The usual method of selecting a beam section is based upon section modulus.
However, the criterion of economy is weight rather than the modulus of section. It
is, therefore, desirable to choose a light beam furnishing the required modulus of
section. Often, it may not be possible to choose the lightest section due to head room
limitations, or if adjacent beams are to be matched. Sometimes deection and occa-
sionally shear may also be the necessary criterion dictating the choice.
fr
Figure 9.2 shows the moment-rotation characteristics of the four classes of cross
sections. The plot is same as that shown in Fig. 3.15, but it also depicts the bending
stress pattern of all the fotu classes of sections. The plastic sections exhibit sufcient
ductility (62 > 6 61, where 61 is the rotation at the onset of plasticity and 62 is the
lower limit of rotation for treatment as a plastic section). The compact sections have
relatively lower deformation capacity than the plastic sections, but are capable of
reaching their fu_ll plastic moment values. For semi-compact sections, the bending
4.50 Limit Stare Llesign of Steel Siructttres
resistance islimited to the (elastic) yield moment only and for slender members local
or lateral buckling occurs in the elastic range." The design moment capacity Md for
the classes of sections dened above is indicated on the plot itself (of course partial
factor of safety for materials should-also be applied to the design moment capacity).
\~
I, -u__ --
i;-it E51
Tension
(a) Simple beam (h) Deection n transverse plane
I/
_
Lateral deflection
Horizontal deflection
Vertical deection
--|.
|< L >|
(c) Buckling of compression flange (d) Buckled position at mid-span
permanent formwork sits on the top flanges and thus needs to be considered as a
destabilising load.
Whichever type of formwork, the weight at the wet concrete stage imposes quite
high stresses in the top flanges of the girders: their strength and stability at this
stage require a detailed evaluation of the progressive changes in structural
behaviour as load is added.
The weight of the concrete cantilevers needs particular attention, because of the
moment (about the longitudinal axis) that is imposed on the outer girder. See
further comment in Section 7.2.3.
Class 4 cross sections are most likely to be encountered at the bare steel stage
in midspan regions, because the web depth/thickness ratio is likely to be high
and at that stage more than half of the web will be in compression, assuming
that the top flange is smaller than the bottom. At an intermediate support, the
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
composite cross section for a deep girder may also be Class 4. For very deep
slender webs longitudinal stiffeners are sometimes provided to improve the
effectiveness of the section, although such stiffeners add considerably to
fabrication costs.
For Class 4 cross sections, the effective cross section is given by EN 1993-1-5,
4.3. The rules allow the determination of an effective width (less than the
gross area) of a panel of a slender web; the difference between the gross area of
the web and the effective area is treated as a hole for the determination of
section properties (see Figure 7.1). Both the size and position of this hole are
given by 4.3, according to the slenderness and the variation of longitudinal
stress across the panel width. The rules for effective areas also cover
longitudinally stiffened sections.
For compression flanges of small slenderness, Equation (12.21) gives lateral torsional buckling lengths
that are excessive because the plastic deformations that the strut may be subjected to reduce its bending
stiffness. For such cases, a more realistic approach is to replace the elastic modulus E by a reduced modu
lus Ered given by the following expression:
(12.22)
Using a reduced modulus requires an iterative approach. Starting with a choice of lD values for Xp
(12.18) and oD (12.14) are calculated. The initial choice of lD is then compared with the value given by
(12.21), and so on.
The displacement v can be calculated using the equations given in Table 14.8 for frame cross bracing.
The greatest value of v, considering those for symmetric and antisymmetric lateral torsional buckling,
governs. In the case of a trough bridge, or when the cross girder is located in the upper part of the cross
section, the displacement v may be calculated to a good approximation using the following relationship,
with notation as defined in Figure 12.3(b):
(12.23)
The uprights of the cross bracing are subject to horizontal forces corresponding to their function as
lateral supports for the compression members. Although this force is theoretically zero for a perfect strut,
numerical analyses that include initial imperfections of the compression member show that the force
needed to provide lateral restraint is around 1% of the compression force acting in the strut itself. For
bridges this force corresponds to the axial force acting in the flange plus part of the web, which has an area
defined by Equation (12.12).
(12.24)
a
ult,k minimum amplification factor to apply to the design loads in order to reach the char
acteristic resistance of the cross section, without taking into account lateral torsional
buckling
( minimum amplification factor to apply to the design loads in order to reach the elastic
^cr,op
critical resistance to lateral buckling of the strut under compression
STEEL BEAMS 291
The critical stress, or the factor cncrop, may be calculated using software that permits variations in the
geometry of the compression flange, and the support conditions due to the cross bracing, to be easily intro
duced. Once the noji-dimensional slenderness is known, then the reduction factor is calculated using
(12.16), replacing by . This then allows the lateral torsional buckling stress (12.14) and the lateral
torsional buckling moment resistance (12.15) to be calculated.
Finally, it should be noted that the general method of a second order elastic analysis of the load car
rying structure, combined with checks on cross sectional resistances, may always be used. This method
usually requires a finite element program that allows equivalent geometric imperfections of the member
to be introduced, and performs a step-by-step calculation that takes into account second order effects. It
is generally assumed that the form of these geometric imperfections corresponds to one of the instability
modes of the system and, according to the Eurocode, the imperfection has a magnitude of //150 (buckling
curve d), where / is the distance between points of inflection for the deformed shape corresponding to the
instability being considered.
(a) Distribution of stresses (b) Deformed section (c) Effective section with
measured during a test a "hole" in the web
Fig. 12.4 Distribution of measured stresses in a plate girder with a slender web [12.3].
To study the bending moment resistance of a beam comprising thin plate elements, and taking into
account the redistribution of stresses noted above, the so-called effective width method is used. This is the
method that was considered in TGC Volume 10, Section 12.3, for the study of plates loaded in compres
sion. It is assumed that the web in compression only resists over an effective depth hce^ which is distrib
uted half adjacent to the compression flange and half adjacent to the neutral axis (Fig. 12.4(c)).
292 ANALYSIS AND DESIGN (BEAM BRIDGES)
(12.25)
A doubly symmetric section subject to a bending moment is stressed the same in compression as it is in
tension, and the buckling coefficient k equals 23.9, therefore:
(12.26)
resulting in:
hf/tw < 126 for S235 steel,
hf/tw< 102 for S355 steel.
If the web slenderness exceeds these limits, which are the same as those given in Table 9 of the stand
ard SIA 263, only the effective depth hceff\s taken into account when calculating the bending moment
resistance. This effective depth is defined by the inequality (12.25) which becomes, for a doubly symmetric
section with hc = hf/2:
(12.27)
(12.28)
For a web with linear, pinned supports (provided by the two flanges), the buckling coefficient k is
defined as follows:
(12.29)
: ratio (including its sign) between the minimum and maximum stresses, = oinflosup,
according to figure 12.5(a).
The effective depth hcejj, which contributes to the section bending resistance up to collapse, is placed
in the parts of the web that do not buckle. These are the regions adjacent to the neutral axis and the
STEEL BEAMS 293
Fig. 12.5 Effective section and distribution of stresses for a singly-symmetric plate girder in bending.
compression flange. The distribution of hce^is approximately 0.6 /zc ^ n e x t to the neutral axis and 0.4 hc^
next to the flange. As a simplification, however, it can be assumed, as is suggested in the SIA standard, that
the effective depth distributes equally between these two regions.
The effective section, therefore, looks like that shown in Figure 12.5(b). This reduced section, with a
"hole" in the web, should be used when calculating the bending moment resistance of the section. This fig-
ure also shows the reduced area of the compression flange. Unless the flange is fully effective, this reduced
area should be considered when calculating the elastic section modulus Weff.
The reduced section, with a "hole" in the web, implies that a revised neutral axis position should be
determined, rather than using that for the gross section. The neutral axis moves towards the tension flange;
therefore, determining the bending moment resistance of the effective section requires calculation of a
revised neutral axis position. A revised value of the second moment of area /^should also be calculated.
The distance e between the neutral axis of the gross section and that of the effective section may be deter-
mined using the following equation:
(12.30)
The second moment of area IejfOf the effective section may be determined using the following equation:
(12.31)
In second moment of area of the gross cross section of the steel beam
Equations (12.30) and (12.31) are only valid provided the compression flange is totally effective. If that
is not the case, then the characteristics of the effective section can still be obtained by adding to the gross
section those areas that are ineffective, as negative areas.
The ultimate bending moment resistance MR about the major axis of such a plate girder, with a reduced
area to allow for local buckling of the compression elements, and with the stress in the compression flange
limited to the lateral torsional buckling stress, is given by:
294 ANALYSIS AND DESIGN (BEAM BRIDGES)
(12.32)
with
(12.33)
Equation (12.32) is only valid for calculating the ultimate bending resistance if hc + e > hj/2. Otherwise,
this resistance should be calculated using the elastic section modulus Wte^-with respect to the mid-thick
ness of the tension flange of the effective section. When lateral torsional buckling is not critical (< 0.4),
fy replaces oD in Equation (12.32).
When the resistance of a plate girder is limited by lateral torsional buckling (aD <fy), the yield strength
fy may be replaced by JoD / , when calculating the effective widths bejf (12.9) and hcejf (12.28). This
approach takes into account the fact that when lateral torsional buckling occurs, the average bending stress
in the compression flange is less than the yield strength of the steely. It leads to less conservative values
of the effective widths, but does require iteration.
Fig. 12.6 Effective section and distribution of stresses for a plate girder with a longitudinal stiffener.
STEEL BEAMS 295
the section. The calculation of this resistance is based on the same concept of effective widths as discussed
elsewhere, with an effective width being calculated for each sub-panel of the web. These sub-panels (hci
and hc2 in Fig. 12.6(a)) are bounded by the compression flange and the longitudinal stiffener (or even two
such stiffeners). A nominally pinned support is assumed at each edge of the panel. If the longitudinal stiff-
eners possess sufficient torsional stiffness (box section stiffener formed from a channel section), the panels
may be considered to be built-in at these supports.
One condition to note, which is essential for this approach to be valid, is that the longitudinal stiff-
eners must be able to provide adequate support to the web as it buckles, in both the pre-buckling and
post-buckling states. This means that the stiffeners must not only be sufficiently resistant, they must also be
sufficiently stiff in bending to provide a true lateral support for the web panels without displacing laterally
themselves. They must be able to do this up to failure of the beam. The conditions that must be fulfilled by
these stiffeners with regard to resistance and stability are given in Paragraph 12.6.4.
The effective depth hci^oi each web sub-panel / (Fig. 12.6(b)) may be calculated using Equation
(12.28):
(12.34)
The yield strength fy may be replaced in this equation by the maximum stress acting on the edges of
the panel being considered. This stress (Jmaxi is calculated for the total area of the web (no loss of effective
area), although any reduction in the compression flange area is taken into consideration. The buckling
coefficient k is defined by Equation (12.29) for a web sub-panel that is simply supported at its edges.
The factor y/ represents the ratio between the stresses omini and omaxi that act at the edges of the panel.
Half the effective width hc>ie^is assumed to be distributed to each edge of the panel /, as shown in Figure
12.6(b).
It is beneficial to use a table when calculating the effective modulus WeffOf such a section. The area of
the longitudinal stiffeners may be taken into account, although their contribution to bending resistance is
negligible.
Given its significant contribution to the bending resistance, it is important that the total flange area is effec
tive for a plate girder. This is not always possible for box girders.
4. Check lateral torsional buckling of the beam (or lateral buckling of the compression flange) by
determining the lateral torsional buckling stress aD in the compression flange (12.14). If this stress
is less t h a n / r then lateral torsional buckling will occur before the ultimate bending resistance can
be achieved. In order to increase the lateral torsional buckling stress, either the second moment
of area (in the transverse sense, about the z axis of the section) of the compression flange can be
increased, or the distance between the lateral supports to the flange could be reduced.
5. Check the slenderness of the web to establish if a reduced effective depth should be used when
calculating the elastic section modulus Wc ^(12.33).
If Checks 3 and 5 do not necessitate either a reduction in the flange width or the web depth, then the
bending moment resistance MR is defined by (12.1). If this is the case, then lateral torsional buckling
should be checked as with rolled sections, taking into account the conditions given in Table 6 of the stand
ard SIA 263 for an EE calculation method.
If Checks 3 and 5 do require a reduction of the effective widths, then the resistance MR is defined by
(12.32).
6. Check the structural safety of the section, taking into account the resistance factor :
(12.35)
M
Ed design value of bending moment M
M
Rd design value of bending moment resistance
MR bending moment resistance
7a resistance factor for steel
Determine the bending moment resistance of the beam assuming that the upper flange is restrained
laterally, by rigid supports, every 10 m.
Web breathing
According to Equation (12.88) for slender beams supporting variable loads, and with a doubly
symmetric cross section, hc = hj/2
STEEL BEAMS 297
>OK
Compression flange
According to Equation (12.8), for a doubly symmetric cross section, check that the total area of the
compression flange contributes to the moment resistance of the section:
> Not OK
Therefore, one must define an effective width of the compression flange according to (12.9):
Web in bending
According to Equation (12.26), for a doubly symmetric cross section, check that the total area of the
web contributes to the moment resistance of the section:
> Not OK
One must define, therefore, an effective depth of web, either using (12.27) for a doubly symmetric
cross section, or the general Equation (12.28) with k = 23.9:
298 ANALYSIS AND DESIGN (BEAM BRIDGES)
The neutral axis is located at a distance e = 1010 - 926 = 84 mm from the y axis, in the direction of the
tension flange. The elastic modulus of the effective section is given by:
This modulus is 15% less than that of the gross section. It would only be 9% less if the total area of the
compression flange were able to contribute (which could be achieved by adopting a flange with an identical
area but different dimensions, such as 500 x 24).
However, the bending resistance of the beam may be limited by lateral torsional buckling. This cor-
responds to lateral buckling of the compression flange (about the z-z axis), which may be considered as a
strut with a cross sectional area equal to the effective area of the compression flange plus that part of the
effective depth of the web located next to this flange. This strut has the following properties:
Design Examples of Steel and Steel-Concrete Composite Bridges 233
Dead Loads
Live Loads
Considering the axle loads on the bridge components according to Load
Model 71 (see Figure 4.4), two cases of loading for the evaluation of maxi-
mum bending moment due to live loads on a main plate girder can be studied.
30.0 m
YA = 414 kN YB = 414 kN
414 kN
+
S.F.D.
414 kN
B.M.D.
+
3105 kN.m
Figure 4.12 Straining actions from dead loads acting on one main plate girder.
234 Ehab Ellobody
30.0 m
I.L. for
B.M.D.
5.5
6.3 5.9
6.7 6.7
7.5
Figure 4.13 Determination of the maximum bending moment on one main plate girder
due to live loads using the influence line method (case of loading 1).
The first case of loading is that the centerline of the main plate girder is located
under one of the intermediate concentrated live loads, with maximum bend-
ing moment calculated at midspan (see Figure 4.13). On the other hand, the
second case of loading is that the centerline of a main plate girder divides the
spacing between the resultant of the concentrated live loads and the closest
load, with maximum bending moment located at the closest load (point a
in Figure 4.14). The maximum bending moment under the first case of load-
ing is calculated using the influence line method (by multiplying the concen-
trated loads by the companion coordinates on the bending moment diagram
and by multiplying the distributed loads by the companion areas on the bend-
ing moment diagram), while that under the second case of loading is calculated
0.95 0.89
1.0 0.84 0.81
I.L. for
S.F.D.
Figure 4.15 Determination of the maximum shear force on one main plate girder due
to live loads using the influence line method (case of loading 1).
analytically using structural analysis. Hence, the bending moments due to live
loads can be calculated as follows:
ML:L: case of loading 1 250 2 6:3 + 2 7:1 + 2 80 0:5
11:8 5:9
12295:2 kN m
ML:L: case of loading 2 1437:5 14:6 80 12:2 8:5 250 1:6
12291:5 kN m
There is only a single case of loading for the live loads to produce a maxi-
mum shear force at the supports of a main plate girder, which is shown in
Figure 4.15. Once again, we can use the influence line method to calculate
the maximum shear force due to this case of loading or analytically by taking
moment at support B and evaluate the reaction at A:
QL:L: 1713:8 kN
Dynamic Factor F
LF 30 m
2:16
F3 p + 0:73 1:139, F3 1:0 and 2:0:
30 0:2
Bending Moment Due to Dead and Live Loads with Dynamic Effect
Added (MD+L+F)
MD + L + F MD:L: gg + F ML:L: gq
3105 1:2 + 1:139 12,295:2 1:45 24,032 kN m
236 Ehab Ellobody
Shearing Force Due to Dead and Live Loads with Dynamic Effect
Added (QD+L+F)
QD + L + F QD:L: gg + F QL:L: gq
414 1:2 + 1:139 1713:8 1:45 3327:2 kN
56
3
3
0.495 28.4
= 38
0.695 = 57
300 cm
298.4
1.6
60
Figure 4.16 Reduced cross-section of plate girder.
Design Examples of Steel and Steel-Concrete Composite Bridges 237
plate of the upper and lower flanges of the cross section is taken as 0.2
the web height, which is equal to 600 mm, while the top plate width is
taken as 560 mm, to allow for welding with the bottom flange plate. The
flange plates have the same plate thickness of 30 mm. The choice of two
flange plates for the upper and lower flanges is intended to curtail the
top flange plate approximately at quarter-span as detailed in the coming sec-
tions. It should be noted that the web height value (L/10) is an acceptable
recommended [1.9] value for railway steel bridges constructed in Great Brit-
ain and Europe. This value is an initial value for preliminary cross-sectional
estimation. The cross section has to be checked, classified, designed, and
assessed against deflection limits set by serviceability limit states. To classify
the cross section chosen,
s r
235 235
e 0:924
fy 275
ks 23:9
lp 300=1:6
p 1:462 > 0:673
28:4 0:924 23:9
1:462 0:0553 1
r 0:633
1:4622
beff 0:633 300=2 95 cm,
56
3
3
38
55
162.8
57
y y
300 cm
y1 y1
1.6
149.2
150
y0 y0
60
Figure 4.17 Calculation of properties of area for main plate girder.
56 3 1:5 + 60 3 4:5 + 60 3 307:5 + 56 3 310:5 + 207
1:6 109:5 + 38 1:6 287
yc
1088
yc 149:2 cm
moment is decreased towards the supports. Therefore, we can stop the top
flange plate at a certain distance to get the most benefit from the material.
This process is commonly called as curtailment (transition) of flange plates.
It should be noted that, theoretically, curtailment (transition) of flange plates
can be conducted by reducing the flange plate width, thickness, or both.
However, in practice, fabricators prefer to keep the flange widths constant
and vary the thickness because this option costs much less than reducing the
flange width that might require a very heavy grinding work. To avoid lateral
torsional buckling of the compression top flange at the reduction zone, it is
recommended practically to reduce the width or thickness by 40% of the
original with a smooth transition zone sloping at 1 (vertical) to 10 (horizon-
tal). It is also recommended that bridges with lengths of 20-30 m are cur-
tailed (transitioned) in one step. While for bridges with spans greater than
30 m, two steps of curtailment (transition) are recommended. For the inves-
tigated design example, we can conduct one-step curtailment (transition) by
reducing the top flange plate of the upper and lower flanges, as shown in
Figure 4.18. To classify the reduced cross section,
s r
235 235
e 0:924
fy 275
38
55
162.9
57
y y
300 cm
y1 y1
1.6
143.1
150
y0 y0
60
Figure 4.18 Calculation of properties of area for curtailed main plate girder.
Design Examples of Steel and Steel-Concrete Composite Bridges 241
lp 300=1:6
p 1:462 > 0:673
28:4 0:924 23:9
1:462 0:0553 1
r 0:633
1:4622
18,804.4
xm
15 m
Figure 4.19 Calculation of curtailed flange plate lengths.
P357 Composite Highway Bridge Design: Worked Examples
For the span girder, the values for the elastic moduli of the gross section will be used
for build up of stresses in the span girder during construction, since classification for
the total stresses will be at least Class 3.
For the pier section, gross section properties will be used for the build up of bending
stresses (since the section is Class 3 in bending) but an effective area will be used the
effects of any axial compression at this stage, since the bare section is Class 4 in
compression (for which Aeff = 105000 mm2)
beff
There is thus a hole in the web 611 582 = 29 mm long centred 913 mm above the soffit
Span girder
Height of NA 547 (mm)
Second moment of area Iy 2.510E+10 (mm4)
Elastic modulus, centroid top flange Wbf,y 3.965E+07 (mm3)
Elastic modulus, centroid bottom flange Wtf,y 4.808E+07 (mm3)
NOTE 3 Elastic-plastic shear lag effects allowing for limited plastic strains may be taken into account
using Aeff as follows:
A eff = A c ,eff A c ,eff (3.5)
The expressions in NOTE 2 and NOTE 3 may also be applied for flanges in tension in which case
Ac,eff should be replaced by the gross area of the tension flange.
4 Plate buckling effects due to direct stresses at the ultimate limit state
4.1 General
(1) This section gives rules to account for plate buckling effects from direct stresses at the ultimate limit
state when the following criteria are met:
a) The panels are rectangular and flanges are parallel or nearly parallel (see 2.3).
b) Stiffeners if any are provided in the longitudinal or transverse direction or both.
c) Open holes or cut outs are small (see 2.3).
d) Members are of uniform cross section.
e) No flange induced web buckling occurs.
NOTE 1 For compression flange buckling in the plane of the web see section 8.
NOTE 2 For stiffeners and detailing of plated members subject to plate buckling see section 9.
(1) The resistance of plated members may be determined using the effective areas of plate elements in
compression for class 4 sections using cross sectional data (Aeff, Ieff, Weff) for cross sectional verifications
and member verifications for column buckling and lateral torsional buckling according to EN 1993-1-1.
(2) Effectivep areas should be determined on the basis of the linear strain distributions with the
attainment of yield strain in the mid plane of the compression plate.
(1) In calculating longitudinal stresses, account should be taken of the combined effect of shear lag and
plate buckling using the effective areas given in 3.3.
(2) The effective cross sectional properties of members should be based on the effective areas of the
compression elements and on the effectives area of the tension elements due to shear lag.
(3) The effective area Aeff should be determined assuming that the cross section is subject only to stresses
due to uniform axial compression. For non-symmetrical cross sections the possible shift eN of the centroid of
the effective area Aeff relative to the centre of gravity of the gross cross-section, see Figure 4.1, gives an
additional moment which should be taken into account in the cross section verification using 4.6.
(4) The effective section modulus Weff should be determined assuming the cross section is subject only to
bending stresses, see Figure 4.2. For biaxial bending effective section moduli should be determined about
both main axes.
NOTE As an alternative to 4.3(3) and (4) a single effective section may be determined from NEd and
MEd acting simultaneously. The effects of eN should be taken into account as in 4.3(3). This requires an
iterative procedure.
13
prEN 1993-1-5 : 2004 (E)
(5) The stress in a flange should be calculated using the elastic section modulus with reference to the mid-
plane of the flange.
(6) Hybrid girders may have flange material with yield strength fyf up to hfyw provided that:
a) the increase of flange stresses caused by yielding of the web is taken into account by limiting the stresses
in the web to fyw
b) fyf (rather than fyw) is used in determining the effective area of the web.
NOTE The National Annex may specify the value h. A value of h = 2,0 is recommended.
(7) The increase of deformations and of stresses at serviceability and fatigue limit states may be ignored
for hybrid girders complying with 4.3(6) including the NOTE.
(8) For hybrid girders complying with 4.3(6) the stress range limit in EN 1993-1-9 may be taken as 1,5fyf.
1
G 2
G
14
prEN 1993-1-5 : 2004 (E)
(1) The effectivep areas of flat compression elements should be obtained using Table 4.1 for internal
elements and Table 4.2 for outstand elements. The effectivep area of the compression zone of a plate with the
gross cross-sectional area Ac should be obtained from:
Ac,eff = Ac (4.1)
where is the reduction factor for plate buckling.
p 0,055 (3 + )
= 2
1,0 for p > 0,673 , where (3 + ) 0 (4.2)
p
p 0,188
= 2
1,0 for p > 0,748 (4.3)
p
fy b/t
where p = =
cr 28,4 k
is the stress ratio determined in accordance with 4.4(3) and 4.4(4)
b is the appropriate width to be taken as follows (for definitions, see Table 5.2 of EN 1993-1-1)
bw for webs;
b for internal flange elements (except RHS);
b - 3 t for flanges of RHS;
c for outstand flanges;
h for equal-leg angles;
h for unequal-leg angles;
k is the buckling factor corresponding to the stress ratio and boundary conditions. For long plates k is
given in Table 4.1 or Table 4.2 as appropriate;
t is the thickness;
cr is the elastic critical plate buckling stress see equation (A.1) in Annex A.1(2) and Table 4.1 and Table
4.2;
235
=
[
f y N / mm 2 ]
(3) For flange elements of I-sections and box girders the stress ratio used in Table 4.1 or Table 4.2
should be based on the properties of the gross cross-sectional area, due allowance being made for shear lag in
the flanges if relevant. For web elements the stress ratio used in Table 4.1 should be obtained using a stress
distribution based on the effective area of the compression flange and the gross area of the web.
NOTE If the stress distribution results from different stages of construction (as e.g. in a composite
bridge) the stresses from the various stages may first be calculated with a cross section consisting of
15
prEN 1993-1-5 : 2004 (E)
effective flanges and gross web and added together. This resulting stress distribution determines an
effective web section that can be used for all stages to calculate the final stress distribution for stress
analysis.
(4) Except as given in 4.4(5), the plate slenderness p of an element may be replaced by:
com ,Ed
p ,red = p (4.4)
f y / M0
where com,Ed is the maximum design compressive stress in the element determined using the effectivep
area of the section caused by all simultaneous actions.
NOTE 1 The above procedure is conservative and requires an iterative calculation in which the stress
ratio (see Table 4.1 and Table 4.2) is determined at each step from the stresses calculated on the
effectivep cross-section defined at the end of the previous step.
(5) For the verification of the design buckling resistance of a class 4 member using 6.3.1, 6.3.2 or 6.3.4 of
EN 1993-1-1, either the plate slenderness p or p ,red with com,Ed based on second order analysis with
global imperfections should be used.
(6) For aspect ratios a/b < 1 a column type of buckling may occur and the check should be performed
according to 4.5.3 using the reduction factor c.
NOTE This applies e.g. for flat elements between transverse stiffeners where plate buckling could be
column-like and require a reduction factor c close to c as for column buckling, see Figure 4.3 a) and
b). For plates with longitudinal stiffeners column type buckling may also occur for a/b 1, see Figure
4.3 c).
16
prEN 1993-1-5 : 2004 (E)
bt bc < 0:
1
beff = bc = c / (1-)
2
b eff
= 2/1 1 0 -1 1 -3
Buckling factor k 0,43 0,57 0,85 0,57 - 0,21 + 0,072
b eff
1 > 0:
1
2
beff = c
c
b eff
< 0:
1
2 beff = bc = c / (1-)
bc bt
17
P325: Introduction to Steelwork Design to BS 5950-1:2000
Discuss me ...
BS 5950-1
Example 3
Consider a 250 150 5.0 hot-finished rectangular hollow section
grade S355, subject to bending about its major axis.
275 275
= = = 0 .88
py 355
The b/t limit for a Class 1 plastic hot-finished flange is 28 but 80 d/t. Table 12
Therefore, the b/t limit for a Class 1 plastic flange is 23.4, which is less than
27. Therefore, the flange is not Class 1 plastic.
Therefore, the b/t limit for a Class 2 compact flange is 28.2, which is greater
than 27. Therefore, the flange is Class 2 compact.
The d/t limit for a Class 1 plastic hot-finished web with the neutral axis at Table 12
mid-depth is 64 = 56.3, which is greater than 47. Therefore, the web is
Class 1 plastic.
Example 4
Consider the same HF RHS (250 150 5.0 S355) subject to a compressive
axial load of 1100 kN and a bending moment about the major axis.
The flange classification limits are the same as in Example 3. Hence, the Table 12
flange is Class 2 compact.
The web is unlikely to be Class 1 plastic with the section subject to 1100 kN
of axial load. Therefore, check the Class 2 compact limit.
22 March
Created on 30 2011
July 2009
Discuss me ...
BS 5950-1
The d/t limit for a Class 2 compact hot-finished web generally is, Table 12
80 Fc
but 40 where r1 = but 1 < r1 1
1 + r1 2 d t p yw
3
1100 10
r1 = = 1.32 > 1 therefore take r1 = 1.0.
2 235 5 .0 355
80 80 0.88
The Class 2 compact d/t limit = = = 35.2
1 + r1 1+1
but 40 = 35.2.
The web d/t equals 47, which is greater than the limit of 35.2, therefore the
web is not Class 2 compact.
120 Fc
but 40 where r 2 =
1 + 2 r2 A g p yw
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
3
1100 10
r2 = = 0 .80 .
2
38 . 7 10 355
The section therefore has a Class 2 compact flange and a Class 4 slender web
when subject to an axial load of 1100 kN. In these circumstances the section
is Class 4 slender.
Discuss me ...
BS 5950-1
The code gives formulae for calculating Seff for various sections. For an I or
H section with equal flanges the formulae are given as:
2 Cl. 3.5.6.2
3w 3f
1 1
d/t
but S x,eff Z x + (S x Z x ) b/T
S x,eff = Z x + (S x Z x )
2 3f
3w 1 1
2f
2w
3f
1
= Z y + (S y Z y )
b/T
S y,eff
3f 1
2f
where:
2f is the limiting value of b/T for a Class 2 compact flange
2w is the limiting value of d/t for a Class 2 compact web
3f is the limiting value of b/T for a Class 3 semi-compact flange
3w is the limiting value of d/t for a Class 3 semi compact web
Sx and Sy are the plastic moduli
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
Effective Area
The effective area is used in determining the compression resistance of a Class Cl. 3.6.2.2
4 slender section (see Section 5.1).
Discuss me ...
BS 5950-1
Figure 8a
Non-effective zone
20 t
15T 15T
20 t
15T 15T
1.5 t 2.5 t
20 t 17.5t
20 t 17.5t
1.5 t 2.5t
Table 3.4 summaries the various cases to consider for calculating the effective
elastic modulus and describes how the effective elastic modulus should be
calculated for a doubly symmetric section.
22 March
Created on 30 2011
July 2009
Discuss me ...
BS 5950-1
Table 3.4 Summary of effective elastic modulus calculation
Section subject to bending Section subject to axial load
and bending
Flange slender Flange not Flange slender Flange not
slender slender
Web slender under Use Figure 8b Use Figure Use Figure 8b Use Figure
pure bending for flange. Use 9 for web. for flange. Use 9 for web.
Figure 9 for Figure 9 for
web. web.
Web only slender Use Figure 8b Zeff equals
under combined for flange. Treat Z.
n/a n/a
axial load & web as fully
bending effective.
Web not slender Use Figure 8b Zeff not Use Figure 8b Zeff not
under pure bending for flange. required. for flange. required.
Notes: Figure 8b of BS 5950-1 is reproduced in part here as Figure 3.7.
Figure 9 of BS 5950-1 is reproduced in part here as Figure 3.8.
When the webs are fully effective, the shaded parts of the flanges shown in
Figure 3.7 are ineffective and should be disregarded when calculating the
effective elastic modulus.
15 t 15 t Figure 8b
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
Compression flange
15 t 15 t
Tension flange
Compression flange
Tension flange
If the web is slender under pure bending an ineffective portion of the web
needs to be determined as shown in Figure 3.8. An iterative process can be
used to determine the ineffective portion of the web because the size of the
ineffective portion is dependent on the position of the elastic neutral axis of
the effective section, which is in turn dependent on the size of the ineffective
2011
portion of the web. Figure 3.8 shows that the size and the position of the
July 2009
Discuss me ...
BS 5950-1
120 t Cl. 3.6.2.4
b eff =
f f tw f
1 + cw 1 + tw
p yw f cw
where:
fcw is the maximum compressive stress in the web subject to pure
bending (should always be taken as positive)
ftw is the maximum tensile stress in the web subject to pure bending
(should always be taken as positive)
pyw is the design strength of the web
t is the web thickness.
The expression for beff simplifies to 60t for a doubly symmetric sections
subject to pure bending.
Figure 9
Compression flange
f cw
0.4 b eff
Non-effective
zone
Effective neutral axis 0.6 b eff
of gross section Effective neutral axis
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
of effective section
f tw
Tension flange
Figure 3.8 Effective width for slender web under pure bending
When calculating effective properties for Class 4 slender sections which are
subject to axial load and bending, each of the load effects should be
considered separately, i.e. the effective area should be calculated assuming the
section is subject to axial load only, the effective modulus about the major
axis should be calculated assuming the section is subject to bending about the
major axis only and the effective modulus about the minor axis should be
calculated assuming the section is subject to bending about the minor axis
only.
Discuss me ...
BS 5950-1
Reduced design strength
As an alternative to calculating effective section properties for Class 4 slender Cl. 3.6.5
sections, a reduced design strength pyr can be calculated (in accordance with
Clause 3.6.5) for which the cross-section would be Class 3. The reduced
strength can the be used with the gross section properties. This approach can
be simpler than calculating effective section properties but can lead to
conservative results.
The effective plastic modulus about the major axis is obtained from:
2 Cl. 3.5.6.2
3w 3f
1 1
d/t
but S x,eff Z x + (S x Z x ) b/T
S x,eff = Z x + (S x Z x )
2
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
3f 1
3w 1 2f
2w
For a 457 152 52 UB, Zx = 950 cm3, Sx = 1100 cm3, d/t = 53.6 mm and
b/T = 6.99 mm.
3f = 15 = 15 1.0 = 15
2f = 10 = 10 1.0 = 10
63.8 2
1
53.6 0 .417
S x,eff = 950 + ( 1100 950 ) = 950 + 150 = 996 cm
3
2
63.8 1.36
1
41.5
15
1
6 .99 1.15
but S x,eff 950 + ( 1100 950 ) = 950 + 150 = 1295 cm
3
15 0 .5
1
2011
10
July 2009
22 March
Discuss me ...
BS 5950-1
Example 2
Consider the 250 150 5.0 hot-finished RHS grade S355 from Example 4,
Section 3.3, subject to a compressive axial load of 1100 kN and a bending
moment.
In Example 4, Section 3.3, the section has been shown to have a Class 2
compact flange and a Class 4 slender web. Therefore, the effective area Aeff
and effective elastic modulus Zx,eff are required.
The effective area should be taken as shown in Figure 8a of the code and
Figure 3.6 and Figure 3.9 of this publication. Only the webs are Class 4
slender and therefore only the webs have ineffective zones.
1.5 t
20 t
Non-effective
zone
20 t
1.5 t
The effective width of the web for each web is taken as,
The total effective area, Aeff = Ag 2 295 102 = 38.7 5.9 = 32.8 cm2
Example 3
Consider the fabricated section grade S275 shown in Figure 3.10 subject to
pure bending. The flanges are Class 1 plastic but the web is Class 4 slender.
Therefore, the effective elastic modulus is required.
To determine the effective elastic modulus, the effective width beff of the
compression zone of the web must be calculated.
For a doubly symmetric section subject to pure bending, beff equals 60t.
Therefore, for the section in Figure 3.10, beff = 60 1.0 8 = 480 mm.
22 March
Created on 30 2011
July 2009
Discuss me ...
BS 5950-1
60
10
Properties
d/t = 125
py = 275 N/mm2
= 1.0
1000
Ix = 97270 cm4
Zx = 1907 cm3
10
10
This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
0.6 b =288
eff
y 1020
4
y
3
y
y bar
2
10
y
1
The d/t Class 3 semi-compact limit for a web with the neutral axis at mid- Table 11
depth is 120. The d/t of the fabricated section is 125, therefore try x equal
to 5t, which is 40 mm.
22 March
Created on 30 2011
July 2009
Discuss me ...
BS 5950-1
( y a )
ybar =
A eff
ybar =
(y a)
Aeff
B D3 B D3 B D3
Ix,eff = + an2 + an2 + + an2
12 Whole 12 Ineffective 12
web web Flanges
8 1000 3 8 28 3
= + 8000 7 2 + 28 8 302 2 + ...
12 12
... +
(
60 1020 3 1000 3)
+ 2 60 10 7 2
12
3. Determine the classification of each element of the section for the design Table 11
loadings. Determine the overall section classification. Table 12
Created on 30
Class 1: Sections develop their plastic bending resistance and have sufficient rotation
capacity.
Class 2: Sections develop their plastic bending resistance but have limited rotation
capacity.
Class 3: Sections develop their elastic bending resistance.
Class 4: Sections are subjected to local buckling and have a resistance lower than the
elastic resistance.
Since local buckling is crucial for cross sections in developing their strength and ductility,
the width to thickness ratios (c/t) of the compressed walls serves as the criterion for the clas-
sification. To classify a cross section under a combination of an axial force and a bending
moment (N, M), a plastic stress distribution is considered first. Subsequently, the c/t ratios
are examined for each wall separately, to classify them in class 1 or 2. If the walls fail to
be classified in class 1 or 2, an elastic stress distribution is considered to examine if the
walls satisfy the limits for class 3. If they do not, these walls belong to class 4. The entire
cross section is then classified in accordance with the largest class of its walls. Class 4 cross
sections can be classified as class 3 if plate buckling verifications are made by the reduced
stress method in accordance with Section 8.4. Alternatively, the effective width method in
accordance with Section 8.5 is employed. The reduced stress method refers mainly to cross
sections with longitudinally stiffened walls. Cross sections with unstiffened walls are almost
always verified for plate buckling with the effective width method.
Tables 9.2 and 9.3 give the limiting c/t ratios for internal elements supported at two edges
and external elements supported at one edge only. It may be seen that for rolled sections,
the width c is composed of the straight part of the element and for welded elements of the
clear part between weld toes. For external elements, c/t limits are given only for compres-
sion, since they mirror stress conditions due to uniaxial bending of the cross section that is
usually relevant for bridge sections. The distinction between classes 1 and 2 is in practice of
limited importance for bridges, since analysis by plastic hinge theory that applies only for
class 1 and not for class 2 may be employed only for accidental loadings.
Tables 9.2 and 9.3 may be used for the classification of steel girders at construction stages
before concrete casting. After concrete casting, the top flanges are rigidly connected to the
341
342 Design of steelconcrete composite bridges to Eurocodes
MR
Class 1
Mpl
Class 2
Mel
Class 3
Mb
Class 4
Bending Rotation
pl max Class resistance capacity
1 Mpl Yes
2 Mpl Limited
3 Mel No
4 Mb<Mel No
Notes:
Cross sections where plate buckling verifications are required:
Class 4 treated as class 3: Class 4 cross sections where plate buckling verifications are made by
the reduced stress method; see Section 8.4.
Class 4: Plate buckling verifications are made by the effective width method; see Section 8.5.
concrete slab through the shear connectors. In such cases, the steel flange attached to the
slab may be classified as class 1 or 2, although it could be class 3 or 4, provided the spac-
ing of connectors is appropriately selected (see Table 9.3), since concrete prevents its local
buckling. Accordingly, the classification of composite bridge sections at the service stage is
relevant mainly for hogging bending where the bottom flange is in compression and the web
partly in compression. In sagging bending, the cross section is usually class 1 or 2, since
the compression flange is connected to the slab, and therefore, class 1 and the compression
part of the web are usually small due to the position of the neutral axis near the top flange.
It is important to note that the class of a composite cross section depends also on the sequence
of construction and the effects due to creep and shrinkage. Therefore, the classification should
be conducted for short- and long-term design; see Tables 9.12 and 9.13. If, for example, the
web is classified as class 3 for the short-term effects, then it may be changed to class 4 for the
long-term ones. Indeed, the factor in Table 9.2 is determined from the direct stresses that are
time dependent due to creep and shrinkage; see Example 9.4. Moreover, the classification of a
box-girder cross section is also influenced from the shear lag effect on the wide flanges.
Table 9.4 gives the classification limits for the bottom flanges of steel cross sections
encased in concrete (filler-beam decks). Obviously, classification takes place only for the
hogging moment areas. It is noted that for the hogging moments of continuous filler-beam
decks with class 1 cross sections, a redistribution at ultimate limit state (ULS) other than
fatigue up to 15% is allowed.
Cross sections with class 1 or 2 flanges and class 3 web may be classified as class 2 pro-
vided that only an effective part of the web in accordance with Figure 9.1 is considered. The
effective part of the web in compression extends 20t w from the plastic neutral axis of
the effective section and 20t w from the compression flange, the remaining part being not
effective.
144 Local buckling of thin-plate elements
= 1 EC3-1-5 T4.1
k = 23.9 EC3-1-5 T4.1
(1540 2 20 2 6)/10
p = = 1.299 EC3-1-5 4.4(2)
28.4 0.825 23.9
= (1.299 0.055 (3 1))/1.2992 = 0.705 EC3-1-5 4.4(2)
bc = (1540 2 20 2 6)/{1 (1)} = 744.0 mm. EC3-1-5 T4.1
beff = 0.705 744.0 = 524.4 mm. EC3-1-5 T4.1
be1 = 0.4 524.4 = 209.8 mm. EC3-1-5 T4.1
be2 = 0.6 524.4 = 314.6 mm. EC3-1-5 T4.1
Local buckling of thin-plate elements 145
Solution.
Using the lower of the web and plate yield stresses of fyw = 355 N/mm2 and
= 0.814,
301.8 10 355
F = = 1.630(>0.5) EC3-1-5 6.4(1)
403.2 103
F = 0.5/1.630 = 0.307 EC3-1-5 6.4(1)
Leff = 0.307 301.8 = 92.6 mm EC3-1-5 6.4(1)
FRd = 355 92.6 10/1.0 N EC3-1-5 6.2(1)
= 328.6 kN < 1400 kN, and so load bearing stiffeners are required.
Stiffener buckling resistance.
Aeff ,s = 2 (16 100) + (2 15 0.814 10 + 16) 10
= 5801 mm2 EC3-1-5 F9.1
3 6 4
Ieff ,s = (2 100 + 10) 16/12 = 12.35 10 mm EC3-1-5 F9.1
ieff ,s = 12.35 106 /5801 = 46.1 mm
1 = 93.9 0.814 = 76.4 6.3.1.3(1)
Lcr = 1.0 1500 mm = 1500 mm EC3-1-5 9.4(2)
= 1500/(46.1 76.4) = 0.426 6.3.1.3(1)
For buckling curve c, = 0.49 EC3-1-5 9.4(2), T6.1
2
= 0.5 [1 + 0.49 (0.426 0.2) + 0.426 ] = 0.646 6.3.1.2(1)
= 1/[0.646 + (0.6462 0.4262 )] = 0.884 6.3.1.2(1)
Nb,Rd = 0.884 5801 355/1.0 N = 1820 kN > 1400 kN OK.
6.3.1.1(3)
150 Local buckling of thin-plate elements
6.2.6(3)
Vpl,Rd = 2789 (355/ 3)/1.0 N = 571.6 kN 6.2.6(2)
hw /(tw ) = 1.2 (351.4 2 9.7)/(7.0 0.814) = 70.0 < 72
EC3-1-5 5.1(2)
and so the flange resistance is not completely utilised in resisting the bending
moment.
EC3-1-5 5.4(1)
bf = 400 mm < 498 mm EC3-1-5 5.4(1)
2
1.6 400 20 345
c = 1800 0.25 + = 469.9 mm
10 15002 355
EC3-1-5 5.4(1)
400 202 355 4000 2
Vbf ,Rd = 1 N = 11 kN
469.9 1.0 4195
EC3-1-5 5.4(1)
Vbw,Rd + Vbf ,Rd = 1733 + 11 = 1744 kN. EC3-1-5 5.2(1)
Table 2.1: Slenderness limits between Class 3 and Class 4 cross sections
Internal compression parts
Axis
of
bending
Axis
of
bending
Part subject Part subject to Part subject to bending and
Class
to bending compression compression
Stress distr.
in parts
(compressi-
on positive)
b 42H
when \ ! 1: d
t 0.67 0.33\
b / t d 124H b / t d 42H
b 62H 1 \
3
when \ d 1*) : d
t \
235 fy 235 275 355 420 460
H
f y (N/mm 2 ) H 1.00 0.92 0.81 0.75 0.71
_____
26 Outstand flanges
_____
27
Fig. 2.14: Effective cross section
2. OVERVIEW OF DESIGN RULES
Fig. 2.16: Class 4 cross section in pure bending
_____
28 If axial force and bending moment act simultaneously, the calculation
of effectivep widths may be based on the resulting stress distribution. EN
1993-1-5 allows a simplified approach where Aeff is calculated only for
stresses due to pure compression and Weff only for stresses due to pure
bending.
In non-symmetrical cross sections subject to an axial force NEd, a shift
eN occurs (of the centroid G of the effective area Aeff relative to the centre of
gravity of the gross cross section G, see Fig. 2.15). This shift results in an
additional bending moment 'M = eN NEd that should be taken into account in
the cross section verification (see section 2.4.6). According to clause 4.3(3)
of EN 1993-1-5 the shift eM (see Fig. 2.16) of the centre of gravity due to
pure bending can be disregarded in the calculation of 'M, even if the cross
section is subject to the combination of axial force and bending moment.
Generally the calculation of effectivep widths requires an iterative
2.4 PLATE BUCKLING EFFECTS DUE TO DIRECT STRESSES
procedure shown in Fig. 2.17 that ends when the differences between two
steps are sufficiently small.
Fig. 2.17: Determination of effectivep area by iterative procedure
The first iteration starts with the stress distribution on the gross cross
section AG1. The effective area for the second iteration Aeff2 is calculated from
this stress distribution and the effective area for the third iteration Aeff3 from
the stresses on Aeff2.
For I-section and box cross section in bending EN 1993-1-5 allows a
simplified approach that ends in two steps. In the first step effectivep widths
in flanges (if they are in Class 4) are determined from the stress distribution
on the gross cross section. In the second step the stresses are determined on
the cross section composed of the effectivep area of the compressed flanges _____
29
and the gross areas of the web and the tension flanges. The effectivep width
in the web is calculated based on these stresses and this is taken as the final
result.
When different stages of construction have to be considered, which is
a normal case in the design of composite bridges, the following simplified
approach proposed in a Note to clause 4.4 (2) of EN 1993-1-5 may be used:
Hybrid girders where the steel grade of the flanges is higher than the
steel grade of the web may represent an economic solution, because the
stronger (and more expensive) material is put where it contributes the most.
The following three requirements should be fulfilled:
Ih d 2.0 (2.11)
_____
30
(N/mm2)
235 N/mm2
Fig. 2.18: Assumed stress distribution in a hybrid girder with Ih = 460 N/mm2 / 235
N/mm2 = 1.96
Pruhlerrr. Determine the aectiun mnment capacityr [if the welded gtje girder {if
535.? ateel ahewn in Fig. 4.33d.
Sela-triers.