RESERVOIR ENVIRONMENT OF THE ONLJh4A GEOTHERMAL POWER PLANT, NORTHEAST JAPAN, ESTIMATED BY FORWARD ANALYSIS OF LONG-TERM ARTIFICIAL-TRACER CONCENTRATION CHANGE, USING SINGLE-BOX-MODEL SIMULATOR

A single-box-model numerical simulator for personal computer analysis was developed in order to estimate macroscopic parameter values for exploited geothermal reservoirs and essential fluids coming from the depth. The simulator was designed to compute history data concerning total production and reinjection fluids at geothermal power plants from the assumed parameter values, based on conservation laws for water mass, heat energy and masses of conservative chemical constituent3 of geothermal fluids. Using two kinds of forward analysis techniques, Le. the cast-net and pursuit methods, programs containing the simulator can semiautomatically select the optimum combination of the unknown parameter values by minimizing the differences between the simulated and measured history data for specific enthalpy and chemical compositions of the production fluids. The forward analysis programs were applied to the history data from the Onuma geothermal power plant (production capacity, 10MWe) where waste hot water reinjection, chemical monitoring and artificial tracer tests have been conducted since 1970, almost the beginning of the geothermal exploitation. Using the history data, enthalpy and iodine concentrations of the total production fluids with the amounts of KI tracer injected as spikes, the macroscopic parameter values for the exploited reservoir and the essential hot water from the depth were uniquely determined as follows: mass of the hot water convecting in the exploited reservoir (MO), 3.23x109kg; recycling fraction of the reinjected waste hot water to the reservoir (R), 0.74; specific enthalpy of the essential water from the depth (Hl), 385kcalkg; iodine concentration of the water (Il) , 0.086mg/kg with chlorine concentration (Cl), 259mgjkg. These results support the conceptual model that the exploited Onuma reservoir mainly in the Tertiary volcanics is supplied with the neutral Na-Cl type hot water of abnormally high B/CI mole ratio of around 1.0 by a large essential reservoir distributed at depth in the Paleozoic to Mesozoic detrital marine sedimentary rocks. INTRODUCTION Producing large amounts of geothermal fluids for power generation causes dynamic changes in reservoir, such as boiling, cold groundwater invasion and essential fluid inflow from the depth. Reinjection of the produced hot waters after steam separation to the reservoir, as at most geothermal power plants in Japan mainly for the purpose of environmental protection, has a large effect on the changes. Therefore, it is very important to optimize allocation of geothermal wells, and their production and reinjection rates, through understanding the nature of these changes and predicting their future, for the purpose of long-term stable operation of geothermal power plants. In Japan, monitoring concentrations of major chemical constituents in production fluids has been conducted at almost all the geothermal power plants. Also, artificial tracer tests have been tried at many Japanese geothermal power plants. These chemical history data have been analyzed by many kinds of methods for various objectives (eg. Shigeno, 1992a). Among these, the studies for the hydrothermal system of the Onuma geothermal power plant by Ito et al. (1978), Matsubaya and Kubota (1987), Kubota and Matsubaya(l987), and Kubota et al. (1989) are very important concerning not only the optimization of geothermal reservoir exploitation but also resource assessments of hydrothermal :systems. They tried to estimate macroscopic values for the exploited reservoir, i.e. total amount of the hot waters convecting in the reservoir and recycling fraction of the reinjected waste hot waters to the reservoir, based on a single box model using artificial tracer (iodine ion, I) test data and chemical (especially chlorine ion, CI) monitoring data. Concerning a well at the Palinpinon geothermal ower plant in the Philippines, Malate and O'Sullivan (991) reported an application of the similar lumped-parameter model to the analysis of the production fluid chemistry changes. We developed semiautomatic forward analysis programs for personal computers in order to estimate macroscopic values for exploited geothermal reservoirs and essential geothermal fluids from the depth, through repetitions of single-box-model numerical simulations, using chemical monitoring data. These programs were applied to the monitoring data from the Onuma area, and the results were compared with those by the above papers, and with the previously proposed hydrothermal-system models for the Onuma area. In this paper, these methods and results were reported. For more details of this paper, refer to Shigeno et al. (1992b) and Shigeno (1992~). ANALYTICAL METHOD Single box model, and difference eauations Fig. 1 shows the very simple single box model of exploited hydrothermal systems used for the numerical simulations in this study. This model is based on Matsubaya and Kubota (1987). An exploited geothermal reservoir is represented by a single homogeneous box, in which only hot water circulates as the fluid phase (shown with suffix 0). Non-steady-state inflow to, and outflow from the reservoir box are as follows: an inflow of an essential hot water from the depth (1); outflows of total hot