Difference between SHORT and SLENDER columns

11.1 Introduction

When a column bends or deflects laterally an M2 = Mmax M2

amount Δ, its axial load will cause an increased
column moment equal to PΔ. This moment will
be superimposed onto any moments already in
the column. Should this PΔ moment be of such
magnitude as to reduce the axial load capacity
of the column significantly, the column will be
referred to as a slender column. Mc = Mmax
Section [Link] of the NSCP states that
= δM2
the design of a compression member should, M1 M1
desirably, be based on a theoretical analysis of
the structure that takes into account the
effects of axial loads, duration of loads, varying
member sizes, end conditions, and so on. If
such a theoretical procedure is not used, the M2 = Mmax M2
NSCP Section 406.6.4 provides an approximate
method for determining slenderness effects.
This method, which is based on the factors just
mentioned for an “exact” analysis, results in a
moment magnifier, 𝛿, which is to be multiplied Mc = Mmax
by the larger moment at the end of the column
denoted as M2, and that value is used in design. = δM2
If bending occurs about both axes, 𝛿 is to be
computed separately for each direction and
the values obtained have to be multiplied by
M1 M1
the respective moment values.
11.2 Nonsway and Sway Frames If we cannot tell by inspection whether we have a Q= ΣPuΔo ([Link].1)
nonsway frame or a sway frame, the code provides Vu 𝓁c
For this discussion, it is necessary to distinguish two ways of making a decision. First, in NSCP, where
between frames without sidesway and those Section 406.6.4.3a, a story in a frame is said to be a ΣPu = total factored vertical load for all of the
with sidesway. In the ACI Code, these are nonsway one if the increase in column end columns on the story in question
referred to respectively as nonsway frames and moments from second-order effects is 5% or less Δo = the elastically determined first-order
sway frames. of the first-order end moments. lateral deflection from Vu at the top of the
For the building story in question, the columns The second method for determining whether a story in question with respect to the
in nonsway frames must be designed according particular frame is braced or unbraced is given in bottom of that story
to Section [Link] of the NSCP, while the the NSCP, Section 406.6.4.3a. If the value of the so- Vu = the total factored horizontal shear for the
columns of sway frames must be designed called stability index Q is ≤ 0.05, the commentary story in question
according to Section [Link]. As a result, it is states that the frame may be classified as a 𝓁c = the height of a compression member in a
first necessary to decide whether we have a nonsway one. (Should Vu be equal to zero, this frame measured from center to center of
nonsway frame or a sway frame. You must method will not apply.) the frame joints
realize that you will rarely find a frame that is
completely braced against swaying or one that
is completely unbraced against swaying.
Therefore, you are going to have to decide
which way to handle it.
The question may possibly be resolved by
examining the lateral stiffness of the bracing
elements for the story in question. You may
observe that a particular column is located in a
story where there is such substantial lateral
stiffness provided by bracing members, shear
walls, shear trusses, and so on that any lateral Nonsway Frame Sway Frame
deflections occurring will be too small to affect Despite these suggestions from the ACI, the individual designer is going to have to make decisions as to what
the strength of the column appreciably. You is adequate bracing and what is not, depending on the presence of structural walls and other bracing items.
should realize that, while examining a particular For the average-size reinforced concrete building, load eccentricities and slenderness values will be small,
structure, there may be some nonsway stories and frames will be considered to be braced. Certainly, however, it is wise in questionable cases to err on the
and some sway stories. side of the unbraced.
11.3 Slenderness Effects Effective Length Factors Columns with different end conditions
To calculate the slenderness ratio of a particular have entirely different effective lengths. For
column, it is necessary to estimate its effective length instance, if there were such a thing as a
The slenderness of columns is based on their geometry
(or the distance between points of zero moment in the perfectly fixed end column, its points of
and on their lateral bracing. As their slenderness
column). For this initial discussion, it is assumed that inflection (or points of zero moment) would
increases, their bending stresses increase, and thus
no sidesway or joint translation is possible. Sidesway occur at its one-fourth points, and its effective
buckling may occur. Reinforced concrete columns
or joint translation means that one or both ends of a length would be 𝓁u∕2, as shown in Figure
generally have small slenderness ratios. As a result, they
column can move laterally with respect to each other. 11.1(b). As a result, its k value would equal
can usually be designed as short columns without
0.5.
strength reductions because of slenderness. If If there were such a thing as a perfectly pinned
slenderness effects are considered small, then columns end column, its effective length would be its Obviously, the smaller the effective length
can be considered “short” and can be designed according unsupported length, as shown in Figure 11.1(a). The of a particular column, the smaller its danger
to Chapter 10 (“Design of Short Columns Subjected to effective length factor, k, is the number that must be of buckling and the greater its load-carrying
Axial load and Bending”). However, if they are “slender,” multiplied by the column’s unsupported length to capacity. Figure 11.1(c) shows a column with
the moment for which the column must be designed is obtain its effective length. For a perfectly pinned end one end fixed and one end pinned. The k
increased or magnified. Once the moment is magnified, column, k = 1.0. factor for this column is approximately 0.70.
the column is then designed according to Chapter 10
using the increased moment.
Calculation of slenderness ratios include
unsupported column lengths, effective length factors,
radii of gyration, and the ACI Code requirements. The
NSCP, Section 406.2.6 limits second-order effects to not
more than 40% of first-order effects.
Unsupported Lengths
The length used for calculating the slenderness ratio of a
column, 𝓁u, is its unsupported length. This length is
considered to be equal to the clear distance between
slabs, beams, or other members that provide lateral
support to the column. If haunches or capitals (see Figure
16.1 in Chapter 16) are present, the clear distance is
measured from the bottoms of the capitals or haunches.
11.3 Slenderness Effects (Cont.)
The concept of effective lengths is simply An example of an unbraced column In Figure 11.2(b), another unbraced column
a mathematical method of taking a column— is shown in Figure 11.2(a). The base of case is illustrated. The bottom of this column is
whatever its end and bracing conditions— this particular column is assumed to be connected to beams that provide resistance to
and replacing it with an equivalent pinned fixed, whereas its upper end is assumed rotation but not enough to be considered a fixed
end-braced column. A complex buckling to be completely free to both rotate and end. In most buildings, partial rotational restraint
analysis could be made for a frame to translate. The elastic curve of such a is common, not pinned or fixed ends. Section 11.4
determine the critical stress in a particular column will take the shape of the elastic shows how to evaluate such partial restraint. For
column. The k factor is determined by finding curve of a pinned-end column of twice the case shown in Figure 11.2(b), if the beam at
the pinned end column with an equivalent its length. Its effective length will the bottom is flexible compared with the column,
length that provides the same critical stress. therefore equal 2𝓁u, as shown in the the k factor approaches infinity. If it is very stiff, k
The k factor procedure is a method of making figure. approaches 2.
simple solutions for complicated frame-
buckling problems.
Reinforced concrete columns serve as
parts of frames, and these frames are
sometimes braced and sometimes unbraced.
A braced frame is one for which sidesway or
joint translation is prevented by means of
bracing, shear walls, or lateral support from
adjoining structures. An unbraced frame does
not have any of these types of bracing
supplied and must depend on the stiffness of
its own members to prevent lateral buckling.
For braced frames, k values can never be
greater than 1.0, but for unbraced frames, the
k values will always be greater than 1.0
because of sidesway.
 Table of Effective length factor, k
11.3 Slenderness Effects (Cont.)
Section [Link].3 of the NSCP
states that the effective length factor k is
permitted to be taken as 1.0 for
compression members in frames braced
against sidesway unless a theoretical
analysis shows that a lesser value can be
used.
Should the member be in a frame
not braced against sidesway, the value of
k will be larger than 1.0 and must be
determined with proper consideration
given to the effects of cracking and
reinforcement on the column stiffness.
ACI-ASCE Committee 441 suggests that
it is not realistic to assume that k will be
less than 1.2 for such columns;
therefore, it seems logical to make
preliminary designs with k equal to or
larger than 1.2.
Effective length factor, k, for braced columns
0.5 ≤ k ≤ 1.0
Effective length factor, k, for unbraced columns
1.0 ≤ k < ∞
11.4 Determining k Factors with
Alignment Charts
The preliminary procedure used for estimating effective lengths
involves the use of the alignment charts shown in the figure. Before
computerized analysis, use of such alignment charts was the traditional
method for determining effective lengths of columns. The chart of part
(a) of the figure is applicable to braced frames, whereas the chart of
part (b) is applicable to unbraced frames.
The 𝜓 factors used to enter the alignment charts, and thus the
resulting effective length factors, are dependent on the relative
stiffnesses of the compression and flexural members. If we have a very
light flexible column and large stiff girders, the rotation of the column
ends will be greatly minimized. The column ends will be closed to a
fixed condition, and thus the 𝜓 values and the resulting k values will be
small. Obviously, if the reverse happens—that is, large stiff columns
framing into light flexible girders—the column ends will rotate almost
freely, approaching a pinned condition. Consequently, we will have
large 𝜓 and k values.

FIGURE 11.3 Effective length factors.

𝜓 = (possible pronunciation – /sī/ ) - ratio of Σ(EI∕𝓁) of compression members to
Σ(EI∕𝓁) of flexural members in a plane at one end of a compression member.
k = effective length factor.
Note: Should one end of the column be fixed, 𝜓 is theoretically equal to 0, and if pinned,
𝜓 = ∞. Because a perfectly fixed end is practically impossible to have, 𝜓 is usually
taken as 1.0 instead of 0 for assumed fixed ends. When column ends are
supported by but not rigidly connected to a footing, 𝜓 is theoretically infinite but
usually is taken as about 10 for practical design.
11.4 Determining k Factors with
Alignment Charts
The preliminary procedure used for estimating effective lengths
involves the use of the alignment charts shown in the figure. Before
computerized analysis, use of such alignment charts was the traditional
method for determining effective lengths of columns. The chart of part
(a) of the figure is applicable to braced frames, whereas the chart of
part (b) is applicable to unbraced frames.
The 𝜓 factors used to enter the alignment charts, and thus the
resulting effective length factors, are dependent on the relative
stiffnesses of the compression and flexural members. If we have a very
light flexible column and large stiff girders, the rotation of the column
ends will be greatly minimized. The column ends will be closed to a
fixed condition, and thus the 𝜓 values and the resulting k values will be
small. Obviously, if the reverse happens—that is, large stiff columns
framing into light flexible girders—the column ends will rotate almost
freely, approaching a pinned condition. Consequently, we will have
large 𝜓 and k values.
To calculate the 𝜓 values, it is necessary to use realistic moments of
inertia. Usually the girders will be appreciably cracked on their tensile
sides, whereas the columns will probably have only a few cracks. If the I
values for the girders are underestimated a little, the column k factors FIGURE 11.3 Effective length factors.
will be a little large and thus on the safe side. 𝜓 = (possible pronunciation – /sī/ ) - ratio of Σ(EI∕𝓁) of compression members to
Several approximate rules are in use for estimating beam and Σ(EI∕𝓁) of flexural members in a plane at one end of a compression member.
column rigidities. One common practice of the past for slenderness k = effective length factor.
ratios of up to about 60 or 70 was to use gross moments of inertia for 0 ∞
Σ(EI/L)C Σ(EI/L)C
the columns and 50% of the gross moments of inertia for the beams. 𝜓= 𝜓=
NSCP Table [Link].1a recommends values for moment of Σ(EI/L)G ∞ Σ(EI/L)G 0
inertias and cross sectional areas for determining 𝜓 values for use in recommended recommended
evaluating k factors. This is discussed in Section 11.6. FIXED END 𝜓 = 1. PINNED END 𝜓 = 10.
11.5 Determining k Factors with Equations As mentioned in Section 11.3 of this chapter, Section [Link].3 of the
Instead of using the alignment charts for determining k values, an NSCP states that k may conservatively be taken to be 1.0 for compression
alternate method involves the use of relatively simple equations. members in frames braced against sidesway unless a theoretical analysis
These equations, which were in the ACI 318-05 Code Commentary shows that a lesser value can be used. In the last paragraph of Section
(R10.12.1) and taken from the British Standard Code of Practice are R6.[Link] of the commentary, use of the alignment charts or the equations just
particularly useful with computer programs. presented is said to be satisfactory for justifying k values less than 1.0 for
For braced compression members, an upper bound to the braced frames.
effective length factor may be taken as the smaller value determined
from the two equations to follow in which 𝜓A and 𝜓B are the values
just described for the alignment charts (commonly called the
Jackson and Moreland alignment charts). 𝜓min is the smaller of 𝜓A
and 𝜓B.
k = 0.7+0.05(𝜓A + 𝜓B) ≤ 1.0
k = 0.85 + 0.05 𝜓min ≤ 1.0
The value of k for unbraced compression members restrained at
both ends may be determined from the appropriate one of the
following two equations, in which 𝜓m is the average of 𝜓A and 𝜓B :
If 𝜓m < 2,
k = (20 − 𝜓m) 1+ 𝜓m/20
If 𝜓m ≥ 2,
k = 0.9 1+ 𝜓m
The value of the effective length factor of unbraced compression
members that are hinged at one end may be determined from the
following expression, in which 𝜓 is the value at the restrained end:
k = 2.0 + 0.3𝜓 FIGURE 11.3 Effective length factors.
11.6 First-Order Analyses Using Special (b) Moments of inertia where Ig = moment of For continuous flexural members, it
Member Properties inertia of gross concrete section about is permitted to use the average
centroidal axis neglecting reinforcement value of I from the positive and
After this section, the remainder of this chapter is (NSCP, Section [Link].1a). negative moment sections. In no
devoted to an approximate design procedure wherein the case is the value of I for flexural
effect of slenderness is accounted for by computing Table [Link].1(a)
members required to be taken less
moment magnifiers that are multiplied by the column Moment of Inertia and Cross-Sectional
than 0.25Ig.
moments. A magnifier for a particular column is a Area Permitted for Elastic Analysis at
function of its factored axial load, Pu, and its critical Factored Load Level Often during the design process,
buckling load, Pc. Member Moment Cross- the designer does not know the
and condition of Sectional final values of member section
Before moment magnifiers can be computed for a dimensions or steel areas when
particular structure, it is necessary to make a first-order Inertia Area
Columns 𝟎. 𝟕𝟎𝑰g making calculations such as those
analysis of the structure. The member section properties in NSCP Table [Link].1(b). This
used for such an analysis should take into account the Uncracked 𝟎. 𝟕𝟎𝑰g
Walls leads to an iterative process where
influence of axial loads, the presence of cracked regions Cracked 𝟎. 𝟑𝟓𝑰g 𝟏. 𝟎𝑨g
the last cycle of iteration assumes
in the members, and the effect of the duration of the Beams 𝟎. 𝟑𝟓𝑰g
the same member properties as the
loads. Instead of making such an analysis, NSCP, Section Flat plates and flat slabs 𝟎. 𝟐𝟓𝑰g
final design. The NSCP, Section
[Link].1 permits use of the following properties for [Link] allows these values to be
Table [Link].1(b)
the members of the structure. These properties may be only within 10% of the final values
used for both nonsway and sway frames. Alternative Moment of Inertia for Elastic
Analysis at Factored Load in the final iteration.
(a) Modulus of elasticity determined from the
expressions given in Section [Link] of the NSCP: Alternative Value of 𝑰 for Elastic Analysis
Member Minimum 𝑰 Maximum
[Link] Modulus of elasticity, 𝑬𝒄 , for concrete shall 𝑨𝒔𝒕 𝑴𝒖 𝑷𝒖
Columns and walls 𝟎. 𝟑𝟓𝑰g 𝟎.𝟖𝟎 + 𝟐𝟓 𝟏− − 𝟎.𝟓 𝑰 𝟎. 𝟖𝟕𝟓𝑰g
be permitted to be calculated as (a) or (b): 𝑨g 𝑷𝒖𝒉 𝑷𝒐 g
(a) For values of wc between 1440 and 2560 kg/m3 Beams, flat plates, 𝒃𝒘
𝟎. 𝟐𝟓𝑰g 𝟎. 𝟏𝟎 + 𝟐𝟓𝝆 𝟏. 𝟐 − 𝟎. 𝟐 𝑰 𝟎. 𝟓𝑰g
𝑬𝒄 = 𝒘𝒄 𝟏.𝟓 𝟎. 𝟎𝟒𝟑 (in MPa)
𝒇′𝒄 ([Link].a) and flat slabs 𝒅 g
(b) For normal weight concrete Note: For continuous flexural members, 𝑰 shall be permitted to be taken as the average of values obtained for the
critical positive and negative moment sections. 𝑷𝒖 and 𝑴𝒖 shall be calculated from the load combination
𝑬𝒄 = 𝟒, 𝟕𝟎𝟎 𝒇′𝒄 (in MPa) ([Link].b) under consideration, or the combination of 𝑷𝒖 and 𝑴𝒖 that produces the least value of 𝑰.
11.7 Slender Columns in Nonsway For nonsway frames, slenderness effects may Avoiding Slender Columns
and Sway Frames be ignored if the following expression is satisfied: The design of slender columns is appreciably more
k𝓁u M1 complicated than the design of short columns. As a
≤ 34 - 12 (406.2.5b)
There is a major difference in the behavior of r M2 result, it may be wise to give some consideration to
columns in nonsway or braced frames and those In this expression, M1 is the smaller factored end the use of certain minimum dimensions so that
in sway or unbraced frames. In effect, each moment in a compression member. It has a plus sign if none of the columns will be slender. In this way,
column in a braced frame acts by itself. In other the member is bent in single curvature (C shaped) and they can be almost completely avoided in the
words, its individual strength can be determined a negative sign if the member is bent in double average-size building.
and compared to its computed factored loads and curvature (S shaped). M2 is the larger factored end
moments. In an unbraced or sway frame, a If k is assumed to be equal to 1.0, slenderness can
moment in a compression member, and it always has
column will probably not buckle individually but usually be neglected in braced frame columns, if
a plus sign. In this equation, the term
will probably buckle simultaneously with all of 𝓁u∕h is kept to 10 or less on the first floor and 14 or
[34 − 12(M1 ∕M2)] shall not be taken larger than 40,
the other columns on the same level. As a result, it less for the floors above the first one. To determine
according to NSCP Equation 406.2.5c.
is necessary in a sway frame to consider the these values, it was assumed that little moment
For sway frames, slenderness effects may be resistance was provided at the footing–column
buckling strength of all the columns on the level ignored if k𝓁u
in question as a unit. ≤ 22 (406.2.5a) connection and the first-floor columns were
r assumed to be bent in single curvature. Should the
For a compression member in a nonsway Should k𝓁u∕r for a particular column be larger
footing–column connection be designed to have
frame, the effective slenderness ratio, k𝓁u∕r, is than the applicable ratio, we will have a slender
appreciable moment resistance, the maximum 𝓁u∕h
used to determine whether the member is short column. For such a column, the effect of slenderness
value given above as 10 should be raised to about
or slender. For this calculation, 𝓁u is the unbraced must be considered. This may be done by using
14 or equal to the value used for the upper floors.
length of the member. The effective length factor, approximate methods or by using a theoretical
k, can be taken as 1.0 unless an analysis provides second-order analysis that takes into account the Should we have an unbraced frame and assume k =
a lesser value. The radius of gyration, r, is equal to effect of deflections. Second-order effects cannot 1.2, it is probably necessary to keep 𝓁u∕h to 6 or
0.25 times the diameter of a round column and exceed 40% of first-order effects (NSCP Section less. So for a 3-m clear floor height, it is necessary
0.289 times the dimension of a rectangular 406.2.6). to use a minimum h of about 3,000∕6 = 500 mm in
column in the direction that stability is being A second-order analysis is one that takes into the direction of bending to avoid slender columns.
considered. The NSCP, Section [Link] permits account the effect of deflections and also makes use of
the approximate value of 0.30 to be used in place a reduced tangent modulus. The equations necessary
of 0.289, and this is done herein. For other for designing a column in this range are extremely
sections, the value of r will have to be computed complicated, and, practically, it is necessary to use
from the properties of the gross sections. column design charts or computer programs.
 Example 11.1
0.30m×0.45m
(a) Using the alignment 2,278×106 mm4)
charts of Figure 11.3, 0.45m
calculate the effective
length factor for 0.50m
column AB of the 0.3m×0.6m 0.30m×0.50m
3.0m 0.875
braced frame shown. 5,400×106 mm4) 3,125×106 mm4)
Consider only
bending in the plane
0.60m 0.50m
of the frame. For
calculating I∕L values, use 0.30m×0.50m 3.6m
𝟎. 𝟕𝟎𝑰g for the columns, 3,125×106 mm4)
𝟎. 𝟑𝟓𝑰g for the girders,
and the full lengths of
members center to center
of supports.
6.0m 7.2m
Compute the value of 𝜓 at each end of column AB:
0.7(3,125×106)
𝜓A = 3,000
= 2.993
0.35(2,278×10 ) 0.35(2,278×106)
6
+
6,000 7,200
0.7(3,125×106) 0.7(3,125×106)
+
3,000 3,600
𝜓B = = 2.315 FIGURE 11.3 Effective length factors.
0.35(5,400×106) 0.35(5,400×106) 𝜓 = ratio of Σ(EI∕𝓁) of columns to Σ(EI∕𝓁) of girders in a
+ plane at one end of a compression member.
6,000 7,200 k = effective length factor.
 Example 11.1
(a) Using the alignment 0.30m×0.45m
2,278×106 mm4)
charts of Figure 11.3, 0.45m
calculate the effective
length factor for
0.50m
column AB of the 0.30m×0.50m
0.3m×0.6m 3.0m
braced frame shown. 5,400×106 mm4) 3,125×106 mm4) 0.875
Consider only
bending in the plane
of the frame. For 0.60m 0.50m
calculating I∕L values, use
𝟎. 𝟕𝟎𝑰g for the columns, 0.30m×0.50m 3.6m
3,125×106 mm4)
𝟎. 𝟑𝟓𝑰g for the girders,
and the full lengths of
members center to center
of supports.
(b) Compute the slen- 6.0m 7.2m
derness ratio of
column AB. Is it a 𝓁u = 3,000 - 450/2 - 600/2 = 2,475 mm
short or a slender For a short column in
k𝓁u 0.875 (2,475)
column? End moments = 0.30(500) a braced frame, the
r
on the column are k𝓁 M
M1 = 60 kN·m and = 14.44 < 26.8 max. u = 34 - 12 1
r M2
M2 = 100 kN·m, 60
resulting in single ∴ it is a short column = 34 - 12 FIGURE 11.3 Effective length factors.
100 𝜓 = ratio of Σ(EI∕𝓁) of columns to Σ(EI∕𝓁) of girders in a
curvature. = 26.8 plane at one end of a compression member.
k = effective length factor.
 Example 11.1b
0.30m×0.45m
(a) Using the alignment 2,278×106 mm4)
chart, re-calculate 0.45m
the effective length
factor for column AB 0.50m
if the frame shown is 0.3m×0.6m 0.30m×0.50m
3.0m
unbraced. Consider 5,400×106 mm4) 3,125×106 mm4)
only bending in the
plane of the frame. 0.60m 0.50m
For calculating I∕L
values, use 𝟎. 𝟕𝟎𝑰g for 0.30m×0.50m 3.6m
the columns, 𝟎. 𝟑𝟓𝑰g for 3,125×106 mm4)
the girders, and the full
lengths of members
center to center of 1.75
supports.
6.0m 7.2m
Compute the value of 𝜓 at each end of column AB:
0.7(3,125×106)
3,000
𝜓A = = 2.993
0.35(2,278×106) 0.35(2,278×106)
+
6,000 7,200
0.7(3,125×106) 0.7(3,125×106)
+
3,000 3,600
𝜓B = = 2.315 FIGURE 11.3 Effective length factors.
0.35(5,400×106) 0.35(5,400×106) 𝜓 = ratio of Σ(EI∕𝓁) of columns to Σ(EI∕𝓁) of girders in a
+ plane at one end of a compression member.
6,000 7,200 k = effective length factor.
 Example 11.1b
(a) Using the alignment 0.30m×0.45m
2,278×106 mm4)
chart, re-calculate 0.45m
the effective length
factor for column AB
0.50m
if the frame shown is 0.30m×0.50m
0.3m×0.6m 3.0m
unbraced. Consider 5,400×106 mm4) 3,125×106 mm4)
only bending in the
plane of the frame.
For calculating I∕L 0.60m 0.50m
values, use 𝟎. 𝟕𝟎𝑰g for
the columns, 𝟎. 𝟑𝟓𝑰g for 0.30m×0.50m 3.6m
3,125×106 mm4)
the girders, and the full
lengths of members
center to center of
supports. 1.75
(b) Compute the slen- 6.0m 7.2m
derness ratio of
column AB. Is it a 𝓁u = 3,000 - 450/2 - 600/2 = 2,475 mm
short or a slender
k𝓁u 1.75 (2,475) For a short column in
column? The max. = 0.30(500)
permissible slender- r an unbraced frame, the
ness ratio for a short k𝓁
= 28.88 > 22 max. u = 22
unbraced column is r
22, as mentioned in ∴ the column is a slender column FIGURE 11.3 Effective length factors.
Section 11.7. 𝜓 = ratio of Σ(EI∕𝓁) of columns to Σ(EI∕𝓁) of girders in a
plane at one end of a compression member.
k = effective length factor.
11.8 NSCP (ACI Code) Treatments The third method is the moment magnifier
of Slenderness Effects procedure (NSCP, Section 406.6.4). Different
procedures for this method are given for sway and
The NSCP permits the determination of nonsway structures. The next two sections describe
second-order effects by one of three methods. the moment magnifier method for these two cases.
The first is by a nonlinear second-order
analysis (NSCP, Section 406.8.1). Such an 11.9 Magnification of Column
analysis must consider nonlinearity of Moments in Nonsway Frames
materials, member curvature and lateral drift,
load duration, volume changes in concrete When a column is subjected to moment along its
because of creep and shrinkage, and unbraced length, it will be displaced laterally in the
foundation or support interaction. The analysis plane of bending. The result will be an increased or
technique should predict the ultimate loads to secondary moment equal to the axial load times the
within 15% or test results on statically lateral displacement or eccentricity. In Figure 11.5,
indeterminate reinforced concrete structures the load P causes the column moment to be
(ACI R6.8.1.2). This technique would require increased by an amount PΔ. This moment will
sophisticated computer software that has been cause 𝛿 to increase a little more, with the result
demonstrated to satisfy the 15% accuracy that the PΔ moment will increase, which in turn
requirement mentioned previously. will cause a further increase in Δ and so on until
The second method is by an elastic second- equilibrium is reached.
order analysis (NSCP, Section [Link]). This We could take the column moments, compute
technique is simpler than the nonlinear the lateral deflection, increase the moment by PΔ,
method because it uses member stiffnesses recalculate the lateral deflection and the increased
immediately prior to failure. Values of Ec and moment, and so on. Although about two cycles
moments of inertia and cross-sectional area for would be sufficient, this would still be a tedious
columns, beams, walls, flat plates, and flat slabs and impractical procedure.
that are permitted to be used in the elastic It can be shown that the increased moment can
second-order analysis are listed in Section 11.6 be estimated very well by multiplying the primary
of the text. This method would also most likely moment by 1/(1 − P∕Pc), where P is the axial load
require a computer analysis. and Pc is the Euler buckling load 𝜋2EI∕(k𝓁u)2.
 Example 11.2
670 kN (a) Primary Moment Resulting from Lateral Load
(a) Compute the primary
moment in the column 90 (4.5)
Mu =
shown in the figure from 4
the lateral 90-kN load. = 101.25 kN·m
(b) Determine the estimated 0.3m × 0.30m (b) Total Moment, Including Secondary Moment
total moment, including column π2EI
2.25m Pc = Euler buckling load =
the theoretical secondary (k𝓁u)2
moment from lateral Δ π2 (21,800)(675×106)
90 kN Pc =
deflection, using the (1.0×4,500)2
appropriate magnification = 7,171.9 kN
factor just presented. E = 2.25m 1
Mc = Magnified moment = M2
21,800 MPa. Assume k =1.0 P
1-
and 𝓁u = 4.5 m. 1 Pc
Mc = (101.25)
670
1-
670 kN 7,171.9
= 111.68 kN·m

300(300)3 Recall:
I= (1) PL3
12 δ δ=
48EI
I = 675×106 mm4
(2) The formula for the sum of an infinite
geometric series is S∞ = a1 /(1-r).
 Difference between SHORT and SLENDER columns
In the previous example, the increased
moment resulting from lateral deflection is
calculated approximately by using the (1 − P∕Pc) M2 = Mmax M2
expression. In NSCP, Section [Link], the
factored design moment for slender columns with
no sway is increased by using the following
expression, in which Mc is the magnified or
increased moment and M2 is the larger factored
end moment on a compression member:
Mc = 𝛿 M2 ([Link].1) Mc = Mmax
Should our calculations provide very small
moments at both column ends, the code provides Mc = δM2
an absolutely minimum value of M2 to be used in M1 M1
design. In effect, it requires the computation of a
moment based on a minimum eccentricity of
(15 + 0.03h) mm, where h is the overall
thickness of the member perpendicular to the
axis of bending.
M2 = Mmax M2
M2 min = Pu(15 + 0.03h) ([Link].4)
A moment magnifier, 𝛿, is used to estimate the
effect of member curvature between the ends of
compression members. It involves a term Cm,
which is defined later in this section.
Mc = Mmax
Cm
Mc = δM2
δ= ≥ 1.0 ([Link].2)
Pu
1-
0.75Pc
M1 M1
The determination of the moment magnifier, 𝛿, involves the following calculations:

1. Ec = 4,700 𝒇′𝒄 MPa for normal-weight concrete (see The alternate expression for (EI)eff that follows is For braced frames without
Section 1.11 for other densities). probably the better expression to use when steel transverse loads between supports, Cm
percentages are low. Notice also that this can vary from 0.4 to 1.0 and is
2. Ig = gross inertia of the column cross section about expression will be the one used if the reinforcement determined with the expression at the
the centroidal axis being considered. has not been previously selected. end of this paragraph. For all other cases,
3. Es = 200,000 MPa. 0.4EcIg it is to be taken as 1.0. (Remember the
(EI)eff = ([Link].4a) sign convention: M1 is positive for single
4. Ise = moment of inertia of the reinforcement about 1 + 𝛽dns
the centroidal axis of the section. (This value equals curvature and is negative for reverse
A third alternative equation (EI)eff is given by NSCP curvature, and M2 is always positive.)
the sum of each bar area times the square of its equation [Link].4c. In this equation, the moment
distance from the centroidal axis of the compression M
of inertia of the column, I, is calculated from NSCP Cm = 0.6 + 0.4 1 ([Link].3a)
member.) Table [Link].1b. M2
5. The term 𝛽dns accounts for the reduction in stiffness EcI
caused by sustained axial loads and applies only to (EI)eff = ([Link].4c)
1 + 𝛽dns
nonsway frames. It is defined as the ratio of the
maximum factored sustained axial load divided by 7. The Euler buckling load is computed:
the total factored axial load associated with the same π2EI
load combination. It is always assumed to have a
Pc = ([Link].2)
(k𝓁u)2
plus sign and is never permitted to exceed 1.0. Should M2,min as computed with
8. For some moment situations in columns, the amplification
6. Next, it is necessary to compute an effective value of NSCP Equation [Link].4 be larger than
or moment magnifier expression provides moments that
(EI) designated as (EI)eff. The two expressions given M2, the value of Cm in this equation shall
are too large. One such situation occurs when the moment
for (EI)eff in the code were developed so as to either be taken as equal to 1.0 or be
at one end of the member is zero. For this situation, the
account for creep, cracks, and so on. If the column based on the ratio of the computed end
lateral deflection is actually about half of the deflection in
and bar sizes have already been selected or moments M1∕M2 (NSCP, Section
effect provided by the amplification factor. Should we have
estimated, (EI)eff can be computed with the [Link].3). If the braced column is
approximately equal end moments that are causing reverse
following expression, which is particularly subjected to transverse loads between
curvature bending, the deflection at middepth and the
satisfactory for columns with high steel percentages. supports, Cm is taken as 1.0 (NSCP
moment there are close to zero. As a result of these and
Equation [Link].3b).
(0.2EcIg + EsIse) other situations, the code provides a modification factor
(EI)eff = ([Link].4b) (Cm) to be used in the moment expression that will result in Example 11.3 illustrates the design of
1 + 𝛽dns a column in a nonsway frame.
more realistic moment magnification.
 Example 11.3
The tied column in the figure has been 1. Is it a Slender Column? Cm 0.9795
approximately sized to the dimensions For a short column in a braced frame, 8. δ = =
Pu 490
300 mm × 375 mm. It is to be used in a k𝓁 M 1 - 0.75P 1 - 0.75(6,140.2)
frame braced against sidesway. The max. u = 34 - 12 1 = 34 - 12 111 c
r M2 117 = 1.096 > 1.0
column is bent in single curvature = 22.62
about its y-axis and has an 𝓁u of 4.8 m. If 9. M2,min = Pu(15 + 0.03h)
k𝓁 0.83(4800)
k = 0.83, fy = 420 MPa, and fc′ = 28 MPa actual u = = 35.41 > 22.62 = 490(15 + 0.03×375)
r 0.3(375)
normal-weight concrete, determine the = 12.86 kN·m
∴ the column is slender
reinforcing bars required. Consider 10. Mc = δM2 = 1.096(117)
only bending in the plane of the frame. 2. Ec = 4,700 𝒇′𝒄 = 4,700 𝟐𝟖 = 24,870 MPa
= 128.23 kN·m
Note also that the unfactored dead axial 1
3. Ig = 12 (300)(375)3 = 1,318.4×106 mm4 Mc 128.23
load PD is 134 kN. 11. Magnified e = =
factored axial dead load 1.2(134) Pu 490
Pu = 490 kN 4. 𝛽dns = = = 0.2617 m
factored axial load (total) 490
= 0.3282 12. γ = 249/375 = 0.664
M1b = 111
kN·m 5. Reinf. unknown, use (406.6.4.4a) for (EI)eff . P 490
300 0.4EcIg 0.4(24,870)(1,318.4×106) 𝑷𝒏 = u = = 753.85 kN
ϕ 0.65
(EI)eff = =
1 + 𝛽dns 1 + 0.3282 P 753.85(1,000)
= 9,874.6 kN·m2 Kn = ′ n =
𝒇𝒄 Ag 28(300×375)
π2EI π2(9,874.6) = 0.2393
63 249 63 6. Pc = = = 6,140.2 kN
M2b = 117 (k𝓁u)2 (0.83×4.8)2 P e 753.85(1,000)261.7
kN·m Rn = ′ n =
375 𝒇𝒄 Ag h 28(300×375) 375
M1 111
7. Cm = 0.6 + 0.4 = 0.6 + 0.4 117 = 0.9795 = 0.1670
Pu = 490 kN M2
 𝒇′𝒄 = 28 MPa 𝒇′𝒄 = 28 MPa
fy = 420 MPa fy = 420 MPa

Pu = 490 kN

WORD OF CAUTION: γ = 249/375

Assuming a frame is fully braced
= 0.664may be quite unconservative.
P n
Kn =
Ag
𝒇′𝒄
0.2393 0.2393
= 0.2393
P e
Rn = ′ n
𝒇𝒄 Ag h 0.1670
0.1670
= 0.1670

checking, using strain 𝜸 when using 4-20mmϕ bars:

compatibility method: 𝟎. 𝟔𝟎 𝟎. 𝟔𝟔𝟒 𝟎. 𝟕𝟎
𝒆 = 261.7 mm: Pn = 762.4 kN 𝝆g 𝟎. 𝟎𝟏𝟗𝟕 𝟎. 𝟎1772 𝟎. 𝟎𝟏𝟔𝟔 𝒆 = 261.7 mm: Pn = 556.6 kN
ϕ = 0.774 ∴ As = 𝝆g𝑨g = 𝟎. 𝟎𝟏𝟕𝟕𝟐 (300×375)= 𝟏, 𝟗𝟗𝟒 mm2 ϕ = 0.90
ϕ Pn = 590.3 kN say, try 4-25mmϕ bars (As = 1,964 mm2) ϕ Pn = 500.9 kN
 In this last equation, ΣPu is the summation
11.10 Magnification of Column Section [Link].2 of the NSCP states that
of all the vertical loads in the story in question,
Moments in Sway Frames the moment magnifier, 𝛿s, can be determined
and ΣPc is the sum of all the Euler buckling
by one of the following two methods. loads, Pc = 𝜋2(EI)eff ∕(k𝓁u)2, for all of the sway-
Tests have shown that even though the lateral resisting columns in the story with k values
1. The moment magnifier may be
deflections in unbraced frames are rather determined as described in NSCP, Section
calculated with the equation given at the
small, their buckling loads are far less than [Link].3. This formula reflects the fact that
end of this paragraph (Equation
they would be if the frames had been braced. the lateral deflections of all the columns in a
As a result, the buckling strengths of the [Link].2a) in which Q is the stability
index previously presented in Section particular story are equal, and thus the
columns of an unbraced frame can be columns are interactive. In calculating Pc ,
decidedly increased (perhaps by as much as 11.2 of this chapter. Should the
computed value of 𝛿s be greater than 1.5, (EI)eff is determined by either NSCP Equation
two or three times) by providing bracing. [Link].4 or NSCP Equation [Link].5 using
it will be necessary to compute 𝛿s by
If a frame is unbraced against sidesway, it 𝛽ds instead of 𝛽dns.
NSCP Equation [Link].2b or by a
is first necessary to compute its slenderness Whichever of the preceding methods is
second-order analysis.
ratio. If k𝓁u∕r is less than 22, slenderness may used to determine the 𝛿s values, the design
be neglected (NSCP Section 406.2.5). For this 1 moments to be used must be calculated with
discussion, it is assumed that values greater 𝛿s = ≥ 1.0 ([Link].2a)
1-Q the expressions that follow.
than 22 are obtained.
2. With the second method and the one M1 = M1ns + δsM1s ([Link].1a)
When sway frames are involved, it is
used in this chapter, the magnified sway
necessary to decide for each load combination
which of the loads cause appreciable sidesway
moments may be computed with the M2 = M2ns + δsM2s ([Link].1b)
(probably the lateral loads) and which do not. following expression:
Sometimes the point of maximum moment
The factored end moments that cause 1 in a slender column will fall between its ends.
sidesway are referred to as M1s and M2s, and δs = ≥ 1.0 ([Link].2b)
ΣPu The ACI Commentary (R6.[Link]) says the
they must be magnified because of the PΔ 1- moment magnification for this case may be
0.75ΣPc
effect. The other end moments, resulting from
evaluated using the procedure described for
loads that do not cause appreciable sidesway,
nonsway frames (NSCP, Section [Link]).
are M1ns and M2ns. They are determined by
first-order analysis and will not have to be Example 11.4 illustrates the design of a
magnified. slender column subject to sway.
 Example 11.4
Select reinforcing bars using the 450 mm × Is it a Slender Column?
450 mm unbraced column shown if 𝓁u =
5.30 m, k = 1.3, fy = 420 MPa, and fc′ = 28 k𝓁 1.3(5,300)
actual u = = 51.04 > 22 ∴ the column is slender
MPa. A first-order analysis has resulted in r 0.3(450)
the following axial loads and moments:
NSCP Pu Mu M2ns M2s ΣPu ΣPc M2
PD = 890 kN MD = 65 kN·m Case Combination δs
Equation (kN) (kN·m) (kN·m) (kN·m) (kN) (kN) (kN·m)
PL = 670 kN ML = 34 kN·m
1 (405.3.1a) 1.4D 1,246 91 91 - 15,000 396,000 1.053 91.00
PW = 290 kN MW = 160 kN·m
Consider the following values of ΣPu and ΣPc 2 (405.3.1b) 1.2D+1.6L 2,140 132 132 - 25,600 396,000 1.094 132.40
for these load combinations:
3 (405.3.1d) 1.2D+1.0W+L 2,028 272 112 160 24,300 396,000 1.089 286.26
ΣPu ΣPc
1.4D 15,000 396,000 4 1.2D-1.0W+L 1,448 -48 112 -160 17,400 396,000 1.062 -57.96
1.2D+1.6L 25,600 396,000
1.2D+1.0W+L 24,300 396,000 5 (405.3.1f) 0.9D+1.0W 1,091 219 59 160 13,100 396,000 1.046 225.88
1.2D-1.0W+L 17,400 396,000 6 0.9D-1.0W 511 -102 59 -160 6,100 396,000 1.021 -104.86
0.9D+1.0W 13,100 396,000
0.9D-1.0W 6,100 396,000 1
δs =
ΣPu
≥ 1.0 ([Link].2b) M2 = M2ns + δsM2s ([Link].1b)
65 1-
0.75ΣPc
Pn 3,120 (1,000)
γ = 320/450 = 0.711 Kn = = Homework:
𝒇′𝒄 Ag 28(450×450)
320 450 P 2,028 = 0.5503 Verify if As of
𝑷𝒏 = u = = 3,120 kN 6-28mm ϕ is
ϕ 0.65 P e 3,120(1,000) 141.2
Rn = ′ n = adequate.
65 M 286.26 𝒇𝒄 Ag h 28(450×450) 450
e= 2= = 0.1412 m
450 Pu 2,028 = 0.1727
 7120
6230
5340
4450
3560
kN.

2670
1780
890
0
113 226 339 452 565
-890
-1780
-2670
kN-m.