4. Geometric nonlinearity

Linear eigenvalue buckling analysis predicts the critical load at which a thin-walled structure becomes unstable, but it says nothing about what happens after that load is exceeded. In reality, thin plates and shells do not simply collapse at the bifurcation point — they continue to carry load, often substantially more, through a geometrically nonlinear postbuckling path. Capturing this behavior requires abandoning the small-displacement assumption and adopting a formulation in which the strain–displacement relations retain second-order terms. This chapter explains why geometric nonlinearity is indispensable for postbuckling assessment, how membrane stretching governs the postcritical stiffness, and how to set up and interpret a nonlinear buckling analysis in practice.

Why Small-Deflection Theory Breaks Down

In linear plate theory, the strain components are approximated as first-order functions of the out-of-plane displacement w: ε_x = −z ∂²w/∂x², and so on. This approximation assumes that slopes ∂w/∂x remain small and that the mid-surface does not stretch. For a thin plate loaded in compression, the buckling mode involves out-of-plane deflections that grow rapidly once the critical load is approached. Even at moderate deflection amplitudes — on the order of the plate thickness — the slope terms are no longer negligible, and the mid-surface begins to stretch. The linear kinematic assumption therefore loses validity precisely in the regime that matters most for postbuckling assessment.

The consequence is not a minor numerical error. Linear theory predicts that the plate has zero stiffness in the postbuckling range, implying immediate collapse. Experiments and nonlinear analysis both show the opposite: a flat plate under in-plane compression develops a stable postbuckling path with increasing load capacity. The missing ingredient is membrane stretching, which is a purely geometric effect captured only when the strain–displacement relations include the quadratic term (∂w/∂x)²/2.

Von Kármán Kinematics and Membrane Stretching

The classical framework for moderately large deflections of plates is the von Kármán plate theory. Its essential feature is the augmented mid-surface strain:

ε_x = ∂u/∂x + ½ (∂w/∂x)²
ε_y = ∂v/∂y + ½ (∂w/∂y)²
γ_xy = ∂u/∂y + ∂v/∂x + (∂w/∂x)(∂w/∂y)

Here u, v are in-plane displacements and w is the transverse deflection. The quadratic terms are the membrane stretching contributions. When a plate buckles into a waveform with out-of-plane deflection w(x,y), the in-plane fibers must lengthen to accommodate the curvature of the deflected surface. This stretching generates membrane stresses that resist further deflection, providing the postbuckling stiffness that linear theory lacks.

The physical picture is straightforward: imagine a flat sheet clamped along its edges and pushed inward. Initially it carries load by in-plane compression. At the critical load it snaps into a wave pattern. As the wave amplitude grows, the sheet's mid-surface stretches, and the membrane tension that develops acts like a set of springs that stiffen the response. The structure does not lose all load-carrying capacity; it transitions from a compression-dominated to a stretching-dominated regime.

Postbuckling Paths: Stable Plates vs. Sensitive Shells

The role of membrane stretching explains a fundamental contrast in thin-walled stability:

• Flat plates: The postbuckling path is stable and rising. After the critical load, the plate continues to carry increasing load, though at reduced stiffness. The membrane stretching provides positive stiffness. Imperfection sensitivity is mild — a plate with a small initial imperfection reaches essentially the same postbuckling capacity as a perfect one.
• Cylindrical shells under axial compression: The postbuckling path is unstable and descending. The shell cannot develop sufficient membrane stretching to stabilize the buckling mode; instead, the deformation localizes into a dimple that grows without additional load. Imperfection sensitivity is severe — a geometric imperfection of one shell thickness can reduce the collapse load by 30–50% or more.

This difference is not a matter of degree but of the topology of the equilibrium path. For plates, the nonlinear equilibrium equations admit a stable postbuckling branch that the structure naturally follows. For axially compressed cylinders, the primary path is unstable beyond the limit point, and the structure must jump to a remote stable equilibrium — often with catastrophic consequences.

Setting Up a Nonlinear Buckling Analysis

A geometrically nonlinear buckling assessment in a finite element code (e.g., Abaqus, ANSYS, Nastran SOL 400) follows a different logic from linear eigenvalue analysis. The goal is not to find a single critical load but to trace the load–deflection equilibrium path, including any limit points or bifurcations.

The standard procedure is:

1. Define the model with nonlinear geometry enabled. Activate the large-deflection option (NLGEOM in Abaqus, NLGEOM,ON in ANSYS). This switches the strain measure to include the quadratic terms.
2. Apply a small initial imperfection. Seed the mesh with a scaled buckling mode shape from a prior eigenvalue analysis, or with a measured geometric deviation. The imperfection amplitude is typically 1–10% of the wall thickness for plates, and 0.1–1.0 times the thickness for shells, depending on the manufacturing tolerance.
3. Use displacement control or arc-length (Riks) method. Load control will fail at a limit point. The Riks method follows the equilibrium path through limit points and snap-backs by treating the load factor as an unknown.
4. Monitor the load–deflection curve. Plot the applied load factor against a representative deflection (e.g., center-point out-of-plane displacement). The curve reveals whether the postbuckling path is stable (rising), neutral (flat), or unstable (descending).
5. Check for bifurcation points. On a perfect structure, the primary path may bifurcate into a secondary branch. Introducing a small imperfection converts the bifurcation into a smooth limit point, which is easier to trace numerically.

Interpreting the Load–Deflection Response

The output of a nonlinear buckling analysis is a curve, not a number. For a flat plate under uniaxial compression, a typical response shows:

• An initially linear rise in load with negligible out-of-plane deflection.
• A knee where the deflection begins to grow rapidly — the nonlinear critical load, slightly below the eigenvalue prediction.
• A postbuckling branch that continues to rise, but with a slope (postbuckling stiffness) that is a fraction of the prebuckling stiffness. For a simply supported plate, the postbuckling stiffness is approximately 40–50% of the prebuckling value.
• At higher loads, secondary buckling modes or material yielding may intervene, flattening or dropping the curve.

For a cylindrical shell, the same plot typically shows a sharp peak followed by a steep drop — the limit point — and then a lower stable branch. The ratio of the limit load to the classical eigenvalue load is the knockdown factor, which is the central quantity in shell design codes.

Practical Implications for Assessment

Three conclusions follow directly from the role of membrane stretching:

First, linear eigenvalue analysis is a screening tool, not a capacity prediction. It gives the load at which the perfect structure first becomes unstable, but it cannot tell you whether the structure will survive beyond that load. For flat plates, the true collapse load may be several times the eigenvalue load; for shells, it may be a fraction of it. The eigenvalue result must always be corrected by a nonlinear analysis or a code-specified knockdown factor.

Second, the postbuckling reserve depends on boundary conditions and aspect ratio. A simply supported plate develops membrane stretching because its edges are restrained against in-plane displacement. If the edges are free to pull in, the membrane effect is lost and the postbuckling path may become flat or unstable. Similarly, a long plate (aspect ratio > 3) behaves differently from a square plate because the buckling mode involves multiple half-waves, and the membrane stretching distribution changes.

Third, the transition from stable to unstable postbuckling is governed by the ratio of bending to membrane stiffness. This ratio scales with (t/R) for shells and (t/b)² for plates, where t is thickness, R is radius, and b is plate width. Thin shells with small t/R are inherently more imperfection-sensitive because their membrane stiffness is low relative to their bending stiffness, and they cannot develop enough stretching to stabilize the postbuckling path.

In summary, geometric nonlinearity is not an optional refinement for thin-walled buckling assessment — it is the mechanism that determines whether a structure has postbuckling reserve or collapses catastrophically. Membrane stretching, captured through the quadratic terms in the strain–displacement relations, provides the stabilizing stiffness for plates and the lack thereof for shells. A properly set up nonlinear analysis with an appropriate imperfection and a path-following solver is the only way to quantify this behavior and to establish a reliable collapse load for design.