This study extends the previous seven-grained embedded packet model for a high Cr heat-resistant ferritic steel to that composed of 23-grained multi-prior austenitic grains, for the purpose representing more realistic multiscale simulations based on FTMP, in terms of the inhomogeneous recovery-triggered accelerated creep rupture processes peculiar to the material systems of concern. The effect of interaction/recovery contributions are systematically examined, where the recovery model is represented by the spatio-temporal mixed component of the incompatibility tensor field in FTMP. Point-wise duality diagram representation of the simulation results reveals that the energy storing region, which needed to be additionally introduced in the previous model, is naturally brought about within the sample, implying one of the significant contributions from the “grain aggregate” scale in the present context.

In this study, the effects of the shape of lugs placed on the outsole of trail running shoes on their mechanical properties was investigated. The finite element method (FEM) analysis were conducted on several analytical models simulating the lugs to evaluate the effect of kinematical shapes on the stiffness of the lugs. The shapes of the lug were defined by setting the dimensions of the three sides and the two angles as parameters. Several models with uniform cross-sectional area were created. The aspect ratio, which is the ratio of the width to the length of the lug, was set to 3, 1, and 1/3, and the stiffness was evaluated for all combinations in which the angle of the V-shaped portion of the arrowhead shape and the tilt angle were varied by 10° between 30 and 90°, respectively. The stiffness of the lugs was evaluated in compression and shear, and the analytical results of each model were compared. The effect of the difference of lug shapes on the stiffness of the lugs was also investigated.

CFRP laminate subjected to out-of-plane load may have complex impact damage which may reduce the compression strength. This study aims to develop an analysis method that can simulate the damage propagation of CFRP laminate under out-of-plane indentation and predict the CAI strength. This paper shows the out-of-plane indentation and the subsequent in-plane compression test analyses using the FEM model of CFRP laminated plate considering the CZM. It was shown that the results obtained by the present method are appropriate.

This paper presents the study on cycle jump method for fatigue crack growth analysis using cohesive zone model. When the cohesive zone model is applied for the high-cycle fatigue problems, it is necessary to cost a large amount of computational time. On the other hand, the application of the cycle jump method into the fatigue crack growth analysis enables us to realize the fatigue crack growth analysis with a practical number of loading cycles. In order to conduct the cycle jump analysis with a sufficient numerical accuracy and acceptable computational time, we examined several types of the extrapolation methods to approximate damage variable that represents material degradation under cyclic loading. The crack growth rates obtained by the analyses using each extrapolation method are compared with that obtained by the cycle-by-cycle analysis, and it could be found that Richardson extrapolation gives the most accurate solution.

Nature-inspired stochastic search techniques such as Evolutionary Algorithms (EAs) are known for their ability to solve complex optimization problems. However, they typically require numerous function evaluations in their search process to find global optimums. This is a drawback if EAs are used in optimization problems that have computationally expensive functions. Surrogate models have been used as cheap approximations to replace these functions. In this study, we propose a dynamically retrained multilayer perceptron-based surrogate model that is coupled with a genetic algorithm (GA) to reduce the number of function evaluations in the optimization process. The proposed method has been successfully applied to some test functions and real-world aerodynamic shape optimization problems. It is also shown that it converges more quickly towards the Pareto-optimal front with fewer function evaluations compared to a stand-alone GA in all optimization problems.

In this study, the effect of metal interface with amorphous structure on mechanical properties is studied by molecular dynamics method. When modeling a metal interface, it is common to create the interface by adjoining separately created atomic models. In this study, atomic structures including an amorphous/amorphous interface and an amorphous/crystalline interface were created by quenching a model with Al and Cu from different temperature ranges. Then, tensile deformation simulations by molecular dynamics were performed for these models. As a result, the model with an amorphous/amorphous interface showed a uniform distribution of atomic strain regardless of the type of atoms or the interface. In the model with an amorphous/crystalline interface, dislocations in the Cu crystal were accompanied by a rapid relaxation of stress. In addition, the stress-strain diagram and the distribution of atomic strain were different depending on the crystal orientation of Cu.

In this study, molecular dynamics simulations were carried out on bundle structures of CNTs in order to obtain fundamental knowledge on the deformation characteristics of them. When CNT bundles are subjected to torsional deformation under periodic boundary conditions, the CNTs farther from the rotation axis are stretched more. Then, axial stresses have inhomogeneity in the twisted bundle structure of periodic boundary conditions. On the other hand, by adjusting the repetition length caused by the six-membered ring of CNTs according to the distance from the rotation axis, a torsional structure with less axial-stress difference can be created even in a periodic structure model. Thus, we investigated the stress state and structure of CNT bundles with different initial structures and analytical conditions.

Thermoplastic styrenic elastomers(TPS) are physical mixtures of polymers. They are used for automobile parts and sports products to suppress vibration because of its mechanical properties. In this study, CFRP / TPS sandwich composite materials with damping TPS sheet were developed, and the mechanical properties were investigated. At first, the biaxial tensile loading tests were conducted to investigate static mechanical properties of the TPS. The hyperelastic model was applied to representation of the experimental results. Next, the 3-point bending tests were conducted to investigate the bending characteristics of CFRP / TPS sandwich laminates. The test result showed that, the bending stiffness was decreased by the TPS in the CFRP specimens. In the FEM analysis based on the hyperelastic model, the bending characteristics of the CFRP / TPS laminated board in the low strain region could be reproduced.

This paper presents the liquid foam formation simulation by bubbles freely rising to liquid-gas interface. The flow is governed by isothermal weakly compressible Navier-Stokes equations under a low Mach number. A coupled Level-Set and Phase-Field method is adopted to capture the interface with high shape accuracy and good mass conservation. Tree-based Adaptive Mesh Refinement is used to improve the mesh resolution around the interface. The surfactant transportation is calculated by using Langmuir model. For high-viscosity liquid and gas, a fractional-step semi-implicit method is proposed to release the constraint to time step and improve the efficiency. Viscous cavity flows under different Reynolds number are simulated to validate the semi-implicit method.

This paper presents a simplified dislocation core structure model for dislocation dynamics (DD) simulations for dislocation precipitate interactions. The model is based on the discrete Peierls-Nabarro model, and deals with the dislocation core influence by a collection of 2-dimensional virtual fractional dislocations (VFD). Using the proposed model, the computational amount and time can be drastically reduced. The model is then applied to the DD simulation of the interaction between a super-dislocation and a g-precipitate in g’-phase. In the simulation, the super-dislocation is modeled with three different dislocation models, namely super-dislocation, perfect dislocation pair and two pairs of partial dislocations. The critical resolved shear stresses (CRSS) calculated by the proposed model and a conventional DD method with a discrete Peierls-Nabarro model are compared. The results suggest that the proposed model can provide the CRSS of the conventional DD method with the discrete Peierls-Nabarro model particularly using the perfect dislocation pair and two pairs of partial dislocations.

In this paper, we develop topology optimization for controlling thermal wave, which is a temperature wave in macro scale. Assuming a porous system in macro scale, the thermal wave is analyzed by the finite element method. The level set function that expresses the structure is optimized by the optimal solution search based on the covariance matrix adaptation evolution strategy (CMA-ES), and the temperature distribution of the thermal wave disturbed by the obstacle is cloaked by the topology optimization using the level set method.

This paper presents a numerical analysis method for surface crack problems based on the dislocation dynamics (DD) method. To satisfy the free surface boundary condition, the superposition principle is utilized. Virtual dislocation loops are placed at the outside of the body of interest to remove the dislocation singularity at the free surface. This study focuses on the numerical accuracy of the DD method for the surface crack problems. In the DD simulation, the number of dislocation loops and the size of the virtual dislocation loop for the surface crack is tunable parameters so that benchmark tests are performed to clarify the appropriate number of dislocation loops and size of the virtual dislocations. Finally, a surface crack interaction problem is solved with the DD method, which illustrate the potential of the DD method for further applications to complicated surface crack problems.

The flow past an archery bare shaft arrow is investigated at the angle of attack 0 degree, by using two kinds of finite difference code: an axisymmetric code with fourth-other accuracy in both space and time, and a 3D code with spatial second-order accuracy and temporal second-order accuracy. The Reynolds number, Re, based on the arrow shaft diameter is varied between 10,000 and 20,000, and the length of the shaft is chosen as 120. The grid dependency of the computed velocity field, the value of drag coefficient Cd, boundary layer thickness are studied. The result of comparison and assessment of numerical errors confirms that the computed velocity field is acceptable for the purpose of analysing the stability properties of the boundary lager flow along the arrow shaft.

In micro-electro-mechanical systems using piezoelectric bimorph actuators such as FWNAV, small displacements given by the actuators can be magnified using a transmission using the flexible structure. Hence, in these systems, coupling among the direct and inverse piezoelectric effects, and the flexible structure occurs. In this study, a partitioned iterative method is proposed based on the block SUR algorithm for this strongly coupled problem, and the basic performances are demonstrated in numerical examples.

This research shows some applications of the Smoothed Particles Hydrodynamics method for hyperelastic-plastic materials subjected to collisions. We utilize the total Lagrangian description for particles under the hyperelastic-plastic regime, while change it to the updated Lagrangian description (with non-Newtonian fluid rheology) for particles under very large accumulated plastic deformation. We also developed a simple yet robust contact algorithm, which relies solely on the internal elastic forces due to inertia to generate any bouncing behavior. This contact algorithm generates an intrinsic energy loss during the impact, which the current authors consider to be a convenient inaccuracy. A simple numerical experiment of two colliding spheres is then conducted to assess this impact-generated energy loss, as well as to show that our method is capable of intentionally dissipate the impact energy through plastic deformation. Finally, we show the full capability of the current method in a catastrophic landslide of a soil mass with randomly distributed large rocks boulders inside.

Airfoil shape, solidity, and number of airfoil blades are optimized for small vertical axis wind turbines (VAWTs) with the Kriging model. VAWT is one of the wind turbine types and has many advantages, e.g. flexibility for wind direction. The VAWT rotor blades are optimized by a global optimization approach based on a surrogate model. As the surrogate model, the ordinary Kriging model is utilized. The promising configuration is explored with the expected improvement criterion. In the optimization, the VAWT performance is evaluated by two dimensional (2D) unsteady computational fluid dynamics simulations using ANSYS Fluent. The rotor having two blades has greater performance than that having other numbers of blades under the specified design condition. The optimal configuration of the rotor is really manufactured using a 3D printer and then its performance is evaluated experimentally.

従来の人間運動計測はパフォーマンス低下や敷居の高さなどの問題が存在していた。本研究ではより簡便な計測法の探索を目的として、映像から深層学習により姿勢推定を行うOpenPoseを用いてゴルフスイング計測を行った。姿勢データをサウンドに変換することで姿勢の状態を可聴化した。前提知識を持たないゴルフ未経験者でも効率的な練習を行うことが可能となる。

In this paper, we present about the structural analysis of mitral valve with valve contact. It is difficult to predict where occur valve contact on mitral valve because it has two cusps connecting smoothly. To allow valve contact at any position, we use penalty method. This method can evaluate the contact force depending on the difference of two integral points.

It is known that the dynamic behavior of polymers depends greatly on not only the strain rates but also the hydrostatic pressure, and the volume change is caused during plastic deformation. It is desirable to apply material constitutive model considering these mechanical properties in numerical simulation for polymers. Therefore, we developed a new constitutive model for polymers and implemented using the user-subroutine of the impact analysis code LS-DYNA. On the other hand, the plastic Poisson’s ratio during plastic deformation of ABS polymer was measured by Mimura et al.

In this study, we focused on the volume change during plastic deformation of ABS polymer, and performed numerical simulation considered Poisson’s effect during the plastic deformation using LS-DYNA to verify the validity of this constitutive model and material parameters.

In design of aero-engine fan blade, it is required to improve the performance such as efficiency and pressure ratio with considering several constraints such as structural strength, resonance avoidance, surge avoidance, and so on. However, it is difficult to find solutions that satisfy all constrains. In this research, feasible and improved solutions than existed solutions are found by multi-objective optimization using genetic algorithm with multiple constraints ranking.

The design and optimization of multi-material structure is attracting attention, because there is a possibility that the performance of the entire structure, which is difficult to achieve with a single material, can be realized by combining multiple materials effectively in one structure. In this research, in order to solve the difficult problem of improving the deformation characteristics as well as reducing the weight of the electric current-driven thermal actuator, density method - based topology optimization considering Joule heat generation, heat transfer and thermal expansion is investigated to determine the distribution of three types of materials including cavities. A validation study is also performed using geometry extracted from the optimization result. Results show that topology optimization is useful to find non-intuitive design of multi-material structure.

This study addresses hierarchical modeling of high Cr steels with lath martensite structures yielding the inhomogeneous recovery-triggered accelerated creep rupture at relatively lower stress conditions, based on field theory of multiscale plasticity (FTMP). From dawn-to-dusk scheme that involves natural modeling of the composing lath block units as well as building-block type constructions of the lath packet structures is proposed, together with the interaction field-based recovery modeling between high dense dislocations in the lath (Scale A) and the lath block structure (Scale B). The contributions of the spatial interaction from Scale A to B via the incompatibility tensor field is extensively discussed in terms of its projection direction in introducing in the hardening law of the constitutive equation used. It turns out that the slip normal direction as the projection direction exhibits the primary role for speeding the localized recovery ultimately leading to the accelerated creep rupture.

Acoustic cloaks that can cloak an object from both air- and water-borne sounds are designed by a topology optimization based on covariance matrix adaptation evolution strategy (CMA-ES). Two objective functions for cancelling the scattering of both air- and water-borne sounds are minimized and the optimum sets of level set functions which express the structures of acoustic cloaks are explored for the scattering cancellations. Acoustic-elastic coupled problems are solved by finite element method for evaluating the performance of acoustic cloaks.

The contribution of this study is to assess dispersion of fracture toughness in ductile-brittle transition region due to interaction of uncertainties between ductile and brittle failure in a metallic material by cohesive traction embedded damage-like constitutive law based on hyperelastic model. For realizing the complex fracture behavior in the ductile-brittle transition region, the cohesive traction embedded damage-like material constitutive law, which has been proposed by combining Gurson-Tvergaard-Needleman (GTN) model with cohesive zone model, is extended into hyperelastic model. GTN model enables us to represent the shrink of the yield surface depending on the void volume fraction. On the other hand, the cohesive zone model realizes stress release process due to the brittle crack propagation after the stress reaches tensile strength. The probability distributions of the initial void volume fraction and the tensile strength are approximated by normal distributions. Several numerical examples demonstrate that our developed damage-like material constitutive law with stochastic collocation method enables us to predict the dispersion of the fracture toughness in the ductile-brittle transition region.

The objective of this study is development of the cohesive traction embedded damage-like constitutive law using the micromorphic approach to improve dependence of finite element size. To develop the cohesive traction embedded damage-like constitutive law, the invariants of the separation-induced strain are selected as the micromorphic variables according to the conventional micromorphic approach. On the other hand, the mricromorphic variables are introduced into the local equilibrium equations between the principal stress and the cohesive traction by replacing the invariants with the eigen values because of the equivalence between them. Moreover, the additional governing equations are obtained by the introduce of the micromorphic variables into the equilibrium equation of internal forces and the free energy function. As a result, the relation between the microscopic variables and the micromorphic variables in the governing equations provides us to realize non-locality without dependence of mesh size. The capabilities of the cohesive traction embedded damage-like constitutive law developed by the micromorphic approach are demonstrated throughout the simple numerical examples.

Since the Great East Japan Earthquake, the share of thermal power generation in the electricity supply in Japan has been about 80%, and there is a need to reduce the environmental impact of thermal power generation. In this study, we describe the development of a large-scale parallel thermal stress analysis system on Fugaku to evaluate the structural integrity of a coal gasifier, which plays an important role in realizing a coal gasification combined cycle power plant that significantly reduces greenhouse gas emissions.

In order to create a product with Kansei value, it is necessary to design a product that has both Kansei value, i.e., design in shape and functionality, such as strength. We developed the shape and layout imagery creation system that uses brain function information obtained from BMI (Brain Machine Interface) to make decisions about the shape and layout of a product based on the designer's initial design image. Since this system is greatly affected by artifacts due to the EEG measurement using BMI, it is necessary to customize the system for each designer. Therefore, for the assessment of customization of the system, we introduced the validation of the system with uncertainty of artifact through ASME V&V 10-2006 Guidelines based on V-model with V&V. In this paper, we present an overview of this system and its uncertainty evaluation method.

While proceeding with the examination of Level 3, we came to think that the concept of cooperative control is effective not only for Level 3 but also for all levels. In Level 3, when the system cannot stand, the driver receives Request to Intervene(RtI) from the system. This will be only possible if the system itself understands why it requests RtI to the driver and the system can the driver what to do. To achieve this situation, It is important for the machine learning to be short calculation time and to get smarter with training. Extended Laplace transform type machine learning HNN(Holographic neural network) is the one that meats these requirements.

The effective regularization method for the inverse problem of identifying cracks inside a structure is developed. We consider the inverse problem whose measurement data is surface deformation. The Joint Estimation-Maximum a Posteriori (JE-MAP) method that considers the physical constraints to be satisfied by the displacement and stress on the crack surface. In this method, Maximum a Posteriori (MAP) method taking account of the prior and GrabCut (GC method), which is an image segmentation algorithm, are performed alternately. In order to demonstrate the effectiveness of the proposed method, the numerical example problem are analized.

In this paper, strength estimation of heterogeneous materials by peridynamics considering random inclusion location variation is discussed. Peridynamics is a method of simulating damage evolution using a non-local failure model. The finite element method is generally used to analyze stresses in composites, it is difficult to simulate damage propagation using the finite element method because of the complexity of damage propagation due to multiple cracks generating from the interface between the reinforcing material and the matrix material. Therefore, the applicability and effectiveness of strength estimation by peridynamics will be investigated for unidirectional fiber reinforced composites.

Introducing DEXCS for preCICE, which makes OpenFOAM / CalculiX coupled analysis by preCICE "easy and ready to use by anyone". Using this, it will be possible to utilize it as a common platform among researchers and to apply it to problems targeting existing structures.

This paper presents a method to reduce the computational time required to solve a shape optimization problem of a hyperelastic body. In the previous report, authors presented a method to use a technique of model order reduction based on the Karhunen-Lo`eve expansion (KLE) in a shape optimization problem of a linear elastic body. In that case, we applied this method only to solve a linear state determination problem. In this paper, we show that this idea is applicable not only to the adjoint problem and the problems seeking domain variations by a gradient based method called the H¹ gradient method but also to nonlinear state determination problem. The method is explained using an end mean compliance minimization problem of a three-dimensional hyper elastic body.