Discrete Fracture Network generation from induced microseismicity data: a method based on 3D Hough transform.

Axelle Pochet and Francois Bonneau and Guillaume Caumon and Judith Sausse. ( 2013 )
in: Proc. 33rd Gocad Meeting, Nancy

Abstract

During hydraulic fracturing, micro-seismic data can be recorded and provide geometric information about some induced fractures. Indeed, micro-seismic events are located along fractures and particularly at their intersection. Assuming that several micro-seisms can belong to the same fracture, a 3D Hough transform method is therefore used to detect sub-coplanar points in the micro-seismic point cloud. Those points lay on planes which can be used to drive Discrete Fracture Network simulations. The usual angles φ and θ that describe planes in the 3D Hough transform method correspond here to fractures dip and azimuth. A Poisson Point Process is proposed in order to select planes in which fractures can be simulated. This probabilistic approach accounts for prior information about fracture size and orientation, hence permits to simulate fractures that are not necessarily underlined by the miscroseismicity. The method is applied to the data of the geothermal heat exchanger of Soultz-Sous-Forêts (France). Fractures implanted with this technique are used e directly to simulate Discrete Fracture Networks, or together with a classical stochastic Discrete Fracture Network simulation in order to have consistent simulation far from seismic events. Prior fractures orientation distribution inferred from borehole and regional characterization at Soultz-Sous-Forêts is well reproduced by the presented method, especially when applying the Poisson e Point Process when selecting planes where to implant fractures.

Download / Links

BibTeX Reference

@inproceedings{Pochet1GM2013,
 abstract = { During hydraulic fracturing, micro-seismic data can be recorded and provide geometric information about some induced fractures. Indeed, micro-seismic events are located along fractures and particularly at their intersection. Assuming that several micro-seisms can belong to the same fracture, a 3D Hough transform method is therefore used to detect sub-coplanar points in the micro-seismic point cloud. Those points lay on planes which can be used to drive Discrete Fracture Network simulations. The usual angles φ and θ that describe planes in the 3D Hough transform method correspond here to fractures dip and azimuth. A Poisson Point Process is proposed in order to select planes in which fractures can be simulated. This probabilistic approach accounts for prior information about fracture size and orientation, hence permits to simulate fractures that are not necessarily underlined by the miscroseismicity.
The method is applied to the data of the geothermal heat exchanger of Soultz-Sous-Forêts (France). Fractures implanted with this technique are used e directly to simulate Discrete Fracture Networks, or together with a classical stochastic Discrete Fracture Network simulation in order to have consistent simulation far from seismic events. Prior fractures orientation distribution inferred from borehole and regional characterization at Soultz-Sous-Forêts is well reproduced by the presented method, especially when applying the Poisson e Point Process when selecting planes where to implant fractures. },
 author = { Pochet, Axelle AND Bonneau, Francois AND Caumon, Guillaume AND Sausse, Judith },
 booktitle = { Proc. 33rd Gocad Meeting, Nancy },
 title = { Discrete Fracture Network generation from induced microseismicity data: a method based on 3D Hough transform. },
 year = { 2013 }
}