Using Outcrop Analogues to Improve Fault Seal Workflows

Abstract

Predicting the hydraulic properties of fault zones has long been a problem for the exploration and production of oil and gas. Several tools exist that attempt to predict the potential impact of fault zones on the movement of fluids in the subsurface. However, reliable predictions are still elusive. To improve the prediction of fault properties this paper analyses data on the structure and contents of 10 faults in the Colorado Plateau (SE Utah).

In this study we compare observations on faults in outcrops to predictions by commonly used fault seal evaluation tools. Outcrops of fault zones have been mapped in centimetre scale detail. By inferring which faults are likely sealing and which are likely non-sealing, we can compare these with predictions by commonly used fault sealing workflows. We compare predictions from SGR, ESG, SSF and CSP to the mapped fault zones. The comparison shows that for the faults in our dataset CSP is the most reliable predictor, correctly distinguishing between sealing and non-sealing faults for 8 out of 10 faults. CSP evaluates the combined effect of smearing of multiple beds of shale. This corresponds well to the architecture we can observe in the outcrops, where smeared shale and silt forms the dominant low permeability fault rock.

In addition to fault sealing we can use the dataset to estimate bulk permeability values for the fault zones. We compare these bulk permeabilities to the estimates of fault permeability provided by established SGR-based workflows. The comparison shows that the two different approaches yield very different results and the data shows no predictive relationship between the outcrop observations and the SGR-based permeability predictions.

The difficulty in predicting fault permeability suggests that more robust tools are required. To reliably evaluate fault zone permeability it is necessary to reliably evaluate fault architecture. We show that fault architectures are the result of a consistent set of geological processes (e.g. shale smearing, formation of sandstone lenses. By evaluating the likeliness of these geological processes, we can estimate the fault architecture most likely to be present at the reservoir interval. The predicted fault architectures can subsequently be used to provide robust upscaled permeability estimates and an estimate of the inherent uncertainty in the prediction.

Acknowledgements

The authors would like to gratefully thank Total for funding this research.

AAPG Datapages/Search and Discovery Article #90323 ©2018 AAPG Annual Convention and Exhibition, Salt Lake City, Utah, May 20-23, 2018