Petrophysical analysis of conventional clastic reservoirs typically focuses on reservoir intervals with scant attention paid to most of the footage that is written off as so-called shale. While prioritizing potential reservoir intervals is understandable, thoughtful consideration of non-reservoir is worthwhile. This paper discusses the value of detailed analysis of shale intervals and introduces a practical scheme for delivering a frame by frame petrophysical analysis from top to bottom of the wells, recognizing that formation properties of interest to the subsurface community differ in sands and shales, so each lithology has its particular evaluation priorities.The underlying principle, that the petrophysical model should address what we need to know about the subsurface, is illustrated by reference to a six well data set from a deep water turbidite environment offshore West Africa. Three main lithology types are present, sand, shale and carbonate, each of which merited its suite of petrophysical models, spliced together via lithology-driven model selection logic.Most of the footage was shale with significant bulk, mineral and textural property variation about regional depth-based compaction trends. Analysis showed that shale at the hard end of the property range was associated with the sands.The range in shale bulk properties correlated with variations in clay mineralogy, chemistry and resistivity anisotropy. Results supported stratigraphic interpretation and illustrated the potential for vertical property variations to cause seismic reflections that could be misinterpreted as sand-shale boundaries.About 20% of the footage was massive feldspathic sand intervals and alternating sand and shale beds unresolved by the logs. We quantified net reservoir footage and the petrophysical properties of the sandy intervals using a generalized variant of the Thomas-Stieber laminated petrophysical model that additionally estimated the significant K-feldspar fraction. Laminated shale in the sands was considered synonymous with the hard end of the property range in the shale intervals, retaining continuity between the sand and shale-focused models after model selection. The carbonate beds were thin, and their chemistry uncertain, so our goal was simply to identify them from their density departure from the regional shale compaction trend.Logging suites varied in comprehensiveness from LWD quad combo to wireline quad combo with NMR, spectral gamma and triaxial induction logs, and data quality varied from good to questionable, so relations observed in data rich intervals were transformed to external inputs guiding the analysis of data poor intervals. Replacing absent logs with locally derived data-driven relations enabled us to deliver continuous volume fractions of sand, K-feldspar, the three shale textures-laminated, dispersed and structural-associated with sand, two shale types in shale intervals and a fluid analysis.The single spliced result simplified data management while addressing the needs of a diverse stakeholder community. The work presented here was executed using commercial software tools on the desktop of most petrophysicists. While space restricts us to one case study, the principles are transferable and have been demonstrated in other clastic reservoir environments.
展开▼
