Fractures ~ Porosity ? Connectivity ~ Permeability ? EGS Flow Stimulation

Abstract

The key to attaining EGS electric power generation is controlling fracture stimulation in low-permeability heat exchange volumes between input and output wellbores. Too few stimulation fractures means too little flow of heat-recharged water between input and output wellbores. Too extensive fracture stimulation could mean significant channeled flow in the heat exchange volume leading to local cooling and lost heat exchange capability. Controlling EGS fracture stimulation requires understanding: i. How/why in situ fractures are distributed; ii. How fracture connectivity controls in situ fluid flow; iii. How raising heat-exchange volume fluid pressure can induce greater fracture connectivity. We understand from the systematics of well-log power-spectra that in situ fracture distributions are characterized by long-range spatially-correlated grain-scale-density fluctuation ‘noise’ that scales inversely with spatial frequency, S(k) ~ 1/k, over five decades, ~1/km < k < ~1/cm. We understand from well-core poroperm fluctuation systematics that spatial changes in porosity ? control spatial changes in permeability ? as ??j ~ ?log?j, j = 1….n, where ??j and ?log?j are respectively n-valued zero-mean/ unit-variance fluctuation sequences of well-core porosity and log(permeability) data. While we do not understand in any controlling detail how raising in situ fluid pressure in a crustal volume can be made to induce fracture connectivity in the volume, we note here that integrating the poroperm fluctuation relation empirically describes fracture connectivity as ? ? ?0 exp(?(?-?0)) in terms of the ratio of standard deviations from sample means ? ? ?(log?j)/?(?j) that reflects the degree to which in situ grain-scale fractures are connected in fracture-networks. Small values of ? correspond to diffuse grain-scale fracture connectivity associated with normally distributed permeability populations. Large values of ? correspond to grain-scale fracture connectivity associated with long-tailed permeability population distributions commonly parameterized as lognormal statistical distributions. The degree of fracture-connectivity in a crustal volume thus appears to be the physical explanation lying behind the lognormal statistical descriptions of well-core permeability distributions. (The physics of fracture connectivity may also lie behind the lognormal statistical descriptions often used to characterise trace element and ore-body population distributions.) Expression ? ? ?0 exp(?(?-?0)) implies that for a fixed porosity distribution ?j increased permeability is associated with increased fracture connectivity in the population of grain-scale fractures. Flow simulations in poroperm media with porosity fluctuations having inverse spectral scaling S(k) ~ 1/k and permeability fluctuations obeying ?? ~ ?log? illustrate the increased permeability due to increased fracture connectivity. We thus argue that inducing finite shear strain in a crustal volume introduces new grain-scale defects in association with existing porosity and hence creates greater permeability through greater fracture connectivity within the volume. Naturally occurring finite-strain-injection processes would generate a range of values of ? testified to by the range of observed normal-to-lognormal distributions for well-core permeability, trace element abundance, and ore body grades. We infer that permeability enhancement for EGS heat-exchange volumes can proceed along similar lines.