On self-potential data for estimating permeability in enhanced geothermal systems

Abstract

We study the use of hypothetical self-potential (SP) data – more specifically streaming-potential data – for the inversion of subsurface permeability distributions, using the enhanced geothermal system at Soultz-sous-Forêts, France, and a synthetic geothermal hard-rock reservoir as examples. Simulations are carried out using the software SHEMAT-Suite. We perform this study based on results obtained via a massive Monte Carlo approach and additionally use the Ensemble Kalman Filter technique for the inversion. In a first step, we perform forward simulations and assume that SP data is measured along the production and injection wells. The SP monitoring data mainly depend on the near-field (150m) permeability around these wells. In this case, the SP signal is in good agreement with the distribution of the hydraulic head. In contrast, Darcy velocity and possible tracer pathways identified by tracer experiments cannot be identified uniquely based on SP data.Alternatively, stochastic inversion is done based on data recorded in deviated wells distributed around the production and injection wells. In this case, principal fluid pathways and permeability magnitudes are reproduced by stochastic inversion of the SP data. The results are comparable to results obtained by tracer experiments. Joint inversion of tracer and SP data yields the best results in terms of small estimation mismatch. Permeability and pathway geometry can be adequately estimated even for an incorrect coupling coefficient as long as it does not differ more than half an order of magnitude from the true value.