In a geothermal reservoir, injected water flows through certain fluid pathways such as natural or induced fractures connecting injection and production wells to harvest the thermal energy from surrounding rocks. The spatial distribution and temporal movement of the fluid in the reservoir is the determining factor of the heat-sweep efficiency. Currently, monitoring of the heat-sweep efficiency is mainly through adding tracers from the injection wells and testing its arrival at the production wells. Such a method can help to estimate the fluid pathway near the wellbore but is unable to see the fluid movement in space farther away from the monitoring well.We propose to inject the bubbled water into the reservoir and use time-lapse vertical seismic profile (VSP) method to map the dynamic distribution of the bubbled water. This method is based on the concept that liquid containing immiscible gas bubbles exhibits distinguished elastic and inelastic properties compared with pure liquid. Such differences can be sensed by seismic waves and eventually be spatially mapped. Petrolern has extensive experience in this area and has developed a ML-based methodology to monitor the CO2 plume movement in carbon sequestration reservoirs.The effect of the fluid property on the elasticity of fractured or porous media has been well studied and applied in oil and gas explorations. In summary, the media will exhibit lower elasticity and velocity when it is filled with more compressible fluid in its fractures. The compressibility of liquid and gas typically differs in order of magnitude. Furthermore, recent studies also prove that the multi-phase fluid saturated rock will exhibit another unique feature when interacting with seismic wave, that is the significantly large attenuation of the seismic wave energy, compared with single-phase fluid saturated rock.In time-lapse VSP survey, the seismic sources are installed and shot at the surface, and multiple receivers are deployed in the boreholes. With a proper design of the experiment, seismic ray paths between the sources and receivers can form a matrix of dataset to enable high resolution tomography analysis both in time-lapse velocity and attenuation variations. Using rock physics transformation, these results can be interpreted into the spatial distribution of the injected bubbled water.The phase I outcome will be a preliminary design of a new field experiment that will acquire and analyze the data to significantly reduce the uncertainty of spatial and temporal movement of fluid in the reservoir. In Phase II, we propose to conduct field test and further develop the technology into a commercial product. The project will be conducted in Atlanta and Houston, USA.
