Tuesday, 27 September 2016
14:30 - 16:00 LIDARs - the zapping competition
Resource assessment  
          

In this highly interactive quick-fire session, participants will scan through 14 LIDAR-related presentations and vote to select the three contributions they would like to hear in full. Presentations will cover a wide range of possible LIDAR applications, both offshore and onshore, such as power-curve validation, resource assessment in complex terrain, turbulence intensity measurements, and more.

You attended this session?

Please give us feedback

 

Learning objectives

Get a wide overview of the latest research and field work involving LIDARs, both onshore and offshore.

 

This session will be chaired by:
Co-chair(s):

Wiebke Langreder Wind Solutions, Denmark
Co-authors:
Wiebke Langreder (1) ^F Bayram Mercan (2)
(1) Wind Solutions, Mårslet, Denmark (2) re-consult Ltd Sti, Ankara, Turkey

Roaming remote sensing: quantification of seasonal bias

Introduction

The possible seasonal bias incurred by the use of a roaming LIDAR/SODAR is being investigated in detail. The advantages of remote sensing (RS) for wind resource estimation has been much discussed during the last years. One possible use is the concept of a roaming RS, where a lidar/sodar complements a conventional measurement mast. In this scenario the RS device is moved around the wind farm site in short intervals of for example three months, while a fixed mast measures in parallel for a longer period, typically one year or more. The purpose of the supplementary RS measurements is to reduce the uncertainty of horizontal extrapolation during flow modelling. The influence of seasonal variations is supposedly being accounted for by the data from the mast, but it turns out that there is a significant remaining uncertainty caused by deployment of the RS for less than a year, and this is for the first time being discussed and quantified in this paper.

Approach

For a site with several measurement masts data from one mast is used to simulate the roaming RS device while the other mast is used as reference. It is then investigated how much a change in period will affect the estimate of the wind speed at the assumed RS position. The results show in some cases significant variations during the year, which suggest that any measurement period shorter than one year with the RS device will be seasonally biased.

Main body of abstract

A total of five sites in various terrain have been analysed. All sites have been equipped with minimum two IEC-compliant measurement masts of a measurement height of minimum 80m which have been measuring for a minimum of one year. The recovery rate of the data is above 95%. The installation of two top anemometers allowed to remove any potential shadowing effect due to the tower and lightning finials on most sites. Delta RIX is smaller than 6% on all sites. The distances between the masts range from 3 to 9km.
In order to quantify the impact of seasonal variations, one mast has been assumed to be the fixed mast, the data from the second mast has assumed to stem from the roaming RS and has been split into either 3- or 6-months subsets. The subsets have been correlated with the fixed mast using linear regression (with residuals). The resulting data string has been compared with the real measured data string. Additionally, different subsets have been combined to simulate seasonal sampling.
To illustrate the impact of the seasonal bias on the AEP (annual energy production), production calculations have been performed for fictive wind farm layouts based on the 1-year data set and the short-term, correlated data set using WAsP. In parallel the AEP using a combination of both masts has been determined as benchmark. It could be shown, that in some cases the introduction of the short-term measurement by a roaming device increased the error (instead of the intended decrease) in comparison to just measuring with one mast, even with seasonal sampling.


Conclusion

Depending on the thermal conditions on the site significant seasonal bias’ even over distances as short as 3km could be shown. The error introduced due to the seasonal bias was as large as 10% in AEP by using roaming RS. In some cases this error is larger than the expected horizontal extrapolation error using only one mast without the RS.
Suggestions are made for which scenario (terrain, climate) the uncertainty due to seasonality could fall below the uncertainty due to horizontal extrapolation.



Learning objectives
This paper gives quantitative information about the pros/cons on the use of roaming RS including an evaluation of possible scenarios where it is advantageous to use a roaming RS.