Linda M. Bonnell¹, Robert H. Lander¹, James C. Matthews¹
(1) Geologica A.S, N-4003 Stavanger, Norway

Abstract: Probabilistic Prediction of Reservoir Quality in Deep Water Prospects Using an Empirically Calibrated Process Model

Presently, deep water reservoirs are generally synonymous with frontier areas. Frontier areas have few wells, little core and a large uncertainty in reservoir character. A key uncertainty in prospect assessment is reservoir quality. In a study by Rose (1987), 40% of the analyzed dry holes were caused by poor prediction of reservoir quality.

In this study we demonstrate a method for pre-drill reservoir quality risk assessment and sensitivity analysis using the Exemplar sandstone diagenetic model combined with Monte Carlo techniques. The Exemplar model is a broadly applicable, empirically calibrated process model that simulates two key reservoir quality reducing processes: compaction and quartz cementation. The model traces the evolution of porosity and permeability through time and uses sandstone composition and texture, and burial history data as input.

Because there is a lack of well control in frontier areas, reservoir composition and texture must be derived using sedimentological models and depositional analogues. The combination of the diagenetic model and Monte Carlo techniques allows us to incorporate uncertainties in the sedimentological interpretation and burial history models in probabilistic reservoir quality prediction. In addition, we can use sensitivity analysis to evaluate which uncertainties have the greatest impact on porosity and permeability prediction. To illustrate this approach for pre-drill reservoir quality prediction, we use examples from the Norwegian shelf where we have done extensive work on both simulating reservoir quality and tying the diagenetic model to facies and provenance models. Using this mature area as an example allows us to test the approach by comparing our predictions to measurements.

AAPG Search and Discovery Article #90914©2000 AAPG Annual Convention, New Orleans, Louisiana