A real case application of a seismic reservoir characterization methodology in a deep water clastic reservoir located offshore West Africa is hereby presented.

The aim of this study is to integrate seismic facies classification into the reservoir static modelling workflow. The proposed probabilistic approach for petroelastic properties estimation and facies identification based on seismic data can be divided into three main steps. First, a Bayesian elastic inversion is performed to obtain seismic impedances from partial-stack seismic data. Then, a statistical rock physics model is applied to estimate the posterior probabilities of porosity, clay content and water saturation from impedances distributions. Finally, a litho-fluid classification is performed combining petrophysical properties probabilities with log facies analysis at well location, to propagate the identified classes in the 3D grid. The final result is a set of probability volumes for each seismic litho-fluid facies that can be used as prior 3D trend for geostatistical simulations. Using Sequential Indicator Simulation algorithm several facies realizations are generated and relevant sets of petrophysical properties are simulated within each facies.

The developed workflow has been applied in a real case study offshore West Africa. The available reservoir model has been updated including the results of the petroelastic inversion, in addition to a set of seismically derived geobodies representing the main geological features (channels, levees and sand lobes). Integrated use of seismic 3D probability trends and geological constraints provides advantages in terms of a more detailed spatial distribution of rock properties, a meaningful geological 3D representation and a better understanding of vertical and lateral connectivity within the reservoir. Moreover, development wells may be optimized in a more reliable way.

The contribution of the presented methodology is of key importance for fields in early development phase with few exploration well data available to constrain 3D models.

You can access this article if you purchase or spend a download.