Chapter 3: Mechanical Testing of Rock Materials
A significant number of tests to predict comminution characteristics have been introduced since the early work of Kick [1885] and Rittinger [1867] and fall into one of three main categories. Grinding methods were covered in Chapter 1 with the remaining types being loading and impact based.
3.1 Factors Affecting Results
A number of variables in both the specimen and the actual test method itself can alter the results obtained. A full review of such variables can be seen in Vutukuri et al [1974].
3.1.1 Specimen Geometry
Depending on the actual details of the test itself a sample can fall into one of a number of geometric shapes. Typical shapes include cylinders, bars, cubes and irregular lumps. The method of force application can lead to different values of strength for a similar shape. Therefore differently shaped particles require the use of modified equations. These variations have lead to the need for a standardisation. Core samples have proved to be the most popular [Brook, 1977]. The effect of shape cannot be overlooked as Lundborg [1976] found that for tests using cylinders the strength decreased with increasing size.
3.1.2 Length to Diameter Ratio (L/D)
It has been recommended [Vutukuri et al, 1974] that to obtain reliable results L/D ratios should be in the region of 2.5-3.0 when tests are compressive in nature. However this is not always possible in practice, and it has been found by Bearman [1998] that values of as low as1:1 can be used effectively. Using values of greater than 3 can lead to additional stresses, for example the effects of twisting. Direct tensile tests differ, as there is a need to grip the specimen, therefore requiring a larger L/D.
3.1.3 Rate and speed of Loading
The rate of loading is where the applied force is increased linearly by adjusting the speed of the loading body. In comparison with a constant speed of loading force where the load observed can increase non-linearly. Willard and McWilliams [1968] found that point-loaded discs loaded at a higher rate were able to absorb more energy than those loaded at slower rates. This is likely to be due to the fact that the samples were not homogenous in nature. In contradiction to this Vutukuri et al [1974] reviewed work that concluded that the rate of loading has no significant effect on compressive strength. It was concluded the rate of loading has little effect on rocks that are homogenous, low in porosity and are elastic in nature.
3.1.4 Rock Fabric
The fabric of a rock is usually multi-phase and as a result cannot easily be defined. For example, some inclusions may be very hard while the main matrix itself is very soft. It is highly likely that the interfaces between any phases within the rock are the weakest points provided that the forces are applied in the correct orientation. Figure 3.1.4 shows the effect of layering and how it could affect experimental results.
Figure 3.1.4 Effect of Rock Fabric
As can be seen, the force applied to sample A has the ability to run along the particle interface, as opposed to sample B, where the force has to cross the interface. This is likely to increase the energy required to break sample B when compared to that of sample A.
Another factor that can be linked with the fabric of the rock is the possibility of flaws arising from porosity. These flaws can take the form of ‘holes’ or microcracks. Microcracks can reduce the observed strength of rock due to the fact that pre-propagated cracks are present. Griffith [1920] showed that cracks could reduce the actual strength of a material when compared to a theoretical strength by as much as two orders of magnitude.
3.1.5 Surface Finish
The surface finish can be an important factor that requires consideration with both the sample and the loading platten. If the loading surfaces are rough in nature, this can give rise to increased stress concentrations, which can subsequently alter the experimental value achieved.
The surface finish of the test specimen can be critical in some cases. For example the chevron bend test has been used extensively in the metallurgical applications and has been subsequently adapted to rock mechanics. It is widely accepted that the quality of the surface affects the final values for KIC, and therefore the surface finish of the test piece must be the same as the bulk sample to ensure that a scientific comparison can be made.