Design of Compression Members

Introduction

Compression members are structural elements that carry predominantly axial compressive force. Columns, posts, struts, diagonal braces, and the individual components of trusses are common examples. Unlike a tension member, whose strength is usually governed by yielding, fracture, or connection failure, a compression member may lose stability suddenly by buckling before the material reaches its yield stress.

The design of compression members therefore requires two simultaneous lines of reasoning. First, the cross-section must have adequate resistance to yielding, local buckling, and section instability. Second, the complete member must remain stable over its unsupported length. The design is consequently controlled not only by steel grade and area, but also by geometry, effective length, end restraint, bracing, residual stress, imperfections, and the axis about which buckling occurs.

This chapter presents the principles used in design standards such as EN 1993-1-1, AISC 360, and comparable national codes. Exact resistance factors, buckling curves, section classifications, and interaction equations must always be taken from the governing code adopted for a project. The equations below are presented as a consistent educational framework; they are not a substitute for project-specific code provisions.

Learning Objectives

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

1. Explain the difference between yielding, local buckling, flexural buckling, torsional buckling, and flexural–torsional buckling.
2. Calculate the slenderness parameters of solid and built-up compression members.
3. Determine the influence of effective length and end restraint on elastic buckling resistance.
4. Select and apply an appropriate design buckling curve or column curve.
5. Design simple columns, struts, and built-up members, including checks on components and connectors.
6. Identify when bracing is required and explain how bracing affects stability.
7. Interpret design results responsibly within a code-based engineering workflow.

1. Structural Behaviour of Compression Members

An ideal pin-ended, perfectly straight, centrally loaded column would remain straight until the Euler critical load was reached. Real members are never ideal. They contain initial crookedness, residual stresses from rolling or welding, load eccentricity, imperfect connections, and sometimes local plate imperfections. These effects cause bending from the beginning of loading and reduce the resistance below the ideal Euler value.

A compression member can fail through several related mechanisms:

• Material yielding or crushing: the average compressive stress becomes too large.
• Local buckling: a slender plate element of the cross-section buckles before the whole member bends.
• Flexural buckling: the member deflects laterally about its weak or strong principal axis.
• Torsional buckling: the member twists about its longitudinal axis.
• Flexural–torsional buckling: lateral displacement and twist occur together.
• Connection or splice failure: the built-up member or its end connection cannot transfer the required force.

The lowest relevant resistance generally governs. A satisfactory design must therefore consider both the cross-section and the member as a structural system.

2. Terminology and Section Properties

The following properties are central to compression-member design:

The radius of gyration is defined by

where I is the relevant second moment of area and A is the area. A larger radius of gyration indicates that area is distributed farther from the buckling axis and generally gives greater resistance to flexural buckling.

For a prismatic member, the geometric slenderness about an axis is

where L_eff is the effective length. The axis producing the larger value of λ_g is usually the critical axis for flexural buckling, although torsional and local modes must also be checked.

3. Euler Elastic Buckling

For a straight, uniform, pin-ended member under concentric axial compression, Euler’s elastic critical load is

Here N_cr is the elastic critical load, E is Young’s modulus, I is the second moment of area about the buckling axis, and L_eff = KL. The equation assumes linear elasticity, ideal geometry, uniform axial loading, and idealized end restraint.

The corresponding critical stress is

The equation is dimensionally consistent: E has units of force per area, while KL/i is dimensionless; therefore σ_cr also has units of force per area.

Euler’s result is most useful for slender members. For stocky members, the predicted elastic buckling stress may exceed the yield strength, which is physically impossible in an elastic model. Real design curves therefore transition from squash or yield resistance for stocky members to Euler-type resistance for slender members.

Effective Length

The effective-length factor represents end restraint and the surrounding structural system. Approximate ideal values include:

These values are idealizations. In frames, effective length depends on joint stiffness, beam-to-column stiffness ratios, sway behaviour, and whether the frame is braced against lateral translation. A designer should not select K from a table without considering the actual restraint and frame analysis.

A lower K increases N_cr because resistance varies with 1/K². Overestimating restraint, however, can be unsafe. Where the restraint is uncertain, a conservative value or a direct second-order analysis should be used.

4. Non-Dimensional Slenderness and Design Buckling Curves

A useful non-dimensional slenderness is

where A f_y is the squash load and N_cr is the Euler elastic critical load for the relevant buckling mode. Equivalently, using σ_cr = N_cr/A,

Thus λ̄ compares the material squash resistance with elastic instability resistance. A small λ̄ indicates a stocky member; a large λ̄ indicates a slender member.

A common form of an imperfection-sensitive reduction factor is

followed by

and

In this formulation, α is an imperfection factor associated with the selected buckling curve, χ is the reduction factor, and γ_M1 is the relevant partial safety factor. The numerical limits and curve selection are code dependent. Some standards use different symbols, reference slenderness limits, or resistance factors.

The design procedure is therefore:

1. Determine A, I, and i for each principal axis.
2. Establish the effective length for each mode.
3. Calculate N_cr and λ̄.
4. Select the correct buckling curve based on cross-section type, fabrication method, axis, and code.
5. Calculate χ and the design resistance.
6. Check local, torsional, connection, and second-order effects.

5. Local Buckling of Cross-Section Elements

A column may fail locally even when its overall length is short. Local buckling occurs when a plate element—such as a flange, web, angle leg, or hollow-section wall—has a large width-to-thickness ratio. Once local buckling occurs, the full gross area cannot safely be assumed to carry compression uniformly.

Code classification systems generally distinguish compact, non-compact, and slender elements. The classification is based on a width-to-thickness ratio, often compared with a limit involving the material parameter

where f_y is in N/mm² and 235 N/mm² is the reference value used in many European classification provisions. The exact limits depend on the element type, stress distribution, and design code.

For slender elements, an effective width or effective area method is used. The effective area A_eff is less than the gross area A, and the design resistance may be represented generally as

The appropriate safety factor and effective-width calculation must be obtained from the governing standard. Local buckling checks are particularly important for thin-walled cold-formed sections, welded box sections, angles with slender legs, and built-up plates.

Local buckling and overall buckling interact. A section with reduced effective area also has altered stiffness and radius of gyration. Consequently, local buckling should be assessed before applying an overall member buckling equation.

6. Flexural Buckling

Flexural buckling is lateral bending about a principal axis. For doubly symmetric sections such as universal columns, it is usually checked about both the major and minor axes. The minor axis frequently governs because I_z and i_z are smaller.

The elastic critical loads are

and

The corresponding non-dimensional slenderness values are calculated independently. The lower design resistance governs, but the member must also be checked for any mode with a lower critical load, including torsional modes.

The buckling curve should reflect the section. Rolled I-sections, welded I-sections, hollow sections, angles, and built-up members may have different imperfection sensitivities. Using a single curve for all sections can be unconservative or unnecessarily conservative.

7. Torsional and Flexural–Torsional Buckling

Torsional buckling involves twisting without significant lateral translation. It is important for sections with low torsional stiffness or weak warping restraint, including cruciform sections, some open built-up sections, and members with certain asymmetric arrangements.

Flexural–torsional buckling combines bending and twisting. It is especially relevant to singly symmetric or asymmetric sections, such as channels, angles, tees, and eccentrically connected built-up members. The centroid, shear centre, and principal axes do not necessarily coincide. A load applied through the centroid may still produce twisting if it does not pass through the shear centre.

A simplified elastic torsional critical-load expression for a centrally loaded member can be written as

where N_cr,T is a torsional critical load, r_0 is the polar radius of gyration about the shear centre, I_w is the warping constant, K_w L is an effective warping length, G is the shear modulus, and J is the St Venant torsion constant. This expression applies only under specified idealizations and conventions; restraint conditions and section properties must be defined consistently.

For asymmetric sections, the flexural and torsional modes are coupled. General eigenvalue analysis or the equations in the governing steel code should be used. A designer should not rely solely on the weak-axis Euler load when the section has substantial torsional flexibility.

8. Bracing and Effective Restraint

Bracing reduces the unbraced length and may restrain translation, rotation, or warping. Examples include floor diaphragms, roof bracing, transverse frames, battens, lacing, bridging, and rigid connections.

A brace is effective only if it has adequate stiffness, strength, continuity, and connection capacity. A nominal tie that cannot develop the required force does not provide reliable restraint. Bracing design should consider:

• the direction and mode of instability being restrained;
• brace stiffness and connection deformation;
• forces generated by initial imperfections and second-order effects;
• continuity through floors, roofs, and adjacent members;
• construction-stage conditions, when permanent bracing may not yet exist.

When a brace divides a member into sublengths, the effective length may be reduced for the restrained axis. The designer must still check the complete member between actual restraint points and the brace itself. Bracing also attracts force; it should not be treated as force-free.

9. Built-Up Compression Members

Built-up members consist of two or more components connected by welds, bolts, battens, or lacing. Examples include twin angles, double channels, laced columns, battened columns, and box sections assembled from plates.

Their design has two levels:

1. Member-level stability: the built-up section is checked for flexural, torsional, and flexural–torsional buckling.
2. Component-level stability: each individual component is checked between connection points, including local slenderness and connector forces.

The gross area of a built-up member is approximately the sum of component areas, less deductions required by the code. The centroid and second moments of area are calculated using the parallel-axis theorem:

where I_i is the component’s own second moment of area, A_i is its area, and d_i is the distance from its centroid to the built-up section axis.

For laced members, shear deformation and the flexibility of the lacing may increase the effective slenderness. Some codes specify an increased equivalent slenderness or an additional reduction. Battened members may behave differently because battens transfer shear through bending and connections. The appropriate code model must be used rather than assuming that the built-up member behaves exactly like a solid section.

Connections between components must transmit forces arising from axial load distribution, imperfections, transverse shear, and relative deformation. Welds and bolts should be checked for shear, bearing, rupture, block shear, eccentricity, and fatigue where relevant.

10. Worked Example: Design Check for a Rolled Column

Illustrative example. A pin-ended steel column is 3.0 m long and carries a design axial compression force of 450 kN. Its gross area is 6,400 mm². The radii of gyration are i_y = 160 mm and i_z = 42 mm. Take E = 210,000 N/mm², f_y = 355 N/mm², K = 1.0, and, for illustration, use α = 0.49 and γ_M1 = 1.0. Assume local buckling is satisfactory and that the selected buckling curve applies to both axes.

Step 1: Slenderness.

The z-axis is more slender.

Step 2: Euler load about the z-axis. Since I_z = A i_z²,

Step 3: Non-dimensional slenderness.

Step 4: Reduction factor.

Step 5: Design resistance.

Since 1459 kN exceeds the applied 450 kN, the illustrative overall z-axis buckling check is satisfactory. The y-axis should also be calculated, although its much larger radius of gyration makes it unlikely to govern. The result is not a complete design: local buckling, end plates, connections, load eccentricity, frame sway, fire design, and code-specific curve selection remain to be checked.

11. Applied Case: Braced Industrial-Frame Column

Illustrative case. A two-storey industrial building uses rolled I-section columns. During preliminary design, the engineer assumes fixed-ended columns and obtains K = 0.5. The resulting column resistance is greater than the factored axial load. Site review, however, shows that the beam-to-column joints are nominally pinned and the roof bracing is incomplete during erection.

Problem. The original design relied on restraint that was neither present in the connection details nor guaranteed during construction. The effective length and sway behaviour were therefore misrepresented.

Analysis. The engineer revises the model in two stages. During erection, the column is checked with the temporary bracing arrangement and a conservative effective length. In the completed structure, the lateral system is modelled as a braced frame only after the roof diaphragm and permanent bracing are demonstrated to have adequate stiffness and connection capacity. The column is checked about both axes, and the design is reviewed for second-order effects. The connections are also checked for the forces required to provide the assumed restraint.

Approach and outcome. Additional roof bracing and a properly detailed transverse tie are installed. The final analysis shows that the completed frame can use a shorter effective length in the restrained direction, while the erection-stage condition remains satisfactory with the temporary bracing. The revised design is slightly heavier but provides a traceable load path and safe construction sequence.

Learning points. Effective length is a structural-system property, not merely a value selected from a table. Bracing assumptions must be shown on drawings, supported by calculations, and maintained throughout construction. A design that is adequate only after completion may still be unsafe during erection.

12. Design Workflow and Professional Practice

A robust compression-member design workflow is:

1. Establish factored axial forces and, where applicable, moments and shear.
2. Identify the design situation, including construction and accidental stages.
3. Define end conditions, frame sway behaviour, and restraint points.
4. Select a trial section and calculate area, centroid, principal axes, radii of gyration, torsional constant, and warping constant where needed.
5. Check cross-section classification and local buckling.
6. Calculate flexural buckling resistance about both principal axes.
7. Check torsional and flexural–torsional buckling where applicable.
8. Check member interactions caused by eccentricity, moments, and second-order effects.
9. Design splices, lacing, battens, welds, bolts, base plates, and braces.
10. Verify serviceability, durability, fire resistance, fabrication tolerances, and constructability.
11. Document assumptions, code clauses, effective lengths, load combinations, and governing utilization ratios.

Computer software is valuable for frame analysis, eigenvalue extraction, second-order analysis, and section-property calculation. It does not remove the need for engineering judgement. A software model can produce a precise answer to an incorrect effective-length assumption. Independent hand checks and sensitivity studies remain essential.

13. Limitations, Safety, and Contemporary Practice

The simplified equations in this chapter assume idealized loading and geometry. They may not adequately represent members with severe eccentricity, significant combined compression and bending, perforated webs, local damage, fire exposure, fatigue, corrosion loss, or complex semi-rigid connections. Advanced projects may require geometrically and materially nonlinear analysis with imperfections, commonly described as GMNIA, or an equivalent code-approved method.

Ethically, the engineer must not conceal uncertainty in restraint, material properties, or construction sequence. Inspection and quality assurance should verify steel grade, weld quality, bolt installation, dimensional tolerances, and the presence of specified bracing. Design changes in the field must be reviewed because removing a brace or altering a connection can change the governing buckling mode.

Contemporary practice increasingly uses three-dimensional finite-element analysis, digital twins, and automated model checking. These tools can improve coordination and identify clashes, but they should be governed by validated assumptions, version control, and professional review. Artificial intelligence may assist with section selection or anomaly detection, but it cannot independently establish code compliance or accept responsibility for a safety-critical design.

Chapter Summary

Compression-member design is fundamentally a stability problem. The designer must distinguish cross-section yielding and local buckling from member-level flexural, torsional, and flexural–torsional buckling. Euler’s equation establishes the elastic instability basis, while non-dimensional slenderness and code buckling curves account for imperfections, residual stress, and inelastic behaviour. Effective length depends on real restraint and frame action, not simply on nominal connection labels.

A complete design checks both principal axes, local slenderness, torsional modes, bracing, component stability, and connections. Built-up members require additional checks for lacing or battens and for the stability of individual components. Worked calculations are useful only when their assumptions are stated. Construction-stage stability, documentation, inspection, and professional judgement are essential parts of safe column design.

Key Terms

Compression member: A structural member carrying predominantly axial compressive force.

Column: A generally vertical compression member in a building or frame.

Strut: A compression member, often diagonal or horizontal, used to resist axial force.

Buckling: Instability involving a change from an initially straight configuration.

Flexural buckling: Lateral bending instability about a principal axis.

Torsional buckling: Instability involving twisting about the longitudinal axis.

Flexural–torsional buckling: Coupled lateral bending and twisting instability.

Local buckling: Buckling of an individual plate element of a cross-section.

Effective length: The length of an ideal pin-ended column having the same elastic buckling load as the actual member.

Effective-length factor: The factor K relating effective length to actual length.

Radius of gyration: √(I/A), a measure of the distribution of area about an axis.

Non-dimensional slenderness: A ratio comparing squash resistance with elastic buckling resistance.

Buckling curve: A code-based relationship reducing squash resistance to account for instability.

Imperfection factor: A parameter representing geometric imperfections and residual stresses.

Bracing: A system that restrains translation, rotation, or warping.

Built-up member: A member assembled from two or more connected components.

Lacing: Diagonal or transverse elements connecting components of a built-up member.

Batten: A transverse plate or member connecting components of a built-up member.

Warping: Longitudinal deformation associated with non-uniform torsion.

Second-order effect: Additional force or moment caused by displacement under existing load.

Review Questions

Short Answer Questions

1. Why can a compression member fail by buckling before the steel reaches its yield stress?
2. Define effective length and explain the meaning of the factor K.
3. Why does the minor axis often govern the design of an I-section column?
4. Distinguish local buckling from flexural buckling.
5. State two reasons why a built-up compression member requires component-level checks.

Long Answer Questions

1. Explain the development from Euler elastic buckling to code design buckling curves, including the roles of imperfections and non-dimensional slenderness.
2. Describe a complete design procedure for a rolled steel column subjected to axial compression.
3. Discuss the conditions that may produce torsional or flexural–torsional buckling.
4. Explain how bracing influences effective length and identify the requirements for a brace to be considered reliable.
5. Compare laced and battened built-up columns, including their structural actions and design concerns.

Analytical / Application Questions

1. A 3.6 m pin-ended column has A = 5,000 mm², i_y = 145 mm, and i_z = 38 mm. Determine the geometric slenderness about each axis and identify the likely flexural buckling axis. State what additional checks are required before design approval.
2. A designer changes the assumed effective-length factor from K = 1.0 to K = 0.7 without modifying the member. Explain qualitatively how the Euler load changes and identify the evidence needed to justify the change.
3. A twin-angle strut is connected only at its ends and has widely spaced components. Develop a checklist covering overall buckling, component buckling, connection forces, torsion, and construction-stage stability.

Multiple Choice Questions with Answers

Question 1

Which parameter directly represents the influence of end restraint in the effective length of a column?

A. A

B. I

C. K

D. f_y

Correct Answer: C. K

Question 2

For an ideal pin-ended column, Euler’s elastic critical load is proportional to which quantity?

A. L²

B. 1/L²

C. A/L

D. f_y L²

Correct Answer: B. 1/L²

Question 3

The radius of gyration about an axis is defined as:

A. I/A

B. √(I/A)

C. A/I

D. I²/A

Correct Answer: B. √(I/A)

Question 4

Which instability involves twisting about the longitudinal axis?

A. Local buckling

B. Flexural buckling

C. Torsional buckling

D. Yielding

Correct Answer: C. Torsional buckling

Question 5

For a typical doubly symmetric I-section column, which axis commonly governs flexural buckling?

A. The axis with the larger radius of gyration

B. The minor axis with the smaller radius of gyration

C. The axis parallel to the load only

D. Neither principal axis

Correct Answer: B. The minor axis with the smaller radius of gyration

Question 6

In the expression λ̄ = √(A f_y/N_cr), increasing N_cr while other variables remain constant will:

A. Increase λ̄

B. Reduce λ̄

C. Make λ̄ equal to K

D. Have no effect on λ̄

Correct Answer: B. Reduce λ̄

Question 7

What is the principal purpose of an effective area in the design of a slender cross-section?

A. To increase the gross area

B. To account for local buckling

C. To eliminate torsion

D. To define the steel grade

Correct Answer: B. To account for local buckling

Question 8

Which statement about a brace is correct?

A. A brace needs no connection design

B. A nominal tie always provides full restraint

C. A brace must have adequate stiffness, strength, and connection capacity

D. Bracing affects only serviceability

Correct Answer: C. A brace must have adequate stiffness, strength, and connection capacity

Question 9

For a built-up member, the parallel-axis theorem is used primarily to determine:

A. Steel yield strength

B. The assembled section’s second moment of area

C. Bolt hole diameter

D. The member’s actual length

Correct Answer: B. The assembled section’s second moment of area

Question 10

A column with significant asymmetry and low torsional stiffness should be checked particularly for:

A. Flexural–torsional buckling

B. Concrete crushing

C. Cable sag

D. Hydraulic buckling

Correct Answer: A. Flexural–torsional buckling