Chapter 4 - Columns and Other Compression Members
Chapter 4 - Columns and Other Compression Members
I. Introduction
Structural members subjected to axial compressive loads are often called by names identifying
their functions. Of these, the best-known are columns, the main vertical compression members
in building frame. Other common compression members include chords in trusses and bracing
members in frames.
For building columns, W shape having nominal depths of 350mm or less are commonly used.
These sections, being rather square in shape, are more efficient than others for carrying
compressive loads. (Deeper sections are more efficient as beams). Built-up sections are needed
in large structures for very heavy loads that cannot be supported by individual rolled shapes.
For bracing and compression members in light trusses, single-angle members are suitable.
However, equal-leg angles may be more economical than unequal-leg angles because their
least radius of gyration are greater for the same steel area.
For top chord members of bolted or welded roof trusses, a pair of angles (usually unequal, with
long legs back-to-back) are used with or without gusset plates. Structural tees may also be used
as top chord members for welded roof trusses.
Column design and analysis are based on the Euler buckling load theory, (Leonard Euler, 1757).
His analysis is based on the differential equation of the elastic curve. However, specific factors
of safety and slenderness ratio limitations are applied from purely theoretical concepts.
When a column is loaded to the buckling load or Euler load, a column will fail by sudden
buckling or bending.
If the column is hinged at both ends, the Euler Critical Load is given as:
𝜋 " 𝐸𝐼
𝑃! = "
𝐿
These equations show that the buckling stress is not a function of material strength. Rather, it is
a function of the ratio L/r known as the slenderness ratio, SR. As the slenderness ratio
increases, the buckling stress decreases, meaning that as the column becomes longer and
slenderer, the load that cause buckling becomes smaller.
Real columns do not have pin-connected ends. The restraints placed on a column’s ends greatly
affect its stability. To counter these effects, an effective length factor, K, is used to modify the
unbraced length. The product kL is called the effective length of the column. This length
approximates the length over which the column actually buckles and this may be shorter or
longer than the actual unbraced length.
Steel columns are usually subdivided into two groups: Long and Intermediate Columns,
depending on their slenderness ratio. These are two values of slenderness ratios, SRx and SRy
𝐾# 𝐿#
𝑆𝑅# =
𝑟#
𝐾$ 𝐿$
𝑆𝑅$ =
𝑟$
The critical slenderness ratio/compression index, Cc, corresponds to the upper limit of elastic
buckling failure, which is defined by an average column stress equal to 0.445Fy. By substituting
0.445Fy to Buckling Stress, Fe.
𝐸
𝐶% = 4.71?
𝐹$
User Note (NSCP 2015): The limiting slenderness ratio KL/r for members whose design is based
on compressive force preferable shall not exceed 200.
V. Design of Members for Compression
Section 505.3 Compressive Strength for Flexural Buckling of Members without Slender
Elements
Local buckling of a plate element in a rolled shape or built-up compression member may occur
before Euler buckling. The ability of plate sections to carry compressive loads without buckling
is determined by the width-thickness ratio, b/t.
Compression members are divided into stiffened elements and unstiffened elements.
Column loads transmitted to masonry and concrete foundations must meet the same bearing
pressure limitations as beam loads.
Note: In most cases, it is desirable to have m = n. This will minimize the required plate thickness
making it most economical column base plate design.
a. For ASD
𝑀𝑐 6𝑀
𝐹& = = "
𝐼 𝑏𝑑
where,
b = 1 mm or 1 unit
d = plate thickness
M = moment at critical section
𝑥 𝑥"
𝑀 = 𝑓' (𝑥) F H = 𝑓'
2 2
𝑥"
6(𝑓' 2 ) 3𝑓' 𝑥 "
𝐹& = = "
1(𝑡 " ) 𝑡
3𝑓'
𝑡 = 𝑥?
𝐹&
where,
Fb = 0.60Fy
fp = Pu / A or Pu / BN
b. For LRFD (NSCP 2015 - Appendix – 1.7)
Z = t2 / 4
𝑥 𝑥"
𝑀 𝑜𝑟 𝑀( = 𝑓' (𝑥) F H = 𝑓'
2 2
2𝑓'
𝑡 = 𝑥?
𝐹&
where,
Fb = 0.90Fy
fp = Pu / A or Pu / BN
1
𝑛) = K𝑑𝑏*
4