Stability classification for CFD simulationsin complex terrain

Dr. Carolin Schmitta Dr. Cathérine Meiβnerb

Andrea Vignaroli b

a juwi Wind GmbH, Energie-Allee 1, 55286 Wörrstadt, Germany

b WindSim AS, Fjordgaten 15, N-3125 Tønsberg, Norway

EWEA 2013 Annual Event, 4-7th February 2013, Vienna

Outline

General Aspects of Atmospheric Stability

Effects of Stability on Windfields in Theory and Measurements

Stability Parameters in Measurement and MERRA Data

Examples from different Sites

Benefits for CFD Modelling

Conclusions und Outlook

General Aspects of Atmospheric Stability

• Atmospheric Stability

The resistance of the atmosphere to vertical motion is dependant on the different stratification parameters.

Simulations don’t fit measurements

• Integration in CFD Modelling

Historically elevation and roughness has been adjusted to compensate for incomplete models but often stability is the real issue.

• Application and Importance for Wind Energy

Modification of wind profile, shear and turbulence up to greater hub heights and flow circulation in complex terrain influence energy yield and also power curve performance.

Theoretical

Effects of Stability on Windfields

Vertical Windprofiles

Measurement

Effects of Stability on Windfields

Theoretical

Stable Neutral

Streamlines in CFD Modelling

Stable Neutral

Vertical Speed in CFD Modelling

Model

Horizontal Effects

Challenges for Wind Energy Applications

Equations/Meteorological relations are known, but:

How is reality and how can this be captured by measurement and model methods?

What influences have to be taken into account and what data and parameters are necessary to reproduce profiles more accurate?

Atmospheric Stability can be determined by various methods, depending on datasets

Temperature Gradient,, Richardson, MOL, Pasquill Classes ...

It is important to establish proper models capable of reproducing reality, otherwise various tuning options will not improving modeling capability and understanding of the flow behavior.

Data for Stability Determination

Measurement Limitations:

Met Tower: No flux measurements; sensor accuracy/mounting for gradient method; short time period not representative.

Lidar: No temperature gradient measured; short time period; approaches like Pasquill not yet validated.

The question is which of the four surrounding MERRA data points should be used to assess the stability parameters or if an inverse distance weighting can be used.

EWC Wind Potential Analysis provide downscaled MERRA data with auxiliary information.

Model Data Possibilities:

MERRA data is free available in the internet. MOL can be calculated by the variables given in the data set.

𝑳=−𝒖𝟑∗𝑻 𝒗𝒍𝒎𝒍∗𝒄𝒑∗𝝆𝟎/(𝒌∗𝒈∗𝒔𝒉𝒕𝒇𝒍)Modern-ERA Retrospective Analysis for Research and Applications, http://gmao.gsfc.nasa.gov/merra/)

Examples from different Sites

Site A - Complex, forest Site B - Medium, no forest

Site C - Flat, forest Site D - Complex, water, no forest

All sites with one or more met towers, A and B with additional Lidar measurements

Site A- Stability distribution

Application of different Stability Classifications Monthly Distribution

Measurement Data

MERRA Data• Similar average ratio of

stable cases over the year

• Ratio neutral/instable depends on classification scheme

• Monthly variability smoothed by MERRA

Site A- Quality

Speed Dir Temp

ρ hour 0.62 0.45 0.98

ρ day 0.83 0.46 0.99

Main stable directions SE/NW not enough represented in MERRA

Around 10% fewer stable cases in MERRA

Good correlation of speed and temperature data but not for wind direction

Site B- Stability Distribution and Quality

Measurement Data

MERRA Data

Speed Dir Temp

ρ hour 0.69 0.31 0.98

ρ day 0.89 0.34 0.99

Good agreement Mast Gradient and MERRA Gradient/MOL

Site C- Stability Distribution and Quality

Measurement Data

MERRA Data

Speed Dir Temp

ρ hour 0.63 0.44 0.96

ρ day 0.79 0.67 0.98

MERRA gives more instable cases and fewer neutral ones, stable distribution ok.

Site D- Stability Distribution and Quality

Measurement Data

MERRA Data

Speed Dir Temp

ρ hour 0.70 0.44 0.89

ρ day 0.83 0.53 0.96

Too much instable cases and around 10% fewer stable cases. Site generally shows more neutral stratification.

Comparison Temperature Cycle C and D

Site C

Site D

Influence of land/sea surface distribution can influence the quality of diurnal temperature cycle in MERRA in comparison to only land surface.

More instable cases are produced

All Sites- longterm MERRA MOL

A- Complex, forest B- Medium, no forest

C- Flat forest D- Complex, no forest

MOL

I N S

A -113 500 132

B -95 -700 119

C -83 -600 118

D -124 2500 231

Stability values and distribution (sectorwise) can be used to CFD model setup

Benefits for CFD Modelling

Results Site D: Sectorwise Wind Profiles

Conclusions and Outlook

Local different stability regimes can be shown to significantly influence windflow behaviour. MERRA has proven to be a valuable data set for the determination of the monthly overall stability conditions of a site as long as the surrounding grid points are representative for the site.

For complex terrain and in coastal sites with land/sea mixing, the model wind direction distribution can be misleading. Often the use of measured wind direction might improve the results.

Application of MERRA MOL helps with proper setup of CFD Modelling and results in better representation of wind profiles. Automatic stability classification from measurements and MERRA data download will be part of WindSim during 2013.

Thank you very much for your attention

Acknowledgements:

Contact:Dr. Carolin Schmitt

juwi Wind GmbH

Energie-Allee 1

55286 Wörrstadt

carolin.schmitt@juwi.de www.juwi.de

Determination and Application of Parameters

How stability classification can be calulated

•Temperature gradient To – gamma*z

•Monin-Obukhov length

•Richardson

• … many more

MERRA Data

The question is which of the four surounding Merra data points should be used to assess the stability or if an inverse distance weighting can be used.

k= karman

g = Gravity

cpl/cp = heat capacity (dry/wet)

Tv = virtual temperature

tt = temperature

spfh = specific humidity

shtfl = surface heat flux

lml = lowest model level

𝑳=−𝒖𝟑∗𝑻 𝒗 𝒍𝒎𝒍∗𝒄𝒑∗𝝆𝟎 /(𝒌∗𝒈∗𝒔𝒉𝒕𝒇𝒍)

MERRA data is free available in the internet.

The MOL is not given directly but can be calculated by the variables given in the data set.