The Effect of Proppant and Fluid Saturation on the Elastic Properties of 3D-printed Rock Models and Eagle Ford Shale

Abstract

Oil and natural gas from unconventional reservoirs is produced through hydraulic fracturing. Hydraulic fracturing process involves injecting fluids with high pressure to induce fractures and uses sand or ceramic proppant to keep open (propped) those induced fractures. Knowing where the propped fractures are is important to predict the production and to optimize the well spacing and completions of future wells. The objective of this research is to estimate the effect of fluids and proppant on the elastic properties of 3D-printed rock models and the Eagle Ford shale outcrop in laboratory experiments. I conducted two different experiments on 3D-printed rock models; one studies the effect of fluids and the second studies effect of proppant, while the Eagle Ford experiment studies the effect of proppant. In the first 3D-printed experiment, I printed two porous models: a layered model with 6% porosity and another model with horizontal inclusions embedded in a porous layered background with 24% total connected porosity. I saturated these models with different fluids (air, water, engine-oil, crude-oil, and glycerol) to determine the effect of these saturants on the rock's elastic properties at ultrasonic frequencies. From air to liquid saturation, P-velocities increased for all liquids in both models with the highest increase being glycerol (57%) in the direction normal to the plane of horizontal inclusions and layering. Unlike P-wave velocity, S-wave velocity decreased for the inclusion model and there was no significant change for the layered model. For the inclusion model, I observed a greater difference between two S-wave velocities Vs1 (parallel to layering) and Vs2 (normal to layering) than between two P-wave velocities Vp0 (normal to layering) and Vp90 (parallel to layering). I attribute this to the S2 wave, which is probably more sensitive than P-wave and has polarization normal to the plane of horizontal inclusions. For the inclusion model, Thomsen's P-wave anisotropic parameter decreased from 26% to 4% in the air to water substitution, and from 26% to 1% in the air to glycerol substitution. The high viscosity of the saturating fluid significantly reduces the velocity anisotropy of the medium, and it behaves as an isotropic velocity medium. I compared our experimental results with theory and found that predictions using Schoenberg's linear-slip theory combined with Gassmann's anisotropic fluid substitution equation were closer to actual measurements than Hudson's isotropic calculations. This work can help inform us about pore fluid effects on the elastic properties of anisotropic rocks and thus assist in the successful development of hydrocarbon reservoirs. The second 3D-printed model experiment and the Eagle Ford shale experiment are related. They investigate how proppant affects the elastic properties of a rock. I used various samples, including 3D-printed models with air-filled plus sand and ceramic proppant-filled fractures, as well as Eagle Ford Shale samples with artificially created fractures with air and sand-proppant. From the 3D-printed model in uniaxial compression experiments, I found that Vs is decreased by 10% for the sand-proppant model, and the Young’s modulus of sand-propped models was lower than the air-filled or unpropped models, suggesting that propped models may be more compliant. Normal compliance calculated from the static data confirms that propped models are more compliant. I extended this experiment with Eagle Ford Shale samples under isostatic stress of 5000 psi (34.4 MPa). For the Eagle Ford shale the S-wave velocity varied from 2061 m/s when unpropped to 2361 m/s when propped for vertical plugs, 2120 m/s to 2398 m/s for 45 deg plugs and 2156 m/s to 2361 m/s for horizontal plugs. The increase in shear velocity could be attributed to the addition of higher velocity material (sand) to the fractures. S-waves attenuated more than P-waves in propped rocks. Propped rocks have higher horizontal apparent Young's modulus than vertical Young's modulus. P-wave velocity anisotropic coefficient, epsilon, increased for propped rock whereas S-wave anisotropic coefficient, gamma, decreased. Observing a 13% increase in shear velocity with two propped fractures at the ultrasonic frequency scale, a difference in shear velocity with dense propped fractures could be observed in the field, at microseismic frequencies. Therefore, shear wave data acquired before and after hydraulic fracturing could be useful to image propped rock and to estimate accurate stimulated reservoir volume (SRV).