Added value of LiDAR correction methods from a user's perspective

Introduction

Wind resource assessment with LiDAR replacing or extending met tower measurements is getting more and more popular during the last years. Hereby, it has to be considered that in comparison to cup anemometry, LiDAR volume measurements may lead to different wind speed values in cases of non-homogenous airflow, especially in complex terrain.
Quantization and correction of this effect is crucial to get reliable wind speed and energy yield estimates. The German “Technical Guideline for Wind Turbines TR6” describes the requirements on precision and uncertainty assessment to be performed on remote sensing measurements to deliver accepted energy yield expertise. With the use of corrected LiDAR data officially permitted by this document, our aim is to find out about the added value of these methods for real world projects.


Approach

At present, two different approaches exist to determine and correct the effect of non-homogenous flow fields on LiDAR wind speeds:

• Online methods use measured vertical wind speed components in different directions in the measuring cone and calculate corrected velocities from a relation of these differences every 10 minutes.

• Offline methods mainly rely on CFD modelled flow conditions to determine the inhomogeneity of the vertical wind speed over the cone and provide an overall correction for the complete measurement period per wind direction sector.

Both methods allow to build LiDAR error maps that propose optimal siting positions for LiDAR. They promise to be helpful tools for error minimization. But of course, they are only as reliable as their underlying methods.
We did extensive comparisons of both methods under different conditions to understand more about the way they work and to see if they can be successfully used as “black box” or if there are limitations or improvements in their applications.


Main body of abstract

Comparisons of LiDAR next to a met tower is done for several sites with different complexity and different magnitude of correction values. The dimension of the mean correction values as well as the sectors with the maximum impact can vary considerably, and not every time a better fit of the remote sensing data is reached. Whereas the online correction has the advantage to be completely automatic and not influenced by the user, the offline correction with a CFD model allows different tuning options.

This can be an advantage to get better results with optimal model setup, but also bears the risk that the amount of uncertainty later appearing in the final energy yield expertise depends a lot on the quality of your model and the choice of parameters.
Investigations of different parameters such as terrain complexity, atmospheric stability, forest parametrization, vertical speed component or minimum speed deliver a background for understanding these effects. Additionally, the TR6 requires that a CFD model should have sufficient horizontal and vertical resolution and at least 10 degree sector size to allow correction estimates. To address this point, we added also checks on the sensitivity of the CFD model set-up to find out which are the crucial points to get good results.


Conclusion

Results for different sites reveal that each of the two approaches has its own advances and limitations and does not lead to identical results for the site. Also they can’t be successfully applied as black box without background knowledge on the critical impact parameters to get satisfying result for all sites.
Considering that the magnitude of necessary correction delivers a significant part to final uncertainty estimation and thus, is directly related to project value, a careful adaption has to be done.



Learning objectives
Audience will learn more about the backgrounds of successful application of LiDAR correction methods in real world projects and the importance of optimal planning and estimating these effects.