| Abstract |
The Habanero Engineered Geothermal System (EGS) in central Australia has been under development since 2002, with several deep (more than 4000 m) wells drilled into the high-heat-producing granites of the Big Lake Suite to date. Multiple hydraulic stimulations have been performed to improve the existing fracture permeability in the granite. The stimulation of the newly-drilled Habanero-4 well (H-4) was completed in late 2012, and micro-seismic data indicated an increase in total stimulated reservoir area to approximately 4 km². Two well doublets have been tested, initially between Habanero-1 (H-1) and Habanero-3 (H-3), and more recently, between H-1 and H-4. Both doublets effectively operated as closed systems and excluding short-term flow tests, all production fluids were re-injected into the reservoir at depth. Two inter-well tracer tests have been conducted since 2008, to evaluate the fluid residence time in the reservoir alongside other hydraulic properties, and to provide comparative information to assess the effectiveness of the hydraulic stimulations. The closed-system and discrete nature of this engineered geothermal reservoir provides a unique opportunity to explore the relationships between the micro-seismic, rock property, production and tracer data. The most recent inter-well tracer test occurred in June 2013, which involved injecting 100 kg of 2,6 naphthalene-disulfonate (NDS) into H-1 to evaluate the hydraulic characteristics of the newly-created H-1/H-4 doublet. Sampling of the production fluids from H-4 occurred throughout the duration of the 3-month closed-circulation test. After correcting for flow hiatuses (i.e. interruptions in injection and production) and non-steady-state flow conditions, tracer breakthrough in H-4 was observed after 6 days (compared to ~4 days for the previous H-1/H-3 doublet), with peak breakthrough occurring after 17 days. Applying moment analysis to the data indicated that approximately 56% of the tracer was returned during the circulation test (vs. approximately 70% from the 2008 H-1/H-3 tracer test). This suggests that a considerable proportion of the tracer may lie trapped in the opposite end of the reservoir from H-4 and/or may have been lost to the far field. Flow capacity:storage capacity plots derived from the H-1/H-4 tracer test indicate that the Habanero reservoir is moderately heterogeneous, with approximately half of the flow travelling via around 25% of the pore volume. The calculated inter-well swept pore volume was approximately 31,000 m³, which is larger than that calculated for the H-1/H-3 doublet (~20,000 m³). This is consistent with the inferred increase in reservoir volume following hydraulic stimulation of H-4. |