An approximate solution of the stress field in unidirectional composites was obtained for tensile and thermal loading conditions based on the Airy stress function. The hexagonal packing of unidirectional fibers was assumed as an ideal arrangement. This solution is applicable to the transverseisotropic fiber/isotropic matrix composite system. Explicit solutions for the local stress and average strain of composite were presented together with the instruction for the numerical process. The average strain became equal to the thermal expansion coefficient of composites when no tensile load was applied. The present results were compared with the previous exact solutions obtained for glass/polymer composites and were found to be satisfactorily accurate. Finally, the stress along the fibermatrix interface and the average axial stress in both matrix and fiber were presented as a function of fiber volume fraction for glass/epoxy and carbon/epoxy composites.
In recent years high-modulus fibers such as boron and graphite have been widely used in many aerospace applications in the forms of hybrid composites along with glass or Kevlar fibers. In this paper presented is an analytical method useful for designing optimum hybrid plates subject to a given compressive load. It is assumed that the hybrid plate is symmetric and consists of two surface layers and a core, and that the ply material for the surface has high performance and is expensive, while the ply material for the core has low performance and is inexpensive. The object of the design problem is to minimize the cost of the hybrid plates, which corresponds to the maximization of the core thickness ratio. The proposed method is based on the flexural lamination parameter diagram of hybrid composites. The feasible design region where the constraint on the buckling load is satisfied becomes smaller as the core thickness ratio increases, and the region finally reduces to a point. The optimum solution can be obtained under this condition. The optimum fiber orientation in both the surface and the core and the optimum core thickness ratio can be obtained easily by using the proposed method. It is found that neither the optimum fiber orientation of the surface nor that of the core is affected by the elastic constants or the fiber orientation of the other material. This means that the optimization of hybrid composites can be done by the independent optimizations of each component materials.
The authors proposed a new failure process simulation model considering interfacial debonding between fibers and matrix as well as random fiber breaking based on a shear-lag theory, in which not only fiber elements but also matrix elements are included. The failure model was applied to characterize not only the static tensile failure process but also the statistical nature of hybrid effect of several types of unidirectional hybrid FRP in the previous reports. The present paper proposes a new dynamic failure process simulation model in which an additional time variable is also incorporated by taking the mass of fiber and matrix elements into account. An exact time-dependent stress redistribution process in a new composite failure model is evaluated by means of a finite difference scheme based on an increment method. The numerical results are shown and discussed for two types of unidirectional carbon (C)/glass (G) hybrid FRP models. As the results, it is shown that a time-dependent stress concentration factor is exactly evaluated by the simulation which is in good accordance with Hedgepeth's solution and that the stress in a carbon fiber is relieved to some extent due to the presence of glass fiber in hybrid C/G FRP. It is also shown that a clear difference exists between the static and dynamic tensile failure patterns for both types of hybrid FRP. It is furthermore shown that the dynamic stress-strain behavior and the acoustic emission are successfully simulated by taking the stress drop and the released strain energy due to fiber and matrix failures into consideration, which gives a more realistic behavior than the static simulation.
Near-threshold growth of delamination fatigue cracks was investigated in unidirectional laminates made from Ciba Geigy prepregs 914C both under mode I and mixed mode of I and II. Double cantilever beam specimens (DCB) were used for mode I tests. Cracked lap shear specimens (CLS) were used for mixed mode tests. The stress ratio, R, of the minimum to maximum stress was kept constant in the tests, and the constant value was between 0.1 and 0.5. Cyclic load was shed automatically by keeping maximum displacement constant. In the region of rates above about 1-5×10-10m/cycle, the crack growth rate is expressed as a power function of fracture mechanics parameters. Below this region, there exists the growth threshold for both mode I and mixed mode. When compared at the same value of fracture mechanics parameters, the crack growth rate for mixed mode is higher than that for mode I. The values of crack growth threshold under mixed mode were smaller than those under mode I. The influence of the stress ratio becomes smaller when the crack growth rate is correlated to the energy release rate range than when the crack growth rate is correlated to the stress intensity range or the maximum energy release rate. The relative contributions of the maximum stress and of the cyclic stress on the crack growth rate are discussed on the bases of fractographic observation and mechanism consideration. The contribution of the maximum stress was high.
The acoustic emission test has been carried out on the Class A-SMC material which has an excellent surface property as a material for automobiles. The AE amplitude increased in the order of the resin fracture, the debonding between resin and calcium carbonate, the debonding between resin and fibers, the fiber pulling out, and the fiber breaking. However, as the AE amplitude depends on the position of AE sensors and the amplification degree, the AE amplitude distribution gives only a relative estimation of fracture modes of the constituent materials in composites. From the results of frequency analysis and the consideration about the AE amplitude distribution, it is found that the frequency ranges corresponding to the resin fracture, the debonding between resin and calcium carbonate, the debonding between resin and fibers and also the fiber pulling out, and the fiber breaking are 80-180, 240-450, 180-250, and 250-400kHz, respectively.
An experimental investigation was conducted to obtain the residual compressive strength of laminate specimens of carbon fiber reinforced epoxy (CFRP), which were damaged by penetrative impact from the side under the axial static load, Pt. After impact, the specimens were inspected visually and ultrasonically. The results obtained are summarized as follows: (1) There was no remarkable change in apparent stiffness after impact loading. (2) The specimen damaged under axial compressive load had a larger interior delamination region than the one damage under tensile load, while the back surface of the former specimen showed a smaller visible region with surface cracks than that of the latter. (3) Both of the damage regions of the back surface and the interior of the specimen became larger with increasing impact velocity. (4) For the case of high compressive load, the residual compressive strength after impact was lower than those for other cases.
In order to know machining characteristics of GFRP, special unidirectional glass fiber reinforced plastic specimens were cut orthogonally under a dry condition. The effects of the direction of yarn with the cut surface and the radius of cutting edge of the tool upon the aspect of failure of the yarn and the cutting force were investigated. The fracture mode of glass yarn was found to be affected by the radius of cutting edge of the tool. The depth of the failed part of yarn from the cut surface became maximum when the direction of yarn was at a certain angle with the cutting direction. The cutting force for the yarn was strongly affected by the direction of yarn with reference to the cutting direction.
A phenomenological theory of superplastic deformation due to isothermal martensitic transformation, constituted of the mechanical constitutive equation and the transformation kinetics, is presented. The transformation kinetics is obtained by modifying the one for the diffusion-type transformation proposed by the present authors. The T-T-T diagrams constructed by the transformation kinetics explain well the qualitative characteristics of the experimental results in Fe-Ni-Mn, Fe-Ni-Cr and Fe-Ni alloys. The proposed theory is shown to be able to describe well the transformation superplastic behavior in Fe-Ni alloy, the stress and temperature dependences of the overall deformationtime behavior, the final accumulated strain and the incubation time of deformation.
The effects of temperature and specimen size on the plastic deformation by crazing were investigated under constant rate tension at different temperatures in kerosene using poly (vinyl chloride) plates with different thickness. An increase in temperature facilitated the initiation and growth of crazes, and as a result reduced the craze yield point. As the specimen thickness increased, the craze yield point rose and the shape of deformation curve became similar to that of shear yielding. The temperature dependence of craze yielding was favourably explained by the theory which was previously developed on the conception that the plastic strain produced by crazes is given by the total volume of crazes per unit volume of the polymer. In order to account for the size effect on craze yielding, a simple model consisting of two crazed surface layers and non-crazed interior was proposed. By applying the above theory on craze yielding to describe the deformation of the surface layers, the calculated results accorded with the experimental results mentioned above.
Low cycle fatigue tests with constant plastic strain range Δεp were carried out on 316 stainless steel at room and elevated temperatures (650°C) in order to investigate a relationship between macroscopic cyclic deformation and microscopic dislocation structures. Macroscopically, cyclic work hardening behaviors under constant temperature and two-step temperature change tests (a room to high temperature test and a high to room temperature test), and cyclic stress-strain curves by means of three different methods (a companion specimens, an increasing step and a multiple step methods) were obtained. Microscopically, the quantitative variation of dislocation structures was measured by transmission electron microscopy observation. The results obtained were summarized as follows, (1) The cyclic work hardening properties under two-step temperature change tests were explained from the dislocation structures at the time of temperature change and the dislocation rearrangement after the change during the fatigue process. (2) The cyclic stress-strain curves by the three different methods at 650°C were almost the same in the range of Δεp>0.3%. (3) The area fraction occupied by cell Sc/S0 increased with the number of cycles N and the Δεp, and the increasing tendency during the fatigue process at 650°C was very different from that at room temperature. The Sc/S0 at 650°C increased at a very early stage of fatigue and the value was remarkably larger than that at room temperature. (4) The relations between the normalized stress amplitude (Δσ/2G) and the average cell size lc, and between the (Δσ/2G) and the total dislocation density ρt were obtained as follows, (Δσ/2G)∝lc-0.5, (Δσ/2G)∝ρt0.5 where G is a shear modulus.
The threshold stress intensity factor, KISCC for stress-corrosion cracking or KIH for hydrogen assisted cracking, is generally determined by one of the cantilever constant-load test, constant-displacement test and rising-load test. In the present paper, problems encountered with these tests, especially on the accuracy, test duration, and/or measurement of crack growth rate, were discussed first. Then the holding-load and fractography test was proposed as a way to improve the existing procedures. An outline of the new testing procedure is as follows: Several precracked specimens exposed in test environments or precharged with hydrogen are loaded constantly at different KI values for a relatively short predetermined test duration. The specimens are then unloaded and separated apart by cleavage fracture or fatigue loading. The amount of subcritical crack growth during the test, ΔaH, is measured by an electron microscopic observation of the fracture surface. The value of KISCC or KIH at the onset of the crack growth and the growth rate are determined from the relation of ΔaH and its rate per test duration to KI value. The test procedure was used to predict the values of KIH and growth rate for hydrogen assisted cracking of 1/2 Mo steel and 2-1/4Cr-1Mo steel. The results indicate that the usual test procedure estimates higher KIH than the proposed one. Thus, it is considered that the reexamination of the existing KISCC and KIH data is required.
Although many attempts have been made to determine the accurate value of Impact fracture toughness KId, no standard method such as the KIc test method has been found. It is desired that a simple method will be established. In the present research, a KId test using a drop weight and 3-point bend specimens was conducted on three metals, i.e., 2124-T6 aluminum alloy, 2124-T6 aluminum alloy with 25%SiC whisker (SiC 25%/2124-T6), and nuclear pressure vessel steel A508S. For the SiC 25%/2124-T6 specimen, which fractured in a brittle manner, the electric potential method was usefully employed to detect crack initiation, and the value of KId was easily determined. The 2124-T6 specimen fractured with a little plastic deformation (but in a “small scale yielding” condition). It was found possible to detect crack initiation by the electric potential method, though it was not so easy as in the SiC 25%/2124-T6 specimen. The value of KId was determined by a similar procedure to that for the SiC 25%/2124-T6 specimen. For the A508S specimen, which fractured after a considerable plastic deformation, it was impossible to detect crack initiation by the electric potential method. By dropping the weight from various heights on several specimens and then observing the fracture surfaces with an electronmicroscope, the critical height for which a crack just initiated was determined. The value of KId was obtained from the maximum value of the J-integral calculated for the critical height.
Low cycle fatigue tests of the dissimilar metal electron beam welded joints between A387, Gr. 22 steel and SUS 405 steel (A387-SUS 405), and between A387, Gr. 22 steel and SUS 316 steel (A387-SUS 316) were carried out under the strain controlled cycling over the welded joints at a temperature of 839K. It was shown that the fatigue lives of the dissimilar metal welded joints were significantly shorter than those of base metals. In order to discuss the fatigue fracture behavior, the distribution of local strain and that of hardness were investigated. In A387-SUS 405, the strain concentration occurred on SUS 405 side during all strain cycling. While in A387-SUS 316, the uniform strain distribution was produced at the early stage of strain cycling, and thereafter the strain concentration occurred on A387 side. The change of strain distribution with strain cycling qualitatively corresponded to that of hardness. From the foregoing results, it was concluded that the shorter fatigue lives of the dissimilar metal welded joints resulted from the strain concentration which occurred on the base metal with lower deformation resistance.
A predicting method of low cycle fatigue life in the dissimilar metal electron beam welded joint (DMEBWJ) at an elevated tempreature was proposed by representing the cyclic softening and hardening behaviors of base metals. The DMEBWJs between A387, Gr. 22 steel and SUS 405 steel, and between A387, Gr. 22 steel and SUS 316 steel were used. The estimated variations of strain distribution in the DMEBWJs with the strain cycling were well consistent with the experimental results. The predicted fatigue lives of DMEBWJs were also in good agreement with the experimental results, except for the case that the heat affected zone with low hardness was formed in the DMEBWJ.
The flow rate of flux for submerged arc welding has been investigated by using SiO2-CaO-MgO fused flux and 60° conical funnels. The circular orifice of the funnels was 4 to 12mm in diameter and the mean grain size of the flux grains was varied from 1.880 to 0.089mm. The test results indicate that the flow rate becomes almost maximum at the mean grain size range of 1.115 to 0.126mm independent of orifice diameter. An attempt to derive a new empirical equation which would represent the flow rate as a function of grain size DP and orifice diameter DO has been made. As a consequence of the study, it was possible to derive an improved equation which has better regression characteristic than other equations proposed previously.
The fracture toughness of austenitic stainless steels decreases so drastically after 5-10 years long time service operation at the temperature of 550-850°C due to grain boundary embrittlement. In this study, therefore, an appropriate NDE procedure of the material characterization of SUS 316H and SUS 321H used as boiler superheater tubes was developed by means of Small Punch test and EPR method. A significant transition behavior from ductile to brittle intergranular fracture was observed in SP energy for the degraded materials, whereas no DBTT for the nondegraded ones. Moreover EPR method was applied to detect the grain boundary embrittlement due to sensitization or σ phase formation in austenite stainless steel. The relation between the material degradation and EPR results was also obtained. Based upon these experimental calibrations of material degradation and NDE parameters, a realistic periodic NDE procedure for austenitic stainless boiler tubes has been proposed.