Sougata Halder, James Keay, Matt Mayer, TGS, Houston; Phil Hargreaves, TGS, Woking; John Xie, Meridee Fockler, Sungho Hong, CDL, Calgary
Carbon sequestration in depleted hydrocarbon reservoirs is attractive since withdrawal of fossil fuels can be balanced by CO₂ storage in the subsurface. This presentation discusses the data, approach, results, and recommendations in development of a large regional database of estimated volumetrics for storage potential of injected carbon dioxide (CO₂), or for an enhanced oil recovery (EOR) in depleted oil and gas reservoirs.
The assessed area extends across heavily explored petroleum producing regions of southern Texas, Louisiana, and Mississippi. The area lies within the heavily industrialized region in the southern US and the US Gulf Coast where large numbers of industrial CO₂ emission sources exist. The US Department of Energy has assessed the costs of CO₂ storage in this region as low, and therefore the southern US may be an attractive area for carbon storage.
Data available for this study includes well performance and reservoir fluid and pressure data from public domain, and proprietary geologic and temperature models. The purpose of this study is to supplement existing regional studies with a detailed consistent dataset useful for reference and screening at the pool scale. The dataset of storage opportunities is presented in GIS format and visualized/referenced to industrial sites, emitters, pipelines, cultural and other spatial layers for context.
Development of the database involved analysis of cumulative and forecast production, and completion records in >150000 wells. Three major workflows are discussed: 1) data sources and cleanup, 2) assignment of production to consistent zones, 3) calculation of volumetrics and Monte Carlo simulation of storage capacity estimates of depleted pools.
Key to the project was assignment of production from well intervals to a standardized set of formations as a process of delineating producing pools and calculating cumulative production from each pool. The producing well intervals from public records required allocation to new standardized formation names. Semi-automated workflows involved depth filtering at the database level and application of a GIS buffering function to generate pool outlines. The process was quality controlled using desktop geoscience interpretation software to visualize the location of perforated zones in wells with gridded stratigraphic surfaces.
The resulting database features storage resources assigned to 27 formation levels. The term ‘storage units’ was introduced to rename individual pools as well as define groupings of laterally contiguous pools in the same stratigraphic zone that may be in communication during CO₂ injection. Monte Carlo simulation proved to be a versatile volumetrics approach as it allowed weighting of various parameters and confidence, providing P10, P50, and P90 ranges of storage outputs. Additional attributes, such as the number of well penetrations in each storage unit, were also calculated as input for further risk analysis.
Project challenges were fundamentally around data standardization and accessing sufficient reservoir attributes to perform consistent calculations across a large area. Reporting inconsistencies of public well records required significant quality control, and accurate reservoir attributes are rare in public data, therefore unique workflows were developed to enable a successful result.