P-wave attenuation measurements from laboratory resonance and sonic waveform data

Authors:

Abstract
We computed compressional-wave velocity (V p ) and attenuation (1/Q p ) from sonic log waveforms recorded in a cored, 30 m thick, dolostone reservoir; using cores from the same reservoir, laboratory measurements of V p and Q p were also obtained. We used a resonant bar technique to measure extensional and shear-wave velocities and attenuations in the laboratory, so that the same frequency range as used in sonic logging (5-25 kHz) was studied. Having the same frequency range avoids frequency-dependent differences between the laboratory and in-situ measurements.Compressional-wave attenuations at 0 MPa confining pressure, calculated on 30 samples, gave average Q p values of 17. Experimental and geometrical errors were estimated to be about 5 percent. Measurements at elevated effective pressures up to 30 MPa on selected dolostone samples in a homogeneous interval showed mean Q p and Q s to be approximately equal and consistently greater than 125. At effective stress of 20 MPa and at room temperature, the mean Q p over the dolostone interval was 87, a minimum estimate for the approximate in-situ conditions.We computed compressional-wave attenuation from sonic log waveforms in the 12.5-25 kHz frequency band using the slope of the spectral ratio of waveforms recorded 0.914 m and 1.524 m from the source. Average Q p over the interval was 13.5, and the mean error between this value and the 95 percent confidence interval of the slope was 15.9 percent. The laboratory measurements of Q p under elevated pressure conditions were more than five times greater than the mean in-situ values. This comparison shows that additional extrinsic losses in the log-derived measurement of Q p , such as scattering from fine layers and vugs or mode conversion to shear energy dissipating radially from the borehole, dominate the apparent attenuation.