Abstract

Despite seemingly good reservoir quality as seen by routine core and log analyses, certain gas reservoirs composed of aeolian dunes from Lower Permian age at around 5000m depth show poor well test results with respect to deliverability and connected volume.

Core analysis of those reservoirs routinely shows porosity values around 10–15% and permeabilities around 100 mD. Even standard measurements of plugs at overburden condition show a reduction of permeability by only 10%. Irreducible water saturation is often reached and has values between 10 to 20%.

Samples of such lithofacies are characterized by pronounced lamination caused by strong grain size variation. Detailed oriented (directional) permeability analyses under overburden conditions focused on such laminated samples have been carried out. It shows that thin bedding/pronounced lamination causes a reduction of bed-vertical permeability by a factor of 2 at ambient but by a factor of 100 and more under overburden conditions.

Capillary pressure curves measured at overburden conditions show a strong increase in irreducible water saturation for densely laminated sections when compared to ambient conditions.

The effective permeability is therefore seen as a function of compaction, lamination count and orientation of sediment bodies in relation to the borehole.

Core data, together with relative sediment dip information from image logs was used to model and successfully predict well test results. The densely laminated sections are too fine to be observed by logging tools directly. A simple porosity threshold can be used, whereas below that threshold such a dense lamination is assumed. This threshold has to be established by calibration to core data and is not uniform. NMR data can be used to detect the high bound water signal from compacted thin laminations.

This logging method was successfully used to define the net reservoir section within a well. The exposed reservoir section was optimized by discriminating "net reservoir" based on the abundance of laminas. Incorporation of a lamination mechanism into a reservoir model yields an improved prediction.

This content is only available via PDF.
You can access this article if you purchase or spend a download.