Structure of the geomodel is always controlled by key depth horizons and faults coming from seismic interpretation. However, these geophysical surfaces carry uncertainties attached to seismic picking and seismic velocities. Some error done on estimation of such velocities can have a great impact on final volumes & on field development as it can change drastically the shape of the reservoirs.
We propose here a multi-realizations approach to manage properly these uncertainties during the geomodeling process, in order to estimate their impact and build alternative distorted geomodels for pessimistic and optimistic scenarios.
This method relies on the alternative velocity models built by the interpreters to depth convert the horizons at the origin of the geomodel grid. These velocity models reflect geophysical uncertainty through alternative hypothesis to the basecase velocity model. The stochastic approach presented mixes those models generating hundreds of intermediate scenarios for top & base reservoir horizons. For instance structure can be tilted, smaller or larger compared to the basecase.
This approach enables the user to decide how far from the basecase top & base reservoir realizations should vary (using confidence weights). So the user can maintain the possible realizations around the most likely one (??).
We propose here to expose & apply each step of this technique to an offshore West-Africa field. The objective of the study is triple: to get a dispersion of the possible GRV values, to get distorted versions of the geomodel according to optimistic or pessimistic GRV realizations so that extreme cases can be analyzed. And eventually to assess possible variations of the flanks of the structure.
This last point is crucial to take into account uncertainties in the producer/injector well development pattern.