GNGTS 2014 - Atti del 33° Convegno Nazionale

GNGTS 2014 S essione 1.3 227 Fig. 1 – a) Top: Sketch of the simulation volume. Blue plane, Earth surface; red plane, injection plane. Bottom: pressure and temperature initial conditions. b) Simplified stimulation functions for the GPK2 and GPK3 Soultz-sous- Forets wells, representing the rates of injected water. (a–f) Times of the stimulation cycle shown. the streaming potential coupling coefficient for the whole injection cycle, following Darnet et al. (2003). Once the value of C and the fluid pressure field are both known, it is possible to compute the electric current sources from Eq. (7) and to solve the electrical problem calculating the electric potential induced by the stimulation process. Method. Our method of analysis consists of a two-step procedure. In the first step, injection of water is simulated (Pruess, 1991, 1999) in a homogeneous medium, approximating a crystalline granite basement compatible with the deep structure of the Soultz-sous-Forets (France) EGS site. The modeled 3D physical domain and the imposed initial conditions are shown (Fig.1a). Water at ambient condition is injected at a variable flow rates, in order to reproduce the effects of a real stimulation experiment realized in the GPK2 and GPK3 wells of the geothermal field during the 2003. An essential scheme of this stimulation process is given in Fig.1b. In such a way the pressure and temperature changes at each point in the medium has been obtained, at six distinct times (Troiano et al. , 2011, 2013). The spatial gradient of the induced pressure field in the whole volume has been successively considered as source of electric potential anomalies, due to the so called electrokinetic effect linked to the pressure gradient in the medium. Fluid flow through a porous medium generates, in fact, an electric potential variation due to the electrical interaction between the fluid and the electrical double layer at the pore-mineral interface (Helmholtz, 1879). Fluid injection and/or circulation within geothermal reservoirs can produce surface Self-Potential (SP) anomalies of several mV that are correlated in space (e.g., Ushijima et al. , 1999) and in time (e.g., Ishido et al. , 1983; Marquis et al. , 2002) to reservoir fluid flow. The electric potential changes induced by fluid injection has been reconstructed resolving the Poisson equation by finite element code. A source term of the kind: (9) has been imposed, where P represents the fluid pressure, σ the electrical conductivity, and l represents the coupling term, expressed in A/m 2 , characterizing the electrical current density produced in response to the unit hydraulic gradient. The coupling coefficient has been assumed as a constant during the stimulation, in agreement with literature data (e.g. Darnet, 2003).