Velocity-effective stress response of CO2-saturated sandstones

Abstract

Three differing sandstones, two synthetic and one field sample, have been tested ultrasonically under a range of confining pressures and pore pressures representative of in-situ reservoir pressures. These sandstones include: a synthetic sandstone with calcite intergranular cement produced using the CSIRO Calcite In-situ Precipitation Process (CIPS); a synthetic sandstone with silica intergranular cement; and a core sample from the Otway Basin Waarre Formation, Boggy Creek 1 well, from the target lithology for a trial CO2 pilot project. Initial testing was carried on the cores at ?room-dried? conditions, with confining pressures up to 65 MPa in steps of 5 MPa. All cores were then flooded with CO2, initially in the gas phase at 6 MPa, 22°C, then with liquid-phase CO2 at a temperature of 22°C and pressures from 7 MPa to 17 MPa in steps of 5 MPa. Confining pressures varied from 10 MPa to 65 MPa. Ultrasonic waveforms for both P- and S-waves were recorded at each effective pressure increment. Velocity versus effective pressure responses were calculated from the experimental data for both P- and S-waves. Attenuations (1/Qp) were calculated from the waveform data using spectral ratio methods. Theoretical calculations of velocity as a function of effective pressure for each sandstone were made using the CO2 pressure-density and CO2 bulk modulus-pressure phase diagrams and Gassmann effective medium theory. Flooding the cores with gaseous phase CO2 produced negligible change in velocity-effective stress relationships compared to the dry state (air saturated). Flooding with liquid-phase CO2 at various pore pressures lowered velocities by approximately 8% on average compared to the air-saturated state. Attenuations increased with liquid-phase CO2 flooding compared to the air-saturated case. Experimental data agreed with the Gassmann calculations at high effective pressures. The ?critical? effective pressure, at which agreement with theory occurred, varied with sandstone type. Discrepancies are thought to be due to differing micro-crack populations in the microstructure of each sandstone type. The agreement with theory at high effective pressures is significant and gives some confidence in predicting seismic behaviour under field conditions when CO2 is injected.