In the analysis of enhanced heavy oil production, it is crucial to have some knowledge of fluid viscosity in order the flow of oil in reservoir simulation. While viscosity can be measured using fluid samples taken from the borehole, it would be advantageous to have some measure of viscosity between wells. We develop a methodology for doing this by using seismic traveltime tomography to estimate velocity, followed by attenuation tomography to estimate seismic-Q. Tomographic inversion methods have been applied to crosswell seismic data from a heavy oil field in Northern Alberta. The Q values are transformed to viscosity values using the Biot Squirt theory (BISQ). While this transformation can be ambiguous, the ambiguity problem can be obviated if we have some borehole information about viscosity to within an order of magnitude.
The estimates of Q could presumably be further improved by the use of full waveform inversion. One of the concerns with full waveform inversion is its sensitivity to noise. We have tested full waveform inversion on noisy model data and have found it to be robust for signal-to-noise ratios of 5, with convergence to a model fit within three iterations.
The next steps in this research should involve the application of the inversion methods and rock physics to time-lapse crosswell seismic data to monitor the changes in viscosity with time. This greatly enhance our ability to characterize dynamic changes is a heavy oil reservoir.
View full article as PDF (0.17 Mb)