Reservoir properties estimation before and after hydraulic stimulation using extended elastic full-waveform inversion and distributed acoustic sensing

ABSTRACT Monitoring the hydraulic stimulation of unconventional reservoirs using seismic data is an active topic with challenging limitations. The effect of stimulation on seismic velocities is transient, often weak, and typically below the sensitivity of conventional seismic acquisitions. Recent studies show that in situ distributed acoustic sensing (DAS) fiber can record seismic waves generated by the perforation shots, which gives access to a cost-effective acquisition in proximity to the reservoir. However, exploiting such data to characterize the entire reservoir layer remains limited to qualitative and geometric analysis, and the potential of advanced geophysical techniques is largely untapped. In this work, we apply high-frequency elastic full-waveform inversion to a field DAS data set with the perforation shots as sources to estimate the reservoir seismic velocities before and after hydraulic stimulation. The location of the sources inside the reservoir layer causes severe footprints in the inversion, particularly when the reservoir is a waveguide. We develop a novel extended inversion, which uses illumination redundancy, to mitigate this issue. We design a workflow to estimate the dynamically changing reservoir properties after stimulation. We reveal a lateral anomaly in the reservoir present before any stimulation, which is consistent with the independent dispersion analysis. The anomaly may indicate a lower pore pressure or a tighter shale region. We determine that shortly after stimulation, there is a substantial drop in seismic velocities. However, the effect fades away rapidly, and only a residual (mild) decrease in S-wave velocity remains hours to days after the stimulation, which is consistent with a partial closing of the induced fractures.