stronger and requires more stress to deform further. The case where $d\boldsymbol{\sigma} : d\boldsymbol{\varepsilon}^{p} = 0$ corresponds to ​​perfect plasticity​​, where the material flows at a constant stress. A violation, where this work becomes negative, describes ​​softening​​—a dangerous state where the material gets weaker as it deforms, a prelude to failure.

These abstract inequalities might seem like dry mathematical bookkeeping. But here is where the magic happens. When combined with another physically reasonable assumption, they paint a beautiful and powerful geometric picture. The assumption is that of an ​​associated flow rule​​, which states that the "direction" of plastic strain in the abstract, multi-dimensional space of strain is perpendicular (or ​​normal​​) to the yield surface in the corresponding space of stress.

What's a ​​yield surface​​? Imagine a graph where the axes represent different components of stress (tension in one direction, shear in another, etc.). The yield surface is a boundary in this space. For any combination of stresses that falls inside this boundary, the material behaves elastically—it springs back if you unload it. If the stress combination reaches the boundary, the material can begin to deform plastically.

Now, let's put the pieces together. Drucker's postulate, in a slightly more general form, states that for a material at a stress state $\boldsymbol{\sigma}$ on the yield surface, and for any other stress state $\boldsymbol{\sigma}^*$ inside the surface, the inequality $(\boldsymbol{\sigma} - \boldsymbol{\sigma}^*) : \dot{\boldsymbol{\varepsilon}}^p \ge 0$ must hold. If we add in the associated flow rule, which says the plastic strain rate $\dot{\boldsymbol{\varepsilon}}^p$ is normal to the surface, this inequality takes on a stunning geometric meaning: the entire elastic domain must lie on one side of the plane that is tangent to the yield surface at any given point.

This is the very definition of a ​​convex set​​.

Think of an egg. It's convex. No matter where you touch it, the rest of the egg is "behind" your finger. Now think of a kidney bean. It's non-convex. It has a dent. You can touch a point on one side of the dent, and part of the bean will be "in front" of your finger. Drucker's postulate, for an associated material, forbids the yield surface from having any such dents. It must be shaped like the egg, not the kidney bean. This profound equivalence—that stability implies convexity, and convexity implies stability—is a cornerstone of modern mechanics. It is a true "if and only if" relationship.

This geometric property is not just an aesthetic curiosity; it is of immense practical importance. Engineers and scientists use computer simulations (like the Finite Element Method) to predict how structures will behave under load. These simulations solve the equations of mechanics step-by-step. In a plastic material, the algorithm must ensure that the calculated stress never goes outside the yield surface. If a "trial" stress step ends up outside, it must be "returned" to the surface.

If the yield surface is convex, this "return mapping" problem has a single, unique solution. There is always one closest point on the surface to return to. This ensures that the simulation is stable and gives a predictable, unique answer.

If the surface were non-convex, a trial stress might have multiple, equally valid points to return to. Which one should the computer choose? The answer becomes ambiguous, and the simulation can become unstable or give nonsensical results.

More frighteningly, this mathematical ambiguity reflects a potential physical reality. A material with a non-convex yield surface is inherently unstable. In some models for materials like soils or concrete, certain parameter choices can produce non-convex yield surfaces. For such a material, when combined with softening, a catastrophic failure mode known as ​​snap-back​​ can occur. Under a controlled increase in displacement, the material might suddenly lose its ability to carry load, causing the stress to plummet and potentially requiring the displacement to be reversed to maintain equilibrium. This is the material-level signature of a sudden, brittle-like collapse. The abstract requirement of convexity is directly linked to preventing catastrophic failure.

Let's zoom in on the yield surface, this boundary between the elastic and plastic worlds. Imagine our stress state is sitting right on the surface. We now apply an infinitesimal stress increment, $d\boldsymbol{\sigma}$. What happens next? The answer depends entirely on the direction of our push relative to the surface geometry. Let $\boldsymbol{n}$ be the outward normal vector to the surface.

• ​​Plastic Loading:​​ If our stress increment $d\boldsymbol{\sigma}$ points "outward" from the surface (i.e., $d\boldsymbol{\sigma} : \boldsymbol{n} > 0$), we push the material into the plastic domain. The material yields, and plastic deformation occurs.

• ​​Elastic Unloading:​​ If our stress increment $d\boldsymbol{\sigma}$ points "inward" from the surface (i.e., $d\boldsymbol{\sigma} : \boldsymbol{n} < 0$), the stress state moves back into the elastic region. The response is purely elastic, and no further plastic deformation occurs.

• ​​Neutral Loading:​​ What if we push exactly tangentially to the surface (i.e., $d\boldsymbol{\sigma} : \boldsymbol{n} = 0$)? In this delicate boundary case, the response is also purely elastic. No plastic flow happens. The stress state skims along the edge of the yield surface without crossing it.

This elegant trichotomy—loading, unloading, and neutral loading—is a direct consequence of the stability framework and its geometric interpretation.

The beautiful synthesis of stability and convexity hinges on the associated flow rule. What happens if a material violates this rule? Some real-world materials, like granular soils, clays, and concrete, are better described by ​​non-associated​​ models, where the direction of plastic flow is governed by a [plastic potential](/sciencepedia/feynman/keyword/plastic_potential) function, $g$, that is different from the yield function, $f$.

In this case, the compass is broken. The direction of plastic flow is no longer tied to the normal of the convex yield surface. The elegant proofs of stability fail. The incremental stiffness of the material is no longer described by a symmetric matrix, which can cause computational difficulties and can be a source of material instability, even if the material is hardening.

An even more dramatic breakdown occurs when a material begins to soften, violating Drucker's second postulate ($d\boldsymbol{\sigma} : d\boldsymbol{\varepsilon}^{p} < 0$). This is not just a theoretical concern; it is the gateway to ​​strain localization​​. Instead of deforming uniformly, the material decides to concentrate all its deformation into a very narrow band, like the crease that forms in a soda can just before it buckles. This shear band is the precursor to a crack or a fault.

The connection is incredibly deep. A mathematical property of the governing equations called ​​strong ellipticity​​ is required for the solutions to be well-behaved. The moment this property is lost, localization can occur. For an associated material, it turns out that the loss of strong ellipticity for a specific, band-like deformation mode is exactly equivalent to the violation of Drucker's stability postulate for that same mode. The stability condition tells us precisely when and how the material will decide to tear itself apart.

We began with a simple idea of stability and discovered its profound connection to geometry and failure. Let us end with one final, elegant consequence that reveals a hidden symmetry in the nature of stable materials.

Starting from the simple premise that the incremental work must be non-negative, $d\boldsymbol{\sigma} : d\boldsymbol{\varepsilon} \ge 0$, one can derive a surprising reciprocity relation. If we consider two different, independent incremental loads from the same state, $(\delta\boldsymbol{\sigma}_1, \delta\boldsymbol{\varepsilon}_1)$ and $(\delta\boldsymbol{\sigma}_2, \delta\boldsymbol{\varepsilon}_2)$, the following must hold:

This means the work done by the first stress increment on the second strain increment, plus the work done by the second stress increment on the first strain increment, must be non-negative. This is not at all obvious from the outset! It is a mathematical echo of the underlying stability, a subtle symmetry woven into the fabric of the material response.

From a simple physical intuition, Drucker's postulates construct a powerful and elegant framework. They give stability a shape (convexity), provide the foundation for reliable computation (uniqueness), classify the material's response at the threshold of change (loading/unloading), and predict the dramatic onset of failure (localization). This is the hallmark of a great physical principle: simplicity at its core, and richness in its consequences.

Now that we have explored the principles and mechanisms behind Drucker's stability postulate, you might be wondering, "What is it all for?" It is a fair question. A physical principle, no matter how elegant, earns its keep by its power to explain the world we see, to predict the behavior of things we build, and to guide us away from disastrous failures. Drucker’s postulate is a champion in this regard. It is not merely a piece of abstract mathematics; it is a sharp lens through which we can understand the stability of everything from a steel beam to a mountainside. It is our reliable guide in the complex world of material deformation and failure.

Let us embark on a journey to see this postulate at work, to witness how this simple rule of "good behavior" for materials brings clarity to engineering design, geophysics, and even the digital world of computer simulation.

First, we must draw a crucial distinction. Our postulate is a law about the stuff an object is made of, not about the object's overall shape or configuration. A material can be perfectly stable, obeying Drucker's postulate to the letter, yet a structure built from it can collapse in an instant.

Consider the classic example of a slender column under compression. Imagine a long, thin ruler pressed from its ends. We know it will suddenly bow outwards and buckle. Let's say the ruler is made of steel, a material that is impeccably stable. The buckling happens at a stress far, far below the stress needed to permanently deform the steel itself. In this elastic regime, any deformation is perfectly reversible, and the net work done over any closed cycle of loading and unloading is zero—a perfect, if trivial, satisfaction of Drucker's postulate.

So why does it buckle? The instability here is structural, or geometric. It is a contest between the material's innate stiffness, which tries to keep the column straight, and the geometry of the load, which encourages it to bend. The total stability of the structure depends on the sum of a "material stiffness" and a "geometric stiffness." Under compression, this geometric stiffness becomes negative, creating a destabilizing effect. When the compressive load is large enough, the overall stiffness drops to zero, and the structure buckles. This has nothing to do with the material becoming unstable; it is about the system as a whole losing its stability. Material stability, therefore, does not guarantee structural stability. Drucker's postulate tells us about the integrity of the material points themselves, and it is on this crucial topic that we shall focus.

With that distinction made clear, let's look at problems where the material's own behavior is paramount. Consider a thick-walled cylinder under immense internal pressure—a boiler, a pipe, or a pressure vessel in a chemical plant. As the pressure rises, the material near the inner wall begins to yield and flow plastically. As engineers, we demand a predictable response. We need to know that for a given increase in pressure, there is one, and only one, resulting state of stress and strain. Without this uniqueness, reliable design is impossible.

Here, Drucker's postulate shines. For a standard ductile metal, whose behavior is captured by an associative plasticity model (where plastic flow is normal to the convex yield surface), the postulate holds. This mathematical "good behavior" guarantees that the solution to the problem is unique. The cylinder expands in a stable, predictable manner.

But what if we use a material that violates the postulate? This is common in the world of geomechanics—the study of soil, rock, and sand. Many geological materials are "frictional" and "dilatant"; think of how a bag of sand seems to expand when you shear it. In these materials, the direction of plastic flow is often not normal to the yield surface—a property called non-associativity. This directly violates the conditions of Drucker's postulate.

And the consequence? The uniqueness of the solution is lost. The material's response becomes unstable. Instead of deforming smoothly, the deformation can suddenly concentrate into a narrow band of intense shear. In the case of the pressurized cylinder made of such a material, we might see a plastic annulus suddenly form and deform wildly, a prelude to catastrophic failure. For a slope of soil or rock, this is the very mechanism of a landslide: the formation of a shear band along which a massive amount of earth suddenly slides. Drucker's postulate, or rather its violation, gives us a profound insight into the fundamental trigger for these instabilities.

The very onset of these failure bands is a fascinating topic. The mathematical equations governing the material's equilibrium change their character from elliptic (which favor smooth solutions) to hyperbolic (which permit sharp discontinuities). This shift, known as a loss of ellipticity, is the mathematical birth of a shear band. It is the violation of Drucker's stability conditions in non-associative materials that can trigger this loss of ellipticity, even when the material is still hardening. So, the abstract postulate is directly linked to the physical patterns of failure we observe in nature.

In the modern world, much of engineering and science is done inside a computer. We build digital models of materials and structures to test them to their limits. But how can we trust these simulations, especially when we are modeling complex processes like fracture? Once again, Drucker's postulate serves as our indispensable guide.

Imagine we are simulating a piece of ductile metal being torn apart. As microscopic voids within the material grow and link up, the material inherently softens—its ability to carry stress decreases as it is deformed further. This softening is a direct violation of Drucker's stability postulate; the incremental plastic work $d\boldsymbol{\sigma} : d\boldsymbol{\varepsilon}^{p}$ becomes negative.

When we put a simple, local model of this softening material into a standard finite element program, a strange sickness appears. The simulation predicts that the failure will localize into a band, but the width of this band shrinks as we refine our simulation mesh. The computed energy required to break the component spuriously drops to zero as the mesh gets finer. The result is meaningless—it depends on the arbitrary choice of our mesh, not on the physics. This is called "pathological mesh dependence."

Drucker's postulate helps us diagnose the disease. The pathology arises precisely because the material model is unstable, and it lacks an intrinsic length scale. The cure is to build more sophisticated models (nonlocal or gradient-enhanced models) that restore a physical length scale to the problem, but the diagnosis comes straight from stability theory.

More than just a diagnostic tool for known problems, the postulate is used as a real-time "sanity check" within simulations. Sophisticated computer codes can be programmed to monitor the state of the material at every point and at every step of the calculation. They compute the incremental plastic work, just as we have discussed. If this value becomes significantly negative (beyond a small numerical tolerance), it raises a red flag. It tells the analyst that the material model is entering an unstable regime, and the results from that point on must be interpreted with extreme caution. It's like a computational doctor checking the vital signs of the digital material to ensure it is behaving physically.

The power of a great principle is often revealed in the unexpected connections it forges. Drucker's postulate beautifully connects the local stability of a material point to the long-term health of an entire structure, and it grounds these theoretical ideas in tangible laboratory experiments.

Consider a bridge under the cyclic loading of traffic, or an aircraft fuselage under repeated pressurization cycles. The loads are not high enough to cause immediate collapse, but they may cause tiny amounts of plastic deformation in localized regions with each cycle. A critical question for a designer is: will the structure eventually "shake down" and adapt, responding purely elastically to subsequent load cycles? Or will plastic strain accumulate with every cycle (a phenomenon called "ratcheting"), leading to eventual failure?

The famous shakedown theorems, which provide the answer to this question, are built directly upon the foundation of Drucker's postulate. Specifically, the postulate gives rise to the "principle of maximum plastic dissipation." This principle states that for a given rate of plastic flow, the actual stress state in a stable material organizes itself to maximize the rate of energy dissipation. This maximization principle is a key ingredient in proving the theorems that tell us whether a structure is safe for long-term cyclic use. Thus, the simple rule of local stability is tied to the fate of a complex structure over millions of cycles. The same principles also guide the formulation of advanced material models that capture complex behaviors like the evolution of the yield surface shape during cyclic loading, always ensuring the model remains physically stable.

Finally, lest we forget that physics is an experimental science, how do we actually check if a new alloy or polymer satisfies our postulate? We must ask the material itself. The theory tells us exactly what question to ask and how to interpret the answer. We take a sample of the material, place it in a testing machine, and subject it to a very slow, controlled cycle of straining and un-straining, well into its plastic range. We meticulously measure the stress and strain throughout the cycle. Then, we compute the net work done on the material, which is the area enclosed by the stress-strain loop, $\oint \sigma d\varepsilon$.

Drucker's postulate for a stable material demands that this net work must be non-negative, $W_{\text{cycle}} \ge 0$. A positive value means the material dissipated energy as heat, which is normal for plastic deformation. But if we were ever to measure a negative net work—meaning the material gave back more energy than we put in over a full cycle—we would have found an unstable material. We would have caught it violating a fundamental rule of thermodynamic stability. This experimental protocol provides the ultimate, tangible grounding for our entire discussion.

So, we see that Drucker's postulate is far from a dry, academic curiosity. It is a profound and practical principle of stability. It draws the line between predictable, reliable materials and those prone to sudden, catastrophic failure. Its elegant simplicity provides a unifying thread that runs through structural analysis, geophysics, computational mechanics, and experimental science, allowing us to better understand, predict, and design the physical world around us.