In slope stability analysis, rock fabric data is usually available from surface and oriented borehole joint surveys. In deep geomechanics cases, this is not the case. For example, to develop an enhanced geothermal energy (EGS) extraction process in a low-permeability but jointed rock mass, few data on rock fabric are usually available, but the rock mass fabric will dictate the success of hydraulic fracturing as an EGS completion method. To assess these probability and risk issues, we need to carry out many analyses to quantify the possible outcomes, so that we can make reasonable decisions. How can we do this efficiently?
I will describe an upscaling approach we have developed for such challenges. We start with a basic distinct element method (DEM) that represents the rock mass and its fabric. We developed a method to develop an upscaled constitutive law (behavior law) so that a continuum model could be used to analyse scenarios. The upscaled model is 100-500 times more rapid than the DEM model because it can be used in a FEM (finite element method) formulation.
We have incorporated this idea into an upscaled hydraulic fracture simulation method that gives excellent results, allowing us to understand the evolution of the stimulated zone at a large scale. The computational efficiency of our approach allows parametric analyses (sensitivity analysis, Monte Carlo methods) in a reasonable amount of time. This helps rock engineers make decisions under conditions of uncertainty, as they can assess different conditions, helping answer the persistent but challenging rock mechanics question: “What if?”