Uncertainty Quantification of 3D Stochastic Fault Modeling in Structurally Complex Reservoirs

Abstract Faults play a pivotal role in the characterization of oil and gas reservoirs. Especially in heavily faulted reservoirs, faults have significant effects on the estimation of reservoir volumes since they define the boundaries of the reservoir and compartmentalize the reservoirs into distinct fault-bounded blocks. This paper presents a method utilizing stochastic simulation of the 3D faults modeling to quantify the uncertainties and minimize their impacts. Fault interpretation uncertainties are represented by creating fault uncertainty envelope volumes that encompass both the footwall and hanging wall sides of the "most likely" base-case fault surfaces. At all interpreted fault points, lateral and vertical uncertainty are defined based on the quality of seismic data and interpretation confidence level. 3D fault uncertainties modeling involves generating hundreds of 3D stochastic fault models to quantify uncertainty. Key fault geometrical parameters that are perturbed include fault lateral position, fault dip, fault strike and fault throw. Inside the fault uncertainty envelope volumes, these fault uncertainty parameters are stochastically sampled using a triangular probability distribution. 3D uncertainties modeling extends beyond faults to horizons, fluid contacts, and the entire model, ensuring preservation of the 3D geological model's integrity. The geometry of the 3D geological grid is updated concurrently with the re-distribution of reservoir properties onto the 3D grid. Moving faults and horizons obviously alters the reservoir volumes. Consequently, bulk and hydrocarbon volumes above fluids contacts are recalculated for each simulation run. Through the uncertainty runs, the asset team was able to measure the effect and sensitivity of individual fault parameter interpretations on reservoir volumes. By ranking hundreds of structural uncertainty scenarios based on reservoir volume, the corresponding 3D geomodels linked to P10, P50, and P90 values were identified and prepared for submission to dynamic flow simulation studies. This approach moves beyond conventional methods that assume simple scalar uncertainties away from wells control and deterministic faults, providing a comprehensive representation of 3D structural framework uncertainties. It enables the quantification of the impact and sensitivity of faults parameters on reservoir volumes within heavily fault-bounded reservoirs.