Cooling-Induced Acoustic Emissions in Westerly Granite: Implications for in Situ Stress Measurement

Abstract

ABSTRACT: Thermally induced stress changes in rock lead to creation and reactivation of microcracks with accompanying acoustic emissions (AEs). Therefore, monitoring AEs during cooling (or heating) of rock potentially allows us to determine the rock’s original stress state. This approach would require predicting time-dependent changes in the thermally induced stress and applying an appropriate AE-based failure criterion, so that the in situ stress can be determined by subtracting the induced stress from the failure stress. The goal of this research is to develop a technology to determine the stress state in geothermal reservoirs by cooling the borehole circumference and detecting the onset of microscale rock failure. As the first step to achieve this goal, we examined spatio-temporal evolution of AEs which were thermally induced in Westerly Granite blocks by abrupt local cooling. Four sets of experiments were conducted with different sample and cooling geometry, stress state, and preexisting damage (cracks). The preliminary results indicate that the onset of AE activity can be correlated best to the thermal front propagation in rock samples and to the applied stress. Interestingly, pre-existing microcrack damage in rock had little impact on the AE onset behavior. The conditions that lead to the onset of AE activity and its relation to the ultimate tensile strength of rocks – a parameter which is determined from conventional rock testing – are still poorly understood, which requires further investigation. 1 INTRODUCTION Sustainable development of geothermal energy resources requires detailed knowledge about the in situ stress state in the target rocks. In deep boreholes, stress estimation commonly relies on analysis of wellbore wall failures, break-outs and drilling induced tensile fractures, which develop spontaneously after the well is drilled (e.g. Zoback et al., 2003). Active methods of stress measurement are usually limited to hydraulic methods for estimation of the least principal stress (De Bree and Walters, 1989). Several methods of active stress measurement, capable of constraining the entire stress tensor have been proposed, e.g. overcoring (Worotnicki, 1993), hydraulic tests on preexisting fractures (HTPF) (Cornet, 1983), or sleeve fracturing (Stephansson, 1983). However, each of these methods has their associated technical difficulties or deficiencies which make them unpopular in practical application in deep environments.