Evaluating uncertainty in empirically derived unconfined compressive strength (UCS) estimates and implications for drilling applications; a case study from the Cooper Basin

Abstract

Accurate estimates of Unconfined Compressive Strength (UCS) are essential for a range of subsurface applications, including drilling wells for subsurface fluid extraction or injection. However, measuring UCS of subsurface rock samples through laboratory-based uniaxial and triaxial testing is time consuming, expensive, and potentially subject to individual sample variance. P-wave velocity logs are routinely obtained in petroleum basins and are commonly used to estimate UCS using empirically-derived correlations. In this study we analysed P-wave velocity data from 43 wells in the Cooper Basin, Australia and created mean ranges of expected UCS through the Tirrawarra, Toolachee, Patchawarra, Murteree Shale and Epsilon formations, using literature-derived P-wave velocity-UCS correlations, and also estimated UCS based on selected available laboratory-determined P-wave velocity measurements. These two suites of P-wave velocity-derived estimates of UCS are then compared to available results from laboratory uniaxial testing of core samples. Our analysis indicates that P-wave velocities, and subsequent derived estimates of UCS from P-wave wireline logs, are typically lower than those from laboratory-analysed samples. This may reflect a sampling bias towards the selection of strong rocks for laboratory UCS testing or lower estimates of UCS due to wellbore damage and pore space content (air or liquids) from P-wave log derived velocities. However, the laboratory-derived P-wave velocities generally agree with laboratory-derived UCS data from the Patchawarra, Murteree Shale and Epsilon formations when using published empirical velocity-strength correlations. A retrospective case study presents the impacts of different UCS estimates on mud weight required to produce observed borehole breakouts in the Epsilon Formation. Breakouts were observed 90 degrees to the orientation of SHmax (117° N) with a mean width of 51°, the breakouts were produced during drilling using a 9.7 ppg mud weight. The back-calculated estimate of UCS from the observed breakout widths at 9.7 ppg is 154 MPa. This sits between the UCS estimate of 171 MPa from uniaxial testing and 145 MPa from the mean of laboratory sample-derived P-wave velocities. Estimates of UCS from empirical correlations based solely on p-wave log-derived are far lower and vary between 87.3 and 114.4 MPa.