Experimental Design and Response Surface Methods: Applications in Geothermal Reservoir Simulations and Probabilistic Resource Assessment

Abstract

Resource estimates in geothermal “green” fields involve significant inherent uncertainty due to poorly constrained subsurface parameters and multiple potential development scenarios. There is also limited published information on probabilistic resource assessments of geothermal prospects. This paper explores the applications of a systematic experimental design (ED) approach to a geothermal reservoir simulation model to generate probabilistic resource assessment results. A Plackett-Burmann design was used to build 12 simulation experiments for a geothermal system, investigating six model parameters at two levels. A probabilistic 30-yr power capacity was successfully generated from the resulting response surface. Analysis of the response also showed that the power capacity was significantly affected by only three of the tested parameters: reservoir temperature, permeability, and well depth. Volumetric methods, both “heat in place” and mass in place, were applied to the same modelled system to compare the resource assessment results. Applying the experimental design (ED) and response surface methodology (RSM) enables the use of geothermal reservoir simulations for probabilistic geothermal resource assessments. The assessment using volumetric heat in place methods, even with similar thermal recovery factors, still results in variations due to differences in conversion of recovered thermal energy at the wellhead to electrical power capacity. It is recommended that the conversion efficiencies be carefully considered and the assumptions clearly stated when translating recovered thermal energy into electrical capacity. Noting the uncertainties in the thermal recovery factor, it is also recommended that production-based methods using numerical simulation or volumetric mass in place be used to estimate electric power capacities.