Thermal transient fatigue tests of five straight pipe models with circumferential weld joints and a nozzle model made of SUS304 steel were performed. Crack locations, directions and densities were observed. Initiation sequences of cracks at different parts corresponded well with the result of analytical prediction using FEM. We evaluated the potential margin on the allowable number of thermal transient cycles according to the ASME Code Case N-47 and the elevated temperature structural design guide developed by the Power Reactor and Nuclear Fuel Development Corporation. The results show a considerable safety margin for the initiation of cracks of 1mm in depth.

View full abstract

The effect of heat treatments (as cast, T4, and T6) on the notch sensitivity of a squeeze-cast Al alloy, AC4CH, is investigated using specimens with a circumferential notch subjected to rotating bending. Although the increases in static strength by heat treatment contribute to the improvement of σ_w2, they hardly contribute to the increase in σ_w1, where σ_w1 and σ_w1 are the crack initiation and propagation limits in notched specimens, respectively. The notch factor K_f2(σ_w0 /σ_w2, σ_w0: fatigue limit of plain specimen) of a squeeze-cast Al alloy is larger than that of a gravity-cast Al alloy, and is smaller than that of a wrought Al alloy.

View full abstract

The creep-fatigue life prediction method based on the overstress concept is exteded for life evaluation under the biaxial creep-fatigue condition. The time-independent damage parameter is described by two geometically different damage parameters, one of which represents the fatigue damage in terms of a resolved overstress, and the other represents the same in terms of the overstress mormal to it. The time-dependent damage parameter is described by the maximum principal overstress. A strain rate effect is also introduced into the multiaxial creep-fatigue damage model in order to evaluate the test results with an environmental effect. The presented theory is applied to the biaxial creep-fatigue test results which were previously reported by the authors, and it is shown that the theory has a great applicability to predict the biaxial creep-fatigue life with an environmental effect within the predicability of a factor of 2.

View full abstract

The statistical property of the fatigue behavior was investigated for normalized 0.21% carbon steel plain specimens. An emphasis is to investigate separately the statistical properties of the initiation and propagation of microcracks through successive observations of the specimen surface by the plastic replica method. Results show that the distributions of crack initiation life and microcrack propagation life are expressed as two- or three-parameter Weibull distributions. More-over, the three-parameter Weibull distribution is well fitted to the crack length distribution, and the value of the shape parameter m in this case decreases with an increase in the relative cycle ratio N / N_f. As the main reason for the decrease in m, we can point out the decrease in the number of newly initiated cracks at each N / N_f.

View full abstract

The electric potential CT method proposed by the present authors was applied to the measurement of a three-dimensional elliptical internal crack embedded in a steel rod. Data of electric potential distribution observed on the surfaces of the rod were computer-processed inversely to identify the crack location, shape and size. As an inversion software, the least residual method was employed. This method was based on the least residual criterion, in which residuals were evaluated between observed potential data and those computed by the boundary element method for various assumed cracks. A hierarchical scheme, which involved two-dimensional scanning inverse analyses and subsequent three-dimensional inverse analyses, was applied successfully for efficient crack identification. The location, shape and size of the elliptical internal crack were determined by the proposed method with good accuracies.

View full abstract

The floating roofs are used in many large-sized oil storage tanks to reduce evaporation. The single deck type floating roofs, considered herein, consist of a thin circular plate (deck) attached at the edge to a buoyant ring (pontoon) of box shaped cross section. Under the accumulated rain water condition, the deck is deflected largely, and both its edge part and the pontoon are compressed tangentially. So, the geometrically nonlinear stress analysis and the bifurcation buckling analysis are required to design the floating roofs. Since the load condition due to the rain water depends on the deflected deck shape, it is difficult to find the unique equilibrium roof condition. This paper describes the bifurcation buckling analysis of the floating roofs by the axisymmetric finite element method. The load incremental method is used in order to modify the water level at each stage.

View full abstract

This paper deals with the applicability of an efficient calculation method using the LDU decomposition process of the stiffness matrix of a tandem-type substructure composed of various substructures in numerical solutions of the finite element method (FEM). This method is applied to an analysis of the two-dimensional elastoplastic strip with semicircular notches. The FEM using this method greatly decreases CPU time and memory capacity in comparison with the conventional FEM applied to the whole structure. The new method also produces good and accurate results, and is therefore effective for microcomputer-aided FEM.

View full abstract

This paper presents an efficient calculation method comprising the following two methods: (1) the method using LDU decomposition process of the stiffness matrix of a tandem-type substructure composed of various substructures and (2) the penalty function method available in the open / closed behavior of crack propagation, in numerical solutions of the finite element method (FEM). This method is applicable to an analysis of the stress intensity factor based on energy release rate theory in a two-dimensional cracked structure. The FEM using this method contributes to the decrease of both CPU time and memory capacity in comparison with the conventional FEM applied to the whole cracked structure. The use of the method also gives good accurate results. Therefore, this method is effective for microcomputer-aided FEM.

View full abstract

The crack propagation behavior under cyclic load was investigated on two types of sintered Si₃N₄. The main results obtained are as follows. (1) The crack propagation rate was accelerated by cycling the load. As the stress ratio decreased or the test frequency increased, the crack propagation rate increased. (2) If the fracture toughness of a specimen was large, the crack propagation rate was small as compared with that of another specimen with small fracture toughness. (3) Debris of the fractured grains and wear mark were left on the fracture surface under the cyclic loads, however, neither such debris nor wear mark were observed on the fast fracture surface. (4) By measuring the crack opening displacement, crack closure was observed. The size of debris (∼1μm) on a fracture surface under the cyclic load was nearly equal to the crack opening displacement at no load. It was concluded that the crack closure of sintered Si₃N₄ was induced by debris between the crack surfaces.

View full abstract

Tensile and fracture toughness tests of a glass-fiber reinforced polyamide 6.6 were carried out to investigate the effects of heterogeneity and loading rate on tensile strength and fracture toughness. The extruding material used had a slight heterogeneity of fiber orientation. The higher strength and fracture toughness were found in the preferred direction of the fiber orientation. The heterogeneity of fracture toughness was more significant than that of tensile strength. A significant effect of loading rate on fracture toughness was also observed: The fracture toughness decreased with reducing the loading rate. This is considered to result from the change in fracture mechanisms with loading rate.

View full abstract

A new development of the slab analogy is proposed for the experimental determination of the K_I and K_II mixed mode stress intensity factors by the combination of the slab analogy and moire measurement. In this method, sharpness of crack tips and arbitrary forms can be more easily given to the crack models to be analyzed in comparison with existent mothods. The present new method was applied to the analysis of theoretically known cracks, such as inclined linear, linear bent and circular cracks, to prove its applicability. The results obtained suggest further effective application to the analysis of Stress intensity factors of the cracks with arbitrary configuration and locations.

View full abstract

This paper presents a new scheme for the estimation of unstable ductile fracture using the J-integral. The proposed method uses a load-versus-displacement diagram which is generated using fully plastic solutions. By this method, the phenomena of the ductile fracture can be grasped visually. Thus, the parametrical survey can be executed far more easily than before. Then, using the proposed method, unstable ductile fracture is analyzed for single-edge cracked plates under both uniform tension and pure bending. In addition, several parametrical surveys are performed concerning (1) J-controlled crack growth, (2) compliance of the structure, (3) ductility of the material (i. e., J-resistance curve), and (4) scale of the structure (i. e., screening criterion). As a result, it is shown that the proposed method is especially effective for the parametlical study of unstable ductile fracture.

View full abstract

Ductile fracture extension of a notched specimen was investigated by means of the observation of the crack growth behavior on the specimen surface. Correlation between the ductile crack growth and the initial notch length was examined in order to understand the feature of the ductile fracture which was extended by the center crack. In this experiment, the crack growth rate, da /dε, for the strain increment depends on the crack length, a, and the value of the nondimensionalized crack growth rate, (da / dε) / a, is higher for the short notch length that the long notch length. Crack opening displacement near the crack tip is also dependent on the crack length and the behavior of the crack opening shows a good relation with that of crack growth.

View full abstract

Failure in composite materials and metal-ceramic joints frequently occurs in the vicinity of a bonding edge. It is therefore necessary to clarify the stress and displacement fields around the bonding edge. In order to develop a suitable boundary element method for such a stress analysis, closed-form solutions are derived in the present paper for point force applied in the interior of a two-phase material consisting of two semiinfinite transversely isotropic elastic media bonded along a plane interface. The interface is parallel to the plane of isotropy of both media, and the solutions are applicable to all combinations of elastic constants. These solutions involve the Green function for infinite and semiinifinite transversely isotropic solids as well as infinite, semiinfinite, and two-phase isotropic solid.

View full abstract

The imperfection sensitivity of elastic-plastic rectangular plates under biaxial compression is investigated. In this analysis, a Reissner-type variational principle in conjunction with the incremental Rayleigh-Ritz technique is used to determine the buckling process of a plate which obeys the Prandtl-Reuss flow rule. The influence of initial imperfections in geometry, the elastic modulus, the work hardening rate and boundary conditions on the maximum support stress are clarified numerically. The results show that a small imperfection considerably reduces the buckling strength of the plate.

View full abstract

In the first report, a new expression op a constitutive law for elastic-plastic material was proposed using two transition parameters which are essential in describing the property of plastic materials. These parameters denote the magnitude and direction of plastic strain increments. Here, Localized necking of rigid-plastic plates under biaxial stretching is studied according to Storen-Rice theory. A constitutive equation for linear-comparison solids is used for Hencky's deformation rule.

View full abstract

The present paper is concerned with the formulation of an anisotropic constitutive model of blood vessels typically observed in peripheral arteries. The hysteresis loops under cyclic loadings are assumed to be caused by the loading path dependence and the loading rate dependence. In view of this, the total deformation is divided into the sum of the elastic part and the inelastic part. The elastic part is established by postulating a strain energy density function of an exponential type whose exponent is the quadratic form of the transversely isotropic type with respect to the elastic strain components. The inelastic part, on the other hand, is formulated by modifying the viscoplastic model proposed in the previous paper, that is, the transverse isotropy is incorporated into the kinematic hardening rule. Comparisons of the simulation results with the literature show that the present model can describe the anisotropic behavior appearing in the cyclic hysteresis loop and in the stress relaxation.

View full abstract

Strengthening of a Fe-27.59% Ni-0. 28%C (Wt%) steel in tension results from both the ordinary strain hardening in austenite and martensite which compose a dual-phase steel and the hardening due to martensite induced during straining of the steel. The volume fractions of martensite induced in the steel at 0°C, -19°C, -34°C and -49°C are measured by the X-ray diffraction method (Motarget, multi-peak method). The stress-strain relations of the steel at these temperatures are analyzed and expressed by the equations of a variable volume model (abbr. as VVM) previously proposed by us. The analytical relations of each strain, each stress, transformation strain, balanced bonding mechanism of two phases, contributions of strain and stress of each phase and so on are discussed in detail. These analytical curves show excellent agreement with the experimental data obtained from a study of the X-ray residual stress measurements on this steel, and the equations of the VVM model are shown to be successfully applied to the Fe-Ni-C steel which induces the lenticular martensite with different tensile properties from those of the lath martensite in SUS 304 stainless steel reported previously.

View full abstract

Quasi-static neck and bulge propagations in polymer are studied for the case of circular cylinders subjected to internal pressure. For a J₂-flow theory material a finite element simulation completely characterizes the development of axisymmetric neck and bulge, and their propagation in axial direction. Furthermore neck development and its propagation in circumferential direction are also investigated for the circular cylinder under internal pressure with axially plane strain condition. The simulation results include the internal pressure versus volume change, the shape of bulges and necks, and strain distribution. Parametric studies are also performed to investigate the effect of variables employed to characterize the uniaxial stress-strain relation of polymeric cylinder on bulge and neck propagations.

View full abstract

Multiaxialhistory effects of cyclic plasticity on creep were studied experimentally for type 316 stainless steel at an elevated temperature. Firstly, after a cyclic stabilization of prior cyclic tension-compression of constant total strain amplitude, constant stress creep tests were conducted under simple tension, simple torsion, and combined tension-torsion (√(3)τ/σ = 0, oo, 1). Three levels of creep stress were chosen which corresponded to the saturated cyclic stress amplitude and to certain values larger or smaller than that. Secondly, in order to elucidate the path shape effect of the prior strain cycle, a non-proportional strain cycle along a circular path was followed by a constant stress creep in tension where the strain amplitude was specified so as to result in the same saturated stress amplitude as in the tension-compression cycle compared. It was observed that the prior tension-compression cycle induced an anisotropic hardening for creep ; creep resistance was enhanced in the torsional direction, while somewhat decreased in the tensile direction. This is similar to the cross hardening effect reported in experiments on non-proportional cyclic strain hardening. The circular cycle, as expected, showed a much more significant hardening effect on creep.

View full abstract

In order to clarify the strain attenuation during shear wave propagation, impact torsion experiments are carried out using a PMMA rod specimen. The strain pulses measured by strain gages are complicated due to the superposition of many reflected and transmitted waves from the supported end of the specimen. To obtain data from such pulses for determining a viscoelastic model, a method which modifies the pulse form is developed taking into account wave propagation behavior in the experimental system. From Fourier components of the modified pulses, complex compliances and viscoelastic constants are evaluated assuming that the three-element standard linear solid model can be applied to the specimen. The attenuation of shear strain with propagation distance is predicted by the one-dimensional viscoelastic wave propagation theory. As a result, predicted values of shear strain agree well with experimental ones.

View full abstract

In the design of a shape memory alloy helical spring, a basic method of designing the spring based on the stress-strain-temperature relation of the material is an important problem to be solved. In the current study, deformation behavior of the shape memory alloy helical spring was analyzed based on stress-strain-temperature relation of the material. The relationship between load and deflection of the spring as well as behavior of transformed region were discussed. The main results are summarized as follows. (1) Relation between load and deflection of the spring. is represented by using a simple model of the stress-strain-temperature relation. (2) For a certain maximum deflection, recoverable force and recoverable strain energy increase with temperature. Dissipated strain energy takes a maximum value at a certain temperature. (3) The transformed region in the cross section of the wire expands into the center from the surface under the loading process, but contracts toward the surface under the unloading process. For a certain maximum deflection, the transformed region is small but the load is high as the temperature is high.

View full abstract

Under the assumption that deformation in a crystal grain is uniform, a method of determining plastic strain in a grain is presented. Information on grain deformation allows us to find the strain distribution of sheet metals. Two line elements, each of which binds two arbitrary points on a grain boundary, are taken as basic vectors before deformation. They change lengths and directions under the process of deformation. Their relationship bofore and after deformation determines the deformation gradient and strain in the grain. A tensile specimen of sheet steel is used as the sample of strain measurement. It is found that uniform deformation is not guaranteed in a grain. Even if two grains are in contact with each other, the average strain in each grain is different. Therefore, in order to determine the macroscopic distribution of deformation, we have to average grain strains in some area.

View full abstract

Uniaxial tensile tests on Ti-Ni alloy wire under cyclic loading and unloading were carried out. The influence of maximum strain, test temperature and shape memory processing temperature upon cyclic characteristics of pseudoelasticity was examined. The results obtained are summarized as follows. (1) Starting stress in stress-induced martensitic transformation decreases with the number of cycles. The rate of decrease takes almost a constast value without depending on the test temperature and maximum strain. (2) In the region of stress-induced martensitic transformation, residual strain in each cycle increases with test temperature. (3) The change of strain energy per unit volume for number of cycles becomes larger in the order, dissipated strain energy, total strain energy, and recoverable strain energy. (4) Yield stresses of stress-induced martensitic transformation and reverse transformation decrease with the shape memory processing temperature. (5) The transformation strain range at each 100th cycle takes a maximum value at a certain shape memory processing temperature. (6) Every characteristic value of pseudoelasticity varies significantly in the early cycles, but becomes stable after these cycles.

View full abstract

An implicit type rigid-viscoplastic finite element program is formulated to simulate a general sheet stretch-draw operation. The analysis is based on the rigid-viscoplastic membrane theory; it accounts for finite strain and displacement and embodies Hill's new theory of normal anisotropy. The features are retained to reduce the CPU time and get a efficint program for industrial applications, such as the sheet metal forming simulation, while keeping the complexity of the material plastic behavior. Special algorithms for dealing with the die contact condition, material unloading, Coulombic friction condition are developed. The accuracy of this simulation code was confirmed by the comparison with the explicit type elastic-plastic finite element analyses in case of simple forming operations such as hemispherical punch stretching of disks and one-inch whdth strips. Further this code is applied to analyse the two-dimensional DRAW-IN problem under the plane-strain hypothesis. It was demonstrated that the results are much affected by the draw-bead resistant force and the friction coefficient.

View full abstract

Nonmetallic inclusions are the cause of a marked reduction and a large scatter of fatigue strength of high-strength steels. The sizes and positions of inclusions influence the fatigue strength. Murakami et al.'s equation has successfully shown the correlation betwween the fatigue strength and the size of the inclusion at the fracture origin, and has enabled us to predict the fatigue strength considering the inclusion size. The maximum size of inclusions contained in a specimen or a machine component must be determined in order to predict the reduction of the fatigue strength. The statistics of the extreme values for the determination of the maximum size have been applied to the experimental microscopic inspection data, though those data give essentially two-dimensional information, because the metallographic inspections were conducted only on a section of a specimen. In order to examine the validity of two-dimensional metallographic inspection data of inclusions, the numerical simulations considering the spatial and size distributions of inclusions are carried out. The prediction errors resulting from conventional metallographic inspections are estimated.

View full abstract

This paper describes two expert systems to provide advice for finite element analysis which were developed through the activities of Research Committee RC-74 on Nonlinear Finite Element Methods under the auspices of the Japan Society of Mechanical Engineers. The elastic analysis of a cantilever beam is taken as the example, the primary objective of the committee work being to examine and establish the fundamental methodology for constructing expert systems combined with finite element analysis. A rule-based system was developed at first by the use of OPS and based on the knowledge of finite element analysis collected from the committee members. This system was revised into another frame-based system written in Prolog with special attention paid to the integrity and consistency of knowledge representation. The prototypes thus developed provide a basic framework for the construction of such expert systems.

View full abstract

Previous expert systems have provided only one type of knowledge representation, even though knowledge representations change from application to application. Therefore, we developed an expert system shell on a personal computer for producing a welding procedure specification. As this task requires a wide variety of knowledge and experience, it is quite a burden even for a highly experienced engineer. WELSYS 2 was developed in order to alleviate this burden. What characterizes WELSYS 2 is that attention is paid to the coupling of symbolic and numerical processing and the linkage to the database. These functions are essential for this kind of application. Although the current system contains knowledge about a pressure vessel, it can be used for a broad range of applications if the content of the knowledge base is replaced.

View full abstract

NDI (Nondestructive inspection) is one of the most important tasks to secure the safety of welded structures. Since there are many kinds of inspection procedures and methods, and structures are becoming more large-scale, complicated and diverse, as well as the fact that inspectors are required to make the procedure specification in a short time, use of NDI has been increasing rapidly. Therefore, with the aim of reducing their workload, an expert system has been developed using Prolog to support inspectors in making decisions and in writing the NDI procedure specification for an oil storage tank. The characteristics of our system include: (1) the user can make an NDI procedure specification esily ; (2) the computer enables ease in obtaining information of NDI and its related laws and codes for the structure; and (3) it is possible to produce an NDI procedure specification quickly.

View full abstract

In bar rolling, a bar is usually made by applying front and back tension, so it is important to analyze the effect of the front and back tension on rolling properties. Further more, it is also important for precise rolling to confirm the ratio of the variation in rolling properties to the change of tensions. In this paper, the analysis of the effect of tension was performed by the energy method using finite element division, which we have proposed previously. First, the effect of the front and back tension on spread, forward slip, roll force and roll torque were proved reasonable by comparison with the experimental formula. Next, the ratio of the variation of the roll torque to the change in tension was obtained approximately using the results of no tension applied. As the roll radius increases, the ratio of the variation of spread, forward slip, roll force and roll torque to the change in tension decreased. As the reduction and height / width ratio of caliber increases, the ratio of the variation of roll torque to the change in tension decreased.

View full abstract