Professional Documents
Culture Documents
Allena Paper 4-14-10 PDF
Allena Paper 4-14-10 PDF
Abstract
This paper presents the development of ultra high strength concrete (UHSC) using
local materials. UHSC mixture proportions were developed using local materials so that
UHSC may be made more affordable to a wider variety of applications. Specifically,
local sand with a top size of 0.0236 in. (600 μm), and locally available Type I/II cement
and silica fume were used in this research. Each of these material selections is seen as an
improvement in sustainability for UHSC. Two mixtures (one with and one without fibers)
were recommended as the UHSC mixtures. The greatest compressive strengths obtained
in this study were 24,010 psi (165.6 MPa) for UHSC with steel fibers and 23,480 psi
(161.9 MPa) for UHSC without fibers. The compressive and flexural strengths obtained
from the UHSC mixtures developed in this work are comparable to UHSC strengths
presented in the literature. Producing this innovative material with local materials reduces
the cost of the material, improves sustainability, and produces mechanical performance
similar to prepackaged, commercially available products.
Introduction
Background
UHSC is composed of cement, fine sand, quartz powder, micro silica, steel fibers
and HRWRA. When used in optimum dosages, the HRWRA reduces the water- to-
cementitious materials ratio while improving the workability of concrete. The addition of
micro silica enhances the mechanical properties of the paste by filling voids, enhancing
rheology, and producing secondary hydrates. The quartz powder is useful for its
reactivity during heat treatment (Dili and Santhanam 2004). The constituents of the
mixture and proportions (by fraction of cement mass) proposed by various investigators
(Richard and Cheyrezy 1995, Shaheen and Shrive 2006, and Matte and Moranville 1999)
are presented in Table 1.
Roux et al. (1996) demonstrated that the mechanical properties of RPC are
obtained by lowering the water-to-cementitious materials ratio, using HRWRA’s, and
including silica fume. The lower water-to-cementitious materials ratio reduces the
porosity of the cement paste and improves durability. Richard and Cheyrezy (1995)
recommended the following principles to develop UHSC:
Materials used in UHSC such as quartz dust and steel aggregate are often shipped
long distances, internationally in many cases, increasing the cost of the material.
Additionally, strict requirements on the chemistry of the cement and silica fume increase
the cost of commercially available, prepackaged UHSC products such as Ductal.
Expensive materials such as ground quartz and fibers that are not available locally are
used in the production of Ductal, resulting in increased cost of the final product.
Therefore, the present work focused on developing UHSC mixture proportions using
local materials so that UHSC may be made more affordable to a wider variety of
applications. Specifically, local sand with a top size of 0.0236 in. (600 μm), and locally
available Type I/II cement and silica fume were used in this research. Each of these
material selections is seen as an improvement in sustainability for UHSC. Reduced
member sizes are also possible with UHSC, which reduces the volume of concrete used
to produce a given structural element.
Experimental Program
Materials. Type I/II portland cement, silica fume, and fine local sand (0.00295-0.0236 in
[75-600 μm]) from Las Cruces, New Mexico were used. Table 2 shows the grain size
distribution for the fine sand. The chemical compositions of the Type I/II portland cement
and silica fume are provided in Table 3. Steel fibers that were 0.5 in. (13 mm) long were
used to provide ductility. To achieve the desired workability, a commercially available
polycarboxylate-based HRWRA (Glenium 3030 NS from BASF Chemicals) was used.
Concrete Mixtures. Table 4 shows the mixture proportions, the water-to-cement ratio
and the water-to-cementitious materials ratio (including silica fume) for a group of 7
mixtures investigated during the development of the UHSC. These mixtures were divided
into five categories. Several other mixture proportions were also evaluated. However,
only these mixtures are presented for the sake of brevity. The mixtures were categorized
as follows:
A : The mixture in this category used Type I/II portland cement and HRWRA.
The aggregates for this mixture were thoroughly washed and dried prior to use.
Steel fibers were used to fill 1.5% of the volume. However, no silica fume was
used in this mixture.
B : The mixtures in this category also used Type I/II portland cement and
HRWRA and the aggregates were thoroughly washed and dried prior to use.
Silica fume and steel fibers were both used in these mixtures.
C and D: The mixtures in these categories used Type I/II portland cement and
HRWRA. The aggregates were thoroughly washed and dried prior to use. Silica
fume was used in these mixtures. However, steel fibers were not used in these
mixtures. The mixture in category D had lower water-to-cement and water-to-
cementitious materials ratios than the mixtures in category C.
Mixtures within each category were assigned a name. Mixture names were four
digit symbols where the first letter referenced the category for the mixture, the second
letter referred to the source of the aggregate, and the last two numbers referred to the
percentage increase in water and cementitious materials content over the base mixture.
For example, mixture BL20 was a mixture from category B that used aggregates from
Las Cruces, New Mexico and had 20% more water, cement, and silica fume than the base
mixture (BL00).
Specimen Preparation. Sand used in this study was passed through a No. 30 sieve to
obtain the desired particle sizes. The aggregates were then thoroughly washed over a No.
200 sieve. After washing, the sand was oven dried at 110°F (44°C) to achieve 0%
moisture content.
The constituents of each mixture were mixed for 15 to 20 minutes using a
laboratory pan mixture. The dry constituents were mixed for 2 minutes and then 75% of
the water was added. After thorough mixing, the HRWRA was added and the remaining
25% of the water was added to the mixture. Steel fibers were added at the end. This
process of mixing seemed to improve the action of the HRWRA.
Compressive strength specimens included 2 in. (50 mm) cubes and 4 by 8 in.
(100 by 200 mm) cylinders. To avoid problems with end preparation of cylinder
specimens, only 2 in. (50 mm) cubes were used as compressive strength specimens for
mixture categories C, D, and E. Modulus of rupture testing was conducted on 3x4x16 in.
(75x100x400 mm) prisms. All specimens were consolidated using a high frequency
vibrating table.
Curing Regimens. Three curing regimens were investigated. For the first regimen,
concrete specimens were cured at room temperature, 65°F (20°C), for the first 24 hours.
After demolding, the specimens were moist cured at a temperature of 73.4°F (23°C) and
a relative humidity of 100% until the day of testing and this curing regimen was
designated as MC.
In the second curing regimen, concrete specimens were cured at room
temperature, 65°F (20°C), for the first 24 hours. After demolding, the specimens were
heat cured in a water bath at 122°F (50°C) until the time of testing. This curing regimen
was designated as WB.
For the third curing regimen, concrete specimens were cured at room temperature,
65°F (20°C), for the first 24 hours. After demolding, the specimens were heat cured in a
water bath at 122°F (50°C). Then, the specimens were removed from the water bath and
dry cured at 392°F (200°C) for two days prior to testing. This curing regimen was
designated as OV.
Modulus of Rupture Testing. Prismatic specimens that were 3x4x16 in. (75x100x400
mm) were tested to evaluate modulus of rupture. Modulus of rupture was determined
using a third point loading test on prism specimens according to ASTM C 78. Specimens
were rotated 90 degrees from the orientation in which they were cast to measure the
flexural strength. Consequently, the 4 in. (100 mm) dimension was aligned with the
direction of loading.
Category A. In the process of developing UHSC, mixture AL00 was produced as the
first trial mixture. AL00 was produced with a water-to-cement ratio of 0.28 and no silica
fume was used in this mixture. Compression tests were conducted on 4 by 8 in. (100 by
200 mm) cylinders after 7 days of moist curing. Mixture AL00 produced a compressive
strength of 6940 psi (47.86 MPa). This compressive strength was considered low in
comparison to the usual compressive strength range for UHSC. The low compressive
strength of mixture AL00 was initially attributed to the absence of silica fume. Therefore,
the subsequent mixtures (categories B, C, D, and E) in the research were produced using
silica fume.
Category C. For the next step in trying to improve the compressive strength, the water-
to-cementitious materials ratio was reduced to 0.22 (from 0.28 for BL00) to produce
mixture CL00.The reduced water-to-cement ratio was produced by increasing the cement
content by 25.31% and decreasing the water content by 7% (compared to BL00). No steel
fibers were used in mixture CL00, and the sand content in CL00 was reduced by 5.5%
compared to BL00.
At this stage, it was decided to use cube specimens for compression testing to
avoid problems with the end preparation of cylinder specimens. Compressive strengths
for all remaining mixtures were measured using 2 in. (50 mm) cube specimens. The size
of the cubes was selected such that the cube dimension was at least four times the size of
the largest particle used in the concrete (0.5 in. [13 mm] steel fibers). CL00 specimens
were cured according to curing regimen WB and tested at 7, 14, and 28 days. The
compressive strengths measured at 7, 14, and 28 days were 14,080 psi (97.10 MPa),
14,250 psi (98.28 MPa), and 16,250 psi (112.06 MPa), respectively.
To produce mixture CL20, CL00 was further modified by increasing the paste
volume. The percent volume of paste was 62.85% for CL00 and 75.44% for CL20. Cube
specimens produced from mixture CL20 were also cured according to curing regimen
WB and tested for compressive strength at 7, 14, and 28 days. The compressive strengths
at 7, 14, and 28 days were 13,560 psi (93.52 MPa), 14,360 psi (99.03 MPa), and 15,200
psi (63.52 MPa), respectively.
The next step in attempting to increase compressive strength was to try to
accelerate the pozzolanic reaction of the silica fume. To accelerate the pozzalonic
reaction of the silica fume, concrete specimens produced from CL00 were cured
according to curing regimen OV. A similar curing regimen was adopted by Reda et al.
(1999). Curing of concrete specimens at autoclave temperatures between 392ºF and
572ºF (200ºC and 300ºC) has been shown to accelerate the pozzalonic reaction and help
in the formation of dense calcium silicate hydrate compounds (Shaheen and Shrive
2006). In this study, curing at 392ºF (200ºC) was achieved in an oven. There were
concerns that dry curing may artificially inflate the compressive strengths. However, no
adverse effects were observed in flexural strength testing. This seems to indicate that
strength gains caused by oven dry curing at 392ºF (200ºC) were primarily due to the
accelerated pozzolanic reaction, not drying. Compressive strengths of the specimens
produced from CL00 (OV) were greater than the compressive strengths of specimens
Category E. Mixture EL00 was produced by introducing steel fibers to the mixture
proportions from DL00. The volume of sand was reduced by the volume occupied by
steel fibers in mixture EL00. Specimens produced from EL00 were cured according to
curing regimen OV. The greatest compressive strengths among all the mixture categories
used in this investigation were achieved using mixture EL00. The compressive strengths
achieved at 7, 14, and 28 days were 21,180 psi (146.06 MPa), 23,420 psi (161. 52 MPa),
and 24,010 psi (165.60 MPa), respectively. The greater compressive strength of the fiber
reinforced UHSC mixture is consistent with results reported by Reda et al. (1999).
Mixtures DL00 (OV) and EL00 (OV) were identified as UHSC and will be considered
for further research.
Strength Gain Versus Time. Compressive strength gain with time was investigated by
considering the plain (CL00 and DL00) and fiber reinforced (EL00) concrete mixtures.
All of the specimens were cured according to curing regimen OV. The ratios of
compressive strength at 7 and 14 days with respect to 28 day strength (assuming
compressive strength at 28 days is 100%) of mixtures CL00, DL00, and EL00 are
presented graphically in Figure 1. For mixtures CL00 and DL00, compressive strengths at
7 days were 84.0 and 85.6 percent of 28-day compressive strength, respectively. This
percentage was 88.2 for mixture EL00. Compressive strengths for CL00 and DL00 at 14
days were 96.4 and 94.6 percent of 28-day compressive strength, respectively, whereas
this percentage was 97.5 for mixture EL00. This strength development is more rapid than
the strength development of normal strength concrete and is attributed to the high
cementitious content of the UHSC mixtures and post-setting heat treatment that increased
the rate of the hydration reaction.
Effects of Steel Fibers. The effects of steel fibers on compressive strength of UHSC
were investigated by considering mixtures DL00 (without fibers) and EL00 (with fibers).
The specimens from both of these mixtures were cured according to the OV curing
regimen. Figure 2 illustrates the influence of steel fibers on compressive strength of
UHSC mixtures. The average percentage increase in compressive strength due to the
presence of steel fibers was 5.43, 5.45, and 2.25 percent at 7, 14, and 28 days,
respectively.
One concern with UHSC specimens cured at elevated temperatures was that drying of the
sample might artificially inflate the compressive strength of the sample and decrease the
flexural strength. However, it was also believed that steel fibers would improve the
ductility of the concrete and could also increase the flexural strength of the concrete. To
investigate these potential effects, modulus of rupture was measured on 3x4x16 in.
(75x100x400 mm) prismatic specimens.
Average flexural strengths from mixture categories A, B, D, and E, measured at 7
days, are presented in Table 6. It can be seen from Table 6 that the greatest flexural
strength was obtained with mixture EL00 (2655 psi [18.31 MPa]). This strength is
consistent with flexural strengths obtained by other investigators for UHSC produced
without dry curing. Again, this indicates that the dry curing at 392°F (200°C) served to
accelerate the pozzolanic reaction of the silica fume.
A significant effect of steel fibers on the flexural strength of UHSC was also
observed (Table 6). The flexural strength of the UHSC with fibers (EL00) was greater
than that of the plain mixture (DL00) by 68.3%. The flexural strengths of plain and fiber
reinforced UHSC mixtures and the percentage increase in flexural strength due to
addition of steel fibers were consistent with the literature (Collepardi et al. 2007 and Dili
and Santhanam 2004).
Conclusion
This paper summarizes several steps in the development of UHSC using local
materials. The conclusions drawn during the course of this work are:
1. UHSC was developed with materials locally available in Las Cruces, New
Mexico that produced a compressive strength of 24,010 psi (165.6 MPa) and a
flexural strength of 2655 psi (18.3 MPa).
2. The strength properties of UHSC produced with local materials were similar to
those provided by prepackaged, commercially available products such as Ductal.
3. Prolonging the mixing period increased the workability of the mixtures.
4. Specimens cured with oven drying attained greater strength than specimens that
were moist cured or cured in a 122°F (50°C) water bath. The greater compressive
strength exhibited by oven dried specimens was attributed to the acceleration of
the hydration reaction when the specimens were kept in a water bath 122°F
References
Collepardi, S., Coppola, L., Troli, R., and Collepardi, M. (2007). “Mechanical Properties
of Modified Reactive Powder Concrete.”
www.encosrl.it/enco%20srl%20ITA/servizi/pdf/high/12.pdf.
Dili, A. S., and Santhanam, M. (2004). “Investigations on Reactive Powder Concrete: A
Developing Ultra High Strength Technology.” The Indian Concrete Journal,
78 (4), 33-38.
de Lerrard, F., and Sedran, T. (1994). “Optimization of Ultra-High Performance
Concrete by the Use of a Packing Model.” Cement and Concrete Research, 24
(6), 997-1009.
Ferraris, C.F., and de Larrard, F. (1998). “Testing and Modeling of Fresh Concrete
Rheology.” Building and Fire Research Laboratory, No. NISTIR 6094,
National Institute of Standards and Technology, Gaithersburg, MD, 1-61.
Matte, V., and Moranville, M. (1999). “Durability of Reactive Powder Composites:
Influence of Silica Flume on the Leaching Properties of Very Low
Water/Binder Pastes.” Cement and Concrete Composites, 21, 1-9.
Reda, M.M., Shrive, N.G., and Gillott, J.E. (1999). “Microstructural Investigation of
Innovative UHPC.” Cement and Concrete Research, 29 (3), 323-329.
Richard, P., and Cheyrezy, M. (1995). “Composition of Reactive Powder Concrete.”
Cement and Concrete Research, 25 (7), 1501-1511.
w/c R, psi
Category Mixture w/(c+sf)
ratio (MPa)
AL00 1675
A 0.28 0.28
(MC) (11.55)
BL00 1450
B 0.37 0.28
(MC) (10.00)
DL00 1585
D 0.25 0.20
(OV) (10.93)
EL00 2655
E 0.25 0.20
(OV) (18.31)