An Innovative Approach to Uncertainty Assessment of Hydrocarbon Estimates in Fluvial Systems: A Case Study

Abstract

Abstract A novel method was developed and applied to a fluvial reservoir for the uncertainty assessment of hydrocarbon estimates related to geological input. Uncertainty assessment is a crucial step in any volumetric study, because it exists in all variables used to estimate hydrocarbons originally in place. Conventional volumetric estimation methods such as Monte Carlo simulation and curve decline methods do not consider the uncertainties in depositional facies proportion. Monte Carlo simulation is a powerful tool in measuring uncertainty in hydrocarbon estimates. Monte Carlo models can be easily built and run extremely fast. The output results can be examined visually as well as statistically. However, the main drawback is the absence of any geological knowledge in the simulation process. The geological knowledge comes from different sources such as drilled cores, wireline well logs, facies maps, modern and ancient analogues and most important of all it grows with the geologists experience. This knowledge is currently employed in the stochastic modeling process only. The new method presented here merges stochastic facies models with the Monte Carlo Simulation method to produce geological driven volumetrics (GDV). The stochastic facies models are built based on different facies proportions associated with their respective uncertainties. Facies volumes were extracted from the different stochastic facies models and were used as a geological input into the Monte Carol simulation algorithm to generate GDV estimates. Using the GDV method gave hydrocarbon estimates that were higher in volume and had less uncertainty compared to the Monte Carlo Simulation and the stochastic modelling method. The Stochastic modelling method is much more time consuming, for uncertainty assessment because it requires running hundreds of stochastic realisations, while the layout of the GDV is equally as efficient as a Monte Carlo Simulation.