Vous êtes sur la page 1sur 19



Prepared by:
Maregesi, Mr Gerald Roosevelt
Impregilo - Kihansi Project

Impregilo Laboratory
Lower Kihansi Hydropower

Monday, February 15, 2010 17:03:00PAPER CONCRETE FOR LOWER KIHANSI DAM



Maregesi, Mr Gerald Roosevelt


This paper presents detailed information on concrete used in construction of Kihansi

dam. The paper outlines the procedure used in concrete production, aggregate
production, cement selection, concrete mix design and measures taken to avoid
thermal cracks in massive dam concrete.


Kihansi dam is 50,000 m3 concrete and 25 metres high gravity dam

which creates a line reservoir of approximately 1,000,000 cubic metres'
storage capacity corresponding to an energy of 300 Megawatts which
will be installed in two phases. The first phase will install three units
which will produce 180 Megawatts and the second phase will install
two units which will produce 120 Megawatts. The underground
power station is located on the Kihansi river at the Udzungwa
The dam is located in U-shaped valley within a unit of Biotitic Gneiss.
The dam was constructed from 1996 and is expected to be
commissioned in March 2000.

The concrete used in Kihansi dam can be divided into two categorises
namely massive concrete and structural concrete. As per technical
specification, the structural concrete was defined as concrete with
construction thickness of less than 2 metres, above this dimension the
concrete was classified as massive.

Table 1: The compressive strength requirements of Kihansi Dam

Concrete class Max. Age of slump Remarks
Agg. size testing (mm)
C25 25 28 70±20 Structural/Massive
C25 64 28 70± 20 Structural/Massive
C30 64 90 30±20 Massive concrete
C25 90 90 30±20 Massive concrete

The compressive strength requirements for the dam concrete are

summarised in Table 1.

Maregesi, Mr Gerald Roosevelt,


Concrete class 25/25 (designated Mix 2 C25/25) was used in all intake
structures, downstream side of overflow section of the dam and at
Concrete class 25/64 (designated Mix 2 C25/64) was used in location
where the concrete was in direct contact with rock/soil or in
unreinforced locations.
Concrete class 30/64 (designated Mix 4) was a massive dam concrete
which was used in the outer part of the dam (Surface).
Concrete class 25/90 (designated Mix 5) was massive dam concrete
which was used in inner part of the dam (Core).
The project specification called for Mix 2 and Mix 4 to be impermeable.

The dam was divided into 15 blocks in form of transversal joints and
the depth of the one lift was 2.5 metres. The lift thickness of 2.5 metres
was selected so as to reduce the number of construction joints which
are likely to be source of structural weakness and seepage through the
dam as the planes of weakness are likely to be created between two
successive lifts.


The project specification favoured the use natural aggregates (sand and
gravel) for concrete production. Since no sources of natural coarse
aggregates were available in the project area, the contractor decided to
use crushed coarse aggregates.

The were several nearby sources of natural sand which were

investigated. It was found that all sources were either excessively
contaminated with mica (biotite and muscovite) or were not
economically viable to exploit. In this case the use of crushed sand was
opted for.

About 80% of aggregates for the concrete used in Kihansi dam were
obtained after crushing biotitic gneiss rock which was hauled from
underground excavation of tunnels in Kihansi. The crushed aggregates
were then transported to the Dam site using dump trucks.
The remaining 15% was obtained from rock hauled from a quarry 10
Kms north of the dam site and 5% were obtained after crushing the
rock coming from shaft excavation at the dam site area.

Maregesi, Mr Gerald Roosevelt,


2.1 Aggregates Quarries and Crushing Plants

Two crushing plants were installed in the project area which were
processing the aggregates to grain size of 0/6, 6/12, 25/64 and 64/90
mm. Typical grading of aggregates produced are shown in figure 1.

The first crushing plant consisting of the primary jaw crusher and
secondary crushing of double cone crusher was installed in Kihansi.
This crushing plant was capable of producing full range of aggregates
needed for dam construction namely 0/6, 6/12, 12/25, 25/64 and
64/90 mm.

The second crushing plant was installed in dam site area. This crusher
plant (consisting of primary jaw crusher and secondary single cone
crusher) was capable of producing aggregates to maximum size of 25
mm. The aggregates produced were 0/6, 6/12 and 12/25 mm.

For economical reasons, all aggregates larger than 25 mm (25-90 mm)

were hauled from Kihansi crusher plant to dam site (16 Kms from
Kihansi) using dump trucks. This was based on the fact that the nearby
quarry was 12 Kms from the dam site in a very rugged terrain and
with road having gradient in excess of 10%. This temporary road to the
quarry needed constant maintenance and during rainy seasons the
road was not passable.
Taking into consideration that abundant rock for crushing was
available in Kihansi as tunnel spoil, it was concluded that it will be
more economical to crush the bigger part of the aggregates needed for
dam construction in Kihansi.

100 0/6 mm
90 6/12 mm
12/25 mm
25/64 mm
64/90 mm
% passing

0.1 1 10 100
Sieve size (mm)

Figure 1: The typical gradings of aggregates used for Kihansi Dam concrete

2.2 Aggregates characteristics

Maregesi, Mr Gerald Roosevelt,

All aggregates for concrete production were obtained after crushing

rock from the Udzungwa scarp. The rock in this area is characterised
by high mica content (biotite and muscovite). The most predominant
rock is biotite gneiss although some tested samples showed a granitic
composition. The main mineral present is quartz, feldspar and biotite
which in most cases is the main constituents.

2.2.1 The Effect of Mica in the Sand

The presence of mica in the sand was affecting the concrete

characteristics by reducing the compressive strength and increasing the
water demand.
From earlier trials conducted on site it was found that the mica in this
area was reducing the cube compressive strength by 8-23%.(2)

Table 2: summary of aggregates characteristics

Test Results Specification
Sodium sulphate Soundness 1.9% 10% Maximum
Sulphate 0.015% 0.25% maximum
Particle density 2.685
Absorption 0.25-0.31%
Aggregate crushing value 25-35% 35% Maximum
Ten percent fine 90-140 KN
Organic impurities Negligible
Alkali silica reaction Not reactive
Flakiness index 16-25% 20% Maximum
Clay lumps 0.58% 2% maximum
Los Angeles abrasion value 42%

Mineralogical analysis

Quartz 20-80%
Feldspar 15-55
Biotite 25-60
Muscovite Trace to 8%

2.2.2 Aggregates’ shape

The shape and the grading of the aggregates were affected by the platy
like structure (rock internal structure) of the mica particles. The
aggregates produced were affected by the cleavage plane and making
the produced aggregates to be flaky and elongated irrespective of
crusher settings. The grading of the aggregates was also affected by the
same. Mica was the weakest mineral within the rock, in this case the
crushed sand produced contained excessive mica and a trend of
increasing the mica content as the grain size decreased was noted. The

Maregesi, Mr Gerald Roosevelt,


mica was crushing down to dust during the crushing process and a
sand with excess of fines was produced.
The physical/chemical/mechanical properties of the aggregates are
summarised in Table 2.


The selection of the cement for production of massive dam concrete is

critical. The cement should be free from free lime, have a low heat
generation and long term mechanical strength. The cement selection
was aimed to get the cement that satisfies the requirements of strength
while having low heat generation so as to minimise the stresses due to
temperature variation.
The specification called for a low heat cement with heat of hydration
of less than 290KJ/g at 7 days as measured using heat of solution
method as per ASTM C186.

Table 3: summary of properties of GGBS

Property Result
(Al+Ca+Mg)/Si 1.72
Ca/Si 0.89
Loss on Ignition (LOI) 0.98 %
Free water 0.00 %
Fineness 3970 cm2/g
Reactivity 7.2 Mpa
Glass content 98 %

Table 4: Chemical composition of GGBS

SiO2 34.3%
Fe2O3 0.87%
Al2O3 15.7%
Mn2O3 1.0%
TiO2 0.6%
CaO2 30.5%
P2O5 0.01%
SO3 0.2%
MgO 12.8%
K2O 1.01%
Na2O 0.2%

Cement brands available locally were tested but none of them was
found suitable for Dam construction. The last and economical solution
was to use imported ground granulated blast furnace slag (GGBS)
which was site blended in ratio of 1:1 with OPC. The OPC was
obtained locally. In order to minimise the variation of cement and
GGBS quality, only one supplier for each OPC and GGBS was chosen.

Maregesi, Mr Gerald Roosevelt,


The OPC complied with BS 12: 1991 as Class 32.5R and the GGBS
complied with the requirements of SABS 1491: 1989.
The summary of the test results of the GGBS used in the Kihansi Dam
are shown in Table 3 and Table 4.


Cement (OPC) and slag(GGBS) were mixed on site manually by

inserting in small hopper one bag of GGBS and one bag of OPC in
repeated sequence which were fed in a motor operated pipe to a
storage silo. After that the blended cement was transported to the
mixing tower silo by compressed air with further mixing taking place
during this transportation operation.
Finally the blended cement with 50% GGBS and 50% OPC was
obtained complying with the requirement of class 32.5N as per BS
12:1991. The efficiency of mixing was checked by sampling the mixed
blend of OPC/GGBS from the mixing tower silo and comparing its
colour with standard laboratory mixed blend of OPC/GGBS.

3.2 Comparison between OPC and OPC/ GGBS blend

After blending OPC and GGBS several characteristics like setting time,
bleeding, heat of hydration, fineness and compressive strength were

3.2.1 Setting time

The setting time of blend of OPC/GGBS (initial and final) was

extended as shown with typical results in Table 5.

Table 5: Comparison between OPC and GGBS/OPC blend

Control 50%:50%
Characteristics OPC OPC:GGBS
Particle density 3.02 2.93
Initial setting time (minutes) 130 190
Final setting time (minutes) 190 240
Fineness (cm2/g) 3145 3450
Standard consistency (%) 25.4 27.4
Le Chatelier Expansion (mm) 0 1
2 days Compressive strength (Mpa) 17.0 7.6
7 days compressive strength (Mpa) 28.0 21.1
28 days compressive strength (Mpa) 42.0 39.1

3.2.2 Bleeding

The measured bleeding of blended cement was higher than that of

OPC apart from the fact that the fineness of GGBS was greater than
that of OPC. The probable reason was the extension of setting time
Maregesi, Mr Gerald Roosevelt,

which increases the time for the concrete to bleed. The bleeding data
are shown in figure 2.

Bleeding %

0 40 80 120 160 200 240 280
Time (minutes)

Figure 2: Bleeding of OPC and OPC/GGBS blend for Mix 5

3.2.3 Fineness

The fineness of GGBS was slightly higher than that of pure OPC. The
resultant fineness after blending was approximately equal to the
weighted average fineness of OPC and GGBS.
The fineness results are shown in Table 5.

3.2.4 Compressive strength

After addition of GGBS to OPC, the early age strength development

was retarded. As early strength is not important for the dam concrete,
this was regarded as an advantage because the slow rate of strength
growth is usually associated with a finer pore structure which results
in concrete with enhanced durability and impermeability.

The compressive strength of the blended OPC/GGBS was retarded at 2

and 7 days. At 28 days, the OPC/GGBS mix attained 93% of strength of
equivalent OPC mix when tested according to BS 12:1991 (EN 196-1).
The typical results are shown in Table 5.

The test results of laboratory trials carried out to compare the strength
development of OPC and blended OPC/GGBS using 350kg/m3 of
binder and water binder ratio of 0.60 is shown in figure 3.
The compressive strength at 3 and 7 days was retarded while the
compressive strength at 28 was 94% in comparison to OPC mix. The
strength at 90 and 180 days was higher than that of OPC mix.

Maregesi, Mr Gerald Roosevelt,



Compressive strength (Mpa)

15 0% GGBS
10 50% GGBS
0 3 7 28 90 180
Age in days (log scale)

Figure 3: Strength development curve for OPC and a blend of OPC:GGBS (Water binder
ratio of 0.60)

3.2.5 Chemical Composition

The comparison of the chemical composition of the OPC and

OPC/GGBS blend is shown in Table 6.

Table 6: The chemical composition of OPC and OPC/GGBS blend

Compound OPC 50% GGBS
SiO2 22.3% 30.4%
Fe2O3 2.1% 1.5%
Al2O3 5.6% 10.6%
Mn2O3 0.07% 0.50%
TiO2 0.27% 0.45%
CaO2 63.5% 48.3%
P2O5 0.04% 0.02%
SO3 2.46% 2.02%
MgO 0.6% 4.3%
Cl 0.01% 0.01%
K2O 0.30% 0.55%
Na2O 0.07% 0.15%

3.2.6 Heat of hydration

After blending OPC and GGBS the heat of hydration was lowered as
shown by test results in Table 6.

Table 7: The total heat under adiabatic condition for OPC and a blend of OPC/GGBS
Age in days OPC OPC/GGBS (1:1)
1 200 120
3 240 210
5 250 240
7 260 245

Maregesi, Mr Gerald Roosevelt,



The concrete mix design was carried out by contractors in their site
laboratory. This involved extensive laboratory testing of aggregates,
cements, mineral additives like silica fume, fly ash, ground granulated
blastfurnace slag (ggbs) and admixtures like water reducer and
The extensive testing of concrete ingredients was aimed at achieving
mixes with better workability at the same time meeting the specified
requirements like strength, durability, permeability and resistance to
The mix design was trying to aim at a mix with maximum aggregate
content with minimum cement content so as to reduce the temperature
rise and thermal stresses. But these were in contradiction with the poor
aggregate quality available in the project area.
The high mica content in the sand and the unfavourable shape of the
fragments of the large particles' sizes led to a tendency of breakage and
high water demand for achieving the desired workability and hence
reduced compressive strength.
In this case the use of high cement content was unavoidable with the
adverse effect of increasing the curing temperature of concrete.

Although it was necessary to use aggregates of poor quality biotitic

gneiss, it was possible to satisfy the concrete requirements by
optimising cement by blending it with GGBS which reduced the heat of
hydration, to use the continuous smooth curve grading to achieve the
maximum density and improved workability, using the water
reducing agent to reduce the water demand and to achieve the
specified compressive strength.

% passing

50 MIX 4
0.1 1 10 100
Sieve size (mm)

Figure 4: Typical grading for massive dam concrete (Mix 4 and 5)

Maregesi, Mr Gerald Roosevelt,


After conducting numerous laboratory trial mixes which took almost

one year the mixes shown in Table 8 were selected.

Table 8: Concrete mixes used for Kihansi Dam

Component Mix 2° Mix 2° Mix 2° Mix 4 Mix 5 Bedding
C25/25 C25/25 C25/64 C30/64 C25/90 Mix
OPC 190 160 165 132.5 105 240
Slag 190 160 165 132.5 105 240
Silica fume 0 20 0 0 0 0
Crushed sand 725 700 495 530 535 1615
Crushed 12/6 285 285 295 325 150 0
Crushed 25/12 850 855 465 445 320 0
Crushed 64/25 0 0 655 730 565 0
Crushed 90/64 0 0 0 0 600 0
Water 215 215 180 155 140 305
Water binder ratio 0.57 0.58 0.55 0.58 0.67 0.64
Plasticizer 1.6 1.4 1.3 1.1 0.8 0
Slump 70±20 70±20 70±20 30±20 30±20 50±20
Permeability <1.0cm <1.0cm <1.0cm 1.0 cm ------- -------
Temp(fresh conc.) 20°C 20°C 20°C 20°C 20°C 20°C
Max. curing temp. 54°C 48 °C 48 °C 43 °C 38 °C -------

7 days strength+ 18.8 18.8 18.8 16.5 12.6 -------

28 days strength+ 30.0 30.0 30.0 26.6 22.3 -------
90 days strength+ 34.1 34.1 34.1 32.2 27.6 -------

Tensile strength * 3.5 3.4 3.5 3.2 2.5 -------

* Tensile splitting strength tested on 150x300 mm cylinder
+ Strength in Mpa

° Statistically all Mix 2 (C25/25 and C25/64) was treated as one mix
The water binder ratio for concrete was defined as whereas ‘w’ is
2k + s + c
water content, ‘k’ is silica fume, ‘s’ is ggbs and ‘c’ is cement content.
The definition of “water binder ratio” implied that the efficiency of silica
fume was set to two while that of GGBS was set to one.


Crack resistance was considered essential and the maximum cement

content and heat of hydration was specified. The specified maximum
heat of hydration for the cement was 290 kJ/g at 7 days. The maximum
specified temperature of fresh concrete was 27 °C and the maximum
temperature during curing was limited to 65 °C.

In most cases the concrete C25 and C30 poses no problem in

production. However, with poor quality aggregate available in the
vicinity of the project, led to concrete to be proportioned with high
cement content to compensate for the aggregate quality with adverse
effect of increasing the effect of heat of hydration of the cement. Due to

Maregesi, Mr Gerald Roosevelt,


this, it was necessary to arrange a method of lowering the temperature

of concrete.
Three critical thermal conditions were considered, these are:-
1. Maximum temperature attained
2. Maximum internal differential
3. Maximum external differential

The maximum temperature attained was measured using thermo

sensors positioned in the middle of the block.
The maximum internal differential was determined by taking the
temperature readings in the middle(core) of the block and another one
at the surface of the block about one centimetre from the concrete
Temperature °C



10 Surface
0 2 4 6 8 10 12
Age in days

Figure 5: checking the internal differential temperature using thermo sensors

The maximum external differential was determined by assessing the

temperature of the lift below by the one on top and the external
difference calculated using the following equation:-

The study carried out on site indicated that precooling the fresh
concrete with chilled water and sprinkling the aggregate with water
will suffice without additional measures of post cooling.
A Koeling containerised water cooling plant with cooling capacity of
23m3/hour and using ammonia as cooling fluid was used to cool the
water to 5 °C which was fed in the inlet at temperature of 20-22 °C. The
cold water was transported to the mixing tower using insulated pipe to
minimise the heat loss. Upon reaching the concrete mixing plant the
water was at 7 °C.

Maregesi, Mr Gerald Roosevelt,




Concrete Temp, Deg


30 50%GG


0 1 2 3 4 5 6

Figure 6: The concrete temperature for a lift of 2.5m and 330Kg/m3 of binder.

The following supplementary measures were also taken to minimise

the adverse effect of temperature:
1) Using minimum cement content to achieve the specified strength
only with very minimum margins.
2) Using the blended OPC and GGBS in ratio of 1:1 so as to lower the
heat of hydration of cement.
3) The use of water reducing agent so as to reduce the water demand
and reduce the cement content.
4) Sprinkling the coarse aggregates with cold water so that some heat
will be lost by evaporation.
5) Managing the construction procedure by limiting the formwork
removal time and monitoring the temperature using thermocouple
in the inner part and the outer part so that the temperature gradient
could be determined.
6) Testing the massive concrete at 90 days instead of 28 days
7) Painting the cement silo with white reflective colour.

Temperature °C

40 OPC
35 50% GGBS
0 1 2 3 4 5 6 7 8
Time (Days)

Figure 7: adiabatic temperature for the sample of OPC and blend of GGBS
Maregesi, Mr Gerald Roosevelt,


The batching of the concrete was carried out using three 2 m3 drum
type mixers with rated maximum output of 50 m3/hour. A single
conveyor was used for feeding the aggregates to the main mixing
plant. The plant was fully computerised to ensure that the concrete is
produced as fast as possible. This type of mixer was selected because of
its ability to mix and discharge the concrete containing large aggregate
size (up to 90 mm) without segregation.

The batching plant comprised of five aggregates bins each having its
own weigh balance. Cement, water and admixture was also equipped
with a separate balance each. The balance was high precision electronic
balance using dynamometers which recorded the deformation of the
metallic specimen under load which produced a resistance change
which was transmitted to the central computer and changed into digit
weight display. The batching results were controlled by a printer
which showed the exact weight of the ingredients batched against the
targeted. The batching precision was set as indicated in Table 9.

Table 9: The batching plant weighing precision

Materials Precision
Cement ±2%
Crushed sand 0-6 ±3%
Crushed 6-12 ±3%
Crushed 12-25 ±3%
Crushed 25-64 ±5%
Crushed 64-90 ±5%
Water ±2%
Admixture ±3%

In order to ensure that the concrete was always batched with a correct
slump, the moisture content of the aggregates was measured every
morning. The moisture content data were inserted in plant computer
and all the adjustments of the weights were done automatically.

The production uniformity was checked on a monthly basis by

calculating the coefficient of variation. The typical data are shown in
figure 7.

Maregesi, Mr Gerald Roosevelt,


Coefficient of variation











Month- year

Figure 8: The monthly 90 days coefficient of variation for massive concrete (Mix 4&5).


Immediately after mixing the concrete was discharged into three

concrete dump trucks each dumper carrying four cubic metres of
concrete. The dump trucks were used to transport the concrete from
the batching plant to a hopper feeding the 42 metres ROTEC Crane
100 x 18 with Port-O-Belt conveyor mounted on Grove crane model
RT75 S.
The concrete was transported and placed in plastic state with slump
ranging from 20-50 mm for Mix 4 and 5. For structural concrete Mix 2,
the slump ranged from 50-90 mm.

The concrete was placed and spread on the mortar (bedding mix) with
a thickness of approximately 30 cm. Effort was made to make sure that
the layer of mortar was covered with concrete before final set of the
mortar. Concrete was placed and vibrated in layers not exceeding 40
cm and the next layer was placed before the setting time of the
previous layer.

After lift completion the surface was cleaned using water and
compresses air to remove the laitance. This operation was carried out
5-8 hours after completing the concrete works.

The lift was left for a minimum of four days before concreting the next
lift. This ensured that heat of hydration of the previous layer has been
dissipated through the upper layer.
The concreting time between two adjacent blocks were set a minimum
of 10 days.

The weather condition at the dam site was favourable allowing the
placement of concrete throughout the year. The monthly placing rate
and cumulative volume is shown in figure 8.

Maregesi, Mr Gerald Roosevelt,


Table 10: Monthly climatic Condition - Dam Site*

Mean Max.temperature(°C) 23 24 23 21 24 22 21 19 21 23 23 25
Mean Min. temperature(°C) 20 19 19 19 17 15 14 16 18 19 21 22
Ave. Temperature(°C) 22 21 21 20 19 17 16 18 19 21 22 23
Ave. Wind speed (m/s) 2 3 4 5 5 4 6 2 2 2 3 1
* Average for the period of 1998-99

6,000 Monthly 50,000

Cumulative volume(m3)
Cumulative 45,000
Monthly volume(m3)

5,000 40,000
4,000 35,000
3,000 25,000
2,000 15,000
1,000 10,000
0 0
Month - Year

Figure 9: Monthly and cumulative volume of concrete in cubic metres


The consolidation of concrete was done using three hand held poker
vibrators having a diameter of 155 mm, amplitude of 2.6 mm and
frequency of 12000 rpm.
The vibration was carried out immediately after placing to ensure that
the concrete compaction was completed before it starts to stiffen. The
vibrator was immersed entirely on the depth of freshly deposited
concrete. The vibration time was ranged from 30-60 seconds in pattern
of about 0.5 metres apart.
The finishing of the vibration was judged by appearance of concrete
after observing mortar flowing up to the surface. Thereafter the
vibrator was gradually removed from the concrete.


The formwork used was quick transfer climbing steel formwork. The
forms for each lift were anchored on the step below with additional
support provided with formwork ties. The formwork was left for at
least 3 days before anchoring to the next lower lift so that the concrete
strength was greater enough to prevent pullout from the anchors.

The removal of formwork was done from 3-6 days depending on the
temperature of concrete during curing so as to reduce the possibility of
having surface cracks caused by thermal shock.
Maregesi, Mr Gerald Roosevelt,

After formwork removal, the concrete was inspected and any surface
irregularities were repaired (depending on the type finishing specified)
using cement mortar mixed with water resistant bonding agent (high
quality synthetic polymer emulsion) to ensure a good adhesion to the


Curing was commenced as soon as possible in order to prevent surface

cracks caused by early drying of concrete. This operation was done
continuously using water sprinkler allowing the concrete to gain
strength before being subjected to tensile stress caused by shrinkage
and volume changes.


Various tests were specified as either pre construction or routine

testing. The testing was carried out using the relevant British,
AASHTO or ASTM standards. Some tests were carried out following
the instruction given in the Technical specification.
The permeability of concrete was tested by subjecting the concrete
block having dimension of 400x400x200 mm to pressure of 2 bar (20
metres head) for 14 days. The concrete was classified as impermeable
when the water penetration was less than 100 mm.

The compressive strength testing was carried out on 150x150 mm

cubes. The large aggregates greater than 32 mm were removed by
screening prior making the cubes. Then the cubes were made and
cured in accordance with BS 1881:Part 111:1983.

The slump testing was also carried out after removing aggregates
greater than 32 mm.


In order to suit site conditions, the following amendments were done

to the Technical Specification :
• The structural concrete strength requirements were relaxed by 5
Mpa and the age of testing for massive dam concrete was increased
from 28 to 90 days because of the poor quality of the aggregates
encountered in the project area.
• In order to facilitate the transportation of the concrete using
conveyor belt the maximum aggregate size was reduced from 120
mm to 90 mm.

Maregesi, Mr Gerald Roosevelt,


• In order to facilitate the hand vibration of concrete, the slump was

increased from 0 to 30 mm.


Apart from the poor aggregate quality encountered in the area, the
construction work of Lower Kihansi dam was carried out successfully.
The concrete produced was generally of acceptable quality with
aggregates embedded in a dense and durable matrix which will
guarantee a satisfactory performance of concrete. The concrete was
produced with overall standard deviation of less than 4.0 Mpa and
coefficient of variation of less than 15% indicating that the quality
control was very good.


The author is very grateful firstly to client, TANESCO for allowing the
details of the project to be published and Mr Charles Kimambo, Site
Manager of TANESCO lower Kihansi for reviewing the paper.


1. Norplan A/S, Lower Kihansi Hydropower Project, Tender

Documents, Contract II Main Civil Works, Volume 2, Technical
Specification, October 1994.
2. Maregesi, G.R, Concrete Making properties of Udzungwa Scarp
Aggregates, 1998 Paper under review.
3. ACI 207.5R-93, Roller Compacted Mass Concrete, ACI Committee
4. ACI 225R-91, Guide to the Selection and Use of Hydraulic
Cement, ACI Committee 225.
5. ACI 305R-91, Hot weather concreting, ACI Committee 305.
6. ACI 207.1R-87, Mass Concrete, ACI Committee 207.
7. ACI 233R-91, Ground Granulated Blast Furnace Slag, ACI
Committee 233.
8. ACI 211.1-91, Standard Practice for Selecting Proportions for
Normal, Heavyweight, and Mass Concrete, ACI Committee 211.
9. ACI 207.4R-93, Cooling and Insulating System for Mass Concrete,
ACI Committee 207.
10. SABS 1491-Part 1:1989, Portland cement extender. Part 1: Ground
Granulated Blast Furnace Slag, Pretoria: South African Bureau of
Standards 1989.
11. ACI 211.3-75, Standard Practice for Selecting Proportions for No-
Slump Concrete, ACI Committee 211.

Maregesi, Mr Gerald Roosevelt,


12. ACI 207.2R-95, Effect of Restraint, Volume Change, and

Reinforcement on Cracking of Mass Concrete, ACI Committee
13. BS 12: 1991, Specification for Portland Cement, British Standard
Institute, 1991.
14. Mnali, S.R, Mineralogical Analysis of the Crushed Sand (Rock)
Sample from Kihansi Crusher Plant, Department of Geology,
University of Dar Es Salaam, Internal Communication to Impregilo,

Maregesi, Mr Gerald Roosevelt,