Yamagawa geothermal power plant, located in the southern Kyushu, commenced its operation in 1995. Since then, geothermal fluids from production wells have been continuously collected mostly at an interval of half a year, and analysed for their physical and chemical properties. The relationships between specific enthalpy and Cl content and also hydrogen and oxygen isotopic ratios of the geothermal fluid revealed the followings. The geothermal fluid is a mixture of a high-enthalpy parent water from deep and a low-enthalpy hot water at shallow depth. The parent water is of seawater origin, of which temperature and Cl content are about 290∼310°C and 18∼19 g/kg. The shallow hot water is also a mixture of seawater and local meteoric water, though its temperature and Cl content are lower than the parent water, and about 150°C and 14 g/kg. The mixing ratios of these two waters are various among individual production wells. In some production wells, underground boiling has taken place under adiabatic condition, while in the other wells under isothermal condition caused by heat supply from rocks. Admixing of re-injection hot water from the power plant is observed in several production wells.
Geothermal production wells with multiple feed zones often show complicated performance over time in terms of deliverability, reflecting changes in surrounding reservoir conditions. For the purpose of clarifying the reservoir properties of each feed zone and their contributions to the productivity of a well with multiple feed zones, a multi-feed wellbore simulator named MULFEWS has been developed. The numerical model at the heart of MULFEWS describes a well surrounded by porous reservoirs and was designed so that it can be applied to calculate not only pressure, temperature and velocity distributions in a well, but also deliverability curves at wellhead, which represent relationship between wellhead pressures and steam, water and total mass flow rates. MULFEWS was applied to the analysis of field data from a production well H-27 at the Hatchobaru geothermal field in Japan, which has four main feed zones and has shown a tendency to increase its water flow rate over time. Consequently, a close matching of measured and calculated results was successfully obtained. The results of the analysis suggest that the reservoirs should be classified into two types, shallower and deeper reservoirs. The shallower reservoirs have changed from a steam and liquid two-phase system to a liquid single-phase system due to a decrease in reservoir temperature. On the other hand, the deeper reservoir has mostly remained liquid single-phase system. The liquid in deeper reservoir has flowed into the wellbore and then begun to flash near the bottom of the hole, although the liquid had initially begun to flash within the formation. Those changes are considered to occur due to the inflow of low temperature fluids from the surrounding formations. The migration of a flash point from the formation to the inside of the well has remarkably increased the water-flow rate. MULFEWS can be used for the interpretation of a well with multiple feed zones to quantitatively explain changes in the reservoir properties of each feed zone and their impact on deliverability.
Injection tests were carried out in 2000 by NEDO (New Energy and Industrial Technology Development Organization) using Well KS-6 at the Mori Geothermal field in Hokkaido, Japan. The pressure interference response at Well NT-303, 460-m apart from the injection well KS-6, was observed. We have conducted inversion analyses of the data using inversion program “DIAGNS”. In addition to conventional line-source models, we assumed point-source models with and without impermeable boundaries. Inversion analysis has been conducted both on the preprocessed interference signals and on the unit response function obtained by the “deconvolution” analysis of the data. As a result, the point-source model with upper and lower impermeable boundaries gives a slightly better estimate than the line-source model. On the basis of the conceptual models, the point source model is preferable, which provides the estimate of the vertical permeability. It is also shown that the quantitative estimation of pressure drift in the data is essential to obtain reliable estimates.
An experimental apparatus has been developed which simulates pore-fluid induced slip on a preexisting fracture. Experiments conducted using the apparatus yield insights into the mechanisms of shear slip induced by fluid injection and relationship between slip behavior and fluid flow. The apparatus construction allows injection of pore fluid directly into a pre-existing fracture in a cylindrical rock sample. Confining pressures and pore pressures up to 25 MPa and 40 MPa, respectively, can be adjusted to impose different stress on the sample. A micro-mass flow meter enables precise measurement of volume of fluid flow into the fracture. Preliminary experimental results are presented which confirm the abilities of the apparatus. Preliminary experimental result shows a characteristic intermittent slip and fluctuation in fluid flow synchronizing with the intermittent slip, implying significant effects of pore fluid injected on the shear slip of a pre-existing fracture.