Periodic Reporting for period 2 - Train2Wind (Training school on entrainment in offshore wind power)
Reporting period: 2022-02-01 to 2024-07-31
The Train2Wind project started out of curiosity about how many wind turbines can be packed close together in the North Sea, giving consideration to effect of wake losses. The plans for buildout offshore, but especially in the North Sea were large and getting even larger, in the tens of GW. While wake modelling behind a single turbine has been studied since the 1980’s, wind farm wakes and the effects of large wind farms only came into focus in the 2000s - 2010s. Wind farm wakes being thought of as independent of the development of the atmospheric boundary layer were investigated then, but for the very large wind farms coming, it is intuitively clear that there also should be an influence on the wider weather and wind resource patterns. Accordingly, Train2Wind chose to concentrate on very extended wind farms or wind farm clusters, which extend with a large density of turbines for so long that the atmosphere reaches a new equilibrium. Not taking the effects of large wind farms on their neighbours sufficiently into account, can lead to tens of billions of euros of misplaced investments, especially in such good wind power areas as the North Sea. The overall objectives were therefore to figure out how closely one can space wind farms, how the atmospheric boundary layer develops in the vicinity of wind farms, investigate the role of turbulence and to further develop experimental and modelling techniques.
The central common investigation we planned was an experimental campaign at a wind farm in Southern Denmark, close enough to shore to fly there with the UAS, measure, and turn back to land, aided by an extensive campaign using floating lidars and scanning lidars installed on the offshore substation. Unfortunately, the flight permissions for the UAS did not arrive in time, leaving the campaign with nearly a year of lidar data from the crew transfer vessel, plus some scanning lidar data for a few months from the substation. The great support from the wind farm owner RWE was paramount to the results. The UAS crews found other areas of investigation.
Additionally, Train22Wind had three main training objectives with early-stage researchers (ESRs) at the centre:
▪ to give the fellows the necessary skills to execute their particular research project successfully,
▪ to give them a broad background in wind energy, so that they are able to frame their own project in the larger picture,
▪ to give them transferable skills for further personal development, including the employment of better research practices and efficient communication/dissemination/exploitation of research.
Of the 19 fellows trained, 7 have gotten further employment elsewhere, while the rest is still employed with the host institution. 4 fellows already finished their PhDs.
The central common investigation we planned was an experimental campaign at a wind farm in Southern Denmark, close enough to shore to fly there with the UAS, measure, and turn back to land, aided by an extensive campaign using floating lidars and scanning lidars installed on the offshore substation. Unfortunately, the flight permissions for the UAS did not arrive in time, leaving the campaign with nearly a year of lidar data from the crew transfer vessel, plus some scanning lidar data for a few months from the substation. The great support from the wind farm owner RWE was paramount to the results. The UAS crews found other areas of investigation.
Additionally, Train22Wind had three main training objectives with early-stage researchers (ESRs) at the centre:
▪ to give the fellows the necessary skills to execute their particular research project successfully,
▪ to give them a broad background in wind energy, so that they are able to frame their own project in the larger picture,
▪ to give them transferable skills for further personal development, including the employment of better research practices and efficient communication/dissemination/exploitation of research.
Of the 19 fellows trained, 7 have gotten further employment elsewhere, while the rest is still employed with the host institution. 4 fellows already finished their PhDs.
A Marie Curie action allows to address a common problem from various angles. We found the necessary expertise in a relatively small consortium, consisting of the Technical University of Denmark (DTU), the Geophysical Institute at the University of Bergen (UiB), the Ecole Polytechnique Federal de Lausanne (EPFL), the Eberhard-Karls-University in Tübingen (EKUT) and the University of Copenhagen (UCPH). Additionally, we invited three of the largest offshore wind developers to the consortium (Vattenfall, Equinor and RWE), a company for vertical axis turbines (SeaTwirl) and Charles Meneveau from the University of Johns Hopkins University (JHU). In the end, we trained 13 PhD students and 6 short-term fellows. One of the fellows had the others as research topic, investigating the collaboration aspects between the students and groups from a humanities perspective (UCPH).
In the consortium, we chose expertise within wind and wind farm modelling with Large Eddy Simulations (LES)(DTU and EPFL), the mesoscale WRF model including wind farm parameterisations inside the model (DTU), built a small wind farm in a wind tunnel (EPFL), had an expert in wind speed observations from satellite (DTU), two groups performing lidar observations (UiB and DTU), and two groups with Uncrewed Aerial Systems (UAS) expertise (EKUT and UiB). One of the developers (RWE) also contributed with access to an offshore wind farm, the Rødsand 2 wind farm near the Danish island of Lolland. Their support for sailing the lidars on the crew transfer vehicles for nearly a year is highly appreciated, as well as their facilitation of access to the substation for installation of the scanning lidar. RWE, as well as Vattenfall and Equinor, hosted students. SeaTwirl collaborated on the adjustment of a load calculation tool for vertical axis floating turbines, and JHU received students.
While the main planned offshore campaign did not work out as planned, Train2Wind created significant progress for being able to investigate the research questions further, both from a conceptual and an experimental side. We were able to (for the first time) measure the wind with two lidars on a ship at many places inside and outside an offshore wind farm, aided by a scanning lidar from the substation for a shorter time period. We also improved satellite based SAR image conversion to wind speeds, and developed a turbulence model as well as improved the wind farm parameterisation in weather models. The turbulence influence in a flow model was investigated, and the measurements of turbulence by UAS was shown to be able in a virtual environment. An improved analytical wake modelling framework was developed, and another one was used for wind farm control. We also improved modelling for vertical axis turbines. For the experimental side, we developed a particle measurement and a much faster hygrometer, a flying sonic anemometer and a motion controlled ship-borne lidar. The time for 5-hole probe calibration was cut from a day to 20 minutes using robotics. Using this, for the first time, Kelvin-Helmholtz billows were detected over a working wind farm. All those individual results and the trained fellows will bring the modelling and measurements of wind farm flows forward.
In the consortium, we chose expertise within wind and wind farm modelling with Large Eddy Simulations (LES)(DTU and EPFL), the mesoscale WRF model including wind farm parameterisations inside the model (DTU), built a small wind farm in a wind tunnel (EPFL), had an expert in wind speed observations from satellite (DTU), two groups performing lidar observations (UiB and DTU), and two groups with Uncrewed Aerial Systems (UAS) expertise (EKUT and UiB). One of the developers (RWE) also contributed with access to an offshore wind farm, the Rødsand 2 wind farm near the Danish island of Lolland. Their support for sailing the lidars on the crew transfer vehicles for nearly a year is highly appreciated, as well as their facilitation of access to the substation for installation of the scanning lidar. RWE, as well as Vattenfall and Equinor, hosted students. SeaTwirl collaborated on the adjustment of a load calculation tool for vertical axis floating turbines, and JHU received students.
While the main planned offshore campaign did not work out as planned, Train2Wind created significant progress for being able to investigate the research questions further, both from a conceptual and an experimental side. We were able to (for the first time) measure the wind with two lidars on a ship at many places inside and outside an offshore wind farm, aided by a scanning lidar from the substation for a shorter time period. We also improved satellite based SAR image conversion to wind speeds, and developed a turbulence model as well as improved the wind farm parameterisation in weather models. The turbulence influence in a flow model was investigated, and the measurements of turbulence by UAS was shown to be able in a virtual environment. An improved analytical wake modelling framework was developed, and another one was used for wind farm control. We also improved modelling for vertical axis turbines. For the experimental side, we developed a particle measurement and a much faster hygrometer, a flying sonic anemometer and a motion controlled ship-borne lidar. The time for 5-hole probe calibration was cut from a day to 20 minutes using robotics. Using this, for the first time, Kelvin-Helmholtz billows were detected over a working wind farm. All those individual results and the trained fellows will bring the modelling and measurements of wind farm flows forward.
Train2Wind progressed experimental and modelling techniques for the investigation of large-scale wind farm flows. The developed tools allow to investigate the possible spacing and the wind resource effects of neighbouring wind farms for developers and marine spatial planning authorities. The impact of Train2Wind is two-fold: the cadre of trained fellows will transfer academic knowledge to industry, either through dissemination or through subsequent employment, and the datasets will be available for further wake model development. Using those improved models, a reduced uncertainty in the wind farm calculations translates directly into savings of financing cost, as the spread between the best estimate (“P50”) and the bank’s safety margin (“P90”) is getting smaller. This enables a low-cost no-regret route towards the EU target of 27% of renewables by 2030. This would tie in with the H2020 Societal Challenge Energy, and would help to keep the current leading position of European companies in the offshore wind sector.
The main results of Train2Wind are published in a book, with chapters for each fellow. Please see Train2Wind.eu for the publication. Additionally, some 30 papers were published, and more are expected to follow. Finally, the Lollex campaign dataset is currently being curated and will be published.
The main results of Train2Wind are published in a book, with chapters for each fellow. Please see Train2Wind.eu for the publication. Additionally, some 30 papers were published, and more are expected to follow. Finally, the Lollex campaign dataset is currently being curated and will be published.