Estimation of Pressure and Permeability Enhancement Distribution Using Induced Earthquake Hypocenter Density for the 2011 Paralana EGS Stimulation

Abstract

To create an Enhanced Geothermal System (EGS), cold water is injected at high pressure, along with acid, with the goal of reactivating pre-existing fractures and enhancing their permeability. Through increases in pore pressure and associated stress changes, shear failure occurs, which is part of the permeability enhancement process, but also results in induced seismicity. In spite of being the primary goal of stimulation, details about the spatiotemporal evolution of permeability are difficult to determine. One measure of its improvement is the increase in well injectivity, which is defined as the injected flow rate divided by the wellhead pressure. However, this measure is sensitive to both the volume of stimulated rock as well as the permeability increase, and so it does not uniquely constrain the stimulation state. To augment this analysis, we present an inverse modelling approach that incorporates both the injection records and the spatiotemporal distribution of induced seismicity. We present an application of the method to the Paralana-2 EGS stimulation undertaken in 2011 in South Australia. High pressure injection is modelled by solving coupled flow and heat transport equations in the reservoir simulator FEHM. In the model, the magnitude of permeability increase is a prescribed function of space and time. The injectivity profile observed at Paralana limits the possible set of permeability evolution scenarios, however, additional constraint is necessary to choose amongst these. As induced seismicity is a consequence of elevated pore pressure, we assume that the density of earthquake hypocenters is proportional to pore pressure rise. By comparing the pressure profiles modelled in the different scenarios to the high-resolution microearthquake data collected during the stimulation, we can pick the permeability enhancement distribution scheme most consistent with the injectivity and seismicity data.