Chapter 2 -Prof.

SMA Nanayakkara

Design of water retaining structures

based on BS 8007

2.1 Introduction

The general requirements for concrete structures are adequate strength,

durability and free from excessive cracking or deflection. In case of water-
retaining structures, the most important requirement is the water-tightness
apart from the other requirements. Liquid loss may occur at cracks or at
joints that have been badly designed or constructed, and also from concrete
surfaces where incomplete compaction has been achieved.

Concrete for liquid-retaining structures must have low permeability. This is

necessary not only to prevent leakage through the concrete but also to
provide adequate durability, and protection against corrosion of the
reinforcement and other embedded steel. An uncracked concrete member of
adequate thickness will be impervious to the flow of liquid if the concrete mix
has been properly designed and compacted into position. However, it is
nearly inevitable that some cracking will be present in all concrete structures
due to various reasons. If a concrete slab cracks for any reason, there is a
possibility that liquid may leak or that a wet patch will occur on the surface.
However, it was found that cracks of limited width undergo selfhealing
process and become watertight after few day of exposure to water. To
achieve this condition, it is necessary to control cracking in reinforced
concrete. Therefore the main objective of design of concrete structures to
retain water is to limit the surface crack widths to a predetermined size.

Cracking of concrete can be due to many reasons and can take place in both
plastic and hardened states. The following section explains the types of
cracks, principles of crack formation, significance of crack width on water
tightness and control of cracking.

2.2 Cracking of concrete

It is a well-known fact that reinforced concrete cracks when the tensile stress
due to externally imposed structural loads exceeds the tensile capacity of
2 Chapter 2

concrete. These are known as structural cracks. By means of appropriate

design and detailing techniques, these cracks can be limited to an acceptable
level in terms of structural integrity and aesthetics.

Concrete is also liable to crack in both plastic and hardened states due to
stresses which it intrinsically sustain by the nature of its constituent materials.
The factors and the concrete properties that influence these intrinsic cracks
may be more variable and less predictable than those that influence structural
cracks.

2.2.1 Types of cracks

Out of the two types of cracks, (i.e. structural and intrinsic cracks) the
intrinsic cracks can be broadly categorized into three categories, namely
plastic, early thermal contraction and long term drying shrinkage. Figure 1
shows the classification of cracking in reinforced concrete structures. Table 1
gives possible locations, causes, remedial actions and time of appearance of
intrinsic cracks. Figure 2 shows the typical locations of cracks in concrete
structures. As indicated in Table 2, some of these cracks can be controlled
while some can be avoided.

Table 1. Types of Intrinsic Cracks (Concrete Society 1982)

Type of Most common location Causes Remedy Time of

cracking appearance

Plastic ◦ Deep sections (over r/f) Excess bleeding Reduce bleeding 10 minutes to
settlement ◦ Top of column + Rapid early (air entrainment) 3 hrs.
◦ Change in depth drying /revibrate
Plastic ◦ Roads and slabs Rapid early drying Improve early 30 minutes to
shrinkage (Diagonal cracks) + r/f near surface curing 6 hrs
◦ Reinforced slabs
(random cracks and over /f)
Early thermal Thick walls and slabs Excess heat Reduce heat 1 day to two
contraction generation generation or three
weeks
Long term Thin slabs and walls Inefficient joints ◦ Reduce water Several
drying content weeks or
shrinkage ◦ Improve curing months

Crazing ◦ Against formwork ◦ Impermeable Improve curing One day to

◦ Floated concrete formwork and finishing seven days.
◦ Over-trowelling Sometimes
much later
Corrosion of Columns and beams Lack of cover More than
reinforcement Poor quality two years
concrete
 Design of water-retaining structures 3

Types of cracks

Fresh state Hardened state

Constructional
Plastic Physical Thermal Chemical Structural
movement

Formwork Early thermal Accedental

Plastic shrinkage Drying shrinkage Corrosion of r/f
movement contraction overload

External seasonal
Subgrade
Plastic settlement Crazing temperature Carbonation Creep
movement
variations

Alkali-aggregate
Design loads
reaction

Figure 1. Classification of cracking in concrete

Figure 2 Typical locations of cracks in concrete structures

(Concrete Society 1982)
4 Chapter 2

Table 2. Control of cracking

State of Type of cracking Possible to Possible to

concrete avoid control

Plastic shrinkage yes

Plastic state Plastic settlement yes
constructional movement yes
Structural yes
Drying shrinkage yes
Crazing yes
Hardened state Corrosion of r/f yes
Alkali aggregate reactions yes
Early thermal contraction yes
Thermal contraction due to yes
seasonal temperature variation

2.2.2 Principles of crack formation

The mechanism of formation of cracks in concrete is somewhat complex

largely due to time dependent factors. There are two basic concepts regarding
the formation of cracks in concrete. i.e. tensile strength and tensile strain
capacity of concrete. Tensile strain capacity of concrete is more relevant in
case of cracking of concrete in plastic stage where as cracking of hardened
concrete can be explained by tensile strength capacity of concrete.

2.2.3 Significance of crack width on watertightness

BS 8110 suggests that for structural members that are visible, the calculated
crack width should be less than 0.3 mm. However, BS 8007 recommends a
maximum design surface crack width of 0.2 mm in width for severe or very
severe exposure condition and 0.1 mm for surfaces where critical aesthetic
appearance is important. These limitations imply that all cracks less than 0.2
mm will prove to be watertight under all circumstances. This can be explained
as follows.

When water percolates through cracks it dissolves calcium hydroxide from

the hydrated cement matrix and then, on contact with carbon dioxide
dissolved in water deposits insoluble calcium carbonate crystals on the crack
surface blocking the flow of water. This action can be very effective at sealing
cracks although the process is likely to produce unsightly white deposits on
the surface. This process is known as autogenous healing or self-healing of
 Design of water-retaining structures 5

cracks. According to the Concrete Society Technical Report No.22, it does

appear that under most circumstances cracks about 0.2 mm wide, and which
are controlled by steel still in elastic state will generally become watertight
after a few days of exposure to water.

2.2.4 Controlling of cracking

Since it is necessary to limit crack widths of concrete to prevent leakage of

water in water retaining structures, it is necessary to control the formation of
both structural and intrinsic cracks. It is clear that a crack will not form unless
there is some form of restraint to movement of concrete and that can be
either internal or external. In reinforced concrete, reinforcing bars act as an
internal restraint to movement of concrete due to external loads, temperature
rise due to heat of hydration of cement and environmental conditions. Since
reinforcement acts as an internal restraint, it controls cracking of hardened
concrete but does not have the same influence in plastic concrete. As
discussed earlier, cracking of concrete in plastic state can be avoided by
selecting proper materials and construction methods. Providing
reinforcement could control the other intrinsic cracks such as early thermal
contraction cracks, long-term drying shrinkage cracks and also structural
cracks.

2.3 Structural cracks

Formation of structural cracks in reinforced concrete members under applied

loads can be analyzed based on the elastic theory. If a reinforced concrete
slab is laterally loaded, the concrete on the side of the tension reinforcement
will extend and, depending on the magnitude of loading, it will eventually
crack with the increase in load. The surface crack width depends on the slab
thickness, cover to r/f, reinforcement size and spacing.

2.3.1 Calculation of crack widths

(1) Due to flexural action (without considering effect of compression

r/f)
Procedure for the calculation of surface crack width due to applied
bending moment MS, involves the following.

(a) Calculation of the depth of neutral axis, lever arm and steel stress by
elastic theory,
6 Chapter 2

(b) Calculation of the surface strain allowing for the stiffening effect of
concrete,
(c) Calculation of the crack width,
b C fC
x/3
x C

h d Z = d –x/3

S T
1 fS
As Stress
Strain
Figure 2 Section subjected to flexural action

The formula to calculate the depth of neutral axis x is derived as follows.

Based on Figure 2 the following relationships can be obtained.

𝑇=𝐶

1
𝑓𝑠 𝐴𝑠 = 𝑓 𝑥𝑏
2 𝑐
𝑓𝑠 0.5𝑏𝑥
= − − − −(𝑎)
𝑓𝑐 𝐴𝑠
𝜀𝑠 𝜀𝑐
=
𝑑−𝑥 𝑥

𝑓𝑠 𝑓𝑐
𝐸𝑠 𝐸𝑐
=
𝑑−𝑥 𝑥
𝑓𝑠 𝐸𝑠 𝑑−𝑥 𝑑−𝑥
= = 𝛼𝑒 − − − −(𝑏)
𝑓𝑐 𝐸𝐶 𝑥 𝑥

0.5𝑏𝑥 𝑑−𝑥
𝐹𝑟𝑜𝑚 (𝑎)&(𝑏) → = 𝛼𝑒
𝐴𝑠 𝑥

0.5𝑏𝑥 2 = 𝛼𝑒 (𝑑 − 𝑥)𝐴𝑠

𝑥 2 𝑥 𝐴𝑠
0.5 = 𝛼𝑒 1 − 𝜌 𝜌=
𝑑 𝑑 𝑏𝑑
𝑥 2 𝑥
+ 2 𝛼𝑒 𝜌 − 2 𝛼𝑒 𝜌 = 0
𝑑 𝑑
 Design of water-retaining structures 7

x  2 
  e   1   1 ------------------------------- (1)
d  e  
where  = AS /bd , e = modular ratio = ES / EC

The tensile stress in steel (fS) and concrete compressive stresses (fcb) are then
given by
MS 2MS
fS  , f cb 
zAS zbx
The average strain at the level surface, m, is the difference of apparent elastic
tensile strain in concrete(1 ) (Figure 2) and the strain due to stiffening effect
of concrete between cracks (ε2).
 m  1   2
fS  h  x
Where 1 =  
ES  d  x 

As per BS 8007, 2 is given by equations (2) and (3) for limiting crack widths
of 0.2mm and 0.1mm respectively.

For a limiting design surface crack

width of 0.2 mm :
d S/2 S/2 h
bt ( h  x )( a  x )
/
 2 ---------- (2)
3Es As ( d  x ) acr Cmin

Figure 3 Reinforcement arrangement

For a limiting design surface crack width of 0.1 mm:

. bt ( h  x )( a /  x )
15
 2 -------------------- (3)
3Es As ( d  x )

Crack width is then computed using the following formula given in BS 8007.

3acr  m
w -----------------------(4)
 a  cmin 
1  2 cr 
 hx 
8 Chapter 2

BS 8007 specifies that the above crack width formula given is valid only if the
compressive stress in the concrete f cb  0.45 f cu and the tensile stress in the
steel under service conditions fs < 0.8 fy .

Where a/ - Distance from the compression face to the point at which the
crack width is being calculated
acr - distance from the point considered to the surface of the nearest
longitudinal bar.
As - area of tension reinforcement
bt - width of the section at the centroid of the tension steel
Cmin - minimum cover to the tension steel
d - effective depth
ES - modulus of elasticity of reinforcement
h - overall depth of member
w - design crack width
x- depth of neutral axis
m - average strain at the level where the cracking is being considered

There are design charts to help the designer to obtain very quickly a range of
concrete sections using different thicknesses, cover and diameter of steel for a
particular service bending moment and crack width.

2.3.2 Calculation of crack width due to tensile forces

In water retaining structures, a crack due to tensile force is of greater

importance than a crack due to flexure since the crack due to a tensile force
penetrates the full depth of the section with a more likelihood to allow
leakage. When M/(T h) < 0.5(1-2a/h)2, then the whole section is under
tension with an applied tensile force and a bending moment (no compressive
strain in section). In such situations, the following formulae can be used.
M T b
f S1  
2bh 1 (0.5h  a ) 2bh 1 a
2 = AS2 /bh AS2 fS2 M
T
1 T  h
f S2     1 f S1 
 2  bh  1 = AS1 /bh AS1
a fS1
The surface strain is given by
Section Stress
f S1 ( f S 1  f S 2 )  a
1   Figure 4 Section subjected to combined bending
ES ( h  2a )  E S
and tension
 Design of water-retaining structures 9

The average surface strain m is the difference of surface strain 1 and the
strain due to stiffening effect of concrete between cracks(2, ). As per BS
8007,
2bh bh
2  for 0.2 mm crack width and  2  for 0.1 mm crack width.
3E S AS ES AS

Design crack width w is calculated for a section under direct tension using the
formula, w  3a cr  m . The definition of acr and other symbols are as same as in
the previous section.

2.3.3 Calculation of crack width due to combined tension and bending

Calculation of crack width when both tensile and compressive stresses occur
across the section due to combined tension/compression and bending could
be performed as follows.

Case 1 : Uniaxial bending and compression

1
1

NC is the compressive force and M is the applied moment. T is the tensile

force in bars in tension. The modular ratio m =Es/Ec

Equating the loads on the section.

10 Chapter 2

Assuming the line of application of load Nc is coinciding with the centroid of

the full concrete section ignoring any reinforcement.

Taking moment about tension steel,

 Design of water-retaining structures 11

From the strain diagram

Case 2 : Uniaxial bending and tension

1
1

Nt is the tensile force and M is the applied bending moment. Using the same
symbols as in case 1 , equation the loads:

C’ + C + Nt = T and e = M/Nt
The expressions for e , g , k2 and k3 are exactly the same as in case 2 and g
may be taken equal to h/2 where the point of application of the tensile force
is at the centroid of the full concrete section ignoring steel.

Taking moment about tensile steel,

12 Chapter 2

As in case 2, check assumed value of x and repeat until convergence is

reached. When a satisfactory solution has been obtained, the surface strain 1
can be calculated. The crack width is then calculated by following the same
sequence given in section 2.3.1.

2.3.4 Limitation on steel stress

Crack width limitation requirement in mature concrete due to external loads

may also be considered satisfactory if the stress in the steel under service
conditions does not exceed the appropriate value specified in BS 8007 (Table
3.). This method of satisfying the limit state of crack control requirement is
known as the “deemed to satisfy” method of design. This method does not
give an economical solution because the steel stress is limited to a low value.
For example, for a wall subjected to a service bending moment of 44
kNm/m, Table 4 shows the necessary reinforcement based on both limiting
crack width and limiting stress design.
 Design of water-retaining structures 13

Table 3 Allowable Steel Stress in Direct or Flexural

Tension for Serviceability Limit States

Design Allowable steel stress (N/mm2)

crack width
(mm) Plain bars Deformed bars
0.1 85 100
0.2 115 130

It shows that, under the limitation of steel stress, the crack width is far below
the allowable value and the steel requirement is much higher. Therefore, in
order to reduce the amount of steel required, the design tensile stress in steel
should be as high as possible. This can only be achieved by adopting limiting
crack width design method. However, it should be noted that relatively low
steel percentages at high stresses are not advisable if the site workmanship is
poor. The recommended steel stress is in the range of 220 – 250 N/mm2
(Anchor 1992) when limiting crack width method is used.

Tables 4 Comparison of results obtained from two design methods

Limiting crack width design Limiting steel stress design

Crack width = 0.2mm Steel stress = 130 N/mm2

1010 mm2/m 1610 mm2/m

Steel provided
(16@ 200 mm c/c) (16 @ 125 mm c/c )
Steel stress fs 201 N/mm2 128 N/mm2
Crack width 0.18 mm 0.11 mm
14 Chapter 2

2.4 Calculation of crack widths in relation to thermal and moisture

effects
2.4.1 Cracking due to heat of hydration and drying shrinkage in
immature concrete

Heat generation due to hydration process of cement causes the temperature

of concrete to rise until the reaction is complete, and the heat is then
dissipated to the surroundings. Results of an experimental investigation
carried out to obtain the temperature variation of a concrete wall panel of
300mm thick with 12mm thick plywood formwork are shown in Figure 5. It
can be seen that the temperature rises in concrete during first few hours and
reduces back to environmental temperature after few days. The value of
maximum temperature reached depends on many factors such as the size of
element, cement content, and type of formwork. This will be further
discussed in section 2.4.6.

When the temperature of concrete is increasing, expansion takes place. When

this expansion is retrained by the adjoining section, some creep will occur in
immature concrete relieving the compressive stresses due to expansion.
During subsequent cooling, concrete will shrink and if there are restraints
(internal or external), tensile strains will develop leading to cracking. Similarly
tensile strain will also develop due to drying of concrete if it is restrained
internally or externally.

If a reinforced concrete member is subjected to contraction either due to

thermal movement or due to drying shrinkage, the free shrinkage is prevented
by the steel reinforcement. The steel is therefore under compression and
concrete is in tension, with longitudinal bond stresses on the surface of
reinforcement (Figure 6). Since reinforcement acts as a form of internal
restraint, reinforcement governs the spacing and width of cracks in hardened
concrete. It does not have the same influence in plastic concrete.
 Design of water-retaining structures 15

45

40
Original length
Temperature (C)

35
Unrestrained
shrinkage

30 Compression in steel

Shrinkage
25 restrained by
reinforcement
Tension in concrete
20
0 0.5 1 1.5 2 2.5 3 3.5 4

Time (days)

Figure 5 Temperature Variation of Figure 6 Contraction of Concrete under

Concrete Internal

2.4.2 Crack distribution

In a reinforced concrete element subjected to shrinkage, concrete is under
uniform tensile stress (Figure 6). When the tensile stress exceeds the tensile
stress capacity of concrete, a crack may form at some random points. As
shown in Figure 7, stress in concrete at the crack must be zero and it
gradually increases to some value. Since the tensile stress of concrete within a
distance  S of the crack is less than the tensile strength of concrete, the next
crack must be formed at a distance more than a magnitude S from the first
crack. If the next crack forms at a distance away from the first greater than
2S, then there is a possibility that another crack could form in between these
two cracks as shown in Figure 7. This indicates that the average crack
spacing is between S and 2S. The distance S depends on the efficiency of
stress transfer between steel and concrete.

Idealized
S S S S stress in
concrete

Idealized
S S1 S1 S2 S2 S stress in
concrete

Figure 7 Principles of Crack Spacing

16 Chapter 2

2.4.3 Critical steel ratio

Figure 8 shows reinforced concrete

element subjected to shrinkage.
Since the shrinkage of concrete is
h
resisted by reinforcement, concrete
is under tension while steel is under
b
compression. The compressive
Cross-section Longitudinal section
force in steel is in equilibrium with
tensile force in concrete. When
section contains a low steel Figure 8 Critical steel ratio
percentage, the strength of steel at
yield will be lower than the ultimate
concrete strength in tension and when cracks form, the steel will yield. Thus,
the crack will be wide and unrestrained. Therefore the critical steel ratio (crit
)can be obtained by equating tensile force in concrete of a section to the yield
force in steel of that area of concrete. Hence, referring to Figure 8,
Asfy=bhfct

 AS  f ct
    crit 
 bh  fy
Since the control of cracking is critical during the early life of the concrete,
the tensile strength of concrete fct should be taken as the tensile strength at 3
days as specified in BS 8007. Table 5 gives the values of crit and fct for grade
25, 30 and 35 concrete based on values given in BS 5337 and BS 8007. It is
clear that crit increases with the increase of concrete grade.

Table 5 Critical steel ratio for different grades of concrete

Direct tensile Yield strength of

Concrete strength of reinforcement crit BS Code
grade concrete at 3 days fy (N/mm2)
(N/mm2) fct (N/mm2)
425 0.27
25 1.15 BS 5337
250 0.46
425 0.31
30 1.3 BS 5337
250 0.52
460 0.35
35 1.6 BS 8007
250 0.64
 Design of water-retaining structures 17

2.4.4 Crack spacing

When the steel ratio in a section is greater than crit, the steel does not yield
while concrete shrinks. As the concrete shrinks, the bond between steel and
concrete resist this movement. The bond force between steel and concrete is
in equilibrium with tensile force in concrete. When the tensile stress in
concrete exceeds the tensile capacity of concrete cracks form and at that
instance the bond force is equal to the tensile strength of concrete. The
idealized bond stress distribution along the steel at this instance is shown in
Figure 9. Since the bond stress depends on the relative movement between
steel and concrete, the maximum bond stress is at the crack and bond stress
at the mid point between two cracks becomes zero. Equating the two forces
gives,
f b S  uS  f ct bh

Where S S Bar diameter = 

u S - Total perimeter of bars in the h

width considered
fb - average bond stress adjacent to b
crack Idealized bond
S - bond length necessary to Smax stress distribution

develop cracking force

Figure 9 Idealized bond stress
fct - early age tensile strength in concrete (i.e. at 3 days)
bh - area of concrete

This can be re-written as,

u S

  ( n )

4
, where n = no. of bars and steel ratio  = As/bh
AS  
   (n)
2

4
From the above equation, the bond length (S) necessary to develop cracking
force could be obtained as,
f    f ct  
S   ct  , and the maximum crack spacing S max  2S    .
 f b  4  fb  2

The ratios (fct/fb) for various types of bars and different grades of concrete at
early age are given in BS 5337 and BS 8007 (Table 6). Interestingly , though
18 Chapter 2

the minimum grade of concrete specified in these two Codes is different, the
ratio of (fct/fb) is same for each type of bar.

Table 6 Ratio of (fct/fb)

 f ct 
 
Concrete  fb 
BS Code
grade Deformed bars,
Plain bars
Type 2
BS 5337 30 1.0 2/3
1.0 0.67
BS 8007 35A
(fb=1.6 N/mm2) (fb=2.4 N/mm2)

According to the BS 8007, the steel ratio ( )required to control cracking in

immature concrete is based on the areas of surface zones as shown in Figure
10. The surface zone for walls and suspended slabs is taken as half the
thickness for sections less than 500mm thick. For sections greater than
500mm thickness, the surface zone is taken as 250mm assuming that each
reinforcement layer controls 250mm depth of concrete, ignoring any central
core beyond this surface depth. This means that the thermal steel in each
surface zone remains constant for slabs and walls over 500mm thickness. The
surface zones for ground slabs of different thicknesses are also shown in
Figure 10.

h/2
h/2 300mm > h
500mm  h h/2

h/2 h/2 h/2

h
300 to 500mm
100 mm
250mm
500mm < h
250mm 250m
h> 500mm
250 250
mm mm 100 mm

Surface zones for Surface zones for ground slabs

wall and suspended

Figure 10 Surface zones for walls and slabs

 Design of water-retaining structures 19

2.4.5 Crack width

As long as the steel ratio is greater than the critical value (crit), steel will not
yield and cracks will form in the concrete as shown in Figure 11. The number
and width of cracks will depend on the total contraction strain that is
unrelieved by joints in the length of the section. The contraction strain is the
sum of shrinkage strain and thermal strain including thermal strain due to
changes in ambient temperatures after the structure is completed. Assuming
the tensile strain in concrete varies linearly from a value of zero adjacent to
crack to the maximum value of ultimate concrete strain (ult ) midway between
cracks at a distance Smax apart, the average concrete strain is half of the
ultimate strain (ult ). Therefore the maximum crack width ( wmax) can be
expressed as,
 1 
w max  S max   cs   te   ult 
 2 
where cs – total shrinkage strain
te - total thermal contraction after peak temperature arising
from thermal effects

According to BS 8007, the ult may be assumed to be 200 10-6 and shrinkage
strain in concrete, minus creep strain, is about 10010-6 unless high shrinkage
aggregates are used. Therefore, the above formula becomes,

w max  S max  te 

Smax

Wmax
Concrete tensile
ult strain

Figure 11 Concrete tensile strain

For immature concrete, the coefficient of thermal contraction less its

associated creep strain (which is very high in immature concrete), may be
taken as one-half the value of mature concrete. Hence the thermal
20 Chapter 2

contraction from hydration peak temperature to ambient temperature may be

assumed as,

 te  T1
2
where  (= 1010-6) is the coefficient of thermal expansion of mature
concrete and T1 is the fall in temperature between the hydration peak and
ambient. In addition to the temperature fall T1 there can be a further fall in
temperature, T2, because of the seasonal variations. According to the BS
8007, the estimated thermal contraction due to T2 in mature concrete could
be assumed to be reduced by half due to low fct/fb , high shear resistance and
creep in mature concrete. Hence the value of wmax for rigid end restraints can
be expressed as,

w max  S max T1  T2 
2
Alternatively the above expression can be written as,

w max  S max R T1  T2 

where R is the restraint factor taken as 0.5 for rigid end restraints, after
allowing for internal creep in concrete. This will be discussed further in
section 2.5.8.

2.4.6 Fall in temperature between the hydration peak and ambient (T1)

The previous code, BS 5337, suggested that fall in temperature (T1 ) values
should not be less than 30 C for summer concreting and it should be
increased if

1. the cement content exceeds 340 kg/m3,

2. the wall thickness exceeds 400mm ,
3. the timber shutters are used .

However, guidelines were not given in BS 5337 to select appropriate values

for different conditions. In the present code, BS 8007, the typical values of T1
for UK conditions are given for different cement content, thickness and type
of formwork (Table 7).

The T1 depends on so many factors such as thickness of the section, size of

the member, cement content, chemical composition of cement, type of
 Design of water-retaining structures 21

formwork, concrete placing temperature and ambient temperature. As

indicated in the Table 7, values for T1 are given for concrete placing
temperature of 25 C and mean daily temperature of 15 C which are quite
different to the conditions prevail in Sri Lanka. Experimental studies carried
out at University of Moratuwa[7] indicate that the temperature rise of a wall
panel is smaller than the corresponding values given in BS 8007. This is
futher discussed in Chapter 3. Furthermore, Code BS 8007 states that for
design purposes T1 should be assumed to be not less than 20 C for walls and
not less than 15 C for slabs.

Table 7 Typical values of T1 for OPC concrete (Table A.2 of BS 8007)

1 2 3 4

Walls
Ground slab:
18 mm plywood OPC content,
Section Steel formwork:
formwork: Kg/m3
thickness OPC content,
OPC content,
Kg/m3
Kg/m3

325 350 400 325 350 400 325 350 400

mm C C C C C C C C C

300 11 13 15 23 25 31 15 17 21
500 20 22 27 32 35 43 25 28 34
700 28 32 39 38 42 49 - - -
1000 38 42 49 42 47 56 - - -

Note 1. For suspended slabs cast on flat steel formwork, use data in column 2
Note2. For suspended slabs cast on plywood formwork, use the data in column 4
The table assumes the following:
(a) that the formwork is left in position until the peak temperature has passed.
(b) That the concrete placing temperature is 20 C
(c) That the mean daily temperature is 15C
(d) That an allowance has not been made for solar heat gain in slabs.

2.4.7 Variation in temperature due to seasonal changes ( T2 )

Temperature records of 14 cities over a 10 year period (1987 –1996) were

analyzed to obtain appropriate values for T2 for various locations in Sri
Lanka. It shows that Vauniya has the highest T2 of 24 C. Figure 12 shows
22 Chapter 2

the monthly variation of maximum and minimum temperatures in Colombo

for the year 1996. The value of T2 was taken as the average value of annual
maximum temperature difference for the period 1987 – 1996. The T2 values
calculated for the main cities with available temperature records are given in
Table 8.
36
34
32
Temperature (C)

30
Max. Temp.
Max. temp difference = 13.9 C
28 Min. Temp.
26
24
22
20
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month

Figure 12 Maximum and minimum temperature variation in

Table 8 Variation in temperature due to seasonal changes (T2)

Mean ambient Mean monthly T2
City Temperature minimum
(°C) Temperature (°C) °C
Anuradhapura 28.5 17.9 11
Badulla 23.9 12.3 12
Bandarawela 20.7 10.1 11
Batticaloa 28.2 20.4 8
Colombo 27.5 19.9 8
Galle 27.2 20.6 7
Hambantota 27.7 20.0 8
Katugastota 24.6 13.0 12
Kurunagala 27.5 17.0 11
Mahailluppalama 27.5 16.3 11
Nuwaraeliya 16.5 4.3 12
Puttalam 27.9 17.9 10
Vavuniya 27.9 16.1 12
Average 10
 Design of water-retaining structures 23

2.4.8 External restraint factor ( R )

Though external restraint factor for rigid restraints may be taken as 0.5, in
certain situations actual restraint factor will be less than 0.5 depending on the
end conditions and proportion of the wall panel or slab. Some typical
situations for thin sections subjected to external restraint are illustrated in BS
8007 and those figures are reproduced and shown in Figure 13(a) ~13(d).
Figure 13(a) shows a typical situation where alternate bay wall construction or
wall with movement joints at the ends. In addition, this situation exists in
cases where temporary open sections of width 1m being left between wall
panels of continuous walls. By selecting this arrangement of construction, the
amount of reinforcement to control early thermal cracking can be reduced.
Figure 13(c) shows the variation of restraint factors of a wall panel in a
situation of a sequential bay wall construction. In alternate bay wall
construction, variation of restraint factors of a wall panel cast between two
already cast panels is shown in Figure 13(d). It is important to note that the
full restraint is effective only at 2.4m from a free edge. According to BS 8007,
field investigations have shown that within 2.4m from free edge of a wall or
floor slab no thermal cracking is likely to occur because the contraction strain
does not exceed the tensile strain capacity of concrete over that length. Code
further states that if not less than 2/3 crit based on the surface zone is
provided and there are no obvious stress raisers, it may be assumed that the
free ends of the members will move inwards without cracking up to where R
= 0.5.
L/H R
1 0
2 0
3 0.05
Variation of 4 0.3
Vertical restraint >=8 0.5
factor within
0.2L
Variation of
Horizontal restraint
factor at top Horizontal
restraint factor
vary linearly
between these
0.2L 0.2L Variation of two values
Vertical
Horizontal restraint
R =0 factor at base
H
2.4m 2.4m
<2.4 m,
< 0.2L
0.5 which ever is the greater

0.25
L

Figure 13(a) Variation of restraint factors for a wall fixed at base and free at other three edges
 24 Chapter 2

Variation of Restraint
Where H < L, factor in x direction
Restraint factor across any part of the
in x direction slab where H < L
= 0.5(1-H/L)

Variation of Restraint
0.5 factor in y direction
0.5
Variation of across the middle
Restraint factor strip when H > L
in x direction
along the fixed
0.5 0.5 Vertical R at this
edge
free edge is zero

Variation of
2.4 m
Restraint factor
in y direction
along the fixed
L 0.5
edge
0.5 L/2 Max. L/2 Max.

2.4 m
y Variation of
Restraint factor
in x direction
H across any part of
the slab
0.5 0.5
x

Figure 13(b) Variation of restraint factors for a horizontal slab between rigid restraints

Variation of R in
the horizontal Vertical R at
direction at top this free edge
free edge is zero

Variation of R in
the horizontal 0.25 Horizontal R
direction at 0.25 at this free
vertical fixed edge edge is zero

0.2H
2.4m
Variation of R in
 0.2H 0.2L the vertical
direction at
`
Variation of R in H  2.4 m
vertical free edge
the vertical whichever
direction at is the
greater 2.4m
vertical fixed edge

 0.2L
 2.4 m
0.5 0.5 L 0.25
whichever
is the
greater

Variation of R in
the horizontal Variation of R in
direction at 0.25 the vertical
bottom fixed edge 0.5 direction at
bottom fixed edge

Figure 13(c) Variation of restraint factors for sequential bay wall construction (with
construction joints)
 Design of water-retaining structures 25

Variation of R in
the horizontal
direction at
anywhere in the
wall
Vertical restraint
factor is zero

2.4m
Variation of R in
the vertical
H direction at
vertical fixed edge

H H

0.5 L 0.5 Variation of R in

the vertical
direction across
0.5
the wall where
L > 2H
Where L  2H,
R = 0.5 (1 – L/2H)
Variation of R in
0.5 the vertical
direction across
the wall where
L < 2H

Figure 13 (d) Variation of restraint factors for alternate bay wall

construction ( with construction joints)

2.5 Joints in water retaining structures

Design of the structure against early thermal movement and shrinkage is

closely associated to the frequency and spacing of all types of joints and
hence it is essential for the designer to specify exactly on the drawings where
construction joints should be located.

The types of joints specified in the Code are as follows. The first five are
classified as movement joints.

1. Expansion
2. Complete contraction
3. Partial contraction
4. Hinged
5. Sliding
6. Construction
26 Chapter 2

2.5.1 Movement joints

Movement joints are designed to provide a break in the continuity of a slab,

so that relative movement may occur across the joint in the longitudinal
direction. The joints may provide for the two faces to move apart
(contraction joints)or, if an initial gap is created, the joint faces are able to
move towards each other (expansion joints). Contraction joints are further
divided into complete contraction joins and partial contraction joints.

Other types of movement joints such as sliding joints are needed at the
junction of a wall and roof slab, and where a free joints is required, to allow
sliding to take place at the foot of the wall of a circular prestressed tank.

Description and details of movement joints are given in BS 8007 for the
purpose of each type of joint and method of construction. The designer has
to choose the position and type of joint considered best for a particular
situation. Some designers prefer closely spaced movement joints in
conjunction with a moderate proportion of reinforcement whereas others
prefer no joints at all and use higher quantities of steel in the form of small
diameter bars at close spacing to control cracking.

Table 9 (extracted from the BS 8007) shows the three main options available
for the designer regarding movement joints. It is important to note that the
options given in Table 9 are considered in terms of horizontal movement, but
vertical movement in walls should also be considered. The two cases in walls
are as follows.

1. It is possible for horizontal cracks to occur at any freestanding vertical end

because of the change in horizontal restraint with respect to height. And
for a bay of any height, vertical  0.5horizontal . The vertical steel ratio should
not be less than crit.
2. Vertical restraint exerted on newly cast bay at a vertical construction joint
may be assumed to developed at a depth of 2.4m from the top surface.
Thus for calculating vertical reinforcement, use freedom of movement
(option 3) for the top 2.4m and partial restraint condition (option2) for
the portion below this depth.
 Design of water-retaining structures 27

Table.9 Design options for control of thermal contraction and

restrained shrinkage

Type of
construction
Option Movement joint spacing Steel ratio Comments
and method of
control

1 Continuous: No joints, but expansion joints Minimum Use small

for full restraint at wide spacing may be of crit size bars at
desirable in wall and roofs that close spacing
are not protected from solar to avoid high
heat gain or where the steel ratios
contained liquid is subjected to well in
a substantial temperature range excess of crit

2 Semi-continuos: (a) Complete joints,  15 m Minimum Use small

For partial (b) Alternate partial and of crit size bars but
restraint complete joints ( by less steel
interpolation)  11.25 m than in
(c) Partial joint  7.5 m option 1

3 Close (a) Complete joints, in meters 2/3crit Restrict the

movement joint  4.8 + w/ joint spacing
spacing: (b) Alternate partial and for options
For freedom of complete joints, in meters  3(b) and 3(c)
movement 0.5Smax + 2.4 + w/
(c) Partial joints 
Smax + w/

Furthermore, when it is not possible to have movement joints in situations

such as in tanks where direct tension occurs, Code BS 8007 recommends to
have temporary open sections between panels (Figure 14). The advantages of
this arrangement are that the amount of reinforcement necessary to control
early thermal cracking is minimized and the only thermal effects to be
considered are those resulting from seasonal variations. It is important to lap
the horizontal reinforcement at temporary open sections for freedom of
movement. In addition, appropriate amount of vertical reinforcement should
also be provided.
28 Chapter 2

1m
6m

Figure 14 Temporary open sections in a continuous panels

2.5.2 Contraction joints

Complete contraction joints have discontinuity of both steel and concrete

across the joint, but partial contraction joints have some continuity of
reinforcement. In partial contraction joints, the reinforcement may all
continue across the joint, or only 50% of the steel may continue across the
joint, the remaining 50% being stopped short of the joint plane as shown in
Figure 15.

Figure 15 Complete and partial contraction joints

Contraction joints may be constructed as such, or may be induced by

providing a plane of weakness which causes a crack to form on a preferred
line. In this case, the concrete is placed continuously across the joint
position, and the action of a device which is inserted across the section to
reduce the depth of concrete locally, causes a crack to form. The formation
of the crack releases the stresses in the adjacent concrete, and the joint then
acts as a normal contraction joint. A typical detail is shown in Figure 16.
 Design of water-retaining structures 29

Great care is necessary to position the

crack inducers on the same line, as
otherwise the crack may form away
from the intended position.

2.5.3 Expansion joints

Expansion joints are formed with a Figure 16 Induced contraction joint

compressible layer of material between
the faces of the joint. The material must
be chosen to be durable in wet
conditions, non-toxic (for potable water
construction), and have the necessary
properties to be able to compress by the
required amount and to subsequently Wall

recover its original thickness. An

expansion joint always needs sealing to
prevent leakage of liquid. In a wall, a
water-bar is necessary containing a bulb
near to the centre which will allow
movement to take place without tearing
Floor
and also prevent leakage of water.
Typical details of expansion joints ate Figure 17 Expansions joints
shown in Figure 17.

The joint also requires surface sealing to prevent the ingress of solid particles.
By definition it is not possible to transmit longitudinal structural forces across
an expansion joint, but designer may wish to provide for shear forces to be
carried across the joint, or to prevent the slabs on each side of the joint
moving independently in a lateral direction. If a reservoir wall and footing is
founded on ground that is somewhat plastic, the sections of wall on either
side of an expansion joint may rotate under load by differing amounts. This
action creates an objectionable
appearance and may tear the
jointing materials.

The slabs on either side of an

expansion joint may be
prevented from relative lateral

Figure 18 Expansion joint including dowel

bar to prevent for shear transfer
30 Chapter 2

movement by providing dowel bars with provision for longitudinal

movement (a similar arrangement to a road slab). The dowel bars must be
located accurately in line (otherwise the joint will not move freely), be
provided with an end cap to allow movement, and be coated on one side of
the joint with a de-bonding compound to allow longitudinal movement to
take place.

2.5.4 Construction joints

It is impossible to build a large liquid retaining structure in one piece. Since

the main source of leakage in water retaining structures is joint positions,
considerable attention needs to be given to construction of joints.

In normal structures, the position of the construction joints is specified in

general terms by the designer, and the contractor is allowed to decide on the
number of joints and their precise location subject to final approval by the
designer.

In liquid-retaining structures this approach is not satisfactory. The design of

the structure against early thermal movement and shrinkage is closely allied to
the frequency and spacing of all types of joints, and it is essential for the
designer to specify on the drawings exactly where construction joints will be
located. Construction joints should be specified where convenient breaks in
placing concrete are required. Concrete is placed separately on either side of
a construction joint, but the reinforcement is continuous through the joint.
At a horizontal construction joint, the free surface of the concrete must lie
finished to a completed level surface. At the junction between a base slab
and a wall, it is convenient to provide a short 'kicker' which enables the
formwork for the walls to be placed accurately and easily.

Construction joints are not intended to accommodate movement across the

joint but, due to the discontinuity of the concrete, some slight shrinkage may
occur. This is reduced by proper preparation of the face of the first-placed
section of concrete to encourage adhesion between the two concrete faces.
Joint preparation consists in removing the surface laitance from the concrete
without disturbing the particles of aggregate. It is preferable to carry out this
treatment when the concrete is at least five days old, either by sandblasting or
b), scrabbling with a small air tool. The use of retarders painted on the
formwork is not recommended, because of the possibility of contamination
of the reinforcement passing through the end formwork. The face of a
construction joint is flat and should not constructed with a rebate. It is found
 Design of water-retaining structures 31

that the shoulders of a rebate are difficult to fill with compacted concrete, and
are also liable to be cracked when the formwork is removed. Any shear
forces can be transmitted across joint through the reinforcement. If a
construction joint has been property prepared and constructed, it will retain
liquid without a waterstop.

Designers are under some pressure to use waterstops in construction joints

for obvious commercial reasons and also because it is though that there is less
responsibility thrown onto the designer if a waterstop is specified than if it is
omitted. There had been instances where waterstops have been used and
leaks have been widespread. In other cases, both waterstops and an external
membrane have been specified and again with completely unsatisfactory
results. These examples suggest that workmanship is critical and that
whatever specification is used this point is valid.

It is perhaps also worth stating the obvious, that it is much cheaper to spend
a little more time initially to make a satisfactory job than to have to make
repairs later.

2.5.5 Waterstops

A waterstop (or waterbar) is a jointing strip made of galvanized steel, copper,

rubber bituminous material, plastic (pvc) or other material imbedded in
concrete during construction. They are located across joints in the structure
to provide a permanent liquid-tight seal.

A wide variety of water bars are available. There are two categories of water
bars in general use, namely rigid and flexible. Rigid waterstops are usually
formed from copper or steel strip. Flexible waterstops may be divided into
three categories (See Figure 19).
32 Chapter 2

Flat Dumbbell-shaped
Contraction, partial contraction and construction)

Centerbulb (expansion contraction and partial contraction)

Surface type

Figure Labyrinth-shaped
19 Various types of water bars

a) Central bulb type – use for expansion and contraction joints

b) Dumbbell type – for contraction and construction joint
c) Flat or surface type – in floor slabs

Non-metallic flexible water bars are the most widely used. The principal
materials are natural rubber and PVC (polyvinyl chloride).

All types of water stops should possess the following characteristics.

a) They should be very durable under the operating conditions.

b) They should be non-toxic when in structures containing drinking
water or other potable liquids.
c) They should be inert and not form a breeding place for fungi and
micro organisms.
d) They should be sufficiently rigid to maintain their position shape
during installation and during placing and compaction of concrete
e) They should bond effectively to the concrete in which they are
embedded.
f) They should posses the necessary degree of flexibility and elasticity
for the operating conditions of the joint which they have to seal.
Usually movement at joints in largely cyclic in response to
environmental changed and this may result in periodic reversal of
stress
g) The joints should be made with solvent welding.
 Design of water-retaining structures 33

The waterstop should be sufficiently wide to insure that water path through
the joint is not unduly short. According to BS8007, the distance of the water
bar from the nearest exposed concrete force should not be less than half the
width of the waterstop.

In wall joints, water bars of the fully embedded type should be placed
centrally in the thickness of the concrete and the wings or end bulbs should
be supported by some form of tying to adjacent reinforcement as shown in
Figure 20. Centrally placed water bars should not generally be used in a flat
horizontal position such as in a floor slab joint. If this is unavoidable care
should be taken to ensure full
compaction of concrete under
the water bar. With slabs, flat
type is preferred.

Continuity of the waterstop

system should be provided
between floor and wall systems.
The correct procedure for
fabricating the running joint on
site using heat fused butt joints
for PVC, Vulcanized joints for
rubber and welded lap joints for
Figure 20 Positioning and fixing of
copper or steel must be adopted. waterstops

According to BS 8007, generally waterstops are not required at construction

joints. However waterstops are provided at certain construction joints such
as
(i) those with high bending moments
(ii) at narrow sections
(iii) congested areas which effectively form narrow sections;

Water bars used at construction joints are the plain type with no provision for
movement. Water bars are not normally required at vertical construction
joint because

(i) They can be compacted well

(ii) Water, dirt will not adhere to the surface

Selection of waterstop depends on the hydraulic pressure head to be resisted.

Depending on the shape, thickness, width and material of the waterstop the
34 Chapter 2

maximum water pressure that can be resisted may vary. Since waterstops are
proprietary items selection should be based on manufacture’s
recommendation.

Tests have shown that 225 mm wide rubber waterbars of the dumb-bell type
will successfully withstand, without leakage, pressure equivalent to 70 in head
of water. Similar water-bars in PVC performed less well, and seepage
occurred at a small extension under a pressure equivalent to an 18m head of
water.

The surface type waterstop in PVC withstood a 70 in head of water with an

extension of 15mm when the water pressure was applied to the underside of
the water bar against the full support of the concrete backing. When the
pressure was reversed and the waterstop was completely unsupported from
beneath, seepage occurred under a pressure equivalent to only 1m of water.
Surface type waterstops should therefore only be used to retain water when
they are rigidly supported from behind, eg. under floor slabs cast on concrete
blinding. In this case, the blinding should be laid smoothly and level, and the
waterstop fixed down firmly so that it is fully in contact with the blinding.
Surface type waterstops should not be used in walls to retain water when only
supported by backfilling, but are suitable in this position to prevent the
ingress of water, eg in basements.

2.5.6 Joint fillers and sealing compounds

Joint fillers are compressible sheets De-bonding Sealing

used in expansion joints. Filler tape compound
board should accommodate
compression and recovers when Filler
concrete faces move apart. It shall board
be non-rotting and non-absorbent
material. Joint filler is also act as a
support to joint sealing compound
as shown in figure below.

Joint sealing compounds are

impermeable ductile material. In
many cases sealing can be delayed
until just before the structure is put
into service - joint opening to be Deformation of sealing compound
accommodated is small when it is de-bonded
 Design of water-retaining structures 35

 Apply bond breaking tape to prevent adhesion of sealant to filler material

References

1. CP 2007: Part 2, Design and construction of reinforced and prestressed

concrete structures for the storage of water and other aqueous liquids,
British Standards Institution. London. 1970
2. BS 5337: Code of practice for the structural use of concrete for retaining
aqueous liquid, British Standards Institution. London. 1976,
3. BS 8110:Structural use of concrete, British Standards Institution. London,
1985.
4. BS 8007:Design of concrete structures for retaining aqueous liquid, British
Standards Institution. London, 1987.
5. ‘Non- structural cracks in concrete’, Concrete Society Technical Report
No. 22, The Concrete Society, 1982.
6. Anchor R.D., Design of liquid retaining concrete structures, Edward
Arnold, London, second edition,1992.
7. Nanayakkara S.M.A, Liyanage W.W. , Use of Grade 25 concrete in
Design of Water Retaining Structures Based on BS 8007
Recommendations , IESL Transactions, 2003, pp 1-8.
8. Nanayakkara S.M.A. Wannigama W.R.K. ,Experimental Investigations on
Temperature rise due to heat of Hydration, IESL Transactions, 2003, pp.
9-15.
9. Nanayakkara S.M.A, Self-healing of cracks in Concrete under water
pressure, International Symposium on New Technologies for Urban safety
in Mega Cities, Tokyo, October, 2003. pp 125-132.