Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

Calculation of crack width and crack spacing

Ingemar Löfgren
Thomas Concrete Group
E-mail: [email protected]

ABSTRACT
The present paper discusses crack propagation and special attention is given
to how the combined effect of reinforcement and fibre bridging influences
the crack spacing and width in the serviceability limit state. Two analytical
approaches, for calculating the crack spacing and crack width, are
presented. The first model is a modification of the conventional crack
spacing model presented in Eurocode 2 and is valid for the case when
cracking is caused by an external load. The second model, which is based
on a bond-slip relationship and a compatibility requirement, is valid for
cracking caused by restraint stresses. Moreover, in the paper some
examples are provided of how the models can be used.
Key words: Fibre-reinforced concrete, Cracking, Restraint, Serviceability, Shrinkage.

1. INTRODUCTION

Concrete has a low tensile strength and tensile strain capacity and cracking is initiated at a
tensile strain of about 0.1 mm/m which can be compared to the drying shrinkage of concrete of
about 0.6 to 0.8 mm/m. Hence, cracks are almost unavoidable and reinforcement is needed to
control the behaviour after cracking and to limit crack widths. Large crack widths are not
aesthetic but may also lead to accelerated reinforcement corrosion in severe environments,
leakage in water-retaining/resisting structures, insanitary conditions, or obstructions and
interruptions in production processes. Cracking may be caused by external applied forces,
imposed deformations, by shrinkage or thermal strains which are externally and/or internally
restrained, or by a combination of these. When cracking is caused by an external applied force
the crack width, if sufficient amount of reinforcement is added, will depend on the applied force.
However, if cracking is caused by an imposed deformation the force in the member depends on
the actual stiffness and the crack width on the number of cracked formed. However, most codes
do not distinguish between these two cases. Furthermore, for structures having both fibre- and
bar reinforcement there exist almost no guidelines exists for structural engineers.

2. THE CRACKING PROCESS

The cracking process differs depending on whether it is caused by an external load, imposed
deformation or restrained shrinkage, see Figure 1. When cracking is caused by an external load
the reinforcement is usually designed such that it is able to transfer the load after cracking
without yielding. For this case the load will cause an immediate cracking process where several
cracks are formed and which are relatively uniformly distributed. For this type of situation the
standard method in Eurocode 2 can be used to determine the minimum reinforcement and for
estimating the crack spacing and crack width. For a member with combined reinforcement

1(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

(fibre- and bar reinforcement) this approach has to be modified. When the cracking is caused by
an imposed deformation a different behaviour can be observed. When a crack is formed this is
accompanied by a sudden drop in the force N and the stiffness of the element also decreases. For
a new crack to be formed the deformation has to be increased so that the force N again reach the
critical value (N > Ncr). However, the force depends on the stiffness of the member and if this is
low a large deformation may be required before a new crack can be formed, compare (b-1) and
(b-2) in Figure 1, and this results in fewer but larger cracks. For this type of cracking process the
standard approach for determining crack spacing and crack width cannot be used.

Crack u
N N

N N N
Force Imposed deformation Imposed deformation

Tension
Ncr stiffening Ncr Ncr

Stadium II
(neglecting tension Large reinforcement Small reinforcement
stiffening) ratio ratio
u u u
(a) (b-1) (b-2)

Figure 1. A reinforced concrete member subjected to: (a) axial force; (b) imposed
deformation, (b-1) with a large reinforcement ratio and (b-2) with a small
reinforcement ratio. Based on Ghali et al [1].

Compared to plain concrete (i.e. without fibres) fibre-reinforced concrete exhibits the ability to
transfer tensile stresses also after cracking, see Figure 2. This material property is referred to as
the residual tensile strength or, for describing the whole curve, the stress-crack opening
relationship (-w relationship). The residual tensile strength increases with increased fibre
dosage but is also influenced by the type of fibre (e.g. slenderness, geometry, material, etc.)


l fct
Fibre
contribution
Residual tensile
w FRC stress
l
w

Concrete
l

w  0.05 mm wc  0.3 mm wc = lf / 2

Figure 2. Schematic description of the fracture behaviour of fibre-reinforced concrete (FRC).

2(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

3. FORCE INDUCED CRACKING

The crack spacing in reinforced concrete structures (without fibres) can be calculated using the
following expression presented in Eurocode 2:

s r . max  k 3  c  k1  k 2  k 4  [mm] (1)
 s ,eff
where:
c is the concrete cover
 is the bar diameter
s,ff is the effective reinforcement ratio,  s ,eff  As Ac,eff and Ac,eff is the effective area of
concrete in tension surrounding the reinforcement
k1 = 0.8 for high bond bars and 1.6 for bars with an effectively plain surface
k2 = 0.5 for bending, 1.0 for pure tension or  1   2  2   1  for eccentric tension
k3 = 3.4
k4 = 0.425

For a section with combined reinforcement a similar expression, which takes into account the
contribution from the fibre reinforcement, can be derived. Consider a reinforced tension rod
loaded with the crack load, Ncr, according to Figure 3. The rod is reinforced with a centrally
placed reinforcement bar, with an area of As, and fibres. The force equilibrium in the region
between two cracks with the maximum crack distance sr,max = 2lt,max is analysed, see Figure 3.

Ncr Ncr

Crack Crack
sr,max New crack
lt,max lt,max Ac
As


fft.res bm ct  fct
Stress acting on the concrete

lt,max
0.5 sr,max
Stress introduced to concrete
through bond, c (x)

Residual tensile strength, fft.res(w)

ct  fctm
Total concrete tensile stress, ct (x,w)
Possible location of new crack

Figure 3. Equilibrium of forces for a tension rod.

At the crack the fibre reinforced concrete transfers a stress fft.res. At the midpoint between the
two cracks the concrete is about to crack and the stress is thus ct  fctm. The increase of stress is
a result of stresses being transferred from the reinforcement to the concrete through bond. The

3(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

bond stress b varies along the transmission length and has an average value of bm which can be
calculated as:
lt , max

 bm 

0
 b ( x)dx
(2)
lt,max

If the tension rod is cut in the middle between the two cracks and along the interface between
the reinforcement and concrete the following equilibrium condition can be formulated:
 bm     (0.5  sr,max )  f ft .res  Ac  f ctm  Ac (3)

The concrete gross cross-sectional area can be formulated as:

A A
Ac  As  c  s (4)
As  s
with s = reinforcement ratio

Inserted in (3) gives

 2
 bm     (0.5  s r,max )   f ctm  f ft.res  (6)
4 s
1  f ctm  f ft .res  
 s r,max    (7)
2  bm s

The minimum crack spacing is equal to half the maximum crack spacing. Accordingly, the
minimum crack spacing can be calculated as:
1  f ctm  f ft .res  
s r,min    (8)
4  bm s
The average crack spacing during the crack formation can be estimated as the average value of
(7) and (8) which gives (in Eurocode 2 it is assumed that sr,max = 1.7×sr,m):
3  f ctm  f ft .res  
s rm    (9)
8  bm s
The stress transfer from the reinforcement to the surrounding concrete depends partly on the
surface properties of the reinforcement and partly on the properties of the concrete. Based on
experimental results, it has been found that the average bond stress can be calculated as:
3
 bm   f ctm (10)
2  k1
If the expression for the average bond stress is introduced into (9), the following expression is
obtained for the crack spacing of a tension rod:
 f ctm  f ft.res  
s rm  0.25  k1  [mm] (11)
f ct s
 f ft .res  
s rm  0.25  k1 1    [mm] (12)
 f ctm  s

The conclusion is that for calculating the crack spacing the basic formula as suggested in
Eurocode 2 can be used but it has to be modified with the relationship between the residual
tensile strength and the tensile strength with the introduced variable as follows:

4(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

 f ft .res 
k 5  1   (13)
 f ctm 

If the effect of concrete cover, the spacing of the reinforcement, and type of loading (tension or
flexural) the following expression can be used to calculate the crack spacing:

s r ,max  k 3  c  k1  k 2  k 4  k 5  [mm] (14)
 s ,eff
1   
 [mm]
s r ,average   k 3  c  k1  k 2  k 4  k 5  (15)
1.7   s ,eff 


3.1 Example
In order to investigate the proposed crack spacing formula full-scale beams were casted and
tested. The experimental program consisted of five series (three beams in each series) with
different fibre dosage and type and amount of reinforcement, sees Table 1. The full-scale beams
were simply supported with 1800 mm span and subjected to a four-point load, see Figure 4. The
full details of the experiments can be found in Gustafsson and Karlsson [2].

Table 1. Test series without and with fibre reinforcement (type Dramix® RC-65/35 from
Bekaert) and amount of conventional reinforcement.
Fibre dosage Reinforcement Number of beams
3
Series [vol-%] and [kg/m ] Number and diameter [mm]
1 Vf = 0 % (0 kg/m ) 3
3 8 3
2 Vf = 0.5 % (39.3 kg/m ) 3
3 8 3
3 Vf = 0.25 % (19.6 kg/ m3) 3 6 3
4 Vf = 0.5 % (39.3 kg/ m3) 3 6 3
5 Vf = 0.75 % (58.9 kg/ m ) 3
3 6 3

600 600 600

h=225

d=200

LVDT Q Q A

C
L b=150
Roller A Roller
1800 Reinf.
2000
ELEVATION A-A

Figure 4. Test set-up (full-scale beams).

In addition to the full-scale beams wedge-splitting test (WST) were conducted, see NT-BUILD
511 [3], and in order to determine the residual tensile strength inverse analyses were carried out,
see Löfgren [4]. In Figure 5(a) the WST-method is outlined and in Figure 5(b) the stress-crack
opening relationships can be seen.

5(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

3.0

2.5 Serie 5 Serie 4

Tensile stress [MPa]

2.0

1.5

1.0
Serie 2
0.5 Serie 1 Serie 3

0.0
0.0 0.2 0.4 0.6 0.8 1.0
Crack opening [mm]

(a) (b)
Figure 5. (a) Schematic description of the WST-method. (b) Obtained stress-crack opening
relationships.

In Figure 6 the calculated crack spacing is compared with the crack spacing obtained in the
experiments. In addition, a comparison is also made with the proposal according to RILEM TC
162-TDF [5], where the crack spacing is calculated as:
   50 
s rm   50  0.25  k1  k 2  b  (16)
  r  l f  f 

As can be seen in Figure 6, the RILEM proposal does not consider the effect of increased fibre
content but whereas the proposal according to equation 15 takes into account the residual tensile
strength of the fibre-reinforced concrete and thus are able to predict that the crack spacing
decreases with increased fibre content, or with increased fibre slenderness as this also increases
the residual tensile strength.

100
Average crack spacing [mm]

90 84 83
81 82
78 77 80
80
69 71
70 66 66
59 60
60 55 54

50
40
38 38 36 36 36
V f = 0% V f = 0.5% V f = 0.25% V f = 0.5% V f = 0.75%
Experiment Model RILEM TC 162-TDF

Figure 6. Comparison between calculated crack spacing and the crack spacing obtained in
the experiments.

6(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

4. RESTRAINT INDUCED CRACKING

Engström [6] has proposed a model for analysing restraint induced cracking and the cracking
process is analysed by modelling the cracks as non-linear springs, see Figure 7. Löfgren [7]
extended the model to include the effect of fibre reinforcement.

Crack, modelled as
non-linear springs

w(s)
l

Forces acting on un-cracked parts N(s) N(s)

(with only bar reinforcement) N(s) N(s)

Forces acting on un-cracked parts for N(fft.res) N(fft.res)

combined reinforcement (fibre and bar
N(s) N(s)
reinforcement), with fft.res as FRCs
residual tensile strength N(s) N(s)

Figure 7. Model for analysing restraint induced cracking.

Engström’s model is based on a bond-slip relationship which has been used to derive an
analytical expression describing the crack width as a function of the reinforcement stress:
0.826
 
 
   s 2  
w s   0.42     s  4 (with  in mm) (17)
  E A   E
 0.22  f cm  E s  1   s 
s s
 
  Ec Aef  
Where  is the bar diameter, s is the stress in the reinforcement, fcm is the average compressive
concrete strength, Es and Ec is the modulus of elasticity of the reinforcement respectively the
concrete, and Aef is the effective concrete area. The effective concrete area can be calculated as
Aef  b  hef , where hef is the part of the tensile zone which has the same centre of gravity as the
reinforcement. The last additional term in (eqv. 17) considers the influence of a zone nearby the
crack where bond is assumed to be fully broken due to radial cracks towards the free surface.

The response during the cracking process can described with the following deformation criteria:
N  s , f ft ,res   l
 1   ef   n  w s   R   cs  l (18)
Ec  AI

where N(s, fft.res) is the force acting on un-cracked parts, l is the length of the member,
AI  Ac  As Es Ec  1 , ef is the effective creep coefficient, n is the number of cracks and R
is the degree of restraint (R=0 for no restraint and R=1 for full restraint). N(s, fft.res) can be
calculated as:
N  s , f ft ,res    s  As  f ft ,res  Aef  As  (19)

7(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

If N(s, fft.res) is larger than the force required to initiate a new crack, N1, more cracks will be
formed. However, if it is smaller only one crack will be formed. The force required to initiate a
new crack, N1, can be calculated as:
 E  
N1  f ctm   Aef   s  1  As  (20)
 E
 c  

where fctm is the average tensile strength.

If N(s, fft.res) > N1 a new crack is initiated (n increases). If N(s, fft.res) < N1 the cracking process
stops and the actual crack width can be determined using expression (17).

4.1 Example
In order to exemplify how the crack width depends on the residual tensile strength, the amount
of reinforcement, and the bar diameter the following example has been analysed, see Figure 8.

Example:
A reinforced “slab-on-grade”, 20 meter long, with full restraint (R=1).
Reinforced with  8, 10 or 12 (0.2% <  < 0.8%)

250

1m c = 30

20 m
Material properties, concrete C30/37 (w/c  0.55):
Tensile strength: fctm = 2.9 MPa (fctk, 0.05 = 2.0 MPa)
Residual tensile strength: 0 MPa < fft.res < 2.5 MPa
Creep coefficient: ef = 2.5
Concrete shrinkage: cs = 600 10-6

Figure 8. Calculation example.

Since the calculation procedure requires iterations, where the number of cracks is step-wise
increased, it is better suited for computer calculations. Hence, the presented model has been
implemented in a small Excel program where the calculation can be made automatically, see
Figure 9.

8(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

Figure 9. Calculation program in Excel.

The calculation results for the calculation example are presented in Figure 10 to Figure 12. As
can be seen the crack width decreases significantly with increasing reinforcement ratio () and
with increasing residual tensile strength. In addition, it can be seen that a small bar diameter is
beneficial; see also Figure 13 which shows how the crack width depends on bar diameter and
reinforcement stress.

9(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

2.5
C 30/37  12
Residual tensile strength [MPa]

2.0

1.5
0.3%
1.0

0.5 0.4%
 = 0.8% 0.6% 0.5%
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Crack width [mm]

Figure 10. Influence of the residual tensile strength and reinforcement ratio () for 12 mm bar.

2.5
C 30/37  10
Residual tensile strength [MPa]

2.0

1.5

1.0
0.3%

0.5

 = 0.8% 0.6% 0.5% 0.4%

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Crack width [mm]

Figure 11. Influence of the residual tensile strength and reinforcement ratio () for 10 mm bar.

2.5
C 30/37  8
Residual tensile strength [MPa]

2.0

1.5
0.25%
1.0

0.3%
0.5

 = 0.8% 0.6% 0.4%

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Crack width [mm]

Figure 12. Influence of the residual tensile strength and reinforcement ratio () for 8 mm bar.

10(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

1
C 30/37, V f = 0% 20
0.8 c = 30 mm
Crack widt [mm] .

16
0.6
12

0.4 10

8
0.2

0
100 200 300 400 500
Reinforcement stress [MPa]

Figure 13. Influence of the bar diameter.

5. DISCUSSION AND CONCLUSIONS

In this paper two models for calculating the crack width for structures with combined
reinforcement (i.e. fibre- and bar diameter) have been presented. The first model is valid for the
case when cracking is caused by an external force while the second model is for structures
subjected to restraint forces. In conclusion it can be said that:
 It is relatively simple to introduce the effect of fibre reinforcement (residual tensile
strength) in models for force induced cracking (crack spacing and crack width).
 Restraint induced cracking, for which models currently is lacking in codes, can be
analysed with the proposed model.
 The Restraint cracking model is more complicated but can easily be implemented in e.g.
Excel for automatic calculations.
 Combined reinforcement (fibre- and bar reinforcement) is effective for crack control.
 However, test methods able to accurately determine the residual tensile strength (or even
better the -w relationship) of FRC is required.

6. REFERENCES

1. Ghali, A., Favre, R. and Elbadry, M.: Concrete Structures – Stresses and Deformations. 3rd
ed, Spon Press, London, 2002.
2. Gustafsson, M. and Karlsson, S.: Fiberarmerade betongkonstruktioner – Analys av
sprickavstånd och sprickbredd (Fibre-reinforced concrete – Analysis of crack spacing and -
width). Examensarbete 2006:105, Institutionen för bygg- och miljöteknik, Avdelningen för
konstruktionsteknik, Chalmers tekniska högskola.
3. RILEM TC 162- TDF: Test and design methods for steel fibre reinforced concrete: --
Design Method Final Recommendation, (Chairlady L. Vandewalle), Materials and
Structures, Vol. 36 October 2003, pp. 560-567.
4. NT BUILD 511: Wedge Splitting Test method (WST) – fracture testing of fibre-reinforced
concrete (Mode I), Nordic Innovation Centre, Oslo, 2005.
5. Löfgren, I.: Fibre-reinforced Concrete for Industrial Construction - a fracture mechanics
approach to material testing and structural analysis. PhD-thesis, Department of Civil and

11(12)
 Presented at Nordic Mini-seminar: “Fibre reinforced concrete”, Trondheim, November 15th 2007.

Environmental Engineering - Structural Engineering, Chalmers University of Technology.

Göteborg, 2005.
6. Engström, B.: Restraint cracking of reinforced concrete structures. Undervisningsmaterial
Institutionen för bygg- & miljöteknik, Chalmers tekniska högskola, 2006.
7. Löfgren, I.: Beräkning av sprickbredd för konstruktioner utsatta för tvångskrafter
(Calculation of crack width for structures subjected to restraint forces). Bygg & Teknik 7/07.

12(12)