ROCKMASS CLASSIFICATION VS CHARACTERISATION. BEYOND NUMERICAL SINGULARITY.

Geotechnical Engineering in mining straddles the fine line between the disciplines of Engineering, Science and Art. In glossy terms (AI explanations) we strategically assign values to localized rock characteristics, carefully blending them through empirical formulas to create a nuanced representation that captures the intricate nature of heterogeneous rock masses. In more cynical terms, we assign arbitrary numbers to localised rock characteristics and then combine them through empirical formulas to provide the illusion that the final number represents a heterogeneous rockmass.

Increasingly, I am seeing a focus on the singularity of “the number”. GSI tables are popping up all over my LI feed. Geotechnical block models are considered necessary and numerical models are being presented as fact.

In reality, the empirical systems we use are one number to describe the rockmass at the point of determination. However, if you are only relying on one number you cannot encapsulate the true geotechnical risks posed to your site.

Both the Q system and the Rock mass rating (RMR) are classification systems. They were developed to try and encapsulate the heterogenous nature of the rockmass. GSI is a visual representation of these numbers. GSI, however is used as an input into many numerical models. Just collecting any one of these numbers without the supporting characteristics does not capture the nuances of the rockmass and does not assist with a true understanding of the rockmass. The following article examines the various parameters that should be considered in a rock mass evaluation and how they relate to values within the classification formulas. It also aims to provides a brief generalised introduction to how these characteristics may be used to determine geotechnical risks and assist in the overall comprehension of the rockmass. The order of the presentation of the parameters below is essentially irrelevant. Each parameter is important in its own right and during an initial evaluation, all parameters should be considered. The initial evaluation may allow some parameters to be ignored (such as stress in a moderately shallow deposit), however, ignoring parameters that are assumed not to be relevant can lead to significant failures.

Lithology, weathering and major structures

Geological factors such as lithology, weathering and major structures are not encapsulated in any of the classification systems. However, they are crucial to understanding your rockmass prior to considering ay classification system. A layered sedimentary rock will behave differently to a massive unaltered granite, which will behave differently to a metamorphosed ultramafic. Lithology should form the basis of all initial geotechnical analysis. Alteration is often overlooked during analysis and is difficult to consider due to inconsistencies in identification. On one site I worked at we had areas where 50m narrow stopes were possible with no issue, however, development in other areas could not proceed more than 10m without significant reinforcement. The difference between the 2 areas was the “colour” of the rock. The good area was black, and the poor area was white. However, when samples were brought to the surface, they looked the same. It was a casual mention from a geologist that suggested that the colour difference was due to the presence or lack thereof talc alteration.

Consideration of weathering may seem logical, but I have seen several examples where the weathering profile has been ignored in the early underground project stages and stope optimisers have placed open stopes into the weathered zone. In open pits, analysis of the weathered zones will be significantly different from those in the hard rock zones. From a technical perspective the test work for weathered materials is significantly different from competent rock with consideration of soil parameters such as particle size distribution, potential for swelling and liquid limits all required amongst others. The type of test work required will depend on the type of soil. Generally, the weathered zones will require flatter slopes, more consideration of the influence of water and there is a high risk of circular failure. Highly decomposed regolith generally cannot be drill and blasted so must be free dug so higher mining rates can be allocated. The transition zone from weathered to fresh rock should not be ignored. Operationally this is one of the most challenging zones to mine. Technically, it is one of the most difficult zones to characterise. Rock mass strengths can vary significantly in short distances, floating boulders can occur in regolith that must be drill and blasted and relic structures often create significant stability risks.

Major structures can consist of faults, shears and intrusions and are often defined from a geological perspective with consideration to how they influence the overall formation and / or interpretation of the geological deposit. Whether major structures influence the geotechnical properties of the rockmass will depend on the nature of the structure. Major features that are well healed and / or have no significant infill can have a negligible influence on the rockmass but may locally influence discontinuity orientations. Structures with significant gouge infill or with significantly broken contacts must be modelled to ensure they are considered in the mining method, infrastructure positioning or pit slope positioning. At times these features can be very thin and may not be considered significant from a geological perspective. To be called a major structure it is assumed that these features are anomalies in the general rockmass and have significant deposit (or domain) scale continuity. It must be noted that even if a geological feature does not appear to influence rockmass conditions locally, it may influence the rockmass at another location. Generally, rather than discarding models that appear to have no relevance, a ranking system of importance should be used. Whilst direct consideration of the less influential structures during mine planning may not be required, in each location, the conditions should be confirmed to ensure that assumptions have not changes.

Another area that is of significant risk is where large-scale features interact. By their nature, large scale features ?have been “active” through various geological phases. As such, where they interact can be some of the most problematic areas within a mine.

Intact rock strength

A quick google search suggests that there are a number of definitions of rockmass. I define a rockmass as a rock material inclusive of all its inherent flaws. This may be discontinuities, voids, pores are any other feature that is likely to change the geomechanical properties of the rock. Intact rock strength testing aims to determine the properties of the rock material separate from its inherent flaws. Intact rock strength is considered as Factor A1 in the RMR system, however it is not considered in the Q system. Arguably, this is a flaw in the Q system and strength should be considered regardless of whether it is an input or not. ?Most numerical modelling programs consider rockmass strength with conversions using Mohr Coulomb or Hoek Brown equations. I will not delve into any of the specific uses of rock testing results in modelling. Rather I will focus on how to use intact rock strength parameters to consider geotechnical risks on any particular site.

In broad terms the uniaxial compressive strength (UCS) along with the elastic properties are used to define the amount of load the rock is capable of bearing. It is important to consider what is influencing lower UCS values (generally <100Mpa). As mentioned in the previous section, alteration and weathering can have a significant impact. The mode of failure of a sample is critical. Any sample that fails in shear along a structure is not providing intact rock strength as a result and should be discarded. However, was the shear failure the result of poor sample selection or finely spaced layering (bedding or foliation) of the rockmass. Understanding what is influencing the failure of the material will assist in identifying the geotechnical risks on a site and providing guidance on the pathways for further analysis.

Minor discontinuities and their properties

Discontinuities are the main cause of weaknesses within the rockmass. For ease of reading, I will generically refer to these as joint. Joint sets are groups of structures with like orientations. Typically, the more joint sets, the more fractured the rockmass. This is why Jn (joint set number) is considered within the Q system. RMR does not directly consider the number of joint sets within the rockmass, however, arguably, RQD is an indirect measure.

One challenge with joint set identification from core logging is the quality of the orientation marks. Modern methods rely directly upon the drilling offsider placing the mark in the correct place. Minor discrepancies can have a significant effect on the converted orientation and increasingly I am observing very scatter results with low concentrations within clusters. It is imperative that drillers and offsiders are trained as to the importance of orientation marks and some measure or orientation mark quality is established.

The number of sets within the rockmass is useful, however it is not the only parameter that is critical when considering discontinuities. The orientation, spacing, cohesion and shear strength of each joint set all influence the potential for instability.

Orientation

The orientation of the joint sets with respect to each other and the excavation is critical to stability. Conjugate sets at 45-60 degrees to each other are common in many geological environments. It only requires another random joint or joint set to form wedges. The probability of these wedges forming depends on the spacing of the joints and excavation parameters. The likelihood of failure will depend on the friction and cohesion of the joint (these parameters will be discussed below). The risk of wedge failures is typically determined separately from rockmass classification systems using kinematic analysis. It is crucial to have sufficient data to determine the joint set orientations across the deposit. Different lithologies or geotechnical domains may have differing joint set orientations and hence the risk of wedge failure will vary across the deposit.

The orientation of the joint sets with respect to the excavation is considered with the RMR system (Factor B) and in the Matthews method for open stoping. In open pits, joints subparallel to the pit walls can present a significant stability risk and designs should consider how they interact with the joint set should this situation occur. In underground environments, mining parallel to bedding or foliation presents many difficulties, especially with respect to drill and blast. Hole accuracy can be difficult to achieve in layered environments resulting in overbreak which further exacerbates the stability issues encountered. If excavations are oriented parallel to the foliation and is combined with a high stress and strain environment, buckling of the walls is often encounter and excavation closure can present a significant challenge.

Spacing

Spacing within the rockmass classification systems are accounted for in several ways. RMR considers joint spacing directly (Factor A3), whereas the Q system does not. Again, arguably RQD is an indirect measure of the joint spacing.

When considering joint spacing, closely spaced joint sets increase the risk of rockmass unravelling. However, largely spaced joints can be a benefit and a hinderance. If intact rock material between the joints is highly competent, few stability issues can be expected. However, when they are spaced a significant distance apart, it can be difficult to identify where they will coalesce to form wedges or blocks of failure. Whilst the risk of wedge failure may be considered low, the consequences of a large wedge forming in either an open pit or underground can be significant. Large wedges are difficult to contain using ground support as the weight is often beyond the practical limits of bolt density and capacity. The location of large wedges in widely spaced joint environments is difficult to determine and is often managed operationally. Newer software analysis programs incorporating discrete fracture networks (DFNs) may be able to overcome this issue.

Joint friction and cohesion

Joint friction and cohesion determine whether the rockmass failure will occur should the joint systems combine with the excavation in a preferential way. It is a direct input into wedge analysis programs and is often considered in other modelling software that considers jointing. It is preferable that these parameters are determined through testing, however it is rare that any testing is complete, let alone sufficient testing to establish statistical validity.

Joint friction and cohesion are considered indirectly in both the Q system and the RMR system (Jr/Ja and Factor A4 respectively). Both systems have values that consider the joint profiles (macro and micro roughness) and infill. Numerical values are assigned to the different conditions. However, simply knowing the numerical values does not provide an insight into the rockmass quality. In the Q system a smooth planar joint with surface staining has the same Jr/Ja value as rough undulating joints with slightly altered joints with sparse mineral coating which have the same rating as rough discontinuous joints with this low friction coatings. Each of these joint conditions present different frictional properties and geotechnical risks. Furthermore, not all discontinuities within an individual joint set are going to exhibit the same characteristics. It is important to understand the overall conditions of the joints as a distribution to ensure that best- and worst-case scenarios are clearly understood. The likelihood of variance can be determined and incorporated into kinematic probability analysis.

RQD

Rock quality designation (RQD) is an empirical “classification” system. In simple terms it measures how broken the rockmass is. It can be used as an indication as to whether rockmass failure is the likely mechanism of failure. It si considered in both the Q and RMR systems.

In general, any value below 25 indicates that the rockmass is severely broken and discontinuity properties are essentially irrelevant. Consistent RQD values less than 25 distributed throughout the rockmass indicate that rockmass failure is likely and that your mining method must be adapted to consider significant stability risks. In mine planning, this is likely to influence your mining method, excavation dimensions, ground support requirements and, in turn, equipment selection. In the operational capacity, drill and blast methods will have to consider the impact to the rockmass and the prevention of significant overbreak to ensure excavation reliability and the safety of operators. Ground support is going to significant and “shell” type systems will need to be considered.

RQD values less than 25 localised to structures (major or minor) have less of an impact. Generally, if a continuous structure is influencing the localised rockmass conditions, it should be modelled as a major structure to ensure adequate consideration within the design process.

RQD values over 75 indicate that rockmass unravelling is an unlikely failure mechanism. The rock will be more forgiving and the restrictions on dimensions and mining methods is less onerous. The other factors mentioned throughout this article will exert a greater influence in determining the stability of an excavation.

Water

Water is considered directly in both the Q and RMR systems (Jw and Factor A5 respectively). Water is a significant consideration in open pit mines and especially when considering soils. The saturation and drying cycle of soils results in expansion and contraction and can lead to failure. Expansion can also be significant when considering swelling clays. In hard rock open pit environments, pore pressure can be significant and cause failure of blocks at the toes of slopes. It also reduces the friction angle of the structures which can also result in failure.

In my experience in underground mines water only directly affects stability in the wettest of mines and in the poorest of rockmass conditions. I have observed the following factors when dealing with water in underground hard rock environments. In most cases hard rock has low porosity and hence pore water pressures rarely accumulate. Major and minor structures form the major conduits for water to the surface of the excavation however, blast damage distributes the pore water and the exposed surface allows for the free flow of water. Often a fault or major structure may be exposed and initially have high pressures, but these dissipate over time if there is no external source of water (e.g. an aquifer). The second factor is the size of the excavation. Underground excavations are generally quite small and joints away from the immediate surrounds of the excavation (5m or so) are confined. Pore water pressure within the joints would have to exceed the confinement of the rockmass to induce failure.

Despite water not being a significant influence in most hard rock underground mines it still must be considered. Indirectly, water can influence stability through the washing out of joint infill causing blocks to dislodge. Furthermore, corrosion of ground support can pose a significant geotechnical risk when capacities are greatly reduced. It is important to consider the conduits for water within the mine and how to manage wet areas should they occur.

Stress

The influence of stresses and strains on the rockmass can have a significant impact on geotechnical risk. Dynamic or high closure environment can change the economic value of a mine and must be managed rigidly to ensure the safety of mine workers. Stress is not considered within the RMR system. While SRF allows the Q system to consider the effect of the magnitude of stress with respect to the rock mass strength, it does not account for the orientation of the stress.? Stress orientation with respect to the major structures and joint orientations can have a significant effect on stability. Stress orientations aligned to induce slip on a major fault can influence on the processes required to control the effects. Drives parallel to the discontinuities but perpendicular to the stress direction can be subject to significant buckling and drive closure.

Strain is not considered in any classification system but in my experience is a much more noteworthy parameter. In the paraphrased words of a colleague, “you may or may not get failure due to high stress, however, in high strain environments failure is likely”. In very simple terms, rock has a lower capacity in tension than in compression. In high stress environments, the UCS of the rockmass determined whether failure will occur. When creating excavations in high strain environments, at the right orientations the walls of the excavation may go into tension causing bucking and substantial closure. Due to the lower load bearing capacity in tension, failure is more than likely.

Conclusion

As mentioned in the opening paragraphs, each of the parameters above should be considered in all geotechnical analysis. The weighting of the parameter will depend on site conditions, but it does not mean that one parameter should be discarded because another has a higher influence over rockmass stability. Overall comprehension of the rockmass and the inputs into evaluations is also critical for an understanding of the assumptions used in analysis. It is the role of the on-site geotechnical engineers to understand these assumptions and ensure that additional analysis is undertaken when conditions vary from these assumptions.