Quantitative Risk Assessment for Enhanced Geothermal Systems

This study uses a quantitative risk-assessment approach to place the uncertainty associated with enhanced geothermal systems (EGS) development into meaningful context and to identify points of attack that can reduce risk the most. Using the integrated geothermal assessment tool, GT-Mod, we calculate the complimentary cumulative distribution function of the levelized cost of electricity (LCOE) that results from uncertainty in a variety of geologic and economic input parameter values. EGS is a developing technology that taps deep (2-10km) geologic heat sources for energy production by "enhancing" non-permeable hot rock through hydraulic stimulation. Despite the promise of EGS, uncertainties in predicting the physical end economic performance of a site has hindered its development. To address this, we apply a quantitative risk-assessment approach that calculates risk as the sum of the consequence, C, multiplied by the range of the probability, ΔP, over all estimations of a given exceedance probability, n, over time, t. The consequence here is defined as the deviation from the best estimate LCOE, which is calculated using the 'best-guess' input parameter values. The analysis assumes a realistic but fictitious EGS site with uncertainties in the exploration success rate, the sub-surface thermal gradient, the reservoir fracture pattern, and the power plant performance. Uncertainty in the exploration, construction, O&M, and drilling costs are also included. The depth to the resource is calculated from the thermal gradient and a target resource temperature of 225 °C. Thermal performance is simulated using the Gringarten analytical solution. The mass flow rate is set to produce 30 MWe of power for the given conditions and is adjusted over time to maintain that rate over the plant lifetime of 30 years. Simulations are conducted using GT-Mod, which dynamically links the physical systems of a geothermal site to simulate, as an integrated, multi-system component, the collective performance of each system over time. It is dynamically linked to the Geothermal Energy Technology Evaluation Model (GETEM - www1.eere.energy.gov/geothermal/getem.html) that calculates the LCOE based on time-series performance output from GT-Mod. A Monte Carlo approach propagates input uncertainties to the output by describing uncertain inputs with probability density functions (PDF's) and then simultaneously varying the PDF's via a Latin Hypercube Sampling (LHS) technique. Exceedance probabilities for the LCOE are calculated as a post-processing exercise. Results show that for the given set of uncertainties, the LCOE assumes a lognormal distribution with the tail skewed towards the higher values and a mean LCOE that is almost 2 ¢/kWh higher than the best estimate; this despite the fact that the 'best-guess' parameter values are the mean values of the input PDF's. This is a result of component feedback that can amplify the system's dynamics and implies that the best estimate LCOE may considerably under-estimate the risk of developing that site. Correlation analysis indicates that reductions in drilling costs and better characterization of the sub-surface environment will reduce risk the most.