• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.26 no.7
• /
• pp.945-950
• /
• 2002

Shape optimization of a flow system is done to obtain the required effects, in the engineering fields. Most of these designs are accomplished by empirical or numerical analysis. In empirical analysis, it is difficult to obtain an optimal shape in the feasible design region. And, in numerical method, it usually needs much calculation expenses for shape optimization, because of design sensitivity analysis. In this study, we used the growth-strain method having only one distributed parameter such as a design variable. It optimizes a shape by making a distributed parameter such as dissipation energy uniform in a flow system, and then applied to two-flow systems. In order to overcome the stability occurred in numerical analysis performed by Azegami, the equation of volumic strain has been modified. Also, the shapes were compared with the known optimal shapes for the flow systems. Consequently, we confirm that the modified growth-strain method is very efficient and practical in shape optimization of the flow systems.

• Journal of Magnetics
• /
• v.16 no.3
• /
• pp.268-270
• /
• 2011

This study introduces a new topology optimization scheme combing the genetic algorithm (GA) with the on/off sensitivity method for the magnetic actuator core and the armature design. The design process intended to maximize the first eigen-frequency of the armature part and the magnetic actuating force acting on the armature simultaneously. GA based optimal design was carried out to obtain the initial structure and the modified on/off sensitivity method was succeeded to accelerate the design process. Final results show tens of percent improvement in actuating force as well as the first eigen-frequency of the armature.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.17 no.1
• /
• pp.11-20
• /
• 1993

Sensitivity information is required in the optimal design process. In structural optimization, sensitivity calculation is a bottleneck due to its complexities and expensiveness. Various schemes have been proposed for the calculation. Analytic and finite difference methods are the most popular at the present time. However, they have advantages and disadvantages in different ways. Semi-anayltic method has been suggested to overcome the difficulties. In spite of the excellency, the semi-analytic method has been found to possess numerical error quite much with respect to shape variables. In this research, the error from each method is evaluated and compared using a shape variable. A two-dimensional beam is selected for an example since it has mathematical solution. An efficient method is suggested for the structural optimization which utilizes finite element method.

• Computational Structural Engineering
• /
• v.7 no.2
• /
• pp.147-153
• /
• 1994

A method for shape design sensitivity analysis for axisymmetric shells of general shapes is developed. The basic approach is to divide the structures into many segments : For each of the segments, the formula for a shallow arch or shell can be applied and the results assembled. To interconnect those segments, the existing sensitivity formula, obtained for a variation only in the direction perpendicular to the plane on which the structure is mapped, has been extended to include a variation normal to the middle surface. The method follows the adjoint variable approach based on the material derivative concept as established in the literature. Numerical examples are taken to illustrate the method and the applicability to practical design problems.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.33 no.3
• /
• pp.243-250
• /
• 2009

The adjoint variable method can reduce computation time and save computer resources because it can selectively provide the sensitivity information for the positions that designers wish to measure. However, the adjoint variable method commonly employs exact analytical differentiation with respect to the design variables. It can be cumbersome to precisely differentiate every given type of finite element. This trouble can be overcome only if the numerical differentiation scheme can replace this exact manner of differentiation. But, the numerical differentiation scheme causes of severe inaccuracy due to the perturbation size dilemma. For assuring the accurate sensitivity without any dependency of perturbation size, this paper employs a complex variable that has been mainly used for computational fluid dynamics problems. The adjoint variable method combined with complex variables is applied to obtain the shape and size sensitivity for structural optimization. Numerical examples demonstrate that the proposed method can predict stable sensitivity results and that its accuracy is remarkably superior to traditional sensitivity evaluation methods.

• The Transactions of the Korean Institute of Electrical Engineers B
• /
• v.52 no.6
• /
• pp.247-253
• /
• 2003

A novel finite element mesh regeneration method is presented for 3D shape optimization of electromagnetic devices. The method has its theoretical basis in the structural deformation of an elastic body. When the shape of the electromagnetic devices changes during the optimization process, a proper 3D finite element mesh can be easily obtained using the method from the initial mesh. For real engineering problems, the method guarantees a smooth shape with proper mesh quality, and maintains the same mesh topology as the initial mesh. Application of the optimum design of an electromagnetic shielding plate shows the effectiveness of the presented method.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.23 no.3
• /
• pp.275-281
• /
• 2010

In this paper, the design optimization of structures with stress constraints is performed using isogeometric shape optimization method. The stress constraints have an important role in design optimization problems since stress concentration could result in structural failure. To represent exact geometry in analysis, the isogeometric analysis method uses the same basis functions as used in the CAD geometry. The geometrically exact model can be used in both stress and design sensitivity analyses so that it can yield more precise optimal design than finite element one. Through numerical examples, the isogeometric approach turns out to be effective in shape optimization problems under stress constraints.

• Proceedings of the KIEE Conference
• /
• 2004.10a
• /
• pp.70-72
• /
• 2004

The ferromagnetic pole piece of permanent magnet assembly for magnetic resonance imaging(MRI) is optimally designed to get high homogenious magnetic field, taking into account the non-linearity of the magnetic materials. In the design, the pole face is kept smooth and axis-symmetric by using B-spline parameterization, and nonlinear design sensitivity analysis is used for search direction.

• Proceedings of the KIEE Conference
• /
• 1994.07a
• /
• pp.150-152
• /
• 1994

In this paper, the optimum shape design of stator slot of induction motors for iron loss reduction is proposed. To obtain the flux distribution in induction motors, 2-D finite element method with voltage source is employed. The iron loss is calculated from the iron loss data given by the iron manufacturer. To calculate the sensitivity of iron loss to shape variation, the sensitivity analysis of discrete approach is used. The proposed algorithm is applied to a 3-phase squirrel cage induction motor. The nodes at stator slot boundary of the induction motor are defined as design parameters. By controlling these parameters under the constant volume of iron, we can minimize the iron loss. Furthermore, the stator copper loss is reduced by increasing the slot area. So the stator slot area is determined at the point that the summation of iron loss and copper loss of stator is minimized. Since the constraint of constant volume of iron is nonlinear to the design parameters, the Gradient Projection method is used as an optimization algorithm.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.651-652
• /
• 2008

This research presents a multi-objective optimal design employing topological approach to maximize magnetic energy while minimizing structural deformation which is caused by magnetic reluctance force and magnetostriction. A design sensitivity formula is derived by employing the adjoint variable method (AVM) to avoid numerous sensitivity evaluations for a coupled magneto-mechanical analysis. The sensitivity analysis is verified using the finite difference method (FDM) in a C-shape actuator. A linear actuator used in a home appliance is examined for optimal design and demonstrates the strength of the proposed topology optimization approach.