Abstract
Wesselkvartalet is a newly constructed, mixed residential/commercial building
complex in the city of Asker, Norway. It integrates a multi-reservoir, shallow
geothermal storage facility to cover its heating and cooling needs, as well as
providing surplus heat energy in the winter to a distributed snow-melting system
for the city streets.
The operation of the geothermal system is complex, with three geothermal
reservoirs at different depths and very different properties, and comprising
more than a hundred wells, set up to provide both constant base load and rapid
release of heat at occasional peak loads.
To guide operations of this complex plant, we are developing a geothermal
simulation and optimization system based on the MATLAB Reservoir Simulation Toolbox
(MRST). The simulator handles fluid flow and heat transport in fractured
reservoirs with background flow and moving water table and supports advanced
well models and controls. It is coupled with an adjoint-based nonlinear
optimization engine that supports model calibration/improvements as well as
developing strategies for optimal use of the geothermal plant.
In this talk, we discuss the numerical geothermal simulator and how the system
is modeled, as well as providing detail on how we constructed the simulation
model, including gridding, fracture representation and calibration to observed
data. We also discuss the computational challenges involved in discretizing and
solving the governing equations, and provide preliminary examples on how
nonlinear optimization could be used to improve the model and guide operations.
complex in the city of Asker, Norway. It integrates a multi-reservoir, shallow
geothermal storage facility to cover its heating and cooling needs, as well as
providing surplus heat energy in the winter to a distributed snow-melting system
for the city streets.
The operation of the geothermal system is complex, with three geothermal
reservoirs at different depths and very different properties, and comprising
more than a hundred wells, set up to provide both constant base load and rapid
release of heat at occasional peak loads.
To guide operations of this complex plant, we are developing a geothermal
simulation and optimization system based on the MATLAB Reservoir Simulation Toolbox
(MRST). The simulator handles fluid flow and heat transport in fractured
reservoirs with background flow and moving water table and supports advanced
well models and controls. It is coupled with an adjoint-based nonlinear
optimization engine that supports model calibration/improvements as well as
developing strategies for optimal use of the geothermal plant.
In this talk, we discuss the numerical geothermal simulator and how the system
is modeled, as well as providing detail on how we constructed the simulation
model, including gridding, fracture representation and calibration to observed
data. We also discuss the computational challenges involved in discretizing and
solving the governing equations, and provide preliminary examples on how
nonlinear optimization could be used to improve the model and guide operations.