SMARTEOLE Field Test 2 is the second field test campaign realized in the scope of French national project SMARTEOLE. It was held at the wind farm Sole du Moulin Vieux (SMV) between June 2017 and February 2018. This campaign mainly focused on the application of advanced wind farm control strategies regarding both wake steering and axial induction control.
The experimental setup considered for this second wind measurement campaign is displayed on the picture below.
All field tests campaigns of the SMARTEOLE project were held at Sole du Moulin Vieux wind farm located in the northern region of France, approximately midway between Paris and Lille. It consists of 7 Senvion MM82 wind turbines at 80m hub height. The first 5 turbines (SMV1 to SMV5 in picture above) were commissioned in 2009, while the remaining two (SMV6 and SMV7) were added as an extension (officially named Les Kerles) in 2013. The wind farm is organized as a North-South layout, while the direction of the prevailing winds in this region of France is mostly South South-West. Consequently the wake losses of the farm are relatively low, however the two turbines SMV6 and SMV6 have an alignment which is much closer to the direction of prevailing winds which, combined with their close spacing, makes it an interesting case study for the analysis of wind turbine wake and wind farm control strategies.
The wind farm was extensively instrumented during this first wind measurement campaign:
Contrary to the first field tests, the wind farm control strategies applied during this campaign were more advanced. They are described in the subsections below together with a small summary of the obtained results.
The axial induction control tests were held between December 2017 and February 2016. A noise reduced operation (NRO) mode was applied to the upstream turbine SMV6 for west-southerly wind directions (180 - 240°). This NRO mode was expected to provide an increase in combined power production of the two turbines of about 1% for the full wake directions and wind speeds between 7 and 9 m/s. It was activated 4 hours per night (when wind conditions where fulfilled) with an hourly toggle in order to compare the controlled and baseline situation in similar atmospheric conditions.
Although approximately 60 hours of curtailment were recorded during the 3-months campaign, only 5 - 10 hours corresponded to the most favourable wind conditions (full wake sector and wind speed range). The analyses drawn from this study were thefore limited, but in general they tended to show that no increase in combined power production of the turbines could be achieved.
However, a reduction in the upstream turbine loading was measured thanks to data collected in the strain gauges, both in terms of thrust and blade root fatigue loads. Regarding the downstream turbine, no changes in the turbine loading were observed.
The wake steering control tests were held during the summer in August and October 2017. They consisted in misaligning the SMV6 upstream turbine by approximately 13 - 15 degrees for about 6 weeks, in order to deflect the wake away from the downstream turbine SMV5. Unfortunately no toggling could be set up for this campaing so the baseline and controlled periods were fully distinct.
A clear wake deflection was observed at the downstream turbine: this could be related to an increase in the combined power production of both turbines, for ~ 15 degrees wind direction sector corresponding to the wake sector. The observed gain in total power production was in the order of 10%. For all other wind directions, the loss was observed since SMV6 was still being misaligned with no benefits for SMV5. Still, it showed that, when applied for the relevant wind conditions, wake steering had a great potential for increasing the total power production of wind farm. Furthermore, no major changes in upstream turbine blade root fatigue loading was observed through the study of the strain gauge data, meaning that this gain in power production could be obtained without endangering turbine blade lifetime.
This wake steering dataset was analyzed in the scope of FarmConners benchmark exercise for code comparison. The participants were invited to use their models to predict both upstream and downstream power turbine productions during the experiments. The results from this exercise can be found in ....
