Modelling of static and fatigue failure in wind turbine blades using a parametric blade model

1

• Modelling of static and fatigue failure in wind
  turbine blades using a parametric blade model
• A G Dutton, M Clarke1, P Bonnet2
• Energy Research Unit (ERU)
• Rutherford Appleton Laboratory (RAL)
• Science and Technology Facilities Council (STFC)
• (now at 1Oxford Brookes University and 2SAMTECH
  Iberica)
• Presented at EWEC 2010, Warsaw, 23 April 2010

2
Background SUPERGEN Wind
To undertake research to improve the
cost-effective reliability availability of
existing and future large scale wind turbine
systems in the UK

• Research Themes
• Baselining wind turbine performance
• Drive-train loads and monitoring
• Structural loads and materials
• Environmental issues

3
Background Blade modelling

• Which are the best materials?
• What is the optimum lay-up?
• What is the best internal structure?
• What are the size limits for wind turbine blades?
• What additional stresses do smart control devices
  generate in a blade?
• How should NDT measurements be interpreted?

Picture credit LM Glasfiber
Picture credit EWEA
4
Parametric blade modelDesign strategy

• Parametric processing tool for creation and
  running of the underlying FE model
• Suitable for sensitivity analyses, flexibility,
  documenting, re-usability
• Python script front end for automation of the
  Abaqus FE package
• Modular program
• Realistic load application, including
  quasi-static aerodynamic loading
• Ultimate strength fatigue analysis
• Developing dynamic implementation

5
Parametric blade modelGeometry definition
6
Parametric blade modelGeometry definition
7
Parametric blade modelLay-up
8
Parametric blade modelFully distrubuted
aerodynamic load
9
Parametric blade modelVariable mesh density...
... at the push of a button
10
Parametric blade model
11
5 MW (61 m) blade model

• Basic lay-up information
• Target mass and stiffness distributions
• Limitations of lay-up information
• Overall mass
• Discretisation of lay-up info
• Required spar-cap stress profile?
• Lay-up modification
• Materials variation
• Static load case (aerodynamic load distribution)
• Fatigue lifetime

12
5 MW (61 m) blade modelSpar-cap stress
distribution (smoothed)
13
5 MW (61 m) blade modelMaterials
Material property Baseline UD material High fatigue strength material
E1T (GPa) 39.0 56.3
E1C (GPa) 38.9 -
?12 0.29 0.25
E2T (GPa) 14.1 9.0
E2C (GPa) 14.997 -
?21 0.95036E-01 0.95036E-01
G12 (MPa) 4.24 4.24
Material property Baseline UD material High fatigue strength material
XT (MPa) 776.5 1757
XC (MPa) -521.8 -978
YT (MPa) 54 54
YC (MPa) -165 165
S (MPa) 56.1 135.4
Fatigue Baseline UD High fatigue strength
S-n curve at R0.1 S0 1176 b 9.74 S0 1250 b 10.59
14
5 MW (61 m) blade modelStatic strength skins
and shear web

• Choice of static failure criteria
• Tsai-Wu
• Tsai-Hill
• Other (user specified)

15
5 MW (61 m) blade modelStatic strength skins
and shear web

• Choice of static failure criteria
• Tsai-Wu
• Tsai-Hill
• Other (user specified)

16
5 MW (61 m) blade modelStatic strength
bonding paste

• Cohesive element model
• Normal stress component
• Shear stress component
• Linear up to characteristic value
• Material softening

17
5 MW (61 m) blade modelFatigue strength
estimation

• Complex loading
• Stochastic / semi-deterministic (cyclic) loading
• Biaxial (triaxial) stress state

• Fatigue characterisation
• Predominantly uni-directional materials data
• Uncertainty in how best to combine different
  stress cycles
• R-ratio (minimummaximum stress in a load cycle)
• Combine into constant life diagram

18
5 MW (61 m) blade modelFatigue strength
estimation
Constant life diagram - Linear Goodman diagram
19
5 MW (61 m) blade modelFatigue strength
estimation
Constant life diagram - Multiple R-values diagram
20
5 MW (61 m) blade modelFatigue strength
estimation
Constant life diagram - Multiple R-values diagram
21
5 MW (61 m) blade modelFatigue strength
estimation

• Complex loading
• Stochastic / semi-deterministic (cyclic) loading
• Biaxial (triaxial) stress state

• Fatigue characterisation
• Predominantly uni-directional materials data
• Uncertainty in how best to combine different
  stress cycles
• R-ratio (minimummaximum stress in a load cycle)
• Combine into constant life diagram
• applies to a single material direction

• How to deal with complex stress states?

22
5 MW (61 m) blade modelBiaxial stress ratio

• Biaxial stress ratio is the ratio between the two
  largest magnitude principal stress components

23
5 MW (61 m) blade modelFatigue strength
estimation
24
5 MW (61 m) blade modelFatigue lifetime

25
Full scale blade testingThermoelastic stress
analysis
Blade test blade with defects
26
Full scale blade testingThermoelastic stress
analysis
Blade test blade with defects
27
Conclusions

• Flexible, parametric blade model for assessment
  of alternative materials
• Simple failure model in blade skin and developing
  damage model in bonding paste implemented
• Fatigue methodology under development
• Initial results also available for application to
  full-scale blade testing, control of smart blades
  and interpretation of condition monitoring data
• Future work planned on dynamic loading
  operation in wakes from upstream turbines
  smart blade devices

28
Acknowledgements

• EPSRC grant no. EP/D034566/1
• SUPERGEN Wind Energy Technologies Consortium

For further information please contact geoff.dutt
on_at_stfc.ac.uk