• 한국소성가공학회:학술대회논문집
• /
• 한국소성가공학회 1999년도 춘계학술대회논문집
• /
• pp.253-256
• /
• 1999

The Streamline Upwind Petrov-Galerkin(SUPG) finite element method is used to solve the two-dimensional laminar and turbulent flow. The flow is simulated by averaged Navier-Stokes equations with a penalty function approach and the lograithmic(k-$\varepsilon$) turbulent model is employed to take into account its turbulent behavior. The near-wall viscous sub-layer model is employed to approach the dominant viscous effects in the near wall zones. To find a good-enough initial guess of the Newton-Raphson iteration solving Nonlinear Matrix the Incremental method is considered for momentum and the Incomplete logarithmic turbu-lent equations for Turbulence. The validation of our method is investigated in comparision with published experimental data.

• 한국안전학회지
• /
• 제13권1호
• /
• pp.40-46
• /
• 1998

This paper is presented that theoretical optimization design method in order to consider mass reduction for the structural safety In this paper, it described methods for reducing vibration in structural safety by the determination of the optimum sizes and locations of tunning masses through formal mathematical optimization techniques. The optimization procedure which employs the tunning masses and corresponding locations is developed. Design variables are systematically changed to achieve low values of shear without a large mass penalty. Three optimization methods ire developed and tested. The first is based on minimizing the modal shaping parameter which indirectly reduce the modal shear amplitudes corresponding to each harmonic of airload. The second method reduces these amplitudes directly and the third method reduces the shear as a function of time during a revolution of the blade. The first method works well for reducing the shear for one mode responding to a single harmonic of the airload but has been found in some bases to be ineffective for more than one mode.

• 한국음향학회지
• /
• 제38권6호
• /
• pp.670-677
• /
• 2019

음성 또는 음향 이벤트 신호에서 발생하는 배경 잡음은 인식기의 성능을 저하시키는 원인이 되며, 잡음에 강인한 특징을 찾는데 많은 노력을 필요로 한다. 본 논문에서는 딥러닝을 기반으로 다중작업 오토인코더(Multi-Task AutoEncoder, MTAE) 와 와설스타인식 생성적 적대 신경망(Wasserstein GAN, WGAN)의 장점을 결합하여, 잡음이 섞인 음향신호에서 잡음과 음성신호를 추정하는 네트워크를 제안한다. 본 논문에서 제안하는 MTAE-WGAN는 구조는 구배 페널티(Gradient Penalty) 및 누설 Leaky Rectified Linear Unit (LReLU) 모수 Parametric ReLU (PReLU)를 활용한 변수 초기화 작업을 통해 음성과 잡음 성분을 추정한다. 직교 구배 페널티와 파라미터 초기화 방법이 적용된 MTAE-WGAN 구조를 통해 잡음에 강인한 음성특징 생성 및 기존 방법 대비 음소 오인식률(Phoneme Error Rate, PER)이 크게 감소하는 성능을 보여준다.

• 대한조선학회지
• /
• 제26권1호
• /
• pp.39-45
• /
• 1989

선체구조물이나 해양구조물의 동적응답중 원치 않는 고유진동수와 고유진동형태를 가지게 되는 경우가 있으며, 이러한 구조물은 동적 구조 재설계가 필수적이다. 본 소고에서는 비감쇄 구조물의 고유진동수와 진동형태를 기진력에 의한 특정한 진동수와 공진하지 않도록 또는 구조물의 중요한 부분이 특정 진동형태의 최대치에 오지 않도록 구조물의 질량과 강성을 최적하게 변화시키는 방법에 대해 논의하고 있다, 이 방법은 기존의 방법에서 사용되는 모든 고유진동형태의 수식포함과 달리 구속된 고유진동형태만을 미지수로 수식중에 사용하여 불필요한 계산과정을 줄이고 있다. 동적 구조 재설계중 최적화 문제에 중점을 두었으며 목적함수로는 구조물의 최소의 변화와 또는 최소의 중량을 취하였고, 예제를 통하여 본 방법의 응용과 효율성이 입증되었다. 예제에서는 간단한 구조물을 다루었으나 본 방법은 상용 유한요소코드의 연계이용으로 각종 선체구조물과 해양구조물의 진동문제해결에 응용될 수 있음은 자명한 일이다.

• 대한기계학회논문집A
• /
• 제22권4호
• /
• pp.811-825
• /
• 1998

An approximate line search is presented for dynamic response optimization with Augmented Lagrange Multiplier(ALM) method. This study empolys the approximate a augmented Lagrangian, which can improve the efficiency of the ALM method, while maintaining the global convergence of the ALM method. Although the approximate augmented Lagragian is composed of only the linearized cost and constraint functions, the quality of this approximation should be good since an approximate penalty term is found to have almost second-order accuracy near the optimum. Typical unconstrained optimization algorithms such as quasi-Newton and conjugate gradient methods are directly used to find exact search directions and a golden section method followed by a cubic polynomial approximation is empolyed for approximate line search since the approximate augmented Lagrangian is a nonlinear function of design variable vector. The numberical performance of the proposed approach is investigated by solving three typical dynamic response optimization problems and comparing the results with those in the literature. This comparison shows that the suggested approach is robust and efficient.

• Journal of Electrical Engineering and Technology
• /
• 제3권4호
• /
• pp.514-519
• /
• 2008

The non-separable quadratic penalty terms create an inherent difficulty when applying the standard augmented Lagrangian relaxation(ALR) method for decomposing the unit commitment problem into independent subproblems. This paper presents a separable augmented Lagrangian relaxation method for solving the unit commitment problem. The proposed method is designed to have a separable structure by introducing the quadratic terms with additional auxiliary terms in the augmented Lagrangian function. Numerical results are presented to validate the effectiveness of the proposed method.

• Steel and Composite Structures
• /
• 제2권6호
• /
• pp.429-440
• /
• 2002

A finite strip method is developed for fatigue reliability analysis of steel highway bridges. Flat shell strips are employed to model concrete slab and steel girders, while a connection strip is formed using penalty function method to take into account eccentricity of girder top flange. At each sampling point with given slab thickness and modulus ratio, a finite strip analysis of the bridge under fatigue truck is performed to calculate stress ranges at fatigue-prone detail, and fatigue failure probability is evaluated following the AASHTO approach or the LEFM approach. After the failure probability is integrated over all sampling points, fatigue reliability of the bridge is determined.

• Nuclear Engineering and Technology
• /
• 제30권3호
• /
• pp.202-211
• /
• 1998

Interfacial stresses at two-material interfaces and initial displacement field over the entire domain are obtained by modifying the potential energy functional with a penalty function, which enforces continuity of the stresses at the interface of two materials. Based on the initial displacement field and interfacial stresses, a new methodology to generate a continuous stress field over the entire domain has been proposed by combining the modified projection method of stress-smoothing and Loubignac's iterative method of improving the displacement field. Stress analysis is carried out on two examples made of dissimilar materials : one is a two-material cantilever composed of highly dissimilar materials and the other is a zirconium-lined cladding tube made of slightly dissimilar materials. Results of the analysis show that the proposed method provides an improved continuous stress field over the entire domain, and accurately predicts the nodal stresses at the interface, while the conventional displacement-based finite element method produces significant stress discontinuities at the interface. In addition, the total strain energy evaluated from the improved continuous stress field converges to the exact value in a few iterations.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2000년도 하계학술대회 논문집 B
• /
• pp.772-774
• /
• 2000

The objective of a cathodic protection system (CP) is to protect the buried metallic structure against the corrosion caused by chemical reaction between the buried structure and the surrounding medium, such as soil. This paper presents a boundary element application to determine the optimal impressed current densities in a cathodic protection system. The potential within the electrolyte is described by the Laplace's equation with nonlinear boundary conditions which are enforced based on experimentally determined electrochemical polarization curves. The optimal impressed current densities are determined in order to minimize the power supply for protection. The solution is obtained by using the conjugate gradient method in which the governing equations and the protecting conditions are taken into account by the penalty function method. Numerical example are presented to demonstrate the practical applicability of the proposed method.

• 대한기계학회논문집A
• /
• 제21권1호
• /
• pp.112-123
• /
• 1997

The optimization techniques have been applied to versatile engineering problems for reducing manufacturing cost and for automatic design. The deterministic approaches or op5imization neglect the effects on uncertainties of design variables. The uncertainties include variation or perturbation such as tolerance band. The optimum may be useless when the constraints considering worst cases of design variables can not be satisfied, which results from constraint variation. The variation of design variables can also give rise to drastic change of performances. The two issues are related to constraint feasibility and insensitive performance. Robust design suggested in the present study is developed to gain an optimum insensitive to variation on design variables within feasible region. The multiobjective function is composed to the mean and the standard deviation of original objective function, while the constraints are supplemented by adding penalty term to original constraints. This method has a advantage that the second derivatives of the constraints are not required. A mathematical problem and several standard problems for structural optimization are solved to check out the usefulness of the suggested method.