A Stochastic Framework for Uncertainty Quantification and Risk Assessment of Enhanced Geothermal Systems

Abstract

A major issue to overcome when characterizing a deep fractured reservoir is that of data limitation due to accessibility and affordability. Moreover, the ability to map discontinuities in the rock with available geological and geophysical tools tends to decrease particularly as the scale of the discontinuity goes down. Geological characterization data include, but are not limited to, measurements of fracture density, orientation, extent, and aperture. All of which are taken at the field scale through a very sparse limited number of deep boreholes. These types of data are often reduced to probability distribution functions for predictive modeling and simulation in a stochastic discrete framework. Stochastic discrete fracture network (SDFN) models enable, through Monte Carlo simulations, the probabilistic assessment of flow and transport phenomena that are not adequately captured using continuum models. Despite the fundamental uncertainties inherited within the probabilistic reduction of the sparse data collected, very little work has been conducted on quantifying uncertainty on the reduced probabilistic distribution functions. Using nested Monte Carlo simulations, we investigated the impact of parameter uncertainties of the probability distribution functions of fracture network on the flow using topological measures such as fracture connectivity, and physical characteristics such as the hydraulic conductivity tensor.