Wind farm blockage and wakes A coupled engineering model
VindKraftNet 29 October 2020
Nicolai Gayle Nygaard
Power variation among front-row turbines Nacelle lidar measurements in wind farm
– Wind speed reduction exceeds the standalone turbine induction
2
Wind farm blockage clues
Mitraszewski, Hansen, Nygaard and Rethoré TORQUE 2012
“Wall effect”
Increasing power trend towards ends of front row
Nygaard and Brink, WESC 2017
3
Measured flow around a single turbine (from above)
Minute by minute evolution
4
Measured flow around a single turbine (from above)
Average over 30 minutes
5
Turbine blockage zone
Blockage zone Wake recovery
– Based on work by Gribben and Hawkes @ Frazer-Nash in Offshore Wind Accelerator project
– Uses Rankine half-body (homogeneous flow + point source)
– Adjusted to match flow in 1D momentum theory
6
Single turbine induction model
Wind direction
Single turbine induction model
Superposition Coupling with wake model
Ground effect
7
Wind farm blockage model
Wind farm
blockage model
Wake model
CT
∆U
Axial induction
Strength given by inflow wind speed and thrust curve
Δ𝑈𝑖 =
𝑗≠𝑖
Δ𝑈𝑖𝑗Aggregated induction from all turbines
Wind direction
1. Solve for wakes from upwind to downwind
2. Use resulting CT values to calculate wind speed reduction from blockage Δ𝑈𝑖
3. Run wake model with 𝑈0𝑖 = U0 + Δ𝑈𝑖
Iterate until convergence
Only wake model applied inside Rankine half-body
Turbine self-induction ignored (this is accounted for in the power curve)
8
Coupling wind farm blockage with wakes
9
From single turbine blockage to global blockage
– Each turbine creates an upstream blockage and a downstream wake
10
From single turbine blockage to global blockage
– Each turbine creates an upstream blockage and a downstream wake
– The aggregated blockage from all turbines creates a stronger wind speed reduction
– Seamless integration with wake model
– TORQUE paper:
– J. Phys.: Conf. Ser. 1618 062072 (2020)
– Uses earlier version of the blockage model
11
Comparison with SCADA data
Compare power variation along front row
SCADA
– Filter on inflow at reference turbine
– Inflow wind speed 8±0.5 m/s
– Wind direction in 20° sector
Model
– Adjust freestream wind speed in model to match inflow at reference turbine
– Calculate model results at 1° resolution
– Frequency-weighted average across sector
12
Comparison with SCADA data
13
Complexity from background flow gradients
14
Comparison with SCADA data
– Working on improvements to the model
– Images turbines aloft to simulate atmospheric boundary layer height
– Momentum conservation
– Other initiatives:
15
Outlook
Offshore Wind Accelerator
SCADA data
https://orsted.com/en/our-business/offshore-wind/offshore-operational-data
Wind data
https://orsted.com/en/our-business/offshore-wind/wind-data
16
Ørsted data sharing