Plates, Shells and Spatial Structural Systems

Introduction and Context

Plates, shells and spatial structural systems are three-dimensional structural forms in which load transfer is governed not only by flexure but also by membrane action, in-plane shear, torsion and geometric curvature. They are therefore fundamentally different from idealized one-dimensional beams and two-dimensional plane frames. Their efficient use of material makes them important in roofs, bridge decks, tanks, silos, aircraft structures, stadiums, offshore platforms, folded roofs and high-rise transfer systems.

A plate is a relatively thin, flat structural element loaded primarily transverse to its middle surface. A shell is a curved, thin structural element in which loads can often be carried predominantly through membrane forces. A spatial structural system is an assemblage whose members and joints participate in three-dimensional load transfer, including space frames, grids, tensegrity systems and hybrid structures.

The advanced analyst must understand both the mechanics and the modelling decisions. The choice of element type, boundary conditions, mesh, material law and load idealization can influence results as strongly as the numerical solver itself.

Learning Objectives

After studying this chapter, the reader should be able to:

1. distinguish plate, shell and spatial structural behaviour;
2. formulate the principal governing equations for thin and moderately thick plates;
3. explain membrane, bending and bending–membrane coupling in shells;
4. analyse common shell geometries and identify their critical actions;
5. formulate stiffness-based models for spatial structural systems;
6. select suitable analytical, numerical and experimental methods;
7. interpret instability, singularity, boundary-layer and nonlinear effects;
8. apply these concepts to a realistic engineering design scenario.

1. Structural Idealizations and Terminology

The middle surface is the surface equidistant from the two faces of a plate or shell. The thickness, denoted by h, is measured normal to this surface. The characteristic in-plane dimension is denoted by a or L. Thin-structure theory generally requires h/L to be small; the precise limit depends on geometry, loading and required accuracy.

Important generalized stress resultants are:

A plate may be isotropic, orthotropic or anisotropic. An isotropic plate has identical properties in all directions; an orthotropic plate has different properties along two mutually perpendicular material axes, as in a reinforced concrete slab or a laminated timber panel. A shell may be developable, such as a cylindrical surface, or double-curved, such as a sphere or hyperbolic paraboloid.

The principal structural actions are:

• membrane action, consisting of in-plane tension and compression;
• plate bending, caused by transverse loading and curvature of the deflected surface;
• twisting, represented by a twisting moment or in-plane shear variation;
• transverse shear, significant in thick plates and near supports;
• geometric stiffness, arising from existing membrane forces;
• warping and distortion, important in space frames and open sections.

2. Classical Thin-Plate Theory

2.1 Kinematic assumptions

Classical Kirchhoff–Love plate theory assumes:

1. the plate is thin;
2. the middle surface is initially flat;
3. normals to the middle surface remain straight and normal after deformation;
4. transverse shear deformation is negligible;
5. strains and rotations are sufficiently small for linear analysis;
6. the material is commonly assumed homogeneous, elastic and isotropic.

Let [[M.x]] and [[M.y]] be in-plane coordinates and [[M.z]] the coordinate through the thickness. The displacement field is approximated by

Here [[M.u_0]] and [[M.v_0]] are middle-surface in-plane displacements and [[M.w]] is transverse displacement. The assumption that transverse normals remain normal leads to zero transverse shear strains in the idealized theory.

For small deflection bending, the curvatures are

The factor 2 in [[M.κ_{xy}]] is a consequence of engineering shear-strain convention. Sign conventions must be stated because alternative structural-analysis texts use opposite signs.

2.2 Constitutive relations

For an isotropic plate under pure bending, the flexural rigidity is

where [[M.E]] is Young’s modulus, [[M.ν]] is Poisson’s ratio and [[M.h]] is thickness. The units of [[M.D]] are force × length. The moment–curvature equations are

These equations apply to a linearly elastic, isotropic, thin plate with small deflection and no significant transverse shear deformation.

2.3 Governing differential equation

For a plate subjected to transverse load [[M.q(x,y)]], equilibrium gives the biharmonic plate equation:

where

The load [[M.q]] has units force per area. The left-hand side has the same units because [[M.D]] has units force × length and four spatial derivatives of [[M.w]] have units 1/length³.

Boundary conditions are essential. A simply supported edge commonly has [[M.w=0]] and zero bending moment normal to the edge. A clamped edge has [[M.w=0]] and zero rotation normal to the edge. A free edge has zero bending moment, zero effective twisting condition and zero transverse shear resultant, subject to the adopted sign convention.

2.4 Navier solution for a simply supported rectangular plate

For a rectangular plate of dimensions [[M.a]] by [[M.b]], simply supported on all four edges, a double-sine expansion is used:

For a load expanded as

coefficient equilibrium gives

This solution is particularly useful for verification of finite-element models and for understanding how aspect ratio and load distribution influence response.

3. Thick Plates and Mindlin–Reissner Theory

When thickness is not negligible, transverse shear deformation and rotary inertia may matter. Mindlin–Reissner theory allows the normal to rotate independently of the deflected middle surface. The generalized unknowns are typically [[M.w]], [[M.θ_x]] and [[M.θ_y]]. The shear strains are approximately

depending on the sign convention for rotations. The transverse shear resultants are

where [[M.G=E/[2(1+ν)]]] is shear modulus and [[M.k_s]] is a shear-correction factor, commonly near 5/6 for a homogeneous rectangular section but dependent on the assumed distribution and formulation.

Mindlin theory is more appropriate for thick plates, sandwich panels, short spans and regions near concentrated loads or supports. Low-order finite elements may suffer from shear locking as the plate becomes thin; reduced integration, assumed-strain methods or higher-order formulations may be needed.

4. Large Deflection and Nonlinear Plate Behaviour

Linear plate theory becomes unreliable when transverse deflections are sufficiently large for in-plane stretching to develop. The Föppl–von Kármán formulation incorporates nonlinear membrane strains. A representative strain component is

with analogous terms in the [[M.y]] direction and a nonlinear contribution to in-plane shear strain. The quadratic terms are responsible for membrane stiffening. Consequently, a plate that is flexible in linear bending may carry additional load through stretching after significant deflection.

Geometric nonlinearity is important in membrane roofs, thin diaphragms, pressure-loaded panels and post-buckling analysis. Material nonlinearity, cracking, plasticity, creep and contact may need to be included separately; geometric nonlinearity alone does not represent these effects.

5. Shell Mechanics

5.1 Membrane and bending theories

A shell can resist load through membrane forces when its geometry and supports permit a compatible in-plane force path. In a membrane idealization, bending moments and transverse shear forces are neglected. This is efficient but restrictive: concentrated loads, abrupt support conditions, edge restraints and changes in curvature generally create bending boundary layers.

Shell bending theory retains membrane forces, bending moments, twisting moments and transverse shear forces. It is needed when the shell is thick, locally loaded, poorly shaped for membrane action or subject to significant boundary restraint.

For a shell with principal radii of curvature [[M.R_1]] and [[M.R_2]], a simplified normal-equilibrium relation for pressure [[M.p]] is

where [[M.N_1]] and [[M.N_2]] are membrane forces per unit length in the principal directions. This equation assumes a smooth shell, pressure normal to the surface, negligible transverse shear and a membrane state. It illustrates why curvature enables a shell to carry pressure without relying entirely on bending.

For a sphere with radius [[M.R]] under uniform internal pressure [[M.p]], symmetry gives equal membrane force resultants:

The result is tensile for internal pressure. For external pressure, the membrane force is compressive and buckling may govern.

For a thin cylindrical shell under internal pressure, the hoop and longitudinal membrane forces are, respectively,

where [[M.r]] is cylinder radius. Dividing by thickness gives membrane stresses [[M.σ_θ=pr/h]] and [[M.σ_z=pr/(2h)]]. These expressions apply to a closed, thin cylinder away from ends, openings, discontinuities and local support effects.

5.2 Curvature and structural efficiency

Double curvature generally improves stiffness and stability because loads can be distributed through two principal directions. Synclastic shells, such as spherical domes, have curvatures of the same sign. Anticlastic shells, such as hyperbolic paraboloids, have opposite principal curvatures and can develop efficient tensile–compressive membrane systems. Developable shells, such as cylinders, have one zero principal curvature and may require stiffening against ovalization or buckling.

5.3 Boundary conditions and edge effects

An ideal membrane solution may predict zero bending in the central region but cannot satisfy all support conditions without localized bending. This is called a boundary layer or edge effect. Ring beams, edge girders, diaphragms and stiffened supports are frequently required to collect shell membrane forces and control deformation.

Support modelling must represent actual restraint. A shell edge fixed against tangential movement may attract large membrane forces; a nominally pinned support may still restrain rotation or out-of-plane movement through connection geometry. Over-restraint in a finite-element model can produce unrealistic stress concentrations.

6. Spatial Structural Systems

Spatial structures include space trusses, space frames, grids, lattice domes, cable systems, tensegrity structures and three-dimensional building frames. Their distinguishing feature is that equilibrium and compatibility require all three translational directions and, where appropriate, rotational degrees of freedom.

A three-dimensional truss idealizes members as two-force elements. A space-frame member generally carries axial force, biaxial bending, torsion and shear. At a structural level, the stiffness method is written as

where [[M.K]] is the assembled global stiffness matrix, [[M.d]] is the vector of unknown nodal displacements and [[M.F]] is the vector of applied nodal loads. After solving, member forces are obtained by transforming nodal displacements to local coordinates and applying the member constitutive relations.

For a three-dimensional truss bar connecting nodes [[M.i]] and [[M.j]], with length [[M.L]] and direction cosines [[M.l,m,n]], the local axial stiffness is [[M.EA/L]]. The global element stiffness is obtained by transformation. The axial extension is

and the axial force is

where [[M.A]] is cross-sectional area and [[M.E]] is Young’s modulus. This linear relation assumes small displacement, elastic material behaviour and a pin-jointed member carrying axial force only.

A space-frame beam-column formulation must include local bending about two axes, torsion and shear where relevant. Warping torsion may be significant for open thin-walled sections. Member releases, eccentric joints, semi-rigid connections and joint offsets should be modelled explicitly when they materially affect force distribution.

Stability and determinacy

A spatial system must be stable in all six rigid-body modes where applicable: three translations and three rotations. A count of members and joints can indicate possible instability but cannot replace a matrix-rank or mechanism check. A structure may satisfy a simple determinacy count yet contain a geometric mechanism, and a numerically stable model may be physically unrealistic if restraints are artificial.

The elastic critical-load problem is commonly expressed as

where [[M.K]] is the elastic stiffness matrix, [[M.K_G]] is the geometric stiffness matrix generated by existing forces, [[M.λ]] is an eigenvalue multiplier and [[M.d]] is a buckling mode. The smallest positive relevant eigenvalue estimates an elastic critical load factor. Real design resistance must additionally consider imperfections, residual stresses, material yielding, connection flexibility and code-specific reduction factors.

7. Finite-Element Modelling of Plates and Shells

Finite-element analysis is the principal practical method for complex plates and shells. Common formulations include:

• plate elements, for flat elements with bending and possibly shear deformation;
• shell elements, combining membrane and bending behaviour on a curved surface;
• solid elements, useful for thick regions, joints and three-dimensional stress states;
• beam and truss elements, for spatial members and edge stiffeners.

A reliable modelling workflow is:

1. define geometry, thickness and material axes;
2. idealize loads as pressure, line load, point load or distributed body load;
3. select an element formulation compatible with thickness and expected deformation;
4. establish realistic supports and connections;
5. refine the mesh in regions of curvature change, openings, supports and load introduction;
6. check convergence of displacements, reactions and meaningful stress resultants;
7. verify global equilibrium and compare with hand calculations or benchmark solutions;
8. assess sensitivity to mesh, boundary conditions and material assumptions.

Nodal stresses at sharp corners, point restraints and singular load applications may not converge. In such regions, design should use force resultants, averaged stresses, structural hot-spot methods or a refined local model rather than the unqualified peak at one integration point.

For orthotropic plates, the constitutive matrix must be oriented with the material axes. In reinforced concrete, the effective stiffness depends on cracking, reinforcement arrangement, tension stiffening, creep and load history. A gross uncracked elastic shell model may be suitable for preliminary force flow but not for serviceability or ultimate assessment without appropriate stiffness modification.

8. Worked Example: Simply Supported Isotropic Plate

Consider a simply supported square steel plate with side [[M.a=2.0 m]], thickness [[M.h=0.010 m]], Young’s modulus [[M.E=200 GPa]], Poisson’s ratio [[M.ν=0.30]], subjected to a uniform pressure [[M.q=2 kPa]]. Estimate the central deflection using the first Navier term.

For a uniform load, the sine-series coefficient for odd [[M.m]] and [[M.n]] is

The first term has [[M.m=n=1]], so

For a square plate,

Since [[M.Q_{11}=16q/π²≈3242 N/m]], and [[M.{2(π/2)²}²≈24.35 m⁻⁴]],

At the centre, the first term has sine factors equal to unity, giving an estimated central deflection of approximately [[M.7.3 mm]]. Higher odd terms reduce this estimate; for a uniformly loaded simply supported square plate, inclusion of additional terms gives a value of approximately [[M.7.0 mm]]. The calculation is elastic, small-deflection and thin-plate based. It does not check yielding, vibration, local indentation, buckling or serviceability criteria.

9. Applied Case: Illustrative Long-Span Lattice Dome

Background

A proposed sports hall has a circular plan, diameter 60 m, and a triangulated steel lattice dome supported on a reinforced-concrete ring beam. The roof carries self-weight, maintenance loading, wind suction and asymmetric snow accumulation.

Problem

A preliminary axisymmetric membrane model predicts modest compression in the dome. However, the actual design includes discrete supports, an oculus, nonuniform snow, imperfect member geometry and semi-rigid joints. The engineering question is whether the ideal membrane force pattern is reliable for member sizing and stability assessment.

Analysis and approach

A staged analysis is adopted:

1. a preliminary surface model establishes global thrust and deformation patterns;
2. a three-dimensional beam-and-joint model includes member eccentricities, connection stiffness and the ring beam;
3. load cases include gravity, asymmetric snow, wind pressure and construction stages;
4. geometric imperfections are introduced using scaled buckling modes and measured fabrication tolerances;
5. a geometrically nonlinear analysis with material yielding is performed for governing combinations;
6. member design checks axial resistance, flexural interaction, connection capacity and fatigue where applicable;
7. reactions and ring-beam forces are independently checked by global equilibrium.

The analysis shows that uniform gravity produces predominantly compressive meridional forces, while asymmetric snow causes bending and circumferential force redistribution. The oculus generates local force concentrations requiring a stiff edge ring. The lowest linear buckling mode is not itself accepted as a design resistance; imperfection-sensitive nonlinear analysis produces a lower and more realistic capacity.

Outcome and learning points

The dome is redesigned with a deeper edge ring, improved joint detailing and additional bracing around the oculus. The principal learning points are that membrane theory is a valuable conceptual and preliminary tool, but support conditions, discontinuities and imperfections determine practical performance. Stability, not merely elastic stress, governs many slender spatial systems.

10. Limitations, Professional Practice and Contemporary Development

Limitations of idealized analysis

Analytical solutions are restricted by geometry, loading and boundary conditions. Finite-element models can create false confidence if they contain inappropriate constraints, poor element quality or unverified material properties. Linear analysis cannot predict post-buckling, cracking, yielding or contact. Shell stress output may depend on local coordinate definitions and smoothing procedures.

Codes, safety and governance

Design must follow the governing structural code and project requirements for actions, combinations, material resistance, fire, durability, robustness and accidental actions. The engineer remains responsible for selecting assumptions and interpreting software output; software verification is not a substitute for engineering judgement. Independent checking is especially important for unusual shells, long-span roofs and imperfection-sensitive systems.

Safety-critical models should preserve an auditable record of geometry, assumptions, releases, load cases, combinations, convergence criteria, solver warnings and design decisions. Ethical practice requires transparent communication of uncertainty and avoidance of presenting highly precise numerical results unsupported by the model.

Digital methods and AI

Parametric modelling can connect geometry, analysis and fabrication constraints. Automated mesh refinement and surrogate models can accelerate design exploration, while machine learning may assist anomaly detection or approximate response. These tools should be validated against mechanics-based models and physical evidence. AI-generated recommendations cannot replace equilibrium checks, code compliance, sensitivity analysis or professional accountability.

Best-practice checklist

• Use membrane solutions as benchmarks, not universal design models.
• Check units, coordinate systems, thickness assignments and material directions.
• Confirm reactions against applied loads.
• Perform mesh and boundary-condition sensitivity studies.
• Distinguish singular numerical peaks from convergent structural actions.
• Include geometric imperfections in stability-sensitive structures.
• Model connection flexibility when it influences load paths.
• Report assumptions, limitations and governing load combinations.
• Use independent calculations for critical quantities.

Chapter Summary

Plates are thin or moderately thick two-dimensional elements whose response combines bending, twisting, transverse shear and, under large deflection, membrane action. Classical Kirchhoff–Love theory provides the thin-plate biharmonic equation, while Mindlin–Reissner theory includes transverse shear deformation. Shells exploit curvature to carry loads through membrane forces, but real shells develop bending near supports, openings, concentrated loads and geometric discontinuities. Spatial structural systems require three-dimensional equilibrium, compatibility and stability assessment; their analysis is commonly expressed through the global stiffness equation and, for buckling, the geometric-stiffness eigenproblem. Reliable practice combines analytical benchmarks, carefully formulated finite-element models, nonlinear and imperfection-sensitive stability analysis, independent equilibrium checks and transparent professional judgement.

Key Terms

Plate: A flat, relatively thin element carrying transverse and in-plane actions.

Shell: A curved thin structural surface capable of membrane and bending action.

Middle surface: Surface midway through the thickness of a plate or shell.

Membrane force: In-plane force resultant per unit length.

Bending moment resultant: Moment per unit length associated with shell or plate curvature.

Kirchhoff–Love theory: Thin-plate theory neglecting transverse shear deformation.

Mindlin–Reissner theory: Plate theory including transverse shear deformation and independent rotations.

Flexural rigidity: The quantity [[M.D=Eh³/[12(1−ν²)]]] for an isotropic thin plate.

Biharmonic equation: Governing fourth-order differential equation for classical plate bending.

Orthotropy: Direction-dependent stiffness with mutually perpendicular material axes.

Boundary layer: Localized region of high bending near a shell edge, support or discontinuity.

Space frame: Three-dimensional frame resisting combined axial, bending, shear and torsional actions.

Geometric stiffness: Stiffness contribution produced by existing stress resultants.

Buckling mode: Deformation pattern associated with an instability eigenvalue.

Shear locking: Artificial stiffness error in some finite elements for thin-plate bending.

Membrane action: Load resistance primarily through in-plane tension or compression.

Imperfection sensitivity: Strong dependence of stability capacity on initial geometric or material imperfections.

Structural singularity: Location where idealized mathematical stresses do not converge to a finite peak.

Ring beam: Circumferential edge member collecting and equilibrating shell forces.

Tensegrity: Spatial system combining discontinuous compression members with continuous tension elements.

Review Questions

5 Short Answer Questions

1. State the principal kinematic assumptions of Kirchhoff–Love plate theory.
2. Distinguish membrane force resultants from bending moment resultants.
3. Why is transverse shear deformation included in Mindlin–Reissner theory?
4. What is meant by a shell boundary layer?
5. Identify the six rigid-body modes relevant to a three-dimensional structural system.

5 Long Answer Questions

1. Derive the classical isotropic plate equation from the moment–curvature relations and transverse equilibrium.
2. Explain the difference between membrane and bending theories of shells, including their assumptions and limitations.
3. Discuss the formulation and interpretation of the global stiffness equation for a space frame.
4. Explain shear locking, stress singularities and mesh-convergence requirements in shell finite-element analysis.
5. Describe how imperfections and geometric nonlinearity influence the buckling resistance of a lattice dome.

3 Analytical / Application Questions

1. A simply supported rectangular plate is twice as long as it is wide and carries uniform pressure. Explain, using the Navier formulation, how aspect ratio influences the dominant deflection term and bending moments.
2. A cylindrical tank is subjected to internal pressure and has a large roof opening. Identify which membrane results remain useful and which local effects require shell bending or a three-dimensional model.
3. Develop an analysis plan for a space truss with semi-rigid joints, asymmetric wind loading and suspected support settlement. Specify model idealizations, checks, nonlinear effects and stability assessments.

Multiple Choice Questions with Answers

Question 1

Which assumption is central to classical Kirchhoff–Love thin-plate theory?

A. Transverse shear deformation is dominant

B. Normals to the middle surface remain straight and normal after deformation

C. The plate carries load only through membrane tension

D. Material yielding occurs throughout the thickness

Correct Answer: B. Normals to the middle surface remain straight and normal after deformation

Question 2

For an isotropic plate, the flexural rigidity is represented by which expression?

A. [[M:D=Eh]]

B. [[M:D=E/h³]]

C. [[M:D=Eh³/[12(1−ν²)]]]

D. [[M:D=12Eh³(1−ν²)]]

Correct Answer: C. [[M:D=Eh³/[12(1−ν²)]]]

Question 3

For a spherical shell under uniform internal pressure, the membrane force resultant in each principal direction is:

A. [[M:pR]]

B. [[M:pR/2]]

C. [[M:2pR]]

D. [[M:p/(2R)]]

Correct Answer: B. [[M:pR/2]]

Question 4

Which theory explicitly includes transverse shear deformation in plate analysis?

A. Euler–Bernoulli beam theory

B. Kirchhoff–Love theory

C. Mindlin–Reissner theory

D. Pure membrane theory

Correct Answer: C. Mindlin–Reissner theory

Question 5

In the global stiffness equation [[M:Kd=F]], the vector [[M:d]] represents:

A. Element stresses

B. Nodal displacements

C. Material properties

D. Geometric imperfections only

Correct Answer: B. Nodal displacements

Question 6

For a thin cylindrical shell under internal pressure, the hoop membrane force per unit length is:

A. [[M:pr]]

B. [[M:pr/2]]

C. [[M:2pr]]

D. [[M:p/r]]

Correct Answer: A. [[M:pr]]

Question 7

What is the principal reason for including geometric stiffness in buckling analysis?

A. To represent paint thickness

B. To account for stiffness changes caused by existing stress resultants

C. To eliminate all material nonlinearity

D. To calculate only thermal expansion

Correct Answer: B. To account for stiffness changes caused by existing stress resultants

Question 8

A high stress at a sharp, idealized point restraint that changes with mesh refinement is most likely a:

A. Structural singularity

B. Material phase change

C. Rigid-body mode

D. Membrane resultant

Correct Answer: A. Structural singularity

Question 9

Which action commonly creates bending in an otherwise membrane-dominated shell?

A. A perfectly smooth uniform pressure field only

B. A concentrated load or abrupt support restraint

C. Zero curvature everywhere

D. A uniform material density

Correct Answer: B. A concentrated load or abrupt support restraint

Question 10

For a three-dimensional truss bar, the small-displacement axial force relation is:

A. [[M:N=EAδ/L]]

B. [[M:N=EL/(Aδ)]]

C. [[M:N=Aδ/(EL)]]

D. [[M:N=Eδ/(AL)]]

Correct Answer: A. [[M:N=EAδ/L]]