© Copyright GustoMSC B.V. 2018. All rights reserved.
Stuttgart Wind Energy (SWE)
@ Institute of Aircraft Design Stuttgart Wind Energy (S
Evaluation of control methods for floating offshore wind turbines
Wei ei YuuYu, Frank Lemmer, David Schlipf, Po Weei YuYu Frank , F
er, Harmen Links, Wen Chen
Neelabhhenngg, arBart isseVi er hh Gupta, Sabrina r H, HHarmen Links
aa Dankemanninkss, nn nn, Neelabhh GuptG
Bernardino ta, Sabrin Gupt
o o Counago
naa ankemanD brin go go, Jose Serna
¾How great is the impact of controller on FOWTs?
¾What makes controlling FOWTs difficult ?
¾How well do the state-of-art control methods work?
2
Background & Motivation
[esteyco]
EU Horizon 2020 project: TELWIND Cost reduction for floating offshore turbine
•Evolved spar concept
•Telescopic tower
•Local and low cost material usage: Concrete
•Simpler manufacturing and installation processes
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design
23.01.2018 23.01.201..001111.222220180180101000011888888888 3
Physical: Negative aerodynamic damping What makes controlling FOWTs difficult ?
3 University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design
Applying conventional on-shore controller to FOWT leads to the instability problem
Larsen, T. J., and Hanson, T. D., 2007. “A method to avoid negative damped low frequent tower vibrations for a floating, pitch controlled wind turbine”. Journal of Physics: Conference Series, 75(1), p. 012073.
Proportional gain: Kp Integral gain: Kp/Ti
-0.03 -0.02 -0.01 0 0.01 0.02 0.03
Open-loop transfer function pole-zero
-0.03 -0.02 -0.01 0 0.01 0.02 0.03
Closed-loop with different gains at 16m/s
low gain median gain high gain
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 4
Control theory: Right-half-plane-zero (RHPZ) What makes controlling FOWTs difficult ?
Transfer Function Wind turbine G(s)
Gen-speedȳ
stable stable unstable
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 5
Selection of theoretical methods
How good do the state-of-art controllers work?
Wind turbine Different control methods used for FOWT by
modifing Baseline controller:
Single-input-single-output (SISO):
Detuning / scheduled detuning
Multi-input-single-output (MISO):
Ptfm damper - feedback of Ptfm-Pitch to Blade-Pitch
Multi-input-single-output (MIMO):
Compensator - feedback of Ptfm-Pitch to Generator torque
Evaluation tool:
Linear analysis: simplified linear mdoel with 5 DOF (SLOW)
Coupled aero-hydro-servo-elastic nonlinear model (Bladed v4.7)
Baseline controller
Detuning method could lead to negative gains at higher wind speed
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 6
Simple approach
1DOF Drivetrain: second order differential system ܫௗ௩߮+ሷ െ߲ܯ
߲ߠ ܭሶ߮+ െ߲ܯ
߲ߠ ܭ
ܶ߮= 0 Eigen-frequency of the drivetrain motion should be lower than the Ptfm eigen-frequency
12 14 16 18 20 22 24
Wind speed [m/s]
0
Wind speed [m/s]
-10 0 10 20
Ti [s]
12 14 16 18 20 22 24 Wind speed [m/s]
0
Wind speed [m/s]
-10 0 10 20
Ti [s]
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 7
Scheduling at different wind speeds SISO: Detuning
Closed-loop with different gains at 16m/s
low gain
Closed-loop with different gains at 24m/s
low gain median gain high gain
RHPZ problem differs from the operating wind speed, thus detuning should be applied according to the operating point
Stable with higher gain at 24m/s unstable
stable
unstable stable
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 8
Trade-off between system stability and control performance SISO: Detuning
Higher stability is at the cost of the control performance.
Closed-loop with different dampers at 16m/s
no damper soft damper median damper hard damper
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 9
How does it work?
MISO: Feedback of Ptfm-Pitch to Blade-pitch
Wind turbine ȳ ߠ increase the pitch stability, however the trade-off between stability and control performance still exist.
Problem with wave
MISO: Feedback of Ptfm-Pitch to Blade-pitch
Wind turbine ȳ ߠ
Due to the difficulty on filtering out the signal in wave frequencies, Ptfm Damper doesn‘t work well for Ptfms with pitch eigen-frequency close to the wave frequencies,
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design
300 350 400 450 500 550 600 650 700
Time [s]
-0.01 -0.005 0 0.005
0.01 PtfmPitch velocity without wave [rad/s]
original filtered
400 420 440 460 480 500 520 540 560 580 600
Time [s]
-0.01 0 0.01
PtfmPitch velocity with wave [rad/s]
original
Bode plot of the Ptfm-pitch velocity filter
low bandwidth
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 11
How does it work?
MIMO: Feedback of Ptfm Pitch to Gen Torque
Wind turbine ȳ ߠ
Closed-loop with different Compensators at 16m/s
no compensate can solve the trade-off problem by moving the positive zero to the left s-place, however will increase the maximum loads on the generator torque.
12
How great is the impact of controller on FOWTs?
Wind: [12, 16, 20, 24] m/s, IEC3-A class Wave:Hs 5.7 [m], Tp 11.5 [s]
0.3 Simple detuning
Scheduled detuning
Wind turbine ȳ ߠ
University of Stuttgart, Stuttgart Wind Energy (SWE) @ Institute of Aircraft Design 13
Conclusion
• System motions and loads are strongly influenced by the controller. These can be significantly reduced by a well designed controller.
• Additional loops can improve the control performance. However, all of the state-of-art approaches have drawbacks.
• Improvement of control performance in wave frequency region is difficult with current sensor and actuators.
phone +49 (0) 711 685-fax +49 (0) 711 685-University of Stuttgart
Thank you!
Wei Viola Yu
68240 yu@ifb.uni-stuttgart.de
EERA DeepWind'2018 - Wednesday, the 17th of January 2018 - V. Leroy 1
V
Vincent LEROY1, 2 PhD Student
J.-C. GILLOTEAUX1, A. COMBOURIEU2, A. BABARIT1, P. FERRANT1
1LHEEA – Centrale Nantes – 1, rue de la Noë – 44321 Nantes - FRANCE
2INNOSEA – 1 rue de la Noë – 44321 Nantes - FRANCE
Impact of the aerodynamic model on the modelling of the behaviour of a Floating Vertical Axis Wind Turbine
Unsteady aerodynamics of a VAWT at sea
EERA DeepWind'2018 - Wednesday, the 17th of January 2018 - V. Leroy 2
D
DeepWind VAWT (Paulsen et al., 2014)
Aerodynamic modelling of VAWTs
AAmongst other theories…
– Inviscid models can usually account for viscous effects with ssemi empirical models
EERA DeepWind'2018 - Wednesday, the 17th of January 2018 - V. Leroy 3
Assumptions Pros Cons
DMS [1]
Double Multiple Streamtube
Steady Inviscid flow Actuator disks
Fast State-of-the-art
Steady Problems at h high TSRs
AC [2]
Actuator Cylinder
Steady, 2D, Inviscid, Incompressible flow
Fast
Accurate cylindrical swept surface
Viscous models added
Steady flow Difficult to go 3D
FVW [3]
Free Vortex Wake + lifting line theory
Potential flow Lifting line
Unsteady aerodynamics Inherent rotor/wake and wake/wake interactions
High CPU cost
CFD
Actuator line + RANS LES, …
Various… More accurate Very high CPU
cost
[1] (Paraschivoiu, 2002) [2] (Madsen, 1982) [3] (Murray et al., 2011)
Which model can we use for a FVAWT ?
EERA DeepWind'2018 - Wednesday, the 17th of January 2018 - V. Leroy 4