In this invited talk at the Annual AAPG Meeting, Geocosm’s Rob Lander and Linda Bonnell review the history and current state-of-the-art in reservoir quality prediction models while also discussing potential future developments.
Over the past 20 years reservoir quality assessment has expanded from a focus on description to encompass process-oriented models. These models predict rock properties in undrilled locations and reconstruct properties through geologic time. Augmenting the discipline with models has reinforced, and not detracted from, the importance of rock characterization given that such data are essential for constraining parameters and testing performance. Existing reservoir quality prediction models build on breakthroughs in process understanding developed largely in the 1990s. These new concepts developed in response to integration of rigorous petrographic characterization, burial history modeling, fluid inclusion analysis, cathodoluminescence imaging, experimental studies, and creative thinking that considered knowledge from allied disciplines.
In addition to compaction and quartz cementation, which were considered in the first generation of models that date from the late 1990s, existing systems simulate a number of other diagenetic processes such as plagioclase albitization, the reaction of kaolinite to form illite, and chlorite formation from volcanic rock fragment alteration. Although models are lacking for other processes, current systems provide accurate predictions for a broad range of sandstone types and geologic settings. Additionally, these diagenetic models have been coupled with rock property models that predict permeability and seismic velocities and may be linked with 3D depositional / stratigraphic and petroleum system models.
We expect that future reservoir quality models will explicitly simulate the micron-scale geometry of sandstones and mudrocks in 3D and will produce compositions and textures that are directly comparable to high resolution 3D images of rocks. These geometric results, like those derived from 3D imaging, could serve as input for “digital physics” simulations of a broad range in rock properties. They also would have the unique ability to provide these rock property predictions for locations or geologic times where no samples are available. A broader range of diagenetic processes including carbonate reactions could be considered by integrating reservoir quality models with reaction-transport methods. Finally, coupling of diagenetic and chemo/mechanical models could improve our ability to predict the properties of faults, deformation bands, and fractures and the impact that they have on fluid-flow and geomechanical behavior.