Serviceability (+ fatigue)

Paul Jackson
Serviceability Issues
• Stress Limitation

• Crack Control

• Deflection Control

• Other issues, e.g. vibration

Serviceability Issues: why check

• Deflection Control
Appearance, drainage, clearance, train
behaviour
Serviceability Issues: why check

• Crack Control
Appearance,
waterproofing,
reinforcement corrosion?
Serviceability Issues: why check
• Stress limitations
[Link]
“To avoid longitudinal cracks, micro-cracks or
high levels of creep”
[Link]
“to avoid inelastic strain, unacceptable cracking
or deformation”
Stress limitations
• Not really a fundamental limit state
• Theoretically possible to design a satisfactory structure
that does not comply
but:

1. Other checks (e.g. crack width) assume linear

behaviour
2. Behaviour only strictly predictable from full stress
history
All analysis at SLS is elastic

Applies to Structure and section analysis

Section analysis can be cracked
(including in prestressed: sometimes)
Serviceability Issues: why check
• Vibration
Resonance with pedestrians
London Millennium Bridge!
High speed rail
Not normally an issue for highway (loading
covers dynamics) and not fully covered in code
Serviceability Issues: effect on design
• Deflection Control
EN 1992-1 gives calculation methods but rarely
critical in bridges
May still need to calculate (for setting out
falsework etc)
Beware! code allows you to check under perm.
load considering tension stiffening. Not right
after live applied!
Serviceability Issues: effect on design
Stress Limits
• Reinforcing Steel: 0,8fyk (0.75 in BS 5400)
1,0 for imposed deformations
• Prestressing steel 0,75fpk
• Concrete : 0,6 fck (0.48 fcu cf BS 5400
0,5 for RC, 0,4 for PSC)
But Calculated on cracked section (sometimes)
Serviceability Issues: effect on design

• Crack Control

Minimum steel area

Ensures reinforcement does not yield on first crack

Crack control rules

Restrict maximum width
Minimum Steel Area
Asminσs = kc k fct,eff Act (7.1)
σs = yield stress
kc = Coefficient (1,0 for pure tension, 0,4 for pure
bending)
k = Coefficient (1,0 for h<300mm)
fct,eff = mean value of tensile strength
(can use early age, e.g. for early thermal
cracking)
Act = Area of concrete within tensile zone
(immediately before crack, i.e. when stress
= fct,eff)

Note: for flanged sections, webs and flanges

considered separately
Table 3.1 Concrete Properties
(abbreviated)
fck (MPa) 25 30 40 50 60 70 90
fck,cube (MPa) 30 37 50 60 75 85 105

fcm (MPa) 33 38 48 58 68 78 98
fctm (MPa) 2,6 2,9 3,5 4,1 4,4 4,6 5,0
fctk,0.05 (MPa) 1,8 2,0 2,5 2,9 3,1 3,2 3,5
Ecm (GPa) 31 33 35 37 39 41 44

εcu2 (0/00) 3,5 2,9 2,7 2,6

εc2 (0/00) 2,0 2,3 2,4 2,6

Minimum Steel Area: simple rectangle in flexure
Asminσs = kc k fct,eff Act (7.1)

Asmin = [Link]/500

= [Link]

Also 9.1
Asmin = 0,26fctmbtd/fyk not less than 0,0013btd

For slab, 9.1 governs!

Minimum steel area

0.25
9.1, h = d
Reinforcement (%bh)

0.2 BS 5400,
h=d
9.1, 160 slab
0.15

0.1 BS5400, 160 slab 7.1

0.05
25 45 65 85 105
Concrete grade (cube)
Table NA.1 /(7.101N)
Recommended values of wmax (mm) and relevant
combination rules

Reinforced members and Prestressed

Exposure Class prestressed members members with
without bonded tendons bonded tendons

Quasi-permanent load Frequent load

combinationc combinationc

X0, XC1 0,3a 0,2

XC2, XC3, XC4 0,2b

XD1, XD2, XD3 0,3 0,2d and

XS1, XS2, XS3 Decompression
Cracking criteria
• Criteria more onerous for prestressed
• Does not actually say you can treat an element (e.g.
deck slab) as prestressed in one direction and RC in
another
• Neither does BS 5400!
• Can still do it
• Is not actually very logical (Said to be for durability but cracks
parallel to tendons more significant)
But: not clear using more severe criteria for
prestressed is logical!
Notes From 7.101 (UK NA version)
a For X0, XC1 exposure classes, crack width has no influence on
durability and this limit is set to guarantee acceptable appearance.
In the absence of appearance conditions this limit may be relaxed.
b For these exposure classes, in addition, decompression
should be checked under the quasi-permanent combination of
loads.
c For the crack width checks under combinations which include
temperature distribution, the resulting member forces should
be calculated using gross section concrete properties and
self-equilibrating stresses may be ignored.
d 0,2 applies to the parts of the member that do not have to be
checked for decompression
Plus

• The decompression limit requires that all

concrete within a certain distance of bonded
tendons or their ducts should remain in
compression under the specified loading. The
distance within which all concrete should remain
in compression shall be taken as the value of
cmin,dur (NA) determined for the relevant surface.
“Decompression” vs BS 5400 Class 1

ε ε

Tendons

Cracked

OK to either OK for decompression,

not class 1
Decompression in thin slabs (UK N.A.)
Nominal cover to tendon
12MPa
> [Link] + Dc + bar diameter
typically 22-35mm more than
[Link]

For example shown, moment to

decompression ≈ 1.3 x moment to
class 1

26mm
-1.6MPa

220mm deck slab, cover = [Link] + 10 +16

Table 7.101N

Reinforced members and Prestressed

Exposure Class prestressed members members with
without bonded tendons bonded tendons

Quasi-permanent load Frequent load

combinationc combinationc

X0, XC1 0,3a 0,2

0,2
XC2, XC3, XC4 (+decompression
0,3 under quasi perm.)
XD1, XD2, XD3 0,2d and
XS1, XS2, XS3 Decompression
Means cracked section analysis needed for prestressed!
But rarely critical for XD case (except reversed moments!)
Main exposure classes
for bridges
XC3/XC4 No Salt Inside voids
Under waterproofing
Bridges not subject to salt
XD1 Moderate Walls etc more than 10m horizontally
from carriageway
Humidity Soffits more than 5m above
carriageway
XD2 Wet, Immersed + buried more than 1m below
carriageway
rarely dry
XD3 Cyclic The rest
wet & dry
Unclear Case
(most tensile face not worst exposure)
(waterproofing)

XS

Resolved by NA
Prestress cracking
BS 5400 EN 1992
Class 1 Decompression
Class 2 No direct equivalent (quite)
Class 3 Crack width check
BS 5400/BD 24 normally requires checks for:
Class 1 HA comb1 + Class 2 Full comb 3

EN 1992-2 normally requires decompression only with no

SV/SOV

There are 2 checks for cases without XD exposure

Cracked SLS check on prestressed

• Needed to make use of savings for XC exposure + cases

where critical tension on opposite face from tendon
• Requires iterative solution as rc column (or external
prestress to BD58)
• Means tensile stress in tendon can govern (which means
reducing jacking load)
• Means you can reduce prestress requirement by adding
secondary steel
• Do not have to do it if max tension < [Link]
• [Link] = fctm (3,5 for 40-50 concrete)
Prestress cracking
BS 5400 EN 1992
Class 1 Decompression
Class 2 No direct equivalent (quite)
But can design as in tension
but uncracked using Clause
7.1(2)
Class 3 Crack width check
BS 5400/BD 24 normally requires checks for:
Class 1 HA comb1 + Class 2 Full comb 3

EN 1992-2 normally requires decompression only with no

SV/SOV

There are 2 checks for XC exposure (0,2mm under

frequent, decompression under quasi-permanent)
For XD/XS exposure

d. 0,2 applies to the parts of the member that do not have

to be checked for decompression

This is apparently only an NA requirement

But: for XC exposure 0,2 applies and having XD exposure in EN
206 does not mean you do not have XC (i.e. chlorides do not
stop carbonation)
Hence NA is only clarifying in
Prestress Cracking
• Tensile stress has to be less than [Link] under
characteristic load to avoid requirement for minimum
steel 7.3.2 (4)
• This brings you back to something similar to BS 5400:
check decompression under normal load and [Link] under
full load

But:
Can include (adjusted) area of tendons within 150mm
Hence normally OK for pretensioned and…
Secondary steel in post-tensioned
(Solid slab)

12MPa (say)

Asminσs = kc k fct,eff Act

Kc = 0.4[1-σc/(1,5X1X3.5)]
-ve for σc >[Link]

No designed secondary steel needed

3,5 MPa
(in theory mult> m crack for σc > [Link]/3)
Secondary steel in post-tensioned
(300mm flange in 1500mm deep box)

12MPa (say)

Asminσs = kc k fct,eff Act

Asmin (in bottom flange)

= 0,5x0,65x3,5x1000X300/500
= 682mm2/m
= B12-150
3,5 MPa(0.4 at top of
Not very onerous
flange)
Determination of
Crack Widths
1. Calculate
or
2. Comply with max bar spacing table
or
3. Comply with max bar diameter table

Same basic approach used for prestressed when you do

not have to comply with decompression
Can’t use 2 for restraint induced cracking
Table 7.3 Maximum bar spacing for crack control

Steel stress * Maximum bar spacing [mm]

[MPa] Wk = 0,3 mm Wk = 0,2 mm
160 300 200
200 250 150
240 200 100
280 150 50
320 100 -
360 50 -
Notes:
1. The values in the table are based on the following assumptions:
c = 25mm; fct,eff = 2,9MPa; hcr = 0,5; (h-d) = 0,1h; k1 – 0,8; K2 =
0,5; Kc =0,4; K = 1,0; Kt = 0,4 and K’ = 1,0
2. Under the relevant combinations of actions
Table 7.2 Maximum bar diameters φ*s for crack control

Steel stress* Maximum bar size [mm]

[MPa] Wk = 0,3 mm Wk = 0,2 mm
160 32 25
200 25 16
240 16 12
280 12 8
320 10 6
360 8 5
400 6 4
450 5 -

Notes:
1. The values in the table are based on the following assumptions:
c = 25mm; fct,eff = 2,9MPa; hcr = 0,5; (h-d) = 0,1h; k1 – 0,8; K2 = 0,5;
Kc =0,4; K = 1,0; Kt = 0,4 and K’ = 1,0
2. Under the relevant combinations of actions
The maximum bar diameter should be
modified as follows:
Bending (at least part of section in compression)
∗
φs = φ (fct,eff /2,9) kchcr/(2(h-d))
s

Tension (all of section under tensile stress)

∗
φs = φ s(fct,eff/2,9)hcr/(8(h-d))
where:
φs is the adjusted maximum bar diameter
φ∗s is the maximum bar size given in the Table 7.2
hcr is the depth of the tensile zone immediately prior to cracking, considering the
characteristic values of prestress and axial forces under the quasi-permanent
combination of actions
kc coefficient for stress distribution, 0.4 for rectangle with no axial force
Cracking in RC
• Only checked under quasi-permanent
• Unlikely to be critical?
• Result: despite apparently radical treating of RC
and prestressed together, still tends to give:

RC: designed at ULS

Prestressed: designed at SLS
Table 7.3 Maximum bar spacing for crack control

Steel stress * Maximum bar spacing [mm]

[MPa] Wk = 0,3 mm Wk = 0,2 mm
160 300 200
200 250 150
240 200 100
280 150 50
320 100 -
360 50 -
Service stress typically ≈ (500/1.15)(1/1.35) ≈ 322MPa
Need permanent load ratio > c 50% for cracking to be critical
Can be critical in larger slab decks, cross heads & aquaducts
Will not be critical in thin deck slabs
Do not forget side faces!
Crack Width Calculation
wk = sr,max (εsm - εcm) (7.8)
where
sr,max is the maximum crack spacing
εsm is the mean strain in the reinforcement under the relevant
combination of loads, including the effect of imposed
deformations and taking into account the effects of
tension stiffening. Only the additional tensile strain beyond
zero strain in the concrete is considered
εcm is the mean strain in the concrete between cracks
 Crack Spacing
Bar spacing ≤ 5(c+φ/2)

sr,max = 3,4c + 0,425k1k2φ /ρp,eff (7.11)

Bar spacing > 5(c+φ/2)

sr,max = 1.3 (h - x) (7.14)

For 25mm bar and 40 mm cover 5(c+φ/2) = 262mm

As + ξ12 Ap
=
ρp,eff Ac,eff
 Crack Spacing

h-x A - Neutral axis

φ

c B B - Concrete tension surface

C - Crack spacing predicted by

w Expression (7.14)

C D - Crack spacing predicted by

D Expression (7.11)

5(c + φ /2)
Ac effective

x
ε2 = 0
h d
A
hc,ef A - level of steel centroid
ε1
B B - effective tension area
Strain, allowing for tension stiffening

fct,eff
σ s − kt
ρp,eff
(1 + α e ρp,eff )
σs
ε sm − εcm = ≥ 0,6
Es Es
Crack widths: calibration

• Despite very different formulae, tends to gives

similar widths to BSs

• Simple rules often more optimistic than

calculations

• Tends to favour using bar spacing rules

Early thermal effects

• EN 1992 gives criteria and crack width approach

• Does not give strains

• PD says EN does not fully cover and refers to

CIRIA 660
Causes of real problems
with cracking
1. Inadequate min steel
2. Early thermal effects (often combined with 1)
3. Reinforcement corrosion (esp. if due to inadequate
cover)
4. Reinforcement yielded
5. Very wide spaced reinforcement (mainly on side faces)
Pretensioned Beams

Section
Pretensioned Beam: Transfer

• Tension (critical for top at ends)

No specific rule:
Decompression checked if tendons close,
(assuming chloride) otherwise crack width?
Gives a paradox: top strand provided to control
tension but checks not needed if no top strand.
Precast Manual proposes using a tensile stress
Pretensioned Beam: Transfer

• Compression (critical for soffit)

BS 5400 0.5fci not greater than 0.4fcu
EN1992 0,6 fck(t) 0,7 fck(t) (subject to NDP)
for pretensioned elements “if it can be justified
by tests or experience that longitudinal cracking
is prevented.”
Transfer Stress: Comparison
• 0,7 fck(t) = 0.56fci ??
• But fci = “cube strength at transfer”
• fck(t) = “Characteristic…”
• Gives c16.8 cf 20 for BS 5400 if fck(t) from fcm(t) and
Table 3.1
• But: with good concrete quality control, and
records to prove it, fck(t) would be greater and
result similar to BS 5400
Table 3.1 Concrete Properties
(abbreviated)

fck (MPa) 25 30 40 50 60 70 90
fck,cube (MPa) 30 37 50 60 75 85 105
fcm (MPa) 33 38 48 58 68 78 98
fctm (MPa) 2,6 2,9 3,5 4,1 4,4 4,6 5,0
fctk,0.05 (MPa) 1,8 2,0 2,5 2,9 3,1 3,2 3,5
Ecm (GPa) 31 33 35 37 39 41 44
εcu2 (0/00) 3,5 2,9 2,7 2,6
εc2 (0/00) 2,0 2,3 2,4 2,6
Jacking force in tendon

• Limited to 0,8 fpk and 0,9 fp0,1k

• May jack to 0,95fp0,1k if controlled to 5%

• Typically fractionally less than BS 5400

• Limited to 0,75 fpk and 0,85 fp0,1k after lock off

Strand Pattern
• Came out identical to BS 5400 design
• If you had no XD/XS (Chloride) Exposure
could save c 25% prestress
• Similar conclusions for rail bridge
With reduced prestress (no chloride)

• ULS might govern

• Increase in tendon force under live loading is
much greater so fatigue or limit on tendon
service stress could govern
In our examples
• ULS did not govern
• Upper limit on in service tendon stress did affect
rail example (i.e. jacking stress had to be
reduced)
• Fatigue limit check did not govern
• Since less prestress is needed and transfer is
the critical condition for concrete compression,
you could reduce section.
For Rail Loading

Design code Initial Number Tendon

and exposure prestress of stress during
class force Strands tensioning
(kN) (N/mm2)
BS 5400 4599 21 1460
EN 1992, XD 4637 22 1405
exposure
EN 1992, XC 2976 16 1240
exposure
Fatigue
• EN 1992-1 says do not check if SLS stress range less
than fixed value
• Fixed value is based on non-propagating stress range so
safe for any loading (e.g. inc. machinery bases)
• PD gives less conservative stress values specific to
highway traffic load: these vary with span
• Our (one!) comparison suggests full fatigue check allows
much higher stress range
• EN 1992-2 identifies some cases where checks not
needed
• NA adds most deck slabs to this
Fatigue: NA stress ranges
Table 1A – Limiting stress ranges – Longitudinal bending for unwelded
reinforcing bars in road bridges (From PD)
Span Adjacent spans loaded Alternate spans loaded
Bars 16 mm Bars Bars ≤16 Bars >16 mm
>16 mm mm
m MPa MPa MPa MPa
≤3.5 150 115 210 160
5 125 95 175 135
10 110 85 175 135
20 110 85 140 110
20-50 90 70 110 85
100 115 90 135 105
≥200 190 145 200 155

Table 2B – Limiting stress ranges– Transverse bending for

unwelded reinforcing bars in road bridges
Span Bars ≤16 mm Bars >16 mm
m MPa MPa
≤3.5 210 160
5 120 90
≥10 70 55
NA gives exemption for deck slabs

Actually work by “compressive membrane

action”
Stress range typically10% calculated
Significant as these are elements with highest
LL ratio
Fatigue likely to be critical for

Non exempt elements with high LL ratio. e.g.

Rail structures: no deemed to satisfy stress range

(could be but work not done!)

Deck slabs which do not comply with rules for exemption

Continuity steel in precast beams made continuous

(steel does not take dead load, so LL ratio higher)

Can be a problem with links (check requires different strut angle)

Large post-tensioned box example

Spans
70-100-70

Support section
Large post-tensioned box example

Mid-span section
Large Box Girder
Prestress Requirement

Prestress Requirement
Code (MN)
Mid-Span Pier
BS 5400 (HA Implementation) 57.4 57.8

EN 1992-2 (XD or XS Exposure) 55.4 60.3

EN 1992-2 (XC Exposure) 36.8* 50.3*

* Varies with secondary steel

External Prestress

• Like BD 58, considers cracking as RC

• Unlike BD 58, does not require
decompression/class 1 under permanent load
• Gives complete freedom for partial prestress!
• Less links than BD 58 (both treat as RC but EN
has more enhancement for compression)
Conclusions
• Design generally comes out similar: some
exceptions
• Savings for prestressed without chloride
• More freedom for unbonded prestress
• Not as radical as it appears
• Crack widths less likely to be critical