Estimating reservoir permeabilities using the seismic response to CO 2 injection and stochastic inversion

A Markov Chain Monte Carlo (MCMC) stochastic inversion tool has been developed that identifies porosity/permeability models that minimize the misfit between observed seismic reflection data, reservoir flow modeling, geostatistical methods, and a novel stochastic inversion technique to identify optimal porosity/permeability models. Reservoir model optimization is accomplished through stepwise refinement of its permeability magnitude and heterogeneity. In each step of the inversion, reservoir conditions and CO2 migration are calculated for the current model realization under prescribed CO2/H2O injection and hydrocarbon/H2O withdrawal. Comparison of observed seismic reflection responses with those calculated for the resultant reservoir conditions determine the associated likelihood and whether the proposed reservoir model is acceptable. This process is repeated until the process converges. The algorithm is demonstrated with a synthetic data example showing that primary features of the known porosity/permeability distribution can be recovered. The inversion algorithm is then applied to observed seismic data example from the IEA GHG Weyburn-Midale CO2 Monitoring and Storage Project with limited success. Shortcomings in applying the methodology to real data is assessed and recommendations for improvements are provided. (C) 2012 Elsevier Ltd. All rights reserved.