CHARACTERIZATION OF THE DRAPABILITY OF REINFORCEMENT FABRICS BY MEANS OF AN AUTOMATED TESTER

Transcription

1 CHARACTERIZATION OF THE DRAPABILITY OF REINFORCEMENT FABRICS BY MEANS OF AN AUTOMATED TESTER Mirko Christ 1, Andrea Miene 1, Ulrich Moerschel 2 1 Faserinstitut Bremen, Am Biologischen Garten 2, Bremen,Germany, 2 Textechno Herbert Stein GmbH & Co. KG, Dohrweg 65, Moenchengladbach, Germany Abstract The behavior of standard and non-crimp fabrics in forming and draping is of importance in any production process of non-flat, fabric-reinforced composite parts. A new automatic drapability tester allows to automatically characterize drapability and the formation of defects during draping and forming. The tester combines the measurement of the force, which is required for forming with an optical analysis of small-scale defects such as gaps and loops by means of image analysis. An optional triangulation sensor can determine large scale defects such as wrinkles. Background and Motivation Due to the rising use of fiber-reinforced plastics (FRP) in automotive and aerospace applications technical textiles are increasingly used for structural load-bearing components. The designers and process developers of such materials have to know the properties of the textiles reinforcements in order to optimize both, the performance of the components and the production processes. A significant difference in the use of reinforcing textiles compared to metals is that the drapability of the fabrics must be taken into account. At the same time, the examination and quantified description of draping is not yet developed to the extent that is necessary to support component design, process development, and simulation. In a collaborative research project test methods and analytical descriptions for analyzing the drapability of non-crimp fabrics (NCF) from glass and carbon fibers were developed. Preforming and Drapability Preforming is the forming of flat textiles to a near-net shape geometry, i.e. the shape generating and, therefore, the most value-adding process step. The drapability of textile structures is a decisive factor for the design and manufacturing of fiber-reinforced plastics and therefore crucial in the manufacturing process of composite parts. In the field of textile reinforcement structures drapability describes the behavior of the textile fabric during the forming process from its flat initial state into a three-dimensional shape preforming processes are still done manually in most instances. A better understanding of the drapability could allow an automated control of draping apparatuses, which may be used for an automated preforming process for complex topologies, in contrast to the simple geometries to which today s automated preforming process are limited [3]. Especially with multiple curved non-euclidean faces, different draping effects and draping modes occur. Shearing leads to fiber disorientation, which can lead to significant deviations in direction of the local stiffness of the final component compared to the original design. Thus, the matrix is exposed to increased stress in these areas, resulting in unacceptably high component strains which can lead to premature failure of the component. [1] For woven fabrics the shearing behavior is the most important draping modus. Regarding Page 1

2 non-crimp fabrics the draping mode of fiber dislocation is important. Gaps will occur between the rovings. This leads to resin-rich areas in these draping defects during the infiltration process. These resin-rich areas can be starting points of cracks. Figure 1: Drapability effects: Gaps (left), loops (middle), wrinkles (right) Drapability Testing The picture frame test is the most well known test of the shearing behavior of textiles. For woven fabrics it can be used to measure the critical shear angle which characterizes the onset of Figure 2: Picture Frame Test [2] wrinkling. However, it cannot be used for testing non-crimp fabrics. Also the fact that picture frames are not standardized and each picture frame test setup is an individual setup that is developed by the institution or company using it lead to poor comparability between the different setups. Only small changes in the setup can lead to substantial changes in the measured forces. The extent and intensity of draping effects can only be measured qualitatively. Other classical mechanical testing, such as friction testing [4] and testing of bending stiffness and further mechanical properties, can give indirect information on the drapability of a fabric [5], [6]. Development of a drapability tester in the frame of a research project Within a public funded research project the new drapability tester was jointly developed with industry- and research partners. This device allows the measurement of draping effects like gaps, loops, wrinkles, and fiber misalignment using optical systems. This makes the draping effects quantifiable and, therefore, available to semi-analytical- as well as numerical methods for the examination of the drapability. In order to get input from all relevant users and researchers, a consortium was formed that encompasses a wide field of necessary knowledge. Page 2

3 The partners (in alphabetical order) and their respective fields of expertise were: Faserinstitut Bremen e.v (FIBRE), Bremen Research institute; preforming and optical analysis Haindl Kunststoffverarbeitung GmbH, Bremen Enterprise; manufacturer of CFRP parts Institute for Textile Technologies (ITA) of RWTH Aachen University, Aachen Research institute; Textile testing J. Schilgen GmbH & Co. KG, Emsdetten Enterprise; manufacturer of woven fabrics SAERTEX GmbH & Co. KG, Saerbeck Enterprise; manufacturer of non-crimp fabrics Textechno H. Stein GmbH & Co. KG, Moenchengladbach Enterprise; manufacturer of testing equipment Together with these partners, the tester was developed, tested on different kinds of fabrics, verified with traditional testing methods, and brought into serial production. The tester combines the multi-axial testing of a fabric over a reference body with the quantitative measurement of several draping effects by means of digital image analysis. Previous testers also employed the method of reference bodies and measured the necessary force, but the measurement and communication of drapability was not sufficient. A variety of technical textiles can be draped with the new tester over generic reference bodies. As a standard, a spherical calotte is used. This corresponds to the multi-axial stresses during the draping of double curvature shapes. During the draping the necessary force is recorded and the resulting draping effects are being captured by a digital camera. The digital image analysis system is capable of detecting the following effects: Gaps between rovings, Fiber misalignment, Loops, Waves, Wrinkles. The image analysis and the laser-triangulation scanner make it possible to turn the occurring effects into hard numbers. The tester, due to its speed and automatic evaluation of results, is well suited for statistical methods. From the use of different reference bodies the transferability of the measured results to different geometries can be derived. Technical realization of the tester The tester is based on an idea and earlier prototype produced by SAERTEX. As a standardized simulation of the draping process a flat circular sample of the fabric to be tested is deformed at its center from below by means of a motor-driven calotte having a ball-shaped top. The sample is held back by means of a clamping ring and a membrane system which is subject to controlled air pressure. All surfaces in contact with the sample have a surface coating with a low coefficient of friction, so that varying the pressure allows to cover a wide range of retaining forces and the sample will not be damaged. The tester is connected to a standard Windows PC for setting test conditions and for data acquisition and storage. Test conditions can be stored in parameter sets, which contain all information about the test, the evaluation, and the form of the protocol to be generated. The number of different parameter sets is unlimited. Page 3

4 Sample diameter can vary from 310 mm to 330 mm, the latter being recommended as at this size the sample will still be completely inside of the retaining mechanism up to an elevation of about 90 mm. Samples having a thickness of up to about 8 mm can be clamped with the standard clamping ring. Clamping rings for thicker samples can be manufactured. A force transducer inside the calotte measures the force which is required to form the sample. Since force is subject to relaxation to a certain extent both maximum- and average force at each elevation level are stored. The calotte drive has a maximum elevation of 100 mm allowing a strong deformation of the sample, if required. Elevation steps are freely programmable in the parameter set in tens of a millimeter. A test could e.g. contain the elevations 0, 20, 40, 60, 75, 85, 95 mm. Figure 3: Principle arrangement of the sample, calotte, and clamping ring at elevations 0 mm, 40 mm, and 80 mm. In contrast to the original prototype by SAERTEX, the new tester is equipped with additional optical detection of the defects arising during forming: A high-resolution camera with an appropriate illumination inspects the sample for fine-scale defects, such as gaps and changes in fiber orientation. The picture quality thus has to allow resolving the individual fibers. The image field has a size of about 30 mm x 40 mm. In order to inspect a maximum percentage of the surface the camera can be positioned in three axis and the sample can rotate by more than 360 degrees. The angle range to be inspected as well as the number of pictures to be taken around can again be defined within the parameter set. A slight overlap of the pictures is recommended. Since the camera can take a 2-dimensional picture of the sample only, an additional lasertriangulation sensor is used in order to define the surface morphology of the sample before and after forming. The sample is rotated while the laser sensor scans the sample continuously. In this way the tester detects large-scale defects such as wrinkles and plies. Page 4

5 Figure 4: Arrangement of detail camera (left), overhead camera (top), and laser-triangulation sensor (right) over the sample at an elevation of 40 mm. A second, medium-resolution camera (overhead camera) takes pictures of the sample before and after the test. At this stage one picture is taken each, with the calotte still at maximum elevation and with the calotte back to elevation 0 mm. This allows evaluating the retraction of a sample in the different orientations after having been draped into a form. A complete drapability test on the instrument comprises the following steps 1. Take the clamping ring out. 2. Insert the circular sample. The line of the laser triangulation sensor is projected onto the sample in order to ease the proper orientation of the sample. 3. Press the green control button to take an overhead picture. 4. Insert the clamping ring and close the locking knobs. 5. Press the green control button in order to establish the clamping pressure and start testing. 6. The tester starts running taking the programmed detail pictures while rotating the sample. When all detail pictures are taken, the sample is rotated again for scanning the surface with the triangulation sensor. Step 6 is repeated for all other programmed elevation steps. The tester indicates when testing at all elevation steps is completed. 7. Take the clamping ring out. 8. Press the green control button in order to take the second overhead picture, drive the calotte down and take the final overhead picture. 9. Take the sample out. Page 5

6 A typical test with several elevation steps takes between 5 and 15 minutes per sample. Figure 5: Realization of the tester. All interactions with the tester during testing are controlled by the control keys on the tester Page 6

7 Results The primary result of the basic instrument is the force measured by the force transducer inside the calotte. This force is composed of the force required to form the sample and the force it needs to pull the sample out of the clamping ring. Measuring force/elevation-curves as shown in figure 6 at different pressure levels shall allow to distinguish between the forming and the friction properties. Figure 6: Average values of the force measured by the force transducer at different elevations for two similar samples. The measurement was performed at high clamping pressure. From the pictures taken by the high-resolution detail camera the image analysis software determines the width, length, and area of gaps, the degree of fiber misalignment [9, 10], all as functions of the elevation and orientation of the camera towards the sample. The image analysis software which determines these parameters has been developed by the project partner Faserinstitut Bremen (FIBRE). No commercial software has been used for this purpose. The algorithms determining these parameters are based on a detection of the fiber orientation as well as on contrast, depending on the respective parameter and the material in question. Area and volume of wrinkles are measured using the laser-triangulation sensor. With the overhead camera the overall deformation of the original circular shape of the specimen can be measured. From this data a complete description of the drapability of a fabric can be derived. The tester automatically groups the measured areas of gaps and fiber misalignment into three user-definable classes. This gives a very quick impression of the draping effects that result from the deformation of the specimen by the spherical calotte. The classes are defined by the gap width and the degree of misalignment. This can be defined as the intensity of the effect. The extent of the effect would then be the overall area affected by e.g. gaps, independent of the width of the individual gaps. For the sake of brevity only results of the gap and fiber alignment detection are shown here. Page 7

8 Figure 7: Results of the fiber orientation measurement after an elevation of 95 mm at 0 (left picture) and 90 orientation relative to the upper fiber layer. Large misalignments are colored red, medium misalignments yellow and small misalignments green. At the 90 position more misalignments have occurred compared to the 0 position. This corresponds to the diagrams in figure 8. The tested sample was a glass-fiber non-crimp fabric with a tricot warp-knit binding. The following data points are the average of five repetitions of the same experiments. The data for the fiber misalignment show the results of greater misalignment at the 90 position where the fibers of the upper layer run perpendicularly to the curvature of the hemisphere. Figure 8: Fiber misalignment at two different orientations and various elevation levels. The data for the gap width distribution are even more interesting. Gaps are already present in the flat specimen due to binding yarn. At 0 mm elevation the gap width distribution is equal at Page 8

9 the 0 and at the 90 position. Any difference is solely due to manufacturing tolerances. When the specimen is deformed, the gaps at the 0 position slowly close, while the gaps at the 90 position are opened wide. This can be traced back to the straightening of the fibers at the 0 position while the fibers at the 90 position fan out to cover the increased area of the hemispherical surface compared to the flat plane. Figure 9: Results of the gap detection after an elevation of 95 mm at 0 and 90 position. Wide gaps are colored red, medium gaps yellow and narrow gaps green. At the 90 position more wide gaps have occurred compared to the 0 position. This corresponds to the diagram in figure 10. Figure 10: Gap width distribution at different orientations and various elevation levels. Page 9

10 Although image analysis on carbon fiber fabrics is tough because of the high reflectivity of carbon fibers, the detection of gaps and misalignments works well for these specimens, too. Figure 11: Results of the classification of misalignment (left picture) and gaps in a carbon fiber NCF For the misalignment classification in woven fabrics the image analysis works similar as for NCF. For the gap classification, it detects the largest diagonal line as the gap width. Figure 12: Results of the classification of fiber misalignment (left picture) and gaps in a glass fiber woven fabric. Page 10

11 The laser-triangulation sensor delivers data, which after filtering can be used to generate a 3D model of the specimen and/or false color images. A statistical evaluation of the elevation at a given radius of the sample can used to describe the unevenness of the sample. Figure 13: 3D-model and false color image generated from data of the laser-triangulation sensor taken on a carbon fiber sample. The pictures from the overhead camera also show the general deformation of the round specimen. This deformation is caused because the different fiber alignments in the non-crimp fabrics are pulled in different directions once the specimen is deformed. This deformation of the sample can be quantified by means of the image analysis software, too. Figure 14 Deformation of specimen for glass (left picture) and carbon as documented in pictures of the overhead camera. The right hand picture also shows results of the image analysis software: The red line follows the specimen contour; the yellow circle describes a circle with the same area as the specimen. Page 11

12 Conclusion and Outlook Measuring and predicting the drapability of a reinforcement fabric has always been a main concern when designing and producing composite parts. With today s trend towards fast and automated large-scale production processes the importance of drapability will further increase. Nevertheless, no drapability testing method was commercially available so far, that let the fabric or part designer understand how a fabric and part shape will interact. With the automatic drapability tester DRAPETEST Textechno Herbert Stein GmbH & Co. KG from Germany fills this gap. Other than the FAVIMAT+, Textechno s outstanding automatic single-fiber tester, which is already used for carbon and glass fibers in about 30 labs world-wide, the new drapability tester is still at the beginning of its career. There is, however, an enormous interest from industry and institutions, recently documented by the JEC Asia 2012 Innovation Award. The image analysis software will be continuously refined in order to allow tests on new basic materials and fabric constructions. Other forming geometries than the spherical calotte can be realized. The knowledge of the draping behavior of the reinforcement textile supports both the development of the component as wells as the optimization of the process. The degree of fiber disorientation can be used as direct input into the laminate design. The fiber orientation of the flat ply stack can be chosen in a way, that the fibers will change their orientation during the preforming into the desired direction. The occurrence and distribution of gaps is of great importance to a stable and reliable process. Gaps that form during the preforming process will influence the later resin infiltration. Single large gaps can lead to resin rich regions that can be the origin of cracks during the use of the part. A large area of small gaps can change the permeability of the textile and lead to dry parts after the infiltration, because the resin flow will be diverted to the area of high gap occurrence. The presented drapability tester can differentiate between such draping effects and guide the engineer during the material selection and process optimization. The benefit of the tester will depend on the field of activity of the user: For the producer of the reinforcement roving, tests of drapability on standardized fabrics can help to optimize the roving properties for best drapability. The major property to be modified will probably be the sizing, but also the single-filament diameter and the shape of the cross-section might be optimized. For the producer of fabrics drapability testing helps with the selection of the raw material, the construction of the fabrics, and the machine settings for optimum drapability For the composite producer the tester or data from the tester provided by the supplier will allow to select the optimum fabric on the world market for a certain application. Institutes working on draping simulation the drapability tester can to a certain deduct input data for the simulation from some of the results. In any case the tester can be used as an ideal tool to validate simulation software and parameter settings, if the geometry of the instrument is simulated by the software. One step still to go is to reduce the numerous results of the tester to a few numbers which precisely characterize the drapability in a simple but precise and reproducible way. The Faserinstitut Bremen (FIBRE) is on a good way in this process. Page 12

13 The existence of a standardized testing method and equipment generating such numbers will simplify the communication between researchers, designers and manufacturers. This will hopefully lead to an international harmonization of drapability testing. Acknowledgements We thank the involved partners for their support and the excellent cooperation, the German Federal Ministry for Economics and Technology for funding the project within the scope of ZIM (Zentrales Innovationsprogramm Mittelstand), and the JEC for granting the JEC Innovation Award References 1. Christ, M.; Sköck-Hartmann, B., Krieger, H., Herrmann, A.S., Gries, T.: Analysing the draping behaviour of glass fiber fabrics. International Glass Fiber Symposia 2012, Aachen. (submitted) 2. Christ, M.; Sköck-Hartmann, B., Krieger, H. Hoffmeister, C., Herrmann, A.S.; Gries, T.: Prüfverfahren und Methoden zur Charakterisierung der Drapierbarkeit von technischen Textilien. Tagungsband zur 13. Chemnitzer Textiltechnik-Tagung 14. und 15. März 2012, p Purol, H.: Entwicklung kontinuierlicher Preformverfahren zur Herstellung gekrümmter CFK- Versteifungsprofile. Berlin: Logos Verlag, Gilhaus, K., Lünenschloß, J.: Die Ermittlung des Reibungskoeffizienten an Einzelfasern. Melliand Textilberichte 2 (1979), S Long, A. C. (Hrsg.): Design and manufacture of textile composites. Cambridge, England. Woodhead Publishing Ltd., Lomov, S., Boisse, P., Deluycker, E., Morestin, F., Vanclooster, K., Vandepitte, D., Verpoest, I., Willems, A.: Full-field straim measurements in textile deformability studies. In: Composites Part A 29 (2008), S Cherif, C.; Drapierbarkeitssimulation von Verstärkungstextilien für den Einsatz in Faserverbundkunststoffen mit der Finite-Elemente-Methode. Aachen: Shaker, 1998; Zugl. Aachen: Techn. Hochsch., Diss., Palinsky, A., Eisenhauer, C., und Ischtschuk L.: Effizientere Faserverbundherstellung durch den Einsatz funktionalisierter Multiaxialgelege. 2nd Aachen-Dresden International Textile Conference. In: DWI Reports (2008) S Miene, A.: Looking right into the Texture. Kunststoffe international 2009/05, p Göttinger, M., Weimer, C., Miene, A.: In-Line Preform Prozesskontrolle in der CFK-Fertigung. Deutscher Luft- und Raumfahrtkongress , Aachen. Page 13