Impact of heterogeneity on high-temperature aquifer thermal energy storage: a case study

High-temperature Aquifer Thermal Energy Storage (HT-ATES) has the potential to reduce greenhouse gas emissions from the heating sector due to its capability of making excess or renewable based heat from the summer period available in winter. The economic viability of HT-ATES is largely determined by the thermal recovery, i.e. the fraction of recovered heat from the previously injected heat. The heterogeneity of the storage formation has an impact on both the induced thermal plume in the subsurface and on the thermal recovery and results in uncertainties of both parameters. This study aims at quantifying these uncertainties resulting from the subsurface heterogeneity.

The case study of Hamburg-Wilhelmsburg (Northern Germany) is considered, where heat can be stored in the Lower Lignite Sands at around 200 m b.g.l. The geostatistical data basis consisted of 26 boreholes in HH-Wilhelmsburg and 476 boreholes distributed over the whole of Hamburg. The lithologies described in these boreholes were grouped into three indicators with high, medium and low permeability, which were parameterized by literature data. Indicator variogram analysis was applied and the thus derived geostatistical parameters were used for conditional sequential indicator simulation, resulting in 30 realizations of indicator distributions for the model domain. HT-ATES operation was simulated for 26 years by a thermo-hydraulically coupled model with OpenGeoSys. In three additional scenarios with 30 realization each, three different exploration boreholes were added to the conditioning data at a distance of 10 m to the warm ATES well, in order to examine the resulting benefit of such an exploration borehole for reducing the prediction uncertainty.

Simulation results show that considering the heterogeneity of the hydraulic permeability leads to substantial variability in the predicted thermal plume. This spatial uncertainty is reduced in all three scenarios which include an exploration borehole. Also, the predicted warm well return temperatures are significantly less variable between realizations, when the exploration boreholes are included in the conditioning data. For the scenario without an exploration well the mean thermal recovery increases from 36% in the first year to 63% in the 26^th year. The standard deviation of the thermal recovery thereby increases from 7% to 10%. The exploration borehole scenarios show different mean thermal recoveries of 29%, 41% and 46% in year one and 53%, 78% and 84% in year 26. The standard deviation is maximum 2% in all three scenarios at all times. This shows, that a large reduction in uncertainty can be expected if an exploration borehole in the direct vicinity of the ATES warm well is available.