X-FEM is an extension method of the standard finite element method (FEM) for analyzing fracture problems. Because a crack surface discontinuities can be defined in the finite element models, it is relatively easy to define crack in the FEM. Therefore, crack propagation analysis by X-FEM is more stable than ordinary FEM. Meanwhile, it is known that it reaches the opposite surface, becomes a through crack, and propagates continuously. In this paper, we report three-dimensional plate fatigue crack propagation analysis that progresses from a small surface crack to a through crack.

We propose a new regularization method using spectral X - ray CT with patient - specific characteristic parameters. In this method, firstly, the characteristic parameter unique to the patient is identified by applying a hierarchical Bayesian model. Secondly, a tomographic image of the X-ray attenuation coefficient is reconstructed by the MAP method with the previous identified characteristic parameters. In order to show the effectiveness of this method, we performed verification using numerical simulation.

Phononic crystal, an artificial structure in which materials having different acoustic characteristics are periodically arranged, has been attracting much attention as a novel means for controlling ultrasonic wave propagation. As an analogy from optical fiber developments, Gradient Index (GRIN) lens based on a refractive index distribution in the phononic crystal have been proposed for an application to acoustic-energy transmission systems. In the present study, we designed a phononic GRIN lens in a silicone rubber (poly-dimethylsiloxane), operating at 470kHz. A quadratic refractive index distribution is derived from controlling group velocity of the crystal by varying smoothly filling factor of the unit cell from 0.380 to 0.495. A 19 layer x 12 row GRIN lens is thus designed. Using the finite-element simulation we show that the designed GRIN lens can focus an elastic wave to a width of its wavelength with the intensity twice as much as that of the incident wave at the frequency (~1Pa).

Flapping wing insects require quick response towards ambient perturbation to maintain stabilized aloft. Passive feathering mechanism which can reduce complexity in wing hinge, is considered to highly possibly exist in some insects. The flight stability of bumblebee hovering with passive feathering mechanism is studied by applying one-directional disturbances, with 8 degrees of freedom including 6 DoFs in body and 1 DoF in each wing seperately. We mainly present the result under unilateral disturbance, which is most significant among all other disturbances. And we find that under asymmetric disturbance, the phase difference between two wings is observed in response to asymmetric disturbances, indicating that the passive feathering mechanism is capable of enhancing flight stabilization within a period of wing beat strokes.

The high frequency electromagnetic field simulation with the E formulation and the edge finite element method leads complex symmetric systems of linear equations. For solving such systems, the iterative method such as COCG method has been widely used, however, it suffers from slow convergence rate and high computational time. To improve the convergence rate, we have been studied to utilize the double-double arithmetic of complex numbers in an iterative process. This study focuses on parallelization and performance optimization of mixed-precision iterative methods for complex symmetric systems. Some numerical examples with OpenMP are demonstrated.

Fracture test of SQV2A (ASTM A533B) C(T) specimens were simulated by elastic-plastic ordinary state based peridynamics with work hardening. This research suggests that under high triaxial stress field, the stretching parameter s₀₀ should be varied depending on its triaxiality. For the C(T) specimen, the distribution of triaxiality and s₀₀ keenly affects unloading process with crack propagation.

We developed a dynamic load balancing algorithm based on the computational time of each particle in the Explicit Moving Particle Simulation method. This time-based algorithm has a potential to settle dynamic load balancing problems caused by polygon wall boundary models. The boundary models obtain a pressure value from the specific particle located closest to respective polygons among other particles and add a variety of computational costs to each particle. Although the former algorithm developed by Murotani et al. decomposes an analysis domain into processing elements based on the number of particles, it does not function efficiently for dynamic load balancing if the number of polygons becomes increased. On the other hand, the proposed algorithm offers various weight values depending on the computational time of respective particles. Consequently, the time-based algorithm provides each processing element with more appropriate weight values compared to the previous one. In this paper, we demonstrate the parallel efficiency of the time-based algorithm by solving a hydrostatic pressure problem and a dam break problem.

In recent years, acoustic diode has attracted attention as a means for controlling sound propagation direction. Recently we have proposed a nonreciprocal propagation mechanism of Lamb wave in hetero-structured phononic thin plates combining two types of phononic structures consisting of piezoelectric bodies and dielectrics. Its operating principle is the following: The forward incident S wave is partially converted into an A wave in a mode converting part and it propagates as a hybridized S + A wave. At an entrance of a mode selecting part, the S-wave component is blocked by an S-wave band gap of the mode selecting part. Thus only the A-wave component propagates through the part. For the reverse incident wave (S wave), it is cut off (reflected) by the S-wave band gap in the mode selecting part, and the rectification effect is thereby realized. This mechanism has, however, a drawback in large transmission loss due to large impedance mismatch between the mode conversion part and the mode selecting part. In the present study, we attempt to reduce the transmission loss by introducing a matching layer which has an intermediate acoustic impedance between the parts at the interface. Using the finite element simulation, we show that the reflection can be suppressed and, in turn, the transmission loss is improved by up to 57%.

In recent years, drone has been utilized in various applications including industrial use. On the other hand, when a drone is to be flying in an area where noise is a problem, such as a residential area, the aerodynamic noise generated by the rotation of the propeller becomes a problem, so it is urgent to take measures to silence the drones in the flight state. Therefore, in this study, the silent design of the propeller was carried out based on the flow distribution of the drones by the propeller rotation and the pressure distribution of the propeller surface obtained from the numerical fluid analysis (CFD). Furthermore, the noise level was evaluated by using acoustic measurement and acoustic analogy method, and quietness of about 2 dB was achieved before and after noise reduction measures.

We are researching to utilize novel machine learning technology and CAE combined for design-estimation and design-optimization in product development. As an example, here is described to verify the fundamental methodology for predicting fluid drag force with a product shape using Deep Learning model which is trained by CFD simulation results.

This paper presents a numerical procedure to evaluate the buckling of cylindrical structures under pressure load with shell element appropriately. In a situation where the internal and external pressures are subjected to the top and bottom surfaces of cylindrical shells, it is difficult to evaluate the influence of the pressure load of the structure appropriately, due to the complicated loading conditions. Numerical solutions, which are obtained by using conventional shell elements, and analytical solutions are expressed by using the difference between the internal and the external pressures as the external force. On the other hand, solid elements can be employed to model the loading conditions properly, but huge computational cost may be required. In this work, the authors propose a numerical procedure for buckling analysis that evaluates the pressure load appropriately in the framework of the structural elements.

To realize an appropriate safety margin, it is necessary to understand a factor of destruction by measuring the deformation behavior in the full field of view. However, it is necessary to put on the strain gauges over the surface for full field measurement. It is expensive and time-consuming. Therefore, 3-dimensional strain measurement with large deformation is performed by using image based method. But, conventional products require severe conditions such as light distribution and image patterns for high precision measurement. Consequently, our research objective is to stabilize high precision measurement in un-uniform lighting conditions by image based method. The method with higher reproducibility and lower calculation cost is developed. Therefore, subpixel unit is modified by using different methods. Hence, the accuracy verification of subpixel unit is performed by the proposed method. An increasement of the correlation coefficient obtained by the proposed method is confirmed. It is assumed that the result has two reasons of the error. First reason is the difference in dot centroid position. The difference of the dot area occur, because xy-stage is not parallel. Second reason is a change in pixel value depending on how photoirradiation. The position of a dot is computed by the centroid of a dot. In short, the computations of the area and cross sectional primary moment largely influence the accuracy of the position. This is corrected by examining the relationship between the direction vector obtained from the multivariate normal distribution and the area change amount.

Meteorological data assimilation is formulated as a large-scale nonlinear optimization problem. At each search, the optimization algorithm requires the gradient of a cost function. The gradient is computed efficiently by the adjoint program. A source computer program that simulates the atmospheric flow is differentiated in reverse mode by the automatic differentiation tool TAPENADE, and the adjoint program is generated. In this paper, we improve the differentiability of the source program in order to generate the more exact tangent program. The generated tangent and adjoint programs are validated by dot product test. Preliminary numerical experiments are presented for data assimilation.

In this work, uniaxial tensile tests of smooth round bar, biaxial tensile tests of plates, and tensile tests of blocks were simulated by using perfect elastic-plastic ordinary state based peridynamics. In the simulations, fracture strain which modified by stress triaxiality scalar state was employed as fracture criteria. As a result, the ratio between equivalent fracture strain of uniaxial tensile tests and that of biaxial tensile tests agreed with theoretical values.

This paper deals with a large-scale visualization of the numerical human body model. In order to treat a 16 billion degrees of freedom model, AVS/Express has been executed on the UV2000 of Information and Communications, Nagoya University. To reduce the load, the data necessary for visualization have been selected, and appropriate file formats have been chosen.

Long-term flood damage occurred during the 2011 Tohoku Region Pacific Offshore Earthquake Tsunami, which hindered rescue and restoration activities. To reduce long-term flooding damage, touhening such as earthquake resistance of drainage facilities including sewage is indispensable. In this research, we investigated the limit of the drainage function by the existing facilities against the supposed tsunami of the Nankai Trough massive earthquake in the Sumiyoshi area of Tokushima City. Tsunami calculation code JAGURS was used, and drainage analysis was performed using flooding analysis software AFREL for calculation inside the dam. At the same time, we evaluated the parallelization performance of tsunami calculation by JAGURS.

This paper deals with three-dimensional non-steady eddy current analysis of a rotating machine as an example of a parallel finite element analysis including moving bodies. To reduce computation time, analysis domain is often reduced using a periodic boundary condition. In general, high efficiency in parallel computing with the periodic boundary condition is difficult to achieve. The hierarchical domain decomposition method (HDDM) is known as an efficient parallel finite element method. In this paper, a new domain decomposition technique for the HDDM that enables us to achieve efficient scalability on massively parallel computers is proposed.

An advanced smoothed finite element method using 10-node tetrahedral elements (SelectiveCS-FEM-T10) is implemented in a general purpose finite element software: ABAQUS. The implementation itself with the user-defined element (UEL) subroutine is successfully conducted; however, it turns out that the UEL is less-useful for ordinary users due to the various restrictions of UEL and thus it needs a native implementation.

This study describes on an application of the machine learning for the particle initial placement problem of the DEM simulation. The particle initial placement problem is to find a regular or irregular packing structure of equal and unequal of spherical and non-spherical particles in a finite space. It is difficult to develop a robust packing algorithm for a three dimensional arbitrary domain. In this study, a particle initial placement problem is regarded as the multi-agent simulation and solved by using the reinforcement learning. At first, a case of single-particle is considered. A single-particle problem has the Markov property and then can be utilized bootstrapping approaches such as a deep reinforcement learning algorithm with the Q-learning. Perhaps, a multi-particle problem is assumed to have the non-Markov property. This study demonstrates numerical examples of the particle initial placement problem by using a deep reinforcement learning techniques.

Peridynamics has some advantage for fracture analysis. Generally, to reduce computational cost, fine mesh is employed around stress concentration region while coarse mesh is employed the external region in FEM. Very fine model is needed to analyze stress concentration problems in PD because general PD formulation is employed uniform horizon. To overcome the difficulties, dual-horizon peridyamics (DHPD) was proposed. DHPD can analyze problems using varying horizon. We investigated the DHPD affects numerical accuracy, then we propose the new modeling for reducing numerical oscillation.

Recently, particle methods such as the SPH method and the MPS methods have been studied for the analysis of free-surface flows. This study focuses on the initial placement of particles in the MPS method. Currently, the initial particle placement on the regular Cartesian grid is widely used. However, it is difficult to represent slopes and curved surfaces of the wall boundary and then has negative effects on the precision of the numerical calculation. Therefore, we propose the initial particle placement determined by using the centroid Voronoi tessellation. Using proposed initial particle placements, some numerical examples with curved wall boundaries are demonstrated.

The breath is to inhale and exhale the air from the nose or mouth to the lung through the throat. Swallowing is to transport food from the mouth to the esophagus via the throat. That means the throat has an intrinsic risk to be obstructed by food known as suffocation or to transport food to the lung as a mistake occurring aspiration pneumonia. In order to avoid both of suffocation and aspiration, breathing and swallowing must act under a precise coordination mechanism to fulfill each function. However, the mechanism has never been revealed yet, since even the latest medical imaging technology cannot visualize the whole motion of the breathing and swallowing at all. The purpose of this study is to produce a numerical simulation model that integrates breathing with swallowing and is to visualize the whole movement of the breathing and swallowing to clarify the coordination mechanism. First, we combined two numerical models of Lung 4Cer CFD for breathing and Swallow Vision® for swallowing. The two models were connected at the larynx of Swallow Vision®. Then the connected model had anatomical structures of the mouth, throat and esophagus as well as trachea, bronchi and lung. To validate the connected model, the lung movement and the airflow during breathing were analyzed. As a result, the flow rate of inflow/outflow surface was confirmed to equal the volume change of the lung.

This paper proposes an isogeometric boundary element method (IGBEM) for three-dimensional doubly-periodic electromagnetic scattering problems where the scatters are dielectric materials, while the previous work considered the perfectly electric conducting scatters. We validate the accuracy of the IGBEM through numerical examples, emphasising the frequency response.

It is important to predict the thermal fatigue life of the wire bonding joint used in power modules. In this study, creep characteristics of aluminum wire which was actually used for power modules were obtained by stress relaxation test. Then, the creep characteristics were used to evaluate the thermoelastic plastic creep behavior of wire-chip joint of power modules. It was found from the FEA results of wire-chip joint during thermal cycle that the stress relaxation caused by the transient creep occurred in aluminum wire.

A method for arterial structural analysis with non-uniform rational B-spline (NURBS) was introduced.⁽¹⁾ We now use T-spline discretization to keep high continuity even with branches. Here, a good parametrization is one of the challenges. In this paper, we use the principal curvatures to reparametrization.

It has been recently recognized that Model Based Development (MBD) could improve development efficiency, and it is now becoming a standard approach especially in automotive industries. In this current research, MBD is adopted into the design process of steam turbines, and an automated design has been put into practice. First of all, optimization of design parameters is carried out using 1DCAE for the purpose of maximizing the efficiency of steam turbines. Secondly, based on the result of the former step, the shape of turbine blades is optimized using CFD. By introducing CAE and optimization methods into different stages of the design process, optimized design and cost down are realized simultaneously.

Silver sintering bonding via reduction reaction of Ag₂O was developed as a direct metal-ceramic bonding technology for next generation power devices. However, the joining mechanism is not clarified. In this study, the mechanical behavior of joint interface was investigated by simulations and experiments in order to research the bonding mechanism between silver and alumina. Multiscale analysis using molecular dynamics (MD) and finite element method (FEM) were used for simulations. In MD, cohesive properties between silver and alumina was investigated by tensile test simulations. The cohesive properties were given for FEM analysis which executed the delamination test. For a validation of simulation results, and mechanical properties during delamination of interface was obtained by nanoindentation test on microscale cantilevers.

In order to keep a safe operation a light water reactor plant, it is necessary to evaluate integrity of a pipe under the assumption of seismic load and a behavior of a crack propagation in a pipe under the load more than the design load. In the previous research, a validity of crack propagation criterion using stress triaxiality and equivalent plastic strain under monotonic loading case is exhibited. However, it is necessary for the condition to consider cyclic loading case to be applied. Objective of research is studying fracture criterion using stress triaxiality and equivalent plastic strain for cracked pipes under cyclic loading comparison with J integral range ΔJ which is existing fracture parameters. For development of fracture criterion, FE analysis using fracture experiment of cracked pipe by CRIEPI is conducted. The conditions of an analysis is almost the same as T47 report by CRIEPI. Stress triaxiality and equivalent plastic strain simultaneously increase as a crack propagates. Because stress triaxiality and equivalent plastic strain with crack propagation show the same trend, the summation of these parameters doesn't become constant value or linear relationship. ΔJ was evaluated excessively in early crack propagation, but it showed the same trend as the experiment as crack propagates. This is caused by the differences between the relationship, which is load point displacement and crack length, and an experiment condition.

Crack growth in the elastic-plastic fracture is an important issue for structural integrity, because seismic wave causes low cycle fatigue in the engineering structure. Many researchers have worked for many experiments and numerical analyses, however general criterion cannot be developed until now. In order to develop a three-dimensional fracture criterion, the fully automated and state-of-the-art FE crack growth simulation should be realized. CT specimen is used for evaluation of crack propagation under Mode I condition. Fatigue experiment, tension and compression experiment, was conducted. An FE analysis is also computed based on the experiment. Quantitative trends of FE analysis result agree with the experimental result well. Maximum load of FE analysis is higher than the experimental one. This is reason why measurements of crack front shapes and material nonlinearity are not enough to reproduce load and load line displacement. The obJectives of this research are to obtain a strain range for determination of cyclic stress-strain relation and precise fracture surface including crack front shape, share-lip and unstable fracture surface. Chaboche model isn't enough for reproduction of low cycle fatigue using CT specimen, because Chaboche model agrees with an experimental stress-strain relation in a limited strain range. Therefore, as Chaboche model is used for the simulation of low cycle fatigue, we have to find an applicable strain range to determine coefficients of Chaboche model.

The contribution of this study is development of ductile and brittle fracture simulation using cohesive zone model and Gurson model depending on temperature. The Gurson model and the cohesive zone model are combined by cohesive-traction embedded damage model⁽⁵⁾ to realize ductile and brittle transition. The Gurson model is employed with a criterion of void nucleation proposed by Kikuchi and Sannoumaru⁽⁴⁾ to represent complex fracture behavior of ductile failure in three-dimension. On the other hand, the cohesive zone model proposed by Rice and Wang⁽⁷⁾ is utilized for propagation of brittle cracks. Furthermore, the change of the critical cohesive traction and the initial yield stress with temperature is approximately determined by experimental results. Throughout the numerical examples, the proposed model enables us to realize the ductile to brittle transition behavior depending on temperature.

In this paper, we developed a new self-excited force sensor with a tunning-fork-beam for measurement of inverse problem. The proposed force sensor has the following characteristics: 1) high heat resistant 2) wireless measurement 3) potential for multiplex 4) possibility of miniaturization The key part of the proposed sensor is the tunning-fork-beam whose frequency changes under tensile force and undergoes self-excited vibration by the flow of air. Firstly, the principle of the proposed sensor is explained. Secondly, the simple theoretical analysis is performed. In order to verify the principle of the proposed sensor, the prototype sensor was built with acrylic material, then experiment was conducted with the prototype sensor. The spectrum of recorded sound was analyzed and the vibration of the sensor with outstanding frequency was observed. Measured relationship between the tensile load and frequency was shown to be consistent with the result of theoretical analysis.

Recently there has been growing interest in traffic because of aggravation of traffic congestion and change in traffic interrelationship. Advanced evaluation is necessary when introducing new policies or technologies in the traffic environment and traffic simulations that reproduce traffic phenomena on computers have been used as a substitute for social experiments in the real world. In this research, we aim to parallelize the microscopic traffic simulator with dynamic load balancing to realize traffic simulation with both refinement and comprehensiveness.

In this paper, dynamic characteristics of Shimajiri Mahji that is a unique soil to Okinawa's main island, were clarified by visualization experiments with ultrahigh-speed camera on the plane wave generated by detonation and the interaction field at the target soil interface. Next, numerical simulation of the detonation experiment using commercial finite element analysis software was carried out and the validity of the soil dynamic characteristics clarified from the experiment was confirmed. Furthermore, in order to clarify the fragment behavior when explosive the unexploded bombs, we have constructed the simple numerical simulation models for soil surface and underground explosion problems by SPH scheme. From the series of computational results, we have confirmed the fragment behaviors is significantly dependent on the explosive amounts and the depth buried of unexploded bomb.

It is well known that the mechanical properties of metal crystallographic materials strongly depend on the micro structure such as the movement of dislocations and their accumulation. There are reports calculating the dislocation density tensor by crystal plasticity theory using local crystallographic orientation by SEM / EBSD method that is, however, the apparent dislocation density calculating two dimensional information. In this research, to visualize the three-dimensional defects field, by introducing a deformation field into a silicon wafer by indentation and performing serial sectioning by chemical polishing, crystallographic orientation maps were obtained and the three dimensional defect fields was reconstructed by calculating the gradient in the depth direction.

We developed a method to identify concentration-dependent characteristics of an electrochemical reaction. Concentration and reaction rate distributions on electrode surface cannot be measured directly and a sum of the reaction rate distribution is observed as a current. Therefore, an inverse analysis approach is applied to identify the concentration-dependent characteristics. The unknown characteristics are expressed with a piecewise linear function with unknown parameters. Concentration field and velocity field are modeled with the advectiondiffusion equation and Navier-Stokes equation. These equations are solved by the finite volume method. In order to estimate the unknown parameters, a residual between the observed current and calculated current is minimized by the Nelder-Mead method. The proposed method is applied to a reduction reaction of a copper ion. The currents calculated from the identified characteristics are good agreement with the observed value. This result shows that identified characteristics approximate the true characteristics.

A large-scale aerodynamics simulation for a running group of bicycle is performed. We employ lattice Boltzmann method with cumulant collision operator to improve numerical stability of the turbulent flow calculation. We also adopt coherent-structure Smagorinski model for eddy viscosity of LES. These methods are implemented with a refined mesh to enhance computational efficiency. Each calculation of the cyclist(s) took one or two day(s) to get the result of 4 seconds in physical duration. We calculate and validate the flow around the single cyclist and a group of four cyclists, and we confirm that the result is suitable compared with the previous wind tunnel experiments. The performance of the calculation is 73% efficiency from 8 GPUs to 128 GPUs in weak scaling. These results show that our method is suitable to estimate aerodynamics of a running group of cyclists.

When a steel undergoes a heat treatment, the maximum strength appears in the middle of transition of carbon-atom states from the solid solution to the precipitation states. Recently, the latest observation technic has revealed the carbon cluster state where the steel shows the maximum strength. However, the strengthening mechanism by the carbon cluster has not been fully understood yet. In this study, to investigate the interaction mechanism between dislocation and carbon clusters, we perform molecular dynamics simulations. We show the strongest carbon cluster's diameter, which well corresponds to the experimental value, and propose strengthening mechanisms for each carbon cluster with different diameters.

High-Entropy Alloys (HEAs) are defined as solid solutions with high mixing entropy due to mixing generally more than five constituent elements in an equiatomic or near-equiatomic fraction, show both high strength and high ductility. The high strength is possibly realized by lattice distortion caused by mixing atoms with different their radius; hence, the strength increases with the atomic radius ratio δ. However, from some experiments, it has been also reported that single-phase solid solutions of HEAs transform into multiphase to amorphous with increasing of δ. These results show that the increase of δ is likely not only to make HEAs stronger, but also to change HEAs amorphous alloys. The purpose of this research is to establish the atomic modeling of HEAs, which is necessary to investigate the strengthening mechanism of HEAs via atomic simulations. We use five constituent-elements systems with different atomic radii to express various values of δ. As the result, we revealed that the proportion of crystal regions maintains near 1 until δ reaches 6, but amorphous regions increases with δ over 6. Furthermore, although the degree of mixing of the first neighbors in crystal regions is constant when δ < 10, the degree of mixing in amorphous regions tends to decrease monotonically regardless of δ. This result implies that the defects in HEAs consist of specific atomic groups.

Complex thermal deformation of the machine tool occurs with increasing heat generation inside the machine because of high speed rotation of the machine tool and asymmetry of the temperature distribution due to the slant structure of the bed portion, etc. However, it is very difficult to predict thermal deformation during operation in the multitasking machine tool having many parts and heat sources. In this study, based on experimental data, the temperature rise process of the actual machine tool and the thermal deformation of the spindle were investigated. The finite element model of the multitasking machine tool with turrets considering bed was made. In the heat conduction analysis, the calorific value of each heat source was estimated and the temperature distribution of the machine tool was obtained. Furthermore, thermal deformation analysis was carried out to get the inclination of the spindle due to thermal deformation and it compared with experimental one.

In the casting processes, in order to predict defect positions and to suppress the defect itself, it is important to clarify the flow process and the solidification process of the molten metal with numerical simulations. For this purpose, it is necessary to analyze the free surface flow and the moving boundaries. Moreover, since pressurization could be performed to reduce defects, it is important to represent the pressure boundary wall in simulations. Particle-based methods, such as the smoothed particle hydrodynamics (SPH) method, are considered to be effective to treat such conditions. Therefore we analyzed the casting process called gravity casting using SPH method and validated its usefulness by comparing with the experiments. We also analyzed the pressurization process with SPH method.

The objective of this study is to predict the fatigue lifetime of structural steel using cohesive-traction embedded elasto-plastic damage model with kinematic hardening. By the cohesive-traction embedded damage model, the elasto-plastic model is combined by introduction of local equilibrium conditions between principal stress and cohesive traction to realize actual crack opening displacement during cyclic loading. Whereas the cohesive zone model proposed by Rice and Wang is utilized with consideration of unloading process, an isotropic and kinematic hardening in the elasto-plastic model is represented by Voce and Chaboche model, respectively. Moreover, an experiment is also conducted with a round notched bar under the cyclic loading condition controlled by a constant amplitude of plastic strain. By the comparison between the numerical and the experimental result, it is confirmed that the proposed model can represent both two kinds of the hardening behavior and predict the fatigue lifetime under cyclic loading.

In material development processes, we often need to control material microstructure that contains metastable phases. For predicting microstructure evolution on the basis of theories and simulations, Gibbs energy parameters of constituent phases are needed. In this study, we propose a method to estimate Gibbs energy parameters of a metastable compound phase from experimental data on microstructure evolution by using phase-field method and adjoint method. We prepared synthetic data of microstructure evolution by a phase-field simulation and used it as experimental data. Twin experiments have shown that the Gibbs energy parameters of the metastable phase can be successfully estimated from the information on microstructure evolution.

In the sports and healthcare technology field, electrocardiogram measurement using the wearable devices has attracted attention. A textile electrode is one of the electrodes used in the clothing type wearable device. The advantage of this electrode is that measurement of biosignals stable for a long period of time does not cause load or damage to the body. On the other hand, fracture or strength due to tension at the time of clothes is not clarified. Evaluation of strength and characteristics in tension caused by wearing is needed for stable electrocardiogram measurement. In this research, a basic examination was conducted to evaluate the strength and characteristic of textile electrodes used for Smart-wear capable of electrocardiography measurement. In order to achieve this evaluate, a measurement system for strength and characterization for textile electrodes was developed. The textile electrode was pulled 50mm at a constant speed of 5mm/s, and the change in electric resistance and elongation wear measured at the same time. It was confirmed that resistance increased nonlinearly with the increase of linear displacement. This result shows that the developed system has applicability of measurement and evaluate for textile electrodes which are high flexibility.

We have conducted numerical simulations of fluid flow in stirred vessels using the Moving Particle Semi-implicit (MPS) method, which is one of the particle methods for incompressible flow. The MPS method originally calculates wall boundary using particles. However, the wall particle method has difficulties in changing boundary shape and spatial resolution. Therefore, in this study, we applied two types of boundary treatments that use polygon mesh for the wall boundary representation, namely the polygon model developed by Harada et al. (2008) and the integral wall model proposed by Matsunaga et al. (2018). We simulated laminar flow stirred by six-blade Rushton impeller in cylindrical tank. The calculation results are compared with those from experimental measurement in literature. It is found that the integral wall model exhibits higher accuracy and higher computational stability.

The tongue deformation is dependent on variation of myofiber orientations among individual subjects. In this study, to investigate the influence of myofiber structures on the tongue deformation, mechanical tongue model considering the personalized structure of tongue myofiber is proposed. The reference geometries of tongue and jaw are constructed from the medical image of magnetic resonance imaging (MRI) for a Japanese male adults. In addition, tongue myofiber orientations obtained from the medical image of diffusion tensor imaging (DTI) are considered for the model. The tongue deformation is calculated by solving a force-equilibrium equation using a continuous Galerkin finite element method with a hyperelastic material, where the deformation is driven by active contraction of the tongue myofibers. Comparison between DTI model and anatomical model shows that initial structure of tongue myofiber make a difference in deformed tongue shape and the proposed models including myofiber orientations of DTI has a potential to identify personally the tongue muscle contractions in speech production.

The wheel flat is the damage occurring between the wheel and the rail. The impact acceleration generated by the wheel flat affects the vehicle and railway track, but the influence of the impact acceleration has not been quantitatively evaluated. Therefore, in this paper, we consider the influence of the difference of analysis model on impact acceleration using three analysis models with wheel flat. As a result, no difference is found between elastic analysis and plastic analysis, but considering the car body mass, introducing a rail pad resulted in a decrease in impact acceleration. Also, looking at the distribution in the contact surface when the maximum acceleration occurs, the normal direction contact force is generated at a high level at the rear of the flat, and the tangential contact force is found to produce high force near the flat rear part and the front part.

In order to realize a low-carbon society, reduction of loss of power devices is the urgent problem to be solved. An essential factor for achieving low-loss bipolar devices, such as Insulated Gate Bipolar Transistors (IGBTs) and pin diodes, is controlling the bulk lifetime of phosphorus (P) doped n-type silicon (Si) crystals. In order to do this, recombination centers, such as vacancy-vacancy (V-V) and vacancy-phosphorus (V-P) pairs are introduced by electron beam irradiation. One of the technological problems of the use for Czochralski (CZ) Si wafers is that the carrier lifetime is extended under the device function. This is probably due to the inactivation of deep energy levels of V-V and V-P pairs because of the interaction of interstitial carbon (C_i), interstitial oxygen (O_i), and C_i-O_i pair. The purpose of this study is to understand the mechanism behind the formation of recombination center (V-V and V-P pairs) and the complexes of C or O (C_s + I, C_i + O_i, C_i + C_s). To achieve this, first principles calculation was performed on possible reactions of intrinsic point defects (vacancy V and self-interstitial Si I), dopant (P), and impurities (C and O), which exist in the P-doped n-type CZ-Si crystal. The main results are as follows: (1) V and V (P) are most stable at the 1^st nearest neighbor and form V-V (V-P) complex. V-V interaction remains in the short distance while V-P interaction reaches more than 9^th neighbor. (2) C_i-O_i pair was easily formed and its formation process was clarified. E_b of C_i-C_s was lower than that of C_i-O_i.

In this research, we focus on the eigenfrequencies of the bending mode of a plate-like elastic body and perform topology optimization aiming at setting them to desired values. The topological derivative required for optimization of this problem is derived from a model approximated as two-dimensional one, assuming the thin plate to governed by the biharmonic equation. Therefore, in this problem, the design domain has a two-dimensional space. In the process of calculating the eigenfrequencies, the plate-like elastic body is modeled in three dimensions and is discretized for the finite element method. By solving the eigenvalue problem for the discretized model, the eigenfrequencies and eigenvectors of the plate-like elastic body are obtained. The eigenfrequencies to be optimized are searched from the obtained eigenfrequencies by using the inner products of the obtained eigenvectors and the virtual external forces. A series of topology optimization procedures are performed on the eigenfrequencies to be optimized. We use the level set method for topology optimization. Simple optimum design is carried out using the obtained topological derivative, and its effectiveness is confirmed.

In this paper, we present fluid mechanics computation of left ventricle including mitral and aortic valves. We use an interface-tracking method, which is based on the space–time slip-interface, topology change and isogeometric discretization methods.

The statistical properties of discontinuous crack propagations in crystalline materials under Mode I loading are investigated via deterministic and stochastic simulations. Molecular dynamics (MD) simulations show that the dislocation emission from the crack tip suppresses the crack cleavage propagation and this phenomenon is a cause of discontinuous crack propagations. The statistical properties of the discontinuous crack propagations observed in MD simulations show clear temperature and strain rate dependences. To investigate the controlling factors of the dependences, Monte Carlo (MC) simulations are performed. The MC simulations taking the evolution of stress field around the crack tip reveal that the controlling factors are the thermal activated processes at the crack tip.