Webber 2015
Webber 2015
Webber 2015
Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct
a r t i c l e i n f o a b s t r a c t
Article history: Codes of practice rely on the effective length method to assess the stability of multi-storey frames. The
Received 5 January 2015 effective length method involves isolating a critical column within a frame and evaluating the rotational
Revised 24 July 2015 and translational stiffness of its end restraints, so that the critical buckling load may be obtained.
Accepted 27 July 2015
The non-contradictory complementary information (NCCI) document SN008a (Oppe et al., 2005) to BS
Available online 24 August 2015
EN 1993-1 (BSI, 2005) provides erroneous results in certain situations because it omits the contribution
made to the rotational stiffness of the end restraints by columns above and below, and to the transla-
Keywords:
tional stiffness of end restraints by other columns in the same storey.
Steel structures
Multi-storey frames
Two improvements to the method are proposed in this paper. First, the axial load in adjoining columns
Effective length method is incorporated into the calculation of the effective length. Second, a modification to the effective length
Buckling ratio is proposed that allows the buckling load of adjacent columns to be considered. The improvements
Design methods are shown to be effective and consistently provide results within 2% of that computed by structural anal-
Design aids ysis software, as opposed to the up to 80% discrepancies seen using the NCCI (Oppe et al., 2005).
Ó 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).
http://dx.doi.org/10.1016/j.engstruct.2015.07.039
0141-0296/Ó 2015 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
A. Webber et al. / Engineering Structures 102 (2015) 132–143 133
Nomenclature
P
C carry-over factor K c;I sum of rotational stiffness of the columns converging at
E Young’s modulus P node I
gx distribution coefficient at node X K b;I sum of rotational stiffness of the beams converging at
I second moment of area node I modified for axial load and far end support con-
IIJ second moment of area of column IJ ditions
j curvature of an elastic member LE,IJ effective length of column IJ
K IJ nominal rotational stiffness of column IJ LIJ physical length of column IJ
K XY nominal rotational stiffness of an adjoining column XY Mx moment at node X
P applied compression
K 0IJ rotational stiffness of column IJ modified for axial load
Pc,IJ critical buckling load for column IJ given by Eq. (3)
K 00IJ rotational stiffness of column IJ at node I modified for PE,IJ Euler buckling load for column IJ given when LE,IJ = LIJ in
axial load and support conditions at node J Eq. (3)
K 00JI rotational stiffness of column IJ at node J modified for PIJ applied compression on column IJ
axial load and support conditions at node I hx rotation at node X
K 00XY rotational stiffness of an adjoining column XY modified S stiffness coefficient
for axial load and far end support conditions v deflection
2. Frame stability ‘C’ (Eq. (5)) for non sway cases [14]. For the sway case (see
Fig. 2), the stability function is ‘n’ and the carry over factor ‘o’,
Being able to determine the effective length of framed columns and equations for these coefficients may be found in the literature
is important, as it provides a simple approach to assessing frame [14]. They define the end moments of a fixed-pin column IJ that is
stability. To find the effective length of an individual column rotated by h at its pinned end I, as displayed in Fig. 2 for sway and
within a frame, the rotational and translational stiffness of its non-sway frames.
end restraints must be considered.
kLIJ sin kLIJ kLIJ cos kLIJ
A braced frame would usually be categorised as a non-sway S¼ ð4Þ
frame (lateral displacements are sufficiently small that the sec-
2 2 cos kLIJ kLIJ sin kLIJ
ondary forces and moments can be ignored) and the translational
stiffness of a column’s end restraints are taken as infinity. kLIJ sin kLIJ
C¼ ð5Þ
In an unbraced frame, the secondary effects caused by lateral sin kLIJ kLIJ cos kLIJ
displacements are usually significant and consequently the trans- qffiffiffiffiffi
P IJ
lational stiffness of a column’s end restraints is taken as zero. where k ¼ EIIJ
, LIJ is the length of column IJ, PIJ is the axial load on
If all connections between beams and columns are assumed to be column IJ, IIJ is the second moment of area of column IJ.
fully rigid (i.e. there is no rotation of the beam relative to the column Since Eqs. (4) and (5) depend only on kLIJ , which can alterna-
at a connection) it may seem sensible to take an effective length ratio tively be given by Eq. (6), they are functions of the ratio of the axial
for a column in an unbraced frame of 1.0. However this could be an load to Euler load (P/PE). Such values have been tabulated exten-
onerous over simplification because the connecting members will sively in the literature [14,15].
deform when the column buckles. The connecting members restrain sffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffi
the buckling column and provide rotational stiffness. It is also possi- P IJ PIJ
kLIJ ¼ LIJ ¼p ð6Þ
ble that adjoining members provide negative rotational stiffness if EIIJ PE;IJ
they too are subjected to significant axial load and have buckled.
A significant source of inaccuracy in the design of columns where PE,IJ is the Euler buckling load of column IJ, given when
using the effective length method is uncertainty in the estimation LIJ = LE,IJ
of rotational boundary conditions for the column. This has been When the axial load P equals zero, the stiffness coefficient S
recognised in the literature [6,7]. The effective length method con- equals four. Therefore the moment required to rotate the column
siders columns individually, even though the presence of other in Fig. 2(a) by theta degrees is given by Eq. (7):
members is crucial to buckling behaviour. The contribution of EI
adjoining members is taken into account indirectly through the MI ¼ 4hI ð7Þ
L
summation of stiffnesses of the members at the top and bottom
Wood’s [8] general definition for rotational stiffness is given by
of a column. This approach was used by Wood [8], whose work
Eq. (8):
forms the theoretical basis for buckling calculation in BS EN 1993
[2]. However, the work of Wood [8] produces unusual results in MI
K IJ ¼ ð8Þ
some situations, which will be considered later. 4EhI
The work of Cheong-Siat-Moy [9] provides early insight into the
The resistance of the fixed-pinned column IJ (Fig. 3) to rotation
need to consider both individual element and overall system beha-
at the pinned node I, when the axial load is zero, is therefore given
viour for accurate buckling analysis, while Bridge and Fraser [10]
by Eq. (9):
extended this to consider negative rotational stiffnesses.
Hellesland and Bjorhovde [11] also show the importance of fully IIJ
K IJ ¼ ð9Þ
considering the contributions of adjacent elements to rotational LIJ
stiffnesses. They propose a ‘weighted mean’ approach to determin- This is called the nominal rotational stiffness and can be modified
ing frame buckling from individual element analyses, and impor- to take account of axial load in the column by using the stability
tantly applied this method to frames in which column stiffnesses function ‘S’, as shown in Eq. (10):
change significantly between storeys. Aristizabal-Ochoa [12] fur-
S
ther examined the effect of uniformly distributed axial loads, and K 0IJ ¼ K IJ ð10Þ
the behaviour of frames with partial side-sway [13].
4
Cheong-Siat-Moy [9] proposed the use of a fictitious lateral load where S is given by Eq. (4).
as a method to evaluate the buckling capacity of columns as an If the end ‘J’ is in fact not fully rigid, but has a rotational stiffness
alternative to the effective length method, but this has not been due to the presence of adjoining beams then this will result in a
adopted by designers. reduced rotational stiffness at end I, which is related to the relative
Ultimately, the critical load of an individual column within a stiffness of the adjoining beams at end J.
frame cannot be obtained without considering the loads in the rest To find this reduction in K 0IJ we need to contemplate the follow-
of the structure as this will affect the stiffness of the column’s end ing scheme, which is illustrated in Fig. 3:
restraints and change its effective length. Codes of practice get
around this by assuming worst-case scenarios such as that adjoin- Consider the column IJ where end I is free and end J is initially
ing columns in the storey above and below buckle simultaneously fixed, Fig. 3(a).
with the column under investigation, and as such reduce the rota- End I is rotated by h, which requires an applied moment of
tional stiffness of its end restraints, and that other columns in the MI = SEKIJ, where S is the general stiffness coefficient for either
same storey buckle simultaneously. Both of these can lead to sway/non-sway case, and KIJ is the nominal rotational stiffness
over-conservative results. from Eq. (9) and E is the Young’s modulus (Fig. 3(a)).
The moment carried over to end J is CMI where C is the general
2.1. Stiffness distribution method carry-over factor.
This step is explained by considering the general case of a
The stiffness distribution method, which is used in SN008a [1] moment applied at a node ‘X’ which has three adjoining
employs the stability function ‘S’ (Eq. (4)) and carry over factor members (one column and two beams, for example). The
A. Webber et al. / Engineering Structures 102 (2015) 132–143 135
Fig. 2. Demonstration of end moments for fixed-pinned struts in (a) non sway and (b) sway frames.
Fig. 3. Moment required to rotate the free end of a column by h, with adjoining beams at far end.
Then, by keeping end I held at h, and replacing the fixed support K 00IJ ¼ 0 ð16Þ
at end J by adjoining beams (Fig. 3(b)) and releasing the However in a frame the rotational restraint provided by any
moment CMI , it follows from Eq. (12) that the moment in the adjoining beams at node I would also need to be considered.
column at end J is CM I gJ and the moment distributed back Therefore the critical axial load in the column is reached when
to end I is CMI gJ C (Fig. 3(b)). Eq. (17) is satisfied.
Therefore the net moment required to rotate end I by h, for the X
column which is not fully rigid at joint J, is: K 00IJ þ K b;I ¼ 0 ð17Þ
P
where K b;I is the rotational stiffness of the beams at node I. This is
M I CM I gJ C ð13Þ found by first combining Eqs. (15) and (17) to give Eq. (18):
!!
Hence, the reduction in K 0IJ is found by combining Eqs. (8) and K IJ S
K IJ S
X
(13). The resistance of a column IJ to rotation at node I, with adjoin- 1 C2 K IJ S
4
P þ K b;I ¼ 0 ð18Þ
4 þ K b;J
ing beams at node J, is then given by Eq. (14) and notated as K 00IJ : 4
136 A. Webber et al. / Engineering Structures 102 (2015) 132–143
1
K b;I ¼ K IJ 1 ð19Þ Deflected shape
gI
I
!
X 1
K b;J ¼ K JI 1 ð20Þ
gJ
L
Substituting these into Eq. (18), noting that the KIJ terms can
now be removed, gives Eq. (21): J
0 0 11
S
S@ 1
1 C2@ 4 AA þ 1 ¼0 ð21Þ
4 S
þ 1 1 gI
4 gJ
MA EK ð4hA þ 2hB Þ hB
sion represents the rotational stiffness of the adjoining members at K 0A ¼ ¼ ¼ K A 1 þ 0:5 ð28Þ
node I. 4EhA 4EhA hA
A. Webber et al. / Engineering Structures 102 (2015) 132–143 137
where hB is the rotation at the far end, and hA is the rotation at the the column effective length ratio (w) to be modified to account
near end. for this effect by Eq. (31).
In the critical buckling mode shape of a sway frame, beams vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffi
uP
bend in double curvature and the rotations at both ends of a beam u Pu Ii 5
w0i ¼t P I P w ð31Þ
are equal hA ¼ hB ! hhAB ¼ 1. Therefore the beams modified rota- Pui w2
8 i0
0
tional stiffness would be: P
where w0i is the modified column effective length ratio; Pu is the
K 0 ¼ K ð1 þ 0:5 1Þ ¼ 1:5K ð29Þ required axial compressive strength of all columns in a storey; P ui is
In the critical buckling mode shape of a non-sway frame, beams the required compressive strength of the column under investiga-
P I
bend in single curvature and the rotation at both ends of a beam tion; w2
is the ratio of the second moment of area to the effective
0
are equal but opposite hA ¼ hB ! hhAB ¼ 1. Therefore the beams length factor of each column in the storey based on the sway align-
modified rotational stiffness would be: ment charts found in [18], Ii is the second moment of area of the
column under investigation; wi0 is the unmodified effective length
K 0 ¼ K ð1 þ 0:5 1Þ ¼ 0:5K ð30Þ factor for the column.
Eq. (31) allows for column end nodes with translational stiff-
2.2. NCCI SN008a ness’ somewhere between zero and infinity, corresponding to the
support conditions of a weaker column that is being braced by a
BS EN 1993 [2] refers to the NCCI document SN008a [1]. The stronger one. It also allows for negative translational stiffness, as
method presented in the NCCI for determining the effective length encountered when the column under investigation is bracing a
of columns is the same as that presented in Annex E of BS 5950 weaker adjacent column or a leaning column (a leaning column
[17]. has pinned supports and therefore has no translational stiffness).
The procedure to determine the effective length of steel col- The application of this modification factor is demonstrated in
umns in frames outlined in SN008a [1] is simple to undertake. Section 3.2.1 below.
Distribution coefficients calculated for the top and bottom of the
column (notated as g1 and g2 in SN008a [1]) allow the effective 3. Application
length ratio to be extracted from design charts for both sway or
non-sway cases. Distribution coefficients can theoretically vary The effective length method will be used to evaluate the stabil-
from zero (analogous of a fully rigid support) to one (representing ity of a variety of framed columns. The results obtained will be
a pure pin). compared to that computed using structural analysis software
The distribution coefficients for the top and bottom column Autodesk Robot Structural Analysis [19] which will be used to per-
nodes are calculated from Eq. (23). form an eigenvalue buckling analysis. This type of analysis is
The rotational stiffness is assumed to be linear elastic. The rota- directly comparable to the Euler formula as it makes all the same
tional stiffness of a member with a fixed far-end and no axial load assumptions, such as initially straight, perfectly elastic members,
can be determined from Eq. (9), which is used in SN008a [1] to cal- and uses a linearised expression for curvature (the one difference
culate column stiffness regardless of their real far end restraint being that for compatibility reasons the curvature of members is
conditions and axial load. The rotational stiffness of adjoining represented using a cubic form instead of sinusoidal).
beams can however be modified in SN008a [1]. At present the
design charts are only appropriate for frames with fully rigid con- 3.1. The effect of a column from an adjacent storey
nections, as the NCCI does not provide any guidance on how to
design for semi-rigid connections. The effective length method will be used to assess the stability
Following the recommendations of BS 5950 Annex E [17], any of the frames in Figs. 5 and 6, considered first as non-sway and
restraining member required to carry more than 90% of its moment then as sway frames. Both frames have regular spaced columns
capacity should be assigned a K value of zero. Similarly, if either and an even distribution of load (which is common in real building
end of the column being designed is required to carry more than structures). Consequently, it is expected that frame instability will
90% of its moment-carrying capacity, then the distribution coeffi- be caused by the buckling of all columns in a single storey as the
cient (g) should be taken as 1. columns in this storey will have reached their critical load. The
In the sway frame design chart, the effective length ratio can effective length method will be used to find the elastic buckling
vary from one to infinity, with an effective length ratio of infinity load of a column in the critical storey. The frames are identical,
corresponding to a sway frame column that has pinned supports other than the height of the top storey.
at both ends (a mechanism). All members are 203 203 60 UC and all connections are
fully rigid. Columns buckle about their weak axis
2.3. The AISC LRFD method (Iminor = 2070 cm4) whilst the beams bend about their strong axis
(Imajor = 6130 cm4), both in the plane of the frame. The beams carry
The American Institute for Steel Construction (AISC) Load and no axial load and do not reach their flexural capacity. The distribu-
Resistance Factor Design (LRFD) manual [18] presents a similar tion of the design loads are shown in Figs. 5 and 6 along with frame
method to the NCCI [1] for the calculation of the effective length geometry. The frames are braced out of plane. Base nodes of the
of columns in multi storey frames. After evaluating stiffness ratios columns are taken as fully rigid (g = 0). Frame B might be expected
at the top and bottom of the column, the effective length ratio is to buckle at a lower load compared to frame A since its top storey
read from design nomographs. Like the NCCI, the AISC method column is more slender.
assumes adjoining columns buckle simultaneously and therefore
adjoining columns provide negative rotational stiffness at 3.1.1. Example 1 – Non sway
restraints. Both methods are examples of linear static analysis of Frames A and B are first considered to be non-sway frames. For
the equilibrium of a column in its deformed state. Frame A (Fig. 5), the NCCI approach suggests that the columns in
In unbraced frames, stronger columns will brace weaker col- the middle storey buckle first and cause frame instability. The elas-
umns in the same storey, and the AISC commentary [18] allows tic critical load of column BC therefore defines the load in the rest
138 A. Webber et al. / Engineering Structures 102 (2015) 132–143
0.5P P A P P 0.5P
1m 0.5P P P P 0.5P
B
4m
0.5P P C P P 0.5P
3m
D
Fig. 5. Frame A: multi-storey fully rigid steel with stiff top storey.
0.5P P E P P 0.5P
6m
0.5P P P P 0.5P
F
4m
0.5P P G P P 0.5P
3m
H
Fig. 6. Frame B: multi-storey fully rigid steel with slender top storey.
of the frame at instability, and is found to be 3846 kN. The axial 3.1.2. Example 2 – Sway frames
load in column BC at frame instability, found using an Eigenvalue Frame A and Frame B are now considered to be sway frames.
analysis in Robot [19], is 7054 kN. The NCCI predicts Frame A will The NCCI predicts that in both frames, the middle storey col-
buckle at a load 45% lower than that computed by Robot. The cal- umns buckle first. The calculated elastic critical loads of BC and
culations for this section are summarised in Table 1. In the FG are calculated as P C;BC ¼ 1211 kN and P C;FG ¼ 1623 kN, contrary
non-sway computational analysis supports were added at all nodes to the expected result that P C;BC > PC;FG . The analysis using Robot
which prevented horizontal movement only. These were removed [19] computes the elastic critical loads of columns BC and FG as
when calculating the sway frames. PC;BC ¼ 2230 kN and P C;FG ¼ 1984 kN.
Applying the NCCI approach to Frame B (Fig. 6) suggests that The error again arises from the calculation of the distribution
column FG will buckle first, at an elastic critical load of 4653 kN. coefficients for the top nodes, where node B is less stiff than node
Using Robot, it is found that the axial load in column FG at frame F. The calculations for this section are summarised in Table 2.
instability is 5622 kN. In this instance the NCCI is closer, but still
predicts a collapse load 17% lower than that found using Robot.
Furthermore the NCCI predicts that Frame A buckles at a lower 3.2. The effect of columns in the same storey
load than Frame B, which from observation does not seem correct.
Robot supports the intuitive prediction that Frame B is less stable Another limitation of the NCCI approach is that it assumes all
than Frame A. columns in a storey buckle simultaneously. One of the conse-
The distribution coefficients calculated for nodes B and F using quences is that in unbraced frames the contribution made by adja-
the NCCI are gB ¼ 0:808 and gF ¼ 0:585. This shows that according cent columns in the same storey to the translational stiffness of the
to the NCCI, node F has more rotational stiffness than node B, column being checked is ignored. In the NCCI, the translational
which is evidently incorrect, as the short column AB is stiffer than stiffness of the end restraints is assumed to be either zero or infin-
slender column EF. The distribution coefficients calculated for ity for sway or non-sway frames respectively [1].
nodes C and G are the same (gC ¼ gG ¼ 0:663) as the members con- In unbraced frames, stronger columns will brace weaker col-
verging on these nodes are identical. So the effective length ratio of umns in the same storey. If the column being analysed is partially
LE;BC braced by a stiffer column then its end restraints will have transla-
column BC and FG, read from the design charts, are LBC
¼ 0:825 and
LE;FG
tional stiffness between zero and infinity. If the column being
LFG
¼ 0:750. The calculations for this section are summarised in checked is partially bracing a less stiff column, or fully bracing a
Table 1. pin ended column, the translational stiffness of the restraints will
Table 1
Calculations for Frame A and Frame B as non-sway frames.
Frame A Frame B
Distribution coefficients K BA þKP KP
gB ¼ K BA þK BC þ
BC
K 0b;B
¼ 0:808 gE ¼ K EF
K 0b;E
¼ 0:360
EF þ
gC ¼ K CB þKP
K CB þK CD þ
CD
K 0b;C
¼ 0:663 gF ¼ K þKK FE þKPFG
K 0b;F
¼ 0:585
FE FG þ
gD ¼ 0 gG ¼ gC ¼ 0:663
Effective length ratios and LE;BC LE;EF
LBC ¼ 0:825 ! P C;BC ¼ 3846 kN LEF ¼ 0:675 ! P C;EF ¼ 2553 kN
elastic critical loads using LE;CD LE;FG
LCD ¼ 0:825 ! P C;CD ¼ 11; 913 kN LFG ¼ 0:750 ! P C;FG ¼ 4653 kN
[1]
Conclusion When column BC buckles the axial load in column CD is: When column EF reaches its critical load, the axial load in column FG is
(3846 1.5 = 5,769 kN < 11,913 kN) therefore BC is critical. (2553 2 = 5106 kN > 4653 kN) therefore column FG buckles first.
A. Webber et al. / Engineering Structures 102 (2015) 132–143 139
Table 2
Calculations for Frame A and Frame B as sway frames.
Frame A Frame B
Distribution coefficients K BA þKP KP
gB ¼ K BA þK BC þ
BC
K 0b;B
¼ 0:585 gE ¼ K EF
K 0b;E
¼ 0:158
EF þ
K CB þKP K FE þKP
gC ¼ K CB þK CD þ
CD
K 0b;C
¼ 0:396 gF ¼ K FE þK FG þ
FG
K 0b;F
¼ 0:319
gD ¼ 0 gG ¼ gC ¼ 0:396
Effective length ratios and elastic LE;BC LE;EF
LBC ¼ 1:47 ! P C;BC ¼ 1211 kN LEF ¼ 1:16 ! P C;EF ¼ 865 kN
critical loads using [1] LE;FG
LFG ¼ 1:27 ! P C;FG ¼ 1; 623 kN
Conclusion From inspection column BC is the critical When column EF reaches its critical load, the axial load in column FG is
column and will buckle first (865 2 = 1730 kN > 1623 kN) therefore column FG buckles first.
be negative and the NCCI can overestimate the elastic critical load,
Frame D - No leaning column
as shown in Example 3 below.
3.2.1. Example 3 P P
The NCCI approach has been used to find the elastic critical
loads of column JK in Frame C, Fig. 7, and column NQ in Frame D, Q R
Fig. 8. The results, given in Table 3, highlight the potential errors
encountered by ignoring the possibility of restraints having nega-
tive translational stiffness. All members are 203x203x60 UC and
all connections are fully rigid. Columns buckle about their weak
axis (Iminor = 2070 cm4) whilst the beams bend about their strong
axis (Imajor = 6130 cm4), both in the plane of the frame.
4m
The critical load of column JK found using the NCCI is almost
double that found using Robot. The critical column for Frame C is
LM, so it would be expected that the buckling load of this column
will define frame instability. However, using the NCCI, the effective
length of column LM would be infinity as both ends are pinned and
it is part of a sway frame; therefore it is unable to take any load.
The designer may decide that it is fully braced by column JK, and
as such read from the non-sway design chart, in which case they N S
would obtain an effective length ratio of one, and an elastic critical 4m
load of 2554 kN, which is evidently much too high. The NCCI gives
the designer no options in these cases and without careful thought
Fig. 8. Frame D: Portal frame without a leaning column.
can lead to potentially unconservative errors.
Applying the AISC modification factor (section 2.3) to Frame C, vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uP qffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
the effective length ratio is given by Eq. (32) and the elastic critical u Pu IAB
w0AB ¼t P I ¼ 2w0;AB ¼ 2ð1:12Þ ¼ 1:50 ð32Þ
load by Eq. (33), which is in very good agreement with the com- PuAB w2 0
puter analysis.
p2 EI p2 EI
PC ¼ ¼ ¼ 1163 kN ð33Þ
k L2 1:5 40002
Frame C - Leaning column
P P 3.3. Conclusion
4. Proposal
J M
4m There are numerous proposed improvements to the effective
length method, but none have replaced the method used in BS
Fig. 7. Frame C: portal frame with a leaning column. 5950 [20] and now the NCCI [1]. The principle reason for this is that
140 A. Webber et al. / Engineering Structures 102 (2015) 132–143
Table 3
Results for Example 3.
many of the proposed methods are far too complicated and 4.2. Semi-rigid connections
time-consuming to apply by hand, and consequently are unsuit-
able for preliminary design purposes and impractical to the The assumption of a fully rigid connection implies that no rela-
Design Engineer. tive rotation of the connection occurs and that the end moment of
BS 5950 [20] and the NCCI [1] also provide formulae to deter- a beam is completely transferred to the column. On the other hand,
mine the effective length ratio from distribution coefficients as a pinned connection implies no rotational restraint is provided and
an alternative to design charts. These formulae are more conserva- the moment is zero at the connection.
tive than the design charts, and more precise formulae have been Several non-linear models have been developed that provide a
developed by Smyrell [21] using curve fitting techniques. closer approximation to the true moment-rotation behaviour of
Lui [22] developed a method to determine the effective length connections, which employ curve fitting techniques that require
ratio of framed columns, which explicitly takes into account trans- the input of connection-dependent parameters that have been tab-
lational stiffness by applying a fictitious horizontal force to the ulated in the literature [7,24].
frame. It also allows for the existence of weaker and stronger col-
umns (or leaning columns) in the same storey. The formula has 4.3. New proposal
been shown to provide reliable results when applied to frames
where columns in a single storey are of different strengths [23]. The main source of discrepancy in the current NCCI method
comes from the evaluation of the rotational stiffness of the adjoin-
4.1. Improved rotational stiffness for adjoining beams ing columns. An opportunity to modify the NCCI method, so that
the rotational stiffness of adjoining columns is appropriately con-
Mageirou and Gantes [7] derived the rotational stiffness’ of sidered, has been identified. Distribution coefficients are proposed
members using the slope-deflection method, similar to Wood’s for use with the design charts of NCCI [1]:
derivation [8] of the rotational stiffness of adjoining beams. The K IJ
options for far-end restraint conditions were expanded to include gI ¼ P ð34Þ
K IJ þ K 00XY;I þ K 00b;I
roller supports with various rotational stiffnesses. These options
are applicable to the far-end restraint conditions of columns in
K JI
sway frames and therefore allow an adjoining column’s rotational gJ ¼ P ð35Þ
stiffness to be modified appropriately for the calculation of the dis-
K JI þ K 00XY;J þ K 00b;J
tribution coefficient. The work of Mageirou and Gantes [7] is also where KIJ is the nominal stiffness of the column IJ which is being
applicable for members with semi-rigid connections, in sway, analysed; K 00XY;I and K 00XY;J are the effective rotational stiffness of the
non-sway and partially sway frames. P 00 P 00
adjoining columns at nodes I and J; and K b;I and K b;J are the
Gantes and Mageirou [6] give rotational stiffness values similar
to those in the NCCI [1], as shown in Table 4. Gantes and Mageirou effective rotational stiffness of the beams converging at nodes I
[6] define rotational stiffness as K = M/h, so that the stiffness of a and J, evaluated from Table 4.
fixed end member with no axial load is given by 4EI/L. Wood used Rearranging the distribution coefficients gives Eqs. (36) and
Eq. (8), so the nominal stiffness is I/L. Therefore, the formulae can (37).
P
be made equivalent by dividing those proposed by Gantes and 1 K 00XY;I þ K 00b;I
Mageirou [6] by 4E (ultimately the methods from both Wood [8] 1 ¼ ð36Þ
gI K IJ
and Gantes and Mageirou [6] provide the same results).
Mageirou and Gantes [7] have also improved the accuracy to
which axial load affects the rotational stiffness and their method
is less conservative than the NCCI method, as is shown by the coef- P
ficients of P/PE used in each equation. C
Table 4
Modified rotational stiffness for beams with various far end restraint conditions.
P
Rotational conditions at far end Mageriou and NCCI [1]
Gantes [6]
A
Fixed support (no rotation, no 4 EIL 1 0:33 PPE 1:0 LI 1 0:4 PPE
translation)
Pinned support (free rotation, no 3 EIL 1 0:66 PPE 0:75 LI 1 1:0 PPE
translation) P
Single curvature (rotation equal and 2 EIL 1 0:82 PPE 0:50 LI 1 1:0 PPE
opposite to that at near end, no
B
translation)
Roller fixed support (free translation, EI
1 0:82 PPE –
no rotation)
L D
Double curvature 6 EIL 1 0:16 PPE 1:50 LI 1 0:2 PPE
Fig. 9. Multi-storey frame design loads.
A. Webber et al. / Engineering Structures 102 (2015) 132–143 141
Table 5 K IJ S
Support conditions at far end Effective rotational stiffness of carried-over moment transferred back into the column under
an adjoining column XY, K 00XY consideration.
2
P
0:5 LIXY 1 0:82 Pd;XY LXY
0:7LIJ
opposite to that at near end, no XY d;IJ
To properly evaluate the rotational stiffness of an adjoining col-
translation) umn, its far-end restraint conditions and axial load need to be con-
2
which will be taken as its real length for sway frames, or 0.7 times
The first term of Eq. (38) represents the effective rotational stiff-
its real length for non-sway frames. This ensures the axial load in
ness of the column under consideration, K 00IJ (Eq. (39)).
adjoining columns is not underestimated.
!!
K IJ S
K IJ S The load in columns AC and BD can then be estimated as
K 00IJ ¼ 1 C2 K IJ S
4
P ð39Þ PAC ¼ 0:5PC;AB and P BD ¼ 1:5PC;AB . A general equation to estimate
4 þ K 00XY;J þ K 00b;J
4 the load in an adjoining column (XY) at the point when the critical
Table 6
Test 1 – Non-sway frame results.
Table 7
Calculations for Frame A and Frame B as non-sway frames using the proposed method.
Frame A Frame B
Column BC Column EF
K BC ¼ LI ¼ 20700000
4000 ¼ 5175 K EF ¼ LI ¼ 3450
2
2
P P
00 I
K AB ¼ 0:5 L 1 0:82 Pd;XY
d;IJ
LXY
0:7LIJ ¼ 9809 K 00FG ¼ 0:5 LI 1 0:82 Pd;XY
d;IJ
LXY
0:7LIJ ¼ 1262
2
P K 0b ¼ 0:5 LI ¼ 3065
K 00CD ¼ LI 1 0:33 Pd;XY
d;IJ
LXY
0:7LIJ ¼ 2979
KP
gE ¼ K EF
K 0b;E
¼ 0:360
EF þ
K 0b ¼ 0:5 LI ¼ 3065ðsingle curvatureÞ
K FEP
K BC P gF ¼ K ¼ 0:415
gB ¼ K 00
K 0b;B
¼ 0:245 00
FE þK FG þ K 0b;F
BC þK BA þ
LE;EF
K CB P ¼ 0:642 ! P C;EF ¼ 2827 kN
gC ¼ K 00
K 0b;C
¼ 0:362 LEF
CB þK CD þ
LE;BC
LBC ¼ 0:650 ! P C;BC ¼ 7146 kN
At this stage the accuracy can be improved by substituting the found value for LE into the K’’
formula for the adjoining column and repeating the analysis.
Column CD Column FG
K CD ¼ LI ¼ 20700000
3000 ¼ 6900 K FG ¼ LI ¼ 5175
2
2
P LXY P
00 I
K BC ¼ 0:5 L 1 0:82 Pd;XY
d;IJ 0:7LIJ ¼ 2544 K 00EF ¼ 0:5 LI 1 0:82 Pd;XY
d;IJ
LXY
0:7LIJ ¼ 1523
2
Conclusion: When column BC buckles the axial load in column CD is Conclusion: When column EF buckles, the axial load in column FG is
(7146 1.5 = 10,719 kN < 12,266 kN) therefore column BC is critical (2827 2 = 5654 kN < 5859 kN) therefore EF is critical.
142 A. Webber et al. / Engineering Structures 102 (2015) 132–143
Table 8
Test 2 – Sway Frame Results.
Table 9
Calculations for Frame A and Frame B as sway frames using the proposed method.
Frame A Frame B
Column BC Column EF
K BC ¼ LI ¼ 5175 K EF ¼ LI ¼ 3450
2
2
P P
K 00AB ¼ 0:25 LI 1 0:82 Pd;XY
d;IJ
LXY
LIJ ¼ 5042 K 00FG ¼ 0:25 LI 1 0:82 Pd;XY
d;IJ
LXY
LIJ ¼ 351
2
P K 0b ¼ 1:5 LI ¼ 9195
K 00CD ¼ 0:25 LI 1 0:82 Pd;XY
d;IJ
LXY
LIJ ¼ 532
KP
gE ¼ K EF
K 0b;E
¼ 0:158
EF þ
K 0b ¼ 1:5 LI ¼ 9195ðdouble curvatureÞ
K FEP
K BC P gF ¼ ¼ 0:155
gB ¼ K 00
K 0b;B
¼ 0:181 K FE þK 00FG þ K 0b;F
BC þK BA þ
LE;EF
K CB P ¼ 1:09 ! P C;EF ¼ 979 kN
gC ¼ K 00
K 0b;C
¼ 0:215 LEF
CB þK CD þ
LE;BC
LBC ¼ 1:12 ! P C;BC ¼ 2087 kN
Column CD Column FG
K CD ¼ LI ¼ 20700000
3000 ¼ 6900 K FG ¼ LI ¼ 5175
2
2
P P
00 I
K BC ¼ 0:25 L 1 0:82 Pd;XY
d;IJ
LXY
LIJ ¼ 36 K 00EF ¼ 0:25 LI 1 0:82 Pd;XY
d;IJ
LXY
LIJ ¼ 67
2
K 0b ¼ 9195 P
K 00GH ¼ 0:25 LI 1 0:82 Pd;XY
d;IJ
LXY
LIJ ¼ 532
K CD P
gC ¼ K 00
K 0b;C
¼ 0:272
CD þK CB þ
K 0b ¼ 9195
gD ¼ 0 K FG
gF ¼ K 00
P ¼ 0:219
LE;CD FG þK FE þ K 0b;F
LCD ¼ 1:09 ! P C;CD ¼ 3917 kN
K GF P
gG ¼ K GF þK 00GH þ K 0b;G
¼ 0:215
LE;FG
LFG ¼ 1:135 ! P C;FG ¼ 2032 kN
Conclusion: When column BC buckles the axial load in column CD is Conclusion: When column FG buckles, the axial load in column EF is
(2087 1.5 = 3131 kN < 3817 kN) therefore column BC is critical (979 2 = 1958 kN < 2032 kN) therefore EF is critical.
load in the column under consideration (IJ) is reached for a frame The first three far-end support conditions listed in Table 5 cor-
with any load distribution is given in Eq. (41). respond to columns in non-sway frames, and as such have the 0.7
effective length ratio included. Only the last rotational condition
Pd;XY
PXY ¼ PC;IJ ð41Þ can be used for columns in sway frames.
Pd;IJ
where PXY is the load in the adjoining column, Pd,IJ is the design load 4.4. Testing
of the column that has reached its critical load PC,IJ, and Pd,XY is the
design load of the adjoining column. The proposed method has been used to assess the stability of
In a sway frame, the rotational stiffness of an adjoining column the frames in Figs. 5 and 6. The results are compared to those found
with a fixed roller support (no rotation, free horizontal translation) using Robot and the NCCI for both non-sway (Test 1) and sway
at its far-end can therefore be approximated from Eq. (42) (see (Test 2) conditions.
Table 4).
4.4.1. Test 1 – Non-sway frames
IXY PXY
K 00XY ¼ 0:25 1 0:82 ð42Þ For Frame A in Fig. 5 the proposed method predicts that column
LXY PE;XY
BC buckles first at a load of 7146 kN, which is within 1.3% of the
Substituting Eq. (41) into Eq. (42), Eq. (43) is obtained: load found using Robot.
2 ! It was shown in section 3.1.1 that using the NCCI approach on
IXY P d;XY LXY Frame B in Fig. 6 suggests column FG buckles first at a load of
K 00XY ¼ 0:25 1 0:82 ð43Þ
LXY Pd;IJ LIJ 4653 kN. The method proposed above predicts column EF buckles
first at a load of 2827 kN, at which point column FG would carry
where LXY is the height of the adjacent storey and LIJ the height of a load of 5654 kN. This is within 0.6% of the load found using
the critical storey. This equation assumes the critical column and Robot (5622 kN). The results are summarised in Table 6, with the
the adjoining column have the same EI value. modified calculations shown in Table 7.
In a similar manner for a non-sway frame, the adjoining column
with a fixed far-end can be approximated by Eq. (44). 4.4.2. Test 2 – Sway frames
2 !
I Pd;XY LXY The proposed method predicts column BC in Frame A (Fig. 5) is
K 00XY ¼ 1 0:33 ð44Þ the critical column, and has an elastic critical load of 2087 kN. The
L Pd;IJ 0:7LIJ
NCCI also predicts BC would buckle first, but at the lower load of
The proposed effective rotational stiffness of adjoining columns 1211 kN. The proposed method is in agreement with Robot, which
are summarised in Table 5. obtained a load in column BC at frame instability of 2230 kN.
A. Webber et al. / Engineering Structures 102 (2015) 132–143 143
Similarly, the proposed method predicts column EF will buckle 5.1. Future work
first in Frame B (Fig. 6) at a load of 979 kN, with a predicted axial
load in column FG at frame instability of 1958 kN. This is within More work is required to determine an appropriate analytical
1.3% of that computed by Robot, as summarised in Table 8 with method for assessing the rotational stiffness of semi-rigid connec-
the related calculations in Table 9. tions in frames. New full size laboratory tests carried out on elastic
The proposed approach correctly recognises the contribution frames could be used to observe the buckling modes of
made by an adjoining column to the rotational stiffness. The results multi-storey frames, and measure points of contraflexure on the
obtained from the proposed method have been shown to be in columns, with the aim of calculating their effective lengths.
good agreement to that computed from a finite element eigenvalue Additional work is required to extend the approach to
analysis. three-dimensional frames.
5. Conclusion All data created during this research are openly available from
the University of Bath data archive at <http://dx.doi.org/10.
The current NCCI method consistently underestimates the crit- 15125/BATH-00131>.
ical load of columns in multi-storey frames because of the conser-
vative assumption that adjoining columns buckle simultaneously References
with the column being investigated. Subsequently, adjoining col-
umns only reduce the rotational stiffness of end restraints. A fur- [1] Oppe M, Muller C, Iles C. NCCI: buckling lengths of columns: rigorous
approach. London: Access Steel; 2005.
ther consequence of this assumption is that the stiffer an [2] BSI. 1993-1. Design of steel structures general rules and rules for buildings.
adjoining column, the greater the reduction in rotational stiffness London: BSI; 2005.
of that end restraint. However it is apparent that a stiffer adjoining [3] Simitses GJ. Fundamentals of structural stability. London: Butterworth
Heinemann; 2006.
column is less likely to buckle. [4] Timoshenko SP, Gere JM. Mechanics of materials. PWS Publishing Company;
To address these issues, a simple improvement to the calcula- 1997.
tion of the distribution coefficients has been proposed, which suc- [5] Patnaik SN. Strength of materials: a new unified theory for the 21st
century. London: Butterworth Heinemann; 2004.
ceeds in accurately assessing the rotational stiffness of adjoining [6] Gantes CJ, Mageirou GE. Improved stiffness distribution factors for evaluation
columns by considering the axial load in these columns when of effective buckling lengths in multi-story sway frames. Eng Struct
the column under investigation buckles. This is shown to signifi- 2005;27:1113–24.
[7] Mageirou GE, Gantes CJ. Buckling strength of multi-story sway, non-sway and
cantly improve the accuracy of the results obtained from design
partially-sway frames with semi-rigid connections. J Construct Steel Res
charts without overcomplicating the design or even changing the 2006:62.
design procedure. [8] Wood RH. Effective lengths of columns in multistorey buildings. Struct Eng
1974:52.
In frames with leaning columns, or where stronger columns
[9] Cheong-Siat-Moy F. Column design in gravity-loaded frames. J Struct Eng
brace weaker columns, it is recommended that the NCCI should 1991:117.
adopt AISC LRFD’s modification factor [18], which has been shown [10] Bridge R, Fraser D. Improved G-factor method for evaluating effective lengths
to be a very simple and effective solution. Without this modifica- of columns. J Struct Eng 1987;113:1341–56.
[11] Hellesland J, Bjorhovde R. Improved frame stability analysis with effective
tion, the critical load of frames with leaning columns could be sig- lengths. J Struct Eng 1996;122:1275–83.
nificantly over estimated, which may lead to the under design of [12] Aristizabal-Ochoa J. Stability of columns under uniform axial load with semi-
frames with potentially disastrous results. rigid connections. J Struct Eng 1994;120:3212–22.
[13] Aristizabal-Ochoa J. Story stability of braced, partially braced, and unbraced
A limitation of the NCCI method is that it can only be applied to frames: classical approach. J Struct Eng 1997:123.
multi-storey frames with fully rigid connections. It has been [14] Livesley RK, Chandler DB. Stability functions for structural
demonstrated that the rotational stiffness of adjoining beams can frameworks. Manchester: Manchester University Press; 1956.
[15] Timoshenki SP, Gere JM. Theory of elastic stability. New York: McGraw Hill;
be modified to account for the rotational stiffness of the 1961.
beam-to-column connection. However, difficulty comes in assess- [16] Bazant PZ, Cedolin L. Stability of structures: elastic, inelastic, fracture and
ing the rotational stiffness of the connections. Providing tables damage theories. New York: Oxford University Press; 1991.
[17] BSI. BS 5950-1. Structural use of steelwork in building Code of practice for
matching the different connection types to the corresponding rota-
design – rolled and welded sections. London: BSI; 2000.
tional stiffness, in conjunction with limiting criteria, is one possi- [18] AISC. AISC manual of steel construction. Load and resistance factor design.
bility. Whilst not a perfectly accurate solution, provided they are Chicago: AISC; 2001.
[19] Autodesk. Robot structural analysis professional. San Rafael, CA; 2014.
less than the real stiffness then the method should be suitable
[20] BSI. BS 5950-1. Structural use of steelwork in building code of practice for
and will improve the accuracy of the results. design in simple and continuous construction: hot rolled sections. London, UK:
It is recognised that simplicity is an advantage in design codes. BSI; 2000.
The methods presented here are intended to clarify and extend the [21] Smyrell AG. Approximate formulae for detemining the effective length of steel
columns to BS 5950. Struct Eng 1993.
current design guidance, retaining simplicity whilst achieving [22] Lui EM. A novel approach for K-factor design. ASCE; 1992.
greater accuracy. It is noted that analytical techniques such as [23] Duan L, Chen WF. Structural engineering handbook. Effective length factor of
those shown in this paper are very useful to design engineers at compression members. London: CRC Press; 1999.
[24] Chen WF, Goto Y, Liew JYR. Stability design of semi rigid frames. New
preliminary design stages. However, where very high accuracy is York: John Wiley and Sons; 1996.
required or very complex structures must be analysed, computa-
tional methods are widely available and these can also be used.