Uniaxial compressive strength (UCS) and the Brazilian tensile strength (BTS) are the most commonly used rock properties. The UCS/BTS ratio is used as an indicator of rock brittleness and to estimate UCS from BTS (or vice versa). However, the relation between UCS and BTS has not been investigated in detail. This paper focuses on UCS/BTS ratio using the raw data pertaining to 1842 rock types from different research institutes. Regression analysis indicated that power law or exponential relations offers a better best fit correlation between UCS and BTS for each data set from different institute than linear function which has been typically offered in the past. Power function was also found to offer a better relationship between UCS and BTS for the all data under evaluation. The correlation coefficient of power function (r = 0.79) was higher than that of the linear function (r = 0.64).
Rock engineers widely use the uniaxial compressive strength (UCS) and Brazilian tensile strength (BTS) of rock in designing surface and underground structures, rock excavation projects, drilling and blasting projects etc. Numerous researchers have investigated the relation between UCS and BTS and suggested linear correlations between two parameters. However, some researchers (Altindag & Guney 2010, Arioglu & Tokgoz 2011, SeshaPhanietal. 2013, Nazir etal. 2013) recently found non-linear relations between UCS and BTS. In this study, the raw data from laboratory testing on 1842 different rock types were analysed to investigate the best formulas for describing the relation between UCS and BTS.
The literature (Farmer 1983, Budavari 1983, Jumikis 1983) indicates the tendency of some researchers to assume rock compressive strength that is approximately 10 times tensile strength for most rocks. However, it has been indicated by others that the UCS/BTS ratio may change from 1.9 to 176.6 (Table 1).
Many researchers presented linear correlations between UCS and BTS. Limited number of recent studies has suggested exponential or power functions to define the relationship between UCS and BTS. The study of SeshaPhani et al. (2013) is on concrete and the studies of Nazir et al. (2013) and Arioglu & Tokgoz (2011) are limited specific rock types, including limestone and rocks from the North Anatolian fault zone, respectively. However, the study of Altindag & Guney (2010) comprises data set containing 143 rock sample compiled from previous studies, ranging from weak rock to very strong rocks, including different types from different areas.
Results of UCS and BTS measurement on 1842 different rock types which were tested over years at different institutes such as Colorado School of Mines (CSM), Pennsylvania State University (PSU), Istanbul Technical University (ITU), University of Melbourne (UM), and Nigde University (NU) were collated and used to evaluate the relation between UCS and BTS. It must be noted that the data was collected from different countries and geographical locations and includes different rock types, covering all rock classes such as sedimentary, igneous, and metamorphic. After receiving of the data from various laboratories, an initial evaluation was performed to eliminate the test results related from structural failure of the samples based on the recorded type of failure. But the classification of the failure as Structural Failure could be subjective and simply based on the observation of the existing planes of discontinuity in the sample prior to testing. If no such defects were observed or reported, a judgment could not be made as to the validity of data and absence of such defects. This could lead to scatter in the data set and presence of some outliers in the data that was inevitable and will be discussed in following sections. CSM data set consists of the UCS and BTS values of 182 different rock types, mainly welded tuff, granite, sandstone, limestone, and argillite. The data have very wide strength range. The UCS values ranges from 1.3 to 468.5 MPa. But, the strength of most samples is less than 150 MPa.