The Effect of Geologic Parameters and Uncertainties on Subsurface Flow: Deepwater Depositional Systems

Abstract

The application of reservoir simulation as a tool for reservoir development and management is widespread in the oil and gas industry. Moreover, it is recognized that the results of any reservoir simulation model are strongly influenced by the underlying geologic model. However, the direct relationship between geologic parameters and subsurface flow is obscure. In this paper we explore this relationship in a deepwater depositional system using data from two reservoir analogs: the shallow seismic dataset from the Mahakam Fan and outcrop data from the Brushy Canyon Formation of West Texas. Shallow seismic data from the Mahakam Fan area shows a high-resolution deepwater channel-levee system consisting of 10 migrating channels. Using an experimental design framework and a series of three increasingly complex models, we investigated the effect of nine different geologic factors on several different measures of the flow behavior. Our results show that, as expected, different geologic factors influence different measures of the flow. Most significant is the clear effect that the proportion and organization of the different internal facies making up the channels have on the recovery factor and net oil production. The Brushy Canyon outcrops used in this work represent sand-rich proximal deposits of a distributary lobe complex. Here we built models on a very small length scale to investigate the effects of sheet-like reservoir architecture as well as internal facies distribution of the sheets on subsurface flow. Again, an experimental design framework was employed, this time to examine the influence of 11 input variables. The proportion and organization of the internal lobe facies has a significant influence on the subsurface flow here but in these distributary lobe complexes other variables, including the stacking of the lobes, were also found to be important. The models in this study address flow behavior in deepwater, sparse well environments. Using models from the simple to the complex, we found that several parameters incorporated in the complex models, and not in simple models, had a significant impact on the predicted flow.

Effective Parametrization for Reliable Reservoir Performance Predictions

“…It also has been observed that seismically conditioned, geostatistical models of deep-water reservoirs were unable to match reservoir behavior observed in 4D seismic or production data. A parallel study in [19] also showed the importance of modeling the organization and architecture of reservoir facies in additional to their proportions. On the other hand, the significance of one geologic feature may vary depending on the measure of subsurface flow.…”

“…It remains to be seen if the compression-based methods can maintain the spatial organization of facies distributions, which can be important to fluid flow [5,19]. Thus far, the conceptual models used to demonstrate these methods are rather simplified.…”

A study of conceptual model uncertainty in large‐scale CO2 storage simulation

“…The number of units in the upscaled models is used as a proxy for model complexity, since data requirement and associated cost of building the model will likely increase with the level of heterogeneity resolution. For example, the FHRM is considered the most complex and expensive to build, followed by the facies model (less expensive, since potentially petrophysical properties of the facies can be borrowed from analog data [ Milliken et al , 2007]), which is followed by the layered model (e.g., thickness‐based division), and the formation model. Note, however, to develop insights into model complexity, this study is conceptual, whereby the FHRM provides a basis to create the simpler models.…”

Stratigraphic and structural compartmentalization of dryland fluvial reservoirs: Triassic Heron Cluster, Central North Sea

“…Ainsworth 2005;Howell et al 2008), but in general, attention has focused on stratigraphic compartmentalization by flooding surfaces in shallow marine and paralic systems (Martinsen 1994;Ainsworth 2005), lobe connectivity in deepwater systems (e.g. Edman & Burk 1998;Milliken et al 2008) and connectivity of channelized sandbodies in fluvial and deepwater settings (Larue & Hovadik 2006). Recent connectivity and modelling studies in these settings have also been used to test the interaction between faults and stratigraphy on fluid flow and compartmentalization (e.g.…”