Académique Documents
Professionnel Documents
Culture Documents
Future construction of concrete floating platforms for offshore oil Esfahani and Rangan4 carried out testing to determine the
exploration off the east coast of Canada will lead to a substantial effects of rib face angle on the bond capacity in high-strength
increase in the use of high-strength lightweight (HSLW) concrete. concrete. The results indicated that the bond strength of bars
HSLW concrete has been extensively used in Norway and other with rib face angles between 23 and 27 degrees is significantly
parts of Europe. HSLW concrete with its high durability and light-
weight characteristics is a very much sought after material in the lower than bars with rib face angles between 40 and 47 degrees.
construction of concrete floating platforms. Azizinamini et al.5 closely examined the bond performance
The main objective of this investigation is to determine the bond and tension development length of reinforcement steel in
strength characteristics of HSLW concrete. The experimental program high-strength concrete. From these tests, it was concluded
consisted of testing 72 pull-out and push-in specimens to evaluate the that increasing the development length was not the most
bond behavior under monotonic and cyclic loading using 25 and efficient way to increase bond capacity in high-strength
35 mm (No. 8 and No. 11) deformed reinforcement embedded in high concrete. Bond behavior of high-strength normalweight
strength of 80 MPa (11.6 ksi) lightweight concrete. concrete (HSNW) under cyclic loading was examined by
The static and cyclic test results indicated that HSLW concrete Alavi-Fard and Marzouk.6
behaves very similar to high strength normalweight concrete. The
bond strength of HSLW was equal to or slightly larger than similar
Structural lightweight concrete is generally considered to
specimens made of high strength normalweight concrete (HSNW) be concrete with compressive strength in excess of 17.5 MPa
at the same lab. The bond stress versus displacement curve indicated (2530 psi) and unit weight of 1950 kg/m3 (3290 lb/yd3) or
a sharp linear ascending portion followed by a steep descending less. In recent decades, however, HSLW has gained popularity,
portion indicating a brittle failure, which is characteristic of high- especially among those designers who strive to design and
strength concrete. The ACI 318 code increases the minimum construct structures from concrete that would traditionally
development length of lightweight concrete by 30%. The ACI be built in structural steel, such as long span bridges and
recommendation of this increase is unjustified for high-strength floating oil platforms. With this increased use of HSLW
concrete made of lightweight aggregates.
concrete, it is of the utmost importance that design engineers
understand the mechanical properties of HSLW concrete.
Keywords: bond; high-strength concrete; lightweight aggregate.
The main intention for issuing the Eurocode 2,7 “Lightweight
Aggregate Concrete with Closed Structure,” is that it becomes
INTRODUCTION the standard for the design of structures to be used throughout
While much research has been performed on the bond Europe. The advantages of using lightweight in a marine
characteristics of normal strength normalweight concrete, environment were illustrated by Fidjestol.8 Recently, a good
little has been done on high-strength concrete, especially seismic review was conducted on the advantageous behavior of
high-strength lightweight (HSLW) aggregate concrete. In high-performance structural lightweight concrete.9 Bond
addition, much of the research and development in the area behavior of HSNW concrete was investigated in similar
of HSLW concrete was performed as part of technical studies research at the same lab by Alavi-Fard and Marzouk.10
associated with the construction of specific structures, mainly
offshore concrete platforms. As a result, much of the
research pertaining to the structural behavior of HSLW RESEARCH SIGNIFICANCE
concrete is confidential between the various research The bond strength between HSLW concrete and steel
institutions and the oil companies. reinforcement is an important factor in designing any rein-
Several research workers, such as Clark,1 used the pullout forced concrete structure under various kinds of loading. The
test to study the effect of various reinforcing bar patterns in experimental result of testing the bond strength of HSLW is
normal strength concrete. It was concluded that the most very valuable because the number of published experimental
effective deformation pattern was one where the shearing data in this area is still very limited. HSLW concrete has
area is less than 10 times the deformation bearing area, with been used extensively in marine applications and offshore
the shearing area defined as the perimeter of the bar multiplied construction of concrete platforms in the North Sea and
by the deformation spacing. Eligehausen et al.2 conducted recently in the construction of the Hibernia platform and the
one of the main comprehensive investigations on the bond- proposed Hebron platform in the Atlantic. Over the last
slip relations using the pull-out test. The results of this decade, interest in the actual performance of marine HSLW
investigation defined the bond behavior of deformed bars concrete has grown, including its durability, mechanical
under monotonic and cyclic loading. Primary and secondary properties, and structural performance.
cracks were observed in research by Kollegger and Mehlhorn.3
From studying the contact surface between the concrete and ACI Structural Journal, V. 104, No. 1, January-February 2007.
MS No. 05-314 received December 6, 2005, and reviewed under Institute publication
reinforcement steel, it was concluded that complete policies. Copyright © 2007, American Concrete Institute. All rights reserved, including
compatibility between concrete and steel is based on steel the making of copies unless permission is obtained from the copyright proprietors. Pertinent
discussion including author’s closure, if any, will be published in the November-
stress, bond stress, and concrete stress. December 2007 ACI Structural Journal if the discussion is received by July 1, 2007.
EXPERIMENTAL INVESTIGATION
A structural steel frame at the structures lab of Memorial
University of Newfoundland was used for the experimental
programs. The frame was designed, fabricated, and erected
as part of previous experimental work on high-strength
concrete and other offshore investigations at the Structures
Laboratory of Memorial University of Newfoundland. The
current test program on HSLW concrete was very similar to
Fig. 1—Test setup.
the program using HSNW concrete conducted at the same
lab using the same test set-up, the same specimen size, and
the same casting position by Alavi-Fard and Marzouk.10 The
current experimental investigation consisted of constructing
36 specimens for each for 25 and 35 mm (No. 8 and No. 11) bars.
A photograph of the test set up is shown in Fig. 1. The
dimensions and configuration of the tested specimens is
shown in Fig. 2. The loading frame was equipped with an
electro-hydraulically controlled testing actuator capable of
applying loads of +670 kN (150 kips). The instrumentation
for the test setup consisted essentially of an actuator with a
load cell attached to measure the load being applied. The
movement at the loaded end of the bar was measured using
the linear variable differential transducer (LVDT) built into
the actuator, whereas the movement at the free end of the bar
was measured using a linear potential differential transducer
(LPDT), which was mounted externally using a magnetic
mounting apparatus.
The load being applied and the corresponding displacements,
along with the strain gauge readings, were continuously
scanned and recorded by the data acquisition system in addition
to being displayed on the monitor. A summary of the test
program is presented in Table 1 to 3. Two bar sizes were Fig. 2—Dimensions of bond test specimen.
used in constructing the specimens. The tested specimen has
a standard coarse thread on one end of the bar to be attached
to the loading frame. The diameter of the tested bars was 25 and (810 and 826 lb/yd3) using 12% silica fume by weight and a
35 mm (No. 8 and No. 11) with a yield strength of 400 MPa water-cement ratio (w/c) of 0.30. A non-chloride water-
(58 ksi) steel confirmed to the Canadian metric standard reducing agent, retarder, and high-range water-reducing
CSA-G40.20-M92. The test specimen shown in Fig. 2 admixture were used in the mixture. The coarse aggregate
represented the confined region of a joint in high-strength for this experiment consisted of a lightweight aggregate that
concrete structures. The dimensions of tested specimen were was imported from North Carolina. It consisted of rotary-
250 mm (10 in.) in length, 15d (tested bar diameter) in width, kiln-dried, high-quality, lightweight aggregates at 590 to
and 5d to 7d in thickness. The bond length is located at the 600 kg/m3 (996 to 1012 lb/yd3). The maximum size of this
middle of the specimen and the rest of the specimen is lightweight aggregate was 19 mm (3/4 in.) with a specific
debonded by the use of two small PVC pipes at the end of gravity of 1.45 and a dry density of 960 kg/m3 (1620 lb/yd3).
each specimen. The test setup, specimen size, number of The fine aggregates consisted of 810 kg/m3 (1376 lb/yd3) of
specimens, and testing parameters were similar to the bond fine sand. The target compressive strength ranged between
investigation conducted on normal strength concrete by 75 and 85 MPa (10,875 and 12,435 psi). The average
Eligehausen et al.2 compressive strength of all batches was 83.1 MPa (12,049 psi).
The concrete was batched over a 2-week period using the
CONCRETE MIXTURE DESIGN 0.1 m3 (0.13 yd3) capacity drum mixer. Three cylinders were
The proportions to produce 1.0 m3 (1.3 yd3) of HSLW taken from each batch. These cylinders, along with the test
concrete for the experimental portion of this research are as specimens, were covered in polyethylene and moist cured
follows. The cement content ranged between 480 and 490 kg/m3 for 28 days.
Fig. 6—Normalized bond stress-slip for 35 mm (No. 11) bar The bond behavior under cyclic loading for 25 and 35 mm
under tension at different loading rates. (No. 8 and No. 11) reinforcement is tabulated in Table 3 and
shown in Fig. 7 and 8, respectively. It can be seen from these
plots that the bond was not severely damaged during the first
affect on the maximum bond stress. The ascending portion of stage of cyclic loading, while increasing the displacement in
the curves for each loading rate is nearly identical followed by the second stage of loading causes a rapid deterioration of
a steep descending portion after the maximum bond stress was bond strength.
achieved. For the 35 mm (No. 11) specimen, however, there The bond behavior of HSLW concrete under cyclic
does appear to be a trend that is directly related to the rate of loading can be summarized as follows. During the initial
loading. The lower the loading rate, the steeper the ascending loading stages of ±3.75 mm/minute (0.148 in./minute), there
portion of the curve, but the maximum bond stress is less. was not a significant reduction in the maximum bond stress.
The first cycle at ±7.50 mm/minute (0.295 in./minute) saw a
Effect of cyclic loading significant reduction in the maximum bond stress. This was
The effect of cyclic loading on the reinforcement in due to the slip that takes place immediately after the bond
concrete structures is to gradually reduce the bond and to was broken and before the ribs of the reinforcement reseated
extend the yielding of the bar to within the development on the concrete. The amount of slip depends on the amount
length region. This effectively reduces the amount of devel- of microcracking and inelastic deformation in the vicinity of
opment length available to develop the yield strength of the the ribs. The bond stress continued to deteriorate more
bar resulting in pull-out of the reinforcement. One sample gradually due to frictional forces and aggregate interlock.
was tested under cyclic loading for each bar size subjected to Lastly, the bond stress decreased to a minimum as the effects
the three loading rates of 1.5 mm/minute (0.059 in./minute), of the frictional forces and aggregate interlock diminish.
75 mm/minute (0.0295 in./minute), and 0.15 mm/minute
(0.0059 in./minute). The loading history was displacement Bond comparison of HSLW and HSNW concrete
controlled with the first 10 cycles set at ±3.75 mm (0.148 in.) The load required to pull a reinforcing bar out of a concrete
and the remaining five cycles set at ±7.50 mm (0.295 in.). block will obviously increase as the length of a bar cast into
The first level of ±3.75 mm (0.148 in.) was set such that the the block increases. When the embedded length becomes
initial response of the bond strength could be studied without long enough, the bar will yield in tension before it pulls out
severe damage to the bond strength, while the second level of the block. The minimum embedded length required to
of ±7.50 mm (0.295 in.) was selected to be close to the develop the yield force of the bar is called the minimum
maximum slip associated with the maximum bond stress. required development length. The development length
requirements are used by all North American codes to indicate 6.314 MPa (776 to 915 psi). Similarly, for all cases of 35 mm
the bond strength of concrete. The ACI 318-0512 code expresses (No. 11) bar in HSLW concrete, the results indicate that the
the bond strength of concrete as a function of the square root mean normalized bond stress will be in the range 5.266 to
of concrete compressive strength. In most European design 5.868 MPa (763 to 850 psi), whereas the mean normalized
codes, the bond strength of concrete is proportional to the bond stress for 35 mm (No. 11) bar in HSNW10 concrete will
cubic root of the compressive strength. The test results of the be in the range of 4.182 to 4.832 MPa (606 to 700 psi). The
current investigation indicated that the bond strength of high- test results indicated that the use of lightweight aggregates
strength concrete, from 70 to 95 MPa (10,150 to 13,775 psi) is for high-strength concrete reduced the bond values by
more appropriately proportional to the cube root rather than approximately 6 to 10%. The current ACI 31812 code, however,
the square root as indicated by Alavi-Fard and Marzouk.10 reduces the bond by 30% for the use of lightweight aggregates.
Therefore, the test results were normalized with regard to the The results show a relatively steep ascending portion of
cube root of the compressive strength. the curve followed by a steep descending portion after the
The mean bond stress along with the standard deviation maximum bond stress level is attained. Albeit, the reduction
and a 95% confidence level for HSLW and HSNW for the in bond capacity in the descending portion of the graphs is
specimens under monotonic loading in tension and significantly more for HSLW than for HSNW. Alavi-Fard
compression were compared. The measured bond stress was and Marzouk10 reported a 30 to 40% decrease in the bond
normalized with respect to the cubic root of the compressive stress during the sharp descending portion of the curve for
strength and the results are tabulated in Table 4 and 5. For all HSNW, whereas the decrease in bond stress for the
cases of 25 mm (No. 8) bar in HSLW concrete, the results descending portion of the HSLW curve is closer to 50 to
indicate within a 95% confidence limit, the mean normalized 60%. This can be attributed to the fact that the lightweight
bond stress will be in the range of 5.18 to 5.773 MPa (751 to aggregates do not exhibit the same aggregate interlock
837 psi), whereas the mean normalized bond stress for 25 mm characteristics of normalweight aggregate. This translates
(No. 8) bar in HSNW10 concrete were in the range of 5.354 to into HSLW concrete being even more brittle and having
much less energy absorption capacity than HSNW concrete. tonic and cyclic testing for each bar size. In addition, under
Therefore, the load carrying capacity of HSLW concrete the monotonic test, the effect of loading in tension was
after the maximum load is reached is less than for HSNW. compared to compression, as well as varying the rate at
A comparison of the test where the rate of loading was which the load was applied. Similarly, the rate of loading
varied indicates that increasing or decreasing the loading rate was varied in the cyclic test to investigate the effect of
has minimal effect on the overall bond capacity of the bar. increasing the rate of loading.
This concurs with the work on HSNW concrete6,10 that The following conclusions can be summarized from these
indicated that changing the rate of loading had no significant tests of HSLW concrete:
effect on the overall bond capacity of the bar. 1. The test results revealed that HSLW concrete behaves
Analysis of the results for the cyclic test conclude that in a manner very similar to HSNW concrete;
cyclic loading does not have significant affect on the bond 2. The maximum bond stress for HSLW concrete is within
strength provided that the maximum cyclic displacement is 10% that of HSNW concrete for the 25 mm (No. 8) bar. The
less than the peak load slip in a static test. However, once the maximum bond stress for HSLW concrete is greater than for
displacement exceeds this peak load slip, then a serious rapid HSNW concrete for the 35 mm (No. 11) bar. The behavior of
deterioration of the bond capacity occurs. A similar conclusion HSLW concrete, however, is more brittle than NSLW concrete;
was reported in research on HSNW concrete. 3. The maximum slip value associated with bond failure
was approximately five times the slip value corresponding to
SUMMARY AND CONCLUSION the maximum bond stress. The shape of the stress-displacement
The experimental investigation consisted of testing 36 speci- curve for HSLW and HSNW concrete is very similar;
mens, each for a 25 and 35 mm (No. 8 and No. 11) bar. To 4. The bond slip curve begins with a sharp, nearly linear
facilitate a direct comparison with the previous work at the ascending portion of the curve, followed by a steep
same lab by Alavi-Fard and Marzouk6,10 on HSNW concrete descending portion indicating very brittle behavior. The
for which the same specimen size, casting position, and test decrease in bond stress for the descending portion of the
setup was used. The test program consisted of both mono- lightweight concrete curve, however, is 50 to 60% as compared