Gryaznov, Andrey (Baker Hughes) | Paludan, Johanne (Baker Hughes) | Bizeray, Morgane (Baker Hughes) | El Menshawy, Ali (Baker Hughes) | Balamaga, Julius (Baker Hughes) | Embry, Jean-Michel (Baker Hughes) | Burns, Chris (Baker Hughes) | Fomin, Roman (Gazprom International) | Aleksakhin, Yuriy (Gazprom International)
A major oilfield services provider was requested by a Russian national oil company to conduct a study of a tight, naturally fractured reservoir in Algeria. The goal was to integrate multiple data types (borehole images, wireline acoustic data, 360-degree core photographs) to generate a representative set of 3D static models describing the natural fracture network (with optimistic, basic and pessimistic cases) and then define fracture permeability. The reservoir is a tight Ordovician sandstone with intensive faulting, a complex facies pattern, and limited well data—all presenting significant challenges.
Fracture interpretation was integrated from different sources, including borehole images, cross multipole acoustic data and 360-degree core photographs. The integration of fracture data from so many sources based on different physical principles enabled fracture modelling with much higher confidence, providing an input for further field development.
The workflow for fracture density determination is divided into several stages: from borehole imaging (including definition of open, mixed and closed fracture types) and acoustic data fracture interpretation to 3D fracture density trend creation and calibration. Image interpretation results show good correlation to acoustic log interpretation results using Stoneley reflectivity and azimuthal anisotropy analysis.
Combining acoustic, core and image logs data allowed organization of the wellbores into fracture classes. Fracture classes are zones of probability about the presence of natural fractures. These classes vary from very high probability (where all data types show presence of fractures) to zero probability (where all data types show no fractures or anisotropy).
The 3D model of fracture density reflects a basic concept: the fracture density decreases away from fault cores, and within the fault cores the fracture density is at a maximum. This observation was supported by many field analogues (including some in Algeria).
There were many intervals of intensive natural fracturing that were identified from images and core photographs. These zones might have contributed significantly to fracture permeability. This idea is supported by well test data analysis: the effective permeability from well tests significantly exceeds the matrix (core) permeability but is within the range of fracture permeability as defined by the continuous fracture network (CFN) modelling.
Various data sets were integrated and calibrated to enable precise identification of fracture density distribution, fracture classes, and dip angle and aperture of natural fractures. These data sets provided input for fracture permeability calculations. Fracture density, fracture aperture and fracture dip angle 3D grids were prepared. Special equations, developed for tight, fractured Algerian sandstones, were applied to calculate fracture properties, e.g., fracture permeability.
Through the close interaction of a multi-disciplinary team it was possible to successfully build a consistent 3D CFN model and to perform fracture uncertainty analysis to determine a variety of high-, mid- and low-fractured permeability cases. This model also supported further 1D and 3D geomechanical modelling studies. This CFN model provided a rapid workflow and robust model for further field development planning and new well placement.