A new computer code for discrete fracture network modelling☆
Introduction
A rock mass comprises rock material and discontinuities that include pores, fractures, joints, faults and bedding planes. In most engineering applications, these discontinuities are the critical factors that determine the performance of the rock, such as the strength of rock structures or the flow characteristics of rock masses. For example, rock fracture networks are the most significant means of fluid transport through rock masses, especially those at significant depths in the subsurface of the Earth. Applications include hot dry rock (HDR) geothermal energy systems in which artificial reservoirs must be created by stimulating fractures to enable geothermal flow; the evaluation and construction of underground repositories for the safe storage and disposal of hazardous wastes for which potential contaminant flow through surrounding natural fractures must be quantified; underground water transport through aquifers in hydrogeological engineering; and movement of oil and gas in hydrocarbon reservoirs. Rock fractures and fracture networks are also the major decisive factors in the stability of rock slopes and underground openings in civil and mining engineering and in block caving mining applications.
Despite the importance of the applications, it is not feasible to map accurately on an engineering scale the fractures and fracture networks in rock masses not least because accurate field measurement of a single discontinuity is difficult and measurement of all discontinuities is impossible. In practical applications there is, therefore, no observable reality on any meaningful scale and the only realistic approach is via a stochastic model informed by sparse data, which normally come from surveys of analogues, such as rock outcrops, or from direct or indirect observations of the rock mass such as drill cores, borehole imaging, geophysical surveys or seismic monitoring during fracture stimulation.
This paper describes software that can be used to generate 2D and 3D fractures and fracture networks and is based on the authors’ previous work in stochastic modelling of fractures in rock masses (Xu et al., 2003, Xu et al., 2006, Mardia et al., 2007, Dowd et al., 2007). The purpose is to provide the research community with a simple, freely available comprehensive tool for stochastic fracture simulation. The software is generic in design and approach and output can be exported for subsequent flow or mechanical analysis. Some common sampling facilities such as plane, window, scanline and borehole sampling are included in the software. The program also includes basic statistical tools such as histogram and rose diagram analysis, probability plots and hemispherical projections.
Section snippets
General approach
The stochastic modelling of fracture networks originated in percolation studies (Robinson, 1983, Sahimi, 1993) and its wider application to rock engineering was promoted in the 1980s by the work of several research groups (e.g., Long et al., 1982; Baecher, 1983; Andersson et al., 1984; Dershowitz and Einstein, 1988). The general approach is to treat locations, persistence (size), orientation and other properties of the fractures as random variables with inferred probability distributions. To
Simulation of fracture locations
Four types of point process models are included in FracSim3D for modelling fracture locations: homogeneous (Poisson) model, non-homogeneous process, cluster process and Cox process.
Fracture generation
The points shown in Fig. 3 are the centroids of the fractures to be generated. In the MPP approach, size and orientations are treated as two marks assigned to the points which are simulated from their respective PDF. In FracSim3D, Monte Carlo simulations are used for this purpose.
It is fairly common practice to assume a negatively skewed distribution for the size of fracture traces from 2D fracture mapping surveys, with the most common being exponential, lognormal and gamma (Zhang and Einstein,
Simulation of additional fracture properties
Following the simulation of fracture geometries, additional fracture properties can be simulated and assigned to the generated fractures. This is done using Monte Carlo sampling of the PDF. The distributions available for this purpose in FracSim3D are uniform, normal, exponential, lognormal and non-parametric. The generated property values are included in the export file to be used for subsequent flow or mechanical analysis. There is an option for colour-coding the generated fractures according
Plane sampling
In FracSim3D a 3D fracture network can be sampled by a sampling plane defined by its dip direction and dip angle. A virtual plane is created by intersecting the 3D fracture system and the intersected fracture traces are displayed on the plane. An example is shown in Fig. 6c.
Window sampling
A rectangular sampling window can be defined for a 2D simulated fracture pattern or fracture traces on a sampling plane. Fracture trace lengths and orientations falling within the window can be analysed, as shown in the
Conclusions
This paper describes FracSim3D, a software package for the stochastic simulation of fractures in two and three dimensions. The fracture locations are generated by point process models while fracture size, orientation and other properties are generated by Monte Carlo sampling of their probability distribution functions.
Virtual sampling tools, comprising planes, windows, scanlines and boreholes, are included in the software. Basic statistics, such as hemispherical projections, histogram and rose
Acknowledgements
The graphical part of the software is programmed using the open source graphics library visualisation tool kit (VTK) (www.vtk.org).
The authors are grateful to Professor Stephen Priest for helpful comments and for encouraging them to publish this paper and to provide the accompanying software to the fracture modelling community.
References (36)
- et al.
Three-dimensional statistical modelling of a fractured rock mass – an example from the Fanay-Augères Mine
International Journal of Rock Mechanics and Mining Science
(1989) - et al.
A three-dimensional fracture network data set for a block of granite
International Journal of Rock Mechanics and Mining Sciences
(2009) - et al.
Inference of discontinuity trace length distributions using statistical graphical models
International Journal of Rock Mechanics and Mining Science
(2006) - et al.
Stochastic fracture geometry modeling in 3D including validations for a part of Arrowhead East Tunnel, California, USA
Engineering Geology
(2003) Derivation of parent fracture population statistics from trace length measurements of fractal fracture populations
International Journal of Rock Mechanics and Mining Science
(2002)- et al.
Fracture connectivity from fracture intersections in borehole image logs
Computers & Geosciences
(2003) A stereographical interpretation of joint trace data
International Journal of Rock Mechanics and Mining Science
(1980)- et al.
Estimating the intensity of rock discontinuities
International Journal of Rock Mechanics and Mining Science
(2000) - et al.
A stochastic model of fractured rock conditioned by measured information
Water Resources Research
(1984) Statistical analysis of rock mass fracturing
Mathematical Geology
(1983)
Fractal and geostatistical methods for modelling of a fracture network
Mathematical Geology
Characterizing rock joint geometry with joint system models
Rock Mechanics and Rock Engineering
Statistical Analysis of Spatial Point Patterns
A comparison of methods for the simulation of rock fractures
Mathematical Geology
Probabilistic and statistical methods in engineering geology
Rock Mechanics and Rock Engineering
Describing the geometry of 3D fracture systems by correcting for linear sampling bias
Mathematical Geology
Stereological analysis of fracture networks along cylindrical galleries
Mathematical Geology
Remarks on the spatial distribution of a reproducing population
Journal of Applied Probability
Cited by (224)
Characterization and reconstruction of rough fractures based on vector statistics
2024, Geoenergy Science and EngineeringRelationship between box-counting fractal dimension and properties of fracture networks
2024, Unconventional ResourcesA novel fractal-statistical scaling model of rocks considering strain rate
2023, Journal of Rock Mechanics and Geotechnical EngineeringAn equivalent line element approach for free surface flow through two-dimensional rock mass including fracture networks
2023, Applied Mathematics and ComputationFourier-based generation method of rough discrete fracture network
2023, International Journal of Rock Mechanics and Mining Sciences
- ☆
Code available from server at http://www.iamg.org/CGEditor/index.htm