An investigation into turbine performance and wind flow modelling

under cold weather driven stable atmospheric conditions

Winterwind 2013 (Östersund, Sweden - 13 February 2013)

GL Garrad Hassan – Independent Renewable Experts

Almost 1000 staff, in 44 locations, across 26 countries

Vancouver

Ottawa

Portland

San Diego

Montreal

Peterborough

Austin

Querétaro

Porto Alegre

Santiago

Beijing

Seoul

Tokyo

Shanghai

Mumbai

Bangkok

Bangalore

Singapore

Newcastle

Wellington

Melbourne

BristolCork

Paris

Izmir

Cairo

CopenhagenHinnerupOldenburgHamburgWarsaw

Lisbon

BarcelonaZaragozaMadrid

Imola

London

Slough

Glasgow

Cape Town

HeerenveenSint Maarten

Kaiser-Wilhelm-Koog

Content

Cold climate driven stable atmospheric conditions in the Nordic Region:

1 - Consequences to the wind flow (challenges) and how it translates

2 - How complex CFD flow modelling can help with the wind flow modelling challenges

3 - Impact of high frequency of stable atmospheric conditions on turbine performance

1– 2 – 3

Stable atmospheric conditions

What is it and how it translates

Radiative cooling driven stable atmospheric conditions

- Unstable, neutral, stable and very stable atmospheric conditions

- High frequency of stable atmospheric conditions

- Actual

- Dry-adiabatic

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

Radiative cooling driven stable atmospheric conditions

- When do stable and very stable atmospheric conditions happen in the Nordic region

High frequency of stable atmosphere and wind conditions

Low TI as a proxy for stable

atmospheric conditions

0.65

0.70

0.75

0.80

0.85

0.90

0.95

1.00

1.05

1.10

0 30 60 90 120 150 180 210 240 270 300 330

Speedup

Direction Sector (deg)

Stable Unstable Total

Wind Flow modelling challenges

Wind speed at mast A/Wind speed at mast B = Speedup

1–2 – 3

Complex CFD flow modelling for

stable wind flow challenges

• Most accurate description of wind flow: Navier-Stokes Equations (below)

• Very difficult to solve!

• Linear models simplify this equation - some accuracy is sacrificed

• CFD makes fewer changes – less loss of accuracy:

• Reynolds-Average Navier-Stokes (RANS)

• Advanced software required

• Modern computing power required

• Like a wind tunnel on your computer

What is CFD (Computational Fluid Dynamics)

Directional speedup errors (95 mast pairs at 16 sites)

Linear

(WAsP)

Neutral CFD

Wind Flow modelling under stable conditions

Average Errors

Linear (WAsP) = 9.6%

Neutral CFD = 9.5%

Stable + Neutral (CFD) = 6.6%

Stable+Neutral

error is lower

Stable + Neutral CFD approach (GL GH approach)

Directional speedup errors (95 mast pairs at 16 sites)

1– 2 –3High frequency of stable conditions

and turbine performance

Performance of wind turbines under stable conditions in the

Nordic Region – What the theory suggests

P = 1/2ρ(u)3(1+3TI2)ACp

0

500

1000

1500

2000

2 4 6 8 10 12 14

Po

wer

(KW

)

Wind speed (m/s)

0%

5%

10%

15%

20%

92%

94%

96%

98%

100%

102%

104%

106%

6 7 8

Low

TI

/ H

igh

TI

me

asu

red

po

we

r cu

rve

pro

du

ctio

n

Annual mean wind speed (m/s)

T1

T2

T3

T4

T5

Average

Performance of wind turbines under stable conditions in the

Nordic Region

Results for IEC Power Curve Measurements of 5 mast/turbine pairs in Sweden

No partial icing cleaning

and no temperature

filtering

92%

94%

96%

98%

100%

102%

104%

106%

6 7 8

Low

TI

/ H

igh

TI

me

asu

red

po

we

r cu

rve

pro

du

ctio

n

Annual mean wind speed (m/s)

T1

T2

T3

T4

T5

Average

Performance of wind turbines under stable conditions in the

Nordic Region

Results for IEC Power Curve Measurements of 5 mast/turbine pairs in Sweden

Data cleaned for partial

icing and filtered

excluding temperatures

below 0˚C

92%

94%

96%

98%

100%

102%

104%

106%

6 7 8

Low

TI

/ H

igh

TI

me

asu

red

po

we

r cu

rve

pro

du

ctio

n

Annual mean wind speed (m/s)

T1

T2

T3

T4

T5

Average

Performance of wind turbines under stable conditions in the

Nordic Region

Results for IEC Power Curve Measurements of 5 mast/turbine pairs in Sweden

Data cleaned for partial

icing and filtered

excluding temperatures

below 0˚C and different

site calibration speedups

applied for stable and

unstable conditions

Performance of wind turbines under stable conditions in the

Nordic Region

Results for IEC Power Curve Measurements of 5 mast/turbine pairs in Sweden

Without different site calibration speedups With different site calibration speedups

80%

85%

90%

95%

100%

105%

110%

4 5 6 7 8 9 10 11 12 13 14 15 16 17

Low

TI

/ H

igh

TI

me

asu

red

po

we

r cu

rve

Wind speed (m/s)

T1 T2 T3 T4 T5 Average

80%

85%

90%

95%

100%

105%

110%

4 5 6 7 8 9 10 11 12 13 14 15 16 17

Lo

w T

I / H

igh

TI

me

asu

red

po

we

r cu

rve

Wind speed (m/s)

T1 T2 T3 T4 T5 Average

Conclusions

� There is high frequency of stable and very stable atmospheric conditions in sites

across the Nordic Region caused by radiative cooling

� This presents a challenge for both wind flow modelling and turbine performance

� Stable CFD reduces the error in wind flow modelling at these sites

� Theory shows that low turbulence reduces turbine performance in raising part of

the power curve

� Preliminary Nordic data supports this, however shows recovery in the “knee” of

the power curve

� Need to gather more Power Curve Measurement data across the Nordic region

� Careful consideration should be given to winter Power Curve Measurement data,

as partial icing affects more low TI data than high TI, and the anemometry used

� Site calibration speedups for stable and unstable conditions should be considered

separately as suggested by the new draft of the IEC standard

Thank you

Contact:

Carla Ribeiro, Senior Team Leader Nordic Region

carla.ribeiro@gl-garradhassan.com

With Thanks to:

Simon Cox

Bob Hodgetts

Andrew Tindal

GL GH CFD Team

GL GH AMOS Team

Any questions?