Integration Of Seismic Lithology Data In The Reservoir Model

Abstract The Marulk field is a gas condensate accumulation located in PL122 in the Norwegian Sea. The field was discovered in 1992: hydrocarbons were encountered in Lysing and Lange formations. The field is operated by EniNorge and it will be put on production with a direct tie-in to the existing Norne FPSO. An update of the reservoir model was based on a new acquired high-quality Q-marine seismic volume, and reservoir properties were distributed using the geophysical attributes: a successful elastic seismic inversion was performed and used as input to the lithology classification study. This supervised Bayesian classification made use of multivariate Probability Density Functions (PDFs) of selected inverted attributes for each litho-class as a-priori information to constrain the classification. Litho-class selection started by analyzing raw and interpreted logs at well locations and petrofacies obtained by means of an integrated petrophysical model. Eventually a new class subdivision was obtained, closely reflecting not only the well data but also the actual geophysical response. This approach resulted in a quantitative estimate of the uncertainty in lithology prediction: significant maps were extracted from Probability volumes of geophysically consistent lithology classes. The obtained probability maps allowed driving the facies distribution in the 3D grid improving the reservoir architecture knowledge and providing a valuable link among seismic data, logs, petrofacies and sedimentological model.