Abstract

Enhanced Geothermal Systems (EGS) use hydraulic stimulation to improve production from high temperature, relatively low permeability formations. EGS projects are usually performed in crystalline lithologies, such as granite. In these settings, there is ambiguity about the relative role of newly forming and preexisting fractures, and there is significant uncertainty about the input parameters that should be used in numerical models. The uncertainties can only be resolved by gathering of field-scale data, and then evaluating modeling approaches and input parameters based on their ability to rationalize the observations. This paper analyzes the data from the Stage 3 stimulation of Well 16A(78)-32 at the Utah FORGE project. The stimulation was performed in granitic rock, from a single perforated interval, with cross-linked gel and microproppant. High-resolution microseismic observations show that the stimulation formed a planar region perpendicular to the minimum principal stress, suggesting that a planar fracture model is appropriate for describing the stimulation. The microseismic observations constrain the size, direction, and aspect ratio of fracture. Hydraulic fracture propagation in granitic rock poses novel theoretical questions. Because granitic formations lack layering and should be relatively homogeneous (compared with sedimentary formations), fracture propagation models may tend to predict extreme vertical height growth; however, this is not observed in actual data. The 16A(78)-32 Stage 3 data suggests that the fracture propagated slightly more laterally than vertically, and with only a modest upward bias. In this paper, a planar fracture model is built and calibrated to match the data. To match the observed fracture geometry with a fracturing simulator, an automated history matching algorithm was used to vary five field-scale calibration parameters that affect fracture geometry. The algorithm converged to a relatively narrow combination of parameters to match the data. The match implies: (a) modestly elevated toughness above typical laboratory values, (b) substantial ‘pressure dependent permeability’ contributing to leakoff from the main fracture, (c) moderately decreased fracture conductivity from the standard cubic law, (d) moderate anisotropy of vertical toughness, and (e) modest anisotropy of vertical conductivity. Future stimulations at Utah FORGE will provide opportunities to evaluate and/or refine the model parameters derived from the Stage 3 match.