FIELD STUDIES


Montpellier University has a strong tradition of field studies and subsequent mechanical analysis. Within Geo-FracNet we aim to study brittle structures which have a considerable impact in terms of reservoir permeability and fault transmissivity. A database of key examples of the various types of structures, combined with corresponding physical models, will allow construction of a classification scheme that will help to constrain the dominant controlling mechanical parameters.

The establishment of tectono-mechanical rules based on that fieldwork will allow the incorporation of sub-seismic scale structures into advanced 3D reservoir models.
 
 

Fold related  fractures in southern Anti-Atlas, Morocco

 
 

BOREHOLE ANALYSIS AND INTEGRATED CORE DESCRIPTION


We have developed an original approach to the analysis of structural heterogeneities in which experimental data are combined with field, borehole and core analysis to provide an improved interpretation. The main improvements are the following:

 By considering analogue experimental data (particularly photoelastic) in combination with borehole fracture data, it is possible to provide additional constraints to elaborate the geometric model of the fracture network at reservoir scale.  
From raw data...
 
Rose diagram for fracture orientation in an horizontal well (from FMS data)
...to data analysis
 
Rose diagrams showing fracture strikes (from FMS) along the well bore in successive beds.
...and fracture pattern model.
 
 
Extrapolation of the fracture orientations on the reservoir zone around an horizontal well by superimposition of interpreted fracture directions (FMS) and modelled stress trajectories (photoelastic model). The two sets (orange and blue) correspond to two separate fracturing events.
 


EXPERIMENTAL STUDIES AND ANALOGUE MODELLING

Brittle Coatings

Brittle coatings have been used in material sciences to study localised deformation for engineering purposes. Brittle coatings behave like a competent layer overlying a less competent one, and the cracks developed in the varnish when the substrate is deformed  are analogous to joints. In particular, they are slightly open, very long with respect to their width and have been shown to grow perpendicular to the minimum principal stress. In Montpellier we adapted the brittle varnish technique in order to study and interpret joint patterns. The evolving fractures in the brittle varnish analogue experiment can interact in a similar way to natural fractures evolving in a geological system. These observed fracture interactions are extremely useful in deciphering the mechanism of jointing. For example, brittle varnish experiments successfully reproduced the complex joint geometries associated with non-cylindrical folding in the Lias of the Somerset Bristol Channel coast, U.K. Concepts resulting from these experiments are now used to generate joints and small scale fractures in reservoir modelling software such as FRAME (Elf EP, Shell International.) or  FRACA (IFP).
Complex natural joint pattern in the 
Liassic sandstones of Lavernock Point (Wales, U.K.).
 Fracture pattern induced in brittle coating  after non-cylindrical folding of the sample.
More generally, the brittle varnish technique has been able to reconstruct and mechanically constrain all of the naturally observed 2D geometries of interacting fractures. Such interactions are a fundamental control on the connectivity of fracture networks within reservoirs.
Aerial photograph of joint pattern in Arches National Park (U.S.A.).
Experimental fracture pattern in Brittle varnish 
obtained by progressive stress field rotation.
A closely related experimental technique consists of using a collophony (violin bow resin) coating on an aluminium substrate. This variation of the brittle coating technique permits accurate modelling of the deviation of fractures around discontinuities, such as asperities on faults, and provides an understanding of the mechanical conditions occurring during joint development. The fracture patterns developed give insights into the perturbed stress trajectories around faults. Recent seismic studies (AVO on shear waves) have revealed open fractures following comparable stress fields at depth. Such integrated analyses of stress/open fracture patterns may be of interest in improving hydrocarbon recovery in the vicinity of faults.
Photoelasticity

Photoelasticity utilises the ability of some transparent materials to exhibit temporary double refraction (analogous to birefringence in crystals) under an applied load. Analysing loaded samples between crossed-polarisers enables the direct derivation of the stress trajectories and differential stress (from isochromatics) and indirect derivation of the individual stress magnitudes. These properties can be used to analyse the distribution of stress around fault models of imposed geometries in polymer plates. This stress determination can also be done using numerical modelling, but photoelasticity allows the analysis of complex and subtle geometries, such as asperities on real fractures, and permits propagation of the faults. This technique can be used to interpret perturbated joint patterns; to extrapolate fracture distributions from borehole fracture data and seismic fault patterns; to study stress concentrations in fault systems and to predict intra-reservoir hydrodynamic behaviour.
Photoelastic fringes associated with vertically loaded overlapping cracks forming a dilational jog
Potential fluid flow pathways predicted for a dilational jog from photoelastic data. Contours are of mean stress; arrows indicate the predicted fluid flow paths (length proportional to gradient).
 
Fracturing in Polymers

Using transparent polymers as rock analogues has been a fruitful way of studying fundamental and applied aspects of fault rupture. For example, brittle-elastic polymers such as Plexiglas (PMMA) are very useful for analysing 2D aspects of mode I and mode II fault propagation, since the same fundamental processes are observed in both rock and PMMA specimens.
Ductile polymers, which develop shear bands from pre-existing defects, can be used to model complex fault propagation geometries in plastic rocks (e.g. clays, marls).
Experiments using polymers adapted to geological conditions are extremely useful for interfacing between the geosciences and material sciences (especially fractures mechanics, tribology). They have the advantage of simple design and instrumentation together with dynamic stress and strain analysis. 3D propagation processes can be studied in adapted weak transparent organic materials (under development).
Branch cracks and stylolites at the tip of reactivated fractures segments in limestone - Pyrenean compression. (Languedoc, Southern France)
Branch cracks and en echelon cracks in a PMMA plate with an open defect submitted to an horizontal  uniaxial compression.
 
3D Fracturing in Multilayer Systems (under development)

 
An ideal model would be a quantitative (experimental or numerical) model of a real geo-object incorporating heterogeneities such as layering and pre-existing structures. This model should be 3D and evolve under the boundary conditions constrained by geological data. Unfortunately, such a model is beyond the capabilities of either the currently available numerical techniques or analogue modelling using existing materials.
To be correctly scaled to the gravitational forces, a modelling material must have a low effective strength. Long term collaboration with chemists has provided us with a class of properly "scaled" ductile materials that have been used to quantitatively model various geodynamic processes (see examples on Alexandre Chemenda's home page) including formation of the faults.


In order to study fracturing-related processes, the modelling material must be brittle, and fail under Mode I, II and III conditions. Such materials have been recently created. They allowed us to develop a new type of mono- and multi-layer models and to obtain encouragingly realistic 3D fracture networks.
Jointing of a mono-layer model under bi-axial compression:cross-sections of two models with different thickness and the same loading conditions 



 
 

Elasto-brittle instability and formation of fracture corridors


 
 

Dynamic fracturing during displacement along a pre-existing thrust fault
Photo and drawing of the fault foot-wall





The 3D fracture architecture and connectivity of the models can be described digitally by post-mortem examination of models that have been injected with a marker dye. Furthermore the models should permit true hydrodynamic tests.
 
 

Multi-scale fracturing in layered model

Further development of this innovative modelling using materials with controllable mechanical properties is a major aim of Geo-FracNet. This new program of multilayer physical experiments adapted to specific geological problems can be the basis for identification of the key parameters defining the tectono-mechanical rules of fracture system development.


NUMERICAL MODELLING

Existing numerical codes allow exhaustive analysis of stress-strain state in the linear elastic mediums and to model some relatively simple cases of brittle fracture propagation.
Numerical model (FRANC2D): propagation of fracture cluster in a three-layer elastic model





Such a numerical analysis is very useful, especially for understanding and interpretation of the experimental results. However, its application to the complex natural abject is quite limited, largely because of theoretical limitations in description (understanding) of the fracture mechanism especially in 3D. Purely elastic representation of real geo-media is also an oversimplification. Therefore when considering fracturing and faulting, we propose to investigate the potential of elasto-(visco)-plastic models (traditionally used for geodynamic modelling) to explore the conditions in which secondary faults develop and interact from pre-existing fractures. This could be extremely useful in understanding sub-seismic fault geometry and characteristics.
The key to numerical modelling of sub-seismic fracturing lies in the use of a continuum theory. Here, the dynamic propagation of a meso-fault, intermediate in scale between the seismic scale fault network (set as an initial boundary condition) and the micro-crack population (to be calculated), can be monitored. For this we analyse the sub-seismic faults by making use of the "smeared crack" approach. Using this approach, we can predict the evolution of the population density of mode I, II and III fractures at the micro-scale, and detect the propagation of the smeared crack at sub-seismic scale. This numerical method also supplies dynamic data on the selective development of instantaneous micro-mechanical crack surfaces in the damage zone as a function of feedback between the pre-existing seismic scale faults and the updated sub-seismic fault network.
One other important development is to integrate the results of 3D analogue experiments on multilayer models into numerical models in order to 1) explicitly represent the fracture network corresponding to specific structures and 2) use them in hydraulic modelling.


STATISTICS

The statistical analysis of fracture spacing data is now a standard technique in the analysis of fracture data. Unfortunately, the results of a purely statistical approach do not further our understanding of the mechanisms controlling fracture development. Geo-FracNet plan to integrate the results of classical statistical analyses of fracture data obtained from field studies with similar and unique data obtained from analogue experiments.

This is therefore a coupled approach in which the field data defines the characteristics of the fracture sets that must be predicted by the analogue experiments. The physical conditions and mechanisms necessary to form such experimental fracture sets can then provide the basis for advancing our understanding of the underlying mechanical processes.