Conclusions and Recommendations

The objective of this study was to collate and condense a significant body of the author=s underground observations on the performance of shotcrete in response to mining. These observations of shotcrete performance were then compared with modes of shotcrete failure and support function identified in the literature review in Chapter 2. From this review and comparison, it was determined that not all observations of shotcrete failure and success could be attributed to the standard radial suppression models applied to shotcrete support design. It was postulated that one mechanism of shotcrete support, for which there was little reported information published, was the ability of the stiff reinforced shotcrete liner to suppress dilation of existing fractures and defects around the excavation periphery, thereby forcing failure to occur through intact rock asperities. In order to demonstrate this mechanism, some simple 2-D block modelling was done using Universal Distinct Element Code (UDEC). While the model analysis was not conclusive, the results of the analysis did not contradict the observations of shotcrete support performance in hard jointed rock. The following section presents conclusions from this study and suggests opportunities for further work in this area of study.

Shotcrete performance observations were categorised by the support function (retain, hold, reinforce), mode of failure (structural or stress related) and key parameter or attribute of shotcrete which contributed to the support function. These were summarised in Table 2.1 (repeated below as Table 1). Methods of designing shotcrete support systems based upon the traditional failure modes of shotcrete are effective means of determining the initial design parameters, however, they do not necessarily account for the long-term effectiveness of the shotcrete liner. In all of the cases presented,

for both empirical and analytical designs, the basic failure mode assumed is as a direct result of radial movement of the tunnel opening and the basic support mechanism is to resist this radial movement. While individual mechanisms or modes of failure are described, it is clear from work of others that shotcrete failure does not occur through one simple mechanism, but rather through a combination of mechanisms, and that failure due to radial deformation does not explain all observations of shotcrete support performance in response to mining. Hoek (1998) described the design of shotcrete support for underground excavations as being a very imprecise process and that one common observation made by operating personnel and engineers is that shotcrete almost always performs better than expected. He further suggested that many examples of the use of shotcrete as a last act of desperation in an effort to stabilise the failing rock have worked, but little documentation exists of these examples. Shotcrete performs through a complex interaction between the failing rock mass and a layer of shotcrete of varying thickness and varying material properties as it hardens, then fails. Such a complex interaction prevents the development of one single theory of how shotcrete works. As such, most design techniques are based upon empirical methods, rules of thumb and sound engineering judgement. The way shotcrete works in highly jointed rock in low or high stress environments is through some combination of mechanisms.

Chapters 3 and 4 presented detailed observations of shotcrete performance in response to mining, under a variety of conditions at INCO Ltd=s Creighton, Frood and Stobie Mines. Rock mass parameters, joint condition, stress state and the character of the ore body all contribute to the varied conditions under which shotcrete performance in response to mining is measured. These factors include: high stress and bursting conditions, stress induced failure and subsequent low stress conditions, geometry resulting from sequencing restrictions due to high stress, blast vibration and corrosive waters due to filling practise and man made effects of equipment damage to primary support. All of these factors contribute to the extreme conditions encountered in mining at depth. The introduction of shotcrete into this environment has provided significant benefits from a ground stability and safety perspective. Initial shotcrete trials in both high stress and low stress conditions, in jointed rock and back fill, demonstrated the benefits of proceeding with the development of further applications of shotcrete support systems for underground mining at depth. As suggested in Chapter 2, the development of shotcrete systems and procedures is an evolutionary, rather than a revolutionary, process. There is a tremendous amount of information contained in the observations reported in this thesis and some key issues have been identified and addressed. The information and observations presented and the resultant conclusions can be divided into three general categories, conclusions related to high ground stress conditions, low ground stress conditions and operational issues and benefits

High Stress or Dynamic Loading Conditions

• Mesh reinforced shotcrete can survive and absorb large peak particle velocities, in the order of 75 to 237 inches per second (1.6 to 6 metres per second) due to VRM blasting less than 10 feet (3 metres) away, and still remain an effective support system. This included repeated cyclic loadings where the rock mass had already undergone substantial damage and dilation prior to the application of the shotcrete, indicating that the shotcrete was able to absorb more energy than a simple rock bolt and mesh support system, by preventing further dilation of an already failed rock mass.
• Even when cracked, mesh reinforced shotcrete can survive high peak particle velocities (burst or blast).
• Mesh reinforced shotcrete provides effective support and reinforcement in dilated and bulked rock masses and absorbs significant amounts of energy in the process, resulting in a ductile support system which can withstand further seismic loading.
• Several cases presented show evidence that rock burst damage stopped or was contained by the mesh reinforced shotcrete support system. This suggests that shotcrete must be reinforced with mesh or steel reinforcement of some type to withstand dynamic loading and that bolt and mesh support, augmented with shotcrete, is a superior support system under rock bursting conditions, relative to bolts and mesh or shotcrete alone. These observations do not include any cases where steel fibre reinforced shotcrete has been subject to rock burst energy release.
• Steel reinforcement is necessary for shotcrete to function effectively in dynamic loading or large displacement, static loading conditions.

• Shotcrete does not perform as well where large closures occur and displacement incompatibility occurs between shotcrete panels (eg: at the shoulders). This results in compressive failure at the shoulder and the possibility of toeing out of the lower wall and loss of contact with the rock mass or back fill. This can result in drummy slabs of shotcrete, which can still provide support functionality so long as the rock or sand behind is not allowed to unravel.
• Failure of shotcrete at the mesh interface does not result in loss of overall integrity of the support system. The mesh will remain an effective containment for the remaining rock and shotcrete as long as the individual wires of mesh do not suffer a sudden and extensive failure. Continuity of the mesh reinforcing is key to the post failure capacity of mesh reinforced shotcrete.
• Rock (or fill) material with very low tensile strength may allow shotcrete to fall away from the surface, despite good adhesion between the shotcrete and rock.
• Unravelling from behind shotcrete will negate its effectiveness as a support/reinforcement system.
• Mesh reinforced shotcrete can sustain significant deformation before failure occurs leaving ample time for reconditioning to be done (ie: failure will not occur in a sudden manner).

• Use of shotcrete as a replacement for timber square set support to prevent unravelling of backfill from the shoulders and backs of topsills has eliminated the need to handle timber in the mining cycle and improved safety of the operations.
• At Frood Mine, initial trials demonstrated that mesh reinforced shotcrete was a more effective support system than the timber square set support traditionally used to mine through old workings where an extremely broken rock mass was encountered.
• Mining under these conditions demonstrated the importance of preventing unravelling of the rock mass or sand fill from occurring behind the shotcrete, to maintain stability of the excavation.
• Testing shotcrete as primary support under these conditions enabled observations to be made on shotcrete performance when the limits of its capacity were reached.
• Slabbing of the shotcrete and spalling at the mesh interface occurred largely because there was too much shotcrete applied. A thinner shotcrete layer allows failure to occur through smaller, more frequent cracks in the liner, allowing shotcrete to behave in a more ductile manner. In a mining environment, the intent is not to stop movement from occurring as much as it is to manage and control the movement long enough to extract the resource.
• Use of shotcrete in extraction draw point brows has eliminated the requirement for cable bolting as secondary support.
• Use of shotcrete in secondary support of VRM brows allows for more effective backfill barricade placement and sealing barricades more effectively and closer to the brows.
• Mesh reinforced shotcrete has proven to be an effective protection against damage from mining operations, mobile equipment and secondary blasting.
• Mesh reinforced shotcrete provides substantial protection against blasting damage in VRM topsills, brows and blasting chambers and protection against re-muck and equipment damage tearing the weld wire mesh and bolts off the walls and allowing the loose fracture zone to fall away.
• Shotcrete as a replacement for timber square set support, maintains the integrity of topsill drill drifts through sandfill and allows the use of topsills for both VRM drilling and blasting below and rib pillar extraction above. This has eliminated the need to redevelop for extraction and recovery of the rib pillar ore above, a common occurrence in square set timbered topsills.
• Shotcreted headings were much easier to recondition when failures did occur. They simply required a mesh Apatch@ with bolts on either side, followed by an additional thin layer of shotcrete.
• Shotcrete also made reconditioning burst damage a lot easier and safer as the initial stabilising shotcrete layer, applied to unstable rock mass, stabilised conditions to allow safe installation of bolts. Areas were normally re-shotcreted for longer term support protection and stability.
• Shotcrete served as protection from corrosion by prolonging the life expectancy of existing support elements such as mesh and friction stabiliser bolts.

The shotcrete trials at Stobie Mine demonstrated the need for steel reinforcing in the shotcrete, in this case #6 gauge weld wire mesh, to withstand the effects of blasting and rock mass dilation associated with mining. Slabbing and Adrummy@ shotcrete were indications of movement along structures, as opposed to blast damage due to blast vibration and peak particle velocity. In most cases, it was the fractured nature of the rock mass which gave the appearance of effective adhesion loss. Even though the shotcrete sounded Adrummy@ and there were obvious tension cracks bounding the Adrummy@ sections, the shotcrete was still able to provide a support and reinforcement function.

The fact that, despite the existence of tension cracking and rock fracturing immediately behind the shotcrete/rock interface, the shotcrete provided functional support at the brow and in the walls of an uppers blasthole mining block demonstrated that indeed, replacement of bolts (the holding element) is possible and effective in this environment. This trial demonstrated the necessity of having some form of steel reinforcement properly placed in the shotcrete to provide effective support in areas of moderate to large deformation and dynamic loads. The nature of this trial, being labour intensive, meant that the up-front cost far exceeded the cost of conventional bolting and screening. That is the single most significant drawback of the two-pass shotcrete system for primary support. To that end, testing of boltless shotcrete for primary support in the sub level cave mining block continued.

While determining that support performance capabilities and re-entry times for primary support applications are important components of research into new ground support systems, determining suitable applications is not. What is intended by that statement is not to say research trials don=t have a role to play in determining suitable applications, but rather that oftentimes, it is the operations department driving the need for finding a suitable application and it is the technical personnel who must strive to Akeep up@. That was more often than not the case in the further development of steel fibre reinforced shotcrete (SFRS) for primary support in sub-level cave development and production. The development of a performance based observational approach to testing and validating shotcrete as an integral component in the support cycle has been a long process. The experience in these simple trials and observations presented here on how shotcrete works are key to the development of a workable and efficient system.

Shotcrete is part of a system, all elements of which must perform as designed in order to get optimum system performance and benefit. To be effective, shotcrete must work with existing support as an integrated system and, while shotcrete can compensate for minor failure of the other support elements, when significant load is imparted upon the support system, failure can still occur. The observations presented in these chapters clearly indicate that shotcrete does more than simply suppress radial movement to retain the rock mass in place. In the examples presented, attributes and properties of the shotcrete, including the strength in compression and shear, adhesion of the shotcrete in a mechanical sense, rather than an adhesive or gluing sense, the ability to distribute point loads due to rock movement, continuity of the mesh reinforcement within the shotcrete layer, the post failure toughness and the ability to withstand additional seismic loading and absorb seismic energy in an already damaged rock mass suggests that the function of shotcrete in this case is as much as a reinforcing element as it is a retaining and holding element. Observations of the failure modes and mechanisms and their relation to shotcrete support function are summarised in Table 2.

Modes of shotcrete failure observed in underground mining, under both static or squeezing loading and dynamic or rock burst/blast vibration loading indicate that resisting radial displacement is not the sole function of the shotcrete liner. The failure mechanism which is postulated in this thesis is one of suppressing joint dilation by the tangential stiffness of shotcrete in resisting shear or tangential movement along the rock/shotcrete interface and helping the rock to support and reinforce itself, requiring failure to break fresh rock (asperities) in order to fail. A small amount of confinement results in significant amount of shear force required to cause this failure, thus the small amount of confinement offered by shotcrete provides significant support pressure by preventing dilation from occurring.

The simple numerical analyses that were conducted to demonstrate the mechanism were not intended to model reality. Models for two conjugate joint sets of 75 degrees were run initially to test the rate of loading and the effect of joint cohesion on results. It was concluded that:

• the response of the model depends upon the prescribed displacement and that for better results, small prescribed displacements are recommended for loading the block.
• the need to have a solid understanding of the effect of confinement as it relates to shotcrete and joint response in both an underground excavation and a surface excavation and how these two different geometries have different effects on the dilation of rough joints and thus on the shear strength of joints and thus the influence of shotcrete on joint strength is very important.

Two scales of roughness are key to the performance of shotcrete in reinforcing the rock mass. The first is the small scale, or joint roughness, and the second is the larger scale of roughness profile of a drill and blast excavation. The profile roughness results in a very tight and tough mechanical interlock between the rock and the shotcrete, which generates an effective adhesion of the shotcrete to the rock, particularly when coupled with rock bolt or tendon support and reinforced with mesh. The roughness of the excavation profile comes into play along with the adhesion of the shotcrete to the rock mass to generate an effective and integrated support system.

The numerical modelling analyses showed that shotcrete placement prevents shear displacement from occurring on the rock joints, resulting in an increase in the strength of the rock mass. While the results show some interesting relationships between the loading conditions, shotcrete properties and dilation angle, the assumptions that were made in order to model the mechanism of dilation suppression do not suggest that this model would be valid for design purposes. The results simply show that there is a relationship between dilation and the ability to prevent dilation from occurring with a thin liner and the load-displacement result. It was not the intent to show definitively and rigorously that this was a valid modelling approach for design purposes. While the results of the model analyses are suspect because of the assumptions that were made about the strength of the shotcrete and the stiffness of the shotcrete rock interface, the important point about the entire analysis is that the results do not contradict the observations of shotcrete performance made in Chapters 3 and 4.

Figure 5.9 summarises how dilation suppression is key to rock mass reinforcement. Mechanism 1 shows that preserving the dilation angle, i, in the relationship (i + f) as being important. This means simply preventing joints from moving and shearing off the asperities which are defined by the dilation angle, i. In some cases, it also means preventing the joints from >jumping= due to dynamic loading. Shotcrete does this in a fractured rock mass by holding the rock mass in place and preventing dilation from occurring. Mechanism 2 simply shows how increasing confinement occurs due to ride-up on undulations in the joints themselves. If no dilation is allowed, then confinement increases and reinforces the rock mass further. The result, if dilation is not allowed to occur, is a net increase in the frictional strength of joints and defects in the rock mass immediately behind the shotcrete liner.

Summary and Further Work

The numerical modelling analysis which was conducted was fairly rudimentary and intended to simply show a relative effect of shotcrete application, represented by material elements in the model, on the load capacity of a rock block through suppression of dilation. Input parameters were selected to allow for stable model conditions and to achieve a solution and do not represent true field values. Further work to define this relationship, particularly with advances in numerical modelling codes in recent years, or alternatively physical modelling, might provide additional insight into the mechanism introduced in Chapter 5.

Hoek (1998) summarised and compiled observations of shotcrete support design, based upon the contributions of numerous authors and upon his own experience (Table 2.5), however, he qualified the use of the observations in the table by saying they A...can only be used as an approximate guide when deciding upon the type and thickness of shotcrete to be applied in a specific application. Modifications will almost certainly be required to deal with local variations in rock conditions and shotcrete quality.@ It became clear to this author, from observation of shotcrete performance in bursting conditions, that shotcrete, and in particular mesh reinforced shotcrete, could play a greater role in rock burst resistant support design, particularly for the moderate to strong rock burst events which occur in mines of the Canadian Shield. These guidelines are repeated in Table 3 of this chapter, with the suggested addition. The final item on the list provides two benefits which are not fully addressed, based upon the author=s observations and experience. First, it accounts for moderate rock bursting in high stress conditions where bulking of the rock mass is commonly subjected to remote large magnitude seismic events and second, it allows for the application of shotcrete over existing ground support, thereby adding value to the support system rather than reinstalling a new support system, such as lacing over an existing system. This would be an intermediate step, before the application of lacing which is often much more labour intensive to install.

An understanding of >all= functions and attributes of shotcrete that allow it to perform well under a range of extreme conditions will benefit the engineer in designing more appropriate support for safer and more economic mining. Better understanding of this theory of shotcrete support function may also provide the mining engineer an additional tool in designing shotcrete support systems more effectively. This can only be accomplished by observation and analysis of the support system performance in order to develop a better understanding of the mechanisms at work and modifying designs based upon the observed performance of the shotcrete system. Numerical tools are an important part of the design process, but can be grossly misleading if used to the exclusion of sound engineering judgement.

Table 3 Summary of recommended shotcrete applications in underground mining, for different rock mass conditions with suggested additional category (after Hoek, 1998)