Composite materials

for wind turbine blades:

issues and challenges
Francesco Aymerich
Department of Mechanical, Chemical and Materials Engineering
University of Cagliari, Italy

SYSWIND Summer school – July 2012 – University of Patras

Department of Mechanical, Chemical and Materials Engineering 1

University of Cagliari, Italy
 Outline of presentation

• Use of composites in wind turbine blades

• Manufacturing processes
• Mechanisms of damage and failure in composite materials
• Strength analysis and damage tolerance approach
• Composites for delamination resistance
• Carbon fibres in wind turbine blades
• Summary

Department of Mechanical, Chemical and Materials Engineering 2

University of Cagliari, Italy
 Wind turbine blades
Wind turbine blades are complex structures whose design involves the
two basic aspects of
• Selection of the aerodynamic shape
• Structural configuration and materials selection (to ensure that
the defined shape is maintained for the expected life)

Modern blades
- consist of different kinds of materials (typically composite materials in
monolithic or sandwich configuration)
- use various connections solutions between different substructures
- include many material or geometric transitions
Department of Mechanical, Chemical and Materials Engineering 3
University of Cagliari, Italy
 Growth of blade mass with blade length
The growth rate of blade mass with length has been reducing in the past decades
Key drivers for reduction:
• Improved manufacturing processes
• Introduction of new materials
• More efficient use of materials and improved structural configurations

[Lekou, 2010]

Department of Mechanical, Chemical and Materials Engineering 4

University of Cagliari, Italy
 Main loads on blades
The main loads on the blades are generated by wind and by gravity.

Wind loads mainly induce both flapwise and edgewise bending.

These loads have both a static and a dynamic component (variations in wind speed and
natural wind shear) that induce fatigue on the blade material.

Gravity loads mainly induce edgewise bending, when the blade is horizontal.
The rotation of the blades cause alternating edge-wise bending and thus fatigue of the
material.

Department of Mechanical, Chemical and Materials Engineering 5

University of Cagliari, Italy
 Cross-section of a blade
The cross-section of a blade consists essentially of :

- Outer shells (ensure the stability of the aerodynamic shape)

- Internal structural support of the outer shells (longitudinal beam
or webs)
OUTER SHELL

LEADING
EDGE

TRAILING
EDGE
OUTER SHELL
INTERNAL
STRUCTURAL
SUPPORT
Department of Mechanical, Chemical and Materials Engineering 6
University of Cagliari, Italy
 Cross-section concepts: main spar
The two aeroshells are bonded to a load-carrying spar-beam (box-beam)
The main spar and the wing shells are manufactured separately and then
joined in a separate bonding process.

FLANGES : THICK
MONOLITHIC COMPOSITE
SHELL: SANDWICH
(0°/±45° fibres) (0° fibres)

MAIN
SPAR
WEBS: SANDWICH
(±45° fibres)
ADHESIVE
BONDING
ADHESIVE BONDING
Department of Mechanical, Chemical and Materials Engineering 7
University of Cagliari, Italy
 Cross-section concepts: Internal stiffeners
The two aeroshells are bonded to two or more internal webs (stiffeners).
The wing shells are manufactured with relatively thick monolithic composite
laminates (spar-caps).
THICK MONOLITHIC
ADHESIVE BONDING COMPOSITE

INTERNAL STIFFENERS
ADHESIVE BONDING
Department of Mechanical, Chemical and Materials Engineering 8
University of Cagliari, Italy
 Cross-section concepts: Integral stiffeners
The entire blade structure, including internal webs/stiffeners, is
manufactured in one single process (no secondary bonding).

THICK MONOLITHIC
COMPOSITE

INTEGRAL STIFFENER

Department of Mechanical, Chemical and Materials Engineering 9

University of Cagliari, Italy
 Key structural design requirements

• the blades must be strong to resist the extreme (ultimate) loads;

• the blades must resist the time-varying (fatigue) loads through
the entire life of the blade;
• the blades must be stiff to prevent collision with the tower under
extreme loads. Local stiffness must be also sufficient to prevent
instability of components under compression (to avoid local or global
buckling)

• the blade construction needs to be as light as possible to

minimize the cost of generated power
• the blades should be stiff and light to avoid resonance

Department of Mechanical, Chemical and Materials Engineering 10

University of Cagliari, Italy
 Materials requirements
The structural design requirements translate to the following
materials requirements in terms of material properties:

• High material strength is needed to withstand the extreme

loads
• High fatigue strength is needed to resist varying loads and
reduce material degradation during service
• High material stiffness is needed to maintain aerodynamic
shape of the blade, to prevent collision with the tower, and to
prevent local instability (buckling) under compressive loads
• Low density is needed to reduce gravity forces and to
minimize the cost of power

Department of Mechanical, Chemical and Materials Engineering 11

University of Cagliari, Italy
 Why composite materials on blades?

MATERIAL
INDEX

m = AL ρ
 ρ 
1/ 2
 12 S 
( )
Mass of the beam

F EI m=  L  1/ 2 
3

Stiffness of the beam S = =k 3  kL  E 

δ L

1/ 2
To minimize mass for a given E
ρ
stiffness S we have to maximize

Department of Mechanical, Chemical and Materials Engineering 12

University of Cagliari, Italy
 Why composite materials?
Minimize mass for assigned stiffness

Line with constant E1/2/ρ values

Ashby plot

Department of Mechanical, Chemical and Materials Engineering 13

University of Cagliari, Italy
 Why composite materials?
Minimize mass for assigned strength

Line with constant σf2/3/ρ values

Ashby plot

Department of Mechanical, Chemical and Materials Engineering 14

University of Cagliari, Italy
 Composite materials
A composite material consists of two or more materials combined to
obtain properties different from those of the individual materials.
- Reinforcing fibres (to add strength and stiffness )
- Matrix (holds and protects fibres, and distributes the load)

Polymer Matrix Composite (PMC) materials are typically used in wind

turbine blades.

Department of Mechanical, Chemical and Materials Engineering 15

University of Cagliari, Italy
 Polymer Matrix
Thermoset materials are obtained from a chemical reaction between the
resin and the hardener to form a hard infusible product.
Polyester : easy to process (does not require post curing), inexpensive
Vinyl ester : cost and strength intermediate between polyester and epoxy
Epoxy : best mechanical properties, less shrinkage, expensive
Advantages Disadvantages
Easy to process Long curing times – Limited toughness
Exotherm during curing (thick components)

Thermoplastic materials soften and melt with heating, then hardening

again with cooling. The softening process can be repeated without any
significant degradation of the material properties.
PP or L-PET: used in film or fibre form and consolidated by heating and vacuum
Reactive thermoplastics (APA-6): suitable for liquid moulding (similar to thermosets)
Advantages Disadvantages
High toughness - Recyclability High process temperature and pressure

Department of Mechanical, Chemical and Materials Engineering 16

University of Cagliari, Italy
Strength and stiffness of polymer resins

[Gurit, 2012]

Department of Mechanical, Chemical and Materials Engineering 17

University of Cagliari, Italy
 Degradation of resin from water ingress
The absorption of water affects the resin and the resin/fibre
interface. leading to gradual reduction of mechanical properties.

Strength
retained: 85%

Strength
retained: 65%

Short-beam shear strength

test on a glassfibre laminate

[Gurit, 2012]

Department of Mechanical, Chemical and Materials Engineering 18

University of Cagliari, Italy
 Reinforcing fibres
Typical reinforcement used in composite materials are stiff, strong and lightweight fibres
such as
Glass fibres (good specific strength, low specific stiffness, relatively inexpensive)
Carbon fibres (high specific strength and stiffness, expensive)
Aramid fibres (hygroscopic, low compression strength, few data on fatigue)

Department of Mechanical, Chemical and Materials Engineering 19

University of Cagliari, Italy
 Properties of reinforcing fibres
Material Tensile Young’s Density Specific
strength modulus E (g/cm3) stiffness
(MPa) (GPa) E/ρ

HS Carbon 3500 160-270 1.8 90 - 150

IM Carbon 5300 270-330 1.8 150 - 180
UHM Carbon 2000 >440 2.0 > 220
E-Glass 2400 69 2.5 28
S-Glass 3450 86 2.5 34
Aramid LM 3600 60 1.45 41
Aramid UHM 3400 180 1.45 125
Aluminium 7020 400 70 2.7 26
Mild Steel 450 210 7.8 26
HS Steel 1250 200 7.8 25

Department of Mechanical, Chemical and Materials Engineering 20

University of Cagliari, Italy
 Comparative fibre cost
Comparison of fibre cost for unidirectional fabrics (300 g/m2)

Approximative cost of E- glass = ~ 2.5 Euro/m2 (~ 3.1 USD/m2)

[Gurit, 2012]
Department of Mechanical, Chemical and Materials Engineering 21
University of Cagliari, Italy
 Final properties of a composite
The final properties of the composite are mainly determined by
• Properties of fibres
• Properties of matrix
• Percentage of fibres (fibre volume fraction, typically ranging from 35% to 65%)
• Orientation and geometry of fibrous reinforcement

For example, stiffness properties are strongly dependent on fibre orientation and fibre fraction

FIBRE ORIENTATION FIBRE FRACTION

Department of Mechanical, Chemical and Materials Engineering 22

University of Cagliari, Italy
 Tensile and compressive properties
of unidirectional composites

Department of Mechanical, Chemical and Materials Engineering 23

University of Cagliari, Italy
 Reinforcement architecture
Common geometries of the fibrous reinforcement include

Continuous or chopped
strand mat (CSM)

Unidirectional fibres

Woven fabrics
Fabric type reinforcement
Stitch bonded (Non crimp) fabrics

Department of Mechanical, Chemical and Materials Engineering 24

University of Cagliari, Italy
 Woven fabrics
Woven fabrics are obtained by interlacing yarns of fibres with
different orientations (usually 0° (warp) , 90°(weft), and ±45°)

Plain weave Twill weave

(1 warp yarn (1 warp yarn
over 1 weft yarn) under 3 weft yarns)

Woven fabrics (vs unidirectional layers)

ADVANTAGES DISADVANTAGES
Higher stability for fibre placement Lower fibre fraction
Laminates have higher resistance to Lower in-plane properties (crimped fibres
crack propagation and stress concentrations)
More difficult to infuse with resin

Department of Mechanical, Chemical and Materials Engineering 25

University of Cagliari, Italy
 Non Crimp Fabrics (NCF)
Non crimp fabrics (NCFs) are obtained by stitching together unidirectional
yarns with different orientations, using non-structural threads.
+45°
90°
NCF fabric
-45°

Woven fabric

Non crimp fabrics (vs woven fabrics)

ADVANTAGES DISADVANTAGES
Higher fibre fraction Stitching may induce fibre fracture
Higher stiffness/strength (straight fibres)
No stress concentration due to fibre
waviness
Easier lay-up (fewer layers)

Department of Mechanical, Chemical and Materials Engineering 26

University of Cagliari, Italy
 Basic structural configurations used in blades

Monolithic laminates
consist of different layers
(plies) of multidirectional
fabrics or unidirectional fibres

Sandwich composites
consist of of a low density core between
thin faces (skins) of composite material.

Polymeric (PVC, PET, PMI) foams with

density in the range 40-200 kg/m3

Department of Mechanical, Chemical and Materials Engineering 27

University of Cagliari, Italy
 Sandwich composites
The insertion of a core increases the thickness of the structure (and thus
flexural stiffness and strength) without increasing its weight.

The skins carry the

tensile and the The core carry
compressive loads the shear load

Weight 1 ~1 ~1
Bending stiffness 1 ~12 ~48
Bending strength 1 ~6 ~12
Department of Mechanical, Chemical and Materials Engineering 28
University of Cagliari, Italy
 Manufacturing techniques
for composite blades
•Wet hand layup (laminating technique)
•Filament winding
•Resin infusion Potential for automation
•Prepregs
WET HAND LAYUP VACUUM INFUSION
Automation
FILAMENT WINDING PREPREG MATERIALS

Minimization of cycle times and cost

Reduction of defects
Improved structural performance
Department of Mechanical, Chemical and Materials Engineering 29
University of Cagliari, Italy
 Wet hand layup (laminating technique)
Dry fibre material (mats, fabrics or unidirectional tapes) are laid in various
layers into the mould of the component.
The layers are then impregnated with resin and cured at room or higher
temperature (70° to 100° C for epoxy).

[Brøndsted et al., 2005]

Advantages Disadvantages

• Production of complex shapes. • The process is labor intensive and

• Fibres can be oriented along preferred time-consuming (hand–made)
directions • Large amount of voids and defects
• Low fibre fraction
Department of Mechanical, Chemical and Materials Engineering 30
University of Cagliari, Italy
 Filament winding
The fibres are passed through a resin bath and are then wound onto a
rotating mandrel.
The process is primarily used for cylindrical components
but can be adapted for blade manufacturing

[Bussolari, 1983]

Advantages Disadvantages
• The process can be carried out in an • Different mandrels must be used to
automatic way. gradually build the airfoil.
• Fibres cannot be easily oriented along
the axis of the blade (0° direction).
Department of Mechanical, Chemical and Materials Engineering 31
University of Cagliari, Italy
 Resin Infusion Techniques
Dry fibres (mats, fabrics or unidirectional tapes) are placed in a mould and
encapsulated in a vacuum bag. Liquid resin is then pulled through the
reinforcement by vacuum and allowed to cure at room or higher temperature.
[Grande, 2008]

Advantages Disadvantages
• Large components can be made in a • Relatively complex process (especially
single step for large components)
• Clean and safe process • Low viscosity resins should be used
• Good final material quality (resulting in lower mechanical properties)
• Potential for automation [Ref]

Department of Mechanical, Chemical and Materials Engineering 32

University of Cagliari, Italy
 Blade Infusion
- The two airfoils and the webs or spar are usually manufactured separately
and subsequently bonded to complete the blade.
- In some technologies however the full blade is infused in a single step.

[Hogg, 2010]

Department of Mechanical, Chemical and Materials Engineering 33

University of Cagliari, Italy
 Resin Infusion Techniques
Key issues for final quality of resin infusion are:
• Improvement of fibre impregnation (to avoid regions with dry fibres)
• Reduction of voids

Possible ways to tackle these issues are

Selection of appropriate fibre coating/sizing to improve wettability

Use of low viscosity resins (at room or moderate temperature) to improve

wettability and reduce the process time for large components

Use of fibre fabrics with special architecture (or special resin distribution
meshes) to facilitate flow of the resin

Improvement of resin flow (optimal placement of inlet and outlet lines for
resin by simulation of the flow and data from sensors)

Department of Mechanical, Chemical and Materials Engineering 34

University of Cagliari, Italy
 Prepreg technology
Prepreg tapes consist of fibre fabrics pre-impregnated with a resin that is not
fully cured.
The prepregs are laid up onto the mould surface, vacuum bagged and then
heated. The pressure required to consolidate the stacked layers of prepregs is
achieved by vacuum. Process temperatures range between 70°C and 120°C.

Advantages Disadvantages
• High fibre ratio and low void content • High cost for prepreg material
• Consistent material properties • Tooling must whitstand process
• Easy control of fibre alignment temperature
• Large components can be made in a • It is difficult to correctly cure thick
single step laminates (temperature not uniform
[Ref]
• Clean and safe process through-the-thickness)
Department of Mechanical, Chemical and Materials Engineering 35
University of Cagliari, Italy
 Automation of blade manufacturing
Manufacturing of turbine blades consists of a combination of manual,
labour-intensive operations
• Fabrics pattern cutting
• Lay-up
• Vacuum bagging
• Infusion
• Demoulding
• Secondary bonding

To reduce labour and manufacturing time, and

improve quality the trend is toward automation
Automated Cutting - Bagging
Automated Tape Layup (ATL)
Automated Bonding
[Black, 2009]
Department of Mechanical, Chemical and Materials Engineering 36
University of Cagliari, Italy
MECHANISMS OF FAILURE IN
COMPOSITE MATERIALS

Department of Mechanical, Chemical and Materials Engineering 37

University of Cagliari, Italy
Mechanisms of failure in composite materials
Failure in composites can be examined at different scales
(fibre/matrix/interface level; ply level; laminate level, etc.).

Strength analyses of composite structures carried out at the laminate level may
often lead to unsafe predictions.
Most adopted criteria have been developed to estimate failure at the ply level.

First-ply-failure (FPF) is often used as a criterion for laminate strength, but

this is often a very conservative approach and there may be a large
distance between the load for FPF and the collapse load of the laminate
(LPF – Last Ply Failure).

Final collapse of a laminate is the result of the accumulation of different

damage modes, which can induce significant degradation of the material
properties during life.

Department of Mechanical, Chemical and Materials Engineering 38

University of Cagliari, Italy
 Idealized stress-strain curve
for a [0/+45/-45/90]s laminate

Department of Mechanical, Chemical and Materials Engineering 39

University of Cagliari, Italy
Main failure modes in composite materials

COMPOSITE LAMINATES
• Fibre failure
• Matrix failure
• Fibre-matrix debonding
• Inter-laminar failure (delamination)
• Buckling instability

COMPOSITE SANDWICHES
• Core failure/crushing
• Core/facesheet debonding

Department of Mechanical, Chemical and Materials Engineering 40

University of Cagliari, Italy
 Fibre Failure
TENSION
Cluster of
Fibres have brittle fracture.
fibre breaks
Failure occurs by unstable growth of a
cluster of adjacent broken fibres.

COMPRESSION
Fibre failure initiates by instability (buckling)
followed by kinking

Department of Mechanical, Chemical and Materials Engineering 41

University of Cagliari, Italy
 Matrix failure
Matrix failure is controlled by tensile or compressive stresses perpendicular to
the fibre direction and by shear stresses.

TENSION loading involves failure of the matrix

perpendicular to the tensile load direction

COMPRESSION loading involves failure of the

matrix along inclined planes

Out-of-plane SHEAR loading involves failure

of the matrix along a 45° plane

Department of Mechanical, Chemical and Materials Engineering 42

University of Cagliari, Italy
 Examples of criteria for ply failure
Maximum stress criteria σy
Matrix failure in tension +
=1
σ y ult
σx
Fibre failure in tension +
=1 σy
σ x ult Matrix failure in compression
−
=1
σ y ult
σx
Fibre failure in compression
−
=1 τ xy
σ x ult Matrix failure in shear =1
τ xy ult

More complex failure criteria include the interaction between the

various stress components.

2
 τ xy 
2 2
 σx   σx  σy   σy 
  −  +  +  =1
Tsai-Hill criterion
 σ x+ ult   σ x+ ult   σ x+ ult   σ y+ ult   τ xy ult 
        

Department of Mechanical, Chemical and Materials Engineering 43

University of Cagliari, Italy
 Delamination (Interlaminar failure)
Delamination is the separation between adjacent plies due to normal (through-
thickness z-direction) or shear stresses at the interface.
It is one of the most common failure processes in laminates, because of the low through-
thickness strength of laminates.

DELAMINATION
DELAMINATION

Delaminations are typically induced in composite laminates during service

- by out-of-plane loads (impacts) or
- by in-plane loads in the presence of strain concentrations such as at
discontinuities (ply drops, wrinkles, material or geometric transitions) or existing
defects.

Department of Mechanical, Chemical and Materials Engineering 44

University of Cagliari, Italy
 Delamination (Interlaminar failure)
Delamination may propagate -under static or cyclic loads- with three
different propagation modes

Growth of delamination may be modelled with a fracture mechanics

approach, assuming that the crack propagates when the energy available
(strain energy release rate) reaches the fracture energy of the material.
Specific FE approaches include the VCC technique or the use of interface
elements implementing a cohesive law of fracture.

Department of Mechanical, Chemical and Materials Engineering 45

University of Cagliari, Italy
 Buckling
Buckling is a mode of collapse occurring under compression, which is
characterized by the appearance, at a critical applied load, of out-of-plane
bending deflections (corresponding to new equilibrium configurations).

Buckling can significantly reduce the compressive strength and stiffness of

composite structures and can lead to the development of other failure modes
(i.e. fibre failure)

Department of Mechanical, Chemical and Materials Engineering 46

University of Cagliari, Italy
 Buckling of delaminated composites
Different buckling modes can be induced in delaminated composite
structures depending on the thickness of the laminate and on the size and
depth location of the delamination.

Local buckling

Global buckling Mixed buckling

• Buckling load is reduced in the presence of delamination.

• Bending of plies due to buckling results in higher stresses and may promote
the growth of delamination

Department of Mechanical, Chemical and Materials Engineering 47

University of Cagliari, Italy
 Failure modes in sandwich structures
Sandwich structures show typical damage modes (in addition to
damage in the composite skins).
Failure shear Face buckling
in the core

Core crushing or
laminate fracture Global or
due to local loads shear buckling Face/core debonding
+ buckling
Department of Mechanical, Chemical and Materials Engineering 48
University of Cagliari, Italy
 Aerospace quality
Typical manufacturing defects

•Voids and dry zones

•Delaminations
•Bonding defects
•Foreign inclusions
•Fibre waviness
PREPREG
•Wrinkles

0° direction INFUSION

Dry regions and fibre waviness INFUSION

Department of Mechanical, Chemical and Materials Engineering 49

University of Cagliari, Italy
 Typical failure modes in blades
- Main spar -

[Sørensen et al., 2005]

The most critical failure modes involve interface failure

Department of Mechanical, Chemical and Materials Engineering 50

University of Cagliari, Italy
 Typical failure modes in blades
- Wing shell -

[Sørensen et al., 2005]

The most critical failure modes involve interface failure

Department of Mechanical, Chemical and Materials Engineering 51

University of Cagliari, Italy
 Main structural design constraints
for wind turbine blades

- Ultimate strength (controlled by material strength)

- Fatigue strength (controlled by material strength)
- Global stiffness (controlled by material stiffness)
- Natural frequencies (controlled by material stiffness and mass)
- Global and local buckling (controlled by material strength and stiffness)

ALL AFFECTED BY DAMAGE/DEFECTS

Department of Mechanical, Chemical and Materials Engineering 52

University of Cagliari, Italy
Basic steps of strength analysis on a blade
Design loads
• Fatigue loads (wind variation, wind shear, blade rotation)
• Ultimate loads (extreme wind, turbulence, emergency braking)
By aeroelastic calculations, loads are quantified in terms of
• Flapwise bending moments
• Edgewise bending moments
• Flapwise shear forces
• Edgewise shear forces
• Axial force (secondary importance)
• Torsional moment (secondary importance)

Stresses and strains are calculated in blade sections and

appropriate failure criterions are used to estimate
strength or life.
Department of Mechanical, Chemical and Materials Engineering 53
University of Cagliari, Italy
 Fatigue of composite blades
Wind turbine blades must withstand the time-varying loads
through the entire life of the blade.
Major cyclic loads are induced by gravity (edgewise
bending) and by varying winds (flapwise bending)
The expected fatigue life of a wind turbine is of the order of
20-25 years (more than 108 cycles)

Department of Mechanical, Chemical and Materials Engineering 54

University of Cagliari, Italy
 Fatigue testing of blade materials
• Fatigue strength data of representative laminates are typically
generated by cyclic tests carried out at constant stress amplitude.
• The cyclic loading is described by the maximum applied stress and by
the R ratio (ratio of minimum applied stress to maximum applied stress).
• Specimens are tested at different maximum stresses and with selected
values of R ratios.
• The fatigue behaviour is described by S-N curves at specific R-ratios.

Department of Mechanical, Chemical and Materials Engineering 55

University of Cagliari, Italy
 Constant Life Diagrams (CLD)
• Constant Life Diagrams (CLD or Goodman diagrams) describe the
influence of R-ratio on the fatigue life. CLD are constructed from S-N curve
obtained at different R ratios.
- Continuous lines connect points of equal fatigue life (constant life lines)
- Radial lines indicate stress states with a given R ratio

Department of Mechanical, Chemical and Materials Engineering 56

University of Cagliari, Italy
 Constant Life Diagrams (CLD)
• Constant Life Diagrams are constructed with fatigue data obtained at
different R ratios.
• R ratios typically characterized are
- R = 0.1 (tension-tension fatigue)
- R = -1 (symmetric tension-compression fatigue)
- R = 10 (compression-compression fatigue)

CLDs built with insufficient fatigue data may result in

highly non-conservative fatigue life predictions
R = -1

R = -1
R=0.5
R=0.1
R = -1 R=-0.35
R=-1
R = -2.5
R=10

Department of Mechanical, Chemical and Materials Engineering 57

University of Cagliari, Italy
 Fatigue analysis with CLD
• Representative load histories are transformed into stress (or
strain) histories at a given blade location;
• A counting procedure (i.e. rainflow) is applied to transform variable
amplitude signals to a collection of blocks of constant amplitude
cycles. Each block is identified by a mean stress value and a
stress range.

Number of cycles Mean Stress (Strain) Stress (Strain) Amplitude

N1 S1 A1
N2 S2 A2
… … …
… … …
Nn Sn An

Department of Mechanical, Chemical and Materials Engineering 58

University of Cagliari, Italy
 Fatigue analysis with CLD
• Fatigue life is estimated by a damage summation rule (i.e. Palmgren-Miner)

(
j
) <1
N actual S applied
∑N (γ S )
j
charact applied

From CLD

• Partial safety factors for materials are applied to account for variability
and uncertainties in material strength

• type of material
• manufacturing method
• post curing
• ply drops
• size effect
• humidity
• temperature
• UV radiation
• ……
DNV-OS-J102 (2006)
Department of Mechanical, Chemical and Materials Engineering 59
University of Cagliari, Italy
 Ultimate strength (static) analysis
Static failure criterions are used to estimate strength at specific locations.
Allowable design stresses/strain for the material are reduced
by safety factors for materials.

DNV-OS-J102 (2006)

Department of Mechanical, Chemical and Materials Engineering 60

University of Cagliari, Italy
Ultimate strength and fatigue life estimation
Main limitations of conventional analyses

The presence of defects is only accounted for by limiting safety

factors (empirically determined). Large safety margins are needed if
the effect of damage on residual strength is not well understood.
The change of properties (stiffness degradation, residual strength)
associated to damage evolution is not predicted or taken into account
in the estimation of strength, life or performance.

Extensive testing is required to develop accurate CLDs. The

approach is conservative only when a sufficient set of fatigue data is
available. (i.e. at least three R=0.1, -1, 10).

Fatigue data are only valid for the tested lay-ups. A change in lay-up
requires new characterization of the material

Department of Mechanical, Chemical and Materials Engineering 61

University of Cagliari, Italy
 Improving accuracy of strength/life estimation
Composites have unique damage modes and exhibit different
levels of sensitivity to the various damage mechanisms.

To address limitations or conventional design tools, more

reliable methods for strength/life assessment of composite
structures should thus be adopted to:
• Account for the main failure and damage mechanisms
• Simulate the growth and interaction of single failure modes
• Predict the degradation of properties due to damage
• Predict the residual strength in the presence of damage

Basic concepts of a DAMAGE TOLERANCE APPROACH

Department of Mechanical, Chemical and Materials Engineering 62

University of Cagliari, Italy
 Damage tolerance approach
A Damage Tolerance approach aims at estimating the ability of
the structure to sustain design loads and to perform its function
in the presence of a specified amount of damage/defects
Specific procedures and tools (i.e. fracture mechanics, cohesive
laws) may be used to predict damage progression during service
(under both static and cyclic loads) and to simulate the effect of
damage on blade performance (strength, stiffness, etc.)

The introduction of damage tolerance concepts would allow to

• optimize the use of materials
• minimize effort (cost, time) for material testing
• reduce safety factors values
• improve the blade reliability

Department of Mechanical, Chemical and Materials Engineering 63

University of Cagliari, Italy
 Damage progression approaches
Possible approaches incorporating damage progression:

• Strength degradation models : residual strength of the composite is

estimated as a function of the number of cycles. Failure occurs when the
residual strength is lower than the maximum applied stress.
• Stiffness degradation models: stiffness degradation is estimated as a
function of the number of cycles by means of appropriate laws. Failure
occurs when the residual stiffness decreases to a predefined stiffness
level.
• Physical damage models: based on the simulation real damage and
fracture mechanisms developing in individual plies of the composite.
Only characterization at the ply level is required; models are potentially
applicable to composites with any stacking sequence.

Department of Mechanical, Chemical and Materials Engineering 64

University of Cagliari, Italy
 Possible basic steps of a physically-based
progressive damage approach
- Mechanical characterization of the basic unidirectional ply in terms of static
strength, fatigue life (damage initiation and propagation) and damage.
- Stress/strain analyses carried out cycle by cycle
- Ply failure is predicted using appropriate multiaxial failure criteria
- Properties of the laminate are estimated as a function of existing damage.
- Final failure is defined from performance requirements (stiffness) or predicted
at the last ply failure.

Main challenges in building physically-based models:

• Development of fracture mechanics criteria for propagation of the different
damage mechanisms
• Development of simple models for associating damage pattern (e.g. crack
density, delamination) to residual mechanical properties of composite.

Department of Mechanical, Chemical and Materials Engineering 65

University of Cagliari, Italy
 Example of damage tolerance analysis
Tensile behaviour of a [03/903] composite plate in the presence of a hole

Matrix cracks in 90° plies

Splitting in 0° plies

Delamination
Static damage
Fatigue damage
Department of Mechanical, Chemical and Materials Engineering 66
University of Cagliari, Italy
 Example of damage tolerance analysis

The growth of axial splitting and

delaminations with static or cyclic loads
can be simulated by a FE model adopting
a fracture mechanics based technique
(Virtual Crack Closure + Paris law).

The stress distribution at the hole is

affected by the entity of damage

The residual strength of the

damaged plate is predicted as the
load inducing failure of 0°plies

Department of Mechanical, Chemical and Materials Engineering 67

University of Cagliari, Italy
 Example of damage tolerance analysis

Prediction of change in residual

strength with fatigue cycling

Damage reduces stress

concentration at the hole

N = 104 cycles N = 105 cycles

Department of Mechanical, Chemical and Materials Engineering 68

University of Cagliari, Italy
 COMPOSITES FOR
DELAMINATION AND
FRACTURE RESISTANCE

Department of Mechanical, Chemical and Materials Engineering 69

University of Cagliari, Italy
 Delamination
Delamination and face/core debonding are two of the
most critical failure modes in composite blades.
The presence of delamination and debonding strongly
affects the compression resistance and the stiffness
of the blade.

In laminated composites, fibres provide no reinforcement

along the thickness direction and crack growth resistance is
largely controlled by matrix properties.
Improving the delamination resistance (interlaminar
fracture toughness) of composites is thus a very effective
way to improve structural reliability and efficiency of
composite blades

Department of Mechanical, Chemical and Materials Engineering 70

University of Cagliari, Italy
 Delamination
Effective routes for improvement of interlaminar toughness
include two approaches

STRUCTURAL APPROACH
• 3D reinforcement (Stitching - Z-pinning)
• Interlayers (Interleaving - Veils)

MATERIAL APPROACH
• Thermoplastic or toughened resins
(disadvantages: expensive; process require high temperature;
toughening additives increase viscosity and make resins difficult to
infuse at low temperatures)
Toughening by Nano-modification of the matrix is an option
being explored
Department of Mechanical, Chemical and Materials Engineering 71
University of Cagliari, Italy
 Stitching / Z-pinning
Stitching consists in sewing a structural thread (kevlar, polyethylene)
through the plies to produce a preform with a 3D fibre structure.
Stitching can be performed on both dry fabric and uncured prepreg tape.

Z-pinning consists in inserting thin and stiff pins (steel, carbon fibres)
through the plies, typically by an ultrasonic process.
Z-pinning is typically performed on uncured prepreg tapes but can be adapted for use
on dry fabrics.

Department of Mechanical, Chemical and Materials Engineering 72

University of Cagliari, Italy
 Stitching / Z-pinning
Stitching /Z pinning are very efficient techniques to improve
interlaminar properties of composites
Stitching delays the growth of delaminations by bridging the delaminated interface and
reducing the interlaminar stresses at the crack tip.

Main disadvantages
Localized fibre fracture
Presence of a weak (resin rich) area around the stitch/pin
In-plane and out-of-plane fibre waviness produced by the process
Environmental effects (increased moisture absorption)

These combined effects may induce a reduction of in-plane static and

fatigue strength of stitched/Z-pinned composites.
Department of Mechanical, Chemical and Materials Engineering 73
University of Cagliari, Italy
Improvement of interlaminar fracture properties by Z-pinning

Mode I fracture test (DCB) Mode II fracture test (ENF)

more then
10 times increase in
fracture toughness

[Partridge et al., 2005]

Department of Mechanical, Chemical and Materials Engineering 74

University of Cagliari, Italy
 Improvement of impact resistance by stitching

5 J impact on a [03/903]s carbon/epoxy laminate

UNSTITCHED

STITCHED
(5 mm x 5 mm stitching pattern)

more than 30% reduction

in damage area

Department of Mechanical, Chemical and Materials Engineering 75

University of Cagliari, Italy
 Improvement of impact resistance by stitching

Bridging
action of
stitches

7 J impact on a stitched [03/903]s carbon/epoxy laminate

Department of Mechanical, Chemical and Materials Engineering 76
University of Cagliari, Italy
 Improvement of compression after impact
(CAI) strength by stitching

Impact damage up to 30% increase in

compression after impact strength

CAI strength of stitched [0/90]3s carbon/epoxy laminates

Department of Mechanical, Chemical and Materials Engineering 77
University of Cagliari, Italy
Buckling of delaminated plies under compression (CAI)

Delamination growth

UNSTITCHED STITCHED

Buckling followed Buckling and delamination growth are

by delamination growth restrained by the bridging action of stitches

Department of Mechanical, Chemical and Materials Engineering 78

University of Cagliari, Italy
 Influence of stitching on static/fatigue strength
Static/Fatigue strength of composite materials is usually reduced by
stitching in fibre dominated laminates (failure controlled by fibre damage)

S-N curves of [±45/0/90]s carbon/epoxy laminates under cyclic tension

Department of Mechanical, Chemical and Materials Engineering 79

University of Cagliari, Italy
 Influence of stitching on static/fatigue strength
Fatigue strength of composite materials may be improved by stitching
in matrix-dominated laminates (failure is controlled by delamination)

In unstitched [+45/-45]s laminates, fatigue damage sequence involves:

1. Matrix cracking 70%

fatigue life

2. Localization of extensive matrix

cracking in specific regions 85%
fatigue life

3. Growth of delamination from

highly cracked regions At failure

UNSTITCHED

Department of Mechanical, Chemical and Materials Engineering 80

University of Cagliari, Italy
 Influence of stitching on static/fatigue strength
Stitching in [+45/-45]s laminates (failure is controlled by delamination)
• Improves fatigue life (up to 3 times increase)

• Improves resistance to stiffness degradation

Delamination initiation

Department of Mechanical, Chemical and Materials Engineering 81

University of Cagliari, Italy
 Influence of stitching on static/fatigue strength
Fatigue strength of composite materials is improved by stitching
In [±30/90]s laminates ((failure is controlled by delamination)

S-N curves of [±30/90]s carbon/epoxy laminates

Delamination
arrested by
stitches

more than 20 times increase in fatigue life

UNSTITCHED STITCHED

Department of Mechanical, Chemical and Materials Engineering 82

University of Cagliari, Italy
 Failure often occur at joints and connections

LEADING
EDGE

TRAILING EDGE

WEBS / SPAR CAP

Department of Mechanical, Chemical and Materials Engineering 83

University of Cagliari, Italy
Improvement of T-joints strength by Z-pinning

Up to 2 times
strength increase

[Koh et al., 2012]

Department of Mechanical, Chemical and Materials Engineering 84
University of Cagliari, Italy
Improvement of lap-joints strength by stitching
[03/903]s carbon epoxy laminate

Up to 5 times increase
in fatigue life

Department of Mechanical, Chemical and Materials Engineering 85

University of Cagliari, Italy
 Stitching in sandwich composites
Stitching can be performed in composite sandwich structures to
reduce skin/core debonding (and thus increasing buckling load)

up to 2 times
increase in
Buckling load
buckling load
[Sharma, 2004]

Department of Mechanical, Chemical and Materials Engineering 86

University of Cagliari, Italy
 Improvement of delamination resistance
by interleaving or veils
Delamination resistance may be improved in laminates by the insertion of
• tough resin films between the layers of the composite (problems for infusion)
• veils between layers (veil is a low weight layer made of randomly oriented
fibres

VEIL

Fibres from the veil migrate to the upper and lower fabric plies and provide
a bridging effect between delaminated layers

Department of Mechanical, Chemical and Materials Engineering 87

University of Cagliari, Italy
 Delamination resistance may be improved
by using tough resins
(high fracture resistance and large deformation at failure)

Examples of TOUGH RESINS

•Rubber modified thermosets (difficult to infuse)
•Thermoplastic resins (difficult to infuse and process)
•Nano-reinforced resins

Main general issue

Higher resin’s fracture toughness does not necessarily translate to
higher composite’s fracture toughness
(large plastic deformation of resin is constrained by the presence of fibres,
poor fibre/matrix adhesion, residual stresses)

Department of Mechanical, Chemical and Materials Engineering 88

University of Cagliari, Italy
Impact resistance of an advanced thermoplastic prepreg
APC2 – Carbon/PEEK (Poly-ether-ether-ketone)

Impact damage in [±45/0/90]2s carbon fibre laminates

8 J impact

Carbon/PEEK

PEEK is a thermoplastic resin with a

processing temperature of about 340°C,
Carbon/epoxy
requiring autoclave or press moulding,
and a quick cooling rate.

Department of Mechanical, Chemical and Materials Engineering 89

University of Cagliari, Italy
Alternative thermoplastics resins and processing routes
Intertwined or commingled fabrics: reinforcing fibres are mixed with fibres
spun from thermoplastic polymers (PET, PP). Impregnation is achieved by
application of sufficient heat and pressure.

Liquid moulding: monomers are heated, activated with a catalyst and

injected into the fabrics as a low viscosity resin; the polymerisation takes
place in situ and the basic process steps are similar to those of thermosets
(vacuum infusion).
Main issues with thermoplastic resins

High processing temperatures More expensive tooling

(160°C to 250°C) Higher energy inputs
Large thermal expansion
Faster heating and cooling rates Special tooling
are required (Brittle matrix with low cooling rates)
High sensitivity to moisture
Reduced mechanical properties
before processing

Department of Mechanical, Chemical and Materials Engineering 90

University of Cagliari, Italy
 Nanomodification of epoxy resin
Dispersion of nanoclay particles (1 nm thick platelets) in
epoxy resins improves both static and fatigue fracture
properties.

Fracture properties of nanomodified epoxy resin

Fracture toughness
Paris curve
[Quaresimin et al., 2012]
Department of Mechanical, Chemical and Materials Engineering 91
University of Cagliari, Italy
Use of nanomodified epoxy resin in laminates
Glass/epoxy laminates infused with nano-clay modified
epoxy resins have worse static and fatigue fracture
properties than conventional laminates.

Fracture properties of nanomodified laminates

Fracture energy
Paris curve
[Quaresimin et al., 2012]

Department of Mechanical, Chemical and Materials Engineering 92

University of Cagliari, Italy
 USE OF CARBON FIBRES
IN WIND TURBINE BLADES

Department of Mechanical, Chemical and Materials Engineering 93

University of Cagliari, Italy
 Potentials for use
of carbon fibres in blades
Carbon fibres have been considered and used in recent years as a way to
reduce weight and increase stiffness in large blades.

Main issues related to the use of carbon fibres

Cost (about 5-10 times more expensive than glass fibres)
Compressive performance of carbon fibres is sensitive to alignment
(prepreg better than wet process).
Fatigue performance of carbon fibres is strongly degraded by stress
concentrations (ply drops; carbon-glass interfaces, voids, wrinkles,
delaminations)

Infusion of thick carbon fabrics is difficult because of the lower

permeability than fibreglass fabrics (this suggests use of prepregs)

Department of Mechanical, Chemical and Materials Engineering 94

University of Cagliari, Italy
 Potentials for use
of carbon fibres in blades
Parametric analyses show potential for significant structural
improvements both for complete and hybrid/selective use of carbon fibres

§ 30% to 40% reduction in mass

§ 20% reduction in tip deflection
for complete replacement of glass with carbon in main spar

§ 50% reduction in root moment

§ 10% reduction in tip deflection
For selective replacement of glass with carbon in the outer span of
the blade
[Griffin et al., 2003]

Department of Mechanical, Chemical and Materials Engineering 95

University of Cagliari, Italy
 Carbon/Glass transition and ply-drops
An issue for the use of carbon fibres is the design of the
transition between carbon and glass fibres and of ply-drops.

Critical aspects include

Large difference between stiffness and strain-to-failure of carbon and
glass layers (higher stress in carbon; load is transferred by matrix shear
and high matrix strength is required)

•Stress concentrations at ply drops/transitions may induce

- strong decrease of fatigue strength in carbon structures (high shear
stress in matrix)
- delaminations at relatively low strains under fatigue

Compression strength of carbon plies is strongly reduced by fibre

waviness

Department of Mechanical, Chemical and Materials Engineering 96

University of Cagliari, Italy
 Design recommendations for transitions/ply drops

STAGGERED PLY DROPS [Griffin et al., 2003]

§ Thin plies are better than thick plies in the presence of ply drops for
delamination resistance.
§ Thicker laminates are better than thin laminates, for the same number of ply
drops, in resisting delamination.
§ No more than one ply drop at the same location (staggered pattern).
§ Internal ply drops are more resistant to delamination than external ply drops

Z-spiked region: fibre tows are driven into adjacent layer

Z-spiking
Department of Mechanical, Chemical and Materials Engineering 97
University of Cagliari, Italy
Carbon fibres for bend-twist coupled blades

Carbon layers are especially effective in achieving bend-twist

coupling because of their high orthotropic ratio

Bend-twist coupling is a form of aeroelastic tailoring in which a

flapwise bending load induces a twist of the blade section.

A change in wind velocity (which induces a change in bending

moment and thus in twist) results in a change in the angle of attack.
The change in the angle of attack provides opportunities for load
mitigation.

This passive approach for load mitigation is especially attractive

because of its simplicity and economy (lighter blade, increased
reliability, less maintenance, etc.).

Department of Mechanical, Chemical and Materials Engineering 98

University of Cagliari, Italy
 Bending-extension coupling of off-axis composites
The bend-twist coupling may be achieved by exploiting the orthotropic
elastic properties of laminated composite materials (off-axis loading)

SHEAR-EXTENSION COUPLING

 N1   A11 A12 A16  ε10   B11 B12 B16  κ1 

     0    
 N 2  =  A12 A22 A26  ε 2  +  B12 B22 B26  κ 2 
N   A A66  ε 60   B16 B66  κ 6 
 6   16 A26 B26

 M 1   B11 B12 B16  ε10   D11 D12 D16  κ1 

     0   
M 2  =  B12 B22 B26  ε 2  +  D12 D22 D26  κ 2 
M   B B66  ε 60   D16 D66  κ 6 
 6   16 B26 D26
BEND-EXTENSION COUPLING BEND-TWIST COUPLING
Department of Mechanical, Chemical and Materials Engineering 99
University of Cagliari, Italy
 Blade with on-axis laminates
The symmetry axis of the composite is aligned with the blade axis.

Extension of the laminates does not

induce shear forces
→ the blade does not twist with bending

[de Goeij, 1999]

Department of Mechanical, Chemical and Materials Engineering 100

University of Cagliari, Italy
 Bend-twist blade with off-axis laminates
The symmetry axis of the composite is not aligned with the blade axis

Extension of the laminates induces shear

forces and generates a torsion
→ the blade twists with bending

[de Goeij, 1999]

Department of Mechanical, Chemical and Materials Engineering 101

University of Cagliari, Italy
 Bend-twist configurations

Bend-twist coupling in
shell, spar cap, or both

Fully (whole span) or

partially (outer span)
coupling.

Parametric analyses suggest partially coupled blades combining off-

axis orientation of fibres both in skin and spar cap as an efficient design
for fatigue load mitigation and mass reduction. [Bottasso 2012]

Department of Mechanical, Chemical and Materials Engineering 102

University of Cagliari, Italy
 Bend-twist configurations

Main issues associated to coupled design

Large off-axis angle for fibres in the spar caps would reduce significantly the
bending stiffness of the spar (thicker spar caps needed)
The presence of off-axis fibres results in increasing importance of transverse
matrix cracks (stiffness degradation; delamination initiation; reduced fatigue
strength)
Lack of fatigue data on off-axis lay-ups

Trade-off in use of carbon fibres : introduction of bend-twist coupling does not

allow to fully exploit high strength/stiffness of fibres (off-axis orientation)

Department of Mechanical, Chemical and Materials Engineering 103

University of Cagliari, Italy
 Summary
The development of innovative composite blades involves a strict interaction between
main design drivers related to
• Materials
• Manufacturing process
• Structural design
Key issues and research trends include:
Materials
Improvement of material properties
Better understanding of material behaviour (reduction of safety factors)
Prediction of the effect of defects (damage tolerance approach)
Optimized use of materials

Manufacturing process
Reduction of defects
Reduction in cycle/production times
Reduction of labour intensive steps (automation)

Structural design
Better design of structural details (ply drops, transitions, bonding)
Better understanding of size effect
Innovative structural configurations

Department of Mechanical, Chemical and Materials Engineering 104

University of Cagliari, Italy
References
S Black. Automating wind blade manufacture, Composites World, June 2009 .
P Brøndsted, H Lilholt, A Lystrup. Composite materials for wind power turbine blades. Annu. Rev. Mater. Res. 35:505–38, 2005.
RJ Bussolari: Fibreglass composite blades for the 4MW-WTS-4 wind turbine. NASA. Lewis Research Center Large Horizontal-Axis
Wind Turbines pp. 259-266, 1983.
WC de Goeij, MJL van Tooren, A Beukers. Implementation of bending-torsion coupling in the design of a wind-turbine rotor-blade.
Applied Energy 63:191-207, 1999.
DNV - Design And Manufacture of Wind Turbine Blades - Offshore And Onshore Wind Turbines, DNV-OS-J102, 2006.
JA Grande. Wind power blades energize composites manufacturing, Plastics Technology (www.ptonline.com), October 2008.
DA Griffin, TD Ashwill. Alternative composite materials for megawatt-scale wind turbine blades: design considerations and
recommended testing, J Solar Energy Eng. 125: 515-521, 2003.
Gurit Guide to Composites, www.gurit.com, 2012.
B Hayman, J Wedel-Heinen, P Brøndsted. Materials Challenges in present and future wind energy. MRS Bulletin, 33:343-353,
2008.
P Hogg. Wind Turbine Blade Materials. SUPERGEN Wind Phase 1 Final Assembly, Loughborough, March 2010
TM Koh, S Feih, AP Mouritz. Strengthening mechanics of thin and thick composite T-joints reinforced with z-pins, Composites: Part
A 43: 1308–1318, 2012
DJ Lekou. Scaling Limits & Costs Regarding WT Blades. UpWind Deliverable 3.4.3, 2010.
IK Partridge, DDR Cartie´. Delamination resistant laminates by Z-Fiber pinning: Part I manufacture and fracture performance.
Composites: Part A 36: 55–64, 2005.
M Quaresimin, M Salviato, M Zappalorto. Fracture and interlaminar properties of clay-modified epoxies and their glass reinforced
laminates. Engineering Fracture Mechanics 81: 0-93, 2012.
SC Sharma, M Krishna, N. Narasimhamurthy. Buckling response of stitched polyurethane foam composite sandwich structures,
Journal of reinforced plastics and composites, 23(12): 1267-1277, 2004
BF Sørensen, K Branner et al. Improved design of large wind turbine blades of fibre composites (phase 2) – summary report, Risø-
R-1526(EN), Risø National Laboratory, Denmark, 2005.
R Stewart. Wind turbine blade production – new products keep pace as scale increases. Reinforced Plastics, Jan/Feb 2012: 19-25.
OT Thomsen. Sandwich materials for wind turbine blades - present and future. Journal of Sandwich Structures and Materials 11:7-
26, 2009

Department of Mechanical, Chemical and Materials Engineering

University of Cagliari, Italy