Investigating the Impact of Different Reservoir Property Modeling Algorithms and Their Associated Uncertainties on Volume Estimation (Gullfaks Field, North Sea)

Abstract

Reporting reliable results for hydrocarbon volume estimation is important for both economic analyses and making key decisions in reservoir management and development. Adequate facies and petrophysical modeling of static reservoir properties are key inputs for the derivation of a robust static reservoir model from which static volume is computed and inherent uncertainties are quantified. However, the choice of geostatistical algorithm for building the model depend on development and production maturity, degree of reservoir heterogeneity and the type, quality and amount of data. This study therefore aims at investigating the impact of the combination of stochastic and deterministic methods of property modeling on volume estimation and also perform uncertainty and sensitivity analyses to quantify uncertainties so as to aid exploration and production decision making process. Facies model were simulated/generated using both stochastic and deterministic algorithms. The resultant facies model formed an input for the petrophysical modeling process also using both stochastic and deterministic algorithms. For each combination, hydrocarbon pore volume was computed. Monte Carlo Simulation method was used to perform the uncertainty analysis where the low case (P10), mid case (P50) and high case (P90) was outputted.

The results show that a combination of Sequential Indicator Simulation (facies) with Sequential Gaussian Simulation (petrophysical) captured a large range of hydrocarbon pore volume for the twenty equiprobable realizations simulated while the combination of Truncated Gaussian Simulation with trend and Gaussian Random Function Simulation gave a limited range. A combination of the deterministic algorithm gave a single estimated and more pessimistic volume. Uncertainty analysis indicated that the facies modeling process and the combination of SIS_SGS algorithm have a higher impact on volumetrics.