Entropy Production as A Comprehensive Indicator to Address Epistemic Uncertainties - A Case Study of the Hague, Netherlands

Abstract

We present a global sensitivity analysis with an entropy production objective function to improve our understanding of transient thermo-hydraulic processes in geothermal reservoirs under epistemic uncertainties. The entropy production describes the contributions of the irreversible heat transfer and the fluid flow friction. As a global sensitivity analysis demands numerous model runs to measure the significance of each physical parameter, we employ the non-intrusive reduced basis method to construct a surrogate model to decrease the computational cost. We test the application for a geothermal reservoir study in The Hague, Netherlands. The use of a surrogate model provides more than eight orders of magnitude speed-up, enabling an efficient global sensitivity analysis. With the use of the entropy production as the objective function, we can capture important parameters contributing to both the thermal and the hydraulic process for all time steps. The use of either only the pressure or the temperature state as an objective function in the global sensitivity analysis fails to identify important parameters contributing to the thermo-hydraulic process.