At a glance
What this guide walks through
Structural design follows a consistent chain: define the loads a structure must resist, find the resulting internal stresses and deformations in each member, check those against material and stability limits, and design connections that transfer force reliably between members. This guide follows that chain in order, then closes with what real structural failures reveal about where design margins actually matter.
The linear-elastic relationship underneath nearly every structural stress calculation.
Doubling span increases a simply-supported beam's deflection eightfold.
Doubling a column's unbraced length cuts its buckling capacity to a quarter.
Essentially constant across ordinary structural steel grades regardless of strength.
Quick navigation
01 // Loads: what a structure actually has to resist
Every structural design starts with defining the loads a structure must carry — get this step wrong and no amount of careful analysis downstream will save the design.
Dead load
The permanent weight of the structure itself and anything permanently attached — framing, floors, fixed mechanical/electrical equipment, cladding. Calculable precisely once materials and dimensions are fixed.
Live load
Variable, non-permanent load — occupants, furniture, stored goods, movable equipment. Building codes specify minimum design live loads by occupancy type (office, residential, storage, assembly) since actual loading varies constantly.
Environmental loads
Wind, snow and seismic (earthquake) load — all site- and location-dependent, defined by codes using statistical design values (e.g. a wind speed or snow load with a defined return period) rather than a single worst-case number.
02 // Load combinations and design philosophy
Real structures rarely see every possible load at its maximum value simultaneously, so codes specify load combinations — different weighted sums of load types — and a structure must be checked against all applicable combinations, with the governing (worst) case controlling the design at each location.
| Method | Approach |
|---|---|
| Allowable Strength Design (ASD) | Required strength from the applicable ASD load combinations is compared with allowable strength, commonly written Rn/Ω |
| Load and Resistance Factor Design (LRFD) | Factored load effects are compared with design strength φRn, using code-defined load and resistance factors |
ASD and LRFD are two code-calibrated ways of checking the same structural limit states. Their load combinations and resistance formats differ, but both explicitly account for uncertainty; for example, AISC 360 incorporates both LRFD and ASD. Which format governs a real project depends on the applicable material standard, jurisdiction and design basis.
03 // Stress and strain: the language of internal force
Once loads are known, structural analysis finds the resulting internal forces in each member, then converts those forces to stress (force per unit area) so they can be compared against material strength regardless of the member's actual size. Strain is the resulting deformation, normalized by original length so it's independent of how big the member is.
| Quantity | Relationship |
|---|---|
| Axial stress | σ = F/A |
| Strain | ε = ΔL/L₀ |
| Hooke's law (elastic region) | σ = E·ε |
| Simple elastic yield ratio (screening only) | Ry = σyield / |σactual| |
Young's modulus (E) is a material property, not a strength value — it describes stiffness (resistance to elastic deformation), not how much load the material can carry before failing. This is why steel and aluminum, with very different strengths, can have very different E values too (steel ≈200 GPa, aluminum ≈69 GPa) — a steel and an aluminum member of identical size will deflect very differently under the same load even before either one is anywhere near its strength limit.
04 // Interactive: axial stress and strain
Enter an axial-force magnitude, cross-sectional area and material properties to find stress, strain, elongation and a simple elastic yield ratio.
05 // Beams: bending stress and deflection
A beam loaded transversely develops internal bending moment, which produces bending stress that varies linearly across the section — maximum tension on one face, maximum compression on the other, zero at the neutral axis in between. The same moment also produces deflection, a serviceability concern distinct from strength: a beam can be strong enough to avoid breaking yet still deflect too much for the floor above it to feel solid or for a door beneath it to still close properly.
| Quantity | Relationship |
|---|---|
| Bending stress | σ = Mc/I |
| Deflection, simply supported, central point load | δ = PL³/(48EI) |
| Deflection, simply supported, uniform load | δ = 5wL⁴/(384EI) |
| Shear stress in a beam | τ = VQ/(Ib) |
Both deflection formulas scale with L³ or L⁴ — span length raised to the third or fourth power. This is why long-span beams need disproportionately more stiffness (larger I, meaning a deeper or otherwise more efficiently-shaped section) than a simple linear scale-up would suggest: doubling span alone, with everything else unchanged, multiplies point-load deflection by eight and multiplies uniform-load deflection by sixteen.
06 // Interactive: beam deflection
Enter a simply-supported beam's span, section and load to find maximum bending stress and deflection, for either a central point load or a uniformly distributed load.
Assumes a solid rectangular section. For other section shapes, see the Bending Stress Calculator.
07 // Columns: why slender members fail by buckling
A short, stocky column loaded in compression fails by crushing — the material simply reaches its compressive strength. A long, slender column under the exact same load and material can fail at a much lower load by buckling: it becomes laterally unstable and bows sideways well before the material itself is anywhere near its compressive strength. This is a stability failure, not a strength failure, and it's governed by geometry and stiffness (EI), not material strength.
| Quantity | Relationship |
|---|---|
| Euler critical buckling load | Pcr = π²EI/(KL)² |
| Radius of gyration | r = √(I/A) |
| Slenderness ratio | KL/r |
K is the effective-length factor, which accounts for how the column's ends are restrained — a column fixed at both ends effectively behaves as a much shorter, stiffer column than the same member pinned at both ends. Because buckling load scales with 1/L², doubling a column's unbraced length (or effective length, if end conditions change) cuts its buckling capacity to a quarter — which is exactly why bracing that shortens a column's unbraced length is often a far more effective (and cheaper) way to increase capacity than upsizing the section itself.
08 // Interactive: column buckling
Enter a column's section, length, end-condition factor and material to find Euler critical buckling load, slenderness ratio, and a comparison against a given applied load.
09 // Connections: critical force-transfer regions
Connections are critical force-transfer regions: forces from one member must pass into another through bolts, welds, bearing, anchors or other details. Their behavior can be more complicated than the idealized member forces alone because eccentricity, prying action, local yielding, slip and connection stiffness may all matter. A sound member design therefore still requires a separately sound connection design.
Bolted connections
Rely on either shear through the bolt shank or friction from bolt preload (slip-critical connections). See EngineerHub's Bolt Torque & Preload Calculator and Thread Engagement Calculator for the fastener side of this.
Welded connections
Fuse members directly together — strong and rigid, but quality depends heavily on procedure, inspection and the welder's skill, and welding introduces residual stress and potential heat-affected-zone material changes that bolted connections avoid.
Bearing connections
Transfer load through direct contact — a column base plate on a footing, a beam seat. Simple and robust, but need adequate bearing area and proper load spreading to avoid crushing or punching failure of the supporting material.
note Several well-known failures in section 10 involve a disconnect between the analyzed design and the structure as fabricated, modified or loaded. Any connection change that affects the force path needs appropriate engineering review; apparently small detailing changes can materially alter connection demand.
10 // What real structural failures teach us
Three well-documented failures, each illustrating a different root cause behind the concepts covered above.
Tacoma Narrows Bridge, 1940
The bridge's slender, solid-plate-girder deck (rather than an open truss that lets wind pass through) developed self-exciting aeroelastic flutter in moderate wind, a dynamic instability entirely different from a simple static overload. The center span collapsed after roughly four months in service. A lesson in checking dynamic and aerodynamic behavior, not just static strength.
Hyatt Regency walkway, 1981
A change from the originally detailed continuous hanger-rod arrangement to two offset rod segments doubled the demand on the upper box-beam connection. The change passed through an inadequate design-review process, and the connection failed during a crowded event, killing 114 people. A lesson in independently checking any fabrication or detailing change that alters the load path.
I-35W Mississippi River bridge, 2007
The NTSB identified inadequate load capacity in gusset plates at a key truss node due to an original design error. Later increases in bridge weight and concentrated construction loads on the day of the collapse increased demand on those already-deficient plates. A lesson in connection design verification, load-rating updates and tracking changes in dead and construction load over a structure's life.
A useful pattern across these three cases is the importance of matching the engineering model to the real structure and real loading: aerodynamic instability can govern even when static strength looks adequate; a connection detail can change the force path; and changes in dead or construction load can consume margin that was never rechecked. Failure investigations are multi-factor, so these short lessons are orientation rather than complete causal accounts.
11 // Background, FAQ, references and limitations
Expand for deeper context and the sources behind the figures used here.
Structural engineering as a formalized discipline grew rapidly through the 19th and 20th centuries alongside the growth of iron and steel construction, railways and taller buildings — each pushing beyond what earlier empirical, experience-based rules of thumb could safely cover. Modern structural codes (AISC 360 for steel, ACI 318 for concrete, ASCE 7 for loads, and their international equivalents) are living documents, regularly updated as research, computational tools and — often most directly — investigations into real failures reveal gaps in prior editions.
Each case influenced engineering practice and review: Tacoma Narrows became a foundational lesson in bridge aerodynamics and aeroelastic stability; Hyatt Regency reinforced the professional responsibility attached to reviewing connection/detailing changes; and I-35W renewed attention to gusset-plate design, inspection and load-rating practices.
What is the difference between dead load and live load? Dead load is the permanent weight of the structure and anything permanently attached. Live load is variable, non-permanent load such as occupants and movable equipment. Both combine with environmental loads (wind, snow, seismic) using factored load combinations.
Why does a beam's deflection depend on the cube of its span? The standard formula for central-point-load deflection is δ = PL³/48EI, so deflection grows with span cubed — doubling span increases deflection eightfold, which is why long spans need disproportionately stiffer sections.
Why do slender columns fail by buckling rather than crushing? A slender column becomes laterally unstable once axial load reaches a critical value, well before the material would fail in pure compression. Buckling load depends on 1/L², so unbraced length and effective length often matter more to capacity than material strength alone.
What usually causes a major structural failure? Rarely simple material weakness. Common root causes include an unreviewed design change during construction, a design error in a critical connection, cumulative load increases exceeding original design margins, or a dynamic effect (like wind-induced flutter) that static checks alone don't capture.
What is the difference between ASD and LRFD design methods? In modern code formats, ASD compares required strength from the applicable ASD load combinations with allowable strength, commonly Rn/Ω. LRFD compares factored load effects with design strength φRn. Both use code-defined load combinations and resistance provisions; ASD is not simply one blanket factor applied to otherwise unfactored loads.
- AISC 360 — Specification for Structural Steel Buildings — current AISC specification page; both LRFD and ASD design methods are incorporated.
- NTSB — Collapse of I-35W Highway Bridge, Minneapolis, Minnesota, August 1, 2007 — official investigation report referenced in section 10.
- ASCE — The Hyatt Regency Walkway Collapse — professional summary of the connection change, review failure and lessons.
- ASCE — Tacoma Narrows Bridge and aeroelastic flutter — professional overview connecting the 1940 Tacoma Narrows collapse with aeroelastic flutter.
Numeric values were reviewed in August 2026. They remain rounded educational comparisons and may differ from specific material grades, code editions or jurisdictions.
This page is a simplified educational overview, not a structural design reference. The axial stress, beam deflection and column buckling calculators use idealized elastic, small-deflection assumptions for simple standard cases (uniform rectangular sections, simply-supported spans, ideal end conditions) and do not account for material nonlinearity, lateral-torsional buckling, local buckling, connection flexibility, dynamic/seismic effects, or code-specific safety factors. The failure-lesson summaries in section 10 are necessarily condensed accounts of complex, multi-factor investigations.
Nothing on this page is validated for real structural design, permitting, or safety-related decisions. Always rely on a licensed structural engineer, a code-compliant analysis, and current design standards for real projects.