Correlation of rock strength between uniaxial compressive, Brazilian and point load tests: a laboratory study

Abstract

Rock engineering design for underground or on-the-ground projects and operations are widely dependent on different properties of rocks among which strength is one of the mainncomponents. The uniaxial compressive strength (UCS) test is still considered to be the mostneffective method, currently used by many industries to estimate either intact or rock massnstrength along with other rock strength empirical indices. The Uniaxial Compressive Strengthnis the test in which a rock sample of either a prismatic or cylindrical shape is compressednbetween two parallel rigid plates under a constant strain rate while the load, axial and lateralndeformations are recorded by a data acquisition device. However, determining the UCS of rocksnis a very complex process due to the need of having quality rock samples, sampling and coringnwhich is time consuming as well as being costly. Furthermore, based on economic reasons, itnhas never been feasible to get a full measurement of every characteristics and other factorsnaffecting the rock behaviour for UCS purpose. Due to its complexity, studies and researchesnare still being conducted to ensure cost and time currently involved are minimised by providingnsome correlations between different other tests which are easy to conduct.nThis paper reviews the Brazilian Tensile Strength Test (BTS) and the Point Load Index Testn(PLI) established by previous researchers by conducting laboratory experiments and correlatenthe findings between these two tests and the UCS tests so that they can be used as alternativenoptions on particular specific rock samples. BTS test is performed by compressing a samplenbetween two curved loading jaws as shown in Figure 2 b). Point load on the other hand, isnperformed manually by compressing a sample between two conical steel platens as shown innFigure 3.nIn order to achieve that, a literature review was conducted to understand rock behaviour undernthe influence of a load and previous findings. In addition to that, three samples (Sandstone,nBasalt and Brisbane Tuff) were sampled, cored, and tested using UCS, BTS and PL machinesnin line with the International Society for Rock Mechanics (ISRM). Previous researches havenshown that there is correlation between UCS, BTS and PL with the most commonly used fornconverting k factor from PL into UCS being (22-24) x (50) and a whole range of othernconversion factors for UCS and BTS. However, over the past years, throughout the wholenprevious researches all these factors were found to changing based on rock types and otherngeological properties in the rocks. Hence, the need to conduct as many experiments as possiblenon different rock types in the aim of correlating the results by either validating previous findings or providing new coefficient factor for future use. Due to the general usage of certain conversionnfactors, this paper focused on these particular three rock samples in order to prove whether thenpast findings also apply to these them. For the conversion of the PL into UCS, the k factor was found to be 17.78, 25.78 and 32.62 fornSandstone, Brisbane Tuff and Basalt respectively while the conversion from BTS to UCS, thenk factor was found to be 10.55, 10.65 and 15.14 for sandstone, Brisbane Tuff and basaltnrespectively. It can be seen that from these data, there is no general k factor to convert fromnBTS or PL into UCS and hence the reason why more investigations still need to be done tonincrease the chance of getting a general k factor for different rocks.n