Experimental testing method to characterise the drapability of UD non-crimp fabrics used in wind turbine blades

This scientific article presents a novel approach for characterising the drapability of fabrics used in wind turbine blade production. This study defines drapability as an intrinsic property of fabric to shear. Specifically, it refers to the potential of the rovings to slide with respect to each other. The evolution of wrinkles has been quantified by the ratio of height-to-width corresponding to a shear angle. The growing industrial interest in binder fabrics, for their preforming ability and improved handling leading to faster blade production, has motivated this study. In this research, two types of non-crimp fabrics, with and without binder, were analysed to study the evolution of wrinkles concerning applied shear angles. A state-of-the-art 3D blue light scanning technique is employed to accurately measure the aspect ratio (height/width) of wrinkles at various shear angles, including 0°, 4°, 6°, 8°, 12°, and 16°. A wrinkle having an aspect ratio of 1/10 was determined to correlate with an applied shear angle of 9° for non-binder fabrics, and 3° for binder-based fabrics. The findings clearly demonstrate the influence of binders on fabric drapability, reducing it by a factor of three. These results provide valuable insights into the influence of different parameters on wrinkle formation, aiding in controlling these factors to avoid manufacturing defects in wind turbine blades.


Introduction
Drapability is a fundamental characteristic that describes a fabric's ability to conform to the surface of moulds [1].The existing standards for drapability, such as BS 5058:73 [2] and AFNOR G07-109 [3], adequately apply to apparel, upholstery, and non-woven fabrics, there is a need for better adaptation to address drapability capabilities in composite preforms [1].For engineering textiles like woven fabrics, unidirectional or biaxial non-crimp fabrics, the draping behaviour is mostly affected by in-plane shear and bending characteristics [4].To characterise the shear deformation behaviour of fabrics, picture frame tests [5,6] and uniaxial bias extension tests [7,8] are commonly used.Several researchers have worked to characterise and analyse of wrinkling behaviour of fabrics with hemispherical [9], square box [10] and double-dome [11] punching preforming methods.
Wind turbine blades possess a complex geometry optimised for aerodynamic efficiency [12] and structural integrity [13].Among the various manufacturing techniques employed, vacuum-assisted resin transfer moulding is widely used to produce longer blades exceeding 55 meters in length.This technique involves placing fabrics in a mould, applying vacuum pressure, and injecting resin into the cavity.The resin fills the space between the fibres, and the resulting component is cured through heat [14].Currently, the most economically feasible approach to wind blade manufacturing involves 1293 (2023) 012020 IOP Publishing doi:10.1088/1757-899X/1293/1/012020 2 manual fabric manipulation by workers to ensure a layup without wrinkles in the mould [15].Preventing out-of-plane deformation during the fabric layup process is crucial.Such deformation occurs when certain fibres are forced to span shorter distances than their parallel counterparts in the fabric, leading to compression and buckling, ultimately forming wrinkles [16].This is especially important in double-curved areas, such as the root section.
A wide range of non-destructive testing (NDT) methods exists for detecting wrinkles in fibrereinforced composites, including ultrasonic techniques, linear phase FIR-filtered ultrasonic array data, laser and X-ray tomography [17], [18], [19].Griffin and Malkin [20] from the DNV have identified fibre wrinkles as a significant manufacturing defect leading to catastrophic failures of wind turbine blades, compromising their reliability and cost-effectiveness.Several studies have reported substantial reductions in strength and stiffness in laminates containing wrinkles [21], [22].Nartey et al. [23] investigated the influence of fibre wrinkles on composite laminate's mechanical properties and failure mechanisms, revealing a significant drop of up to 21% and 37% in tensile and compressive strength, respectively, for the specimen with wrinkle angle of 7.8°.
The use of binder fabrics has gained industrial interest due to their superior handling characteristics compared to binder-free fabrics.Binder fabrics offer the potential for faster wind turbine blade manufacturing by preparing fabric stacks outside the mould and subsequently placing them inside.Automation of the blade manufacturing process has also gained considerable attention, with attempts made to equip robotic arms with standard layup tools to replicate the manual movements of workers during fabric seating in the mould.These robotic arms employ sensory input to apply appropriate pressure to manipulate the fabric effectively.Controlling wrinkle formation during automated fabric installation is critical for enhancing manufacturing quality.
Despite extensive research on the detection of fibre wrinkles and their impact on the mechanical performance of fibre-reinforced composites, there remains a notable research gap concerning the assessment of fabric wrinkle formation during composite manufacturing.This paper presents a method for characterising fabric drapability, defining it as an intrinsic property related to the shearing potential between rovings without the formation of wrinkles.The study focuses on analysing two types of noncrimp fabrics, one with a binder and one without, investigating the evolution of wrinkles in response to applied shear angles.A 3D blue light scanning technique is employed to measure the out-of-plane movement of the fabrics and quantify the aspect ratio of wrinkles for various shear angles (0°, 4°, 6°, 8°, 12°, and 16°).

Materials
The two investigated non-crimp fabrics, with and without binder, have a total glass fibre areal weight of 1382 g/m 2 .Each fabric consists mainly (96 wt %) of H-glass fibres with a fibre diameter of 17 to 24 µm oriented at 0°.In addition, there are 4 wt % E-glass backing fibres, with a fibre diameter of 9 µm, oriented at ±80°.The fibre angles are given relative to the warp direction of the fabric.The stitch to keep fibres in place is a combined tricot-chain polyester thread stitch with an areal weight of 15 g/m 2 .The preform fabrics have an additional 15 g/m 2 of a soluble polyester binder dispersed on the roving side of the fabric.Three fabric specimens of each fabric type with a width of 400 mm have been tested.

Methods 2.2.1. Experimental setup.
The experimental setup for the drapability test is depicted in Figure 1.The setup consists of an Instron universal testing machine (UTM) fitted with a 1 kN load cell, depicted on the left-hand side of Figure 1.A parallelogram fixture, similar to the previously referred method [24] forms the subsequent component of the setup.While one side of the parallelogram fixture is affixed to the frame, the other side can move freely in the in-plane direction.A metallic wire links the top grip of the UTM to the movable side of the parallelogram fixture through a pulley mechanism.The test fabric 1293 (2023) 012020 IOP Publishing doi:10.1088/1757-899X/1293/1/0120203 is then placed on the fixture and secured using bolts.A constant torque of 7 Nm is applied to the bolts utilising a torque wrench.
The upward movement of the top UTM grip initiates a shear action, causing the fabric to form wrinkles.To monitor and analyse the fabric's behaviour during shearing, the top surface of the fabric is scanned at different shear angles (0°, 4°, 6°, 8°, 12°, and 16°) using a blue light 3D scanner (EviXscan 3D).
Before the scans, the 3D scanner's projector is configured for brightness and distance, and the camera's exposure is calibrated.The initial state of the fabric is set as the zero height reference.Subsequently, the acquired 3D scans are processed and analysed using the Geomagic Design X software developed by 3D Systems. Figure 2 provides an illustrative example of a 3D scan obtained during the drapability test.

Testing configurations.
In developing this novel test method, two different configurations representing distinct boundary conditions of the test setup have been investigated (see Figure 3).In configuration 1, the fabric is completely constrained to the frame using a rectangular metal plate and bolts.This configuration simulates fully constrained boundary conditions.It is similar to the biasextension test to investigate the in-plane shear kinematics of a fabric [25].The main concern for this configuration is that the boundary conditions might lead to the full locking of the fabric, especially preventing its shear and instead testing the backing fibres in tension.Therefore, this configuration does not represent a fabric layer laid out in the wind turbine blade mould.In configuration 2, the fabric is 4 partially constrained to the frame at the diagonal positions.It gives freedom to the fabric to move around, replicating the fabric placement scenario in a wind turbine blade.The analysed area of the fabric is 65 mm away from the clamps, following the Saint Venant principle [26] to leave the boundary conditions far enough from the measurement area (see Figure 1).

Results and discussion
Figure 4 shows the load-displacement curve for the same non-binder fabric tested with two configurations.In configuration 1, the resultant load is higher than in configuration 2. In configuration 1, the backing fibres oriented at ±80° are stretched due to the fully constrained boundary conditions.Therefore, the measured load is not describing the shear load necessary to shear the fabric but also the tensile load carried by the backing fibres.Configuration 2 shows a lower applied load than configuration 1, representing the shear load needed to shear this specific fabric.In this configuration, shearing between the unidirectional (UD) rovings is thus possible, leading to wrinkle formation after a certain angle.Therefore, the rest of the study used configuration 2 of the drapability tester setup.The decision has been taken to observe the wrinkle height and width as a function of the applied shear angle rather than monitoring load and displacement, as it is more relevant and practical for wind turbine blade manufacturing applications.This study defines drapability as an intrinsic property of fabric to shear.Specifically, it refers to the potential of the rovings to slide with respect to each other without creating wrinkles.Two non-crimp fabric types, with and without binder, have been analysed in terms of the evolution of wrinkles with respect to the applied shear angle.Figure 5 shows an example of wrinkle evolution with increasing shear angles of 0°, 4°, 6°, 8°, 12°, and 16° for a non-binder fabric.Figure 6A shows the relationship between the average maximum height of fabric wrinkles and the applied shear angle (°) for a standard non-binder fabric.The "average maximum height" refers to the arithmetic mean of maximum height values for all the observed wrinkles.Notably, no wrinkles were observed at a shear angle of 4°.This observation indicates that at low shear angles, the fibre rovings possess the ability to slide, thereby avoiding wrinkle formation.This finding aligns well with production and design data.After 4°, the backing fibres are stretched, creating a locking phenomenon on the UD rovings.The UD rovings are no longer capable of sliding next to each other, and as the displacement continues to increase, the fabric wrinkles.In other words, wrinkles appear when the shear is prevented, and Poisson's effect becomes dominant.The non-binder fabric shows an increasing trend in the average maximum height of the wrinkle with an increasing shear angle.Figure 6B shows the average width of fabric wrinkle at the maximum height of the wrinkle.The width of the wrinkle also shows an increasing trend with an increasing shear angle.Figure 7A presents the average maximum height of fabric wrinkle as a function of the applied shear angle for a binder fabric.Wrinkles were observed at a shear angle of 4°, with an average maximum height of 10 mm.It is highly likely that wrinkles begin to appear at a shear angle lower than 4°.The binder fabric demonstrated an increasing trend in the average maximum height of wrinkles as the shear angle increased.However, at a shear angle of 12°, the height of the wrinkles reached a constant value.Figure 7B shows the average width of fabric wrinkle at the maximum height of the wrinkle.The width of the wrinkle shows an increasing trend until a shear angle of 6°, then decreases with increasing shear angle.In comparison to the non-binder fabric, the binder fabric exhibited lower drapability.The binder material acts as a glue between the UD rovings and prevents their movement relative to each other.Due to this phenomenon, coupled with the locking effect from the backing bundles, binder fabrics exhibit a higher average wrinkle height and width than standard fabrics with no binder.Figures 8A and 8B show the height-to-width ratio of the wrinkle for non-binder and binder fabric types, respectively.The severity of a wrinkle can be measured by the aspect ratio of its height and width [27].Mendonça et al. [28] evaluated the effect of wrinkle defects on the stiffness response of composite laminates for wind turbine blades.According to their definition of aspect ratio (height/halfwidth), a wrinkle with a 1/5 aspect ratio was the most severe one.Compared to flat composite laminate, the laminate with wrinkle showed a knockdown of 54% in stiffness properties.In this study, on the fabric level, the aspect ratio used by Mendonça et al. corresponds to the aspect ratio of 1/10.This ratio is also plotted in the dashed line in Figure 8.In this study, an aspect ratio of 1/10 corresponds to an applied shear angle of 9° and 3° for non-binder fabric and binder fabric, respectively.These results clearly showed the effect of binder on the fabric, reducing the drapability by a factor of 3. The aspect ratio measured on the fabric level will not directly translate to the same aspect ratio on a composite level, given the fact that the fabric layers will undergo the vacuum infusion process, potentially reducing the overall size of the wrinkle.There is a scope to study the characteristics of wrinkles prior to and after infusion stages.The purpose is to bridge a link between a wrinkle on a fabric to the wrinkle on the composite laminate level, drawn up to the structural level.Manufacturers implement various procedures to detect and repair these defects.Therefore, avoiding these defects and ensuring a high level of manufacturing quality and productivity is vital.This study provides a basis for identifying the potential zones for repair exceeding the permissible limit.The results show that to resolve challenges associated with the upcoming fabric types and avoid wrinkles forming, it would be necessary to define and qualify drapability more precisely.

Conclusion
A novel methodology has been introduced for characterising fabric drapability as an intrinsic property associated with the shear angle.In this study, two types of non-crimp fabrics, with and without binder, have been examined to investigate the evolution of wrinkles in response to varying shear angles.A state-of-the-art 3D scanning technique was employed to quantify the aspect ratio at shear angles ranging from 0° to 16°.For the non-binder fabric, a critical aspect ratio of 1/10 corresponded to a shear angle of 9°, while for the binder fabric, it corresponded to a shear angle of 3°.The findings clearly demonstrated the detrimental effect of binders on fabric drapability, resulting in a three-fold reduction.These results provide valuable insights into the impact of binders and other parameters on wrinkle formation, enabling better control of these factors to avoid manufacturing defects in wind turbine blades.Consequently, this novel experimental method could assist the industry in selecting suitable fabrics according to their drapability for specific blade designs, thereby minimising postproduction repairs.However, this study also highlights challenges, including the need for a precise and universally accepted definition of drapability to prevent wrinkle formation, particularly in regions with highly curved geometry.Future investigations will focus on refining the aspect ratio measurement at lower shear angles and establishing the relationship between fabric wrinkles and laminate wrinkles, as well as their influence on the mechanical properties of laminates at a structural level.

Figure 1 .
Figure 1.Experimental drapability set-up enabling fabric shear and wrinkle detection by 3D blue light scanning.

Figure 2 .
Figure 2. Example of a 3D scan showing wrinkling of the non-crimp fabric at 6° shear.

Figure 3 .
Figure 3. Two configurations of the drapability tester set-up.The left side represents a completely constrained configuration, whereas the right represents a partially constrained configuration.

Figure 4 .
Figure 4. Load vs displacement curves of non-binder fabrics tested with two configurations of the drapability tester set-up.

Figure 5 .
Figure 5. Example of the evolution of wrinkles in the non-binder fabric with increasing shear angle.

Figure 6 .
Figure 6.A) The average of maximum wrinkle height (in mm) for the non-binder fabric as a function of the applied shear angle B) The average wrinkle width (in mm) at the point of maximum wrinkle height.

Figure 7 .
Figure 7. A) The average of maximum wrinkle height (in mm) for a binder fabric as a function of the applied shear angle B) The average wrinkle width (in mm) taken at the point of maximum wrinkle height.

Figure 8 .
Figure 8.The height-to-width ratio of the wrinkles for A) non-binder fabric and B) binder fabric.