We investigate the physical mechanisms governing the activation of faults during hydraulic fracturing. Recent studies have debated the varying importance of different fault reactivation mechanisms in different settings. Pore pressure increase caused by injection is generally considered to be the primary driver of induced seismicity. However, in very tight reservoir rocks, unless a fracture network exists to act as a hydraulic conduit, the rate of diffusion may be too low to explain the spatio-temporal evolution of some microseismic sequences. Thus, elastic and poroelastic stress transfer and aseismic slip have been invoked to explain observations of events occurring beyond the expected distance of a reasonable diffusive front. In this study we use the high quality microseismic data acquired during hydraulic fracturing at the Preston New Road (PNR) wells, Lancashire, UK, to examine fault triggering mechanisms. Injection through both wells generated felt induced seismicity—an ML 1.6 during PNR-1z injection in 2018 and an ML 2.9 during PNR-2 in 2019—and the microseismic observations show that each operation activated different faults with different orientations. Previous studies have already shown that PNR-1z seismicity was triggered by a combination of both direct hydraulic effects and elastic stress transfer generated by hydraulic fracture opening. Here we perform a similar analysis of the PNR-2 seismicity, finding that the PNR-2 fault triggering was mostly likely dominated by the diffusion of increased fluid pressure through a secondary zone of hydraulic fractures. However, elastic stress transfer caused by hydraulic fracture opening would have also acted to promote slip. It is significant that no microseismicity was observed on the previously activated fault during PNR-2 operations. This dataset therefore provides a unique opportunity to estimate the minimum perturbation required to activate the fault. As it appears that there was no hydraulic connection between them during each stimulation, any perturbation caused to the PNR-1z fault by PNR-2 stimulation must be through elastic or poroelastic stress transfer. As such, by computing the stress transfer created by PNR-2 stimulation onto the PNR-1z fault, we are able to approximate the minimum bound for the required stress perturbation: in excess of 0.1 MPa, orders of magnitude larger than stated estimates of a generalized triggering threshold.
Bibliographical noteFunding Information:
TK is supported by the NERC UKUH Challenge Grants SHAPE-UK project (Grant Number NE/R018006/1). JV contribution to this work is supported by NERC (Grant Number NE/R018162/1). The authors would also like to thank the UK Oil and Gas Authority for partial funding of this research.
We would like to acknowledge Cuadrilla Resources Limited, the operator of the PNR site, and Schlumberger Ltd., who conducted the processing of the microseismic data that are presented in this work, and Nanometrics Ltd., who conducted the processing of the surface derived data acquired by Nanometrics Ltd. and the British Geological Survey. We would like to thank Michael Kendall for insightful discussion of the results, as well as Ryan Schultz and Louis De Barros for their constructive feedback. This work was a product of the Bristol University Microseismicity Projects (BUMPs)?a research consortium whose sponsors include several hydrocarbon operators and service providers.
© Copyright © 2021 Kettlety and Verdon.
- induced seismicity
- stress modeling
- hydraulic fracturing