Abstract:
Currently, New Zealand generates around 18% of its total electrical capacity using geothermal
sources and could benefit significantly from simulation for optimisation (Ministry of Business,
Innovation & Employment, 2017). Current geothermal process and reservoir simulations are
conducted separately with data manually parsed between the different simulators. Delays in the
modelling process and the inability to efficiently model effects that reservoir changes have on
the plant over the asset’s lifetime . Coupling both process and reservoir simulators would
enable accurate prediction of both reservoir and plant issues. In this thesis, a proof of concept
is developed. The reservoir simulator AUTOUGH is coupled with the process simulator
VMGSim using Python and PyTOUGH. A demonstration based on a plant in New Zealand
was built. The aim of this study is to demonstrate and compare the effects of coupled models
in predicted plant performance.
Geothermal fluid mass flow, pressure and temperature data is passed between AUTOUGH and
VMGSim where both the wellbore and plant are simulated. Brine injection data is passed back
to AUTOUGH. This cycle is run until either a simulated plant failure occurs or the simulation
is terminated. In an ORC plant, typical failures relate to temperature drops in the geothermal
fluid that lead to the inability to vaporise the working fluid used to power the turbines. As a
result, plant changes are required to maintain production, which could reduce power generation
or require drilling an additional production well.
During coupling CO₂ depletion within the reservoir can be studied and the effects
characterized. A reduction of CO₂ produced for a binary plant promotes heat transfer but at the
cost of increased pressure drop within the wellbore; as a result the power generation decline
occurs much earlier than anticipated.
Coupling models adds additional benefits to the modelling process that support the
optimisation of both reservoir and surface-related activity. Addition of historical ambient air
temperature allows for more accurate results but at the cost of increased simulation time, but
in-turn allows for more accurate economics when conducting FEED (Front End Engineering
Design)......