An Investigation of a Vertical Test Method for Large Deformation Bending of High Strain Composite Laminates

Abstract

Research to date indicates that traditional composite material failure analysis methods are not appropriate for thin laminates in flexure. Thin composite structures subjected to large bending deformations often attain significantly higher strain-to-failure than previously anticipated tensile and compression coupon test data and linear material model assumption predict. At NASA Langley Research Center, a new bend test method is being developed for High Strain Composite (HSC) structures. This method provides an adequate approximation of a pure moment, large deformation bend test for thin-ply, high strain composites to analyze the large strain flexure response of the laminates. The objective of this research was to further develop this new test method to measure the true bending stiffness and strain-to-failure of high strain composite materials. Of primary importance is the ability to characterize composite laminates that are of interest for current NASA deployable structures in both materials and layups. Two separate testing campaigns were performed for the development of the testing procedure. Initially six laminates were bend tested in three different fiber orientations. These laminates were some combination of unidirectional intermediate modulus (IM) carbon, high tenacity (HT) carbon plain weave, and astroquartz plain weave composite materials. The second test campaign was performed as a more detailed look into the simplest composite laminates at thicknesses that better represented deployable boom structures. The second campaign tested three basic, thinner laminates, again in three different fiber orientations. All testing was monotonic loading to failure. The thickness of the laminates tested ranged from 0.166mm (campaign 2) to 0.45mm (campaign 1). The measured strains at failure for the unidirectional material were approximately 2.1% and 1.4% at the compression and tension sides, respectively, failing as fiber tensile fracture. Both of these values differ from what would be expected from considering much thicker coupons tested under pure compression and tension, that show a strain-to-failure of 1.0-1.1% and 1.6-1.7%, respectively. The significant differences in strain values obtained at the outer surfaces of the coupon is thought to be related to the shift in neutral axis that the specimen experiences during the large deformation bending test as a result of fiber material nonlinearities at higher strains. The vertical test nature of the CBT when compared to other test methods proves to be helpful for visually capturing with Digital Image Correlation the distinct behavior of the flexure on both the compressive and tensile sides. It was found that the thinner the laminate tested, the more confirmation of a nonlinear response of this classification of composites. The moment versus curvature curves were predominantly nonlinear resulting in a near linear bending stiffness versus curvature response. At these large strains, carbon fibers are highly nonlinear resulting in the laminate flexure modulus increasing by up to 5x. The theoretical bending stiffness values calculated using Classical Lamination Theory analysis are within small differences with respect to the experimentally measured values: errors of approximately 5-10% for both D11 and D22. The error between the finite element model computed strain response and the experimental values was on average around 22%, with 35% of the laminates and orientation having errors less than 7%. Comparison between CLT, FEA, and experimentation show that the Column Bend Test appears to be a promising candidate for characterization of large deformation bending behavior of thin-ply high strain composite laminates.