|
Improving blade design requires a profound knowledge of possible structural failure mechanisms leading
to blade collapse. Risø DTU does research on these mechanisms. The tests conducted provide the data
for thorough blade analysis and, consequently, design improvement.
Motivation
Illustration of the safety margins for current and future blade designs. For more details, visit Home. The weak point of current wind turbine blades is the structural strength. The safety factor with respect to material strength is unnecessarily high compared to the one with respect to structural strength, see Publications and Structural Blade Design and Testing. The goal is therefore to find a better balance between structural strength collapse and material failure. During full-scale tests and simulations, Risø DTU has observed new nonlinear behaviour. For more details on the observed phenomena, please visit F.M. Jensen's PhD Thesis.
Brazier effect - Nonlinear geometric deformation
The so- called 'ovalisation', a non-linear phenomenon caused by bending of a pipe section, is known as Brazier
effect. When a wind turbine blade bends in the flapwise direction, the compressed panel produces a downward
crushing load normal to the surfaces. The opposite occurs on the lower panel, see Figure below. This nonlinear
deformation or ‘flattening’ of the cross section is also known as the Brazier effect and is most pronounced for long
thin hollow beams subjected to bending.
It has been recognized that thin-walled, hollow structures under bending suffer a flattening of the cross-section.
Flapwise bending results in crushing pressure. Left: Longitudinal curvature results in an inward crushing force. Right: Crushing pressure on a wind turbine blade section. This flattening manifests itself as an ovalisation in circular sections and a ‘sucking in’ for squared tubes.
Left: Inward crushing forces on a thin walled structure subjected to bending, leading to flattening of the cross-section. Ovalisation of a circular section as well as ‘sucking in’ of a squared tube. Right: When large deformations of a box girder cross section subjected to a bending moment are taken into account, the forces in the flanges can be resolved in transversal and longitudinal components. The transversal components are responsible for the flattening effect. The flattening effect is caused by internal forces. When large deformations of e.g. a box girder cross section subjected to a bending moment are taken into account, the forces in the flanges can be resolved in transversal and longitudinal components, see figure above. The transversal components are responsible for the flattening effect. Only structures with a longitudinal bending curvature have these resulting forces. “Crushing pressure” rises quadratically with the longitudinal curvature. It is expected that the future blades will be significantly more flexible. Then the importance of this problem will increase. For more details on the phenomenon, please visit F.M. Jensen's PhD Thesis. The ovalisation causes several failure modes observed in the full-scale tests. - Web failure
Web failures have been observed as the main reason for collapse in two full-scale tests. The web towards the trailing edge collapsed just before ultimate failure.
Collapse of the shear web in a full-scale test of a 34m blade from SSP-Technology A/S. Left: A non-linear FE-prestudy indicated that the shear web connection between the two half parts would be a weak point. The Figure presents the 10m section. Right: Photo of the shear web towards trailing edge after the blade has failed also indicates that the shear web has caused ultimate failure of the blade.  Sketch of failure mode observed in two full-scale tests. Shear webs collapse at the first full-scale test which had no additional reinforcement where the two half parts were connected. Collapse of the shear web in the first full-scale test. Measured web deflections at 10.3m. It is clearly visible here that the web towards the trailing edge collapsed just before ultimate failure, since a dramatic change in web deformation is observed at 99% of ultimate load.  For more details concerning the figures, please visit F.M. Jensen's PhD Thesis. Crushing test of a box girder section
The experiments were performed in displacement control using a 250 kN Instron material testing machine. The upwind side (bottom flange) is supported by two cylinders symmetrical about the longitudinal midplane. The downwind side (top flange) is loaded by a cylinder in this midplane (see Figure below). Both the support and loading cylinders are slightly angled in order to account for the longitudinal tapering of the box girder. Three specimens cut from the load carrying box girder of a 25m wind turbine blade were tested and deformations and strains were measured. The specimens were tested until failure and they all failed ultimately in the sandwich webs towards the trailing edge as shown in the Figure below (right). As the specimen is loaded, bending of the top flange is building up below the loading cylinder. The inner side is experiencing tensile strain and the outer side compressive strain. The flange is constructed with mainly UD-laminates combined with some Biax-laminates. For the UD- laminates the fibers are in the longitudinal direction of the blade and as transverse tensile strain are building up the inner UD-laminates start to fail perpendicular to fibers. A crack is then propagating from loading point towards the webs (see Figure below). No failure is noticed in the Biax-laminates. Close to final failure of the trailing edge web, localization at the webs is observed. The core is failing for all specimens and the inner skin is failing for 2 out of 3 specimens. Left: Test setup. The top is fixed and the bottom with the two supports is moving upward during loading.
Right: Shear web failure of box girder section. Picture showing specimen deformation after loading to failure.
Figure from ref. [2].
When the specimens were loaded under a constant displacement ramp there was a linear region which consisted of ‘flattening’ of the flanges, and bowing of the webs. Typically this would lead to a number of non-linear effects, one of which being cracks forming in the UD layers of the flange, which would build up and induce separation of the layers, commonly known as delamination. In the webs, high levels of shear would appear between foam and skin, and eventually lead to either buckling or the skin de-bonding from the sandwich core, see Figure below. This load in this test mimics the Brazier crushing pressure. In the unreinforced 2D section test the single skin and the adhesive connection of the two half parts buckled, see Figure below (left). For the test of the reinforced section, the failure also happened in the shear web, but this time in the sandwich area, see Figure below (right).
2D-section experiments of original box girder (left) and reinforced box girder (right). Figure from ref. [1].
Full scale test of a load carrying box girder
The full-scale test involved only the reinforced load carrying box girder. The Figure below shows a frozen frame picture from a video recording of the first failure mode of the box girder full-scale test at 10.1m from the root. It shows initial face debonding of the outer skin on the shear web’s sandwich section, leading to ultimate collapse of the box girder. The Figure to the right shows a sketch illustrating the failure and the strain gauges attached in that area.
Skin debonding of the sandwich web towards leading edge observed in a full-scale box girder test. Left: Frozen frame picture showed skin debonding of the sandwich web towards leading edge. Right: Sketch which illustrates the failure and the strain gauges attached in that area. Figure from ref. [1].
- Interlaminar failure of load carrying cap specimen
Interlaminar shear failure is a mechanism that occurs between the layers in the load-carrying cap laminate, see Figure below. The failure is caused by interlaminar shear stresses. The interlaminar shear failure can be developed by the crushing pressure causing biaxial stress distributions, interlaminar and peeling stresses due to the curved structure being flattened out, see Figure below. Brazier forces, causes the out-of-plane deformation of the cap. The cap deflections are increased by the lay-up, since the fibres are mainly placed in the longitudinal direction of the blade. The lack of fibres in the transverse direction causes the cap to be relatively flexible in the lateral direction, see below.
Interlaminar shear failure of the load-carrying cap. Left top: Sketch of cap deformation and failure between layers.
Right: 3 and 4-point bending tests of two cap specimens. The bending tests were performed at Imperial College – Department of Mechanical Engineering. The overlaid contour plots show the strains along axis of the test specimen (referring to the transverse strains in the box girder) caused by the bending loads
Left bottom: The bending tests lead to initial interlaminar failure after 8mm and 4mm deflections in the 3 and 4-point bending tests.
For more details concerning the figure, please visit F.M. Jensen's PhD Thesis.
Local deflections
Local deflections were measured on the skin, that is the deflection on the middle of the main spar relative to the webs of the main spar. In flapwise loading the webs are following the blade deflection so placing the measurement system fixed over the two stiff webs of the main spar and then measuring the deflection in the middle of the main spar gives the local deflection of the flange of the main spar as shown on the figure below.
This local deflection was shown to be a good quantity for describing the state of the blade, i.e. how close the blade is to an interlaminar failure.
A measurement setup including five sensors for measuring local deflections is shown in the figure below for the test of section three.
Left: System for measuring local deflections. Right: Local deflection at at4.5m from the root with load applied at 11.3m. Figure from ref. [4].
In another full-scale test of a 34m SSP-blade the cap deflections were in the range of 1-6mm, so lateral bending stiffness may be too small to avoid interlaminar failure.
Cap deflections of the centreline in different sections in the first full-scale test. Figure from ref. [5].
The results above lead to a conclusion that the load carrying laminate may be critical to interlaminar failure in the load carrying laminate caused by the non-linear crushing pressure
- Transverse shear distortion
It has been noticed that transverse shear distortion gains importance when nonlinear behaviour is taken into account.
Transverse shear distortion of a wind turbine blade cross-section is an important failure mechanism.
 
Distorted profile. Left: Sketch of the phenomenon. Right: Finite Element Model of the blade in combined flap- and edgewise loading, showing both, deformed and undeformed shape.
A thin walled structure without any internal reinforcement will try to distort the profile in the transverse direction. This is even more prominent when the blade is non-symmetric, both in geometry and in lay-up, since the blade will try to twist. Furthermore, the lay-up is highly orthotropic with the majority of fibres in the longitudinal directions, so the circumferential stiffness is small.
The combined flap- and edgewise load direction, which is the most realistic load case, causes the blade to distort. In the future, larger blades are expected to be more critical for this failure mode.
 |
 | Combined gravity and aerodynamic forces result in a load component different from the traditional flap- and edgewise loads applied in full-scale test causing the blade to distort.
Furthermore, if future wind turbine blades are lighter due to optimization, the longitudinal curvature will increase, which will raise the crushing pressure and this could be critical for a distorted blade section, see Figure below.
 |
 | Left: Aerodynamic loads on wind turbine blade section (pressure and drag) including gravity loads. Right: Crushing pressure from the longitudinal bending (Brazier loads) at a distorted profile is very critical.
Also other failure modes have been observed during full-scale tests:
- Panel deformations cause failure
Minimizing deformation of the aerofoil is needed for several reasons e.g. trailing edge fatigue problems and aerodynamic efficiency. The aerodynamic efficiency will not be discussed here, only the fatigue problem will be addressed in F.M. Jensen's PhD Thesis. Today forces are carried by the double curved airfoils, which leads to out of plane deformation, see Figure below. Out of plane deformations cause peeling stresses in the trailing edge, and failures in the trailing edge often occurs. The load-carrying box girder is attached to the outer skin (aerofoil). The connection between the outer skin and the box girder is an adhesive joint sensitive to peeling stresses.  Sketch of the trailing edge shells with ‘out of plane’ deformations. The close ups show fatigue failure in the trailing edge as well as debonding of outer skin from the box girder.
- Buckling
Buckling is a structural instability phenomenon for structures, which are loaded in compression. The behaviour of shell-like structures under buckling is characterized by limit points rather than a bifurcation point. With linear buckling analysis, the bifurcation point is obtained as the solution to an eigenvalue problem. Accordingly, linear buckling analysis is a guideline for the design load, to which a suitable reduction factor is called for. The size of the reduction factor may be obtained by further analysis and/or substantial testing. In the analysis of shell imperfection sensitivity it is necessary to investigate the geometrically non-linear structural response.
Measurement of the local cap deformations (out-of-plane) in the first box girder test. The results from DIC measurement are superimposed to a picture of the cap surface.
Other buckling failure modes which have been considered are:
- skin debonding from box girder
- local buckling of the panels in the trailing edge
- global buckling of the trailing edge
Left: Picture from full-scale test presenting skin debonding caused by buckling. Figure from ref. [4]. Middle: Buckling of the trailing edge shell. Sketch where the trailing edge panel has out of plane deformations, which lead to redistribution of the forces. Right: Global buckling of trailing edge.
References:
[1] “The Brazier effect in Wind-Turbine Blades and its Influence on Design”. F.M. Jensen, P.M. Weaver, L.S. Cecchini, H. Stang, R.F. Nielsen. (Submitted November 2008)
[2] ”Modelling Failure in Cross-Section of Wind Turbine Blade” K. Branner. Risø National Laboratory, NAFEMS-Seminar. Denmark (June 2006)
[3] “Non-destructive analysis of wind turbine blade structural integrity” A.Puri, F.M. Jensen, M.McGugan, 2009 ASME Pressure Vessels and Piping Division Conference, July 26-30, 2009, Prague, Czech Republic
[4] “Full scale testing of wind turbine blade to failure – Flapwise loading”. E. R. Jørgensen, K. K. Borum, M. McGugan, C. L. Thomsen, F. M. Jensen, C. P. Debel, B. F. Sørensen Risø-R-1392(EN). (2004).
[5] “Structural testing and numerical simulation of a 34m. Composite wind turbine blade”. F. M. Jensen, B.G. Falzon, J. Ankersen, H. Stang, Composite Structures 76. (2006)
Go to the top |