Documente Academic
Documente Profesional
Documente Cultură
B
S 8007 includes recommendations Es modulus of elasticity of reinforce- cracking is being considered
for the calculation of design crack ment εm1 average strain at face 1
widths for sections under flexure fc compressive stress in concrete εm2 average strain at face 2
and for sections under direct tension. It fs stress in reinforcement εsm average strain in tension reinforce-
does not provide recommendations for fs1 stress in reinforcement at face 1 ment
sections under the combined forces. fs2 stress in reinforcement at face 2 εsm1 average strain in tension reinforce-
This Technical Report examines the fstif stiffening tensile stress in concrete ment at face 1
procedures given in the code, and shows fstif1 stiffening tensile stress in concrete εs strain in tension reinforcement
the separate equations for flexure and at face 1 εs1 strain in reinforcement at face 1
direct tension to be based on similar fstif2 stiffening tensile stress in concrete εs2 strain in reinforcement at face 2
premises. The situation where tension at face 2 ∆εs strain reduction in tension rein-
exists across the whole of the section is Fstif total stiffening tensile force in forcement due to tension stiffening
examined, and the limiting values of the concrete of concrete
depth of the neutral axis are calculated. Fstif1 portion of stiffening tensile force ∆εs1 strain reduction in reinforcement
Equations are developed for surface acting at the level of the steel at at face 1 due to tension stiffening of
strains and the stiffening effect of the face 1 concrete
concrete. Fstif2 portion of stiffening tensile force ∆εs2 strain reduction in reinforcement
Similar equations are developed for acting at the level of the steel at at face 2 due to tension stiffening of
the second case where some compres- face 2 concrete
sion is present on one face of the h overall depth of section ρ1 ratio of reinforcement at face 1
section. It is shown how the design
a constant < = F
a2
e= o
A
crack widths for these cases can be k1 s1
h - a1 bh
determined. Equations are developed
with variables allowing for different ρ2 ratio of reinforcement at face 2
a constant < = F
h - a2
reinforcement ratios and concrete cover k2
e= o
a1 A s2
at each face.
bh
K a constant for a particular section
Notation under a certain configuration of Note: Generally subscripts 1 and 2 refer
a’ distance from the compression face moment and direct tension to faces 1 and 2 of the section respec-
to the point at which crack width is R V tively.
being calculated S e+ h -a W 2
modular ratio b = l
d effective depth tion of walls of rectangular or square
Es
J M N αe
Ec
tanks), a need for calculating design
e eccentricity K= O crack widths for this case exists. The
T
L P ε1 strain at level considered ignoring Technical Report proposes a method for
Ec modulus of elasticity of concrete (1/2 the stiffening effect of concrete calculating the design crack widths, w1
the instantaneous value when used ε11 strain at face 1 ignoring stiffening and w2, at the two faces of a section
to determine αe) effect of concrete under these loadings.
BS 8007 approach bt (h - x)
In a section under direct tension,
The procedure for the calculation of Df s = ....(2) the value of the depth of the neutral
3Es As
design crack widths given in Appendix axis is x = –∞ or x = ∞. By substituting
B of BS 8007 can be summarised as The average value of the steel strain, these values in eqn (7) the design
follows: therefore, is: crack width for a section under direct
tension is:
bt (h - x)
• calculate the average strain in the f sm = f s - ....(3)
3Es As
section at the level where cracking is w = 3acr f m ....(8)
being considered allowing for the The average value of the strain at a
stiffening effect of the uncracked distance al from the compression face, Since this is the same as eqn (4) of BS
concrete between cracks where the crack width is to be calcu- 8007: Appendix B, it can, therefore, be
• calculate the design crack width lated, is: assumed that eqn (7) will also apply to
using this value of the strain. the case of combined flexure and direct
(al- x) f sm
fm = tension.
The average steel strain may, as an ad - xk
approximation, be determined by calcu- Combined flexure and direct
lating the steel stress on the basis of a
(al- x) f s b t a h - x k a al- x k tension
cracked section, and reducing this by the = - ...(4) The procedure for calculating the design
tensile force due to tension stiffening in ad - xk 3E s A s a d - x k crack width for sections under combined
the concrete (BS 8110: Part 24 Clause flexure and tension, can be summarised
3.8.3, Assessment of crack widths). =f1 - f 2 as follows:
By considering the cracked concrete bt a h - x k a al- x k
section in flexure as shown in Fig 1 (See where f2 = • determine the position of the neutral
BS 8110: Part 2 Fig 3.1), Eqn (2) of BS 3Es As a d - x k axis in the cracked section, x
8007: Appendix B, which defines the • determine the strain due to tension
stiffening effect of the concrete for a Similarly, by considering the cracked stiffening of concrete between cracks,
design crack width of 0.2mm, can be concrete section in direct tension as ε2
derived as follows5. shown in Fig 2, Eqn (5) of BS 8007: • determine the average strain at the
The value of the tensile stress at the Appendix B for the stiffening effect of level where cracking is considered, εm
tension face between cracks is assumed concrete in direct tension for a design • determine the crack width according
as 2/3MPa. The stiffening force of the crack width of 0.2mm, can be derived as to eqn (7).
concrete in tension, therefore, is: follows:
The stiffening force of the concrete in Two cases can be considered:
bt (h - x)
Fstif = ....(1) tension is:
3
• Case 1: Complete section in tension
2
The effective strain reduction in the Fstif = bt h ....(5) • Case 2: Section partially in compres-
3
steel is: sion.
f s1 = - a e h - xx - a1 f c ....(36)
Fstif2 = Fstif - Fstif1 ....(21) a2 _ f s1 - f s2 i
f 12 = f s2 - ....(29) f s2 = a e x -x a2 f c ....(37)
When the neutral axis position is at x _ h - a1 - a2 i
≥ h (Fig 5(b)): Therefore, according to BS 8007 Determining the stiffening force
Clause B3, the average strains are as As shown in Fig 7, the maximum stiff-
f stif2 = 2 3 MPa for w = 0.2mm, and follows: ening tensile stress in the concrete is:
....(22)
f stif2 =1MPa for w = 0.1mm At face l: εm1 = ε11– ε21 ....(30) f stif1 = 2 MPa for w = 0.2mm, and
3
At face 2: εm2 = ε12– ε22 ....(31)
f stif1 = f stif2 x -
x
h ....(23) f stif1 =1MPa for w = 0.lmm ....(38)
Determining the design surface crack
As previously, the total stiffening width When 0 < x <a2:
force is as given in eqn (18), the The total stiffening force is:
3a cr f m
centroid of the stiffening force as in When x # 0: w = ....(32)
2 _ a cr - c min i
eqn (19), and the values of Fstif1 and Fstif2 1+ Fstif = 1 fstif1 _ h - x i b ....(39)
_h - xi
2
as in equations (20) and (21) respec-
tively. When x ≥ h, this equation has to be The centroid of the stiffening force is
modified by replacing (h – x) in the at:
Determining the average strain denominator by x. Therefore:
The effective reduction in strain in the x l= x + 2h ....(40)
3a cr f m 3
reinforcement due to the stiffening when x $ h: w = ....(33)
2 _ a cr - c min i
effect of the concrete is: 1+ Taking moments about the reinforce-
x
ment in face 2 yields:
Fstif1
Df s1 = , and ....(24) By substituting the values for εm1 or
E s A s1
Fstif1 = Fstif d n
x l- a2
εm2 in these eqns, the values of the ....(41)
Fstif2 h - a1 - a2
Df s2 = ....(25) design crack widths, w1 and w2 at faces
E s A s2
1 and 2 respectively, can be calculated. From horizontal equilibrium:
The effective reduction in strain at It should be noted that when A2 << A1,
the faces of the section is: w2 can actually be larger than w1, even Fstif2 = Fstif - Fstif1 ....(42)
a1 _ Df s2 - Df s1 i if the bending moment is applied in an
f 21 = Df s1 - ....(26) anti-clockwise direction as shown in Fig When a2 ≤ x < h:
_ h - a1 - a2 i 5. The complete stiffening force acts on
a2 _ Df s2 - Df s1 i the reinforcement in face 1 as in the
f 22 = Df s2 + ....(27) Case 2: Section partially in case of flexure only. Therefore,
_ h - a1 - a2 i compression
The strain in the concrete at the Determining the neutral axis position Fstif1 = 1 f stif1 _ h - x i b ....(43)
2
faces, ignoring the stiffening effect, Consider a cracked section with a
is: width, b, which is partially under b _h - xi
a1 _ f s1 - f s2 i compression, as shown in Fig 6. Define =
3
for w = 0.2mm, and
f 11 = f s1 + ....(28) fc and fs2 as negative to indicate
_ h - a1 - a2 i compression. By considering horizontal b _h - xi
= for w = 0.1mm
2
Determining the average strain
When 0<x<a2:
The effective reduction in strain in
the reinforcement due to the stiffening
effect of the concrete is:
Fstif1
Df s1 = ....(44)
E s A s1
Fstif2
Df s2 = ....(45)
E s A s2
Fig 7. Due to the section being partially in
Stiffening effect of compression, cracks can form only in
concrete, section face 1 and the crack width in face 2
partially in need not be calculated. The effective
compression reduction in strain at face 1 is:
where f 21 = Acknowledgement
3E s A s1 _ h - a1 - x i
Conclusions The author thanks Robert Anchor for
for w = 0.2mm, and 1. It has been shown that the expression his assistance in the development of
b _h - xi for design crack widths in flexure as given this method, both in discussions and by
2
_ h - a 1i c e - m + a 2 c e + 2h - a 2 m
In both cases, ie when 0 < x < a2 and a2 t1 h
<x < h the crack width at face 1 can be + a1
t2 2
determined as before from eqn 7. x= ....(9)
ce - 2 m + ce + 2 - a 2m
t1 h h
+ a1
Transition t2
Consider a section with the following
characteristics:
b = 1000mm; h = 350mm; c1=c2 =40mm n1 d 3 - 2n1 + 6e n + 6t2 _ ae - 1i d n1 - a2 n d1 - 2a2 + 2e n + 6t1 ae d1 - a1 - n1 n d1 - 2a1 - 2e n = 0
2
Queries,
Q comments,
correspondence,
and curiosities…
tion to location in a structure, thus based.) sections for columns is inefficient.
Codes of practice showing that the use of partial This covers the general format S.H.S., R.H.S. and C.H.S. sections
safety factors negates the objective of the current codes. However there should be used, with built up or
I have received a lengthy letter of structural analysis which is to are also specific mistakes in the fabricated box section columns for
from Mr S. L. Hammond of calculate accurately the distribu- individual codes. I will try and heavier loads. With these type of
Whitstable, Kent who has tion of forces in a structure due to briefly list some of the worst errors sections the treatment and formu-
given much thought to the use a given set of loads. that I have come across in each of lae given in Chapter 1 of
of partial safety factors in Investigation of the statistical the codes: Timoshenko and Gere, ‘Beam
Codes of Practice and the basis of the ‘partial factors’ also columns’ can be applied.
design principles on which the reveals faulty mathematics. We are BS 8110: To achieve the stress As far as existing structures go,
codes are based. For the sake told that the partial factors are set distribution assumed in the section the most appropriate form of reme-
of comprehension Mr according to the degree of risk of design for beams, a considerable dial work to the columns is to weld
Hammond’s letter is published the particular type of load exceed- amount of deflection must occur, plates across the toes of the flanges,
in full. He writes: ing the factored load (whether any sufficient to lead to a rearrange- thus transforming the open section
proper assessment of real loads has ment of forces in the structure, and into a closed section and avoiding
When I entered the field of struc- ever been carried out we are not possible collapse. Certainly the the problem of torsional instability.
tural engineering the new design told). Thus for the dead load we rotations required to develop the
codes BS 8110 and BS 5950 were assume that, say, 99% of the dead design stresses in the beam negate BS 5400: Generally the same
coming into use, since then I have loads will be less than or equal to any question of elastic analysis, comments as BS 5950 apply, with
used these codes along with the 1.4 times the calculated dead load, although the code assumes an the added observation that the
masonry code which is in a similar and for the live load, 99% of the elastic distribution of forces in ratios between dead and live loads
format. As an engineer I have tried actual live loads will be less than or subframe analysis. vary more in a bridge structure
to make sense of these publications equal to 1.6 times the calculated than building structures, and
and understand the theoretical live load. Thus the actual probabil- Column design: as far as I have therefore the actual factors of
basis of the design rules. ity that the calculated factored load been able to find out (Kong and safety vary more greatly across the
Unfortunately the more closely will be exceeded depends on a Evans) the design of column structure.
these documents are examined, the combination of probabilities that sections also assumes yield in the
less sense they make from the varies with the proportion of dead steel and failure in the concrete. BS 5628: Apart from the format of
point of view of basic physical laws to live load (as the factors of safety This involves the structure in partial safety factors the method
and mathematics. are not equal). In fact the highest further contortions in order that it given for the design of elements
All these codes, and the new probability of failure occurs when may achieve the assumptions on under vertical load is wrong. The
‘Eurocodes’, make use of so called the load is all of one type (dead or which the section design is based. capacity of masonry walls and
‘partial safety factors’ in order to live load), and the minimum prob- No account is taken of elastic insta- columns is dependent on the elastic
calculate the design loads acting on ability of failure occurs when the bility or deflection of the column. properties of the masonry as well
the structure, or a particular two loads are approximately equal. The same remarks apply to the as its crushing strength. No
section of a member, the idea being –There is an appreciable difference design of concrete walls. account is taken of this.
that the various ‘partial factors’ in the probabilities of the order of a
represent different risks of the factor of 100, thus showing that the BS 5950: Allowing for the incorrect Mr Hammond says in his
calculated load being exceeded in stated objective of providing a format of safety factors the beam covering note that his
the actual structure. consistent probability of the struc- design appears to be adequate and conclusions are based on
Take the following example of a ture not failing is not met. The cover more design situations than sound mathematical
simply supported beam loaded introduction of further partial previous codes. However, the reasoning and that not only is
with a variety of non-uniform factors such as for wind load only column design is based upon he alarmed by the inadequacy
imposed and dead loads, these serves to complicate the statistics inelastic buckling, which for open of the current codes but by the
different types of load having even more. sections such as channels and H fact that they do not offer an
different partial factors applied to (As an elementary example of section columns is inadequate. adequate means of assessing
them. Then calculate the design the probability of combined events Examination of Timoshenko & existing structures for
(factored) moments and shears for consider the probability of throw- Gere’s Theory of elastic stability strength. Is there support for
the beam, and the actual (non ing two sixes in succession with a shows that the stability of open Mr Hammond’s conclusions
factored) moments and shears in dice, the probability is one in section columns is treated in among the membership? If so
the beam. Unless both types of load thirty-six as opposed to one in six Chapter 5 ‘Torsional buckling’ and please write in.
are applied in an identical way, the for a single six. Loads are continu- that for the commonly used H-
ratio (factor of safety) between ous variables and the probabilities section, failure occurs in a
actual and design forces will vary that the load will exceed the total combined torsional and lateral Promoting the profession
from location to location in the factored load varies with the ratio buckling mode. (This is not to be
beam. In the same way a calculated of the loads, but unless there is confused with the torsional buck- Andrew Sandford has sent in
example will show that the ratio only one type of load, the probabil- ling of beams). No account is taken this letter after reading the
between the calculated design ity of exceedance bears no rela- of this in allowable stresses for interview with Professor
loads and the actual loads acting tionship to the probabilities on columns either in BS 449 or BS David Blockley which
on a structure will vary from loca- which the partial factors are 5950. In any case the use of open appeared in NCE 8-15 August.