What are the best methods to identify and characterize fracture networks in different reservoir types?

1 Fracture identification methods

Fracture identification methods are techniques that can detect the presence, orientation, and density of fractures in the subsurface. They can be classified into direct and indirect methods. Direct methods are based on observing the fractures directly from cores, outcrops, borehole images, or seismic data. Indirect methods are based on inferring the fractures from other measurements, such as well logs, pressure tests, tracer tests, or production data. Both methods have their advantages and limitations, depending on the fracture size, spacing, connectivity, and contrast with the matrix.

2 Fracture characterization methods

Fracture characterization methods are techniques that can quantify the geometry, distribution, and properties of fractures in the reservoir. They can be classified into static and dynamic methods. Static methods are based on analyzing the fracture attributes, such as aperture, length, density, orientation, and connectivity, from the fracture identification methods. Dynamic methods are based on modeling the fracture effects on the fluid flow, stress, and deformation of the reservoir, using numerical or analytical approaches. Both methods require calibration and validation with field data and laboratory experiments.

3 Fracture network models

Fracture network models are representations of the fracture system in the reservoir, based on the fracture characterization methods. They can be classified into discrete and continuum models. Discrete models are based on explicitly describing each fracture as a geometric object, such as a line, plane, or polygon. Continuum models are based on implicitly describing the fracture system as a continuum medium, with equivalent properties, such as porosity, permeability, or stiffness. Both models have their pros and cons, depending on the complexity, scale, and purpose of the analysis.

4 Fracture network analysis

Fracture network analysis is the process of evaluating the fracture network models in terms of their impact on the reservoir performance and behavior. It can be classified into deterministic and probabilistic analysis. Deterministic analysis is based on using a single or a few fracture network models, and applying deterministic inputs and outputs, such as pressures, stresses, or production rates. Probabilistic analysis is based on using a large number of fracture network models, and applying probabilistic inputs and outputs, such as distributions, uncertainties, or scenarios. Both analysis types can provide valuable insights and recommendations for reservoir management.

5 Fracture network optimization

Fracture network optimization is the process of improving the fracture network models or analysis in terms of their accuracy, efficiency, or applicability. It can be classified into calibration and sensitivity optimization. Calibration optimization is based on adjusting the fracture network models or parameters to match the field data or observations, using inverse or history matching methods. Sensitivity optimization is based on identifying the fracture network models or parameters that have the most influence on the reservoir performance or behavior, using sensitivity or uncertainty analysis methods. Both optimization types can enhance the reliability and robustness of the fracture network models or analysis.

6 Fracture network challenges

Fracture network challenges are the difficulties or limitations that arise when dealing with fractured reservoirs in formation evaluation. They can be classified into data and model challenges. Data challenges are related to the availability, quality, and resolution of the data that are used for fracture identification and characterization. Model challenges are related to the complexity, scale, and uncertainty of the models that are used for fracture network representation and analysis. Both challenges require careful data acquisition, processing, and interpretation, as well as model selection, validation, and optimization.