Carlsson, Leif A.

This research focuses on deformation constraints of honeycomb core cells in a sandwich imposed by bonds to the face sheets. Specifically, the influence of one-sided core constraints on the bending stiffness of a single-face honeycomb core sandwich is examined. To characterize the unconstrained in-plane compressive response of honeycomb core, a range of honeycomb cores was experimentally examined. Cores with a thin cell wall displayed extensive bending deformation of inclined cell walls while cores with thicker walls failed by a shear-type instability of the cells indicated by tilting of vertical cell wall segments. The modulus and compressive strength of the core were compared to the predictions from unit cell models. The results show that geometrical imperfections such as deviation from the intended cell wall angle cause in-plane anisotropy and have strong influence on modulus and strength of the core. Modulus and strength were in reasonable agreement with predictions from unit cell models for cell wall modulus and strength between 5-12 GPa and 72-171 MPa for the set of cores examined.

The objective of the research presented in this thesis is to develop analysis and test procedures for the characterization of disbonding crack growth in a honeycomb (HC) core sandwich structure. Face sheet-to-core disbonding are of particular interest to aircraft certification authorities due to several in-service occurrences. Experimental investigation was initially focused on the mode I dominated Single Cantilever Beam (SCB) test method. Various data reduction methodologies were employed to determine the fracture toughness. The MBT method produced the most consistent and conservative results. Finite element analysis (FEA) a double periodic array of hexagonal cells was conducted to determine the effective in-plane extensional modulus and Poisson ratio of the HC core. It was shown that deformation constraints on the core, due to attachment of the core to rigid face sheets, will drastically change the behavior of the HC core. The response changes from being governed by bending to stretching which substantially elevates the effective in-plane modulus. Fracture mechanics analysis of a face/core interface crack in a HC core SCB specimen was performed using FEA. The influence of in-plane properties of the constrained core on energy release rate and mode mixity phase angle was examined. Use of plane strain conditions and an elevated modulus of the constrained core in the analysis is recommended. The approach is substantiated by testing of HC core SCB sandwich. Test results showed good agreement with FEA prediction of compliance and kink angle.

The influence of voids on the moisture uptake of epoxy has been studied. Specimens with void contents from 0 to about 50% were prepared. Void geometry and content were analyzed using microscopy and density methods. Void containing dry samples were characterized by Differential Scanning Calorimetry and Dynamic-Mechanical Analysis which verified consistency of chemistry of the epoxy network. The moisture uptake of specimens immersed in distilled water at 40 °C was monitored. The rate of absorption and saturation moisture content increased with increasing void content. The moisture uptake of void-free and void containing specimens was non-Fickian. The Langmuir model provided good fits to the experimental results for specimens with low to medium void content, although the moisture uptake of the high void content specimens showed substantial deviations from the Langmuir diffusion model. The moisture diffusivity agreed reasonably with predications from the Maxwell inclusion model over a range of void contents from 0 to 50%. The state of sorbed water was examined using mass balance calculations and DSC analysis. Only 6-8% of the void volume is occupied by water at saturation. Absorbed water may be classified as free and bound water. For void-free specimens, only bound water was found. The medium and high void content specimens contained water in three states: free water, freezable bound water, and non-freezable bound water. The DSC results show that the proportions of free water and freezable bound water increase with increasing void content, while the content of non-freezable bound water decreased. Moisture induced swelling decreased with increasing void content. The swelling is attributed to the content of non-freezable bound water.

Flexible thermoplastic p1pes under field and laboratory loading conditions have been
examined in the present study. The flexible pipes were tested under truck loading
application with shallow soil cover. The pipe-soil system response includes soil stresses
around and above the buried pipes, vertical pipe crown diametral strain, and
circumferential pipe wall strains. Modeling the pipe-soil system is made using plane
strain and thin ring assumptions. A thin ring model using Castigliano's theorem is
developed to analyze the behavior and response of a flexible pipe under well defined
loading conditions and simulate the behavior of the buried pipe under the live load
application. Laboratory work was carried out to study the pipe behavior and response
under two-point, three-point, and four-point loading configurations. The thin ring model
predictions show good agreement with classical solutions specially valid for two-point
and three-point loading configurations. Laboratory results were also in good agreement with the predictions. Laboratory results show that the maximum tensile strain for the
four-point loading test occurs at inner pipe crown region. Comprehensive efforts were
made to correlate the thin ring model predictions with the field test results; however, it
appears that the thin ring model cannot be used to simulate the effect of the live load
application. A major source of the differences between the predicted and measured
values is attributed to the applied load magnitude. A further investigation was carried out
to examine the applicability of the model to study the general pipe behavior. The
predicted hoop pipe wall strain profile was found to be similar to that of the reported
strain profile by Rogers under overall poor soil support condition. Comparison of soil
stress distribution shows that the 2D prediction approach provides nonconservative
results while the FE analysis agrees more favorably with the measured pressure data.
Overall, FE analysis shows that a linearly elastic isotropic model for the surrounding soil
and flexible pipes with a fully bonded pipe-soil interface provides a reasonable prediction
for soil pressures close to the buried pipes.

Degradation of the critical components of polymer matrix composites in marine
environments had been experimentally investigated. Water absorption behavior of neat
resin and composite specimens was examined. The tensile strength of fibers was
monitored using the single filament test. The mechanical properties of the resins were
monitored by tensile, flexure, and dynamic-mechanical tests. In addition, matrix
shrinkage during cure and matrix swelling after immersion in water were monitored. The
integrity of the fiber/matrix (F/M) interface of the composite systems was studied using
the single fiber fragmentation test (SFFT). Macroscopic composites were examined using
transverse tensile and transverse flexure tests to study the influence of the integrity of the
matrix and F/M interface on the macroscopic response. In addition, for characterization
of F/M debonding in the SFFT, a fracture mechanics model and modified test procedure
were developed.

The objective of this work is to characterize the moisture transport in a unidirectional,
transversely isotropic carbon/vinylester composite. Diffusion occurs when the material is
immersed in sea water and the moisture is transported through the voids, and interface gaps. This
uptake of moisture can cause problems including matrix degradation and swelling, reduction of
fiber/matrix interface strength, etc. To characterize water transport, three diffusivities are
required, D1, D2, and D3. However, transversely isotropic material can be characterized by two
diffusivities, along and transverse to the fibers (D1=DL, and D2=D3=DT). Composite materials
may absorb moisture along the fiber/matrix interface, especially if the ends of the fibers are
exposed. This mechanism of moisture transport inside the composite is known as “wicking”,
which would increase the value of DL. Wicking is promoted by voids and unbonded regions
between the fiber and matrix.
Experiments will be conducted on specimens made from vinylester resin reinforced with
unidirectional carbon fibers. A range of specimens will be prepared and immersed in 40°C sea
water. The moisture content will be monitored until maximum saturation. Previous studies
indicate that the Fickian diffusion model is a reasonable descriptor of the moisture absorption
process. From the plots of moisture content versus time1/2, the apparent diffusivity will be
reduced for the different size panels. The longitudinal and transverse diffusivities will be
determined using a Matlab algorithm. The amount of moisture absorbed into the composite due
to wicking will be quantified from mass balance analysis and related to the longitudinal and
transverse diffusivities.

The objective of this research was to characterize the seawater transport and its effect on the transverse tensile strength of a carbon/vinylester composite. The moisture contents of neat vinylester and unidirectional carbon/vinylester composite panels immersed in seawater were monitored until saturation. A model for moisture up-take was developed based on superposition of Fickian diffusion, and Darcy’s law for capillary transport of water. Both the predicted and measured saturation times increased with increasing panel size, however the diffusion model predicts much longer times while the capillary model predicts shorter time than observed experimentally. It was also found that the saturation moisture content decreased with increasing panel size. Testing of macroscopic and miniature composite transverse tensile specimens, and SEM failure inspection revealed more fiber/matrix debonding in the seawater saturated composite than the dry composite, consistent with a slightly reduced transverse tensile strength.

The tensile strength of the fiber/matrix interface was determined through the development of an innovativetest procedure.Aminiature tensile coupon with a through-thickness oriented, embedded single fiberwas designed. Tensile testing was conducted ina scanning electron microscope (SEM)while the failure process could be observed.Finite element stress analysis was conducted to determine the state of stressat the fiber/matrix interface in the tensile loaded specimen, and the strength of the interface.Test specimensconsistingof dry E-glass/epoxy and dry and seawater saturatedcarbon/vinylester510Awere preparedand tested.The load at the onset of debondingwascombined withthe radial stressdistributionnear thefree surface of the specimento reducethe interfacial tensile strength. For glass/epoxy, was 36.7±8.8MPa.For the dryand seawater saturated carbon/vinylester specimensthetensilestrengthsof the interface were 23.0±6.6 and 25.2±4.1MPa, respectively.The difference is not significant.

An aggregate (mosaic) model is proposed to represent the structure of paper and model the mechanical properties. The model treats paper as an aggregate of three subregions of characteristic materials, viz. bonded regions, unbonded regions (free fiber segments) and voids. A computer simulation based on the Monte Carlo method is performed to generate random and oriented paper sheets and input parameters for the aggregate model. The number of fiber crossings, total bonded area, average free fiber segment length and volume fractions of bonded material and free fiber segments and apparent sheet density are obtained from the statistical geometry description of the paper structure. The upper and lower bounds on the elastic moduli and moisture swelling coefficients of void-free paper are derived based on anisotropic elasticity theory and a fiber orientation distribution parameter. The finite element method is applied to generate effective elastic moduli and moisture swelling coefficients of the aggregate model consisting of fiber crossings and segments, but no voids. The elastic moduli of paper so obtained are corrected for the voids present in paper. The predictions are compared with previously published experimental results, and it is demonstrated that the results generally fall within the theoretical bounds. The mosaic model was shown to approximate the mechanical properties of paper.

The influence of voids on the hygroelastic properties of paper has been investigated using analytical and numerical methods. Paper was modeled as a laminate made of cell-wall layers. A continuous fiber orientation distribution was introduced into the laminate model to derive the baseline properties of the papersheet. The voids in the papersheet were modeled as reinforcements with zero elastic properties. The reduction of elastic stiffnesses of isotropic materials containing different shapes and volume fractions of pores were analyzed using Voigt, Reuss, foam and combination models. Hashin's two-phase bounding model and Christensen's three-phase self-consistent models were also used to predict the elastic stiffnesses of isotropic porous materials. The influence of voids on the engineering constants of orthotropic materials was analyzed using 2-D and 3-D finite element models. The invariance of hygroexpansion in the presence of voids was demonstrated using analytical and numerical methods. The theoretical model predictions were correlated with previously published experimental results.