Influence of turbulence on, and turbulence created by tidal power plants in tidal currents using Large Eddy Simulations

SNIC 2018/1-40


SNIC Large Compute

Principal Investigator:

Göran Broström


Göteborgs universitet

Start Date:


End Date:


Primary Classification:

10509: Oceanography, Hydrology and Water Resources

Secondary Classification:

20306: Fluid Mechanics and Acoustics

Tertiary Classification:

20304: Energy Engineering




Using tidal power for energy production is seen as one component for a sustainable future energy production. Sweden do not has important tidal energy resources along its coastline (although there are places with relatively strong ocean currents e.g. on the Swedish west coast). However, Sweden has a long tradition of innovative projects and start-up companies in tidal power production (and wave energy utilization). This stems from early initiative in alternative energy production systems, and a tradition of innovative engineering. The Minesto company is one example and uses a special kind of power plants characterized of a moving kite in the tidal stream. The overall goal with this project is to provide realistic simulations of tidal water turbulence to ensure that tidal power plants (such as Minesto Deep Green technology, see are adapted to the actual conditions. The main project is funded by the Swedish Energy Agency (Energimyndigheten), and runs over 2.5 years. The knowledge of tidal turbulence generated in this project will be used both for evaluation of environmental impact, and to optimize the control system, power generation and construction with regard, for example, to dynamic loads. The environmental impact will be studied through modelled changes in turbulence levels downstream of a powerplant, this will also require development of novel turbine models. We will perform large eddy simulations (LES) of fully developed tidal driven turbulent flows. We will then incorporate a model of the tidal power plant, which in this case is a kite moving in a horizontal 8. The domain must be discretized fine enough to resolve the small-scale turbulence and the domain size must be large enough to encompass the moving kite (power plant) and the long wake behind it. This will make this study computationally demanding. The study will therefore use two codes dedicated for different purposes (one pseudo-spectral ”NCAR-LES” and one finite volume based ”OpenFOAM”) in order to make efficient use of the computational resources. For the purposes of this study we request 350000 core hours/month on Trilith and 200000 core hours/moth on Hebbe.