| Abstract |
Accurate imaging of geothermal system around hot springs and fumarole manifestations from surface to deep zones is an interdisciplinary problem that contributes to geothermal resource exploration, volcanology, and geodynamics. The system is composed of various aspects including: temperature, lithology, and fracture distributions; hydrothermal fluid conditions on flow pattern, velocity, and pressure; rock physical properties of permeability, thermal conductivity, and porosity; and chemical properties of fluids and rocks. Most of these data are obtained by well logging and laboratory tests using borehole cores. In general, the distribution of geothermal wells is biased, and their data are limited in amount and depth ranges. Therefore, a sophisticated spatial modeling technique is indispensable to 3D imaging of geothermal system. For this purpose, we developed a 3D modeling method by combining an interpolator analogous to minimization principle of mechanical potential energy and a stochastic simulation. The 22 km?18 km area in the Hohi geothermal region in central Kyushu, southwest Japan, was chosen as the test site. This area is well known to have two geothermal power stations, Hatchobaru and Otake. By applying the technique to the well logging dataset, the imaging down to 3000-m depth was successfully drawn based on the three modules: temperature modeling, geologic modeling, and parameter estimation. |