• Transactions on Electrical and Electronic Materials
• /
• v.18 no.5
• /
• pp.287-297
• /
• 2017

Novel power system stabilizers (PSSs) have been proposed to effectively dampen low frequency oscillations (LFOs) in multi-machine power systems and have attracted increasing research interest in recent years. Due to this attention, recently, fractional order controllers (FOCs) have found new applications in power system stability issues. Here, a tilt-integral-derivative power system stabilizer (TID-PSS) is proposed to enhance the dynamic stability of a multi-machine power system by providing additional damping to the LFOs. The TID is an extended version of the classical proportional-integral-derivative (PID) applying fractional calculus. The design of the proposed three-parameter tunable TID-PSS is systematized as a nonlinear time domain optimization problem in which the tunable parameters are adjusted concurrently using a modified group search optimization (MGSO) algorithm. An integral of the time multiplied squared error (ITSE) performance index is considered as the objective function. The proposed stabilizer is simulated in the MATLAB/SIMULINK environment using the FOMCON toolbox and the dynamic performance is evaluated on a 3-machine 6-bus power system. The TID-PSS is compared with both classical PID-PSS (PID-PSS) and conventional PSS (CPSS) using eigenvalue analysis and time domain simulations. Sensitivity analyses are performed to assess the robustness of the proposed controller against large changes in system loading conditions and parameters. The results indicate that the proposed TID-PSS provides the better dynamic performance and robustness compared with the PID-PSS and CPSS.

• Earthquakes and Structures
• /
• v.11 no.6
• /
• pp.943-966
• /
• 2016

Recently, magnetorheological elastomer (MRE) material and its devices have been developed and attracted a good deal of attention for their potentials in vibration control. Among them, a highly adaptive base isolator based on MRE was designed, fabricated and tested for real-time adaptive control of base isolated structures against a suite of earthquakes. To perfectly take advantage of this new device, an accurate and robust model should be built to characterize its nonlinearity and hysteresis for its application in structural control. This paper first proposes a novel hysteresis model, in which a nonlinear hyperbolic sine function spring is used to portray the strain stiffening phenomenon and a Voigt component is incorporated in parallel to describe the solid-material behaviours. Then the fruit fly optimization algorithm (FFOA) is employed for model parameter identification using testing data of shear force, displacement and velocity obtained from different loading conditions. The relationships between model parameters and applied current are also explored to obtain a current-dependent generalized model for the control application. Based on the proposed model of MRE base isolator, a second-order sliding mode controller is designed and applied to the device to provide a real-time feedback control of smart structures. The performance of the proposed technique is evaluated in simulation through utilizing a three-storey benchmark building model under four benchmark earthquake excitations. The results verify the effectiveness of the proposed current-dependent model and corresponding controller for semi-active control of MRE base isolator incorporated smart structures.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.1
• /
• pp.109-115
• /
• 2010

A main technology of opening and closing a sluice gate is accurate synchronous and position control for the two cylinders when they are moving with the sluice gate together over 10[m]. Since the supply flow and supply pressure of cylinders are not constant and a nonlinear friction force of the piston in cylinders exists, a difference will be made between the displacement of two cylinders. This difference causes the sluice gate to deform and abrade, and even it may be out of order. In order to solve this problem we design two kinds of fuzzy PI controllers. The former is for a position control of two cylinders, the latter is for their synchronous control. We show some simulation results compare the performance of fuzzy PI controller to the conventional PID controller.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.2
• /
• pp.243-250
• /
• 2010

As the wind has become one of the fastest growing renewable energy sources, the key issue of wind energy conversion systems is how to efficiently operate the wind turbines in a wide range of wind speeds. Compared to fixed speed turbines, variable speed wind turbines feature higher energy yields, lower component stress and fewer grid connection power peaks. Generally, measurement of wind speed is required for the control of variable speed wind turbine system. However, wind speed measured by anemometers is not accurate owing to various reasons. In this work, a new control algorithm for variable speed wind turbine system based on Kalman filter which can be used for the estimation of wind speed and artificial neural network which can generate optimum rotor speed is proposed. Also, to verify the feasibility of the proposed scheme, various simulation studies are carried out by using Simulink in Matlab.

• The Journal of the Acoustical Society of Korea
• /
• v.19 no.1
• /
• pp.4-10
• /
• 2000

The work presented in this paper is about a postprocessor, which improves the performance of automatic speech segmentation system by correcting the phoneme boundary errors. We propose a postprocessor that reduces the range of errors in the auto labeled results that are ready to be used directly as synthesis unit. Starting from a baseline automatic segmentation system, our proposed postprocessor trains the features of hand labeled results using multi-layer perceptron(MLP) algorithm. Then, the auto labeled result combined with MLP postprocessor determines the new phoneme boundary. The details are as following. First, we select the feature sets of speech, based on the acoustic phonetic knowledge. And then we have adopted the MLP as pattern classifier because of its excellent nonlinear discrimination capability. Moreover, it is easy for MLP to reflect fully the various types of acoustic features appearing at the phoneme boundaries within a short time. At the last procedure, an appropriate feature set analyzed about each phonetic event is applied to our proposed postprocessor to compensate the phoneme boundary error. For phonetically rich sentences data, we have achieved 19.9 % improvement for the frame accuracy, comparing with the performance of plain automatic labeling system. Also, we could reduce the absolute error rate about 28.6%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.10a
• /
• pp.463-466
• /
• 2008

With the development of Information Technology in recent years, the image has been an important means to store or express information. Generally, during the process of acquiring and storing images, the images can be corrupted by noise of which typical types are Impulse(Impulse Noise) and AWGN(Addiction White Gaussian Noise). Impulse noise shows irregularly in black and white over the length and breadth of the image by sharp and sudden disturbance of the image signal. In the Impulse noise environment, SM(Standard Median) filter would be used because of its good noise removal performance and simple algorithm. However, when SM filter removes noise, it also produces error at the edge of image and causes whole image quality deterioration. In this paper, we propose a method based on modified nonlinear filter operation scheme which enhances the features of noise removal and detail image preservation when restoring image in Impulse noise environment. And, we compared it with existing methods and the performances through simulation.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.7 no.5
• /
• pp.427-436
• /
• 2002

This paper presents an implementation of high-dynamic performance of position sensorless motion control system of Reluctance Synchronous Motor(RSM) drives for an industrial servo system with direct torque control(DTC). The problems of high-dynamic performance and maximum efficiency RSM drives controlled by DTC are saturation of stator linkage flux and nonlinear inductance characteristics with various load currents. The accurate estimation of the stator flux and torque are obtained using stator flux observer of which a saturated inductance $L_d$ and $L_q$ can be compensated by adapting from measurable the modulus of the stator current and rotor position. To obtain fast torque response and maximum torque/current with varying load current, the reference command flux is ensured by imposing $I_{ds} = I_{qs}$. This control strategy is proposed to achieve fast response and optimal efficiency for RSM drive. In order to prove rightness of the suggested control algorithm, the actual experiment carried out at $\pm$20 and $\pm$1500 rpm. The developed digitally high-performance motion control system shown good response characteristic of control results and high performance features using 1.0kW RSM which has 2.57 Ld/Lq salient ratio.

• Journal of KIISE
• /
• v.45 no.1
• /
• pp.22-29
• /
• 2018

Over the decades, neural networks have been successfully used in numerous applications from speech recognition to image classification. However, these neural networks cannot explain their results and one needs to know how and why a specific conclusion was drawn. Most studies focus on extracting binary rules from neural networks, which is often impractical to do, since data sets used for machine learning applications contain continuous values. To fill the gap, this paper presents an algorithm to extract logic rules from a trained neural network for data with continuous attributes. It uses hyperplane-based linear classifiers to extract rules with numeric values from trained weights between input and hidden layers and then combines these classifiers with binary rules learned from hidden and output layers to form non-linear classification rules. Experiments with different datasets show that the proposed approach can accurately extract logical rules for data with nonlinear continuous attributes.

• Journal of Astronomy and Space Sciences
• /
• v.29 no.4
• /
• pp.351-362
• /
• 2012

We made a study on real-time determination method for relative position using the laser-measured distance data between satellites. We numerically performed the determination of relative position in accordance with extended Kalman filter algorithm using the vectors obtained through nonlinear equation of relative motion, laser simulator for distance measurement, and attitude determination of chief satellite. Because the spherical parameters of relative distance and direction are used, there occur some changes in precision depending on changes in relative distance when determining the relative position. As a result of simulation, it was possible to determine the relative position with several millimeter-level errors at a distance of 10 km, and sub-millimeter level errors at a distance of 1 km. In addition, we performed the determination of relative position assuming the case that global positioning system data was not received for long hours to see the impact of determination of chief satellite orbit on the determination of relative position. The determination of precise relative position at a long distance carried out in this study can be used for scientific mission using the satellite formation flying.

• Korea-Australia Rheology Journal
• /
• v.13 no.1
• /
• pp.29-35
• /
• 2001

The development of filament stretching extensional rheometers over the past decade has enabled the systematic measurement of the transient extensional stress growth in dilute and semi-dilute polymer solutions. The strain-hardening in the extensional viscosity of dilute solutions overwhelms the perturbative effects of capillarity, inertia & gravity and the kinematics of the extensional deformation become increasingly homogeneous at large strains. This permits the development of a robust open-loop control algorithm for rapidly realizing a deformation with constant stretch history that is desired for extensional rheometry. For entangled fluids such as concentrated solutions and melts the situation is less well defined since the material functions are governed by the molecular weight between entanglements, and the fluids therefore show much less pronounced strain-hardening in transient elongation. We use experiments with semi-dilute/entangled and concentrated/entangled monodisperse polystyrene solutions coupled with time-dependent numerical computations using nonlinear viscoelastic constitutive equations such as the Giesekus model in order to show that an open-loop control strategy is still viable for such fluids. Multiple iterations using a successive substitution may be necessary, however, in order to obtain the true transient extensional viscosity material function. At large strains and high extension rates the extension of fluid filaments in both dilute and concentrated polymer solutions is limited by the onset of purely elastic instabilities which result in necking or peeling of the elongating column. The mode of instability is demonstrated to be a sensitive function of the magnitude of the strain-hardening in the fluid sample. In entangled solutions of linear polymers the observed transition from necking instability to peeling instability observed at high strain rates (of order of the reciprocal of the Rouse time for the fluid) is directly connected to the cross-over from a reptative mechanism of tube orientation to one of chain extension.