| Keywords |
geothermal lithium, solute mining, depletion, lifetime, fluid turnover, fluid residence time, inter-well, tracer test, artificial tracer, Upper Rhine Rift Valley, Oberrheingraben |
| Abstract |
The co-operative research project UnLimiteD (www.geothermal-lithium.org/en), initiated by EnBW Karlsruhe, deals with lithium co-production from geothermal reservoirs in the Upper Rhine rift valley (URRV) and the N-German sedimentary basin (NGB). The Göttingen working group uses fluid tracers, natural and artificial, to estimate the overall lithium amount extractable by fluid mining in various geological settings, quantify the evolution of solute co-production rates during fluid turnover, and predict the final mass and energy output over the solute mining lifetime (which may or not exceed the thermal lifetime of a given reservoir, depending on its operation schemes). Behrens et al. (2022, SGP-TR-223) presented a tracer-based method to forecast the depletion of a co-produced solute during fluid turnover, and illustrated its use for various reservoir settings, with a special focus on Soultz-sous-Forêts in the URRV and Horstberg in the NGB. The method is MODEL-INDEPENDENT, viz. the measured signals of any conservative artificial tracer, from single-well or inter-well circulation tests conducted under representative flow conditions, can be used to predict the future co-production output of any fluid-mined solute (in particular: lithium), and its gradual depletion during fluid turnover, irrespective of the availability and parametrizing of a reservoir model (distributed- or lumped-parameter, numerical or analytical). Complementarily, we now rely on the incipient (nil) signal from the inter-well tracer test which started in late February 2022 at a geothermal well doublet on the German side of URRV, to estimate lower-bounds on fluid residence times (RT) and reservoir turnover volume (TOV). Knowing that the tracer signal at the production well stays below a certain value DL (detection limit) for a duration of at least T_nil since the tracer was added at the re-injection well, thereby derived values for RT and TOV are then used to predict lithium depletion and the total lithium output expectable over the reservoir lifetime. This approach, however, is MODEL-DEPENDENT in that it relies on certain assumptions on reservoir structure and boundaries (constraining the large-time asymptotic tracer recovery). Lowering the DL level (due to progress in laboratory-instrumental capabilities) from 0.2 ppb to 20 ppt for all fluid samples collected during the same T_nil (~6 months) raises the predicted lithium output from 3.9 to 4.3 kilotons for the first three decades. We further explain two generic (model-independent) differences between solute depletion rates in petrothermal versus hydrothermal reservoirs, with major implications for the economic efficiency of any solute mining (geothermal co-production) endeavor. |