• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2001년도 추계학술대회 논문집
• /
• pp.129-132
• /
• 2001

Recently, a ZVT boost converter is embedded in a power factor correction system. The control circuit of the converter assures soft-switching for all the MOSFETs and load regulation. The PFC system contains additional control circuits which assure the input voltage in a sinusoidal form and feed-forward line voltage regulation. In this paper, a soft switching boost converter with zero-voltage transition(ZVT) main switch using zero-current switching(ZCS) auxiliary switch is proposed. Operating intervals of the converter are persented and analyzed. The proposed results show that the main switch maintains UT while auxiliary switch retains ZCS for the complete specified line and load conditions.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 추계학술대회 논문집 학회본부 B
• /
• pp.812-814
• /
• 1999

In this paper, we propose a new control algorithm based on the neural network(NN) feedback compensation with a desired trajectory modification. The proposed algorithm decreases trajectory errors by a feed-forward desired torque combined with a neural network feedback torque component. And, to robustly control the tracking error, we modified the desired trajectory by variable structure concept smoothed by a fuzzy logic. For the numerical simulation, a 2-link robot manipulator model was assumed. To simulate the disturbance due to the modelling uncertainty. As a result of this simulation, the proposed method shows better trajectory tracking performance compared with the CTM and decreases the chattering in control inputs.

• 비파괴검사학회지
• /
• 제28권3호
• /
• pp.254-262
• /
• 2008

The objective of the proposed study is to produce a tool-condition monitoring (TCM) strategy that will lead to a more efficient and economical drilling tool usage. Drill-wear monitoring is an important attribute in the automatic cutting processes as it can help preventing damages of the tools and workpieces and optimizing the tool usage. This study presents the architectures of a multi-layer feed-forward neural network with back-propagation training algorithm for the monitoring of drill wear. The input features to the neural networks were extracted from the AE signals using the wavelet transform analysis. Training and testing were performed under a moderate range of cutting conditions in the dry drilling of steel plates. The results indicated that the extracted input features from AE signals to the supervised neural networks were effective for drill wear monitoring and the output of the neural networks could be utilized for the tool life management planning.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2007년도 춘계학술대회논문집
• /
• pp.1106-1111
• /
• 2007

This paper is focused on design of a new active control engine mount (ACM), which is compact in size and cost effective. The ACM, consisting of an electrodynamic actuator as the active element, flat springs and a sliding ball joint, is different in structure from the previous ACM designs based on the conventional hydraulic engine mount. Dynamic characteristics of the proposed ACM are extensively investigated before a prototype ACM, which meets the design specifications, is built in the laboratory. For cost effectiveness, a feed-forward control algorithm without a feedback sensor is used for reduction of the transmitted force through the ACM from the engine. The prototype ACM is then harmonic-tested with a rubber testing machine for verification of its control performance as well as adequacy of modeling. Experimental results show that the proposed ACM is capable of reducing the transmitted force by 20 dB up to the frequency range of 60 Hz.

• Structural Engineering and Mechanics
• /
• 제32권1호
• /
• pp.71-94
• /
• 2009

In this paper a procedure for Monte Carlo simulation of univariate stationary stochastic processes with the aid of neural networks is presented. Neural networks operate model-free and, thus, circumvent the need of specifying a priori statistical properties of the process, as needed traditionally. This is particularly advantageous when only limited data are available. A neural network can capture the "pattern" of a short observed time series. Afterwards, it can directly generate stochastic process realizations which capture the properties of the underlying data. In the present study a simple feed-forward network with focused time-memory is utilized. The proposed procedure is demonstrated by examples of Monte Carlo simulation, by synthesis of future values of an initially short single process record.

• Computers and Concrete
• /
• 제8권5호
• /
• pp.491-510
• /
• 2011

Before deciding if structures exposed to high temperature are to be repaired or demolished, their final state should be carefully examined. Destructive and non-destructive testing methods are generally applied for this purpose. Compressive strength and color change in mortars are observed as a result of the effects of high temperature. In this study, ordinary and pozzolan-added mortar samples were produced using different aggregates, and exposed to 100, 200, 300, 600, 900 and $1200^{\circ}C$. The samples were divided into two groups and cooled to room temperature in water and air separately. Compression tests were carried out on these samples, and the color change was evaluated by the Munsell Color System. The relationships between the change in compressive strength and color of mortars were determined by using a multi-layered feed-forward Neural Network model trained with the back-propagation algorithm. The results showed that providing accurate estimates of compressive strength by using the color components and ultrasonic pulse velocity design parameters were possible using the approach adopted in this study.

• International Journal of Automotive Technology
• /
• 제7권5호
• /
• pp.547-554
• /
• 2006

This paper is concerned with an optimal control suspension system using the preview information of road input based on a quarter car model. The main purpose of the control is to combine good vibration isolation characteristics with improved attitude control. The optimal control law is derived with the use of calculus of variation, consisting of three parts. The first part is a full state feedback term that includes integral control acting on the suspension deflection to ensure zero steady-state deflection in response to static body forces and ramp road inputs. The second part is a feed-forward term which compensates for the body forces when they can be detected, and the third part depends on previewed road input. The performance of the suspension is evaluated in terms of frequency domain characteristics and time responses to ramp road input and cornering forces. The effects of each part of the suspension controller on the system behavior are examined.

• 한국항공우주학회지
• /
• 제36권3호
• /
• pp.244-249
• /
• 2008

본 논문에서는 적분비례항법 유도방식을 사용하는 유도탄 유도루프에 대한 안정된 시간영역을 유도하였다. 고려된 유도탄 유도루프는 한 개의 선형 시불변요소와 시변 궤환이득으로 구성된 제어시스템으로 형식화될 수 있다. Circle criterion 을 적용함으로써 안정성이 보장되는 비행 잔여시간에 대한 한계치를 총 비행시간의 함수로 유도할 수 있었으며, Popov criterion 을 통해 얻는 결과에 비해 덜 보수적인 결과를 얻었다.

• Structural Engineering and Mechanics
• /
• 제26권4호
• /
• pp.377-392
• /
• 2007

A discrete sliding mode control (SMC) method based on hybrid model of neural network and nominal model is proposed to reduce the vibration of flexible structures, which is a robust active controller developed by using a sliding manifold approach. Since the thick boundary layer will reduce the virtue of SMC, the multilayer feed-forward neural network is adopted to model the uncertainty part. The neural network is trained by Levenberg-Marquardt backpropagation. The design objective of the sliding mode surface is based on the quadratic optimal cost function. In course of running, the input signal of SMC come from the hybrid model of the nominal model and the neural network. The simulation shows that the proposed control scheme is very effective for large uncertainty systems.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 1995년도 추계학술대회발표논문집; 서울대학교, 서울; 30 Sep. 1995
• /
• pp.325-333
• /
• 1995

Back-propagation is the most widely used algorithm for supervised learning in multi-layer feed-forward networks. However, back-propagation is very slow in convergence. In this paper, a new weight initialization method, called rough map initialization, in multi-layer perceptrons is proposed. To overcome the long convergence time, possibly due to the random initialization of the weights of the existing multi-layer perceptrons, the rough map initialization method initialize weights by utilizing relationship of input-output features with singular value decomposition technique. The results of this initialization procedure are compared to random initialization procedure in encoder problems and xor problems.