Seismic 4D Inversion for Quantitative Use in Automated History Matching

Abstract

We use 4D seismic inversion on equal footing with other types of dynamic data (well water cuts and GORs along with initial RFTs) to match the reservoir production history and improve predictive capability of the models. Automated history matching focuses on minimising the mismatch between simulated and observed data. This comparison can be performed on multiple levels: migrated seismic data vs. synthetic seismic data; elastic properties calculated from simulation model vs. inverted elastic properties; or simulated pressures and saturations vs. inverted pressures and saturations. We acknowledge the existence of uncertainties in all the data and at every level of the workflow, and choose to work within the Bayesian framework to capture multiple realisations in reservoir simulations and seismic inversion. The workflow requires a petro-elastic model (PEM) which links the reservoir simulation to elastic parameters. Initial comparisons demonstrate that the inverted elastic properties are consistent with the simulated elastic properties obtained using PEM. Preliminary indications are that this data can be directly included in an automated history matching workflow.