Please use this identifier to cite or link to this item: http://hdl.handle.net/1946/4616
A distributed-parameter numerical model of the Námafjall-Bjarnarflag geothermal reservoir has been developed. Instead of following the most common approach of modeling the wellbores as constant wellbottom pressure sinks, they are modeled as variable wellbottom pressure sinks, with constant wellhead pressure, through the use of coupled reservoir-wellbore simulation. The purpose of the work is to study the efficiency of this kind of coupling and to predict the reservoir response to three different exploitation scenarios: 40 MWe, 60 MWe and 90 MWe. The flow of mass and heat in the reservoir is modeled through the theory of non-isothermal multiphase flow in porous media implemented by the TOUGH2 code, and an inverse estimation of reservoir parameters is made through the use of automatic parameter estimation capabilities available in the iTOUGH2 code, using a least-squares objective function and the Levenberg-Marquardt minimization algorithm. The HOLA wellbore simulator is used to model the flow within the wells, and the pre- and post-processing tools were based on Linux Shell scripts using freely available software. The automatic parameter estimation was found very useful in finding a set of parameters which produced a reasonable match with available field data for both the natural state and the production response data. The model derived can be regarded as almost closed, and hence pessimistic since the natural fluid recharge into the reservoir is only 14% to 25% of the extracted mass. For the 90 MWe scenario, simulations predict extended boiling throughout the reservoir, pressure drawdown values close to 44 bar and cooling of 35 to 40 °C around the wells. An average decline rate in electrical output of 7.55 MW/yr is expected and by year 2045, 30 wells will be required to maintain 90 MW electrical production. Differences between 15% and 20% were found in the reservoir electrical output if variations in well bottomhole pressures are taken into account through the use of coupled reservoir-wellbore simulation. The coupling method employed in this work is relatively simple and computationally inexpensive, but has the disadvantage that only single feedzone wells can be modeled.