| Abstract |
Heat extraction from deep 'engineered' fractured formations is currently under investigation at many places in the world. The challenge is to develop a reservoir in deep and hot rock masses and to circulate a fluid in order to recover heat and produce electricity. After a significant success at a moderated depth, the European project now evolves toward a three well system at a depth of 5 km, and a temperature close to 200 C. Hydraulic tests obtained in the recent wells GPK2 and GPK3 during the 2000 - 2003 field campaign have shown from the micro-seismic signature, that fractures in the new targeted reservoir are still easy to re-activate. However the overall gain of impedance at the reservoir scale remained poor. The discussion now focuses on the hydraulic significance of the shear failure mechanism, considered as the source of the acoustic emissions. To improve our understanding of these coupled hydrau-mechanical processes, a numerical model was presented [1], based on a 3D random description of fracture networks. Local flow rules along equivalent 1D channels connecting the fractures can account for a normal closure versus effective stress dependency together with a dilatant behaviour during shearing motion when a failure criterion is met. The purpose of the present work is to simulate hydraulic tests in some synthetic fracture networks with different fracture size distributions (log-normal, power law), to characterise their hydraulic diffusivity from low rate injection tests and to analyse the spatio-temporal growth of the sheared zones, assuming that this process is analogue to the development of the induced seismicity. An evaluation of the so called SBRC reservoir characterisation method [2] is then performed and we show how sensitive it is to the hypothesis of critically stressed pre-existing fractures. The method is found generally successful and we discuss cases where a set of sub-critically oriented fractures exist that does not form a connected sub-network. The examples also confirm that the migration rate of the seismic front is not affected by the irreversible changes experienced by the fractures in between the injection zone and the failure front. Dealing with Hot Dry Rock projetcs, the main conclusion derived from the above agreement is that the local hydraulic properties of a 3D discrete fracture network can be calibrated to reproduce the virgin upscaled reservoir permeability before stimulations are performed. This numerical procedure is also found appropriate to discuss the occurrence of delayed seismic events, triggered in the far field after an injection phase has terminated. [1] Bruel, D., 2002, Impact of induced thermal stress during circulation tests in an engineered fractured geothermal reservoir. Oil & Gas Science and Technology - Rev. IFP, 57, no. 5, 459-470 [2] Shapiro, S.A., Royer, J. and Audigane, P., 1999, Large scale in situ permeability tensor of rocks from induced microseismicity. Geophysical Journal International, 137, 207-213. |