A system of rigid bodies liked together at joints is called a link mechanism. Motion of the link mechanism is modeled as an initial value problem of differential-algebraic equations constructed from equation of motion of each rigid body and motion constraints. In the previous paper, the solution to the shape optimization problem of the rigid bodies with which an initial value problem is define was presented. In the shape optimization problem, the objective function to maximize was constructed from the external work done by a given external force, which agrees with the kinetic energy of the link mechanism, for an assigned time interval. The total volume of all the links formed the constraint function. The Frechet derivatives of these cost functions with respect to the domain variation, which we call the shape derivatives of these cost functions, was evaluated theoretically. A scheme to solve the shape optimization problem was presented using the H^1 gradient method (the traction method). By the program originally developed, a numerical example of shape optimization for a crank mechanism was shown. In the present paper, we examine to put on replacing the solver of the initial value problems by a general-purpose CAE software, and becoming solvable of the program to various problems of link mechanism. For a simple pendulum under the gravity, the shape variation that the center of mass moves to the tip of the pendulum is observed. Moreover, for a crank mechanism, the similar result to the previous study is obtained.

View full abstract

This paper presents a numerical solution to shape identification of unsteady natural convection fields to control temperature to a prescribed distribution. The square error integral between the actual temperature distributions and the prescribed temperature distributions in the prescribed sub-domains is used as the objective functional. Shape gradient of the shape identification problem is derived theoretically using the Lagrange multiplier method, adjoint variable method, and the formulae of the material derivative. Reshaping is carried out by the traction method proposed as an approach to solving shape optimization problems. The validity of proposed method is confirmed by results of 2D numerical analysis using FreeFem++.

View full abstract

The author considers the issues about the robust optimization raised by Prof. Arakawa at the last JSME D & S symposium in 2014. The most known definition of the robust design is that variation of the objective function value is the smallest under the variations of the design variables as well as the minimization of the objective function values. Many researhers have formulated that the objective function for the robust desig as the weighted sum of the deterministic objective function value and the variations. Then, many researches concerning the robust design would like to emphasize the difference between the deterministic and robust optimum design to show the importance of considering uncertainties. On the other hand, his raised issue is that there exists the problem that the robust optimum design is largely shifted by slightly change of the design conditions according to the robust design definition above. The large change might be desirable under the usual definition of the robust design, but the change itself is the issue. In this study, the author shows another gap between the obtained results under and the expected results and then, discisses the expectation to the robust design.

View full abstract

This study proposes a system modeling method based on FE model and measured data. The performance of the control system designed with the model obtained by this method is evaluated. Usually, a state space equation derived FE model is not enough accurate to achieve high control performance. On the other hand, when high-order MIMO model is identified by measured data, it's difficult to identify the precise model due to the complexity of the calculation. Therefore, this study introduces a method that tunes the state space equation derived FE model to fit the measured data. Simultaneous perturbation stochastic approximation (SPSA) is used as tuning algorithm. The vibration control experiment by using this tuned model is carried out and demonstrates the high vibration control performance.

View full abstract

There are a lot of studies on Robust Optimization. It seems these techniques are matured but basic formulation does not really express the idea of what they really like to do. At first, questions on formulation are listed and methods to deal with all those questions are argued. Then to have real meaning of robust solution, worst case optimization is proposed, and its effectiveness are discussed.

View full abstract

There are a lot of surrogate optimization methods, but we can summarize them into three steps: data detection step, approximation step and recommendation of experimental points step. With these steps, it becomes effective and has a chance to obtain global solution within a relatively small number of experiments and/or analyses. However, when we have a number of design variable and/or extensions of design space, it becomes more and more difficult to have its efficiencies. In this paper, we propose zooming technique which automatically limits both searching area and data around local optima that have been obtained during the surrogate optimization processes and tries to keep effectiveness of the surrogate optimization.

View full abstract

Hybrid or electric automobiles which is environmentally-friendly has been becoming obtaining. In these automobiles, noise induced by an engine is decreased, but road noise has relatively come to stand out. Damping sheet and sound-proof materials are widely used in multi-layered structures to suppress vibration and noise in automobiles. The number of combinations of multi-layered structures are too enormous to calculate all possible combinations. Therefore, this paper calculates efficiently for lighter and better sound proofing multi-layered structures than existing multi-layered structures. Finite element method that has already verified the accuracy quantitatively with the experimental results is used to calculate vibration, but it takes a very long time to calculate the analysis result. We employ transfer matrix method that is shorter time than finite element method. In transfer matrix method, multi-layered structures are expressed by transfer matrix using Biot's model to analyze efficiently. Lighter and better sound proofing efficiency multi-layered structures were efficiently calculated by using a multi-objective genetic algorithm having four objective functions that are mass per unit area, sound pressure level, transmissivity and absorption coefficient.

View full abstract

Structural optimization has been successfully used in many industries, such as automotive industries. In particular, topology optimization, first proposed by Bendsoe and Kikuchi, has recently been applied to a variety of complex problems, including multiobjective optimization problems. To handle the multiobjective optimization, a scalarization technique is often used, in which a multiobjective optimization problem is converted to a single objective optimization problem, but adjustment of the scalarization parameters depends on a cumbersome iterative process. On the other hand, there exist a number of solutions, called as Pareto optimal solutions, in multiobjective optimization problems. Pareto optimal solutions, which are also called as non-dominated solutions, help us to analyze the trade-off conditions among the conflicting objective functions, and once Pareto optimal solutions are obtained, design engineers can easily choose a particular solution that best meets their needs among the Pareto optimal solutions. Here, we propose a new multiobjective topology optimization method that can efficiently explore the Pareto frontier. The proposed method adopts a population-based approach in which many searching points are updated considering the distribution of the non-dominated solutions, that is, the weighting coefficients are adaptively given during the optimization process, considering the distribution of the non-dominated solutions. To illustrate the applicability of the proposed method, it is applied to a numerical example for design of a fluid device.

View full abstract

The purpose of this study is to generate the three dimensional gait of a two-legged robot to avoid obstacles. It is expected that two-legged robot can avoid obstacle more smoothly in the same way that animal and human adjust stride naturally to step over obstacles. Stepping points are determined optimally under the condition of maximizing walking speed and minimizing energy for walking. The gait generation problem is reduced to a combinatorial optimization problem solved by using genetic algorithm. Orbits of toes and hip between stepping points are generated by means of parametric modeling. The stable walking patterns are obtained under the condition of the maximizing walking speed and the minimizing energy consumption. The pareto front of the multi-objective optimization for the given robot model is visualized in advance by the MOGA, the optimum walking pattern is determined by using the satisficing trade-off method. The effectiveness of the proposed method is shown by simulation results.

View full abstract

SDSI (System Design System Integration)-Cubic is proposed which enables optimal design of large-scaled and complex system. However, SDSI-Cubic cannot derive the global optimal solution when it's applied to the optimization problem with huge design space. So, we propose new hierarchical optimization method to solve this problem for SDSI-Cubic. In the proposing method, design problem is decomposed in terms of system structure. By defining the value of common parameter (interaction) between subsystems in advance, solutions derived in each subsystem problems necessarily have equivalent interaction and are consistent. Furthermore, by evaluating the set of solutions integrating outputs of subsystem problem, it can be recognized whether defined value of interaction is appropriate or not. Thus, we propose optimization method hierarchizing two kinds of optimization problems; the upper-system-level optimization problem evaluating outputs from lower problems and searching the most appropriate value of interactions, and the lower-subsystem-level optimization problem deriving pareto solutions in the condition of interaction defined in upper problems. We also propose the objective function to evaluate a set of Pareto solutions in the upper-system-level optimization problem. This method is applied to the System LSI design problem in order to show the availability of this method.

View full abstract

This research addresses a value evaluation method of installation of dental computer system via developing a product-service simulator to describe the dental service processes. The value of dental service comprises three: (1) quantitative aspects such as time and costs, (2) qualitative aspects such as emotion and satisfaction, and (3) service effects. Based on the experimental study, the value of dental computer service is structure d as the three-le vel description. This paper especially reports a scientific experiment result about (1) quantitative aspects such as operation time, interaction processes between patient and doctor, and spatial movement of each actor.

View full abstract

The highly informative society is progressing. In connection with this, the circuit scale and functions of system LSIs are increasing. Therefore, development of the system design methods for designing a large-scale and complicated system LSI efficiently are needed. To this subject, writers have developed the SDSI-Cubic method as the efficient system design method. Using this method, we have derived the proper 3D system LSI for next-generation cellular phones by constructing a parametric physical model. However, this physical model is insufficient of consideration of a business investment because system LSI cost is evaluated by chip cost and investments for equipment is not evaluated. This means that business evaluation is not made. Therefore, in this paper, by constructing the economic model which evaluates total profit and loss dissolution time as the index of business, we derived proper composition of the 3D system LSI for next-generation cellular phones considering investments for system LSI equipment.

View full abstract

This research proposes an integrated method of scenario planning method which consider future scenarios actively and qualitatively, and discovery driven planning (DDP) method which consider product development value quantitatively and learningfully. This integration enables manager to acquire contextual changing various environmental factors considering various futures, and evaluate the project value in which a firm select an option. A prototype system is developed to confirm whether proposed method can select the option to maximize the expected project value under changing management environment.

View full abstract

Disaster sites often put rescue workers in danger. While many universities and companies are researching how to develop rescue robots, their efforts are not always as practical as needed. Therefore, this research conducted a survey of rescue workers in order to determine exactly what they need their robots to do. And I performed development of practical robot in use. In order to develop successful products, it is essential to ask: "Who are customers?" "What do they want?" and "How can we implement their requirements in our product?". It is necessary to find proper system configuration at early design stage. We employed the 1DCAE design support tool which can be used in upstream as well as in downstream design. As a result, this rescue robot system got praise from rescue workers. This paper introduces design process using needs analysis and 1DCAE. In addition, propose this process for helping MOT.

View full abstract

This paper describes a proposal on a new technique that enables the appropriate and practical bidding decision, before submitting the proposal to participate in the bidding for a large-scale international infrastructure project. Such technique has been required and demanded by potential bidders in order to improve the recent trend that Japanese companies given up in participating in the international infrastructure projects in spite of opportunities for Japanese companies to extend their business in highly potential overseas market. Specifically, we have established a comprehensive check list of risk assessment for international infrastructure projects, by reference to one of leading Japanese consultants who has participated in a number of overseas infrastructure projects and has extensive experiences in designing, bidding management and project/construction management. Furthermore, we have conducted a post-evaluation with realistic assumptions, for the construction project of Airport Rail Link (ARL) in Bangkok, Thailand, which was carried out through an international competitive bidding ICB process with participants from international companies. Based on the results of evaluation and assessment, we have concluded the validity of the check list with suggestion for further improvement of the evaluation method.

View full abstract

Recently, Product-Service System (PSS) that is consisting of tangible products and intangible services have been regarded as a new business model for manufacturing firms to gain competitiveness. In order to increasing efficiency of PSS business design process, it is important for designer to evaluate design solutions in the early-stage of the design process. To do so, in this study, the concept of qualitative simulation is applied to PSS business design. Especially, we propose a modeling method based on Balance Score Card (BSC) so that designers can develop PSS business strategy in the early-stage of PSS business design process.

View full abstract

This paper proposes a dynamic table assignment method for a restaurant to reduce customer waiting time and inequality of lengths of waiting among each customer. It is developed a simulation model to estimate remaining time until departure of each customer and the model evaluates effective sequences of table assignment to improve waiting time and availability of tables virtually. Gini coefficient value is used as an evaluation index for measuring inequality. In this paper, two conventional table assignment rules and the proposed method are described. Computer experiments are demonstrated that the proposed method is superior to the conventional model which manages to assign in First In First Out (FIFO) way or assigning smaller customer groups in priority to larger customer groups.

View full abstract

Product-Service Systems (PSS), systems which forces on generating higher value by integrating physical products and services, have been attracting much attention in manufacturing. To design PSS business model, it is important to select best one for your company after reviewing many business model in conceptual business model design stage. However, it is difficult for traditional product-selling manufacturing that inventing PSS business model which includes much service element. This is also shown in the workshop by the authors have been performed multiple times so far. Therefore, the final goal of this study is to support PSS business model design. To accomplish this final goal, this paper analyzes the characteristics of design process that has achieved variety of PSS business models and acquires the knowledge that relates characteristics of design process.

View full abstract

Recently, variety of products and services are expanding than ever, and requirement to creativity is getting bigger. For the reason, numeral method of idea generation is proposed, but most of the processes are not established appropriately and outcome depends on designers themselves. This is because process of idea generation is still remained unexplained. We tried to propose idea generation model from the cognitive approach. In cognitive science, human memory is assumed as structure of verbal and nonverbal system. It is notable that nonverbal system has much more abundant information compared to verbal system, activating nonverbal memory be useful for idea generation. On the other hand, there is attention: the system which controls the visual information processing which will suppress nonverbal memory activation. In order to reinforce the nonverbal memory activation, we propose that dispersing the attention makes nonverbal memory system more activate and we created attention dispersion system and attempted an experiment to confirm that dispersing attention makes more access to memory and affect to participant's idea generation. From the experiment, participants gave more terms when their attention was dispersed, compared to no attention dispersion.

View full abstract

Fault tree analysis (FTA) is a method of analyzing and visualizing the causes of a fault using a fault tree diagram (FT diagram), which has a tree structure with logical steps. Design engineers developing a new product generally use FTA to analyze many fault events, calculate their probability, and include redundancy systems in the design process. Furthermore, FTA has been used to analyze problems with products and to prevent the occurrence of problems in the design phase. However, it is not easy for design engineers to produce an accurate FT diagram in the actual design process. We have developed a computer-aided knowledge management system for creating FT diagrams (FTAid) as part of a collaborative group (The University of Tokyo, National Institute of Advanced Industrial Science and Technology (AIST), and Jatco Ltd.). This system has been verified by the design engineers of Jatco Ltd. in actual product development. We report its effectiveness for predicting mechanical, electrical, and heat transfer failure, the verification of the system, and its validation in an actual design process. We conclude that the system can help design engineers to effectively and efficiently create FT diagrams in reliability engineering, although some existing ability in FTA and engineering is required. We also describe some outstanding issues regarding the improvement of FTAid, engineering education, and ensuring reliability.

View full abstract

The authors have been studying physical quantity dimension indexing as a knowledge management technology about physical phenomena such as faults and their causes in fault tree analysis (FTA). Although the technology enables us objectively and systematically describe and retrieve knowledge about physical phenomena that can be represented with physical quantities, the technology is not applicable to phenomena or event that is not represented with physical quantities such as "a connector falls out". In this study, the authors try to propose an indexing system that is applicable to more various fault events and their causes.

View full abstract

Recently, view point of service has been widespread. In learning domains, service providers have been required to adopt it and to provide values co-created by teachers and learners. The value co-creation takes place in joint sphere where providers and customers make interaction, and the value is subjectively perceived by receivers (learners in learning service). That is, a mental state of leaner is one of possible value in the co-creation. The current study aims to develop a conceptual framework which describes mental states of learners for specifying values in the co-creation. As a result, we proposed ontological descriptions of learning event which consists of a teacher, a learner, learning materials, and a learning action by the learner. In concrete terms, we proposed ontological descriptions for receiver states in a learning activity, on the basis of a conceptual model of value co-creation. The ontological descriptions are useful for sharing knowledge on values perceived by learners to design learning services.

View full abstract

In recent years, the number of overseas design site is increasing. We estimate that the mass of the communication data exceed the capacity of the communication line. When the design data is not transferred appropriately because communication line congestion, we cannot use design data to execute design work. Therefore, design completion might be late. To construct the design process to make a design period minimize, we developed the data communication prioritizing simulation tool when plural design works are carried out in parallel. The design work is comprised of two kinds of tasks of a design task and the data communication task. And data communication task includes data transfer task and real time data share between two design sites. The simulation process involves five main steps. STEP1: Making a node for both sides of the communication task and last design task, and connect all tasks from start node to goal node. When the communication task includes real time data share, confirm the design task consistency. STEP2: Input the design or data transfer time of each task. STEP3: Integrating the same node and edge combination. STEP4: Rearranging node position. STEPS: Calculating the shortest course using the Dijkstra method. This method was implemented to calculate the data communication priority for simple design process model. As an evaluation result, we confirmed that the proposed method could prioritize data communication task and construct design process with minimizing design period.

View full abstract

We have developed welding support system based on the cases consisting of welding beads image, welding condition, evaluation results and notes. This paper discusses an extraction of image features from welding beads, which indicate technique of well-skilled welders. We examine two type of features such as Speeded Up Robust Features (SURF) and Gray-Level Co-Occurrence Matrix (GLCM). We have tried to use these features and examined the features can be applied to for the beginners to obtain welding technique.

View full abstract

The rocket engine turbopump is the high-rotational speed and high-power turbomachinery to feed the propellant from the tank to the combustion chamber. Lateral vibration on turbopump rotor is a very serious problem and it has brought about the unplanned reworks in the development phase. In order to avoid these rework process, we have proposed the morphology design method in the upstream design phase. This design method is characterized by considering the layout and permutation of the elements (inducer, impeller, turbine, bearing, seal, etc.) as design parameter. In this paper, the adopting results to the liquid hydrogen turbopump by using the morphology design method are presented, and the usability of this method is described.

View full abstract

Wrinkling and tearing are major defects in deep drawing, and blank holder force (BHF) plays an important role for the successful sheet forming. Variable blank holder force (VBHF) that the BHF varies through the punch stroke is useful for prevent these defects. On the other hand, the pulsing blank holder force (PBHF) is an attractive approach for the successful sheet forming. In this paper, both the VBHF and the PBHF trajectory are optimized, the formability window (FW) that represent the relationship between the BHF and the punch stroke is clarified. The identification of FW is one of the important issues to difficult-to-draw material. A design optimization problem is constructed to identify the FW and improve the limit drawing ratio (LDR). The object function is taken as the punch stroke. Several parameters in the VBHF and the PBHF are considered as the design variables. Numerical simulation in sheet metal forming is so numerically intensive that a sequential approximate optimization using a radical basis function network is used. In the numerical simulation, deep drawing of a cylindrical cup based on NUMISHEET2011 (BM1) is used. It is found from the numerical result that the proposed approach is useful for identifying the FW of a difficult-to-draw material. In addition, the optimum PBHF can reduce the maximum punch load that affects the tool life.

View full abstract

This paper proposes optimal vibration-proof design of reactor core based on coupled analysis of magnetics and harmonic response. A reactor is a part of the in-vehicle DC-DC converter to boost the electric voltage, and it is composed of cores, coils, and so on. Because of electromagnetic and harmonic response characteristics of the reactor, the reactor core vibrates when the reactor core is in use, and as a result, noises are generated. To reduce these noises, a topology optimization method for reactor core design is proposed in this paper. Here, the main difficulty is to compute the electromagnetic force appropriately. Roughly speaking, standard topology optimization methods, such as density based methods, are not suitable to compute the electromagnetic force because grayscale elements smear out the structural boundary in which the electromagnetic force is imposed. Therefore, in this paper, boundary tracking type level-set based topology optimization is used where grayscale elements can be suppressed completely. Thanks to this grayscale-free topology optimization, the coupled analysis of magnetics and harmonic response is successfully performed, and as a result, topology optimization of the reactor core for reducing the noises is realized. The usefulness of the proposed method is confirmed using some numerical examples.

View full abstract

The purpose of this study is the generation of a two-legged robot gait. Stepping points are determined optimally under the condition of maximizing walking speed and minimizing energy for walking. The gait generation problem is reduced to a combinatorial optimization problem solved by using genetic algorithm. Trajectories of toes and hip between stepping points are generated by means of parametric modeling. The stable walking patterns are obtained under the condition of the maximizing walking speed and the minimizing energy consumption. In this report the walking motion is generated to satisfy the target transition behavior in which the walking speed changes from acceleration motion to constant speed and to deceleration motion.

View full abstract

Recently, cloaking devices have collected a lot of attentions as an aplication of metamaterials. The cloaking is a technology to hide an object by arranging cloaking materials around the object. The light goes around the object by the cloaking materials and the object will be invisible. Many researchers have been trying to develop the cloaking devices by periodical arrangement of micro structures, which is not necessarily "optimal" since the arrangement is often designed by trial and error prcess. To realise an "optimal" cloaking device, the topology optimisation is considered to be a promissing design tool. The topology optimisation is the most flexible configuration optimisation method and can find an optimal configuration based on mathematical verification. In this method, the finite element method (FEM) is generally used. The FEM, however, takes a lot of numerical costs to generate a mesh and cannot treat an open domain strictly. From these points of view, the boundary element method (BEM) can be said more suitable for the topology optimisation for cloaking devices. In the BEM, mesh is generated only on the boundary and open domain can be evaluated exactly. In this study, we apply a BEM-based topology optimisation to the design problem of carpet cloaking devices. In the optimisation problem, an objective function is defined on a lot of observation points around the object to be hidden. The configuratin will therefore be complicated, which causes the large number of optimisation steps and accumulation of analysis error. In order to avoid this problem we employ the quadratic element to discretize the boundary. Through numerical examples, we confirmed the efficiencly of the proposed method.

View full abstract

Structural optimization has been used not only for structural design but also, recently, for material microstructural design. Microstructural optimization methods are categorized into two types: one aims to control material properties while the other aims to improve the performance of a structure. Most microstructural optimization methods are based on the homogenization method which assumes that the microstructures are infinitesimally small, but these microstructures actually have a fmite size. On the other hand, micropolar continuum theory enables us to handle materials whose microstructures are finite in size as homogeneous materials. This theory introduces rotational degrees of freedom in addition to the usual translational degrees of freedom and requires new material constants, such as characteristic length. Micropolar constants are derived in cases where the microstructure has a lattice pattern, although they are generally unknown. We therefore propose a method for optimizing the lattice widths of a microstructure at each point by approximating the material as a micropolar elastic solid so that the performance of a structure can be improved. The proposed method is applied to several numerical examples in order to demonstrate its effectiveness.

View full abstract

We propose a fluid topology optimization method using a Newton-gradient-hybrid scheme. The method converges in small computation time. In addition, the boundary between fluid and solid region is clearly distinguished. In the method, the domain is updated concurrently with solving a flow field. The flow field is solved by the lattice Boltzmann method (LBM). Due to the formulation of LBM and the optimization algorithm, the Hessian matrix is a diagonal matrix. The Hessian matrix is not generally positive semidefinite since the flow topology optimization problem is a nonconvex problem. Hence, we employ a Newton method for positive definite part of the Hessian marix, and employ a gradient method for negative semidefinite part.

View full abstract

This paper proposes a topology optimization method for high power motors, which is used in hybrid electric vehicles. The Permanent Magnet Synchronous motor (PMSM) is often used in hybrid electric vehicles, and its performances directly affect to performances of hybrid electric vehicles. Therefore, it is valuable to research topology optimization of PMSM. Although there are a few research papers discussing topology optimization of PMSM, these researches consider only torque performances. However, we consider that mechanical stiffness to hold permanent magnets should be also considered. If we consider only torque performances and ignore the mechanical stiffness, it will be possible to yield invalid optimal design which is meaningless in structural engineering viewpoints. In this paper, therefore, we propose a topology optimization method for PMSM, in which both torque and mechanical stiffness are maximized. In the proposed method, material distribution of the laminated electrical steel sheet of the rotor is optimized on the basis of the density method. The usefulness of the proposed method is demonstrated using some numerical examples.

View full abstract