Post conversion of Lidar data on complex terrains

Stéphane SANQUER (1), Alex WOODWARD (2)(1) Meteodyn France, (2) ZephIR Lidar

27/09/2016 LIDARs - the zapping competition – WIND EUROPE SUMMIT 2016 1Stéphane SANQUER

229/09/2016

Context…

Wind turbines generator higher and higherNeed Remote Sensing Devices (RSD) precise enough to be compared to

standard anemometers

Strong inhomogeneity of wind on complex terrains Some discrepancies may appear with standard anemometersPost conversion of Lidar data in such situations

Numerical methodology to give efficient post conversion factorMay we use CFD ?When should we use the post-conversion : level of terrain complexity ?

Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Context and purposes

Why conversion are needed in complex terrains ?

Velocity vectors aren’t be constant across the Lidar scan disk

We need the 3D vectors of the wind upstream and downstream the Lidarscan disk center in order to project them on the beams

329/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Context and purposes

𝑉𝐿 =𝑉𝑟 𝑑𝑜𝑤𝑛 − 𝑉𝑟 𝑢𝑝

2 sin(𝜃)=

𝑉𝑑 . 𝑖 − 𝑉𝑢 . 𝑗

2 sin(𝜃)

1.00

1.05

1.10

1.15

1.20

0 1 2 3 4 5

Co

rre

ctio

n f

acto

r

Wind incidence (°)

𝑉𝐿 =1

2 sin(𝜃) 𝑉𝑑 sin 𝜃 + 𝜑𝑑 + 𝑉𝑢 sin(𝜃 − 𝜑𝑢)

CF=

VC/V

L

429/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Interpretation of site complexity

F. Bingöl (2009) : Complex terrain and wind Lidars – Risoe-PhD

Context and purposes

529/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Context and purposes

Differences between RSD and mast data depends on complexity

629/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Context and purposes

Purposes…

Assessing differences between ZephIR 300 and mast data in various complexity classes

Highlighting the categories where CFD conversion is needed

Providing an efficient methodology for conversion of Lidar data on complex terrains

729/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Computations of 8 sites : from the simplest to the highest complex

Assessing the conversion factor

Orography Roughness

Mo

der

atel

y co

mp

lex

Hig

hly

co

mp

lex

CFD Conversion

Factor

𝑉𝐿 =1

2 sin(𝜃) 𝑉𝑑 sin 𝜃 + 𝜑𝑑 + 𝑉𝑢 sin(𝜃 − 𝜑𝑢)

829/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Post-conversion of Lidar data

At each time step, by using the conversion factor (depends on the wind incidence)

RSD data are Post-converted

Correlation factor R² unchanged – Linear regression factor increased close to unity

929/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Post-conversion of Lidar data

0.86

0.88

0.90

0.92

0.94

0.96

0.98

1.00

Flat terrain Moderatelycomplex

Complex Highly complex

Before conversion After conversion

Linear regression factor ULidar vs UMast

1029/09/2016Challenges of forest modelling – WIND EUROPE SUMMIT 2016

Conclusions

Applying CFD conversion to data from RSD in complex terrain improves the agreement between wind speed measurements from RSD and masts

CFD post conversion needed for complex and extremely complex terrain.

Same effect of terrain than forest on the linear regression factor (0.97). Conversion improve the factor close to unity (0.99)

After correction for highly complex site, the factor is 0.98. Method of post conversion depends on the ability of CFD to accurately predict the flow deviation over the ground.

Post-conversion of Lidar data