Studies assembling high quality datasets of fracture systems (joints and faults) from four reservoir analogues are described. These comprise limestones (Ireland), sandstones (Norway and Saudi Arabia) and chalk (Denmark). These are used with existing information from the literature to review the major controls and scaling behaviour of fracture systems expected in reservoir rocks. Lithological layering was found to be important and two end-member fracture systems have been identified. In "stratabound" systems, fractures are confined to single layers, sizes are scale restricted, and spacing is regular. In "non-stratabound systems", fractures show a wide range of sizes (often power-law), are spatially clustered and vertically persistent. In nature, variations between and combinations of these systems exist. These end-member systems have contrasting implications for fluid flow, including the scale of fracture that controls flow and the existence of a representative elementary volume, and thus on appropriate modelling approaches.

Over the past 20 years, the Soultz experimental geothermal site in Alsace, France, has been explored in detail by the drilling of five boreholes, three of which extend to 5 km depth. Data on geology, fluid geochemistry, temperature, microseismicity, hydraulics and geomechanics have been collected and interpreted by the various teams from the participating European countries and their international collaborators. Two reservoirs have been developed within granite at depths of 3.5 and 5 km. The reservoir at 3.5 km was formed from two wells, 450 m apart, both of which were subjected to hydraulic stimulation injections. The system was circulated continuously for 4 months at 25 kg/s in 1997 using a downhole pump, and yielded results that were extremely encouraging. The impedance reduced to 0.1 MPa/l/s, the first time this long-standing target had been attained. Construction of a deeper system began shortly afterwards with the drilling of 3 deviated wells to 5 km true vertical depth, where the temperature was 200 °C. The wells were drilled in a line, 600 m apart at reservoir depth, and all were hydraulically stimulated and subjected to acidization injections. The 3-well system was circulated under buoyancy drive for 5 months in 2005 with injection in the central well, GPK-3, and production from the two outer wells, GPK-2 and GPK-4. This showed good linkage between one doublet pair, but not the other. Further acidization operations on the low-productivity well led to its productivity increasing to almost the same level as the other wells. Construction of a power plant at the site was completed in 2008 and a trial circulation with a production pump in one well and the other shut-in was conducted with power production. Downhole pumps are now installed in both production wells in preparation for long-term circulation of the system. In this article we present an overview of the principal accomplishments at Soultz over the past two decades, and highlight the main results, issues identified, and lessons learnt.

Summary. The European Enhanced Geothermal System (EGS, formerly Hot Dry Rock, HDR) programme of Soultz-sous-Forêts is organized around three wells drilled to a

An extended description of the implementation of the deep geothermal wells GRT-1 and GRT-2 respectively drilled in 2012 and 2014 in the framework of the ECOGI project is given. These wells located in Rittershoffen (France) offer a unique opportunity to gather high quality datasets on a deep geothermal system in the Upper Rhine Graben at the transition between the Buntsandstein sandstone and the Palaeozoic granite basement. We present the extensive logging and well testing program that was applied and focus on hydraulic and thermal characterization of the targeted deep reservoir. Well architectures of GRT-1 and GRT-2 are described. Temperature logs in both wells are discussed in details. In particular, temperature in the production well GRT-2 is shown to reach 177°C at 3196m MD at thermal equilibrium. Production tests of both wells, reservoir development strategy applied in well GRT-1 and circulation test realised between wells are analysed. Productivity and injectivity indexes of GRT-1 and GRT-2 are estimated: Post-stimulation injectivity of GRT-1 and productivity of GRT-2 are respectively estimated to 2.5l/s/bar and 3.5l/s/bar. Hydraulic properties of the reservoir are inferred from production tests. Implications for the characterization of the large scale natural hydro-thermal system are discussed