Experimental study on fiber-reinforced concrete panels subjected to shear , Conclusions - Tesi

The goal of this experimental work was to study the behavior of SFRC and MSNFRC

panels subjected to pure shear, and to perform an investigation of compressive, tensile

and flexural behavior of the same concrete specimens, comparing one to each other with

referring to previous experimental programs.

An experimental program was undertaken at University of Toronto involving

experimental tests on small-scale specimens (compression cylinder tests, uniaxial dog-

bones direction tests and modulus of rupture bending tests), as well as on larger scale

in-plane shear panel specimens.

For the analytical part of this work, the primary goal was to understand the behavior of

the macro-synthetic fibers and to find a right way to model the MSNFRC specimens.

Many theories were studied to improve this behavior , with simplified relationship and

to adopt reasonable bond constitutive laws.

Lastly, SFRC and MSNFRC were implemented into the finite element program,

VecTor2, and a short verification has been done to investigate the accuracy of the

predictions. 269

8. CONCLUSIONS

7.2 Material tests

7.2.1 Cylinder Compression Tests

As shown and previously discussed, the pre-peak behavior was not affected by the

addition of fibers.

Adding fibers into the concrete matrix exhibited an improving in the post-peak behavior

for the all FRC specimen than the plain concrete specimen: for plain concrete the

compression resistance drop suddenly after the peak, whereas the FRC post-peak

behavior has a softer and more controlled curve so, adding fibers into the concrete

matrix, led to improved ductility and toughness of the concrete.

DC-P2 specimens, made of steel fibers, shows a better post-peak behavior than DC-P3,

made of macro-synthetic fibers that, for the same strain level, has a lower stress level.

Some difficulties were obtained in the workability due to accommodate the fibers into

the concrete matrix: the space occupied by fibers is removed for the concrete, and the

resistance pays the consequences with the creation of larger voids, reducing the overall

strength.

7.2.2 Uniaxial Direct Tension Tests

When 0.5% fiber by volume was added to the matrix, although a strength reduction was

observed immediately after cracking, the concrete still retained a small amount of

residual strength (Susetyo, 2009). Increasing fibers up to 1.0% and 1.5% did not alter

the pre-cracking response of the concrete than by 0.5% of fiber by volume, but the post-

cracking behavior was improved remarkably.

The influence of fiber type is viewable by the main differences obtained: shorter steel

fibers were more effective than the longer fibers, at volume fraction of 1.0%, maybe due

to the fact that there is a large number of individual fibers in the mix for shorter steel

fibers, at a given volume fraction.

At large crack width, the SFRC with shorter fibers began to lose the load-carrying

capacity rapidly, and the residual stress dropped below that of the longer fibers. 270

8. CONCLUSIONS

The response of MSNFRC specimens was different: the drop after cracking was larger

than SFRC regardless to the volume fraction; moreover, a large crack width was

required before the macro-synthetic fibers began engaged.

However, despite to this initial drop, the MSNFRC response regained some strength: in

most cases the maximum residual tensile stress occurred at a much greater crack width

than with the SFRC specimen, as much as 150% of the stress at engagement.

At a crack width of 2.4 mm to 2.8 mm, the steel fibers began to lose bond strength due

to the straightening of the end-hook; the macro-synthetic fibers performed more

favorably at this level of cracking (Carnovale, 2013).

The flexibility of the fiber is a significant properties that affect the engagement of the

macro-synthetic fibers. At first cracking, some fibers were oriented in non-orthogonal

directions to the crack. These fibers had to become bent around the matrix entrance

points at both sides of the crack and become aligned with the direction of the load,

before becoming effective: this does not happen instantly and some crack opening is

required to allow this alignment to occur. This explain the requirement of a relatively

large crack opening before that these fibers become engaged.

Therefore, at a small crack widths, only fibers perfectly aligned perpendicular to the

crack can transmit significant tensile stress across the crack.

Watching the post-racking behavior of the specimens, they seem to don’t be

significantly influenced by the strength of the concrete matrix: the only difference is the

interfacial shear strength between the fibers and the concrete matrix (higher in high

strength concrete matrix than in the normal one) so at the onset of first cracking, the

fibers in high strength concrete specimens were subjected to a higher tensile stress than

those in normal strength concrete specimens.

7.2.3 Modulus of Rupture Tests

It is not viewable in this research, but Carnovale found that the short steel fibers were

more effective in residual load-carrying capacity, because more fibers were still present

to transmit load across the crack: with as little as 1.0% by volume of the short fibers

experienced elevated amounts of strain hardening, and attained the greatest peak load at

all the mid-span displacement. 271

8. CONCLUSIONS

Longer fibers exhibited more ductility.

For the macro-synthetic fibers responses, at low crack widths these fibers did not

become sufficiently engaged: this can be attributable to their low stiffness; as dog-

bones, at a high mid-span displacement of over 4 mm, specimen containing MSNFRC

showed the greatest residual flexural load-carrying capacity.

It was evident that the macro-synthetic fibers, despite a more sudden drop in load after

cracking, provided significant improvements in residual load-carrying capacity,

toughness and ductility over plain concrete; this improvement was similar to that of

1.0% by volume of end-hooked steel fibers with the same length, consistent with the

experimental findings of Richardson et al. (2010).

7.2.4 Panel Tests

Monotonic in-plane shear panel tests showed that shear strength similar to that of low

percentages of conventional transverse reinforcement ( = 0.42%) could be attained

using 1.0% by volume of steel fiber reinforcement; the same results could not be

obtained with macro-synthetic fibers in 2.0% by volume.

The pre-peak compressive behavior was somewhere affected by the presence of fibers:

28

modulus of elasticity and -day compressive strengths exhibited some differences,

with a lower value of elastic modulus and a worse strength in the compressive peak;

shear strength attained by the MSNFRC panels was only the 67% of the shear strength

for low percentage of conventional steel, while the 1.0% by volume of SFRC panels

obtained at least the 90% of the shear strength for low percentage of conventional steel.

The post-peak compressive behavior of concrete was improved: strain at peak stress was

increased for all FRC specimens, with greater results for ductility and toughness in

compression; short steel fibers exhibited the best behavior and the greatest

improvements in toughness.

After cracking, all FRC specimens exhibited a gradual and ductile release of load as the

fiber were pulled out or ruptured; 272

8. CONCLUSIONS

Few of macro-synthetic fibers ruptured, meaning that a higher ultimate strength of these

fibers would be needed if stiffness and anchorage are improved: this would help to

preserve the ductility exhibited by MSNFRC specimens;

MSNFRC panel exhibited a better ductility than the conventionally reinforced concrete

and SFRC.

The average and maximum crack widths were greater in the MSNFRC panels than in

the SFRC one: crack spacing were also larger, meaning that the degree of multiple

cracking was greater for the end-hooked steel fibers.

Benefits in ductility using the MSN fibers cannot be denied, yet the greater bond

strength would be beneficial in improving shear strength and promoting distributed

cracking. To maintain the ductility of the response, this improved bond strength must be

coupled with higher fiber tensile strength to prevent brittle fiber failure.

7.3 Analytical Modeling

The VEM and DEM were shown to reasonably predict the behavior of SFRC

specimens. Thus, they have been adapted to also model the MSNFRC tensile response.

Despite the observation of a few fiber ruptures during the MSNFRC dob-bone tests,

elastic fiber deformation was ignored: it was postulated that these ruptures occurred in

fibers that were aligned in the loading direction immediately upon cracking; these fibers

carried too much tensile loads before other fibers became aligned and effective.

Watching the literature, it was found that the MSN fibers could be modeled as the SFR

one: frictional bond component and mechanical anchorage component could be

separately represented.

For the slip at peak mechanical anchorage strength, 0.5 mm was chosen as experiments

showed that peak fiber bond strength for macro-synthetic fibers occurred at larger

cracks widths than end-hooked steel fibers.

Using this new model, the slope of the degradation in direct tension after cracking was

over-predicted. An attenuation factor was well used, considering the effect of fiber

engagement on the energy released during cracking. 273

8. CONCLUSIONS

7.4 Finite Element Modeling

Panel specimens were tested, after the series of SFRC beams and MSNFRC beams, to

verify the reliability of the FE Program, VecTor2, on these kinds of fibers.

Predictions of SFRC were well done, with a good range of acceptability; MSNFRC

specimens predictions were quite well done, at least as successful as those obtained for

SFRC or plain concrete. However, a number of discrepancies exist, mostly surrounding

the predicted load-carrying capacity and the degree of strain hardening immediately

after cracking.

7.5 Conclusions

This experimental program, the analytical modeling and the FE Modeling, aboard with

the literature, shows the following conclusions:

1. Steel Fiber addition in the concrete matrix gives many improvements on the

behavior of the all material (post-cracked residual strength, tensile ductility and

control crack widths); the same results were obtained with the Macro-synthetic

fibers;

2. The panels tested with the 2.0% of MSN fibers shown a strain hardening

behavior and multiple cracking: it is worth notice that nothing can be said about

less percentage (2.0%, in some states, can be considered a too high fiber

percentage);

3. Moreover, 2.0% by volume is a very high quantity of fibers, leading in problems

with the workability of the concrete and in the certain correct distribution of the

fibers in the all cement paste; shorter fiber lengths may improve the distribution;

4. Comparing steel fibers with macro-synthetic, the first has the best behavior in

terms of peak strength; the ductility was quite the same, but the macro-synthetic

shows a better post-peak decay in terms of ductility and toughness (these fibers

can transmit relatively high amou