Some of recent developments in Boundary Element Methods (BEMs) are overviewed, specifically stressing on the following four intrinsic research items in the boundary element and/or integral equation methods: (1)Transformation of domain integral to corresponding boundary integral; (2)Hypersingular integral and its reguralization; (3)Error analysis and adaptive elements; (4)Tools for the inverse analysis.

抄録全体を表示

This paper presents a shear deformable finite element method for the plate bending analysis of laminated composite plates. The finite element for laminated composite plates is based on the first-order shear deformation theory and is obtained with a modified stiffness matrix in the condition in which the out-of-plane shear strain is constant in two directions within an element. The element employed is a triangle with 15 degree-of-freedom which include extension, bending and transverse shear deformation states. Several numerical examples show that good results can be obtained, which demonstrate the validity of the method.

抄録全体を表示

In the bending analysis of thin elastic plates subjected to lateral load, formulas based on the linear theory are quite satisfactory, provided that the deflections are small compared to the plate thickness. If deflections are of the same order of magnitude as the plate thickness, the linear theory may yield results which are erroneous and therefore a large deflection theory should be applied. The nonlinear Berger equation for large deflection problems of plates and the point-matching method are utilized to obtain solutions for deformation in rectangular isotropic plates with a circular rigid body at the center. Numerical solutions are presented in order to illustrate the influence of the rigid body size on the deflection distribution in plates subjected to concentrated load at the center of the rigid body. The numerical procedure is easier than other procedures, such as the boundary element method, and reasonable results are obtained.

抄録全体を表示

The authors proposed a new practical optimum design method (=statistical optimization method) that consists of two steps of an effectivity analysis (using a combination of the design of experiments and the structural analysis) and mathematical programming. The design of experiments is used to generate the approximate evaluation functions for the controlling behavior depending on the changes of design variables of the object structure, using finite element analysis (FEA). The evaluation functions can also be used as tools for directly estimating the behavior of the designed structure. A successive quadratic programming (SQP) method is employed to solve the optimization problem of the approximate evaluation functions. It is confirmed that the proposed method call be used in solving various problems. In this study, the authors applied the present method to the optimum design of a frame column subjected to axial crushing. This study deals with many complex behaviors such as large deformation, contact and material nonlinearities as well as impact behavior. However, this problem has been solved easily using the present method, and satisfactory results of the effectivity analysis and optimum design are obtained.

抄録全体を表示

This paper describes an interactive FEA (finite element analysis) mesh refinement technique in an automatic hexahedral mesh generation system based on the shape recognition technique and boundary-fit method. In this system, a solid model is analyzed, and cubes are put together to construct a corresponding mapping model. Hexahedral meshes are then generated by mapping the cubes onto the solid model. If users are not satisfied with the automatic mesh generation results, they can remake the intermediate models interactively. Furthermore, if this system fails to construct the intermediate models automatically, users can interactively aid its construction. The advantages of this technique are: (1) more types of solid models can be meshed, and (2) users can direct the creation of higher quality meshes.

抄録全体を表示

A method is proposed for estimating the interval of structural responses with respect to changes in structural parameters. The parameters are assumed to change within a closed region, and the first-order approximation is applied to the response change on the basis of finite element sensitivity analysis. The method is based on convex modeling of the parameter changes. The maxima and minima of the responses are located on the boundary of a convex hull by means of linear programming, since the first-order approximation forms a convex function of the parameters. The validity of the proposed method is demonstrated by a numerical example related to a planar portal frame subjected to static loads.

抄録全体を表示

This paper presents an inverse analysis method for the searching and the identification of unknown boundary condition in a finite element method (FEM) on a small memory capacity machine such as a personal computer. This method is formulated by extending an influence function method. Following two calculation processes are adopted in the method: (1) an alternative calculation process which searches the unknown condition and obtains a relative nodal load distribution, and (2) a calculation process which determines a nodal load ratio of the distribution in (1). Both the distribution and the ratio are obtained through a same stiffness matrix in the FEM equation. Then, the paper also describes an efficient calculation method in the determination of both the distribution and the ratio using the answer of the extention of the influence one. Therefore, the extention of the influence one makes it possible to execute the efficient calculation of the inverse analysis for the identification of unknown boundary condition. Finally, the validity of the inverse analysis method is comfirmed from the numerical solution of a three-dimensional inverse finite element model, as an example of an elastic structure with reinforced members under tension, carried out on a small memory capacity machine.

抄録全体を表示

The computational procedures in bifurcation problems with finite displacements of structures may be reduced to solution methods of nonlinear equilibrium equations. The nonlinear stability problems are discussed from the viewpoint of globally convergent nonlinear solution methods. A series of useful path equations is introduced to define trajectories connecting the starting point and the target. Numerical examples are computed to show the versatility of the globally convergent nonlinear solution procedures.

抄録全体を表示

A method for determining hammering sequences in incremental forming of sheets is proposed on the basis of the genetic algorithm. The determination of the sequence for forming a sheet into a desired distribution of curvature in incremental forming is treated as a combinatorial optimization problem. The sheet is modeled to have many numbered points worked with a hammer, and is deformed under a constant load. In the genetic algorithm, a group of individuals having a combination of hammering points is set, and the combinations are optimized by selection, crossover and mutation operations in order to minimize the difference from the desired curvature distribution. The hammering sequences in incremental forming of aluminium square sheets with a hemispherical punch and a flat die are chosen as an example.

抄録全体を表示

High-speed networks such as the Fast Ethernet are currently available for workstation clusters. They have the potential for solving communication capacity problems for large-scale parallel finite element analysis on a workstation cluster. Parallel performance of the domain decomposition method, the domain partitioned conjugate gradient method and the parallel Gaussian elimination method is evaluated on a high-performance cluster system with Fast Ethernet and compared with different communication and computing performance. The optimization of parallel parameters for the Fast Ethernet environment is also performed based on the parallel performance evaluation. A large-scale problem with 350000 degrees of freedom has been solved in about one hour using ten workstations.

抄録全体を表示

Large-scale finite element analyses generally require a very long time for computation. Most of that time is for solving the stiffness equation, so it is important that the computation time for that part is reduced in practical use. We discuss how the Cholesky decomposition, which is used to solve the stiffness equation of FEM, is parallelized in order to reduce computation time. We point out that the synchronization with barrier, which is used popularly, results in the loss of efficiency in the case of band matrix decomposition, and then show that another synchronization mechanism, eventcount, has the ability to construct an efficient algorithm for parallel decomposition. As a result, a shared memory parallel computer exhibits near-linear speedup by the algorithm.

抄録全体を表示

The meshless method is expected to be an effective procedure for realizing a "CAD/CAE seamless" system from modeling to computation. In this study an effective meshless approach that uses node-by-node procedures to discretize the weak formed governing eguations derived by the principle of virtual work is proposed. To realize node-by-node procedures, three key methodologies are utilized. They are interpolation by MLSM, nodal integration and iterative solver using the stiffness at nodes. We describe the formulation and implementation, and investigate the applicability of the method in two-dimensional elastic problems.

抄録全体を表示

Asynchronous massively parallel genetic algorithm (AMPGA) based on a newly proposed "artificial society model" was implemented for relatively large scale optimization problems; and its parallel processing efficiency was evaluated using several optimization problems such as shape optimization and traveling salesman problems in this paper. Because of its asynchronous characteristics produced by the artificial society model which is an agent-oriented model, an emergent behavior such as timings of mating for crossover/mutation and even intergenerational crossover was also observed through shape optimization problems. For traveling salesman problem, large population model was also included in the proposed AMPGA and the effect of the number of agents on each processor was observed in this study. As a result, two types of shape optimization problems and 105-city traveling salesman problem were tested on CRAY T3D massively parallel processor system using the proposed method, and its applicability and parallel processing efficiency were verified.

抄録全体を表示

It is a serious problem that many students neither show any interest in traditional-style lectures, nor do they read traditional-style textbooks. One of the reasons for this is that the traditional teaching method does not fit the younger generation of the TV-era. As a result, university students may fail to understand the fundamental concepts of engineering sciences which are essential in solving real engineering problems. In order to overcome this problem, we have developed a CAI system based upon recent developments in computer hardware and software enabling use of multimedia data, which include text data, graphical data, and sound data. The present system makes use of interactive digital movies, and it runs on inexpensive personal computer systems so that university students can use it personally and learn the engineering fundamentals intuitively. In the present study, the system is applied to a couse in strength of materials including interactive movies on material experiments such as a Charpy test and tensile tests, where it demonstrates the effectiveness and advantages of a multimedia instruction system.

抄録全体を表示

The effects of the fiber volume fraction of crystallized CFRTP (Carbon Fiber Reinforced Thermoplastic polyimide) on creep behavior were investigated. The bending creep tests for 31wt.% crystallized CFRTPs where fiber volume fractions were 0, 3.7, 7.6 and 15.6% were carried out at constant temperatures from 230 to 265°C. For all materials, master curves of creep compliance could be drawn and the reciprocation law of time and temperature of Arrhenius type was applied. With shift not only on the time axis but also on the compliance axis, these 4 master curves were found to correspond to each other. Hence it was concluded that the shape of master curves of CFRTPs mainly depended on the viscoelastic property of matrix resin. The effects of the fiber volume fraction on creep compliance of the master curves are related to the law of mixture. And the relationship between the fiber volume fraction and the shift factors on time axis was linear.

抄録全体を表示

Influences of the recycling process on the fatigue crack propagation (FCP) property of glass fiber-reinforced PEEK were examined. In the present research the influence of the recycling process on the property of matrix resin was focused on. The main results were as follows: (1)It was noted that the FCP property of recycled PEEK/GF materials were as superior to that of virgin material. (2)The length of glass fibers in PEEK resin as not markedly reduced during the recycling process, due to the good fluidity of PEEK resin. (3)FCP curves on a da/dN-ΔK diagram for virgin and recycled PEEK/GF materials were expressed as one curved line on a da/dN-ΔK/E diagram. It was found that the elasticity modulus had been improved by repeated melting and crystallization during the recycling process.

抄録全体を表示

Fatigue tests were conducted on aluminum alloy (2024-T6) reinforced with 20 volume percent and two different sizes of SiC particles using a specially designed servo-hydraulic fatigue testing equipped machine in a scanning electron microscope. The influence of reinforcement-particle size on the initiation and propagation behavior of small fatigue cracks in a smooth specimen was examined. In the case of fine particle-reinforced composites, crack initiation was associated with matrix slip in a particle-rich region. When the crack propagated from the particle-rich region to a matrix-rich region, sub-crack and crack bridging by particles were observed. In the case of coarse particle reinforced composites, cracks were mainly initiated by particle cracking, and particle fracture ahead of a main crack and crack bridging by an uncracked ligament were observed.

抄録全体を表示

The influence of a corrosive environment on fatigue-strength distribution was studied for a slightly notched specimen (stress concentration factor α=1.03) and a sharply notched specimen (α=4.60) of S35C carbon steel. Fatigue strength is markedly lower and the standard deviation is also lower for S35C steel in synthetic seawater than those in laboratory air, because many cracks initiate on the specimen surface. Cathodic protection restores the fatigue strength and increases the standard deviation to those obtained in laboratory air. Since many cracks initiate in the sharply notched specimen in laboratory air, the standard deviation of the fatigue strength is lowered to one-third that of slightly notched specimen. Cathodic protection in seawater reduces the number of cracks and increases the standard deviation to a value twice as large as in the air. The influence was also studied for three kinds of stainless steels with slight notch (SUS304, SUS316 and SUS403). Among stainless steels, low-corrosion-resistance SUS403 in seawater exhibits a large decrease in fatigue strength compared with that in air. However, the standard deviation is not lowered by the kind of stainless steel, because the number of cracks is extremely low compared with the carbon steel. The dispersion of the fatigue life has a good correlation with that of the fatigue strength in terms of standard deviation. Therefore, the standard deviation is a significant parameter for the corrosion fatigue. The standard deviation of the fatigue strength has a good correlation with the number of cracks which led to the final fracture.

抄録全体を表示

In this paper, we present a technique to generate axial shear acoustic resonance in a cylindrical rod or pipe by a magnetostrictively-coupled electromagnetic acoustic transducer (EMAT). We apply this ultrasonic technique to monitor rotating bending fatigue damage. The technique is based on the electromagnetic acoustic resonance (EMAR), which is a combination of the resonant technique and a noncontacting EMAT. The EMAT for the axial shear wave consists of a solenoidal coil to apply a static field along the sample rod (or pipe) and a meanderline coil to induce the dynamic field, relying on the magnetostrictive effect in ferromagnetic metals. The axial shear wave propagates along the circumference of the cylindrical rod or pipe with axial polarization. The resonant mode is determined by the sample dimensions and the wavelength. The major advantage is the capability of evaluating the inhomogeneous materials character in the radial direction using the different resonant modes. The attenuations and the resonant frequencies of the first three resonant modes are measured during the rotating fatigue test. The attenuation shows a remarkable sensitivity to damage accumulation, mainly due to the evolution of the dislocations and the grain size. The ultrasonic evaluation is interpreted in terms of the photomicrograph and transmission electron micrograph observations.

抄録全体を表示

Fatigue tests of full-scale axles with flaws on the wheelseats were performed up to the cycle of 10⁸. Shinkansen vehicle axles, which had a straight bore of 60mm in diameter or a bore of 60mm at both journal parts and 110mm between two wheelseats for the purpose of testing, had been induction-hardened almost over their entire length. The result of fatigue tests was evaluated from the calculated stress intensity factor by regarding the flaw as a crack, whose depth was estimated by adding the compressive yielded zone size to the depth of the flaw. The crack at the tip of a flaw which originated a fatigue fracture was divided into two regions. The measured size of the region adjacent to the flaw was nearly equal to the calculated compressive yielded zone size. On the other hand, the crack length at the tip of a flaw which did not cause a fatigue fracture was almost 30% of the calculated compressive yielded zone size.

抄録全体を表示

A method of lens-assisted holographic microscopy with high spatial resolution has been developed for taking microscopic photographs of rapidly propagating cracks, the speed of which is 200-600m/s or more. The spatial resolution is more than 350 lines/mm, which is about three times as high as those of previous methods of pulsed holographic microscopy utilized in studies of dynamic fracture. One can apply the method to rapidly propagating cracks in PMMA, take microscopic photographs of them and, measure craze opening displacements which are less than 10 microns, the measurement of which has been difficult so far. The method will also be applicable to measuring the opening displacements of rapidly propagating cracks in ceramics, the displacements of which are about 10 microns or less.

抄録全体を表示

The tension-softening model has been employed to evaluate the crack growth behavior in rocks. In this paper, the tension-softening relation was determined taking the crack tip toughness into account. The extended J-based technique is applied to determine the tension-softening curve for a granite which has crack tip toughness. Compact tension (CT) specimens were used to conduct J-based fracture toughness tests on the granite. The crack tip toughness was measured using K-resistance curves in the granite. Three sizes of CT specimen were used to examine the specimen size effect on the deduced tension-softening curves. It was shown that the extended J-based technique provides reasonable tension-softening curves irrespective of the specimen size, suggesting the validity of the extended J-based technique. Finally, the necessity of the introduction of the crack tip toughness is demonstrated by performing numerical simulations of the load-displacement behaviors of the granite.

抄録全体を表示

In this paper, the growth behavior of fracture process zone in rock is examined from the viewpoint of the microscopic strength distribution. Numerical calculations were performed using a network model in order to analyze the microcracking behavior which is accompanied with the crack extension. CT specimen configuration was used in the network model analysis. The network model is composed of brittle stick elements, and microscopic inhomogeneity is represented by the strength distribution of the stick elements. The calculation results show that the microcracking behavior depends on the microscopic strength distribution. The geometrical characteristic of the fracture process zone is observed to change from a distributed microcrack zone type to a localized zone type, depending on the microscopic strength distribution.

抄録全体を表示

The velocity gradient is discontinuous on the necking plane that forms a boundary between the shear band and its outside. On the other hand, a characteristic plane is the plane beyond which solutions cannot be extended without additional information, because some of the derivatives normal to the plane cannot be determined by their governing equations alone. Therefore, on the characteristic plane some derivatives of stress rates become indefinite and such property must be also required for the necking plane where some of the velocity gradients are discontinuous. We examine the mode of the characteristic plane in the well known Cam-clay model, which has only two modes under plane strain rate, that is, modes P and O. Mode N that the characteristic plane is a principal plane cannot be realized in non-cohesive soil. The angle between the shear band and the maximum compressive principal direction is predicted to be 45 degrees under plane strain compression when the hardening modulus h vanishes, though it is smaller than 45 degree at the dilatant state. On the assumption that plastic instability occurs when h=0, the predicted angle does not agree with the experimental result of about 30 degrees.

抄録全体を表示

Hydraulic fracturing rock stress measurement utilizing shut-in pressures of transverse cracks, the reliability of evaluated in situ stress state depends upon the independence of crack orientation, qualitatively speaking, the reliability improves when orientation of cracks differs significantly from each other. In order to evaluate the relationship between independence and reliability, we introduced the condition number of a matrix which is completely determined by the orientation of transverse cracks induced by hydraulic fracturing. We examined 30 sets of artificial data, each set of which consisted of ten cracks generated by random sampling, under two given stress states, and investigated the relationship between the condition number and the reliability of the estimated stress state. It was revealed that the reliability and the condition number had strong correlation and, thus, the condition number is a very useful measure of the reliability of the estimated stress state.

抄録全体を表示

Stresses and stiffnesses of weld-bonded box section beams subjected to bending moments have been studied analytically. The influences of spot-welded spaces, Young's modulus and thickness of the adhesives on the stress distributions and the bending rigidities were investigated analytically and experimentally. Maximum shear stresses in the adhesives occur at the inner edge of adhesive layers and decrease with decrease of Young's modulus of the adhesives. The analytical results were consistent with the results of FEM and experiments.

抄録全体を表示

Using piezoelectric actuators which have the capability to induce various amounts of strain upon activation, it is possible to reduce stress concentrations under varying external load conditions by distributing them in high stress regions in advance. In this paper, we verify experimentally the relaxation of stress concentrations in an adaptive plate with a circular hole using piezoelectric actuators. The adaptive plate is a simple actuator/substrate system in which the substrate is an acrylic acid resin plate with a circular hole and actuators are PZT piezoelectric ceramics bonded to the plate surfaces around the hole. The optimal voltage calculated to minimize maximum stresses in the substrate subjected to uniaxial tensile load was applied to each piezoelectric actuator. The experimental result of circumferential stress distributions on the circular hole boundary was compared with the numerical one. Both results appeared in agreement and relaxation of stress concentrations in the adaptive plate was verified.

抄録全体を表示

The stress distributions of single-lap adhesive joints of dissimilar adherends subjected to external bending moments are analyzed as a three-body contact problem using the two-dimensional theory of elasticity. In the analysis, dissimilar adherends and an adhesive are replaced by finite strips, respectively. In the numerical calculations, the effects of the ratio of Young's modulus, thickness and length of dissimilar adherends on the interface stress distributions are examined. As a result it is found that the stress singularity occurred near the edge of the interface and it increases at the edge of the interface of the adherend with greater Young's modulus. It is also seen that the singular stress decreases at the edge of the interface of the adherend with larger thickness. In addition, experiments are carried out and fairly good agreement is found between the analytical and the experimental results.

抄録全体を表示

Thermal stress distributions in adhesive butt joints under uniform temperature changes are analyzed using a two-dimensional finite element method. The effects of hole defects and fillers in the adhesive layer on the stress singularity and the delamination growth at the edge of the interface are examined and the amount of delamination growth is predicted using the tensile stress near the edge of the interface. Photoelastic experiments are carried out at room temperature. The delamination growth is measured by a scale when the joints are cooled stepwise. A fairly good agreement is seen between the measured values and the predicted results for the delamination growth and between the results of the photoelastic experiments and the analyses. As a result, it is shown that the hole defects reduce the delamination growth but the effect of the fillers on the delamination growth is small.

抄録全体を表示

Fractured surfaces of epoxy adhesives under mode I loading have fractal characteristics. The fractal dimensions of fractured surfaces can quantitatively characterize the fractured surface topology. The effects of rubber modification, adhesive thickness and cross-head speed on the static fractured surfaces of epoxy adhesives were examined using the fractal dimensions. The fractal dimensions become high due to rubber modification. The fractal dimensions are related to the static crack growth properties for both rubber-modified and unmodified adhesives. The following equation gives the relationship between the crack extension resisitance, G, and the fractal dimension, D: log G=a·D+b, where a and b are experimental constants. Regardless of whether the adhesives contained rubber particles, fractal dimensions were not affected by adhesive thickness and cross-head speed.

抄録全体を表示

Optimum design of cruciform specimen was performed for biaxial loading-fatigue test on the basis of the calculations through elastic finite element method. The stress must be greater in its central thickness-reduced region than in any other regions including four corners in the specimen, since fatigue crack should originate at this site. The stresses calculated in the central thickness-reduced region of the cruciform specimen with the determined configuration and dimensions are in quite good agreement with those measured with strain gauges, evidencing that the present calculation method is reasonable for designing the cruciform specimen. Several fatigue crack growth tests were performed under uniaxial and equi-biaxial loading using the cruciform specimen. The tests showed that crack growth was not influenced by the in-phase cyclic loadings parallel to crack plane, being controlled only by mode-I cyclic loading. The background of such test results was examined in terms of plastic deformation calculated around the crack tip by means of elastic-plastic finite element method.

抄録全体を表示

A compact dynamic loading device for the strain rate range from 300/s to 1500/s which enables us to measure a long stress pulse without wave disturbance due to wave reflection and interference has been developed and examined. A special load cell mechanism, a stress sensing block, which consists of the small cylindrical knob followed by a relatively large mass block was employed, and the displacement of the specimen was measured by means of the magnetic linear encoder. In this device, the dimension of input and output bar system can be reduced to 1/2 or 1/3 of that of a ordinary split Hopkinson bar device, while the accuracy of the obtained stress-strain curves retains the same level of that obtained by the split Hopkinson bar device. In order to verify the applicability of the device, stress-strain rerations of several fcc, bcc and hcp metals were measured, and their experimental tendencies are discussed.

抄録全体を表示

We present a unified analysis of plate bending problems based on classical theory containing an elliptic hole or elliptic rigid inclusion. Theoretical closed-form solutions under singular loads as well as bending moments at infinity are shown for cases of arbitrary-shaped elliptic hole or rigid inclusion. The results of stress intensity (singular) factors under the same loads are also derived. Many numerical examples are given by graphic representation.

抄録全体を表示

A computer program was developed to deal with a wide range of structural problems of generally laminated FRP rectangular plates. It covers problems of static bending, elastic buckling and free vibration. The analysis uses the Ritz method which is based on the classical plate theory, and a unified solution is derived to accommodate arbitrary support conditions (any combinations of classical boundary conditions, elastic point supports and elastic foundation) and loading conditions (in-plane and out-of-plane loads). Numerical examples are given for symmetric cross-ply and antisymmetric angle-ply laminated graphite/epoxy plates. Convergence study is presented by increasing the number of terms in the present solutions, and the effectiveness of the present method is well demonstrated by comparison with exact solutions and finite element solutions.

抄録全体を表示

In estimating the reliability of a structural system; a limit-state function is needed to relate the structural state (failure or safety) to random variables of the system. However, it is not easy to obtain such an explicit function for complex structures. As a consequence, structural analysis must be performed repeatedly to check the structural state, which is very expensive. We develop an approximate limit-state function by using a neural network. Orthogonal factorial designs are selected as learning data for the network. An "active learning algorithm" is proposed to enable the network to determine important failure regions by itself and also to do further learning at those regions to achieve a good fitness with the real structural state there. The validity of the method is illustrated through numerical examples.

抄録全体を表示

The deformed mechanism of pushing a long rod into a pipe has been analyzed from the point of view of elastic stability. In the case of pushing a long rod into a pipe, the rod is in a complicated manner deformed. At first, the rod is a linear figure, then, it is transformed into an s-shape, and finally, into a spiral figure. We describe the relationship between the critical pushing distance in a pipe and the dynamic characteristics of a rod, such as weight per unit distance, bending stiffness, and friction coefficient to the pipe wall through theoretical and experimental investigations. A design standard of a long rod structure has been clarified by these studies.

抄録全体を表示