• 수산해양기술연구
• /
• 제27권4호
• /
• pp.321-327
• /
• 1991

Experiments of quenching were made with cylindrical specimens of carbon steel S45C of diameters from 12 to 30mm were performed. The specimens were heated by electric furnace and quenched by immersion method. In order to analyze the temperature profile(cooling curves) of carbon steel including the latent heat of phase transformation, nonlinear heat conduction problem was calculated by the numerical method of inverse heat conduction problem using the apparent heat capacity method. The difference between the calculated and the experimented cooling curves was caused by the latent heat of phase transformation, and the effects of the latent heat were especially manifest at the cooling curves of center of specimens. The temperature and the quantity of the latent heat of phase transformation depend on the cooling speed at A sub(1) transformation point, and the region for cooling speed to become zero was caused by the latent heat of phase transformation.

• Journal of Electrical Engineering and Technology
• /
• 제8권2호
• /
• pp.271-281
• /
• 2013

In order to improve the performance of analysis, it is important to consider the nonlinearity in power system. The Carleman embedding technique (linearization procedure) provides an effective approach in reduction of nonlinear systems. In the approach, a group of differential equations in which the state variables are formed by the original state variables and the vector monomials one can build with products of positive integer powers of them, is constructed. In traditional Carleman linearization technique, the tensor matrix is truncated to form a square matrix, and then regular linear system theory is used to solve the truncated system directly. However, it is found that part of nonlinear information is neglected when truncating the Carleman model. This paper proposes a new approach to solve the problem, by combining the Poincar$\acute{e}$ transformation with the Carleman linearization. Case studies are presented to verify the proposed method. Modal analysis shows that, with traditional Carleman linearization, the calculated contribution factors are not symmetrical, while such problems are avoided in the improved approach.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2005년도 학술발표회(2)
• /
• pp.916-917
• /
• 2005

• 대한기계학회논문집
• /
• 제16권6호
• /
• pp.1036-1055
• /
• 1992

본 연구에서는 일단앞섬 상태추정기(one-step ahead state estimator)를 사용 하는 방법을 제안하였다. 이 방법은 비선형 시스템을 궤환 선형화하여 등가 선형모 델을 구한 후, 이 등가 선형 모델에 일단앞섬 상태추정기를 구성하여 디지틀 제어를 실현하는 방법으로서, 계측할 수 없는 상태를 추정할 수 있을 뿐 아니라, 선형시스템 에 대한 상태 추정기이므로 다음 단계의 상태를 비교적 정확하게 추정할 수 있어서 궤 환 선형화시 요구되는 연속적인 상태 궤환에 근사시킬 수 있고, 직접 궤환 선형화된 등가 선형 모델의 상태를 추정할 수 있으므로, 변환시의 연산량도 줄일 수 있다. 이 로서 Grizzle이나 Lee의 방법보다 간단하게 궤환 선형화에 의한 디지틀 상태 제어기를 구현할 수 있었다.

• 한국해양공학회지
• /
• 제19권2호
• /
• pp.29-33
• /
• 2005

The accuracy of predicting wave transformation in the nearshore is very important to wave hydrodynamics, sediment transport, and design of coastal structures. Numerical experiments are conducted to identify the shoaling and breaking characteristics of a fully nonlinear Boussinesq equation-based model. Simulated shoaling showed good agreement with the Shouto's formula, and the results of the breaking experiment agreed well with experimented data, over several beach profile.

• Journal of applied mathematics & informatics
• /
• 제22권1_2호
• /
• pp.223-235
• /
• 2006

By applying the Landau-type transformation, we transform a Stefan problem with nonlinear free boundary condition into a system consisting of a parabolic equation and the ordinary differential equations. Fully discrete finite element method is developed to approximate the solution of a system of a parabolic equation and the ordinary differential equations. We derive optimal orders of convergence of fully discrete approximations in $L_2,\;H^1$ and $H^2$ normed spaces.

• Journal of applied mathematics & informatics
• /
• 제27권5_6호
• /
• pp.1165-1176
• /
• 2009

This paper is concerned with the interval oscillation of the second order nonlinear ordinary differential equation (r(t)|y'(t)|$^{{\alpha}-1}$ y'(t))'+p(t)|y'(t)|$^{{\alpha}-1}$ y'(t)+q(t)f(y(t))g(y'(t))=0. By constructing ageneralized Riccati transformation and using the method of averaging techniques, we establish some interval oscillation criteria when f(y) is not differetiable but satisfies the condition $\frac{f(y)}{|y|^{{\alpha}-1}y}$ ${\geq}{\mu}_0$ > 0 for $y{\neq}0$.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.48.1-48
• /
• 2001

In this paper, we present a global adaptive output feedback control scheme for a class of uncertain nonlinear systems to which adaptive observer backstepping method may not be applicable directly. The allowed output feedback structure includes quadratic and multiplicative dependency of unmeasured states. Our novel design technique employs a change of coordinates and adaptive backstepping. With these proposed tools, we can remove linear and quadratic dependence on the unmeasured states in the state equation. Also, the multiplication of the two unmeasured states can be eliminated ...

• 한국지능시스템학회논문지
• /
• 제19권6호
• /
• pp.865-871
• /
• 2009

본 논문은 이산 시간 비선형 시스템을 이산 시간 T-S 퍼지 모델에 의해 표현되는 새로운 궤환 선형화에 대해서 논한다. T-S fuzzy 모델의 국부적인 선형 모델들은 각각 가제어 표준형으로 변환되어지고, 그것들의 T-S 퍼지 결합은 궤환 선형화 가능한 T-S fuzzy 모델이 된다. 이 모델을 토대로 비선형 상태 궤환 선형 입력이 결정된다. 비선형 상태 변환은 가제어 표준형에 대한 선형 상태 변환으로부터 추론된다. 본 논문에서 제안하는 방법은 충분한 수학적 배경이 요구되는 고전적인 궤환 선형화 기법과 비교하여 수학적으로 보다 직관적이고 이해하기 쉽다. 본 논문의 궤환 선형화 조건은 고전적인 궤환 선형화와 비교하여 더 완화되었다. 이것은 고전적인 선형화방식 보다 더 큰 범주의 비선형 시스템이 선형화가 가능해진다는 것을 의미 한다.

• ETRI Journal
• /
• 제34권6호
• /
• pp.847-857
• /
• 2012

Intrusion detection systems (IDSs) have an important effect on system defense and security. Recently, most IDS methods have used transformed features, selected features, or original features. Both feature transformation and feature selection have their advantages. Neighborhood component analysis feature transformation and genetic feature selection (NCAGAFS) is proposed in this research. NCAGAFS is based on soft computing and data mining and uses the advantages of both transformation and selection. This method transforms features via neighborhood component analysis and chooses the best features with a classifier based on a genetic feature selection method. This novel approach is verified using the KDD Cup99 dataset, demonstrating higher performances than other well-known methods under various classifiers have demonstrated.