Plate & Shell Theory — Engineering Reference

1. At a glance

Plate theory and shell theory are the two-dimensional continuum-mechanics formulations engineers use to analyse thin structures — bodies whose thickness h is small compared to in-plane dimensions (typically h/L < 1/20 for “thin”, 1/20–1/10 for “moderately thick”). The midsurface (the locus of points equidistant from the two faces) becomes the analytic domain; the three-dimensional displacement field is expressed through a small number of midsurface quantities by a kinematic assumption about how normals to the midsurface deform.

The distinction:

• Plate — flat midsurface, h ≪ L. Loaded transversely → bends. In-plane response decouples from bending response in the linear, isotropic, homogeneous case.
• Shell — curved midsurface, h ≪ R (radius of curvature). Loaded by pressure, thrust, or transverse load → in-plane (membrane) action and bending action couple via the curvature. A pressurised cylinder carries pressure primarily by hoop tension (membrane); a flat circular plate of the same thickness under the same pressure would deflect by orders of magnitude more and develop large bending stresses.

Where these theories live in industry: pressure vessels (ASME BPVC, EN 13445), storage tanks (API 650/620, AWWA D100), ship hulls + decks (ABS, DNV, LR class rules), aircraft fuselage + wing skin (FAR 25, EASA CS-25), building floor slabs + roofs (ACI 318, Eurocode 2), pipelines (ASME B31.3/B31.4/B31.8), silos + bunkers (EN 1991-4), domes + cooling towers (ACI 372/334), automotive body panels (FEA-driven), MEMS diaphragms, satellite solar arrays, and printed-circuit-board substrates.

Place in the design stack: continuum mechanics → mechanics-of-materials (1D bar/beam/shaft) → plate & shell theory (2D thin-walled) → fem-fea (numerical solution on arbitrary geometry) → code design (ASME/EN/API).

2. Why it matters

Most engineered structures are made of thin walls. A beam idealisation is too simple — it forces uniaxial stress and ignores the load-spreading benefit of two-dimensional action. A full 3D solid analysis is overkill — for a wall whose thickness is 1 % of its other dimensions, the through-thickness variation of stress is dominated by a low-order polynomial (linear for Kirchhoff, linear-plus-shear for Mindlin) and the additional 3D DOFs are wasted. Plate and shell theories sit between these, reducing the 3D elasticity problem to a 2D PDE on the midsurface and recovering full 3D stress via the kinematic assumption.

Modern finite-element codes (ANSYS Mechanical, Abaqus, Nastran) ship with extensive shell-element libraries (S4R, MITC4, CQUAD4) that automate the solve, but the engineer must still understand the assumptions, the boundary-layer behaviour at edges and junctions, the buckling-imperfection sensitivity of compressed shells, and the code-formula vs design-by-analysis trade-off. Failures that came from misapplying or under-respecting shell behaviour:

• Aloha Airlines Flight 243 (1988) — 18-foot section of fuselage skin tore off a Boeing 737-200 at altitude. Multi-site fatigue damage along a row of lap-joint rivets propagated into a controlling crack; the pressurised shell (Δp ≈ 60 kPa) lost containment. Drove industry-wide adoption of FAA AC 91-56B aging-aircraft and damage-tolerance reanalysis.
• Sleipner A platform (1991) — Condeep concrete oil platform sank during ballast-down off Norway. Root cause: a tricell cell wall (concrete shell element) was under-reinforced for the local high-bending boundary-layer near the slot; an unconservative shear assumption in the FE post-processing missed it.
• Lac-Mégantic (2013) — DOT-111A tank cars (cylindrical shells, ~13 mm wall) ruptured catastrophically during derailment and burned 47 people. The tank-car shell was insufficient under impact + thermal — Transport Canada then mandated TC-117 with thicker shells and head shields.
• Hartford Civic Center roof collapse (1978) — space-frame roof shell collapsed under snow; nodal-eccentricity moments and plate-buckling of compression-chord webs were misrepresented in the design analysis.

These are not exotic failures. They are the everyday consequence of the same theory the engineer uses to size a tank head, a floor slab, or a satellite cylinder.

3. First principles — plate

3.1 Geometry and definitions

Take the midplane at z = 0; thickness h; upper and lower faces at z = ±h/2. For thin-plate theory the standard limits are h/L < 1/20 (Kirchhoff strictly valid) and h/L < 1/10 (Mindlin acceptable up to). Loading is decomposed into:

• In-plane (membrane) loads — N_x, N_y, N_xy [N/m]
• Transverse load — q(x,y) [Pa = N/m²]
• Bending moments per unit length — M_x, M_y, M_xy [N·m/m]
• Transverse shear forces per unit length — Q_x, Q_y [N/m]

3.2 Kirchhoff–Love thin-plate (CPT, 1850)

Three kinematic assumptions:

1. Straight normals — material normals to the undeformed midplane remain straight after deformation.
2. Inextensible normals — they do not stretch (σ_z is neglected, ε_z = 0).
3. Normal-perpendicularity — they remain perpendicular to the deformed midsurface (γ_xz = γ_yz = 0).

Consequence: the only kinematic unknown is the midsurface transverse deflection w(x,y). In-plane displacements at height z are u(x,y,z) = −z·∂w/∂x, v(x,y,z) = −z·∂w/∂y.

Constitutive relation (isotropic, plane-stress in z):

M_x  = − D · (∂²w/∂x² + ν·∂²w/∂y²)
M_y  = − D · (∂²w/∂y² + ν·∂²w/∂x²)
M_xy = − D · (1 − ν) · ∂²w/∂x∂y

Flexural rigidity:

D = E · h³ / (12 · (1 − ν²))         [N·m]

Governing PDE (substitute moments into transverse equilibrium):

D · ∇⁴w  =  q(x,y)              (biharmonic plate equation)
∇⁴ = ∂⁴/∂x⁴ + 2·∂⁴/∂x²∂y² + ∂⁴/∂y⁴

This is the biharmonic equation with right-hand side q/D. It is fourth-order in space, so each plate edge requires two boundary conditions: typical pairs are (w, ∂w/∂n) on a clamped edge, (w, M_n) on a simply-supported edge, and (M_n, V_n) on a free edge — where V_n = Q_n + ∂M_ns/∂s is the Kirchhoff effective shear (the ghost-shear correction that arises from reducing three natural BCs to two on a fourth-order PDE).

3.3 Mindlin–Reissner moderately-thick plate (1945–1951)

Eric Reissner (1945) and Raymond Mindlin (1951) relaxed the perpendicularity assumption: normals stay straight but rotate independently from the midsurface slope. This admits transverse shear strain:

γ_xz = ψ_x + ∂w/∂x
γ_yz = ψ_y + ∂w/∂y

Three midsurface unknowns: w(x,y), ψ_x(x,y), ψ_y(x,y). Constitutive relation adds the shear-rigidity term:

Q_x = κ · G · h · γ_xz
Q_y = κ · G · h · γ_yz

with shear correction factor κ ≈ 5/6 for rectangular cross-sections (Reissner) or π²/12 (Mindlin’s frequency-matched value). The governing system is three coupled second-order PDEs, requiring only one BC per edge per DOF.

Mindlin–Reissner is the workhorse of modern FE shell elements (MITC4, MITC9, S4R, CQUAD4) because (a) it generalises naturally to thick shells, (b) it requires only C⁰ continuity (Kirchhoff needs C¹, which is hard with standard isoparametric elements), and (c) it converges to Kirchhoff as h/L → 0 provided the formulation handles shear locking correctly (MITC = Mixed Interpolation of Tensorial Components; assumed-natural-strain; reduced/selective integration).

3.4 Higher-order plate theories

For composite laminates and sandwich panels, where shear flexibility is high and warping is significant, third-order theories (Reddy 1984) assume a cubic through-thickness distribution of in-plane displacement: u(x,y,z) = u₀ − z·∂w/∂x + z²·ζ_x + z³·η_x. This satisfies zero shear on top and bottom surfaces automatically, eliminating the need for a shear correction factor. Layerwise theories treat each ply (or sub-laminate) with its own kinematic field — more accurate, far more DOFs.

Theory	DOFs / point	Shear deformation	h/L validity	C-continuity	FE shell elements
Kirchhoff–Love (CPT)	1 (w)	Ignored	< 1/20	C¹	DKT, DKQ, BCIZ
Mindlin–Reissner FSDT	3 (w, ψ_x, ψ_y)	First-order	< 1/10	C⁰	MITC4/9, S4R, CQUAD4, ANS
Reddy TSDT	5 (+ higher)	Third-order (cubic)	< 1/5	C⁰ (or C¹)	Research / composites
Layerwise / 3D-shell	n × 3 per layer	Full	Any	C⁰ per layer	SC8R (Abaqus continuum shell), C3D8

4. First principles — shell

A shell is a curved thin body. The midsurface carries Gaussian curvature K = κ₁·κ₂ (where κ₁, κ₂ are the principal curvatures). The presence of curvature couples in-plane membrane action to out-of-plane bending — push downward on a dome and the midsurface develops circumferential compression that resists the deflection.

4.1 Membrane theory

Membrane theory assumes the shell carries load entirely through in-plane stress resultants, with zero bending. Valid in the interior of a smoothly-loaded thin shell with no abrupt boundaries or concentrated loads. For axisymmetric shells of revolution loaded by internal pressure p:

σ_meridional / r_1   +   σ_hoop / r_2   =   p / t          (Laplace shell equation)

with r_1 = meridional radius of curvature, r_2 = circumferential radius of curvature, t = wall thickness. Specialised:

Shell type	σ (hoop)	σ (meridional / axial)	Notes
Thin cylinder, internal p, dia 2R	p·R / t	p·R / (2·t)	r_1 = ∞, r_2 = R
Thin sphere, dia 2R	p·R / (2·t)	p·R / (2·t)	r_1 = r_2 = R; biaxial-equal
Cone, half-apex α, slant r	p·r / (t·cos α)	p·r / (2·t·cos α)	r varies along meridian
Hemispherical head on cylinder	p·R / (2·t)	p·R / (2·t)	Matches cylinder if t_head = t/2
Torus, R = centre-to-tube axis, a = tube radius	p·a·(2R+a·sin φ)/(2·t·(R+a·sin φ))	p·a/(2·t)	φ = circumferential angle

4.2 Bending theory and the boundary layer

Membrane theory fails where:

• The shell meets a flat closure (cylinder-to-flat-head, cylinder-to-dished-head)
• Concentrated point loads or moments are applied
• A support / skirt / nozzle introduces a discontinuity
• The shell is stiffened by a ring or stringer
• A free edge exists (open cylinder, edge of a dome)

In a narrow strip near these features the shell develops bending stress that decays exponentially into the interior. For a thin cylinder, the bending boundary-layer characteristic length is:

λ = ( 3 · (1 − ν²) / (R² · t²) )^(1/4)             [1/m]
ℓ_bl = 1/λ  ≈  ( R · t )^(1/2) / 1.285             (for ν ≈ 0.3)

Engineer’s rule of thumb: bending effects die out within ≈ 2.4·√(R·t) of a discontinuity. A 2 m diameter, 8 mm wall cylinder has ℓ_bl ≈ 2.4·√(1.0·0.008) = 0.21 m — discontinuities within 200 mm of each other interact; further apart, treat independently.

4.3 Classical shell theories

Theory	Year	Key feature	Industry use today
Love	1888	First consistent 1st-approximation shell theory (Love’s “first approximation”)	Historical foundation
Flügge	1934 / 1962	Retains 1+z/R curvature terms; better for moderately thick	Pressure vessel textbooks
Donnell–Mushtari–Vlasov (DMV)	1933 / 1951	”Shallow shell” simplification; drops some curvature-coupling terms	Cylindrical-shell buckling
Sanders–Koiter	1959 / 1960	Consistent strain-energy, satisfies rigid-body-motion exactly	Modern aerospace shells
Naghdi–Reissner	1963	Including transverse shear (FSDT for shells)	FE thick-shell elements
Koiter post-buckling	1945 (Dutch thesis), 1963 (English)	Initial-postbuckling analysis + imperfection sensitivity	Compressed shell design

Modern FE codes implement variants of Naghdi-type FSDT with MITC or ANS treatment of shear locking; the literature distinction between Love / Flügge / Sanders matters mostly for analytical solutions and for very-thin / very-precisely-loaded shells.

5. Plate bending — analytical and numerical solutions

5.1 Navier solution (rectangular, simply-supported on all four edges)

For a rectangular plate of sides a × b, simply supported on all four edges, the deflection under transverse load q(x,y) is expanded as a double Fourier sine series:

w(x,y)  =  Σ_m Σ_n  A_mn · sin(m·π·x/a) · sin(n·π·y/b)

Substituting into D·∇⁴w = q and matching coefficients:

A_mn  =  q_mn  /  ( D · π⁴ · ((m/a)² + (n/b)²)² )
q_mn  =  (4/(a·b)) · ∫₀ᵃ ∫₀ᵇ  q(x,y) · sin(m·π·x/a) · sin(n·π·y/b)  dx dy

For uniform load q₀: q_mn = 16·q₀ / (m·n·π²) for m,n odd, zero for even. Maximum deflection at the centre, for a square a=b plate with ν=0.3:

w_max  =  α · q₀ · a⁴ / D ,    α ≈ 0.00406 (square)
                                α ≈ 0.00772 (a/b = 2)

Maximum bending moment (centre): M_max ≈ 0.0479·q₀·a² (square, ν=0.3).

5.2 Lévy solution

For a plate simply supported on two opposite edges (x = 0 and x = a) and arbitrary BCs on the other two, the deflection is expanded as a single Fourier series in x with y-dependent coefficients:

w(x,y)  =  Σ_m  φ_m(y) · sin(m·π·x/a)

Substitution gives an ODE for φ_m(y) whose solution is a sum of hyperbolic + particular-integral terms; the four constants are fixed by the BCs at y = 0 and y = b. Faster convergence than Navier; standard textbook treatment (Timoshenko & Woinowsky-Krieger 1959).

5.3 Strip method and yield-line method (reinforced-concrete slabs)

For practical RC slab design the analytical Navier/Lévy solutions are augmented by two limit-state methods:

• Hillerborg strip method (1956) — a lower-bound (safe) plasticity method. Divides the slab into strips spanning between supports; each strip is designed as a beam. ACI 318 and Eurocode 2 permit it.
• Johansen yield-line theory (1943) — an upper-bound (unsafe-side) method. Postulates a kinematically admissible collapse mechanism (yield lines as plastic hinges) and equates external work to internal plastic dissipation. Gives a closed-form ultimate load. Used for irregular slabs, openings, and capacity checks of existing slabs.

These coexist with elastic Navier/Lévy because RC slabs are designed at the ultimate limit state (plastic) for strength and the serviceability limit state (elastic) for deflection / cracking.

5.4 Finite-element solution

Modern industry practice is FE. Two element families:

• Kirchhoff (thin-plate) elements — DKT (Discrete Kirchhoff Triangle, Batoz/Bathe/Ho 1980), DKQ (Discrete Kirchhoff Quad). Enforce zero transverse shear at discrete points instead of in the variational form. C¹-continuity is faked; works very well for thin plates.
• Mindlin (thick-plate) elements — MITC4/9 (Bathe), S4R (Abaqus), CQUAD4 (Nastran). C⁰-continuous; vulnerable to shear locking as h/L → 0 unless reduced/selective integration or assumed-natural-strain interpolation is used. MITC4 is the de-facto reference.

NAFEMS benchmarks LE1 (square plate, simple-supported, uniform load) and LE10 (clamped square plate, central point load) are the canonical tests; commercial codes are validated against these.

5.5 Plate buckling

A plate under in-plane compression buckles into a half-wave pattern. The classical elastic critical stress:

σ_cr  =  k · π² · E / (12 · (1 − ν²)) · (t/b)²

where b is the loaded width, t is the thickness, and k is a non-dimensional plate-buckling coefficient depending on aspect ratio and edge BCs:

Edge condition (loaded edges S-S)	k_min	k formula or note
All four edges simple-supported	4.0	k = (m·b/a + a/(m·b))² for aspect a/b
Two unloaded edges clamped, loaded S-S	6.97	Tighter restraint raises k
One unloaded edge S-S, other free	0.425	Free flange — controls AISC plate-girder design
All edges clamped	10.07	Maximum

This formula underpins the plate-slenderness limits in AISC 360-22 (b/t for compression flanges, h/t_w for webs of W-shapes) and Eurocode 3 Part 1-5 (plate-girder design). The effective-width method (von Karman 1932, Winter 1947) accounts for post-buckling reserve: beyond σ_cr the plate continues to carry load along the supported edges as the middle “unbuckles” — b_eff = b·√(σ_cr/σ).

6. Shell — analytical solutions and buckling

6.1 Pressurised cylinder

Internal pressure p, internal radius R, wall thickness t, ν = 0.3:

σ_hoop  (membrane)   =  p · R / t
σ_axial (membrane)   =  p · R / (2·t)
σ_radial             =  − p (inner surface) to 0 (outer surface)  [thin-wall: neglected]

Near a closure (flat head, dished head, flange), bending boundary-layer develops with decay length ℓ_bl ≈ 2.4·√(R·t). The peak edge bending moment (clamped to a rigid base) is:

M_edge  ≈  p · R · t / (2 · √(3·(1 − ν²)))    (per unit circumference)

Combined with the membrane stresses, the edge bending stress can exceed the far-field hoop stress by 30–50 %. This is why ASME Section VIII Div 1 has explicit nozzle-reinforcement (UG-37) and head-to-shell junction (UG-32, UG-33) rules.

6.2 Pressurised sphere

σ  =  p · R / (2 · t)    (equal in any meridian direction; "membrane-only" loading)

The sphere is the most efficient pressure-container shape — for a given p, R, and σ_allow, t is half what a cylinder requires. LPG bullets, LNG storage spheres (Moss/Kvaerner type), satellite propellant tanks, and small-scale industrial accumulators exploit this.

6.3 Cylinder buckling under axial compression

The classical (Donnell) critical axial stress:

σ_cr,classical  =  E · t  /  ( R · √(3 · (1 − ν²)) )
                ≈  0.605 · E · t / R     (ν = 0.3)

But real cylinders fail at 15–30 % of this value due to geometric imperfections — initial out-of-roundness, weld-line distortion, dent damage. Koiter (1945) showed that the cylinder post-buckling path drops sharply, making the cylinder extremely imperfection-sensitive. NASA SP-8007 (1968) tabulates empirical knockdown factors γ:

σ_cr,design  =  γ · σ_cr,classical
γ  ≈  1 − 0.901 · (1 − exp(− (1/16) · √(R/t)))     (NASA SP-8007)

For R/t = 100 → γ ≈ 0.59; for R/t = 500 → γ ≈ 0.36; for R/t = 1000 → γ ≈ 0.30. Current ESA / NASA design practice (Hühne/Rolfes 2008, Wagner 2017) prefers the Single Perturbation Load Approach (SPLA) — apply a calibrated lateral pinch load to a nonlinear FE model and use the resulting buckling load rather than empirical knockdown.

6.4 Cylinder buckling under external pressure

Long cylinder, simply-supported ends (between heavy frames):

p_cr  ≈  2.6 · E · (t/R)^(5/2)  /  ( (L/R) · (1 − ν²)^(3/4) )    (von Mises / Donnell)

Short cylinder or between rings: forms ovalisation buckling mode with n = 2,3,4,… circumferential waves. ASME BPVC Section VIII Div 1, UG-28 gives chart-based external-pressure design that codifies this.

6.5 Sphere buckling

p_cr,classical  =  2 · E · t²  /  ( R² · √(3 · (1 − ν²)) )
                ≈  1.21 · E · (t/R)²    (ν = 0.3)

Empirical knockdown ≈ 0.2–0.3 (extremely imperfection-sensitive due to multiple coincident buckling modes — Koiter’s theorem on simultaneous modes). Submarine pressure hulls (e.g. titanium DSV hulls) design at γ ≈ 0.7 because they are precision-formed and rigorously inspected; commercial steel storage spheres design at γ ≈ 0.2.

7. Worked examples

7.1 Example A — Reinforced-concrete flat plate (Kirchhoff)

Floor slab, simply supported on four edges, a = 5.0 m, b = 4.0 m, h = 0.15 m, RC with E = 30 GPa, ν = 0.20, uniformly distributed load q₀ = 10 kPa (self-weight + live).

D = E · h³ / (12 · (1 − ν²))
  = 30·10⁹ · (0.15)³ / (12 · (1 − 0.04))
  = 30·10⁹ · 3.375·10⁻³ / 11.52
  = 8.79·10⁶  N·m

Aspect ratio a/b = 1.25. From Timoshenko & Woinowsky-Krieger Table 8 (uniform load, all edges S-S, ν = 0.3 — close enough), w_max coefficient α ≈ 0.00138 (using shorter span b in the formula w_max = α · q₀ · b⁴ / D):

w_max = 0.00138 · 10·10³ · (4.0)⁴ / 8.79·10⁶
      = 0.00138 · 10000 · 256 / 8.79·10⁶
      = 3.53·10⁶ / 8.79·10⁶
      = 0.40 mm     (centre deflection)

Sanity check via simpler aspect-ratio formula: w/(q·b⁴/D) ≈ 0.00138 for a/b = 1.25 (Roark Table 11.4, case 1a). Result is in the range L/8000 — much less than the L/360 serviceability limit (b/360 = 11.1 mm). Strength check uses ACI 318 / Eurocode 2 for moments:

M_x,max  ≈  β · q₀ · b²    with β ≈ 0.050 for a/b = 1.25 (Timoshenko Table 8)
         =  0.050 · 10·10³ · 16
         =  8.0 kN·m / m  width

Compare to RC slab moment capacity from steel reinforcement. With 12 mm bars at 200 mm spacing, top + bottom, d = 120 mm: capacity ≈ 25 kN·m/m — adequate.

7.2 Example B — Cylindrical pressure vessel (membrane + ASME Div 1 check)

A 2.0 m inside-diameter (R = 1.0 m), 5.0 m long carbon-steel cylindrical vessel, t = 8 mm wall, design pressure p = 1.5 MPa internal, material SA-516 Gr.70 (σ_y = 260 MPa, σ_u = 485 MPa at room temperature), operating temperature 50 °C, corrosion allowance 0.

σ_hoop  = p · R / t  = 1.5·10⁶ · 1.0 / 0.008   = 187.5 MPa
σ_axial = p · R / (2·t)                         =  93.8 MPa
σ_radial ≈ 0 (thin wall)

Von Mises (plane stress, principal axes):
σ_VM = √(σ_hoop² − σ_hoop·σ_axial + σ_axial²)
     = √(187.5² − 187.5·93.8 + 93.8²)
     = √(35156 − 17588 + 8798)
     = √26367
     = 162.4 MPa

ASME BPVC Section VIII Div 1 allowable for SA-516 Gr.70 at 50 °C is the lower of σ_u/3.5 and σ_y/1.5 (per current ASME II-D):

S = min(485/3.5, 260/1.5) = min(138.6, 173.3) = 138.6 MPa

The required thickness from ASME UG-27 (cylindrical shell):

t_req  =  p · R / ( S · E_joint − 0.6 · p )       (UG-27 (1))
      =  1.5 · 1000 / (138.6 · 1.0 − 0.6 · 1.5)   [mm units, p MPa, R mm]
      =  1500 / (138.6 − 0.9)
      =  1500 / 137.7
      =  10.9 mm

(assuming full-radiography joint efficiency E_joint = 1.0). The 8 mm wall fails Code — required is 10.9 mm. Engineering action: increase t to 12 mm (next stock plate gauge), or change to SA-537 Cl.1 (S = 165 MPa). Worked Section VIII Div 2 design-by-analysis (allowable based on σ_y/2.4 ≈ 108 MPa for SA-516 Gr.70 plus elastic-plastic protection-against-collapse criteria) gives a slightly different but comparable answer.

7.3 Example C — Cylindrical-shell axial buckling (with imperfection knockdown)

A satellite payload-adapter cylinder, R = 0.5 m, t = 4 mm, mild steel E = 200 GPa, ν = 0.3, R/t = 125. Axial compression load capacity wanted.

Classical critical axial stress:
σ_cr,classical = E · t / ( R · √(3·(1 − ν²)) )
              = 200·10⁹ · 0.004 / (0.5 · √(3·0.91))
              = 800·10⁶ / (0.5 · 1.654)
              = 800·10⁶ / 0.827
              = 967 MPa

NASA SP-8007 knockdown for R/t = 125:
γ = 1 − 0.901·(1 − exp(−(1/16)·√125))
  = 1 − 0.901·(1 − exp(−0.699))
  = 1 − 0.901·(1 − 0.497)
  = 1 − 0.901·0.503
  = 1 − 0.453
  = 0.547

σ_cr,design = 0.547 · 967 = 529 MPa

But this exceeds steel σ_y (~ 250 MPa for plain mild steel) — material yields first:
σ_design = min(529, 250) = 250 MPa

Axial load capacity:
P_cr = σ_design · A = σ_design · 2π·R·t
     = 250·10⁶ · 2π · 0.5 · 0.004
     = 250·10⁶ · 0.01257
     = 3.14 MN

For a design factor of 1.5 against buckling/yield combined, the allowable load is 2.1 MN. A real qualification programme would (a) measure shell out-of-roundness, (b) run nonlinear FE with measured imperfection field or SPLA, and (c) confirm with a test article — for launch vehicles this is mandatory.

8. Pressure-vessel and tank application

8.1 Code landscape

Code	Scope	Allowable basis	Design philosophy
ASME BPVC Section VIII Div 1	Unfired pressure vessels, p ≤ ~20 MPa	S = min(σ_u/3.5, σ_y/1.5)	Design-by-formula. Charts + closed-form.
ASME BPVC Section VIII Div 2	Higher-stress vessels	S = σ_y/2.4 (typ), FE-based	Design-by-analysis. Requires elastic-plastic FE, fatigue, ratcheting protection.
ASME BPVC Section VIII Div 3	Very-high-pressure, p > 70 MPa	Fracture-mechanics based	Autofrettage, fracture, design-by-analysis mandatory.
ASME BPVC Section I	Power boilers	Lower allowable, creep regime	High-temperature steam.
EN 13445	European unfired pressure vessels	Parts 1–8 covering design / fabrication / inspection	Parallel to ASME VIII; Part 3 = design.
PD 5500	UK unfired pressure vessels (legacy)	Similar to ASME VIII Div 1	Still used in petrochemical retrofit.
JIS B 8265	Japan unfired pressure vessels	ASME-aligned	Used in Japanese petrochem + power.
API 650	Atmospheric welded steel oil tanks	Variable-design-point method	Bottom-up shell-course thickness.
API 620	Low-pressure (≤ 100 kPa) tanks	Membrane theory + bending allowance	LNG tanks (inner shell), LPG storage.
AWWA D100	Welded carbon-steel water tanks	Allowable-stress design	Municipal water storage.
API 12F / API 12D	Shop-welded production tanks	Standardised sizes	Onshore oilfield.

8.2 Heads and closures

A cylinder must close at each end. The geometry choice is a stress–cost trade:

• Hemispherical head — best stress distribution (σ = pR/(2t) = half cylinder hoop). Thinnest head for given p. Highest forming cost. ASME UG-32(f).
• 2:1 Ellipsoidal head — semi-major-to-minor 2:1. Standard for most pressure vessels. Discontinuity stress at the knuckle is the controlling region. ASME UG-32(d).
• Torispherical head (“flanged-and-dished”, “F&D”) — large spherical crown + small knuckle radius + cylindrical flange. Cheapest. Highest stress concentration at the knuckle — failures (e.g. Salem boiler 1965, Avonmouth NH3 vessel 1971) traced to knuckle plastic deformation. ASME UG-32(e).
• Flat head — bolted blind flange or welded flat plate. Rarely used above 1 MPa because thickness scales as √p (vs t ∝ p for cylinder). ASME UG-34.
• Conical head — for transitions between cylinders of different diameter; reinforced at the cone-cylinder junction.

8.3 Nozzle reinforcement

A penetration (manway, inlet, level-gauge connection) removes shell material and creates a stress concentration of ~3× (Lamé thick-walled / Kirsch hole effect, but shell-modified by curvature). ASME UG-37 / UG-39 require area-replacement reinforcement — the cross-sectional area of metal removed by the opening must be replaced by metal within a defined limit-of-reinforcement zone (a reinforcing pad welded around the nozzle, plus the credit of nozzle-wall and shell excess thickness). EN 13445-3 Annex Q is the European equivalent (limit-load-based, more efficient for large openings). Modern practice for complex nozzles or high cyclic service: 3D FE per ASME Section VIII Div 2 Part 5.

8.4 Special service

• Cryogenic LNG tanks — inner shell of 9 % Ni steel (good toughness at −196 °C) or 304L / 304 austenitic stainless, double-wall construction, perlite or vacuum insulation, outer carbon-steel containment. API 620 Appendix Q.
• Hot pressure vessels (creep regime, T > 0.4·T_melt) — allowable stresses drop with time-at-temperature; ASME Section II-D Table 1A gives creep-controlled S values down to 10⁻⁵ /h creep rate. Design typically incorporates 100,000-hour creep rupture criterion.
• Hydrogen service — hydrogen-induced cracking + hydrogen embrittlement; Nelson curves in API 941 govern material selection vs hydrogen partial-pressure / temperature.
• Sour service (H₂S environment) — NACE MR0175 / ISO 15156 limits steel hardness (≤ 22 HRC for C-Mn carbon steel) and dictates PWHT to prevent sulphide stress cracking.

9. Edge cases and gotchas

• Boundary layer near edges and discontinuities — bending moments are significant within ≈ 2.4·√(R·t) of a junction. Stress-classification per ASME Section VIII Div 2 Part 5: P_m (primary membrane, allowable S), P_L (local membrane, 1.5·S), P_b (primary bending, 1.5·S), Q (secondary, 3·S), F (peak, fatigue only). Misclassifying secondary stress as primary leads to over-design; the reverse is dangerous.
• Shell-to-flat-head junction — flat heads on cylinders cause large discontinuity bending. Either avoid (use dished head) or reinforce with a thick “blind flange” head + face groove.
• Wind girders and stiffeners on tall thin shells — open-top API 650 tanks need a wind girder near the top to prevent ovalisation under wind suction; intermediate stiffeners on tall slender silos.
• Local thinning from corrosion — drives buckling capacity down disproportionately. Pit-array thinning modelled per API 579-1 / ASME FFS-1 (Fitness-for-Service); remaining-life prediction.
• Out-of-roundness as imperfection — a 1 % diametrical out-of-roundness on a cylinder under external pressure can drop p_cr by 30–50 % vs perfect shell. ASME UG-80 limits initial out-of-roundness to (D_max − D_min) / D ≤ 1 %.
• Localised support loads — saddle supports on horizontal pressure vessels concentrate stress at the contact zone. Zick analysis (1951) gives closed-form stresses at saddles; embedded in ASME Section VIII Div 2 Annex 4-D and in pressure-vessel software (COMPRESS, PVElite).
• Welds in pressure vessels — Category A (longitudinal seam, hoop stress), Category B (circumferential seam, axial stress), Category C (nozzle-to-shell), Category D (head-to-shell). Joint efficiency E_joint depends on weld type and inspection level: E = 1.0 full-RT, 0.85 spot-RT, 0.70 no-RT. ASME UW-12.
• Fatigue in cyclic and thermally-cycled vessels — ASME Section VIII Div 2 Part 5 fatigue analysis using elastic stress-range range and Miner’s rule against published S-N curves (smooth-bar curves for low-cycle, welded-joint curves for fatigue Class FAT 71–125 per IIW). Required if N > 1000 cycles of significant range.
• Pipe ovalisation under bending (von Karman effect) — bending a thin pipe flattens the cross-section, dramatically lowering its bending stiffness and ultimately leading to local “Brazier” collapse. Important for offshore pipelines installed by S-lay reel-lay (DNV-ST-F101).
• Composite-laminate plate stacking — symmetric laminate ([0/45/−45/90]_s pattern) decouples membrane from bending; balanced + symmetric also eliminates extension-shear coupling. Designer-friendly laminates avoid B (bending-extension coupling) ≠ 0 because it complicates buckling and warping prediction.
• Sandwich panel — thin facesheets (CFRP or aluminium) on a low-density core (aluminium honeycomb, Nomex, PMI foam). Bending stiffness scales with face-to-face distance squared; mass scales linearly with face thickness. Extremely high specific bending stiffness; used in aircraft interiors, satellite panels, and racing-yacht hulls. Failure modes: face wrinkling (compressive face buckles into the core), intracell dimpling (face buckles between cell walls), core shear, face-to-core disbond. Allen 1969 / Zenkert 1995 are the classic monographs.
• Stiffened panel (aircraft fuselage, ship deck) — ring frames + longitudinal stringers carry compression in column-buckling mode; skin plate carries pressure / shear. Effective-width method: only a portion of skin near each stringer acts compositely under compressive load.
• Yoshimura diamond pattern — observed post-buckling shape of a thin cylinder under axial compression (Yoshimura 1955). Useful for thin-shell origami structures (deployable booms, crash energy absorbers).
• Punching shear in flat slabs — concentrated column loads punch through RC flat plates. ACI 318 / Eurocode 2 give critical-perimeter approach (b_0 at d/2 from column face, where d = effective depth). Failure mode is local — cross-link to reinforced-concrete for design rules and shear-reinforcement (studs, stirrups).

10. Tools and software

Tool	Class	Plate / shell strengths	Typical user
ANSYS Mechanical	General FE	SHELL181 (4-node, FSDT), SHELL281 (8-node), SOLSH190 (solid-shell). Composite Pre/Post.	Mainstream industry: aerospace, auto, energy.
Abaqus / SIMULIA	General FE (nonlinear-heavy)	S4R (reduced-int), S4 (full-int), STRI3 (Kirchhoff triangle), SC8R (continuum shell), conventional and continuum shell coupled with Riks/arc-length for post-buckling.	High-end nonlinear: aerospace, defense, biomedical.
MSC Nastran	Linear-dominant FE	CQUAD4, CQUAD8, CTRIA3, CTRIA6 — Nastran’s shell legacy is the de-facto aerospace standard for linear stress / vibration.	Aerospace (Boeing, Airbus internal flow).
COMSOL Multiphysics	General multi-physics FE	Shell + Plate + Membrane physics. Composite-laminate sub-module.	R&D, multi-physics coupling (acoustic / piezo).
Femap / Patran	FE pre/post + Nastran	Geometry-driven shell modelling	Aerospace.
COMPRESS (Codeware)	Pressure-vessel specialist	ASME BPVC VIII Div 1 + 2 by formula; nozzle UG-37 area replacement; saddle Zick analysis.	Pressure-vessel fabricators (chemical, refining).
PV Elite (Hexagon)	Pressure-vessel specialist	ASME VIII Div 1 + 2, EN 13445, PD 5500, ASME B31.3 piping integration. Wind/seismic.	EPC engineers, oil & gas.
NozzlePRO (Paulin Research)	Nozzle FE specialist	FE-based WRC 107/297 + ASME Section VIII Div 2 nozzle stress.	Critical-service nozzles.
TANK (Hexagon)	Storage-tank specialist	API 650, API 620, API 653 (inspection / fitness-for-service).	Tank design + integrity engineers.
AutoPIPE / CAESAR II	Pipe stress	ASME B31.x; thin-shell formulas internally for stress intensification factors.	Plant piping.
CalculiX	Open-source FE	Abaqus-compatible input syntax; shell elements; post-buckling.	Academic, hobbyist, OSS toolchain integration.
code_aster (EDF)	Open-source FE	DKT/DKQ + DST shell elements; nonlinear; nuclear-grade verification.	French nuclear (EDF); academic.
Roark’s Formulas for Stress and Strain (Young / Budynas / Sadegh, 9th ed 2020)	Reference handbook	Closed-form solutions for canonical plates + shells + pressure vessels — 30+ chapters.	Sanity-check before / after FE.
ESACOMP / HyperWorks Composite / Compro	Composites specialist	Laminate-mechanics (ABD matrix), failure criteria (Tsai–Wu, Tsai–Hill, max-strain), cure-induced residual stress.	Aerospace + wind-blade composites.
STAAD.Pro, RFEM / RSTAB, SAP2000	Structural civil FE	Plate / shell finite elements for slabs, walls, silos, tanks.	Building engineers.

11. Cross-references

• mechanics-of-materials — 1D stress/strain foundation that plate-shell theory generalises to 2D.
• beam-theory — 1D analogue. Euler–Bernoulli is to Kirchhoff what Timoshenko is to Mindlin–Reissner.
• structural-analysis — system-level methods (stiffness, finite-difference, finite-element) that consume plate/shell elements.
• fem-fea — companion numerical-implementation note. Shell-element families (S4R, MITC4, DKQ) live there in detail.
• materials-steel — pressure-vessel and tank materials (SA-516, SA-537, 9% Ni, austenitic stainless).
• materials-composites — composite-laminate plates and shells, classical laminate theory.
• fracture-mechanics — leak-before-burst pressure-vessel design, fatigue-crack growth in shells.
• fatigue-analysis — ASME Section VIII Div 2 Part 5 fatigue, weld classification IIW FAT.
• reinforced-concrete — flat-slab design, yield-line and strip methods, punching shear.
• aerodynamics — aircraft skin, pressure differential, aeroelastic plate flutter.
• heat-transfer — thermal stress in pressure vessels (Section VIII Div 2 thermal-transient analysis).
• structural-dynamics — plate/shell vibration modes, blast and seismic response.

12. Citations

• Timoshenko, S. & Woinowsky-Krieger, S. Theory of Plates and Shells. 2nd ed., McGraw-Hill, 1959. The canonical reference; coefficient tables for Navier/Lévy and shell-of-revolution solutions still used industry-wide.
• Reddy, J. N. Theory and Analysis of Elastic Plates and Shells. 2nd ed., CRC Press, 2007. Includes higher-order theories and laminated composites.
• Ventsel, E. & Krauthammer, T. Thin Plates and Shells: Theory, Analysis, and Applications. CRC Press, 2001.
• Donnell, L. H. Beams, Plates and Shells. McGraw-Hill, 1976.
• Flügge, W. Stresses in Shells. 2nd ed., Springer, 1973.
• Calladine, C. R. Theory of Shell Structures. Cambridge University Press, 1983. The most physical / engineering-intuition-driven shell text.
• Bushnell, D. Computerized Buckling Analysis of Shells. Martinus Nijhoff, 1985. Standard reference on BOSOR / shell-of-revolution buckling.
• Young, W. C., Budynas, R. G., & Sadegh, A. M. Roark’s Formulas for Stress and Strain. 9th ed., McGraw-Hill, 2020.
• Love, A. E. H. “The small free vibrations and deformation of a thin elastic shell.” Phil. Trans. Roy. Soc. London A 179, 491–546 (1888).
• Kirchhoff, G. “Über das Gleichgewicht und die Bewegung einer elastischen Scheibe.” J. reine angew. Math. 40, 51–88 (1850).
• Reissner, E. “The effect of transverse shear deformation on the bending of elastic plates.” J. Appl. Mech. 12, A69–A77 (1945).
• Mindlin, R. D. “Influence of rotatory inertia and shear on flexural motions of isotropic elastic plates.” J. Appl. Mech. 18, 31–38 (1951).
• Sanders, J. L. “An improved first approximation theory for thin shells.” NASA TR R-24 (1959).
• Koiter, W. T. “A consistent first approximation in the general theory of thin elastic shells.” Proc. IUTAM Symposium on the Theory of Thin Shells, Delft, 12–33 (1960).
• Koiter, W. T. The Stability of Elastic Equilibrium. Ph.D. thesis, Delft (1945); English translation: AFFDL-TR-70-25, 1970. Foundational text on post-buckling and imperfection sensitivity.
• Donnell, L. H. “Stability of Thin-Walled Tubes Under Torsion.” NACA Report 479 (1933).
• von Kármán, T. & Tsien, H. S. “The buckling of thin cylindrical shells under axial compression.” J. Aero. Sci. 8, 303–312 (1941).
• Yoshimura, Y. “On the mechanism of buckling of a circular cylindrical shell under axial compression.” NACA TM 1390 (1955).
• Bathe, K. J. Finite Element Procedures. 2nd ed., 2014. MITC element formulation.
• NASA SP-8007. Buckling of Thin-Walled Circular Cylinders. NASA Space Vehicle Design Criteria, 1968 (rev. 1968). Empirical knockdown curves still cited.
• Hühne, C., Rolfes, R., et al. “Robust design of composite cylindrical shells under axial compression — Simulation and validation.” Thin-Walled Structures 46, 947–962 (2008). SPLA method.
• ASME Boiler and Pressure Vessel Code, Section VIII Division 1, Division 2, Division 3. Current edition (2023). American Society of Mechanical Engineers, New York.
• EN 13445:2021 Unfired pressure vessels — Parts 1 to 8. CEN, Brussels.
• API 650 (13th ed., 2020). Welded Tanks for Oil Storage.
• API 620 (13th ed., 2020). Design and Construction of Large, Welded, Low-pressure Storage Tanks.
• AWWA D100-21. Welded Carbon Steel Tanks for Water Storage.
• API 579-1 / ASME FFS-1 (2021). Fitness-For-Service.
• NACE MR0175 / ISO 15156 (2020). Petroleum and natural gas industries — Materials for use in H₂S-containing environments.
• WRC Bulletin 107 / WRC Bulletin 297 — Local Stresses in Spherical and Cylindrical Shells due to External Loadings. Welding Research Council; canonical nozzle-stress reference until superseded by ASME Section VIII Div 2 Part 5 FE methods.
• Allen, H. G. Analysis and Design of Structural Sandwich Panels. Pergamon, 1969.
• Zenkert, D. An Introduction to Sandwich Construction. EMAS Publishing, 1995.