--> Training Image Characterization and Multipoint Statistical Modeling of Clastic and Carbonate Formations

2014 Rocky Mountain Section AAPG Annual Meeting

Datapages, Inc.Print this page

Training Image Characterization and Multipoint Statistical Modeling of Clastic and Carbonate Formations

Abstract

In partnership with the U.S. Department of Energy, the Energy & Environmental Research Center and the Plains CO2 Reduction Partnership Program have built several 3-D geocellular models for the purposes of studying CO2 storage and CO2 enhanced oil recovery (EOR). A training image (TI) library was established to represent a broad range of both clastic and carbonate formations in various structural settings. 3-D facies models were built using the TIs in a multipoint statistics (MPS) workflow to represent depositional patterns that correlate to geologic interpretation and features seen in modern analog environments. With project budget and time constraints, using the MPS workflow allowed multiple geocellular models to be built with realistic results and proved to be useful in creating facies models for field- to basin-sized projects with little or no data regarding facies distribution. Classic geological interpretation in the form of analogs, maps, well logs, and cross sections was used to create the TIs. The TIs were created from the geologic interpretations by “painting the image” in comparison to simulated techniques. The MPS workflow then uses the TI (in addition to soft data and hard data if applicable) to populate the facies property. Applying a similar workflow will improve characterization and reduce uncertainty in areas where limited data are available to assess CO2 storage and EOR potential. In addition, the workflow will increase project efficiency when multiple formations are being investigated.