Unconventional reservoirs are inherently heterogeneous, often possessing complex mineralogy. As such, their characterization must begin with the accurate estimation of rock component volumes. Multi-mineral inversion-based workflows solve tool-response equations to calculate depth-by-depth component volumes based on log-specific end-members. These end-members are poorly constrained and may vary with depositional environment, provenance, rock fabric, logging tool design and physics. Existing multi-mineral workflows rely on predefined end-members that are manually optimized until interpreted mineral component volumes are consistent with core measurements. This manually intensive workflow is poorly defined, highly iterative, and generally limited in geologic scope and scale.
The proposed solution to develop a more robust and predictive multi-mineral and fluid interpretation workflow is comprised of two parts: (A) a scalable multi-well calibration step in which the well-log end-members are predicted for individual logs using core mineral volumes and porosity data across multiple wells and intervals; and (B) an interpretation step in which the calibrated end-members (part A) are used to solve the specific tool-response equations at each depth. Any combination of logs can thus be included in the interpretation step (part B) as long as a calibration exists. The inversion workflow captures the full range of component volumes that honor the input log measurements (uncertainty) which is a critical element because multi-mineral inversion problems are generally under-determined.
The workflow was benchmarked against core data, NMR, and spectroscopy-based mineral and porosity volumes using quad-combo log measurements in dozens of Permian wells. Compared to conventional workflows, this novel approach to multi-mineral interpretation is faster, more accurate, and more robust, benefiting from the systematic optimization of the end-members.
Reliable characterization of unconventional reservoir rocks is critical to the accurate identification of productive intervals, estimation of resource density and production capacity, and the development of geologic, reservoir, and hydraulic fracture models. By systematically optimizing component end-members against core measurements, a more robust and effective rock component volume may be delivered for reservoirs with complex mineralogy.
