Pierre Rossi, International fiber reinforced concrete consultant, Paris

The use of fiber reinforced concrete in the tunnelling industry is a field of continuing research and development with valuable dialogue generated by sharing information among professionals at conferences. In that spirit, international expert in the field, Pierre Rossi, presents a critical analysis of the conference paper entitled The Use of Macro-synthetic FRS for Safe Underground Hard Rock Support by E.S. Bernard that was presented at the World Tunnel Congress WTC 2014 in Iguassu Falls, Brazil in May 2014, raising several points of theoretical inaccuracy and technical appreciation concern.

In the WTC 2014 conference paper entitled The use of macro-synthetic fiber-reinforced shotcrete (FRS) for safe underground hard rock support, the author summarizes the benefits of using macro-synthetic fiber-reinforced shotcrete for underground works carried out in rocky ground. These benefits are demonstrated with respect to certain concerns, such as performance at the beginning of the structure's lifetime, long-term safety, and long-term loss of ductility with respect to fragilization, corrosion, and creep mechanisms.

Steel fibers for concrete reinforcement

The paper contains theoretical inaccuracies and basic errors that limit its technical value and usefulness for practical purposes. It is best characterized as a commercial document, the purpose of which is to claim that synthetic fiber-reinforced concretes are superior to metal fiber-reinforced concretes. This critique describes, point by point, the issues raised by the author and addresses the errors committed in his treatment of them.

Performance early in the lifespan
The author suggests that fibers of any kind - synthetic or metal - do not offer much advantage with respect to structural performance of a shotcrete retaining wall or support system during the early stage of the lifespan since the performance is mainly related to the performance of the cement matrix. This is not contestable.

From this first considered statement, all further points in the paper are intended to confirm that synthetic macro-fibers are superior to metal fibers in treating each concern.

Polyfiber being dosed into the concrete mix

Long-term safety
According to the author, there are two main causes of long-term ruptures in underground supports: under design of the structure because of changes over time in the stress levels to the underground support wall or lining and corrosion of the metal reinforcements.

With respect to the rupture of the underground support related to overloading, the author suggests that "fiber-reinforced shotcrete can have great ductility in the interior of cracks when they are very open," which can be used to facilitate the redistribution of forces and therefore to warn of an imminent unstable rupture of the structure. The author adds that shotcrete reinforced with macro-synthetic fibers offers greater ductility (as he defines it) than shotcrete reinforced with metal fibers or traditional reinforcement.

In fact, the fiber-reinforced concretes of which he speaks generally display post-cracking behavior, decreasing in flexibility when the structural behavior is isostatic. This is because the percentages of fibers used are relatively low. With this type of fiber-reinforced concrete, the only way to ensure the stress-hardening behavior of the material, and therefore to permit redistribution of forces, is to have a hyperstatic mechanical system. This is the case when the underground support interacts with rocky ground with hard rocks. In this situation, the redistribution of forces is even more important because the fibers respond rapidly and effectively to cracks as they are created (i.e., cracks that are barely open). It is well known that synthetic fibers only become effective when the cracks are already relatively open because of their low Young's modulus. This is not the case for metal fibers, which respond to cracking much more quickly.

Therefore, for a hyperstatic mechanical system (such as an underground support that is in contact with hard ground), metal fibers allow for better distribution of forces within the retaining wall or support lining than do synthetic fibers, which leads to better structural ductility with the former.

The notion of ductility in the interior of cracks is nonsensical and demonstrates a lack of basic understanding of physics and mechanics.

Reduction of ductility with age

Steel fiber in a concrete mix

Steel fiber across a concrete crack

The author returns to the notion of ductility, but does not seem to have mastered the concept. He suggests that the difference in capacity to absorb energy in the interior of cracks that exists between synthetic fibers and metal fibers (to the advantage of the synthetic fibers) increases because of three problems: fragilization, corrosion, and creep.

Concerning fragilization, the author claims that because of too great an improvement in strength, over time, of the anchorage in the matrix of the hooked metal fiber, the fiber breaks when it is crossed by a crack instead of dissipating the energy by friction or by sliding in the matrix.

In fact, if we know steel fiber-reinforced concretes, we know that we should always choose a fiber geometry relative to the compactness of the matrix, as characterized by its compression resistance, in order to avoid rupture of the fiber when the matrix cracks.

Moreover, I assert that the steel fibers usually used in shotcrete, having a length of around 30mm, can be used in a matrix for which compressive strength can reach 90MPa with only a small percentage of them breaking.

The point that must be taken into consideration is that the more compact the matrix, the more brittle it is, and the higher the percentage of fibers needed to generate the same ductile behavior as the fiber-reinforced concrete. This is easy to manage when we know the development of the compression resistance of a concrete over time.

Steel-fiber and polyfiber shotcrete lining

Corrosion of metal fibers
Here we see the usual argument given by those promoting vast superiority of synthetic fibers over metal fibers. What the author says on this subject is confusing, at the very least.

To reiterate what must be understood on the subject, the cracking-durability relationship of fiber-reinforced concretes is only meaningful when the structure is in a service state situation. In addition, it is generally considered that, in service, crack openings must not exceed 100µm in a very aggressive environment and 200µm or 300µm in a less or only slightly aggressive environment. Consequently, the structure is sized so that, in service, these crack opening limits are never exceeded. As demonstrated in good scientific literature, it is necessary to know that:

•  When a crack opening does not exceed 300µm in a
    fiber-reinforced concrete, it presents a very tortuous
    and, at times, discontinuous path, which makes the
    circulation of aggressive agents difficult.

•  When a crack opening does not exceed 300µm,
    self-healing can occur in the matrix and, in the case
    of steel fibers, corrosion products can be deposited
    on the inside of the cracks. These two physical
    mechanisms plug up the cracks and therefore block
    the circulation of aggressive ions.

•  Because of the very small diameter of the fibers,
    their possible corrosion cannot cause damage to the
    matrix that surrounds them as the volume of the corrosive products would be so small as to create insignificant pressure on the porosity of the matrix.

It is evident that, in a case where poor design, calculation, or production of an underground support is implemented, and consequently the crack openings are abnormally large, the fibers inside of the cracks can corrode and thus lose their mechanical efficiency.

Creep fracture
To begin with, it is important to note that the author's treatment of the behavior of different concretes is not very precise. Therefore, a brief reminder is necessary: the creep of a material is characterized by an increase, over time, of its deformation under constant stress. Alternately, the relaxation of a material is characterized by a reduction, over time, of the stress that it undergoes under constant strain.

Corrosion in a steel-fiber reinforced shotcrete lining

The author indicates that, thanks to the existence of creep, the shotcrete retaining wall or support is capable of withstanding the development of ground deformations, over time, without cracking. I would suggest that, if this is the case, it is not due to the creep of the shotcrete but to its relaxation.

To consider the problem of the propagation of a crack in steel fiber-reinforced concrete subjected to stress that is held constant over time, this is a situation where one may speak legitimately of rupture by creep. Serious experimental studies on this subject have shown that, for the crack to propagate, it is necessary for the initial opening to be ≥0.5mm, and that the force, held constant over time, be ≥80% of the static force that created the crack. This is an extreme case since, generally, the constant force over time is never this high and the crack opening in service must not exceed 0.3mm.

The problem raised by the author is, in fact, quite different. He mentions cracking created during deformation of the ground (i.e. service cracking) which continues to develop because deformation of the ground also continues slowly over time. Therefore, we have a case of propagation of a crack under very slow static stress and not a problem of crack propagation under creep stress. This creates further confusion by the author. He suggests that, in this situation, the rupture occurs more rapidly than in a case of more rapid static stress. How can that be affirmed? I am not aware of studies published in scientific journals dealing with the influence of the rate of static stress on the cracking of steel fiber-reinforced concretes.

In fact, the problem cited by the author concerns the following situation: the service stress that the retaining wall support of the paper's case study example must overcome is poorly understood because this stress develops slowly over time. Consequently, we are confronted with the classic problem connected with the difficulty of defining dimensioning or design stresses, which is a problem that is generally resolved by using a safety factor. This is not a difficulty specific to steel fiber-reinforced concrete.

As for synthetic fibers, the fact that they are capable of resisting more force than steel fibers across very open cracks, which occur when the real stress is greater than the design stress, does not provide any greater safety if it is not within the framework of a mechanical hyperstatic system. In this mechanical hyperstatic system, steel fiber-reinforced concretes function better than the synthetic fiber-reinforced concretes for redistributing the forces, as we have mentioned previously, provided that the cement matrix is not too young for its adhesion to the steel fibers to be correct.

Corrosion of isolated steel fiber at an intersection with a crack

Finally, we note that the author forgets to mention that the creep of synthetic fibers (when the cracked underground support is truly subjected to creep stress) between the lips of the cracks leads to a significant increase in the opening of the cracks, and therefore the underground support can no longer ensure its sealing function.

Finally, it is a surprise to know of the author's virulent criticism concerning the studies by a large group of international experts (of which I was part) for incorporation into the 2010 Model Code for Concrete Structures by the Special Activity Group 5 of fib - fédération internationale du béton - Federation for Structural Concrete. Perhaps, however, with hindsight, we can consider ourselves lucky that a colleague has illustrated that there remains confusion and misunderstanding about the benefits and special requirements for the design and application of fiber-reinforced concretes.

Response to macro-synthetic fibre critique - TunnelTalk Feedback, Nov 2014
TunnelTECH: Fibre-reinforced concrete for precast segments - TunnelTalk, May 2014
20 years of fibre concrete linings in the UK - TunnelTalk, Oct 2010
TunnelTECH: PP fibres to resist fire-induced concrete spalling - TunnelTalk, Nov 2010
fib Model Code for Concrete Structures 2010 - TunnelTalk, Books & Reports