Statics and Structural Mechanics

Intuition

Statics studies bodies in equilibrium: ΣF = 0, ΣM = 0. This assumption works for the majority of load-bearing structures. The instantaneous force balance tells you what you need to predict the structure survival in the next second.[fn::This is not a deep truth about the universe. It is a deep truth about the regime of operation. Statics works when accelerations are small relative to gravity. Strains are small relative to geometry. The structure's first natural frequency is well above any excitation. The moment you step outside that envelope, the assumption breaks. A long-span bridge in vortex shedding, a milling spindle at 20 krpm, or a building in an earthquake will not obey the static formulas. The analysis becomes dynamics in disguise. The boundary is the whole game.]

Why this matters: every machine routes force through stress fields. Mechanical design controls where forces flow. The equations are elementary but govern everything from watch hairsprings to 600 m towers. Students treat the subject as solved. Practitioners forget when the assumptions break.[fn::A reviewer's objection is that statics is trivial. The rules of chess are also trivial. The combinatorial space of real structures is where the actual difficulty lives. Real structures involve redundancy, boundary ambiguity, contact, residual stress, and joint compliance. None of these topics appears in the textbook theorems.]

When it fails (in roughly increasing order of embarrassment):

• Geometric nonlinearity: deflections exceed ~L/200 and load moment arms shift with the deformed shape (P-Δ).
• Material nonlinearity: σ exceeds ~0.7σ_y and Hooke's law stops working (see Fatigue & Fracture).
• Stability: the load path bifurcates and the structure reaches a lower-energy configuration (see Buckling).
• Dynamics: when ΣF = 0 stops being valid.
• Contact compliance: boundary conditions assumed rigid are actually flexible.

First principles

Equilibrium and the free-body diagram

The subject reduces to two equations and one discipline: draw the diagram:

ΣF = 0, ΣM_P = 0 for any point P.

The free-body diagram isolates a body and replaces every external interaction with a force or moment at the point of action. A missing reaction is not a numerical error; it is a category error.

Euler-Bernoulli beam bending

Assumptions, stated so they can be violated on purpose: plane sections remain plane and perpendicular to the neutral axis; small deflection; linear, isotropic, homogeneous material; the beam is prismatic. From pure bending of a segment with end moments M, the strain varies linearly through the depth:

ε(y) = y / R,

and with Hooke's law σ = Eε we get the flexure (bending) equation:

M / I = σ / y = E / R,

where I = ∬_A y² dA is the second moment of area about the neutral axis. The three ratios are the same statement in three languages: geometric (curvature), kinematic (strain), and stress. The leverage of this single relation is hard to overstate — it is the analytical engine behind every beam, shaft, plate, and flexure in the discipline.

For a beam of width b and height h, I = b h³/12, so bending stress scales with h but stiffness scales with h³ — which is why deep beams are wildly more efficient than wide ones, and why I-sections put material at the flanges and starve the web.[fn:: The counterpoint: a deeper beam is also a less stable beam in the lateral-torsional sense, and a more expensive beam to fabricate and ship. Engineering is the discipline of not optimizing one equation in isolation.]

Shear and moment relations

For a beam under transverse load w(x),

dV/dx = -w(x), dM/dx = V(x).

Point loads produce jumps in V, distributed loads produce slopes in V and parabolas in M, and concentrated moments produce jumps in M. Reading the qualitative M/V diagram from the loading is a skill worth over-learning.

Deflection: integration and superposition

Integrate the moment-curvature relation twice to find deflection v(x). For point loads, step moments, and overhangs, Macaulay's method uses angle-bracket notation to handle discontinuities in a single expression. Integrate termwise and apply boundary conditions once.

Superposition extends this to arbitrary loadings, provided the system is linear — i.e., small deflection, linear material, fixed boundaries. Tabulated solutions for common cases (cantilever end load, simply-supported central load, uniform load) are added with coefficients. The moment this stops working is the moment one of those provisos breaks, and it breaks more often than people admit: a contact-supported beam that lifts off its support is no longer superposable.

Torsion

For a solid circular shaft with radius r, length L, and polar second moment J = Ï€rÂ⁴/2 under torque T:

T / J = τ / r = G θ / L,

where G = E/(2(1+ν)). This equation matches the bending formula with renamed variables. The same failure modes apply: non-circular sections cause warping, stress concentrations matter, and failure occurs when τ exceeds ~0.6σ_y.

For thin-walled closed sections (e.g. torsion tubes), the simpler Bredt's formula applies: τ = T / (2 A_m t), where A_m is the area enclosed by the median line and t the wall thickness. This is the basis of why a thin-walled tube is remarkably efficient in torsion per unit mass.[fn:: Note the inverse: an open thin-walled section — an I-beam, a split ring, a channel — has terrible torsional stiffness, scaling as t³ not t². This asymmetry is the single most important thing to know about torsion that the circular-shaft chapter does not tell you.]

Mohr's circle for plane stress

For a 2D stress state (σ_x, σ_y, τ_xy), the stresses on a plane at angle θ from the x-axis are:

σ_θ = (σ_x + σ_y)/2 + (σ_x - σ_y)/2 cos 2θ + τ_xy sin 2θ, τ_θ = -(σ_x - σ_y)/2 sin 2θ + τ_xy cos 2θ,

and the principal stresses (the extremes of σ_θ) are

σ_1,2 = (σ_x + σ_y)/2 ± √[ ((σ_x - σ_y)/2)² + τ_xy² ],

at principal angle θ_p = ½ arctan( 2τ_xy / (σ_x - σ_y) ).

Mohr's circle makes principal stresses visible. It shows how normal and shear stresses rotate through any plane.

Combined stress and failure criteria

Real components see bending, torsion, and axial loads simultaneously. The question shifts from "what is the stress?" to "does this stress state fail the material?" This is the domain of Failure Criteria: von Mises for ductile (σ' = √(σ_x² - σ_x σ_y + σ_y² + 3τ_xy²)), max-shear (Tresca) for a conservative lower bound, and the various Mohr/Coulomb variants for brittle. The statics part ends at the stress tensor; the failure part begins there.

Worked example

Example A — simply supported steel beam, central point load

A simply supported beam, length L = 2.0 m, cross-section 25 mm × 50 mm (b × h, strong axis), mild steel E = 200 GPa, σ_y = 250 MPa. Central point load P. Find P at first yield and the tip — i.e. midspan — deflection at that load.

Geometry and section: I = b h³ / 12 = (0.025)(0.050)³ / 12 = 2.604 × 10⁻⁷ m⁴. c = h/2 = 0.025 m. Z = I / c = 1.042 × 10⁻⁵ m³.

Reactions (symmetry): R_A = R_B = P/2.

Maximum moment at midspan: M_max = P L / 4.

First-yield condition (extreme fiber reaches σ_y): σ_y = M_max / Z => P_y = 4 σ_y Z / L = 4 (250 × 10⁶)(1.042 × 10⁻⁵) / 2.0 = 5210 N ≈ 5.21 kN.

Midspan deflection (from the standard superposition result δ = P L³ / (48 E I)): δ = (5210)(2.0)³ / (48 × 200 × 10⁹ × 2.604 × 10⁻⁷) = 4.17 × 10⁻⁴ m / (48 × 5.21 × 10⁴) ≈ wait — recomputing carefully with units: Numerator: 5210 × 8 = 41 680 N·m³. Denominator: 48 × 200e9 × 2.604e-7 = 48 × 5.208e4 = 2.500e6 N·m². δ = 41 680 / 2.500 × 10⁶ = 0.01667 m = 16.7 mm.

Sanity check: δ/L = 16.7/2000 ≈ 1/120. This is already outside the strict Euler-Bernoulli small-deflection regime (commonly taken as δ/L < ~1/200), so the result is an over-estimate of stiffness and an under-estimate of true deflection — a flag, not an answer. A design margin would push the section or shorten the span, and one would switch to a nonlinear (P-Δ) or finite-element check.[fn:: This is the kind of example that, taken at face value, teaches a student to plug in numbers and trust the calculator. The honest move is to compute the deflection, notice it violates the assumption, and report that the number is suspect — which is exactly what a practicing engineer does and an undergraduate almost never does.]

Example B — power-transmission shaft in torsion

A solid circular steel shaft, diameter d = 30 mm, length L = 500 mm, transmitting torque T. G = 79 GPa, τ_y (shear yield) ≈ 0.6 σ_y ≈ 150 MPa. Find torque at first yield and the angular twist.

J = π d⁴ / 32 = π (0.030)⁴ / 32 = 7.952 × 10⁻⁸ m⁴. c = d/2 = 0.015 m.

Yield condition: τ_max = T c / J = τ_y. T_y = τ_y J / c = (150 × 10⁶)(7.952 × 10⁻⁸) / 0.015 = 795.2 N·m.

Twist at yield: θ = T L / (G J) = (795.2)(0.500) / (79 × 10⁹ × 7.952 × 10⁻⁸) Numerator: 397.6 N·m². Denominator: 6.282 × 10³ N·m². θ = 0.0633 rad ≈ 3.63°.

For a rotating shaft this is also a stress state under combined loading (bending from gears/belts plus torsion), and one would normally apply a von Mises check or a Shigley-design fatigue criterion (see Fatigue & Fracture and Shigley, *Mechanical Engineering Design*). The pure-torsion number above is a limiting case, not a service load.

Pitfalls

These are the things that have, in aggregate, cost me or someone I worked with at least a week each. Stated as warnings, not rules.

1. Sign conventions on Mohr's circle. Two competing conventions exist. Mixing them flips the principal planes by 90°. Pick one convention and stick to it.

2. Assuming linearity past yield. The flexure equation M/I = σ/y = E/R applies only to the elastic regime. Once the extreme fiber yields, the linear formula under-predicts moment capacity by up to ~1.7 for ductile sections.

3. Ignoring shear in short beams. The Euler-Bernoulli model throws away shear deformation; for L/h ≳ 10 this is fine, for L/h ≲ 3 it is not, and a Timoshenko correction (or just an FEA run) is required. The error is ~ (h/L)², which is small but compounds with other small errors, and a short, deep beam fails in shear — a mode the textbook bending formula literally cannot see.[fn:: The trap is that a short beam "feels" stiff, so people under-estimate the load and over-design the moment capacity, then watch it fail in diagonal tension. The shear-dominated regime is where intuition from slender beams inverts.]

4. Boundary-condition mistakes. A "fixed" support may be compliant, a "pin" may transmit moment. Treat supports as stiffness matrices. Ask: is support stiffness at least 10× larger than the constrained member? If not, you do not have the support you drew.

5. Forgetting residual and thermal stresses. Welding, forming, and fit-up create residual stresses not in your free-body diagram. These change the failure calculation. Failure criteria see total stress, not just applied stress.

6. Treating "Euler-Bernoulli" and "beam" as synonyms. Four beam theories exist (Euler-Bernoulli, Timoshenko, shear-deformable, large-deflection). Each fails differently when mis-applied. Knowing which one you use matters more than finding the "correct" one.

References

• Shigley, *Mechanical Engineering Design* — best reference for combined stress and fatigue.
• Gere & Timoshenko, *Mechanics of Materials* — clearest Mohr's-circle exposition.
• Hjelm, *Mechanics of Materials* — modern, FEA-aware treatment.
• Beer & Johnston, *Mechanics of Materials* — useful for worked examples and sign conventions.
• Pilkey & Pilkey — stress concentration corrections for geometry.

Related

• Failure Criteria — where statics ends and "does it break" begins.
• Fatigue & Fracture — the regime where static stress is necessary but not sufficient.
• Buckling — when equilibrium becomes unstable and ΣF = 0 has two solutions.
• Finite Element Analysis — what you reach for once hand-calc assumptions stop being honest.
• Beam Analysis Calculator — the tool that should be used only after one can do the calculation by hand, and never before.