This paper describes a stochastic homogenization analysis of a particle reinforced composite material using an approximation technique. In order to analyze the influence of a microscopic random variation of an elastic property of a component material on the homogenized elastic property of a particle reinforced composite material, the Monte-Carlo simulation is employed. Since the conventional Monte-Carlo simulation sometimes involves a higher computational cost, an approximate stochastic homogenization method using the Monte-Carlo simulation combined with a polynomial-based approximation technique is employed, and accuracy of the approximate Monte-Carlo simulation is investigated. In order to apply a lower order approximation to the approximate Monte-Carlo simulation effectively, the weighted least square method is proposed, and its effectiveness is discussed with the numerical results.

View full abstract

The dynamic response of non-uniform functionally graded beams subjected to a variable speed moving point load is studied by using the finite element method. The material properties of the beams are assumed to be graded in the thickness direction by a power law. A beam element, taking the effects of shear deformation, cross-sectional variation and the shift in the neutral axis position, is formulated by using exact polynomials obtained from solutions of the governing differential equations of a homogeneous Timoshenko beam element. The dynamic responses of the beams are computed by using the implicit Newmark method. The numerical results show that the dynamic characteristics of the beams, including the maximum mid-span deflection, mid-span axial stress distribution are greatly influenced by the acceleration and deceleration of the moving load. The effects of the moving speed, material non-homogeneity, cross-section variation as well as aspect ratio on the dynamic response of the beams are investigated in detail and highlighted.

View full abstract

Design Informatics has three points of view. First point is the efficient exploration in design space using evolutionary computation. Second point is the structuring and visualizing of design space using data mining. Third point is the application to practical problems. The investigation of efficient evolutionary-based optimizer for the above first point is essential in order to generate hypothetical database for data mining. In the present study, the performance regarding diversity and convergence has been compared among pure and their hybrid methods using three standard mathematical test problems with/without noise. The result indicates that the hybrid method between the genetic algorithm based on the elitist non-dominated sorting genetic algorithm and the differential evolution is better performance for efficient exploration in the design space under the condition for large-scale engineering design problems within 10² order evolution at most. Moreover, the comparison among eight crossover indicates that the principal component analysis blended crossover is good performance on the hybrid method between the genetic algorithm and the differential evolution.

View full abstract

A realistic wall boundary treatment method for viscous flow computations using Cartesian grid method is proposed to simulate compressible flows around complex geometries at low and high Reynolds numbers. In this method, a Cartesian grid is used as a background grid to cover the whole computational domain and a wall boundary is treated by gridless method to express a smooth wall boundary of input geometries in a Cartesian grid. Subgrid which is similar to a prismatic layer in conventional unstructured grid method is introduced near the wall boundary to resolve a boundary layer. The proposed method is applied to the Building-Cube Method (BCM) which is a block-structured Cartesian grid approach proposed by one of the authors. Capability of the proposed method is demonstrated through several two-dimensional test cases.

View full abstract

We proposed a new improvement of self-adaptive strategy for controlling parameters in differential evolution algorithm (ISADE). The differential evolution (DE) algorithm has been used in many practical cases and has demonstrated good convergence properties. It has only a few control parameters as number of particles (NP), scaling factor (F) and crossover control (CR), which are kept fixed throughout the entire evolutionary process. However, these control parameters are very sensitive to the setting of the control parameters based on their experiments. The value of control parameters depend on the characteristics of each objective function, so we have to tune their value in each problem that mean it will take too long time to perform. We present a new version of the DE algorithm for obtaining self-adaptive control parameter settings that show good performance on numerical benchmark problems and constrained engineering optimization problems.

View full abstract

Recent studies of the flapping flight of insects have succeeded in solving the unsteady aerodynamics of hovering and contributed to realizing bio-inspired micro aerial vehicles (MAVs). However, the effect of wing deformation on the aerodynamics has not been investigated because of a lack of appropriate analysis methods. As an initial step to creating a “total” simulator for flapping flight, we developed a free-flight simulator by combining fluid-structure interaction finite element analysis based on the arbitrary Lagrangian-Eulerian method, which can quantitatively treat the strong interaction between the wing deformation and its surrounding airflow, and a rigid body dynamics analytical solver. With biologically-inspired flapping motion, which mimicked the changes in the stroke motion of the wing, the numerical model of the dragonfly performed rapid turning over 1200°/s of yaw angular velocity. Although the flapping motion for the left wing on the trigger flapping and the right wing on the resumed flapping (or its inversed combination) are identical, a considerable difference in the deformation of the wing during this identical flapping between the former and latter halves of the turn was observed. Thus, while these actuations were identical, the directions of the aerodynamic forces were largely controlled by passive deformations of the wings. These results meant that the effect of wing deformation on its aerodynamics should be taken into account and thus fluid-structure interaction analysis is required to effectively design the actuation of the wing on an artificial MAV.

View full abstract

A MEMS-viscosity sensor with dual spiral structure has been developed based on a novel measurement method unlike traditional method for viscosity measurement. This viscosity sensor, passing completely through a silicon wafer using MEMS (Micro Electro Mechanical Systems) fabrication processing, was made up the spiral structure with a vibrating body and a sensing body. When a large deflection was generated toward each direction of the vibration body of spiral structure due to the applied external load, it can be the cause of performance deterioration of the MEMS-viscosity sensor. And the analytical methods were proposed to fine the deflections by a computational analysis using FEM (Finite Element Method). A computational analysis was carried out by coupled analysis using modal analysis method and harmonic response analysis assuming air environment. In the spiral model of the MEMS-viscosity sensor on which normal loads of F_z=100 [µN] and F_z=1000 [µN] were applied, the maximum resonance points occurred at about 1400[Hz] of 1st mode along with vertical direction (Z-direction) respectively. When F_z=1000 [µN] was applied, the maximum value of deflection was obtained about 3.0x10^-4 [m] in vertical direction, but the normal load of more than F_z=1000 [µN] should be avoided for safety and reliability of this MEMS-viscosity sensor because the deflections of horizontal directions (X and Y-direction) were near to limit of the design values. We found that the deflections of horizontal direction were small enough to be negligible compared to the vertical direction, and the spiral structure can be stably maintained against less than F_z=1000 [µN] of the external load. We also found out the waviness phenomenon in the deflections along the spiral beam. As a solution of the waviness phenomenon in the spiral structure with rectangular shape beam, it seems that the spiral beam must be given so that the spiral radius is continuously compensated by using other beam shapes. It was demonstrated that the approach using computational analysis allows us to deduce visually the deflection of the spiral structure for viscosity sensor.

View full abstract

This paper discusses the multiscale stochastic stress analysis of a resin-based porous material having a quasi-periodic microstructure. For the purpose of the analysis on the influence of a microscopic geometrical random variation of holes, the perturbation-based multiscale stochastic stress analysis method is employed. As the microscopic geometrical random variables, the volume fraction, shape and location of the holes are taken into account, and in order to compute the perturbation term of the microscopic stresses for the microscopic geometrical random variation, the finite difference method is used. With the numerical results of the multiscale stochastic stress analysis of the porous material, the probabilistic response of the microscopic stresses for the microscopic randomness is investigated. Also, the accuracy of the perturbation-based approach is discussed in comparison with the results of the Monte-Carlo simulation.

View full abstract