Integration of prebend optimization in a holistic wind turbine design tool
Download as PPTX, PDF1 like437 views
The document discusses the integration of prebend optimization into a wind turbine design tool, focusing on enhancing the design framework to allow large blades to optimize shape parameters such as cone, tilt, and prebend. It presents analysis and results from two test wind turbines (3.4 MW and 10 MW), showcasing the impact of prebend on performance metrics like annual energy production and cost of energy. The document concludes with the potential for further studies to refine design constraints and enhance overall wind turbine performance.
![POLI
diMI
tecnico
lano
tecnico
lano
The Science of Making Torque from Wind, 5-7 October 2016
Integration of prebend optimization
in a holistic wind turbine design tool
L. Sartori[1], P. Bortolotti[2], A. Croce[1], C.L. Bottasso[1,2]
[1] Politecnico di Milano
[2] Technische Universität München](https://image.slidesharecdn.com/lucasartoriintegrationofprebendoptimizationinaholisticwindturbinedesigntool-161116082848/85/Integration-of-prebend-optimization-in-a-holistic-wind-turbine-design-tool-1-320.jpg)
![PreBendOptimization
7/17
Prebend optimization setup (ii)
Formulation
Complete MB model of the WT
Prebend shape described by parameters 𝝑
Out-of-plane deflections known from N sensors:
𝜹 𝜼, 𝝑 = [0, 𝛿1 η1, 𝝑 , 𝛿2 η2, 𝝑 , … , 𝛿 𝑁 1, 𝝑 ]
𝒓 𝜼 is the geometric location of the rotor plane
The following optimization problem is performed:
𝝑
𝒎𝒊𝒏
(
𝒊=𝟏
𝑵
𝛿𝑖 η𝑖 , 𝝑 − 𝑟𝑖 η𝑖 )](https://image.slidesharecdn.com/lucasartoriintegrationofprebendoptimizationinaholisticwindturbinedesigntool-161116082848/85/Integration-of-prebend-optimization-in-a-holistic-wind-turbine-design-tool-7-320.jpg)
![PreBendOptimization
8/17
Prebend optimization setup (iii)
Parameterization
• Prebend described by Bézier curves
• The coordinates of the Control Points
are the unknowns of the problem
• Dependency on the number of CPs is
investigated
CP Prebend [m] J [-]
4 6.49 0.0201
5 6.43 0.0190
6 6.49 0.0020
7 6.33 0.0180
8 5.53 0.0376
Out-of-planeposition[m]
Out-of-planeposition[m]
Nondimensional spanwise location [m]Nondimensional spanwise location [m]](https://image.slidesharecdn.com/lucasartoriintegrationofprebendoptimizationinaholisticwindturbinedesigntool-161116082848/85/Integration-of-prebend-optimization-in-a-holistic-wind-turbine-design-tool-8-320.jpg)
![PreBendOptimization
11/17
• Design constrained by both displacement and frequency
• Larger clearance, lower mass (-2 %)
• Lower cone & tilt, higher AEP (+0.4%), lower cost (-0.4%)
• No specific constraints on prebend shape
Applications (ii)
3.4 MW Rotor
3.4 MW rotor Units Straight Prebent Variation
Nacelle uptilt angle [deg] 6.00 4.84 -19.3 %
Rotor cone angle [deg] 4.00 2.02 -49.5 %
Prebend at tip [m] 0.00 3.55 -
Blade/tower clearance [m] 11.70 12.00 +2.5 %
Max tip displacement [m] 8.22 8.62 +4.9 %
Blade mass [kg] 13143 12880 -2.0 %
AEP [GWh/yr] 14.58 14.64 +0.4 %
CoE [$/MWh] 39.67 39.51 -0.4 %](https://image.slidesharecdn.com/lucasartoriintegrationofprebendoptimizationinaholisticwindturbinedesigntool-161116082848/85/Integration-of-prebend-optimization-in-a-holistic-wind-turbine-design-tool-11-320.jpg)
![PreBendOptimization
13/17
• Design constrained only by displacement
• Larger displacement, lower blade mass (-10%)
• Lower tilt but cone unchanged (possible conflict?)
• Prebend higher than technically feasible (further investigations with constraints)
Applications (iv)
10 MW Rotor
10 MW rotor Units Straight Prebent Variation
Nacelle uptilt angle [deg] 4.90 4.20 -14.2%
Rotor cone angle [deg] 2.50 2.50 0.0%
Prebend at tip [m] 0.0 6.50 -
Blade/tower clearance [m] 14.80 20.20 +36.5%
Max tip displacement [m] 10.40 14.40 +38.5%
Blade mass [kg] 45324 40782 -10%
AEP [GWh/yr] 48.98 49.09 +0.6%
CoE [$/MWh] 69.94 69.70 -0.3%](https://image.slidesharecdn.com/lucasartoriintegrationofprebendoptimizationinaholisticwindturbinedesigntool-161116082848/85/Integration-of-prebend-optimization-in-a-holistic-wind-turbine-design-tool-13-320.jpg)
![PreBendOptimization
15/17
Applications (vi)
10 MW Rotor
• Larger torsion caused by
higher torsional response
during DLC6.2
• Further investigation needed
DLC 6.2 @ -30° yaw
BladerootTorsion[kNm]BladerootTorsion[kNm]
10MWPrebend10MWStraight](https://image.slidesharecdn.com/lucasartoriintegrationofprebendoptimizationinaholisticwindturbinedesigntool-161116082848/85/Integration-of-prebend-optimization-in-a-holistic-wind-turbine-design-tool-15-320.jpg)
Integration of prebend optimization in a holistic wind turbine design tool
- 1. POLI diMI tecnico lano tecnico lano The Science of Making Torque from Wind, 5-7 October 2016 Integration of prebend optimization in a holistic wind turbine design tool L. Sartori[1], P. Bortolotti[2], A. Croce[1], C.L. Bottasso[1,2] [1] Politecnico di Milano [2] Technische Universität München
- 2. PreBendOptimization 2/17 Outline Background Cp-Max design framework • Program workflow • A modular approach Prebend optimization setup Applications: 3.4 MW wind turbine Applications: 10 MW wind turbine Conclusions
- 3. PreBendOptimization 3/17 Background Design of large blades is often constrained by tip deflection Current Cp-Max architecture only supports straight blades Prebend is considered through an equivalent cone angle: γgeom Δtip 𝜸 𝒄𝒐𝒏𝒆 = 𝜸 𝒈𝒆𝒐𝒎 + 𝜹𝜸 𝜹𝜸 = 𝒂𝒕𝒂𝒏( ∆ 𝒕𝒊𝒑 𝒍 ) The goal is to include the prebend within the optimization variables, and to find an optimal tradeoff between cone, tilt and prebend.
- 4. PreBendOptimization 4/17 Cp-Max design framework (i) Workflow Background Cp-Max design framework • Global design • Detailed design Prebend optimization Applications: 3 MW wind turbine Applications: 10 MW wind turbine Outlook
- 5. PreBendOptimization 5/17 Cp-Max design framework (ii) A modular approach Macro design: • Some variables are defined at Macro level (cone, tilt, rotor solidity, radius…) • Each evaluation requires a full loop of all the modules • Variations of the macro variables are sensed by the CoE merit function. Detailed (modular) design • Each module performs a detailed design of some features of the WT (aerodynamics, structure, control laws, tower) • Each module requires an individual merit function, in order to optimize the desired performance • Each evaluation requires only to run the module
- 6. PreBendOptimization 6/17 Prebend optimization setup (i) Formulation Prebend optimization located at detailed design level Goal is to maximize the swept area under rated loads Rated conditions known from the synthesis of control laws This criterium is seen by the outer loop as an indirect AEP-maximization
- 7. PreBendOptimization 7/17 Prebend optimization setup (ii) Formulation Complete MB model of the WT Prebend shape described by parameters 𝝑 Out-of-plane deflections known from N sensors: 𝜹 𝜼, 𝝑 = [0, 𝛿1 η1, 𝝑 , 𝛿2 η2, 𝝑 , … , 𝛿 𝑁 1, 𝝑 ] 𝒓 𝜼 is the geometric location of the rotor plane The following optimization problem is performed: 𝝑 𝒎𝒊𝒏 ( 𝒊=𝟏 𝑵 𝛿𝑖 η𝑖 , 𝝑 − 𝑟𝑖 η𝑖 )
- 8. PreBendOptimization 8/17 Prebend optimization setup (iii) Parameterization • Prebend described by Bézier curves • The coordinates of the Control Points are the unknowns of the problem • Dependency on the number of CPs is investigated CP Prebend [m] J [-] 4 6.49 0.0201 5 6.43 0.0190 6 6.49 0.0020 7 6.33 0.0180 8 5.53 0.0376 Out-of-planeposition[m] Out-of-planeposition[m] Nondimensional spanwise location [m]Nondimensional spanwise location [m]
- 9. PreBendOptimization 9/17 Prebend optimization setup (iv) Constraints • Bézier curves ensure the regularity of the shape • Possible constraints account for manufacture and transportability • High-order curves may result in bumps or local minima Max (+) prebend Max (-) prebend Max steepness Horizontal tangent
- 10. PreBendOptimization 10/17 Applications (i) Optimization set-up • 2 test rotors: o 3.4 MW from IEA Task 37 o 10 MW from INNWIND.EU • Macro design: o variables: Tilt, Cone • Prebend design: o variables: Control Points o constraints: None (to check "analytical" optimum) • Structural design: o variables: Thickness of components o constraints: Frequencies separation , Tip displacement, Ultimate stress/strain, Fatigue. DLC Condition Wind Faults 1.1 Power production NTM - 1.3 Power production ETM - 2.1 Power production + fault NTM Grid loss 2.3 Power production + fault EOG Grid loss 6.1 Parked, idling EWM (50y) - 6.2 Parked, idling EWM (50y) Grid loss 6.3 Parked, idling EWM (1y) -
- 11. PreBendOptimization 11/17 • Design constrained by both displacement and frequency • Larger clearance, lower mass (-2 %) • Lower cone & tilt, higher AEP (+0.4%), lower cost (-0.4%) • No specific constraints on prebend shape Applications (ii) 3.4 MW Rotor 3.4 MW rotor Units Straight Prebent Variation Nacelle uptilt angle [deg] 6.00 4.84 -19.3 % Rotor cone angle [deg] 4.00 2.02 -49.5 % Prebend at tip [m] 0.00 3.55 - Blade/tower clearance [m] 11.70 12.00 +2.5 % Max tip displacement [m] 8.22 8.62 +4.9 % Blade mass [kg] 13143 12880 -2.0 % AEP [GWh/yr] 14.58 14.64 +0.4 % CoE [$/MWh] 39.67 39.51 -0.4 %
- 12. PreBendOptimization 12/17 Applications (iii) 3.4 MW Rotor • Key Loads globally reduced: Blade root Mxy (-3%) Hub Myz (-26%) • Further mass reduction prevented by frequency constraint
- 13. PreBendOptimization 13/17 • Design constrained only by displacement • Larger displacement, lower blade mass (-10%) • Lower tilt but cone unchanged (possible conflict?) • Prebend higher than technically feasible (further investigations with constraints) Applications (iv) 10 MW Rotor 10 MW rotor Units Straight Prebent Variation Nacelle uptilt angle [deg] 4.90 4.20 -14.2% Rotor cone angle [deg] 2.50 2.50 0.0% Prebend at tip [m] 0.0 6.50 - Blade/tower clearance [m] 14.80 20.20 +36.5% Max tip displacement [m] 10.40 14.40 +38.5% Blade mass [kg] 45324 40782 -10% AEP [GWh/yr] 48.98 49.09 +0.6% CoE [$/MWh] 69.94 69.70 -0.3%
- 14. PreBendOptimization 14/17 Applications (v) 10 MW Rotor • Spar mass heavily reduced • Some Key Loads are higher • This is probably related to higher torsional moments along the blade
- 15. PreBendOptimization 15/17 Applications (vi) 10 MW Rotor • Larger torsion caused by higher torsional response during DLC6.2 • Further investigation needed DLC 6.2 @ -30° yaw BladerootTorsion[kNm]BladerootTorsion[kNm] 10MWPrebend10MWStraight
- 16. PreBendOptimization 16/17 Conclusions Remarks: Prebend design has been integrated in Cp-Max Aero-structural advantages can be obtained from the simultaneous design of cone, tilt and prebend The max swept area criteria has an impact on the AEP, so it follows the overall cost-minimization strategy Outlook: More studies with the technical/manufacture constraints on prebend shape (ongoing) Include the design of blade sweep (ongoing) Complete WT design including radius, chord, twist, blade thickness (short- term) Test if prebend at a macro instead of detail level can lead to better results (medium-term)
- 17. PreBendOptimization 17/17 Thank you for your attention