Estimating the Confidence in UCS Measurements with Wireline Logging Data

Rock mechanics analysis depend on understanding the mechanical behavior of the rock mass. This typically involves taking core samples and testing them for unconfined compressive strength (UCS) and Brazilian tensile strength (BTS). However, the critical question is: to what extent are these measurements representative of the surrounding rock mass? How confident can an engineer be when extrapolating these localized results across a mine site?

Wireline full-waveform sonic (FWS) logs from a 300 m borehole near Adelaide, Australia, were used to estimate confidence in sample representativeness. Laboratory tests showed significant variability in UCS, whereas BTS was more consistent. Intact velocities were robustly extracted from the FWS logs and converted into dynamic elastic modulus. Cluster analysis was applied to the wireline logs to define geological domains for mapping UCS. By using the relative median absolute deviation (RMAD) of dynamic modulus with these domains, the study introduces a method to assess the representativeness of UCS measurements.