Natural fracture networks are a major control on fluid flow and geomechanics in the subsurface. They are therefore important for applications including mining and tunnel stability, groundwater flow and pollution monitoring, geothermal energy, CO2 sequestration and hydrocarbon exploration and production.

However, they generally cannot be imaged and mapped directly – geophysical methods such as seismic have too low resolution to image fractures, while wellbores give only a 1D viewpoint and very limited coverage.

Currently, if natural fractures are modelled at all, they are modelled stochastically – placed at random locations with arbitrary geometry. The resulting models are typically inaccurate and give poor results when used in geomechanical and fluid flow models

To improve this and generate more accurate fracture models, we have developed a technique for simulating the growth of large fracture networks based on fundamental geomechanical principles. This forward modelling technique, analogous to new sedimentological forward modelling techniques, allows more accurate and realistic fracture models to be generated more quickly, even in cases with limited input data. They can be easily integrated into fluid flow and geomechanical models. This method also allows for more rapid uncertainty modelling.