Deepwater Reservoir Uncertainty Evaluation Through Seismic Petrophysics and Multi-Scenario Approach

Abstract

A major challenge in reservoir characterization is the integration of information at different scale from different sources. Coupling seismic data and depositional model is a way forward to capture uncertainty of complex geological heterogeneities. The main object of this paper is to present an application of reservoir uncertainty evaluation through seismic petrophysics and multi-scenario approach to a green field in early production phase, to handle large-scale geological features variability. Starting from a shared petro-elastic facies model for geological and seismic characterization, advanced geostatistical techniques conditioned by seismic elastic inversion attributes are used to generate multiple scenarios in an automated way. In the proposed methodology, core data, well logs and synthetic elastic curves, are key input to a statistical cluster analysis in order to discriminate facies in the petrophysical, petro-elastic and elastic space of seismic inversion. The petro-elastic facies classification allowed a reliable sedimentological and petrophysical characterization, while assuring the maximum discrimination within the elastic space of inversion with a minimum number of classes. The petro-elastic facies model plays the role of common hard data for reservoir modelling and classification of seismic inversion attributes to facies probabilities. A consistent picture of reservoir heterogeneity is then achieved coupling seismic probabilities and Multiple-Point Statistics (MPS). Seismic probabilities provide low frequency trends for reservoir heterogeneity among wells, while process-based modelling is used to generate Training Images (TI) for MPS, capturing depositional patterns and medium-scale geological heterogeneities. MPS modelling, conditioned by seismic trends, reconstructs turbidite Geo-bodies, honoring depositional patterns and well data. A nested Sequential Indicator Simulation conditioned by seismic is then performed to model reservoir internal architecture by distributing productive facies within Geo-bodies realizations. Seismic also drives modelling of porosity by facies. Petrophysical geostatistical simulations based on facies complete the process. Following this methodology, multiple depositional scenarios were generated, combining different Training Images and seismic probabilities weights in MPS simulations. Process automation makes it possible: 1) to investigate a wide distribution of cases and 2) to converge on a subset of realizations whose reservoir dynamic simulation were fitting to the observed production data. The same workflow is also suitable to handle reservoir static properties variability on the selected scenario, capturing the impact of uncertainties on fluid flow behavior. This integrated approach was successful in increasing the geological realism and consistency compared to a traditional deterministic one. Nevertheless, the presented workflow also fits the requirements of a fast-track strategy accelerating model computation and updating.