Search

Elastic-plastic finite-element analyses were used to calculate stresses and displacements around a crack in finite-thickness plates for an elastic-perfectly plastic material. Middle- and edge-crack specimens were analyzed under tension and bending loads. Specimens were 1.25 to 20 mm thick with various widths and crack lengths. A global constraint factor alpha(sub g), an averaged normal-stress to flow-stress ratio over the plastic region, was defined to simulate three-dimensional (3D) effects in two-dimensional (2D) models. For crack lengths and uncracked ligament lengths greater than four times the thickness, the global constraint factor was found to be nearly a unique function of a normalized stress-intensity factor (related to plastic-zone size to thickness ratio) from small- to large-scale yielding conditions for various specimen types and thickness. For crack length-to-thickness ratios less than four, the global constraint factor was specimen type, crack length and thickness dependent. Using a 2D strip-yield model and the global constraint factors, plastic-zone sizes and crack-tip displacements agreed reasonably well with the 3D analyses. For a thin sheet aluminum alloy, the critical crack-tip-opening angle during stable tearing was found to be independent of specimen type and crack length for crack length-to-thickness ratios greater than 4.

A 24-ply composite double cantilever-beam specimen under mode I (opening) loading has been analyzed by a 3D FEM code that calculated along a straight delamination starter for several different specimen materials. An isotropic specimen was found to have a strain-energy release rate distribution which varied along its delamination front due to the boundary-layer effect and another effect associated with the anticlastic curvature of the bent specimen arms. A 0-deg graphite-reinforced epoxy specimen had a nearly-uniform strain-energy release rate distribution which dropped only near the edge, due to the boundary-layer effect, and a +/- 45-deg graphite/epoxy specimen exhibited a pronounced strain-energy release rate variation across the specimen width.

Fiber/matrix (F/M) cracking was analyzed to provide better understanding and guidance in developing F/M interface fracture toughness tests. Two configurations, corresponding to F/M cracking at a broken fiber and at the free edge, were investigated. The effects of mechanical loading, thermal cooldown, and friction were investigated. Each configuration was analyzed for two loadings: longitudinal and normal to the fiber. A nonlinear finite element analysis was performed to model friction and slip at the F/M interface. A new procedure for fitting a square-root singularity to calculated stresses was developed to determine stress intensity factors (K sub I and K sub II) for a bimaterial interface crack. For the case of F/M cracking at a broken fiber with longitudinal loading, crack tip conditions were strongly influenced by interface friction. As a result, an F/M interface toughness test based on this case was not recommended because nonlinear data analysis methods would be required. For the free edge crack configuration, both mechanical and thermal loading caused crack opening, thereby avoiding frictional effects. A F/M interface toughness test based on this configuration would provide data for K(sub I)/K(sub II) ratios of about 0.7 and 1.6 for fiber and radial normal loading, respectively. However, thermal effects must be accounted for in the data analysis.

The mixed-mode bending (MMB) test uses a lever to simultaneously apply mode I and mode II loading to a split-beam specimen. A nonlinear, iterative analysis that accurately predicts the measured load-displacement response and the strain energy release rate, G, of an MMB test, has shown that the errors in G calculated using linear theory can be quite large. The nonlinear analysis was used as a guide in redesigning the MMB apparatus. With the redesigned apparatus, loads are applied through a roller attached to the lever and loaded just above the midplane of the test specimen. The redesigned MMB apparatus has geometric nonlinearity errors of less than 3 percent, even for materials substantially tougher than APC2 (AS4/PEEK). This apparatus was demonstrated by measuring the mixed-mode delamination fracture toughness of APC2. The data from the redesigned MMB apparatus were analyzed with a linear analysis which yielded results similar to those found with the original apparatus and the nonlinear analysis.

The Mixed Mode Bending (MMB) test uses a lever to simultaneously apply mode I and mode II loading to a split beam specimen. An iterative analysis that accounts for the geometric nonlinearity of the MMB test was developed. The analysis accurately predicted the measured load displacement response and the strain energy release rate, G, of an MMB test specimen made of APC2 (AS4/PEEK). The errors in G when calculated using linear theory were found to be as large as thirty percent in some cases. Because it would be inconvenient to use a nonlinear analysis to analyze MMB data, the MMB apparatus was redesigned to minimize the nonlinearity. The nonlinear analysis was used as a guide in redesigning the MMB apparatus. With the redesigned apparatus, loads were applied through a roller attached to the level and loaded just above the midplane of the test specimen. The redesigned apparatuus has geometric nonlinearity errors of less than three percent, even for materials substantially tougher than APC2. This apparatus was demonstrated by measuring the mixed mode delamination fracture toughness of APC2.

A dual-control test system measuring the strength of graphite/epoxy specimens under tension and compression for a wide range of bearing-bypass load ratios is presented. The bearing-bypass strengths of single-fastener specimens are measured for damage onset and ultimate failure and are plotted against one another. The results indicate that an initiated damage grows to failure in the same mode for most cases. However, if the damage is initiated in the compression-reacted-bearing mode, the specimens fail in the offset-compression mode. It is concluded that this type of transition can occur in multi-fastener joints under compression and complicate the strength predictions.

A combined experimental and analytical study was conducted to investigate and predict the failure modes of a graphite/epoxy laminate subjected to combined bearing and bypass loading. Tests were conducted in a test machine that allowed the bearing-bypass load ratio to be controlled while a single-fastener coupon was loaded to failure in either tension or compression. Onset and ultimate failure modes and strengths were determined for each test case. The damage-onset modes were studied in detail by sectioning and micrographing the damaged specimens. A two-dimensional, finite-element analysis was conducted to determine lamina strains around the bolt hole. Damage onset consisted of matrix cracks, delamination, and fiber failures. Stiffness loss appeared to be caused by fiber failures rather than by matrix cracking and delamination. An unusual offset-compression mode was observed for compressive bearing-bypass loading in which the specimen failed across its width along a line offset from the hole. The computed lamina strains in the fiber direction were used in a combined analytical and experimental approach to predict bearing-bypass diagrams for damage onset from a few simple tests.

A simple strength of materials analysis was developed for a double-cantilever beam (DCB) specimen to account for geometric nonlinearity effects due to large deflections and T-tabs. A new DCB data analysis procedure was developed to include the effects of these nonlinearities. The results of the analysis were evaluated by DCB tests performed for materials having a wide range of toughnesses. The materials used in the present study were T300/5208, IM7/8551-7, and AS4/PEEK. Based on the present analysis, for a typical deflection/crack length ratio of 0.3 (for AS4/PEEK), T-tabs and large deflections cause a 15 percent and 3 percent error, respectively, in the computer Mode 1 strain energy release rate. Design guidelines for DCB specimen thickness and T-tab height were also developed in order to keep errors due to these nonlinearities within 2 percent. Based on the test results, for both hinged and tabbed specimens, the effects of large deflection on the Mode 1 fracture toughness (G sub Ic) were almost negligible (less than 1 percent) in the case of T300/5208 and IM7/8551-7; however, AS4/PEEK showed a 2 to 3 percent effect. The effects of T-tabs G sub Ic were more significant for all the materials with T300/5208 showing a 5 percent error, IM7/8551-7 a 15 percent error, and, AS4/PEEK a 20 percent error.

Within a multi-fastener joint, fastener holes may be subjected to the combined effects of bearing loads and loads that bypass the hole to be reacted elsewhere in the joint. The analysis of a joint subjected to search combined bearing and bypass loads is complicated by the usual clearance between the hole and the fastener. A simple analysis method for such clearance-fit joints subjected to bearing-bypass loading has been developed in the present study. It uses an inverse formulation with a linear elastic finite-element analysis. Conditions along the bolt-hole contact arc are specified by displacement constraint equations. The present method is simple to apply and can be implemented with most general purpose finite-element programs since it does not use complicated iterative-incremental procedures. The method was used to study the effects of bearing-bypass loading on bolt-hole contact angles and local stresses. In this study, a rigid, frictionless bolt was used with a plate having the properties of a quasi-isotropic graphite/epoxy laminate. Results showed that the contact angle as well as the peak stresses around the hole and their locations were strongly influenced by the ratio of bearing and bypass loads. For single contact, tension and compression bearing-bypass loading had opposite effects on the contact angle. For some compressive bearing-bypass loads, the hole tended to close on the fastener leading to dual contact. It was shown that dual contact reduces the stress concentration at the fastener and would, therefore, increase joint strength in compression. The results illustrate the general importance of accounting for bolt-hole clearance and contact to accurately compute local bolt-hole stresses for combined bearings and bypass loading.

Attention is given to the stresses and the strain energy release rate along the delamination front in the present three-dimensional elastic analysis of a 24-ply, cocured double-cantilever beam specimen by means of 20-noded parabolic-isoparametric finite elements. At the free surface, the strain energy release rate was found to be substantially smaller than the plane strain value; this is suggested to be due to the free-surface effect that exists where the delamination meets the surface edge.

A combined experimental and analytical study was conducted to investigate and predict the failure modes of a graphite/epoxy laminate subjected to combined bearing and bypass loading. Tests were conducted in a test machine that allowed the bearing-bypass load ratio to be controlled while a single-fastener coupon was loaded to failure in either tension or compression. Onset and ultimate failure modes and strengths were determined for each test case. The damage-onset modes were studied in detail by sectioning and micrographing the damaged specimens. A two-dimensional, finite-element analysis was conducted to determine lamina strains around the bolt hole. Damage onset consisted of matrix cracks, delamination, and fiber failures. Stiffness loss appeared to be caused by fiber failures rather than by matrix cracking and delamination. An unusual offset-compression mode was observed for compressive bearing-bypass laoding in which the specimen failed across its width along a line offset from the hole. The computed lamina strains in the fiber direction were used in a combined analytical and experimental approach to predict bearing-bypass diagrams for damage onset from a few simple tests.

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 elastic stress analysis was conducted for a double cantilever beam (DCB) specimen using finite-element methods. The purpose of this study was to identify the important parameters that influence stresses ahead of the delamination front. The study focused on an aluminum DCB specimen, typical of adhesively-bonded joints, and on a graphite/epoxy specimen representing a cocured composite. Opening mode sigma sub y stresses ahead of the crack tip were calculated and compared with those for a monolithic reference specimen. Beyond the singularity-dominated region very near the crack tip, the sigma sub y distribution was elevated compared to the monolithic case. Both the adhesive thickness and the adherend transverse (thickness-direction) stiffness were found to influence the elevation of sigma sub y. In contrast, adherend thickness and longitudinal stiffness has very little effect on this stress distribution. Estimates for adhesive yielding beyond the aluminum DCB crack tip showed that both the area and height of the plastic zone increased to a peak value for increasing adhesive thicknesses. Results from this study would provide insight for comparing data from different DCB specimens and for designing new DCB specimens.

The effects of simultaneous bearing and bypass loading on a graphite-epoxy (T300/5208) laminate were investigated. Onset damage and ultimate strengths were determined for each test case. A finite element stress analysis was conducted for each test case. The computed local stresses were used with appropriate failure criteria to analyze the observed failure modes and strengths. An unexpected interaction of the effect of the bypass and bearing loads was found for the onset of compression reacted bearing damage. The interaction was caused by a decrease in the bolt-hole contact arc and a corresponding increase in the severity of the bearing loads. The amount of bolt-hole contact had a significant effect on local stresses and, thus, on the calculated damage-onset and ultimate strengths. An offset-compression failure mode was identified for laminate failure under compression bearing-bypass loading.

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.

Crack extension in elastic-plastic material involves energy dissipation through the creation of new crack surfaces and additional yielding around the crack front. An analytical procedure, using a two-dimensional elastic-plastic finite element method, was developed to calculate the energy dissipation components during a quasi-static crack extension. The fracture of an isotropic compact specimen was numerically simulated using the critical crack-tip-opening-displacement (CTOD) growth criterion. Two specimen sizes were analyzed for three values of critical CTOD. Results from the analyses showed that the total energy dissipation rate consisted of three components: the crack separation energy rate, the plastic energy dissipation rate, and the residual strain energy rate. All three energy dissipation components and the total energy dissipation rate initially increased with crack extension and finally reached constant values.

An elastic stress analysis was conducted for a double cantilever beam (DCB) specimen using finite-element methods. The purpose of this study was to identify the important parameters that influence stresses ahead of the delamination front. The study focused on an aluminum DCB specimen, typical of adhesively-bonded joints, and on a graphite/epoxy specimen representing a cocured composite. Opening mode sigma sub y stresses ahead of the crack tip were calculated and compared with those for a monolithic reference specimen. Beyond the singularity-dominated region very near the crack tip, the sigma sub y distribution was elevated compared to the monolithic case. Both the adhesive thickness and the adherend transverse (thickness-direction) stiffness were found to influence the elevation of sigma sub y. In contrast, adherend thickness and longitudinal stiffness has very little effect on this stress distribution. Estimates for adhesive yielding beyond the aluminum DCB crack tip showed that both the area and height of the plastic zone increased to a peak value for increasing adhesive thicknesses. Results from this study would provide insight for comparing data from different DCB specimens and for designing new DCB specimens.

A combined experimental and analytical study was conducted to determine the behavior of a graphite/epoxy laminate subjected to combined bearing and bypass loading. Single-fastener quasi-isotropic specimens were loaded at various bearing-bypass ratios until damage was produced at the fastener hole. Damage-onset strengths and damage modes were then analyzed using local hole-boundary stresses calculated by a finite-element analysis. The tension data showed the expected linear interaction for combined bearing and bypass loading with damage developing in the net-section tension mode. However, the compression bearing-bypass strengths showed an unexpected interaction involving the bearing mode. Compressive bypass loads reduced the bearing strength by decreasing the bolt-hole contact arc and thus increasing the severity of the bearing loads. The bearing stresses at the hole boundary were not accurately estimated by superposition of the stress components for separate bearing and bypass loading. However, superposition produced reasonably accurate estimates for tangential stresses especially near the specimen net-section.

Quasi-isotropic graphite-epoxy laminates (T300/5208) were tested under bolt bearing loads to study failure modes, strengths, and failure energy. Specimens had a range of configurations to produce failures by the three nominal failure modes: tension, shearout, and bearing. Radiographs were made after damage onset and after ultimate load to examine the failure modes. Also, the laminate stresses near the bolt hole calculated for each test specimen configuration, and then used with a failure criterion to analyze the test data. Failures involving extensive bearing damage were found to dissipate significantly more energy than tension dominated failures. The specimen configuration influenced the failure modes and therefore also influenced the failure energy. In the width-to-diameter ratio range of 4 to 5, which is typical of structural joints, a transition from the tension mode to the bearing mode was shown to cause a large increase in failure energy. The failure modes associated with ultimate strength were usually different from those associated with the damage onset. Typical damage sequences involved bearing damage onset at the hole boundary followed by tension damage progressing from the hole boundary.

The results of an experimental study to provide long-term durability data on detailed full-scale graphite/epoxy wing-skin joint designs under environmental exposure and cyclic loading associated with commercial transport aircraft are reported. The specimens consisted of a single-row bolt configuration fabricated from T300/5208 and a double-row bolt configuration fabricated from T300/5209. The unpainted specimens were exposed to the outdoor environment under a sustained tensile load, and at yearly intervals, they were subjected to fatigue loading. Experimental results showed a slight reduction in residual tensile strength for both graphite/epoxy joints under the exposure times and fatigue loadings reported. A 7.5-percent decrease in residual strength was observed for the T300/5208 single-row joint after 5 years exposure and two lifetimes of fatigue loading. A 5.3-percent decrease in residual strength was observed for the T300/5209 double-row joint after 7 years exposure and 2.8 lifetimes of fatigue loading. The 5208 epoxy material was more susceptible to degradation by ultraviolet radiation than the 5209 epoxy material.

In the present, simple analysis method for bolt hole clearance stresses in mechanically fastened joints, an inverse formulation is used with a finite element analysis. Conditions along the bolt hole contact arc are specified by displacement constraint equations. The method was applied to a single fastener clearance-fit joint with typical clearance values, for the cases of a rigid, frictionless bolt and a quasi-isotropic graphite/epoxy laminate. Results indicate that the contact arc, as well as the peak stresses around the hole (and their locations), were strongly influenced by the clearance. After a minor initial nonlinearity, peak stresses varied linearly with applied load. Clearance levels are found to have only a minor influence on overall joint stiffness.

The methods for predicting the strength of composite laminates with fastener holes are reviewed for the cases of unloaded as well as loaded holes. Three categories of methods are discussed: (1) the hole-boundary stress (HBS) methods which compare the peak stresses and the laminate strength; (2) the linear elastic fracture (LEFM) methods which assume cracks emanating from the hole and predict failure when the stress intensity factor for a crack equals the laminate fracture toughness; and (3) two different two-parameter (TP) methods: the average-stress (ASTP) method, which predicts failure when the average stress over a distance from the hole equals the laminate strength; and the point-stress (PSTP) Whitney-Nuismer (1974) approach, which predicts failure when the stress near the hole equals the material strength. The PSTP is relatively accurate and is by far the most widely used of all the prediction methods. However, for large ranges of variables and cases, the current strength predictions may not always be accurate.

Quasi-isotropic graphite/epoxy laminates (T300/5208) were tested under bolt bearing loads to study failure modes, strengths, and failure energy. Specimens had a range of configurations to produce failures by the three nominal failure modes: tension, shearout, and bearing. Radiographs were made after damage onset and after ultimate load to examine the failure modes. Also, the laminate stresses near the bolt hole calculated for each test specimen configuration, and then used with a failure criterion to analyze the test data. Failures involving extensive bearing damage were found to dissipate significantly more energy than tension dominated failures. The specimen configuration influenced the failure modes and therefore also influenced the failure energy. In the width-to-diameter ratio range of 4 to 5, which is typical of structural joints, a transition from the tension mode to the bearing mode was shown to cause a large increase in failure energy. The failure modes associated with ultimate strength were usually different from those associated with the damage onset. Typical damage sequences involved bearing damage onset at the hole boundary followed by tension damage progressing from the hole boundary.

A simple bolted joint was analyzed to calculate bolt clampup relaxation for a graphite/epoxy (T300/5208) laminate. A viscoelastic finite element analysis of a double-lap joint with a steel bolt was conducted. Clampup forces were calculated for various steady-state temperature-moisture conditions using a 20-year exposure duration. The finite element analysis predicted that clampup forces relax even for the room-temperature-dry condition. The relaxations were 8, 13, 20, and 30 percent for exposure durations of 1 day, 1 month, 1 year, and 20 years, respectively. As expected, higher temperatures and moisture levels each increased the relaxation rate. The combined viscoelastic effects of steady-state temperature and moisture appeared to be additive. From the finite-element analysis, a simple equation was used to calculate clampup forces for the same temperature-moisture conditions as used in the finite-element analysis. The two sets of calculated results agreed well. Previously announced in STAR as N82-19320

(Previously announced in STAR as N82-25327)

Stress distributions were calculated near a circular hole in laminates, using a three dimensional finite element analysis. These stress distributions were presented three ways: through the thickness at the hole boundary, along radial lines at the 0/90 and 90/0 interfaces, and around the hole at these interfaces. The interlaminar normal stress, and the shear stress, distributions had very steep gradients near the hole boundary, suggesting interlaminar stress singularities. The largest compressive stress occurred at about 60 deg from the load axis. A simple procedure was introduced to calculate interlaminar stresses near the hole boundary. It used stresses calculated by an exact two dimensional analysis of a laminate with a hole as input to a quasi three dimensional model. It produced stresses that agreed closely with those from the three dimensional finite element model.

A simple bolted joint was analyzed to calculate bolt clampup relaxation for a graphite/epoxy (T300/5208) laminate. A viscoelastic finite element analysis of a double-lap joint with a steel bolt was conducted. Clampup forces were calculated for various steady-state temperature-moisture conditions using a 20-year exposure duration. The finite element analysis predicted that clampup forces relax even for the room-temperature-dry condition. The relaxations were 8, 13, 20, and 30 percent for exposure durations of 1 day, 1 month, 1 year, and 20 years, respectively. As expected, higher temperatures and moisture levels each increased the relaxation rate. The combined viscoelastic effects of steady-state temperature and moisture appeared to be additive. From the finite-element analysis, a simple equation was developed for clampup force relaxation. This generalized equation was used to calculate clampup forces for the same temperature-moisture conditions as used in the finite-element analysis. The two sets of calculated results agreed well.

Stresses were calculated for finite size orthotropic laminates loaded by a frictionless steel pin in a circular hole of the same diameter. The calculations were based on finite element analyses for six laminates. Stress concentration factors, based on nominal bearing stress, were determined for wide ranges of the ratios of width to diameter, w/d and edge distance to diameter, e/d. An infinite laminate case was analyzed for each laminate. Orthotropy had a significant influence on the tensile stress concentration at the hole. For example, the stress concentration factors for the infinite laminate cases ranged from 0.82 to 2.16, compared with 0.87 for the quasi-isotropic laminate. The finite widths and edge distances strongly influenced the tensile stress concentration. For the practical range w/d or = 3, the peak tensile stresses were as much as 80% larger than the infinite laminate reference value. For e/d or = 3, these stresses were amplified by as much as 50%. In contrast, the finite width and edge distance had little effect on shear-out and bearing stress concentrations.

Graphite/epoxy laminates (T300/5208) were tested under bolt-bearing loads for a range of bolt clampup torques and for several test conditions involving water. High clampup torque improved both the static strength and fatigue limit by about 100 percent compared to a simple pin-bearing case, which had no through-the-thickness constraint. The static strength improvement was explained in terms of failure modes. Bolt clampup also influenced the hole elongation that developed before failure. For both the static and fatigue tests, the hole elongations were much larger for the low-torque-clampup case than for the pin-bearing case or for tests with moderate clampup torques. Tests in water degraded static bearing strength only slightly, but reduced the fatigue limit about 40 percent below that for tests conducted in air.

A quasi-three-dimensional finite-element analysis was used to analyze the edge-stress problem in four-ply, composite laminates. The seven laminates that were considered belong to the laminate family where the outer ply angle is between 0 and 90 deg. Systematic convergence studies were made to explore the existence of stress singularities near the free edge. The present analysis appears to confirm the existence of stress singularities at the intersection of the interface and the free edge. The power of the stress singularity was the same for all seven laminates considered.

Cyclic debonding rates were measured during constant-amplitude loading of specimens made of graphite/epoxy bonded to aluminum and S-glass/epoxy bonded to aluminum. Both room-temperature and elevated-temperature curing adhesives were used. Debonding was monitored with a photoelastic coating technique. The debonding rates were compared with three expressions for strain-energy release rate calculated in terms of the maximum stress, stress range, or a combination of the two. The debonding rates were influenced by both adherent thickness and the cyclic stress ratio. For a given value of maximum stress, lower stress ratios and thicker specimens produced faster debonding. Microscopic examination of the debonded surfaces showed different failure mechanisms both for identical adherends bonded with different adhesive and, indeed, even for different adherends bonded with identical adhesives. The expressions for strain-energy release rate correlated the data for different specimen thicknesses and stress ratios quite well for each material system, but the form of the best correlating expression varied among material systems. Empirical correlating expressions applicable to one material system may not be appropriate for another system.

A fatigue analysis, based on finite-element calculations and fatigue tests, was conducted for an aluminum-alloy sheet specimen with a steel interference-fit bolt. The stress analysis of the region near the bolt hole showed that the beneficial effect of an interference-fit bolt can be interpreted as the combined result of two effects: (1) load transfer through the bolt and (2) the compressive interference stresses in the sheet. Results of the fatigue tests show that progressively higher interference levels produced longer fatigue lives. The tests also show that a high level of interference prevents fretting at the bolt-sheet interface and that interferences larger than this level produced little additional improvement in fatigue life.

Strain energy release rates were used to correlate cyclic debonding between metal sheets and composite reinforcement. An expression for the strain energy release rate was derived and applied to fatigue test results for three material systems: graphite bonded to aluminum with both a room temperature and an elevated temperature curing adhesive; and, S-glass bonded to aluminum with an elevated temperature curing adhesive. For each material system, specimens of several thicknesses were tested with a range of fatigue loads. Cyclic debonding was monitored using a photoelastic technique. A close correlation was found between the observed debond rates and the calculated strain energy release rates for each material system.

A semianalytical method was developed for determining elastoplastic cyclic stresses and strains at notch roots in metallurgically bonded metal laminates. The method is based on the Neuber equation, which was used with an effective stress-strain curve for the laminate. It was applied to laminates containing a circular hole which were subjected to one cycle of reversed loading. The laminates consisted of two elasto-perfectly-plastic materials with different yield strengths and with either equal or different Young's moduli. A laminate of high-strength titanium alloy with alternate layers of commercially pure titanium was also analyzed. The accuracy of the method was evaluated by comparing the stresses and strains with those calculated from a finite-element analysis. The results estimated by the simple method based on the Neuber equation agreed closely with the results computed from the more elaborate finite-element analysis.

The applicability of acoustic emission technique to measure fatigue cracks in aluminum alloy specimens was investigated. There are several variables, such as the metallurgical and the physical treatment of the specimen, that can affect the level of acoustic activity of a fatigue specimen. It is therefore recommended that the acoustic emission technique be supplemented by other nondestructive evaluation methods to obtain quantitative data on crack growth.

The stresses and strains in a uniaxially loaded sheet with an unloaded interference-fit bolt were calculated by an elastoplastic finite-element analysis. The material properties represented a 7075-T6 aluminum alloy sheet and a steel bolt. The analysis considered the two ideal cases of no slip and no friction at the bolt-sheet interface for a single combination of bolt diameter, interference level, and cyclic loading. When the bolt was inserted, the sheet deformed plastically near the hole; the first tensile load cycle produced additional yielding, but subsequent cycles to the same level caused only elastic cyclic stresses. These stresses together with fatigue data for unnotched specimens were used to estimate crack initiation periods and initiation sites. The cases analyzed with interference-fit bolts were predicted to have crack initiation periods which were about 50 times that for a clearance-fit bolt. Crack initiation was predicted to occur on the transverse axis at a distance of about one radius from the hole.

A quasi three dimensional finite element analysis was used to analyze the edge stress problem in four-ply, composite laminates. Convergence studies were made to explore the existence of stress singularities near the free edge. The existence of stress singularities at the intersection of the interface and the free edge is confirmed.

An improved procedure for the three dimensional finite element analysis of an angle ply laminate with a circular hole is discussed. The procecure exploits polar symmetry and is the basis for a new finite element computer code. For the broad class of laminates that contain angle plies, this code requires only one half as much computing capacity as conventional codes.

Graphite/epoxy laminates were tested under bolt-bearing loads for a range of bolt clampup torques and for several test conditions involving water. High clampup torque improved both the static strength and fatigue limit by about 100 percent compared to a simple pin-bearing case, which had no through-the-thickness constraint. The static strength improvement was explained in terms of failure modes. Bolt clampup also influenced the hole elongation that developed before failure. For both the static and fatigue tests, the hole elongations were much larger for the low-torque clampup. Tests in water degraded static bearing strength only slightly, but reduced the fatigue limit about 40 percent below that for tests conducted in air.

Stresses were calculated for finite-width orthotropic laminates with a circular hole and remote uniaxial loading using a two-dimensional finite element analysis with both uniform stress and uniform displacement boundary conditions. Five different laminates were analyzed. Computed results are presented for selected combinations of hole diameter/sheet-width ratio d/w and length-to-width ratio L/w. For small L/w values, the stress-concentration factors K sub tn were significantly different for the uniform stress and uniform displacement boundary conditions. Typically, for the uniform stress condition, the K sub tn values were much larger than for the infinite strip reference condition; however, for the uniform displacement condition, they were only slightly smaller than for this reference. The results for long strips are also presented as width correction factors. For d/w less or = 0.33, these width correction factors are nearly equal for all five laminates.

Fatigue-crack initiation and early stages of growth in notched 2024-T3 aluminum alloy sheet specimens were monitored during constant-amplitude fatigue tests conducted for a range of stress ratios. A video tape system was developed to detect fatigue cracks 0.1 mm in length on the sheet surface and to monitor their growth. Fatigue cracks initiated either at the notch corner or on the notch surface. Most of the first fatigue cracks started as notch-corner cracks, and the crack-initiation period was defined as the number of cycles for which the crack was 0.1 mm in length on the sheet surface. For each of the stress ratios used, the results showed that at the low stress levels, the crack-initiation period constituted at least 90 percent of the fatigue lives. However, for the negative stress ratios at the high stress levels, cracks started as early as at about 40 percent of the fatigue lives.

The effects of loading sequence on crack-initiation period were investigated for notched aluminum-alloy specimens under high-low two-step loading with special emphasis on local cyclic stresses and strains at the notch root. Local stress and strain were determined by a procedure based on an equation proposed by Neuber which relates elastoplastic stress and strain at a notch. Local stress and strain were also measured experimentally to verify the Neuber equation. The effects of initial high load on the crack-initiation periods were demonstrated with notched specimens and were simulated in unnotched specimens fatigue tested with local stress sequences. An analysis of the results indicated that sequence effects were not caused solely by local residual stresses, as is usually assumed; the existence of a damaging effect, resulting from the high local strain cycles, was demonstrated. The sequence effects observed with notched specimens were interpreted as the combined result of residual stresses and high local strain cycles.

Notched metal crack initiation, determining high cyclic loads effects by prior local plastic behavior at stress concentration

Elastoplastic stress-strain behavior at notch roots in sheet specimens under constant amplitude loading

Fatigue cracking at stress concentration under cyclic loading

Cyclic plastic stresses at notch roots in aluminum alloy sheet specimens under constant- amplitude repeated tension and reversed loading

Cyclic plastic stresses in notched aluminum alloy sheets under constant-amplitude repeated tension and reversed loading - fatigue life data

Cyclic plastic stresses at notch roots of aluminum alloy sheets under repeated tension and reversed loading

Strain energy release rates were used to correlate cyclic debonding between metal sheets and composite reinforcement. An expression for the strain energy release rate was derived and applied to fatigue test results for three material systems: graphite bonded to aluminum with both a room temperature and an elevated temperature curing adhesive, and S-glass bonded to aluminum with an elevated temperature curing adhesive. For each material system, several thicknesses were tested with a range of fatigue loads. Cyclic debonding was monitored using a photoelastic technique. A close correlation was found between the observed debond rates and the calculated strain energy release rates for each material system.