Ocean tides deliver a steady and reliable source of renewable energy that can be harvested with so-called hydrokinetic turbine farms. It is, however, not fully clear yet how a turbine affects the efficiency of another turbine placed in its wake. To better understand this, scientists at the Institute of Marine Engineering in Rome recently performed Large-Eddy Simulations using PRACE resources to analyse a turbine’s wake flow. They not only found valuable insights for efficient hydrokinetic farms, but they also created a basis for more accurate modelling of whole turbine farms.
Renewable energy is produced not only from wind or the sun, but also from the kinetic energy of rivers, tides and oceans. The Earth’s largest bodies of water produce powerful and reoccurring currents that, similar to winds, can power turbines to produce electricity. In fact, in the next decades, experts expect the so-called hydrokinetic energy to become an important part in the mix of renewable energy sources.
“This method of energy production has significant advantages,” says Riccardo Broglia, a research scientist at the Institute of Marine Engineering (CNR-INM) in Rome. Since water is much denser than air, tidal energy is more powerful than wind energy. In addition, tides and currents are more predictable and stable than winds, and therefore able to yield a more reliable stream of electricity.
Finding the balance
A number of different designs for tidal stream turbines have already been developed by companies. However, in order to be an effective alternative to conventional energy sources, hydrokinetic turbines need to operate in farms composed of many individual devices, and it is not yet clear how exactly a specific turbine affects other turbines located in its wake.
It is understood that a producing turbine will absorb the kinetic energy of the water flow, which will recover its motion and regain its kinetic energy only after a certain distance downstream of the turbine — the scientists call this process “momentum recovery”. If the flow has not fully recovered its momentum, a downstream turbine will yield less energy than it otherwise could. “Increasing the distance between turbines will reduce this interference, but also result in a larger farm size and, therefore, in higher environmental impact and costs,” explains Broglia. “In order to find the perfect balance, the turbine wakes and their effects have to be understood in detail.”
The image shows how the flow momentum recovers downstream of the turbine: The colour gradient from blue to red denotes increasing momentum deficit, that is, the loss of flow velocity and kinetic energy in the wake of the turbine. Momentum starts to recover after a distance of roughly three times the diameter of the turbine (z/D=3).
Image source: A. Posa and R. Broglia
Resolved with two billion points
Riccardo Broglia and his co-researcher Antonio Posa performed Large-Eddy Simulations using PRACE resources to analyse the wake flow generated by a specific axial-flow hydrokinetic turbine — a small scale model with a blade length of 70 centimetres that had already been characterised experimentally by colleagues at the CNR-INM, the French Research Institute for Exploitation of the Sea (IFREMER) and the University of Strathclyde in Glasgow.
The simulations were conducted on a grid composed of almost two billion points, making them by far the most resolved ones to date. This ultra-high resolution was possible thanks to parallel computing and the use of the so-called Immersed-Boundary methodology, as Posa explains. This enabled the scientists to perform their simulations using a grid in cylindrical coordinates, which, compared to the more typical body-fitted grids, reduced the memory requirements. In addition, this approach enabled the scientists to use more efficient numerical methods, which improved the efficiency of parallel computing.
This way, Broglia and Posa were able to simulate the turbine wake, as well as visualise and analyse the speed of momentum recovery of the water flow, in fine detail. They found that, after the turbine has absorbed the kinetic energy from the water, momentum starts to return after a distance of roughly three times the diameter of the turbine, and it will be reasonably recovered — to roughly 80 percent — at a distance of around 10 times the turbine diameter.
The turbine wake flow features strong, helical vortices arising from the tips of the rotating turbine blades. As soon as these tip vortices start to become instable and break down, the wake becomes narrower and mixes with the surrounding water stream, and flow velocity begins to recover.
Image source: A. Posa and R. Broglia
Even though the simulations were conducted on a model scale turbine of only 70 centimetres in diameter, the results provide valuable insights for full scale hydrokinetic turbines, which may reach diameters of 10 to 20 metres. In particular, the scientists found that a big role in momentum recovery is played by the tip vortices, that is, the helical water vortex structures that arise from the tips of the rotating turbine blades
Smartly placing turbines
In fact, the stability — or rather the instability — of these tip vortices triggers the momentum recovery process: As long as the vortices are stable, they act like a shield, isolating the turbine wake from the surrounding water streams. It is only when these previously regular vortices start to break down that the wake becomes narrower. Then, surrounding water is mixed into the wake, fuelling it with fresh kinetic energy, so that flow momentum starts to recover. “These kinds of behaviours in flow physics are very difficult to measure or estimate from experiments. That’s why high-fidelity simulations like ours are crucial to understand them,” says Broglia.
The image shows the rate of momentum recovery in the wake of the turbine, in particular the correlation between the instability of the shielding tip vortices and the recovery of flow momentum: At a distance of about twice the diameter of the turbine (z/D=2), the tip vortices (blue streaks) break down; simultaneously, momentum recovery is triggered, meaning the flow starts to recover (visible in red).
Image source: A. Posa and R. Broglia
From their results, Broglia and Posa could also deduce how the unsteadiness of the water flow affects the turbine wake: The more turbulent the flow into the turbine is, the faster the tip vortices become unstable, which will, in turn, accelerate momentum recovery. “In practice, this means that placing turbines in highly turbulent currents will allow for a closer positioning of downstream turbines and, therefore, for more compact turbine farms than a placement in less turbulent currents,” says Posa.
The analysis of the results also revealed a finding that is important from a methodological point of view: There is a strong deviation in the behaviour of the tip vortices from what is usually assumed when modelling the flow in turbine wakes. In particular, two physical quantities, the so-called deformation tensor and the Reynolds stress tensor, are typically assumed to be proportional in turbulence modelling. But, for the tip vortices of hydrokinetic turbines, this turned out to be largely incorrect — a valuable insight for the turbulence modelling community. “In turbulence modelling, especially large-scale turbulences are strongly dependent on the specific problem and setup at hand, and, consequently, their accuracy in simulations relies on high-quality, problem-specific reference values”, Posa explains.
Training for other models
In addition, Broglia’s and Posa’s results can be used to train simplified, lower-fidelity models to represent not one, but several hydrokinetic turbines in a single simulation — at affordable computational cost. Same as turbulence models, such so-called actuator models representing the action of several turbines require reference data from experiments and high-fidelity simulations to be properly tuned. Such precise reference values for a hydrokinetic turbine wake and momentum recovery are now provided by the CNR-INM scientists’ results and can be used to create more accurate simulations of whole turbine arrays.
Project title: 19th call: wakehydroLES – Characterization of the wake of an axial-flow hydrokinetic turbine via LES
The resources awarded were:
19th call: 39 million core hours on Joliot-Curie (KNL) hosted by GENCI at CEA, France
Research field: Engineering
A. Posa and R. Broglia: Momentum recovery downstream of an axial-flow hydrokinetic turbine. Renew. Energy (2021). DOI: https://doi.org/10.1016/j.renene.2021.02.061
A. Posa and R. Broglia: Characterization of the turbulent wake of an axial-flow hydrokinetic turbine via large-eddy simulation. Comput Fluids (2021). DOI: https://doi.org/10.1016/j.compfluid.2020.104815
A. Posa, R. Broglia and E. Balaras: Instability of the tip vortices shed by an axial-flow turbine in uniform flow. J. Fluid Mech. (2021). DOI: https://doi.org/10.1017/jfm.2021.433