GNGTS 2014 - Atti del 33° Convegno Nazionale

Numerical modelling of self-potential for Enhanced Geothermal Systems A. Monetti 1 , A. Troiano 1 , M. G. Di Giuseppe 1 , D. Patella 2,1 , C. Troise 1 , G. De Natale 1 1 INGV-Osservatorio Vesuviano, Napoli, Italy 2 Dipartimento di Fisica, Università Degli Studi ‘Federico II’, Napoli, Italy Introduction. Geothermal resources represent a sustainable and potentially competitive alternative to fossil fuels. Enhanced geothermal system(EGS) technologies, in particular, provide a powerful way to produce geothermal electric energy in almost every area of the world. EGSs exploit hot rock systems with low water content, with the economic feasibility depending on the drilling costs needed to reach a suitable temperature. Despite its great potential, EGS exploitation is still perceived as environmentally threatening, because of the problems posed by unwanted induced seismicity above a certain magnitude threshold (MIT Report, 2006). Such events can more frequently occur due to hydraulic stimulation that is aimed at creating a permeable reservoir in EGS systems. Such a negative perception of EGSs is mainly due to the Basel earthquake of magnitude ML 3.4 that occurred in 2006 December. Although this event did not produce serious damage, it was strongly felt by the population because the geothermal site was located in the center of the city (Haring et al., 2008; Ripperger et al., 2009). Less known but equally interesting cases have also been described in the literature over the last few decades (see Majer et al., 2007, and references therein). Hence, interpreting the mechanisms of induced seismicity and understanding ways of mitigation is important to allow the promotion of geothermal EGS exploitation worldwide (Giardini, 2009). The hot-dry-rock site of Soultz-sous-forets is one of the best examples of the experience of EGSs. The permeability enhancement of this reservoir was obtained through the drilling and subsequent stimulation of four wells that reached depths of up to 5km (Portier and Vuataz, 2009). A complex sequence of uid injection was performed over several years, to enlarge the fracture system of the basement rock, composed mainly of granite, and to enhance its permeability. The stimulation of this multi-well structure allowed the creation of natural heat exchangers and the generation of stable commercial electricity. The development of the Soultz power plant and its several related scienti c projects have been fully described in the literature. The geophysical prospecting consisted of a series of passive seismic tomography investigations (Charlety et al., 2006; Cuenot et al., 2008; Dorbath et al., 2009), as well as electromagnetic imaging (Geiermann and Schill, 2010). Also, geochemical data were collected (Sanjuan et al., 2010), the rock permeability and the fracture system were described (Genter et al., 1997; Evans et al., 2005a, 2005b), and the regional stress eld was estimated (Cuenot et al., 2006). Furthermore, the whole drilling process was accurately described for each well through a series of technical reports (Baria et al., 2004). These reports thus provide highly detailed records of the different phases of the arti cial stimulation that was carried out to create the permeable reservoir, including the ow rates, the head pressures of the boreholes, the temperature pro les and the distribution and magnitude of induced seismic events. In this paper we use this large amount of information as a basis for testing the capability of a classical applied geophysical method, so called self-potential (PS) to forecast the induced seismicity related to deep uid injection during well stimulation. The self-potential method (SP) is one of the oldest of all the geophysical techniques. It consists of monitoring or mapping passively the electrical field existing at the ground surface of the Earth. In this paper the distribution of SP associated with a real pumping stimulation at Soultz-sous-Forets has been numerically evaluated and successively compared with induced seismicity occurred during the stimulation process. To this aim a numerical procedure has been used allowing the reconstruction of the thermodynamic evolution of the reservoir in terms of pressure and temperature. This kind of procedure has been already employed to evaluate the Coulomb Stress changes on preferred fault mechanism (Troiano et al. , 2013) and its matching with the induced seismicity recorded during the fluid injection. The reconstructed changes in pressure and temperature are subsequently considered 224 GNGTS 2014 S essione 1.3