• Computational Structural Engineering : An International Journal
• /
• 제1권1호
• /
• pp.31-38
• /
• 2001

In the case of the non-proportionally damped system such as the soil-structure interaction system, the structural control system and composite structures, the eigenproblem with the damping matrix should be necessarily performed to obtain the exact dynamic response. However, most of the eigenvalue analysis methods such as the subspace iteration method and the Lanczos method may miss some eigenvalues in the required ones. Therefore, the eigenvalue analysis method must include a technique to check the missed eigenvalues to become the practical tools. In the case of the undamped or proportionally damped system the missed eigenvalues can easily be checked by using the well-known Sturm sequence property, while in the case of the non-proportionally damped system a checking technique has not been developed yet. In this paper, a technique of checking the missed eigenvalues for the eigenproblem with the damping matrix is proposed by applying the argument principle. To verify the effectiveness of the proposed method, two numerical examples are considered.

• 한국게임학회 논문지
• /
• 제16권1호
• /
• pp.119-128
• /
• 2016

본 논문에서는 IFS(iterated function system) 프랙탈에서 점의 코드를 활용한 변형의 개선방법을 고려한다. 기존의 변형방법에서는 부분적 변형이 랜덤 반복에 의해 임의로 전파되므로 대개 단조로운 느낌을 주는 모양이 나타나고 있다. 이러한 점을 개선하기 위하여 코드를 활용하여 맵의 선택을 제어하는 방법을 제안한다. 제안된 방법을 적용한 결과 프랙탈의 모양 특성이 적절히 반영된 흥미로운 변형을 얻을 수 있었다. 또한 코드에 따라 변화하는 상태변수를 도입하여 좌표의 변환 외의 다른 속성의 변형을 손쉽게 구현하는 방법도 제안한다.

• 대한기계학회논문집A
• /
• 제31권2호
• /
• pp.211-220
• /
• 2007

This work presents an iterated improved reduced system (IIRS) procedure combined with sub-structuring scheme for large structures. Iterated IRS methods are usually more efficient than others because the dynamic condensation matrix is updated repeatedly until the desired convergent values are obtained. However, using these methods simply for large structures causes expensive computational cost and even makes analyses intractable because of the limited computer storage. Therefore, the application of sub-structuring scheme is necessary. Because the large structures are subdivided into several (or more) sub-domains, the construction of dynamic condensation matrix does not require much computation cost in every iteration. This makes the present method much more efficient to compute the eigenpairs both in lower and intermediate modes. In Part I, iterated IRS method combined with sub-structuring scheme for undamped structures is presented. The validation of the proposed method and the evaluation of computational efficiency are demonstrated through the numerical examples.

• 한국산업정보학회논문지
• /
• 제8권4호
• /
• pp.26-30
• /
• 2003

대칭이고 큰 희소 행렬(Large Sparse Matrices)에 대한 가장 작거나 또는 가장 큰 고유치(Eigenvalues)들을 구하기 위해서 Lanczos 방법이 많이 이용된다. MP(Message Passing) 병렬 컴퓨터에서 global communications은 계산 속도를 떨어뜨린다. 본 논문에서는 s-step Lanczos 알고리즘을 소개하였으며 이 s-step 방법은 기존의 Lanczos 알고리즘에 의해 생성된 행렬에 유사한 축소 행렬을 생성하며 s-step Lanczos 알고리즘에서 한번의 반복은 기존의 Lanczos 알고리즘의 s 번 반복에 해당한다. s-step 방법은 global communications을 최소화하였으며 기존의 알고리즘에 비해 뛰어난 병렬 성질을 가진다. 알고리즘들은 Cray T3E에서 수행되었으며 그 결과를 볼 수 있다.

• 산업공학
• /
• 제20권2호
• /
• pp.112-120
• /
• 2007

As the market rapidly changes, the speed of new product development is highlighted as a critical element which determines the success of firms. While firms endeavor to accelerate the development speed, frequent iterations in a development process hinders the effort of acceleration. For this reason, many previous researches tried to find the optimal structure of the development process which minimizes the number of iterations. However, such researches have a limitation in that they can be applied to only a single-project environment. In a multi-project environment, waiting time induced by lack of resources also delays the process as well as the iterations do. In this paper, we propose dynamic sequencing method focusing on both iterations and waiting time for reducing the durations of development projects in a multi-project environment. This method reduces the waiting time by changing the sequence of development tasks according to the states of resources. While the method incurs additional iterations, they are expected to be offset by the reduced waiting time. The results of simulation show that the dynamic sequencing method dramatically improves the efficiency of a development process. Especially, the improvement is more salient as projects are more crowded and the process is more unbalanced. This method gives a new insight in researches on managing multiple development projects.

• 로봇학회논문지
• /
• 제17권2호
• /
• pp.221-229
• /
• 2022

Sliding mode control (SMC) is a robust control method to control a robot arm with nonlinear properties. A high switching gain of SMC causes chattering problems, although the SMC allows the adequate control performance by giving high switching gain, without the exact robot model containing nonlinear and uncertainty terms. In order to solve this problem, SMC with sliding perturbation observer (SMCSPO) has been researched, where the method can reduce the chattering by compensating the perturbation, which is estimated by the observer, and then choosing a lower switching control gain of SMC. However, optimal gain tuning is necessary to get a better tracking performance and reducing a chattering. This paper proposes a method that the Q-learning automatically tunes the control gains of SMCSPO with an iterative operation. In this tuning method, the rewards of reinforcement learning (RL) are set minus tracking errors of states, and the action of RL is a change of control gain to maximize rewards whenever the iteration number of movements increases. The simple motion test for a 7-DOF robot arm was simulated in MATLAB program to prove this RL tuning algorithm. The simulation showed that this method can automatically tune the control gains for SMCSPO.

• 전력전자학회논문지
• /
• 제27권6호
• /
• pp.471-480
• /
• 2022

Inductive power transfer (IPT) is an attractive power transmission solution that is already used in many applications. In the IPT system, optimal coil design is essential to achieve high power efficiency, but the effective design method is yet to be investigated. The inductance formula and finite element method (FEM) are popular means to link the coil geometric parameters and circuit parameters; however, the former lacks generality and accuracy, and the latter consumes much computation time. This study proposes a novel coil design method to achieve speed and generality without much loss of accuracy. By introducing one-turn permeance simulation in each FEM phase combined with curve fitting and optimization by MATLAB in the efficiency calculation phase, the iteration number of FEM can be considerably reduced, and the generality can be retained. The proposed method is verified through a 100 W IPT system experiment.

• 대한기계학회논문집B
• /
• 제40권11호
• /
• pp.705-715
• /
• 2016

본 연구의 목적은 에어컨 사이클 성능해석에 있다. 응축기, 증발기, 팽창밸브 및 압축기는 냉동사이클을 구성하는 핵심요소이다. 사이클의 개별적인 구성요소들에 대한 해석 기법들을 합리적으로 통합하여 다양한 운전조건에서 에어컨 시스템 성능예측이 가능하도록 하였다. 응축기 압력은 압축기 질량유량과 팽창밸브 유량이 일치하도록 반복계산에 의해 획득되며, 증발기 압력은 목표 흡입과열도가 획득되도록 압축기 흡입엔탈피를 반복계산에 의해 획득되었다. 더 나아가서 복수 실내기를 장착한 에어컨 시스템의 성능이 예측될 수 있도록 알고리듬들이 마련되었으며, 이들 모델들에 대한 해석결과를 제시하였다. 소프트웨어의 정확성은 실험결과에 의해 증명 되었다. 특히, 8.3 kW급 모델의 실험결과와 비교함으로써, 소프트웨어의 정확성이 다양하게 검정되었다. 해석결과로써, 정확성은 대체적으로 10% 이내에 있는 것으로 확인되어 우수한 신뢰성이 확보되었다.

• 한국공간구조학회논문집
• /
• 제20권2호
• /
• pp.51-58
• /
• 2020

Recently, deep learning that is the most popular and effective class of machine learning algorithms is widely applied to various industrial areas. A number of research on various topics about structural engineering was performed by using artificial neural networks, such as structural design optimization, vibration control and system identification etc. When nonlinear semi-active structural control devices are applied to building structure, a lot of computational effort is required to predict dynamic structural responses of finite element method (FEM) model for development of control algorithm. To solve this problem, an artificial neural network model was developed in this study. Among various deep learning algorithms, a recurrent neural network (RNN) was used to make the time history response prediction model. An RNN can retain state from one iteration to the next by using its own output as input for the next step. An eleven-story building structure with semi-active tuned mass damper (TMD) was used as an example structure. The semi-active TMD was composed of magnetorheological damper. Five historical earthquakes and five artificial ground motions were used as ground excitations for training of an RNN model. Another artificial ground motion that was not used for training was used for verification of the developed RNN model. Parametric studies on various hyper-parameters including number of hidden layers, sequence length, number of LSTM cells, etc. After appropriate training iteration of the RNN model with proper hyper-parameters, the RNN model for prediction of seismic responses of the building structure with semi-active TMD was developed. The developed RNN model can effectively provide very accurate seismic responses compared to the FEM model.

• 한국인터넷방송통신학회논문지
• /
• 제10권5호
• /
• pp.1-5
• /
• 2010

본 논문은 열연 공정을 거친 철강 강판에 형성된 산화철 층, 즉 scale 층의 두께를 유전체 렌즈 안테나를 이용하여 측정하는 방법을 소개하였다. 유전체 렌즈 안테나는 X 밴드 대역에서 주파수에 독립적인 특성을 가지며, 혼 안테나에서 방사되는 구면파를 초점이 형성되는 평면에 평면파를 형성하는 역할을 한다. 이러한 동작원리를 이용하여 철강 강판에 형성된 scale 층에 완전 도체와 유전체로 형성된 two-layer 구조에 직각 입사하는 평면파의 이론적 해석이 적용될 수 있다. Scale의 두께를 도출해 내는 과정에서 유전체 렌즈의 영향을 최소화하기 위한 calibration 과정이 삽입되었으며, 이로 인한 반사 계수 위상의 오차가 발생하였다. 이러한 위상 오차에 의한 scale 두께의 오차를 줄이기 위하여, 수치적으로 regression 방법을 사용하였으며, 기존의 iteration 방법과 비교하여, 주기적으로 얻어지는 두께의 값이 아닌 단일 두께 값을 얻어낼 수 있었다.