Speaker
Description
Reservoir modeling and simulation play a fundamental role throughout the entire reservoir exploitation chain, as they are essential for decision-making, reserve evaluation, and the development of field abandonment plans. One of the most important properties influencing the outcome of a reservoir simulation is relative permeability.
Relative permeability can be obtained through laboratory experiments, such as displacement and centrifugation tests, or through tools related to Digital Rock. A common practice in the industry is to perform tests under unsteady conditions and recover relative permeability using numerical simulators through history matching. However, history matching is prone to uncertainties and non-uniqueness issues. Therefore, it is crucial to develop methods that consider laboratory uncertainties and their impact on the obtained relative permeability curves. Recently, Bayesian inference and stochastic methods have been applied to estimate relative permeability, considering the modeling and parameterization errors mentioned.
In this work, approximately 300 samples that underwent relative permeability tests under unsteady conditions were first classified according to their main features identified in micro-CT images—heterogeneity, laminations, presence of preferential pathways, presence of vugs, and presence of barite—and the tests performed—was there a pressure increase during the test? Is the pressure stable? A thorough quality control of the results was then carried out.
After that, the relative permeability results under unsteady conditions were reprocessed using an MCMC algorithm. These analyses provide uncertainty intervals of the relative permeability curves used in flow modeling. This additional perspective allows for a better clarification of the question regarding under which conditions and protocols we can expect to determine relative permeability curves from a transient experiment, and when such attempts are associated with unacceptably large uncertainty intervals.
Through this analysis, clearer controls of the porous medium over the relative permeability results were identified, such as the influence of heterogeneity on water fractional flow curves, absolute permeability on the relative permeability curves of spherulites, and laminations in grainstones.
The confidence intervals of the main features of the relative permeability curves, such as Swi and Sor and curve parameters, could then be used, after this analysis, to better inform petrophysical well analyses and reservoir simulations in a way that is more suitable for scenario analysis. This work also shows that many of these features are highly correlated, which is often ignored in scenario analyses but has a significant impact on the obtained fractional flows.
| Country | Brazil |
|---|---|
| Acceptance of the Terms & Conditions | Click here to agree |
