<p>Vortex-Induced Vibrations for Energy Harvesting</p> Gauss Centre for Supercomputing e.V.


Vortex-Induced Vibrations for Energy Harvesting

Principal Investigator:
Oriol Lehmkuhl

Barcelona Supercomputing Center

Local Project ID:

HPC Platform used:
SuperMUC of LRZ

Date published:

Over the last decade, the renewable energy sector has been growing at a much faster pace than the rest of the economy in Europe. Wind power especially has seen substantial growth, and a lot of research is currently being done to work out the best way to extract this bountiful energy source.

Nearly all wind energy is currently harvested using what are known as horizontal axis turbines – the large windmill-liked turbines that are already common across Europe. However, there are other ways of harnessing wind energy that do not use this type of structure. A relatively new type of wind generator that does not use rotational movement to capture energy is now being studied extensively across the world. The basis of this technology is a phenomenon called vorticity. Bladeless structures, resembling little more than tall poles, convert wind power into electricity through the oscillation that happens when the structure of the device reaches the same frequency of resonance as the wind vortexes created behind. This is known in fluid dynamics as vortex-induced vibration (VIV).

Oriol Lehmkuhl, a postdoctoral researcher from the Barcelona Supercomputing Centre, won the PRACE Best Industrial Presentation during the PRACEdays19, for his work on the topic. The research project, titled “VIVALdI - HPC of Vortex Induced VibrAtions for flow controL and energy harvestIng”, was awarded 27 million core hours on the GCS supercomputer SuperMUC hosted at the Leibniz Supercomputing Centre in Garching near Munich.

“When you have a fluid interacting with a solid, vibrations are generated and the solid begins to move,” explains Lehmkuhl. “This phenomenon is well known in bridges. If the wind is the right speed and direction, it can cause the supporting structures of the bridge to vibrate. In some cases, this has even caused bridges to collapse.” Because of the inherent danger of these types of vibrations, most research into them has revolved around safety. But Lehmkuhl’s research has taken a different view of the phenomenon – one of opportunity. He and his colleagues are interested in seeing whether it is possible to harness the energy of these vibrations, using tall, flexible structures that move in the wind.

This topic has been investigated to a certain extent, but using small cylinder structures and with low wind speeds. Lehmkuhl’s research has extended this to more realistic wind speeds with cylinders closer to the size of standard wind turbines. “Our hope is that in the future this type of technology could be complementary to standard wind energy turbines.”

In the VIVALdI PRACE project, high fidelity simulations of a cylindrical body oscillating in a free-stream from sub-critical to super-critical Reynolds number have been carried out by means of wall-resolved large-eddy simulations using thousands of CPUs and meshes of hundreds of million of elements. This is the first time these kind of simulations have been performed at this level of modelisation, providing a step forward in the understanding of the physics of fluid-structure interaction in the range of industrial applications.

“Thanks to the PRACE Tier-0 resources, we have been able to carry out large scale simulations that will help build bigger and more efficient VIV wind turbines. The impact of the results will cover very basic research aspects of the turbulent fluid-structure physics, but will also help industry to build more competitive wind-harnessing structures like these”, says Lehmkuhl. “In our work we simulate the fluid and we simulate the structure, and we couple them together in the same simulation domain,” says Lehmkuhl. “For the fluids, we use large eddy simulation models that directly simulate the larger scales of the flow and then use turbulence models for the smaller structures of the flow.”

The simulations are challenging computationally. The model of the fluid in itself is extremely computationally intensive and, on top of this, because the solid is moving all of the time, the mesh needs to be changed throughout the simulation to make the calculation. As such, High-Performance Computing is needed, with many CPUs working at the same time to solve the system. The team is still finishing some of the calculations, but overall Lehmkuhl is pleased with what they have achieved so far in the project. “We have identified how these vortex-induced vibration mechanisms happen at very high Reynolds numbers, and have gained more insight into other mechanisms involved with these technologies. Hopefully this research will help others to push this technology further towards becoming a reality.”


D. Pastrana, I. Rodriguez, J. C. Cajas, O. Lehmkuhl, and G. Houzeaux. On the formation of Taylor-Görtler structures in the vortex induced vibration phenomenon, submitted to International Journal of Heat and Fluid Flow (2019).

I Rodriguez, O. Lehmkuhl, D. Pastrana, J.C. Cajas and G. Houzeaux. Wakes and instabilities of static and freely vibrating cylinder. Direct and Large Eddy Simulations XII, ERCOFTAC Series, Springer, Cham.

O. Lehmkuhl, I Rodriguez, D. Pastrana, J.C. Cajas and G. Houzeaux. High fidelity simulation of vortex induced vibrations for flow control and energy harvesting, EuroHPC Summit Week 2019, PRACE Days, Poznan, Poland, 2019.

Scientific Contact

Oriol Lehmkuhl
Postdoctoral Researcher
CASE - Large-scale Computational Fluid Dynamics
Barcelona Supercomputing Center (BSC)
08034 Barcelona (Spain)
e-mail: oriol.lehmkuhl [@] bsc.es

more information

NOTE: This  is a reprint of the article published in PRACE Digest 2019, p.18-19. The simulation project was made possible by PRACE (Partnership for Advanced Computing in Europe) allocating a computing time grant on GCS HPC system Hazel Hen of the High Performancee Computing Center Stuttgart (HLRS).

LRZ project ID: pn69fa

April 2020

Tags: CSE Energy Energy Efficiency LRZ