High strength concrete generally has high viscosity, and hence high segregation resistance even for large slump. On the other hand pumping efficiency is low despite large slump. Thus it can be concluded that the slump may not be a good measure of workability for high strength concrete. A study was therefore conducted to establish a new index to evaluate workability of high strength con- crete by examining its rheological characteristics. A conclusion was that use of rheology constants themselves, such as plastic viscosity or yield value, is more desirable than the direct use of various consistency test results. Figure 3.16 shows that rheology constants can be obtained from the combined results of slump test and ASTM flow test. It was also shown that casting performance
600
500
400
?
£ 300 o
&
I 2 0 0
J 100
"0 400 800 1200 1600 Yield value ri (Pằ)
Fig. 3.16. Estimation of rheology constants from the combination of current consistency tests.
of fresh concrete in the form could be analyzed by viscoplastic divided space element method using rheology constants. Thus the use of rheology constants leads to a good prediction of casting performance of high strength concrete.
3.1.2.2. Standard Test Method for Compressive Strength
Aiming at a proposal of standard test method for the compressive strength of high strength concrete, various factors that would affect the compression test results were examined. They are as follows: characteristics of testing machines such as stiffness or swivel detail, loading speed, end surface treatment of cylin- der by grinding or capping, shape and size of cylinder and its dry or moist condition at testing, manufacture of cylinder such as forms or method of com- paction, and so on. Figure 3.17 shows results of compression test of four kinds of concrete by 16 testing machines (A through P). Some testing machines show consistently low values. It may be the consequence of calibration problem, but it may also be resulted from some difference in some of the above-mentioned influencing factors.
For high strength concrete
J 1 J _ I
130 120 110
2 10°
§ 90 f 80
| 70
•1 60 3
I 50
j 40 30 20 10
° A B C D E F G H I J K L~M~N 0 P Testing machines
Fig. 3.17. Effect of testing machines on the compressive strength.
Based on these studies, a proposal for making compression test cylinders was made based on JIS A 1132, and that for test method for compressive strength was made based on JIS A 1108. As for end surface treatment, un- bonded capping method was developed which does not require any specific end surface finishing. Effect of rubber pad quality and hardness, and that of steel frame diameter was examined by comparing the test results with machine-ground cylinders. For chloroprene rubber pads, increase of compres- sive strength was observed with the increase of rubber hardness in case of high strength concrete. For polyurethane or NBR pads, compressive strength dropped with the increase of rubber hardness when the diameter of steel frame was large. In both cases many cylinders reached the compression failure ac- companied by end chipping or vertical splitting. In conclusion, conditions of unbonded capping that would give equivalent compressive strength and failure mode to machine-ground cylinders were presented.
3.1.2.3. Mechanical Properties
Stress-strain relationship, Young's modulus, and failure characteristics of high strength concrete, basic mechanical properties of confined concrete, and tensile strength were the major items of the series of investigations into mechanical properties of high strength concrete. Results are summarized below.
F,=80 (MPa)
2000 4000 6000 Strain (X10~*) (a) Kent & Park
2000 4000 6000 Strain (X10"') (b) Falitia & Shah
2000 4000 6000 Strain (X10-*)
(e) Muguriuzut
2000 4000 6000 Strain (X10-ô) (d) Popovics
Fig. 3.18. Comparison between measured (full lines) and calculated (dashed lines) stress- strain curves.
Coarse aggregate content — Max.size
a / n3) (nn>
5000| - o 0 ^ 200-20 D 400-20 • 400-15 5, *. 400-10
5 4 500
! i
4 000
3S0O
55 3 000 -
2 500
20 60 100
Compressive strength (MPa) 140
Fig. 3.19. Influence of coarse aggregate on strain at compressive strength.
There are several proposals for the stress-strain relationship of high strength concrete, and some of them represent the test results quite accurately as shown in Fig. 3.18. One problem is the estimation of the strain associated with the maximum stress, which increases gradually as the compressive strength increases, but the trend differs for different coarse aggregate. Furthermore Fig. 3.19 shows the different interdependence of strain at maximum stress on
Compressive strength (MPa)
(a) Without considering ki, k2 and y variations
X
<N 50000
X
\ 40000
sa
-a 3
O : a
00
"60
c a
T3 T3
*z
/ A +
, - °
+ x v
* o
* '
A
<?
A ' * >
*• *
o
• ° •
r .c • "
A River O
• Cnishe<
O Cnished D Crashed O Crushe V Blast F
•
E-;
(k,
ivel Graywacke Quartzile Limestone Andesite nace Slag
.
J*^ l':
A ' RC Equation 3500X1^ xk,,x kj=i, T=2.4)
H Calcine
•^ Crushed d Cnished
• Crashed + Lightw X Lightw
'^f 5 ,
T/2.4)ix(aJ
Bauxite Cobble Basalt Ctaystone ght Coarse Ag ght Fine + Co
e
SO)"1 [MPa]
regale se Aggregate 40 60 80 100 120
Compressive strength (MPa) (b) Considering ki, k2 and y variations
Fig. 3.20. Relationship between compressive strength and Young's modulus.
the compressive strength for different coarse aggregate content or maximum aggregate size. Thus it would be necessary for a stress-strain model to incorpo- rate coarse aggregate related parameters, and further study is needed in this regard.
Available test data of compression test of cylinders were collected to in- vestigate the relationship between Young's modulus and compressive strength of high strength concrete. Figure 3.20(a) shows the straight results. In the figure, the Architectural Institute of Japan (AIJ) equation, which is basically the same as the American Concrete Institute (ACI) equation, is shown in the range of concrete strength less than 36 MPa, in which mass of unit volume 7 is put equal to 2.3 t / m3. A new equation developed in the New RC project is shown in the range of concrete strength greater than 36 MPa, with the con- stant mass of unit volume 7 of 2.4 t / m3. The new equation is different from the AIJ equation in that the exponent to mass of unit volume is 2.0, the exponent to compressive strength is 1/3, and that two coefficients fci and &2 are intro- duced to account for the type of coarse aggregate and mineral admixture. In Fig. 3.20(a), it is clear that the data for lightweight aggregate concrete fall far below that for normal weight concrete, indicating the significance of the term for mass of unit volume. Figure 3.20(b) shows the modified Young's modulus taking into account not only the mass of unit volume but also two coefficients k\ and kz. The scatter of data becomes much smaller than the previous figure, indicating the effectiveness of the New RC equation in predicting the Young's modulus of high strength concrete of wide variety. The detail for coefficients fci and fo can be found in Chapter 8.
Effect of confinement was observed similar to the normal strength concrete, but the confining effect decreased as height-diameter ratio of the specimen increased, and the effect was not influenced by the aggregate type. Effect of confinement is further discussed in Sec. 3.3 of this chapter.
3.1.2.4. Drying Shrinkage and Creep
Research projects aiming at long term behavior of high strength concrete such as drying shrinkage and creep characteristics produced following results.
From mix tests and unified tests for chemical admixtures, it was found that drying shrinkage of high strength concrete is strongly influenced by the rock type of aggregate, water-binder ratio, and dosage of chemical admixture.
Figure 3.21 shows variation of drying shrinkage strain with respect to
1 0 -
X 8 l _
g '* fil—
H °
W=170kg/ms
I
Sandstone
10 15 20 25 30 35 40 45 50 Water-binder ratio (%)
Fig. 3.21. Relationship between water-binder ratio and drying shrinkage.
10 2 8 x
'I 6
2 -
0 —
V=4.10 + 0.90^(n=30)
X _L J _ _L
1.0 2.0 3.0 4.0 5.0 Dosage of admixture (% to cement)
Fig. 3.22. Relationship between dosage of chemical admixture and drying shrinkage strain.
water-binder ratio for two rock types of coarse aggregate. When hard sandstone is used, drying shrinkage increased in proportion to water-binder ratio, but for river gravel it remained high regardless of water-binder ratio. Figure 3.22 shows that drying shrinkage increases when the dosage of chemical admixture is increased. Within the examined test data, unit water content was not found to be influential on the drying shrinkage.
Shrinkage cracks were tested based on the proposed JIS of drying shrinkage crack test method. It was found that high strength concrete develops large shrinkage strain at relatively early age, and shrinkage cracks appear in early days. Figure 3.23 shows the age at crack appearance for various water-cement
SR 40
~ 35
nee (day CO o
S 25
1
•3 20
era
= 15
1 10
5 0
- _
~ - _
~
—
S R ằ •
A
SR o
S Ra
oằ
o
A • A A *
i i i
25 30 35 Water-
S R
Confining plate thickness o
• A A
SR S R
1 40
• l - 6 m m
* 2.0mm - 2.4mm : 2.9mm '. Shrinkage
reducing agent o
•
1 1 1
45 50 55 -cement ratio (%)
O
A •
A
1 60
Fig. 3.23. Age of shrinkage crack appearance.
ratio. With the use of confining plate thickness 2 mm or greater, crack age increased almost proportional to water-cement ratio. The use of shrinkage reducing agent was found to be effective in delaying the shrinkage crack ap- pearance as shown by marks SR in the figure.
Compressive creep test was conducted using concrete with water-cement ratio of 25 to 60 percent in the form of plain concrete columns varying from 20 cm square section to 60 cm square section as well as 10 cm diameter and 20 cm high cylinders. Free shrinkage strain and creep strain tended to be smaller for higher compressive strength of concrete. For 60 MPa concrete smaller creep strain was observed for larger column sections, but creep of 100 MPa concrete did not depend on section size.
3.1.2.5. Durability
In order to evaluate durability of high strength concrete, frost resistance and alkali-aggregate reaction was tested, leading to the following findings.
Freezing-thawing test as specified in ASTM C666 Method A, in-water freezing and in-water thawing method, was conducted for concrete with water-cement ratio of 28 to 55 percent, air content of 2 to 5 percent, and with various curing conditions, as shown in Fig. 3.24. This is the case of concrete using andesite and river gravel combined for coarse aggregate. From the figure it can be seen that the effect of low water-cement ratio on the
< ằ > W / C = 2 8 % ( b ) w / C = 3 2 % ( d w / C = 3 7 % ( d ) w / C = 4 5 % (e) w / C =
- /
= 5 5 %
Curing conditions
• • • • a 2yra. sir exposed
£r • -A 8 week*
/ inair / • — • 2 weeks / in water f Jp o — O 4 weeks JF in water
Fig. 3.24. Relationship between air content and durability factor after different curing.
freeze-thaw resistance is as high as the entrained air, and a clear difference can be seen between water-cement ratio of 28 percent (Fig. 3.24(a)) and 37 percent (Fig. 3.24(c)). However even in case of concrete with water-cement ratio of 28 percent some deterioration can be observed in freeze-thaw resistance after being exposed two years in the outdoor air (dotted line in Fig. 3.24(a)).
Although low water-cement ratio was effective under moist curing conditions, it did not compensate for the low air entrainment under dry condition. Thus it was concluded that certain air content was necessary for frost resistance even for high strength concrete.
The entrained air has conspicuous effect also in preventing frost damage at early ages, and so it is recommended to insure air content of at least 3.5 percent at concrete casting. For concrete with air content of 3.5 percent or more, the minimum curing time to prevent frost damage at early ages is up to the age at which compressive strength of 3.2 MPa is obtained.
Concrete expansion due to alkali-aggregate reaction was measured by the Japan Concrete Institute (JCI) concrete bar method for high strength concrete with unit cement content of 650 kg/m3 and water-cement ratio of 26 or 36 percent, and for normal strength concrete with unit cement content of 350 kg/m3 and water-cement ratio of 56 percent both using reactive or non- reactive aggregate and varying alkali content by adding varying amount of sodium hydrate. Figure 3.25 shows the case of reactive aggregate both for normal and high strength concrete. It is seen that expansion due to alkali- aggregate reaction depends not on the concrete strength, but on the alkali content in the concrete. High strength cannot prevent expansion due to alkali- aggregate reaction. On the other hand, nonreactive aggregate under normal usage has little possibility of producing harmful expansion when used for high strength concrete within usual condition. It was also confirmed that an appropriate replacement of cement with mineral admixtures such as ground
0.10
J
g 0.05
I
0
0.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Age (months)
(a) Normal strength concrete, unit cement content 350 kg/m3
0.10 a
I 0.05
&
0
0.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Age (months)
(b) High strength concrete, unit cement content 650 kg/m3
Fig. 3.25. Comparison of expansion due to alkali-aggregate reaction of normal and high strength concrete using reactive aggregate.
granulated blast furnace slag, silica fume or fly ash fume in case of high strength concrete was effective in preventing alkali-aggregate reaction just as in case of normal strength concrete.
3.1.2.6. Fire Resistance
To evaluate fire resistance of high strength concrete, explosive fracture under varying heating speed was examined of 10 cm by 20 cm cylinder of concrete with water-cement ratio of 25 to 65 percent and unit water content of 140 to 200 kg/m3. Explosive failure occurred most often to concrete with the lowest water-cement ratio of 25 percent. In another heating test of 15 cm by 30 cm cylinders of concrete with varying kind of coarse aggregate and moisture con- tent of concrete, it was found that the moisture content had dominant influence on the fire resistance, and concrete with moisture content less than 3.5 percent did not explode even with the water-cement ratio of 25 percent.
Alkali addition (kg/m3)
• 1.8
* 2.4
• 3.0
• 3.6
J _ I I I 1 1 I I I I I I 1 I L
1100 1000 900
§ 800
- Lengths outside parentheses indicate depths of measuring point w / c : 60% _^_ ] W/C.35% QO02--* " t
IV/C:259S(*) _ - . - • " -^1'." II
^+ằ^ ^0*~'
s
6 0 0 - 5 500
400 300 200 100
1/ / ^ "
I / i&m***^
90 120 150 180 Time after start of heating (min)
Fig. 3.26. Measured interior temperature of concrete during fire resistance test.
Fire resistance test of 50 cm concrete cube specimens was conducted at two months age of natural drying condition, and specimens with water-cement ratio of 35 percent did not explode but those with water-cement ratio of 25 percent exploded. But the same kind of specimens after one-year exposed in the outdoor air under rain shelter showed much milder behavior in the fire resistance test. Figure 3.26 shows measured time history of internal temperature during the fire resistance test at two months age. In the figure results for three different water-cement ratio, i.e. 60, 35 and 25 percent, are shown. The specimen with 25 percent water-cement ratio showed violent explosion and temperature measurement is shown only for reference. It is seen that water- cement ratio has little influence on the temperature rise, and it can be con- cluded that 2 to 4 cm cover is necessary for three-hour fire resistance, in order to keep the steel temperature of a reinforced concrete member below 500 degree Celsius. Also from the fact that temperature rise did not depend on water- cement ratio, it can be inferred that the heat conductivity of high strength concrete is similar to that of normal strength concrete.