Embed Size (px)
Citation preview
“The science of making torque from wind”, Oldenburg, October 2012
Wind turbine control applications of turbine-mounted LIDAR
E A Bossanyi, A Kumar and O Hugues-Salas
An independent study
• Much recent interest in LIDAR for wind turbine control
• Some dramatic claims of e.g. increased energy capture
• Potential for reduced loads
• Need for an independent and objective study:
• Co-funded by GL-GH and two leading LIDAR suppliers
• Completed Summer 2012
Two key objectives:
• To evaluate the likely benefits of adding LIDAR to the wind turbine controller
• To provide advice to LIDAR manufacturers about the characteristics of LIDAR
systems which are most likely to be of value for this application
Project outline
• Develop enhanced simulation modelling capability, covering many
LIDAR types
• Develop algorithms for processing the raw LIDAR signals
• Initial screening of many LIDAR configurations by testing their ability to
estimate rotor-averaged quantities (wind speed, direction and shear
gradients)
• Develop simple control algorithms to make use of a few selected
configurations to improve the wind turbine control action
• Evaluate the performance of the LIDAR-assisted controllers using
detailed loading simulations carried out in accordance with the IEC
Edition 3 standard.
Actual at hub(=measured ifFrozen Turbulence)
Measured (UnfrozenTurbulence)
Lon
gitud
inal
velo
city [
m/s
]
Time [s]
9
10
11
12
13
14
15
16
17
18
0 2 4 6 8 10 12 14 16 18 20
Enhanced simulation modelling capability
• Along-wind decorrelation: abandoning Taylor’s frozen turbulence hypothesis
• Paper presented at EWEA 2012
• Avoids ‘cheating’ in simulations
• Many different LIDAR types …
LIDAR types modelled
• CW LIDAR with various focal distances and values
• Pulsed LIDAR with various numbers of simultaneous focal distances (up to 10)
• Various sampling rates
• Single staring beam
• 2, 3, 4, 6 or 8 fixed beams (pulsed only)
• Simultaneous or sequential switching of focal distances
• Single circular scanning beam with various angles between beam and centreline, and
various numbers of samples per scan
• Single beam performing rosette or Lissajous scan
• Nacelle mounted, spinner or blade mounted options
Weighting function with alpha = 0.00055 for different ranges
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 50 100 150 200 250
Distance from lens (m)
50 75 100 150 200
Weighting function for different Alpha at 75m range
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 20 40 60 80 100 120 140 160 180 200
Distance from lens (m)
0.00055 0.001 0.0025 0.005 0.01
0
0.05
0.1
0.15
0.2
0.25
-30 -20 -10 0 10 20 30
Distance from focus point (m)
Weig
hti
ng
fu
ncti
on
(no
rmalised
)
050
100150
-100
-50
0
50
100-80
-60
-40
-20
0
20
40
60
80
050
100150
-100-50
050
100-80
-60
-40
-20
0
20
40
60
80
Processing the raw LIDAR signals
• LIDAR measures component along the beam-line only (VLOS)
• Useful to measure at many points in rotor swept area
• Many possible algorithms to extract rotor-average values from these measurements:
• Longitudinal wind speed (U)
• Wind direction () • Vertical and horizontal shear gradients (UZ UY)
• Least-squares algorithm chosen: assumes uniform U, , UZ, UY and that varies
more slowly than UY
}Difficult to distinguish , UY
Initial screening of LIDAR configurations
• 10-minute simulation
• 13 m/s with IEC class 1A turbulence
• Superimposed sinusoidal direction transient: +15º over 10 minutes to exercise the
direction estimation
• Compared LIDAR estimate against true rotor-averaged quantities
Initial screening: examples
True rotor-averaged value Lidar value
Longitudin
al w
ind s
peed [
m/s
]
Time [s]
8
10
12
14
16
18
20
0 100 200 300 400 500 600
True rotor-averaged value Lidar value
Vert
ical shear
[1/m
in]
Time [s]
-1
-2
-3
0
1
2
3
4
5
6
7
0 100 200 300 400 500 600
True rotor-averaged value Lidar value
Hori
zonta
l shear
[1/m
in]
Time [s]
-1
-2
-3
-4
0
1
2
3
0 100 200 300 400 500 600
True rotor-averaged value Lidar value
Dir
ection [
deg]
Time [s]
-5
-10
-15
-20
-25
0
5
10
15
20
25
0 100 200 300 400 500 600
Longitudinal wind speed Vertical shear gradient
Horizontal shear gradient Direction
Initial screening: conclusions
• Both CW and pulsed systems can give good results
• Need good coverage of swept area: at least 10 points in 1 second
• Trade-off between no. of points and time taken to scan them all
• No advantage of complex scans over circular scan
• Large beam angle to centreline is better for direction, worse for wind speed & shear;
a single configuration won’t be the best for everything
• Sharp focus is not always advantageous
• Spinner mounting is as good as nacelle mounting: can correct for azimuth & tilt, and
no blade blockage
• Blade mounting also works well (one example at 70% radius)
• Easy to get good estimate of longitudinal wind speed
• Vertical shear: not bad, assuming mean upflow is known
• Horizontal shear and direction: get reasonable separation if direction is assumed
slowly-varying.
Possibilities with LIDAR-assisted control
• Improved energy capture due to better yaw tracking?
• Probably not much – but very useful for wind vane calibration!
• Yaw control is usually slow (yaw motor duty, gyroscopic loads, etc.)
• Pay attention to convention yaw tracking strategies first
• Improved energy capture due to better Cp tracking?
• Tiny improvement, outweighed by large power & torque variations
• Reduced extreme loads due to anticipation of extreme gusts?
• Promising but difficult to assess
• Reduced fatigue loads due to anticipation of approaching wind field?
• Improved collective pitch control yields easy benefits
• More marginal for individual pitch control
Reduced loads implies a potential for re-optimisation of turbine design
Improved cost-effectiveness for future designs
Collective pitch control:
Very simple feed-forward implementation
-
LIDAR measurements Modification to
control action
generator
speed
set-point
Measured
generator speed,
tower top
acceleration, etc. pitch angle
Measured generator speed
Other measured signals
Wind field
estimation
algorithms
Bladed simulation
(or real turbine)
LIDAR-based
feed-forward
control action
Feedback controller
Improved collective pitch control with LIDAR
• Immediate improvement in speed regulation
• Take the benefit by reducing control gains
→ Calmer pitch action
→ Lower loads (especially tower bending moments)
Base PI Base PI + Lidar Reopt + Lidar Reopt, no Lidar
Roto
r speed
[rpm
]
Time [s]
10.0
10.5
11.0
11.5
12.0
12.5
13.0
13.5
0 100 200 300 400 500 600
Improved collective pitch control with LIDAR
• Re-optimised PI with LIDAR (red line) achieved similar speed control to baseline
controller but with lower PI gains
• Pitch movements are reduced
• The pitch controller anticipates the increases in wind speed
• The pitch movements start earlier and so have smaller peaks
Base PI Reopt + Lidar
Mean p
itch a
ngle
[d
eg]
Time [s]
-2
0
2
4
6
8
10
12
14
0 100 200 300 400 500 600
Improved collective pitch control with LIDAR
• Reduced pitch movements result in reduction in thrust variation
• Thrust related loads on turbine are reduced, for example: tower base
bending moment.
Base PI Reopt + Lidar
Tow
er
My
[M
Nm
]
Time [s]
20
30
40
50
60
70
80
90
100
110
0 100 200 300 400 500 600
Improved collective pitch control with LIDAR
Blade root load reduction
0
0.51
1.52
2.53
3.54
4.5
Mx My Mz Fx Fy Fz
% r
ed
uct
ion
SN 4 (Steel)
SN 10 (GRP)
Shaft load reduction (SN 4)
0
2
4
6
8
10
12
14
Mx My Mz Fx Fy Fz
% r
ed
uct
ion
Yaw bearing load reduction (SN 4)
-2
0
2
4
6
8
10
12
14
Mx My Mz Fx Fy Fz
% r
ed
uct
ion
Tower base load reduction (SN 4)
-2
0
2
4
6
8
10
12
14
Mx My Mz Fx Fy Fz
% r
ed
uct
ion
Even very simple methods achieve significant reduction in thrust-related fatigue loads
• 20% reduction in above-rated wind speeds
• 14% lifetime fatigue load reduction, e.g. tower base bending moment
Improved collective pitch control with LIDAR Extreme load reduction is much harder to assess:
• Extreme gusts not realistic – and how do they convect and evolve?
• LIDAR must be working at moment of extreme load
• Affected by meteorological conditions? (Fog, precipitation, lack of
aerosols)
• Extreme gusts may not be design drivers
Now more emphasis on extreme turbulence:
• DLC1.1: indicates reduction in extreme tower base overturning
moment:
0.001
0.01
0.1
1
35000 55000 75000 95000 115000
Tower base My (kNm)
Pro
bab
ilit
y o
f exceed
an
ce
Base case
LIDAR (typical range)
0.001
0.01
0.1
1
5000 7000 9000 11000 13000 15000 17000
Blade root My (kNm)
Pro
ba
bilit
y o
f e
xc
ee
da
nc
e
Base case
LIDAR (typical range)
Improved IPC with LIDAR?
0 20 40 60 80 100 120 140
Blade root My moment (steel)
Blade root My moment (GRP)
Shaft My moment (steel)
Shaft Mz moment (steel)
Tower top nod moment (steel)
Tower top yaw moment (steel)
Increase in pitch travel
Decrease in loads or increase in pitch travel (%)
Conventional IPC
LIDAR IPC
Both together
• LIDAR estimates the vertical & horizontal shear
• Very simple strategy some reduction of asymmetrical loads (without
needing load sensors)
• Not as effective than using load sensors, but more sophisticated strategies
would be possible.
Improved CP-tracking with LIDAR?
Rotor speed tracks wind speed better
Needs huge power/torque swings to accelerate/decelerate rotor
Tiny fraction of % increase in power production – not worth it! RPM (No Lidar) RPM (Lidar) Rotor average wind
speed, m/s
m/s
o
r
RP
M
Time [s]
5
6
7
8
9
10
11
12
13
0 100 200 300 400 500 600
No Lidar Lidar
Ele
ctr
ical pow
er
[M
W]
Time [s]
0
1
2
3
4
5
6
0 100 200 300 400 500 600
Improved yaw tracking with LIDAR?
• Probably not much – but could be a very useful commissioning tool for wind vane
calibration! (Calibration required as a function of operating point.)
• Yaw control has to be slow (yaw motor duty, gyroscopic loads, etc.)
• Pay more attention to convention yaw tracking strategies first: should not be losing
more than 0.5 – 1% of energy compared to ‘ideal’ continuous yawing
15s
30s45s
0
2
4
6
8
10
12
14
0 0.02 0.04 0.06 0.08 0.1
Mean absolute yaw rate [deg/s]
RM
S y
aw
mis
ali
gn
men
t [d
eg
]
No Lidar
Lidar
Mixed
10-minute simulation
(but really depends
on low-frequency
variations which are
site-dependent)
Conclusions
• Enhanced LIDAR modelling capabilities in Bladed
• LIDAR can reduce loads significantly, even with very simple control algorithms
• Collective pitch control enhancement: 14% lifetime tower fatigue reduction
• Extreme load reduction, with caveats
• IPC also possible
• Cp-tracking and yaw control: benefits much less clear (but LIDAR could certainly be a
useful commissioning tool, e.g. for wind vane calibration
• Both pulsed and CW LIDAR are suitable if configured appropriately
• Need at least ~10 points in swept area, sampled every second, to give a few seconds
of look-ahead time
Acknowledgements
Many thanks to:
• Natural Power (Mike Harris, Chris Slinger)
• Avent Lidar Technology (Samuel Davoust, Thomas Velociter)
for financial contributions and many valuable discussions.
