Abstract

Conventional approach to 4D seismic inversion consists of parallel inversions applied to seismic vintages. Only then, the inverted changes of seismic attributes are converted into petrophysical properties using rock physics. The paper develops a robust approach to 4D seismic inversion based on a Bayesian approach along with rock physics constraints. This means that observed time-lapse seismic response along with the baseline amplitudes are inverted directly into rock properties via pre-defined relations to the seismic properties. To this end, we extend the functional of Delivery - an open-source stochastic inversion software. We illustrate efficiency of Delivery4D using synthetic 4D dataset generated for Stage 2C of CO2CRC Otway project, Victoria. Complexity of the synthetic wavefield resembles field data acquired for the Otway project while all unknown sources of noise/uncertainty are excluded and we have ‘ground-truth’ subsurface properties. Despite the relatively thin CO2 plume, the 4D inversion reduced detected time-lapse anomaly to the location that closely corresponds to the actual CO2 plume. Estimated distributions of the plume characteristics (thickness, saturation and CO2 mass) are overall similar to the static and dynamic geomodels. However, the values inverted at a particular trace may differ significantly. We attribute these discrepancies to the limited seismic resolution and imperfections of the amplitude-preserved seismic processing.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call