STEELWORK / CON4334
Worked examples

Column example 1: a pinned column in axial compression

Diagrams, explanations and answers are presented in reading order. No playback or response is required.

← Read this lecture example beside its chapter concepts

All answers shownChinese–English terminology
Diagrams, explanations and answers are presented in reading order. No playback or response is required.

Select a Grade S355 UC for a 6m column pinned at both ends, carrying a design axial compression of 2,500kN. Check the selected section’s compression resistance.

Original source: LectureNotes/Ch 4_Column.pdf — p. 21, p. 22.

A given value comes from the question or diagram.

A lookup value comes from the named table.

A calculated value follows from the working.

Each assumption is stated where the calculation needs it.

Known information, diagram labels and unknowns

Keep load categories separate

A floor carries permanent material and changing occupancy loads. Find the design load for the stated check.

The two boxes show G and Q with their separate factors. The plus sign joins their factored values.
The two boxes show G and Q with their separate factors. The plus sign joins their factored values.

Prediction, then explanation. Increasing imposed load changes its own term. It does not change the dead-load factor or the steel stiffness.

First convert mass to weight where required. Next collect G and Q in compatible units. Apply the combination specified by the course problem.

Wu=1.4G+1.6QW_u=1.4G+1.6Q

G is characteristic dead load; Q is characteristic imposed load. Wu is the design load for this course gravity combination.

Scope. This combination is not universal. Use the specified service load for deflection and the stated factors for other standards or cases.

Check and takeaway. Illustration: G = 10 kN and Q = 5 kN give 22 kN. Factoring their sum by one guessed factor loses the distinction.

A table value has conditions

Find a design property without mixing a row, column or standard from a different case.

The row and column meet at one cell. The table heading supplies the units and the conditions that give the cell meaning.
The row and column meet at one cell. The table heading supplies the units and the conditions that give the cell meaning.

Prediction, then explanation. A different flange thickness can select a lower design strength even when the steel grade name remains unchanged.

Match the complete section designation. Read the relevant axis and unit. For strength tables, select material, thickness and curve before interpolating.

Given values come from the question. Lookup values come from a named table. Assumed values require an explicit condition.

Scope. A missing row or omitted restraint cannot be recovered from a schematic’s scale. Keep the result conditional when required data are absent.

Check and takeaway. Record the table, row, column and unit beside the substitution. That makes a later check possible.

Use the property for the physical question

A section table contains several numbers. Choose the property that matches the calculation and bending axis.

Read the three labels beside the I-section: I describes geometric stiffness, Z elastic section modulus and S plastic section modulus.
Read the three labels beside the I-section: I describes geometric stiffness, Z elastic section modulus and S plastic section modulus.

Prediction, then explanation. Moving equal area away from an axis increases I about that axis. It does not increase total area.

Each small area contributes its area multiplied by squared distance. Sum those contributions to obtain I.

I=∫y2 dA,r=I/AI=\int y^2\,dA,\qquad r=\sqrt{I/A}

I is second moment of area in mm⁴. y is distance from the chosen axis in mm. A is area in mm². r is radius of gyration in mm.

Scope. Z is elastic section modulus; S is plastic section modulus in this course. Neither is interchangeable with I.

Check and takeaway. A nominal section designation is a row identifier. Use the actual dimensions and units in that exact row.

Apply the reasoning to this section’s data and conditions

InputSource and meaning
Design load

Given Fc=2,500kN, already factored.

No dead/imposed breakdown is needed.

RestraintsPinned-pinned, translation prevented at ends; Table 8.6 gives K=1 in both axes.
Selected rowData File p.11, 305×305×118 UC: flange T=18.7mm, bT=8.22, dt=20.6, rx=13.6cm, ry=7.77cm, A=150cm2.
Strength and curves

S355,T=18.7, giving py=345Nmm2.

Hot-rolled H-section, T≤40, so x axis uses curve b; y axis uses curve c.

Why this method applies

A strong cross-section can belong to an unstable member

A column carries axial compression and bending. Explain why more than one check is required.

The three boxes name separate checks: Section, Flexural and LTB. The bottom caption explains that passing one does not pass the others.
The three boxes name separate checks: Section, Flexural and LTB. The bottom caption explains that passing one does not pass the others.

Prediction, then explanation. At fixed moments and resistances, increasing compression increases utilisation. A section pass cannot remove the length-dependent instability checks.

Separate cross-section, flexural-buckling and LTB interaction equations. Use each equation’s own denominators and first-order or amplified moments.

AgpyA_g p_y is a section axial resistance. AgpcA_g p_c is a member compression resistance. Elastic pyZp_y Z and prescribed McM_c are different bending denominators.

Scope. Preserve the course barred-moment definitions and Eq. 8.81 convention. Do not substitute plastic moduli into an elastic denominator.

Check and takeaway. Report each requested utilisation and the governing check. The largest properly calculated ratio identifies the closest limit.

Check buckling about both axes

Determine which buckling direction limits the compressed member.

Compare the outlines labelled Straight and Bowed. The dashed line beside Bowed marks its original straight reference.
Compare the outlines labelled Straight and Bowed. The dashed line beside Bowed marks its original straight reference.

Prediction, then explanation. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Calculate slenderness about each axis. Select each axis’s curve from section type and thickness. Then obtain and compare the two compressive resistances.

λx=LEx/rx,λy=LEy/ry,Pc=pcAg\lambda_x=L_{Ex}/r_x,\quad\lambda_y=L_{Ey}/r_y,\qquad P_c=p_cA_g

Effective lengths and radii use mm. pc is compressive strength in N/mm²; Ag is gross area in mm²; Pc is in N.

Scope. Do not select the governing axis using radius alone. Effective length and buckling curve can differ. A combined column also needs interaction checks.

Check and takeaway. A short-section yield resistance is not the same as long-member buckling resistance. Keep py and pc distinct.

Apply the reasoning to this section’s data and conditions

There is no stated bending moment.

Therefore, the core numerical task is axial member buckling.

First ensure the local plates are not slender, allowing gross area.

Then calculate both axis slendernesses and select the correct curve for each.

The area alone is not enough: a long weak column can buckle well before reaching A𝗀pᵧ.

1. Choose a section and confirm its strength

Use the same mechanism: Check buckling about both axes. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Use the same mechanism: A table value has conditions. A different flange thickness can select a lower design strength even when the steel grade name remains unchanged.

Apply the reasoning to this section’s data and conditions

The lecturer selects a trial 305×305×118 UC.

Its nominal mass 118kgm is part of the section designation.

The strength calculation uses the tabulated gross area 150cm2.

The flange thickness is 18.7mm, falling within the S355 thickness band 16<T≤40mm.

py=345Nmm2Ag=150cm2×100=15,000mm2Gross-section squash resistance: Agpy=15,000×3451,000=5,175kN

5,175kN exceeds the load but is not the member resistance.

Proceed to local-slenderness and overall-buckling checks.

Mechanism detailsEffective length and buckling axes2 concepts · 2 source expressions

The complete mechanism is explained below.

Read each stage together with its stated conditions; playback and interaction are not required.

  1. A column can bow sideways before its section reaches the crushing resistance.
  2. Each axis has its own radius of gyration and restraint spacing.
  3. , with compatible length units.

    A tie affects only the directions it actually restrains.

  4. Select each buckling curve and compressive strength before forming Pc.

    The governing axis is determined by resistance, not slenderness alone.

2. Establish a non-slender section

Find which plate buckles first

Before using plastic bending resistance, check whether the flange and web can sustain the required stress distribution.

The short waves are local plate buckling. They differ from sideways movement of the whole member.
The short waves are local plate buckling. They differ from sideways movement of the whole member.

Prediction, then explanation. At fixed width, a thinner plate has a larger width-to-thickness ratio. Local buckling then becomes more restrictive.

Calculate the flange and web ratios separately. Compare each with the limits for its stress condition. The worse class governs.

ε=275/py\varepsilon=\sqrt{275/p_y}

py is the design yield strength in N/mm². ε is a dimensionless scale used in the supplied classification limits.

Scope. Pure bending and uniform compression use different web limits. Class 2 allows plastic resistance but not the Class 1 rotation capability.

Check and takeaway. A Class 1 web and Class 2 flange give a Class 2 section. Do not average the class numbers.

Apply the reasoning to this section’s data and conditions

Physical interpretation: A thin part can wrinkle first

A thin plate may wrinkle before the whole steel member reaches its intended resistance.

A thin part can wrinkle first — A thin plate may wrinkle before the whole steel member reaches its intended resistance.

Sketch, not to scale; use the original dimensions.

  1. Check flange and web slenderness using their own definitions.
  2. Compare each ratio with the correct class limits.
  3. The element with the more restrictive class determines the section class.

Remember: Bending limits and uniform-compression limits are different.

Related concept and full method

ε=275345=0.892805Semi-compact limit for a uniformly compressed flange: 13ε=11.6065bT=8.22<11.6065Semi-compact limit for a uniformly compressed web: 40ε=35.7122dt=20.6<35.7122Both pass; the section is non-slender and the gross-section compression-resistance method applies.

This establishes Class 3 or better, not necessarily a plastic flange.

The pure-compression example does not need plastic bending resistance; checking the non-slender boundary is sufficient for its stated method.

Mechanism detailsWhy thin elements buckle locally1 concept · 1 source expressions

The complete mechanism is explained below.

Read each stage together with its stated conditions; playback and interaction are not required.

  1. The flange outstand and web have different widths, thicknesses and edge support conditions.
  2. A thinner plate can wrinkle locally before the complete member loses stability.
  3. Class 1 allows plastic rotation.

    Class 2 reaches plastic resistance.

    Class 3 reaches elastic resistance.

    Class 4 requires effective properties.

  4. Check every relevant compression element with the supplied limits and stress distribution.

    The deformation shown is qualitative.

3. Calculate slenderness about both axes

Use the same mechanism: Check buckling about both axes. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Apply the reasoning to this section’s data and conditions

Physical interpretation: Why a long column can fail before crushing

Push a long thin ruler from both ends: it may bow sideways first.

Why a long column can fail before crushing — Push a long thin ruler from both ends: it may bow sideways first.

Sketch, not to scale; use the original dimensions.

  1. Find effective length and radius of gyration for each axis.
  2. Calculate slenderness for both directions.
  3. Use the appropriate buckling curve before forming resistance.

Remember: Compare the final resistances; slenderness alone may not identify the controlling axis.

Related concept and full method

LEx=LEy=1×6,000=6,000mmrx=13.6cm=136mm;ry=7.77cm=77.7mmλx=6,000136=44.117647λy=6,00077.7=77.220077

The smaller weak-axis radius makes its slenderness much larger.

Still read both curves rather than assuming the larger slenderness by itself proves which resistance controls.

Mechanism detailsEffective length and buckling axes1 concept · 1 source expressions

The complete mechanism is explained below.

Read each stage together with its stated conditions; playback and interaction are not required.

  1. A column can bow sideways before its section reaches the crushing resistance.
  2. Each axis has its own radius of gyration and restraint spacing.
  3. , with compatible length units.

    A tie affects only the directions it actually restrains.

  4. Select each buckling curve and compressive strength before forming Pc.

    The governing axis is determined by resistance, not slenderness alone.

4. Read and interpolate the two compressive strengths

Use the same mechanism: Check buckling about both axes. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Move between two known values

The required slenderness lies between two table rows. Estimate the strength on the straight segment between those rows.

xy(70, 200)(74, 192)(80, 180)70748040% of input interval → 40% of output change
The marked position lies partway along the horizontal interval. Use the same fraction of the vertical change.

Prediction, then explanation. If the second strength is lower, moving toward that row must reduce the interpolated strength.

Subtract the lower input to find distance travelled. Divide by the input interval. Multiply that fraction by the output change.

f=x−x0x1−x0,y=y0+f(y1−y0)f=\frac{x-x_0}{x_1-x_0},\qquad y=y_0+f(y_1-y_0)

x is the required input. The two known pairs are (x₀, y₀) and (x₁, y₁). f has no unit.

Scope. Use one table, curve and strength column. This interpolation does not permit extrapolation or inventing an intermediate steel section.

Check and takeaway. Illustration: between strengths 200 and 180, a fraction of 0.4 gives 192 N/mm². The result lies between both endpoints.

Apply the reasoning to this section’s data and conditions

Physical interpretation: A table lookup needs several keys

A table is like an address: knowing only the street is not enough.

A table lookup needs several keys — A table is like an address: knowing only the street is not enough.

Sketch, not to scale; use the original dimensions.

  1. Select the curve using section type, thickness and axis.
  2. Select the material-strength column.
  3. Bracket the slenderness and interpolate if required.

Remember: Do not jump between steel grades or extrapolate past supplied data.

Related concept and full method

Table 8.7, rolled H section with maximum thickness 18.7≤40mm: x–x uses 8.8(b), y–y uses 8.8(c).

In each use the pᵧ=345 column of Data File p.8.

Major axis, curve b: 44 row gives 302,46 row gives 298Nmm2. Interpolation fraction: 44.117647−4446−44=0.058824pc=302+0.058824×(298−302)=301.764706Nmm2Minor axis, curve c: 76 row gives 196,78 row gives 191Nmm2. Interpolation fraction: 77.220077−7678−76=0.610039pc=196+0.610039×(191−196)=192.949807Nmm2

The source rounds these to 302 and 193Nmm2.

The minor axis has the smaller pc and controls for the same area.

Mechanism detailsEffective length and buckling axes2 concepts · 2 source expressions

The complete mechanism is explained below.

Read each stage together with its stated conditions; playback and interaction are not required.

  1. A column can bow sideways before its section reaches the crushing resistance.
  2. Each axis has its own radius of gyration and restraint spacing.
  3. , with compatible length units.

    A tie affects only the directions it actually restrains.

  4. Select each buckling curve and compressive strength before forming Pc.

    The governing axis is determined by resistance, not slenderness alone.

5. Form resistances and conclude

Use the same mechanism: Check buckling about both axes. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Apply the reasoning to this section’s data and conditions

Pcx=15,000×301.7647061,000=4526.471kNPcy=15,000×192.9498071,000=2894.247kNPc=min(Pcx,Pcy)=2894.247kNFc=2,500<Pc; adequate under the specified axial compression. Utilisation: 2,5002894.247=0.8638

Using the lecturer’s rounded 193 gives 150×19310=2,895kN, agreeing with the original answer.

The governing failure mode is minor-axis overall buckling, not yielding of the full gross area.

Mechanism detailsEffective length and buckling axes1 concept · 2 source expressions

The complete mechanism is explained below.

Read each stage together with its stated conditions; playback and interaction are not required.

  1. A column can bow sideways before its section reaches the crushing resistance.
  2. Each axis has its own radius of gyration and restraint spacing.
  3. , with compatible length units.

    A tie affects only the directions it actually restrains.

  4. Select each buckling curve and compressive strength before forming Pc.

    The governing axis is determined by resistance, not slenderness alone.

Result and concise exam wording

Use the same mechanism: A strong cross-section can belong to an unstable member. At fixed moments and resistances, increasing compression increases utilisation. A section pass cannot remove the length-dependent instability checks.

Use the same mechanism: Check buckling about both axes. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Apply the reasoning to this section’s data and conditions

Select S355 305×305×118 UC; T=18.7, giving py=345.

bT=8.22<13ε,dt=20.6<40ε, non-slender.LxE=LyE=6m,λx=44.1176,λy=77.2201.

Curves b/c give respectively pcx=301.765, pcy=192.950Nmm2.

Pc=2,894.247kN>2,500kN, adequate.

Rounded source result 2,895kN.

Misconceptions and checks

Use the same mechanism: Check buckling about both axes. At fixed radius, longer effective length raises slenderness. The supplied buckling curve generally gives lower compressive strength.

Apply the reasoning to this section’s data and conditions

  • Do not use 355Nmm2 solely because the steel is S355.
  • Do not halve the effective length just because both ends are supported.
  • Do not use the same strut curve for both UC axes.
  • Do not treat the lower reaction arrow as a second applied load.

How the method transfers to a new variant

Vertical bending can trigger sideways motion

A beam has enough local bending resistance. Check whether its compression flange can move sideways and twist.

The dashed straight line is the initial reference. The curved flange lines illustrate the labelled sideways movement and twist.
The dashed straight line is the initial reference. The curved flange lines illustrate the labelled sideways movement and twist.

Prediction, then explanation. Adding a valid lateral restraint shortens a stability segment. It need not change the beam’s vertical reaction or moment diagram.

Identify the compressed flange and effective restraints. Calculate the specified slenderness, read buckling strength and compare resistance with the segment demand.

λ=LE/ry,Mb=pbSx  (Class 1/2),mLTMmax≤Mb\lambda=L_E/r_y,\qquad M_b=p_bS_x\;\text{(Class 1/2)},\qquad m_{LT}M_{max}\le M_b

LE and ry use the same length unit. pb uses N/mm²; Sx uses mm³; Mb uses N·mm. mLT is dimensionless.

Scope. Use the specified section class, effective-length branch and loading condition. Load height alone does not prove a destabilizing case.

Check and takeaway. Removing a lateral restraint requires a new segment stability check. The old shorter-segment result cannot establish the longer segment’s adequacy.

A force couple carries moment

Separate the force couple used to carry moment from the force used to carry web shear.

The flange forces act in opposite directions. Their separation z gives the lever arm. The web arrow represents shear.
The flange forces act in opposite directions. Their separation z gives the lever arm. The web arrow represents shear.

Prediction, then explanation. For the same moment, a larger lever arm requires a smaller couple force. It does not remove the web shear.

Moment is force multiplied by perpendicular separation. Divide the required moment by the specified separation to obtain each couple force.

F=M/zF=M/z

M is moment in kN·mm; z is the specified force-line separation in mm; F is force in kN.

Scope. Use the lever arm defined by the source model. Do not silently replace it with overall section depth or another geometric distance.

Check and takeaway. Check flange welds for the couple forces and web welds for the stated shear model. Both force paths need a satisfactory detail.

Use the same mechanism: A strong cross-section can belong to an unstable member. At fixed moments and resistances, increasing compression increases utilisation. A section pass cannot remove the length-dependent instability checks.

Apply the reasoning to this section’s data and conditions

  1. Identify axial load, any moments, actual length and restraints in each axis.
  2. Choose/read a section and confirm strength from the actual thickness.
  3. Check the appropriate flange and web local-slenderness limits.
  4. Find λx and λy separately.

    Select each axis curve and interpolate pc.

  5. Calculate both axial resistances and identify the governing axis.
  6. If moments exist, complete section, flexural and axial/LTB interactions using their own capacities and moment definitions.

Prediction with explained answer

Use the same mechanism: Find which plate buckles first. At fixed width, a thinner plate has a larger width-to-thickness ratio. Local buckling then becomes more restrictive.

Use the same mechanism: A strong cross-section can belong to an unstable member. At fixed moments and resistances, increasing compression increases utilisation. A section pass cannot remove the length-dependent instability checks.

Apply the reasoning to this section’s data and conditions

Prediction and explained answer.

Invented variant: the same column now carries 3,000kN design axial compression, with the same length and restraints.

Does it pass?

Answer and reasoning

No. Non-slender classification and buckling resistance are unchanged.

Utilisation=3,0002,894.247≈1.0365>1.

The gross squash resistance 5,175kN cannot override the member buckling failure.

Mechanism detailsEffective length and buckling axes2 concepts · 1 source expressions

The complete mechanism is explained below.

Read each stage together with its stated conditions; playback and interaction are not required.

  1. A column can bow sideways before its section reaches the crushing resistance.
  2. Each axis has its own radius of gyration and restraint spacing.
  3. , with compatible length units.

    A tie affects only the directions it actually restrains.

  4. Select each buckling curve and compressive strength before forming Pc.

    The governing axis is determined by resistance, not slenderness alone.