An LE-7 rocket engine main injector is subjected to both internal pressure stress and high thermal stress due to large temperature difference between the fuel manifold (600°C) and LOX dome (-180°C). Therefore thermal ratchetting was considered to be one of the problems. A failure-mode simulation test was conducted using the structure model of an INCONEL718 welded joint, which ended with cracking of the weldment after 70 cycles of axial mechanical load and thermal load. Measured accumulated strain was smaller than that estimated from Bree's ratchetting diagram. Elastic-plastic thermal analysis showed a fairly good agreement between estimated life and test result, and this failure mode was due to mainly creep damage. It is confirmed that the thermal ratchetting effect on this failure mode is negligible, and the analytical method of creep-fatigue life evaluation is adequate for the prediction of structure life.

View full abstract

Nondestructive creep and fatigue life evaluation technology was investigated for INCONEL718 in the LE-7 rocket engine. Metallurgical microstructural change is not considered to be a criterion for life prediction because it is very small in high-stress and short-term creep and fatigue damage processes such as in the LE-7 engine. Crack length observation on the metal surface is the most effective method to estimate damage rate. Correlation of maximum crack length and life consumption rate was shown for the parent metal and welded joint based on the interrupted creep and fatigue test specimen. Furthermore, the analytical life prediction method by means of structural strength analysis and reference material data was checked using notched specimen creep and fatigue test data.

View full abstract

Welded joint structural failure has been observed in the LE-7 engine main injector, which is subjected to high thermal stress during an engine start and stop sequence. Failure analysis showed that these failures were caused by extremely-low-cycle fatigue due to a low number of hot-fire tests in a very high-strain condition. It was revealed that the welded joint has a considerably short low-cycle fatigue life and low fracture toughness, contrary to expectation. Based on several kinds of material tests, countermeasures such as design change of the welded structure, modification of the welding process, strengthening during non-destructive inspection and application of solution heat treatment were taken to improve welded joint strength characteristics. Using the modified main injector, a complete qualification test of LE-7 engine was performed without occurrence of failure and the first H-II rocket was launched successfully.

View full abstract

The LE-7 rocket engine main injector has undergone a few occurrences of welded joint structural failure because it is subjected to high thermal stress during the engine start and stop sequence. The low-cycle fatigue test and high-strain fatigue test which were conducted as part of failure analysis revealed that the welded joint has a considerably shorter fatigue life than expected. To improve the ductility and toughness of the welded joint, high-temperature-solution heat treatment was applied after welding. The low-cycle fatigue life of a welded joint subjected to this solution heat treatment became three times longer than that without it. A design criterion was established for low-cycle fatigue life estimation, and the life requirement was found to be satisfied.

View full abstract

A thermal shock fatigue test was conducted as a part of low cycle fatigue life evaluation of the INCONEL718 welded joint in the LE-7 rocket engine main injector. It was revealed that specimen thermal shock fatigue test data were in agreement with standard low cycle fatigue test data. Thus thermal shock fatigue life, which is determined predominantly by thermal bending stress, can be analyzed and evaluated appropriately by the same method which is used for low cycle fatigue life evaluation under the control of membrane stress. Application of high-temperature solution heat treatment after welding improved low cycle fatigue life of the welded joint as well as ductility and toughness. Another thermal shock fatigue test of a structural model, which was exposed to thermal stress as large as the actual main injector, showed longer fatigue life than required.

View full abstract

Rotating bending fatigue tests were carried out on the notched specimens of cast stainless steels. We prepared fine and coarse grain size materials in order to investigate the effect of grain size on the notched fatigue strength. The main results are summarized as follows : (1) The notch radii of the branch point ρ₀ are nearly 0.3 mm, independent of grain size. (2) The master curves of K_tσ_w2 vs 1/ρ of the two materials are nearly equal, but the master curve of K_tσ_w1 vs 1/ρ of fine-grain-size material is somewhat higher than that of coarse-grain-size material. (3) When the notch radius is larger than about 2.5 mm, the notched fatigue strength of cast stainless steels decreases by the effect of a microshrinkage cavity.

View full abstract

The localized strain around a crack tip and the crack tip opening displacement were measured under displacement- and load-controlled cyclic loadings with displacement and load ratios of -1.0 which involved single overload and single overdisplacement applications, respectively. Special attention was given to the two behaviors during one cyclic loading. The crack remains open at the compressive tip of a load-displacement hysteresis loop under displacement-controlled loading as well as under load -controlled one. The strain around the crack tip and the crack tip opening displacement largely generated by single overdisplacement application are not fully restored in the following compressive load stroke to their original amounts at the compressive tip of the controlled displacement wave just before the overdisplacement application. A small amount of tensile strain and of the crack opening displacement are retained in the vicinity of the crack tip at the compressive tip of the hysteresis loop immediately after a single overdisplacement application. A single overdisplacement application causes the delay of crack growth due to the two behaviors in the displacement-controlled test, which, however, is quite small compared with that in the load-controlled test involving a single overload.

View full abstract

In this study, the small fatigue crack initiation and propagation behavior of A6061-T6 metal matrix composites reinforced with 20 vol.% SiC particles was investigated under cyclic axial loading at the stress ratio R of 0.1. The results obtained are summarized as follows. (1) The microscopy indicates that the fatigue cracks initiate preferentially at the sharp end of a particle in the matrix aluminum material, because the stress near the end of a particle appears to be much higher than the average stress. (2) The propagation behavior of small cracks (crack length 2c<150μm) was quite irregular. The crack propagation ceased at the end of a particle at some point of the loading cycle. This may be interpreted in terms of the connection of the main crack with the small second crack. (3) At Δσ=280 MPa, the relationship between dc/dN and ΔK_eff for small cracks (crack length 2c>150μm) is nearly identical to that for long cracks. It is considered that, due to crack closure, dc/dN of a small crack is greater than that of a long crack for the same ΔK.

View full abstract

The fracture mechanisms of FRPC (fiber reinforced polycarbonate) composites which are molded using various screw designs and resin feed methods with experimental values for the number of screw rotations are studied by the acoustic emission technique. The optimal screw design and resin feed method are found from the loads P_b and P_c at which changes in the fracture mechanisms are observed by the AE energy gradient method. The specimens which are molded using the screw of lower compression ratio and the direct resin feed show the highest values of loads P_b and P_c, and hence the experimental values for the number of screw rotations are important factors. It is found that the effect of these factors on the tensile strength and loads P_b and P_c can be explained by the shearing stress of melted resin at the metering zone of screw.

View full abstract

In order to simulate three-dimensional dynamic crack propagation, a three-dimensional moving finite-element method was developed. A new expression was also derived for the dynamic J integral (J') in terms of the equivalent domain integral (EDI) method. The moving finite-element procedure together with the dynamic J integral evaluation procedure made it possible to evaluate the distribution of the dynamic energy release rate along the propagating crack front at each time step. Furthermore, a new formula for converting the dynamic J integral to the stress intensity factor along the crack front was derived by introducing a three-dimensionality parameter. Numerical simulation was carried out for several sample problems.

View full abstract

An elastic-plastic finite-element analysis of fracture with debonding in a ceramic (silicon nitride)/metal (steel) joint was performed. The following results were obtained. (1) The experimental results that debonding occurs first, followed by fracture in ceramics, were explained by taking account of the residual stress due to bonding. (2) The experimental result that an interlayer (a copper sheet) reduces the fracture in ceramics after debonding, was explained by the plastic deformation of the interlayer. (3) It was suggested that stable or unstable debonding may occur depending on the compliance of the brazing filler (e.g., Ti-Ag-Cu). (4) It was shown that the wake of plastic deformation due to debonding has little effect on the stress distribution.

View full abstract

This paper presents a new inverse analysis method using an influence function method in a finite element method (FEM) on a small memory capacity machine such as a microcomputer and a personal computer. First, we discuss the problems of the inverse analysis using the FEM. Especially, the size of unknown vectors is larger than that of known ones in the FEM equation including unknown boundary condition. This FEM is usually unsolvable. Second, an inverse compliance matrix is introduced into the FEM. Then, we formulate a solvable FEM for inverse problems like unknown boundary condition. Although a compliance matrix is obtained using the influence function method in the ordinary FEM on the small machine, the compliance matrix must be divided into smaller submatrices for the small machine. We calculate the inverse compliance submatrices. From the above procedures, we establish a FEM for inverse analysis on the small machine. Finally, it is verified that this FEM for inverse analysis is exact from the numerical solution of a three-dimensional elastic finite element model on the small machine.

View full abstract

The dependence of the ultrasonic velocities on the creep damage can be used for its nondestructive evaluation. This study discussed the validity of the oblate ellipsoidal void model to predict their relationship. In the experiment, the ultrasonic velocities and the density change in creep damaged coppers were measured. The prediction by the spherical void model gives good agreement with the experimental results in the initial stage of the damage process. As the damage progresses, this model is no longer valid because of the anisotropy in ultrasonic velocities. The oblate ellipsoidal void model is thus used to explain the anisotropy. However, this model shows an opposite tendency in some cases of velocity rate changes with decrease in density. The density decrease estimated from the ultrasonic velocities is about two times larger than the direct measurement. Metallographical observations show that the clustering of voids on the grain boundaries and their orientation distribution should be considered.

View full abstract

The classical plate theory neglects transverse shear strains, which brings about considerable errors in deflections, strains and stresses for the analysis of laminated plates. Now, a simple higher-order theory proposed by Reddy, which is one of the higher-order shear deformation theories of laminated plates, is widely used. In this paper, a uniformly loaded thick circular laminated plate with a clamped edge is analyzed using 3-dimensional finite-deformation theory, and from this result, the applicability of the higher-order shear deformation theory to both infinitesimal deformations and large deformations is examined. As a numerical method, the 3-dimensional finite element method which is formulated using the incremental theory based on the convected coordinates is used.

View full abstract

A constitutive law of plastic deformation has been proposed by Goya and Ito, and can incorporate the stress increment directional dependence of plastic strain increment ε^p. The law is developed through the introduction of two transition parameters μ(α) and β(α), which denote the magnitude and the direction angle of the plastic strain increment, respectively, and are defined to be functions of the direction angle of the stress increment. In this report, the law is implemented in an F. E. M. code for deformation analysis. A plane strain tension problem is analyzed for investigation of the effect of these parameters on the occurrence of the shear band. The results show that the clear shear does not appear when the calculation is performed using B_max=π/6, which corresponds to the degree of the directional dependence for the corners of the Tresca yield locus.

View full abstract

The criteria of localized necking were proposed by Hill and Storen and Rice, and they are very famous for their ability in predicting sheet formability. Though Hill's analysis cannot be applied to sheet formability under biaxial stretching, the one by Storen and Rice gives the impression that it can be applied to biaxial stretching. However, detailed examination reveals that both analyses are identical in their assumptions of equilibrium and deformation rate. Therefore, the analysis by Storen and Rice remains within the sphere of Hill's analysis, and it also cannot be applied fundamentally to biaxial stretching. Localized necking is a phenomenon where velocity gradient or strain rate becomes discontinuous on a necking plane, and force equilibrium must be maintained on both sides of the plane. However, the published analyses were insufficient with respect to the equilibrium and mode of the necking plane. Rigorous treatment for them is demonstrated. Diffuse necking proposed by Swift is represented by the criterion that in-plane loads become simultaneously stationary under plane stress. However, the loads obtained based on the criterion, in general, are not stationary. Then, the origin of the unreasonable results is discussed, and an alternative criterion is proposed which can yield a rigid solution in addition to plastic solutions. All the criteria in diffuse and localized necking are involved in the universal criteria obtained from the permissible condition for multiple solutions.

View full abstract

In this study, lateral compression tests and tensile tests for Mullite circular tubes were carried out at static rates. Theoretical and numerical analyses for the deformation of ceramic tubes in lateral compression were also performed. In considering the deformation of ceramic circular tubes in lateral compression, elastic contact with flat plate and tube surface was taken into account as well as the bending of the curved wall of the tubes. Young's modulus of ceramic tubes obtained from lateral compression tests agreed well with that obtained from tensile tests. A simple equation which enables easy calculation of the fracture stress in lateral compression tests was obtained using FEM analysis. From the comparison with the two-dimensional analytical solution, it was found that the equation was applicable to lateral compression tests of relatively thick-walled tubes up to h/Do=1/3 (Do : outer diameter, h : wall thickness of circular tubes). It was also found that the fracture stress in lateral compression was about 1.5 times greater than the tensile strength of ceramic tubes.

View full abstract

A solution is presented to an axisymmetric contact problem of a circular rigid punch penetrating an elastic half-space in which a spherical inclusion coaxial with the punch is embedded. By expressing the contact stress under the punch and the interface displacements and stresses between the half-space and the inclusion as appropriate series, the doubly mixed boundary-value problem is reduced to the solution of an infinite system of simultaneous equations. Numerical results are obtained for stress distributions under the punch and around the inclusion as well as for the force transmission underneath the punch due to the presence of the inclusion. Comparisons with the results in the absence of the inclusion are made in order to illustrate the effect of the presence of the inclusion on the stress field.

View full abstract

The tensile fracturing process of a nanoscale polycrystalline pure iron, which was newly developed in a previous paper for MD simulation analysis, was examined. The simulation analysis was performed in the temperature range from 100 K to 700 K. Temperature dependences of tensile strength σ_B and Young's modulus are first studied, and σ_B is found to decrease with increasing temperature. Regarding the fracture mechanism, defects are generated on the grain boundaries as the strain increases at room or low temperature, until the defects reach a cross point of three grain boundaries. Moreover, trangranular defects are generated as the strain increases. At high temperature, the fracture is caused by grain boundary diffusion and large-scale slip formation near the grain boundaries, which finally leads to the defect formation. In other words, increasing temperature results in the fracture mode transition from a cleavage fracture to a ductile one.

View full abstract

A new method for generating nanoscale polycrystalline materials for molecular dynamics (MD) simulation analysis was proposed using pure iron. MD relaxation calculation was first performed on a cluster of single crystals, consisting of ninety atoms and having random orientation, and then the nanoscale polycrystal was obtained. In order to confirm the validity of the polycrystal generated, thermodynamic properties such as coefficient of thermal expansion and specific density were examined in the temperature range from 100 K to 3000 K. The results obtained showed fairly good agreement with those reported by other investigators.

View full abstract

Volume resistivity of carbon-fiber-filled silicone oil was measured under shear flow. A rheometer with parallel-plate-type sample cell was employed to measure the volume resistively, where both upper and lower parallel plates were used as electrodes. The effects of shear rate, volume fraction of fibers, and distance between electrodes on volume resistivity were investigated. The critical values for a sudden increase in volume resistivity were found for these three factors. These experimental results were interpreted on the basis of a quantitative model for the formation and destruction of an electric circuit by fibers or the interaction and orientation of fibers. A new method of measuring volume resistivity was found to be effective for inspection of distribution and orientation of fibers in the matrix for conductive-fiber-filled insulating matrix composite materials.

View full abstract

In a previous study, micromechanical analysis was performed on a unidirectional fiber-reinforced composite showing interfacial debonding between a fiber and the matrix. The criteria for both the matrix crack and the debonding crack propagation were derived, and it was found from the numerical calculations that the unstable propagation of a matrix crack was arrested at some critical size of the crack and then an interfacial debonding occurred. In the present study, based on the previous result, the overall non-linear stress-strain behavior, i.e., stiffness reduction, of a composite whose constituents are all brittle materials is analyzed, assuming such a complicated fracture path that the previously starting interfacial debonding mentioned above ceases to propagate under certain conditions and then propagation and suspension of a matrix Crack recur, and the influence of material parameters such as the volume fraction of a fiber and the surface energy of the interface on the stiffness reduction of a composite is also investigated.

View full abstract

An electromagnetic field of an eddy current thickness sensor was analyzed for the purpose of improving measurement accuracy. Maxwell equations were discretized with the boundary element method. In order to facilitate the calculation, a quasi-stationary approximation and an axisymmetric one were employed. A numerical integration formula with high accuracy was applied to evaluate a singular integral. A special element was developed for effective analysis of a thin film. It is found from numerical results that the radial component of magnetic flux density produced by the eddy current is greater than the axial component in nickel-plated steel. Using this property, a new thickness sensor for a metal film on a metal substrate was realized.

View full abstract

An attempt has been made to measure precisely the leaky surface wave velocity, V_R, within a very localized region using a scanning acoustic microscope (SAM). To achieve the relative precision of V_R, e.g., 10^-4, required for acoustoelastic stress measurement, two major modification of the conventional SAM were made. The defocus distance was measured with a high-resolution displacement sensor. For each measurement, the distance as well as the output of V(z) were averaged many times. The very low as well as very high frequency components in V(z) curves were eliminated by the digital filtering technique, and temperature of water was controlled within 0.02 K. Thus the relative standard deviation of V_R of less than 10^-4 was achieved for fused silica.

View full abstract

We investigated a method for automated measurement of micro Vickers hardness using image processing techniques. A TV image of the indentation from a microscope was input into an image processing system, and a method was developed to measure the size of Vickers indentation. When an indentation is marked on an unclean surface, the difference of gray levels in the material surface region and the indentation region is small in the input image of the specimen, and the boundary of the indentation is not clear in this case. It is not easy to separate the indentation part from the material surface. Since the features of the gray level distribution in the indentation region are different from those at the material surface region, we applied the image feature extraction methods to detect the indentation region. The result measured by this method was similar to the result measured manually. It was concluded that the accuracy of this method is sufficient for practical use.

View full abstract

In this study, YBa₂Cu₃O_y superconducting wires fabricated by cold extrusion. This process is expected to orient the grains and to produce a high J_C value of the wires. The wire samples, which were extruded with the ratios of extrusion 75, 92, 140, 234, 500 and 2000, were sintered in flowing O₂ gas. As a result, all samples were confirmed to exhibit the Meissner effect and to have TC of 85-93 K. The J_c values were increased with increasing the ratio of extrusion, and the maximum J_c value of 490 A/cm² was measured in liquid nitrogen. However, many pores remained in the sintered body, and the oriented grains could not be recognized from SEM micrographs since the direction of grain growth was unrestricted for the pores.

View full abstract