Subcritical growth of natural hydraulic fractures

Joints are the most common example of brittle tensile failure in the crust. Their genesis at depth is linked to the natural hydraulic fracturing, which requires pore fluid pressure in excess of the minimum in situ stress [Pollard and Aidyn, JSG1988]. Depending on the geological setting, high pore pressure can result form burial compaction of interbedded strata, diagenesis, or tectonics. Common to these loading scenarios is slow build-up of pore pressure over a geological timescale, until conditions for initiation of crack growth are met on favorably oriented/sized flaws. The flaws can vary in size from grain-size cracks in igneous rocks to a fossil-size flaws in clastic rock, and once activated, are inferred to propagate mostly subcritically [Segall JGR 1984; Olson JGR 1993].

Despite many observational studies of natural hydraulic fractures, the modeling attempts appear to be few [Renshaw and Harvey JGR 1994]. Here, we use boundary integral formulation for the pore fluid inflow from the permeable rock into a propagating joint [Berchenko et al. IJRMMS 1997] coupled with the criteria for subcritical propagation assisted by the environmental effects of pore fluid at the crack tip to solve for the evolution of a penny-shape joint, which, in interbedded rock, may eventually evolve to short-blade geometry (propagation confined to a bed).

Initial growth is exceedingly slow, paced by the stress corrosion reaction kinetics at the crack tip. During this stage the crack is fully-drained (i.e. the fluid pressure in the crack is equilibrated with the ambient pore pressure). This “slow” stage is followed by a rapid acceleration, driven by the increase of the mechanical stress intensity factor with the crack length, towards the terminal joint velocity. We provide an analytical expression for the latter as a function of the rock diffusivity, net pressure loading at the initiation (or flaw lengthscale), and parameters describing resistance to fracture growth. Due to a much slower rate of the crack volume expansion of short-blade joints compared to that of penny-shape joints, the former would propagate much faster than the latter under otherwise identical conditions. Finally, we speculate about possible relation of the predicted patterns of joint development with morphology of joint fracture surfaces observed in sedimentary rock.