Predicting Fluid Flow in Stressed Fractures: A Quantitative Evaluation of Methods

Abstract:

Reliable estimation of fracture stability in the subsurface is crucial to the success of exploration and production in the petroleum industry, and also for wider applications to earthquake mechanics, hydrogeology and waste disposal. Previous work suggests that fracture stability is related to fluid flow in crystalline basement rocks through shear or tensile instabilities of fractures.

Our preliminary scoping analysis compares the fracture stability of 60 partly open (apertures 1.5-3 cm) and electrically conductive (low acoustic amplitudes relative to matrix) fractures from a 16 m section of a producing zone in a basement well in Bayoot field, Yemen, to a non-producing zone in the same well (also 16 m). We determine the Critically Stressed Fractures (CSF; Barton et al., 1995) and dilatation tendency (T_d; Ferrill et al., 1999). We find that:

1. CSF (Fig. 1) is a poor predictor of high fluid flow in the inflow zone; 88% of the fractures are predicted to be NOT critically stressed and yet they all occur within a zone of high fluid flow rate

2. T_d (Fig. 2) is also a poor predictor of high fluid flow in the inflow zone; 67% of the fractures have a LOW T_d(< 0.6)

3. For the non-producing zone CSF is a very reliable predictor (100% are not critically stressed) whereas the values of T_dare consistent with their location in non-producing interval (81% are < 0.6) (Fig. 3 & 4).

In summary, neither method correlates well with the observed abundance of hydraulically conductive fractures within the producing zone. Within the non-producing zone CSF and T_d make reasonably accurate predictions. Fractures may be filled or partially filled with drilling mud or a lower density and electrically conductive fill such as clay in the producing zone and therefore appear (partly) open. In situ stress, fluid pressure, rock properties (friction, strength) and fracture orientation data used as inputs for the CSF and T_d calculations are all subject to uncertainty. Our results suggest that scope exists to systematically quantify and explore the impacts of these uncertainties for better predictions of geomechanical stability and fluid conductivity in the subsurface.