Accuracy of EMD-WRF reanalyis for wind power estimations, using WindPRO
(2020) In CODEN:LUTEDX/TEIE EIEM01 20201Industrial Electrical Engineering and Automation
- Abstract
- Developing wind farms is a long and costly task. Assessing the potential of a site is a major part of the development of a wind project. Reanalysis, that are modelling of the wind speed made from scattered wind data across the world (from satellites, weather stations for example), can help to fulfill this step as they are improving their accuracy with each new release. Among them is EMD-WRF, a reanalysis dataset released in the summer 2019. It was chosen to assess this particular reanalysis due to its novelty and because no study was done on it yet. The reanalysis data is compared to 23 measurement masts located in France, that are at heights ranging from 42 m to 122 m above ground level. The comparison is performed on the software WindPRO... (More)
- Developing wind farms is a long and costly task. Assessing the potential of a site is a major part of the development of a wind project. Reanalysis, that are modelling of the wind speed made from scattered wind data across the world (from satellites, weather stations for example), can help to fulfill this step as they are improving their accuracy with each new release. Among them is EMD-WRF, a reanalysis dataset released in the summer 2019. It was chosen to assess this particular reanalysis due to its novelty and because no study was done on it yet. The reanalysis data is compared to 23 measurement masts located in France, that are at heights ranging from 42 m to 122 m above ground level. The comparison is performed on the software WindPRO which
allows to handle wind data by moving them to one location to another, by the process called downscaling, and to do wind power estimations. The reanalysis overestimates the measurements by an average of 18.6 % and with a standard deviation quite high of 12.52 %. The overestimation does not seem to be linked to the correlation coefficient between the two dataset or the geographical proximity of the dataset, except for flat
terrain with very few trees. However, the impact of the correlation between terrain complexity, ie. orography and roughness, and the overestimation is investigated and shows promising results. It shows that categorizing the sites based on terrain criteria can help to reduce the scattering of the results. The measurement sites with simple
terrain are generally having the least overestimation from the reanalysis data. This led to test modifications, based on the type of terrain, applied to the studied reanalysis in order to calculated the wind power of known sites. The wind power estimation was improved in all the sites but it led to some underestimation of the site’s potential. (Less) - Popular Abstract
- Wind development is a long and costly phase, sometimes wind power plants can take more than four years to be build. In that context, one year is already used to record the wind data (speed and direction) and after a full year of recording, it is possible to notice that the site does not have enough potential to install a wind farm.
One solution looked into here is to used computer generated wind data, reanalysis, to do a wind estimation of a site. Those are available immediately and cover different years of data.
Unfortunately, their precision still need to be improved, my work focused on looking for criteria to reduce the uncertainty of those reanalysis. Terrain complexity was noticed as having an impact on
the reanalysis precision.
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9009656
- author
- Bruneau, Mathieu
- supervisor
- organization
- course
- EIEM01 20201
- year
- 2020
- type
- H3 - Professional qualifications (4 Years - )
- subject
- keywords
- Wind power, wind speed estimation, reanalysis
- publication/series
- CODEN:LUTEDX/TEIE
- report number
- 5436
- language
- English
- id
- 9009656
- date added to LUP
- 2021-04-27 14:47:48
- date last changed
- 2021-04-27 14:47:48
@misc{9009656,
abstract = {{Developing wind farms is a long and costly task. Assessing the potential of a site is a major part of the development of a wind project. Reanalysis, that are modelling of the wind speed made from scattered wind data across the world (from satellites, weather stations for example), can help to fulfill this step as they are improving their accuracy with each new release. Among them is EMD-WRF, a reanalysis dataset released in the summer 2019. It was chosen to assess this particular reanalysis due to its novelty and because no study was done on it yet. The reanalysis data is compared to 23 measurement masts located in France, that are at heights ranging from 42 m to 122 m above ground level. The comparison is performed on the software WindPRO which
allows to handle wind data by moving them to one location to another, by the process called downscaling, and to do wind power estimations. The reanalysis overestimates the measurements by an average of 18.6 % and with a standard deviation quite high of 12.52 %. The overestimation does not seem to be linked to the correlation coefficient between the two dataset or the geographical proximity of the dataset, except for flat
terrain with very few trees. However, the impact of the correlation between terrain complexity, ie. orography and roughness, and the overestimation is investigated and shows promising results. It shows that categorizing the sites based on terrain criteria can help to reduce the scattering of the results. The measurement sites with simple
terrain are generally having the least overestimation from the reanalysis data. This led to test modifications, based on the type of terrain, applied to the studied reanalysis in order to calculated the wind power of known sites. The wind power estimation was improved in all the sites but it led to some underestimation of the site’s potential.}},
author = {{Bruneau, Mathieu}},
language = {{eng}},
note = {{Student Paper}},
series = {{CODEN:LUTEDX/TEIE}},
title = {{Accuracy of EMD-WRF reanalyis for wind power estimations, using WindPRO}},
year = {{2020}},
}