SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
The fracture of thick laminated graphite/epoxy composites has been the subject of an extensive research program. The program was divided into three major areas of investigation which included laminate thickness, laminate stacking sequence, and part-through surface flaws. The results from this program are reviewed with emphasis placed on their applicability to the design of thick laminated composite structures.
The fatigue response of a (0/24/90/-45)s4 T300-5208 graphite-epoxy laminate with a drilled centerhole subjected to fully reversed tension-compression (R=-1, T-C) constant amplitude loading was investigated. Damage evaluation techniques such as stiffness monitoring, penetrant-enhanced X-ray radiography, C-scan, laminate deply and residual strength were used to establish the mechanisms of damage development as well as the relations between this damage and the stiffness, strength and life of the laminate. Damage initiated at the hole as matrix cracking parallel to the fibers in all plies. Matrix cracks had a significant effect on delamination initiation and growth. Delaminations initiated near the surface in the densely cracked region at the hole and grew along major matrix cracks. Compressive properties degraded more rapidly than tensile properties. At the stress levels used, residual tensile strength increased early in the fatigue life and remained approximately constant to near the end of life, when failure was precipitated by excessive laminate instability during the compressive portion of the loading.
The objectives are to present the derivation of the new virtual crack closure technique, evaluate the accuracy of the technique, and finally to present the results of a limited parametric study of laminates with a postbuckled delamination. Although the new virtual crack closure technique is general, only homogeneous, isotropic laminates were analyzed. This was to eliminate the variation of flexural stiffness with orientation, which occurs even for quasi-isotropic laminates. This made it easier to identify the effect of geometrical parameters on G. The new virtual crack closure technique is derived. Then the specimen configurations are described. Next, the stress analyses is discussed. Finally, the virtual crack closure technique is evaluated and then used to calculate the distribution of G along the delamination front of several laminates with a postbuckled delamination.
Strain energy release rates for edge delaminated composite laminates were obtained using quasi 3 dimensional finite element analysis. The problem of edge delamination at the -35/90 interfaces of an 8-ply composite laminate subjected to uniform axial strain was studied. The individual components of the strain energy release rates did not show convergence as the delamination tip elements were made smaller. In contrast, the total strain energy release rate converged and remained unchanged as the delamination tip elements were made smaller and agreed with that calculated using a classical laminated plate theory. The studies of the near field solutions for a delamination at an interface between two dissimilar isotropic or orthotropic plates showed that the imaginary part of the singularity is the cause of the nonconvergent behavior of the individual components. To evaluate the accuracy of the results, an 8-ply laminate with the delamination modeled in a thin resin layer, that exists between the -35 and 90 plies, was analyzed. Because the delamination exists in a homogeneous isotropic material, the oscillatory component of the singularity vanishes.
An arbitrarily laminated, anisotropic cylindrical shell of finite length, under uniform internal pressure, is analyzed using Love-Timoshenko's kinematic relations and under the framework of classical lamination theory. The previously obtained solutions for asymmetrically laminated orthotropic (cross-ply) as well as unbalanced-symmetric and balanced-unsymmetric (angle-ply) cylindrical shells under the same loading conditions have been shown to be special cases of the present closed-form solution. Numerical results have been presented for a two-layer cylindrical shell and compared with those obtained using finite element solutions based on the layerwise constant shear-angle theory. These are expected to serve as benchmark solutions for future comparisons and to facilitate the use of unsymmetric lamination in design.
A two-dimensional finite-element model has been developed to calculate interlaminar stresses and strain energy release rates for the study of delamination in composite laminates subjected to various loads. This paper addresses the formulation, implementation, and verification of a suitable 'element for uniaxial tension, bending, and torsion', as well as parametric studies of the interlaminar stresses and strain energy release rates for various composite laminates. The study concentrates on establishing relationships of GI, GII, and GIII as functions of crack length and stacking sequence in laminates due to bending. Preliminary results of laminates subjected to torsional loads are included to increase the applicability and completeness of the present formulation.
The reduction in the compressive and tensile strengths of graphite-epoxy laminates with thickness discontinuities due to dropped plies was studied by experiment and analysis. The specimens were fabricated with all the dropped plies lumped together in the center of a sixteen-ply quasi-isotropic layup, such that one surface was flat and the slope of the opposite surface changed abruptly at the dropped ply location to accommodate the thickness change. Even though the thick and thin sections are symmetrically laminated, there exists bending-extension coupling due to the geometric eccentricity of the middle planes of the thick and thin sections. Experiments were conducted on fifty-four specimens that differed in the configuration of the dropped plies only. The strength of a laminate with dropped plies is less than the strength of its thin section, and the compressive strength of a laminate with dropped plies is less than its tensile strength. The reduction in strength is directly related to the axial stiffness change between the thick and thin sections. To examine the mechanism of failure, the three-dimensional state of stress in the dropped ply region was evaluated by the finite element method. A tensile interlaminar criterion predicted the correct location of failure, but underestimated the failure load.
Compressive failure mechanisms in quasi-isotropic graphite/epoxy laminates were characterized for both unnotched and notched specimens and also following damage by impact. Two types of fibers (Thornel 300 and 700) and four resin systems (Narmco 5208, American Cyanamid BP907, and Union Carbide 4901/MDA and 4901/mPDA) were studied. For all material combinations, failure of unnotched specimens was initiated by kinking of fibers in the 0-degree plies. A major difference was observed, however, in the mode of failure propagation after the 0-degree ply failure. The strength of quasi-isotropic laminates in general increased with increasing resin tensile modulus. The laminates made with Thornel 700 fibers exhibited slightly lower compressive strengths than did the laminates made with Thornel 300 fibers. The notch sensitivity as measured by the hole strength was lowest for the BP907 resin and highest for the 5208 resin. For the materials studied, however, the type of fiber had no effect on the notch sensitivity.
Natural vibration frequencies of orthotropic and anisotropic simply supported right circular cylinders are predicted using a higher-order transverse-shear deformation theory. A comparison of natural vibration frequencies predicted by first-order transverse-shear deformation theory and the higher-order theory shows that an additional allowance for transverse shear deformation has a negligible effect on the lowest predicted natural vibration frequencies of laminated cylinders but significantly reduces the higher natural vibration frequencies. A parametric study of the effects of ply orientation on the natural vibration frequencies of laminated cylinders indicates that while stacking sequence affects natural vibration frequencies, cylinder geometry is more important in predicting transverse-shear deformation effects. Interaction curves for cylinders subjected to axial compressive loadings and low natural vibration frequencies indicate that transverse shearing effects are less important in predicting low natural vibration frequencies than in predicting axial compressive buckling loads. The effects of anisotropy are more important than the effects of transverse shear deformation for most strongly anisotropic laminated cylinders in predicting natural vibration frequencies. However, transverse-shear deformation effects are important in predicting high natural vibration frequencies of thick-walled laminated cylinders. Neglecting either anisotropic effects or transverse-shear deformation effects leads to non-conservative errors in predicted natural vibration frequencies.
A series of angleplied graphite/epoxy laminates was fabricated from overaged prepreg and tested in tension to investigate the effects of overaged or advanced cure material on the degradation of laminate strength. Results, which include fracture stresses, indicate a severe degradation in strength. In addition, the fracture surfaces and microstructural characteristics are distinctly unlike any features observed in previous tests of this prepreg and laminate configuration. Photographs of the surfaces and microstructures reveal flat morphologies consisting of alternate rows of fibers and hackles. These fracture surface characteristics are independent of the laminate configurations. The photomicrographs are presented and compared with data from similar studies to show the unique characteristics produced by the overage prepreg. Analytical studies produced results which agreed with those from the experimental investigations.
Strain energy release rates for edge delaminated composite laminates were obtained using quasi 3 dimensional finite element analysis. The problem of edge delamination at the -35/90 interfaces of an 8-ply composite laminate subjected to uniform axial strain was studied. The individual components of the strain energy release rates did not show convergence as the delamination tip elements were made smaller. In contrast, the total strain energy release rate converged and remained unchanged as the delamination tip elements were made smaller and agreed with that calculated using a classical laminated plate theory. The studies of the near field solutions for a delamination at an interface between two dissimilar isotropic or orthotropic plates showed that the imaginary part of the singularity is the cause of the nonconvergent behavior of the individual components. To evaluate the accuracy of the results, an 8-ply laminate with the delamination modeled in a thin resin layer, that exists between the -35 and 90 plies, was analyzed. Because the delamination exists in a homogeneous isotropic material, the oscillatory component of the singularity vanishes.
Experimental and analytical investigations of the fracture of a graphite/epoxy laminate with part-through semielliptic surface notches have been conducted. Experimental values of notched strength were determined for a variety of notch aspect ratios and notch depths. Analytical predictions of laminate failure were obtained from three different models. The three models were equally successful in predicting the fracture of laminates with deep notches. The experimental results of laminates with shallow notches indicated more notch sensitivity (that is greater strength reduction) than did the analytical models.
An analytical investigation of the effects of holes on the moment distribution of symmetric composite laminates subjected to bending moments is described. A general, closed-form solution for the moment distribution of an infinite anisotropic plate is derived, and this solution is used to determine stress distributions both on the hole boundary and throughout the plate. Results are presented for several composite laminates that have holes and are subjected to either pure bending or cylindrical bending. Laminates with a circular hole or with an elliptical hole are studied. Laminate moment distributions are discussed, and ply stresses are described.
Theoretical models describing the mechanical behavior of fiber-reinforced composite laminates are developed analytically and demonstrated. The curing of composite structures and the out-of-plane shapes of unsymmetric laminates are discussed and illustrated with photographs; the derivation of the governing equations is outlined; numerical results for typical laminates are presented in graphs; and particular attention is given to stability considerations and the effects of material properties, laminate thickness, aspect ratio, and stacking sequence.
The fundamental mechanics of free-edge effects in laminated fiber-reinforced composites is examined, reviewing the results of recent experimental and analytical investigations. The derivation of the governing equations for the basic problem is outlined, including the equilibrium and mismatch conditions and the elasticity formulation, and experimental data on axial displacement and shear strain in angle-ply laminates are summarized. Numerical predictions of free-edge deformation and interlaminar and through-thickness stress distributions are presented for cross-ply, angle-ply, and quasi-isotropic laminates, and the mechanisms of edge damage and failure in angle-ply laminates are briefly characterized. Extensive diagrams, drawings, graphs, and photographs are provided.
Tapered (0 deg) laminates of S2/CE9000 and S2/SP250 glass/epoxies, and IM6/1827I graphite/epoxy were tested in cyclic tension. The specimens usually showed some initial stable delaminations in the tapered region, but these did not affect the stiffness of the specimens, and loading was continued until the specimens either delaminated unstably, or reached 10(exp 6) to 2 x 10(exp 7) million cycles with no unstable delamination. The final unstable delamination originated at the junction of the thin and tapered regions. A finite-element model was developed for the tapered laminate with and without the initial stable delaminations observed in the tests. The analysis showed that for both cases the most likely place for an opening (Mode 1) delamination to originate is at the junction of the taper and thin regions. For each material type, the models were used to calculate the strain energy release rate, G, associated with delaminations originating at that junction and growing either into the thin region or tapered region. For the materials tested, cyclic G(sub Imax) values from DCB tests were used with the maximum strain energy release rates calculated from the finite-element analysis to predict the onset of unstable delamination at the junction as a function of fatigue cycles. The predictions were compared to experimental values of maximum cyclic load as a function of cycles to unstable delamination from fatigue tests in tapered laminates. For the IM6/1827I and S2/SP250 laminates, the predictions agreed very well with the test data. Predicted values for the S2/CE9000 were conservative compared to the test data.
Delamination failure in a unidirectional curved composite laminate was investigated. The curved laminate failed unstably by delaminations developing around the curved region of the laminate at different depths through the thickness until virtually all bending stiffness was lost. Delamination was assumed to initiate at the location of the highest radial stress in the curved region. A closed form curved beam elasticity solution and a 2-D finite element analysis (FEA) were conducted to determine this location. The variation in the strain energy release rate, G, with delamination growth was then determined using the FEA. A strength-based failure criteria adequately predicted the interlaminar tension failure which caused initial delamination onset. Using the G analysis the delamination was predicted to extend into the arm and leg of the laminate, predominantly in mode I. As the initial delamination grew arould the curved region, the maximum radial stress in the newly formed inner sublaminate increased to a level sufficient to cause a new delamination to initiate in the sublaminate with no increase in applied load. This failure progression was observed experimentally.
Delamination is a common failure mode of laminated composite materials. This type of failure frequently occurs at the free edges of laminates where singular interlaminar stresses are developed due to the difference in Poisson's ratios between adjacent plies. Typically the delaminations develop between 90 degree plies and adjacent angle plies. Edge delamination has been studied by several investigators using a variety of techniques. Recently, Chan and Ochoa applied the quasi-three-dimensional finite element model to the analysis of a laminate subject to bending, extension, and torsion. This problem is of particular significance relative to the structural integrity of composite helicopter rotors. The task undertaken was to incorporate Chan and Ochoa's formulation into a Raju Q3DG program. The resulting program is capable of modeling extension, bending, and torsional mechanical loadings as well as thermal and hygroscopic loadings. The addition of the torsional and bending loading capability will provide the capability to perform a delamination analysis of a general unsymmetric laminate containing four cracks, each of a different length. The solutions obtained using this program are evaluated by comparing them with solutions from a full three-dimensional finite element solution. This comparison facilitates the assessment of three dimensional affects such as the warping constraint imposed by the load frame grips. It wlso facilitates the evaluation of the external load representation employed in the Q3D formulation. Finally, strain energy release rates computed from the three-dimensional results are compared with those predicted using the quasi-three-dimensional formulation.