De-trending of turbulence measurements

Kurt Schaldemose Hansen (Invited author), Gunner Chr. Larsen (Invited author)

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    Abstract

    Traditionally, turbulence is considered as a stationary stochastic process imposed on a given constant mean wind speed. However, measured (raw) turbulence intensities often display the characteristics of a non-stationary process, where the mean wind speed changes slowly with time. The change in mean wind speed appears as a trend in the wind speed time series, and often a linear trend is assumed. Wind resource measurements typically include statistics of ten-minute mean and standard deviation, and for such data it is not possible to calculate the trend contribution directly, because this requires access to the basic time-series. However, including a suitable modelling of the mean wind speed time variation, it is possible to estimate an approximate (linear) trend correction based on statistical data only. This paper presents such an algorithm for de-trending of turbulence standard deviation based on time series statistics only. The performance of the proposed de-trending algorithm is assessed using huge number of time series recorded at different types of terrain and orography. The strategy is the following: Based on the available time series information a conventional (linear) time series de-trending is performed and subsequently compared with the prediction from the proposed algorithm. The de-trended turbulence intensities are reduced in the range of 3 – 15 % compared to the raw turbulence intensity. The performed analysis shows that the proposed model, based on statistical information only, accounts for approximately 80% of the “true” (linear) trend as evaluated on basis of the full time series information.
    Original languageEnglish
    Title of host publicationEuropean Wind Energy Conference & Exhibition 2007 : Milan, Italy, 7-10 May 2007
    Number of pages8
    Place of PublicationBrussels
    PublisherEuropean Wind Energy Association (EWEA)
    Publication date2007
    Edition1
    Publication statusPublished - 2007
    Event2007 European Wind Energy Conference and Exhibition - Milan, Italy
    Duration: 7 May 200710 May 2007

    Conference

    Conference2007 European Wind Energy Conference and Exhibition
    Country/TerritoryItaly
    CityMilan
    Period07/05/200710/05/2007

    Fingerprint

    Dive into the research topics of 'De-trending of turbulence measurements'. Together they form a unique fingerprint.

    Cite this