• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.9
• /
• pp.47-52
• /
• 2020

With the 4th industrial revolution, condition monitoring using machine learning techniques has become popular among researchers. An overload due to complex operations causes several irregularities in MOSFETs. This study investigated the acquired voltage to analyze the overcurrent effects on MOSFETs using a failure mode effect analysis (FMEA). The results indicated that the voltage pattern changes greatly when the current is beyond the threshold value. Several features were extracted from the collected voltage signals that indicate the health state of a switched-mode power supply (SMPS). Then, the data were reduced to a smaller sample space by using a principal component analysis (PCA). A robust machine learning algorithm, the support vector machine (SVM), was used to classify different health states of an SMPS, and the classification results are presented for different parameters. An SVM approach assisted by a PCA algorithm provides a strong fault diagnosis framework for an SMPS.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.11 no.6
• /
• pp.923-934
• /
• 1987

A flexible one-link robotic manipulator is modelled as a rotating cantilever beam with a hub and tip mass. An active control law is developed with consideration of the distributed flexibility of the arm. Equation of motion is derived by Hamilton's principle and, for modal control, represented as state variable form using Galerkin's mode summation method. Feedback coefficients are chosen to minimize the linear quadratic performance index(PI). To reconstruct the complete state vector from the measurements, an observer is proposed. In order to suppress vibration of the manipulator arm to desirable extent and to obtain accuracy of the positioning, weighting factor of input in PI is adjusted. Spillover effect due to the controller which controls several important modes is examined. Experiment is also performed to validate the theoretical analysis.

• Journal of computational fluids engineering
• /
• v.10 no.4 s.31
• /
• pp.51-58
• /
• 2005

The unsteady supersonic flow over two- and three-dimensional cavities has been analyzed by the integration of unsteady Reynolds-Averaged Navier-Stokes(RANS) with the k-$\omega$ turbulence model. The unsteady flow is characterized by the periodicity due to the mutual relation between the shear layer and the internal flow in the cavity. An explicit 4th order Runge-Kutta scheme and an upwind TVD scheme based on the flux vector split with the van Leer limiters are used for time and space discritizations, respectively. The cavity has a L/D ratio of 3 for two-dimensional case, and same L/D and W/D ratio of I for three-dimensional case. The Mach and Reynolds numbers are 1.5 and 450000 respectively. In the three-dimensional flow, the field is observed to oscillate in the 'shear layer mode' with a feedback mechanism that follows Rossiter's formula. In the two-dimensional simulation, the self-sustained oscillating flow has more violent fluctuation inside the cavity. The primary fluctuating frequencies of two- and three- dimensional flow agree very well with the 2nd mode of Rossiter's frequency. In the three-dimensional flow, the 1st mode of frequency could be seen.

• Proceedings of the KIPE Conference
• /
• 2019.11a
• /
• pp.96-97
• /
• 2019

Pulse Width Modulation (PWM) is a widely adopted technique to drive the motor using the voltage source inverters. Since they generate high frequency Common Mode (CM) Voltage, a high shaft voltage in induction motor is induced which leads to parasitic capacitive currents causing adverse effects such as premature deterioration of ball bearings and high levels of electromagnetic emissions. This paper presents an Active Cancellation Circuit (ACC) which can significantly reduce the CM voltage hence the common mode current in the three phase induction motor drives. In the proposed method the CM voltage is detected by the capacitors and applied to the frame of the motor to cancel the CM voltage hence the CM current. Unlike the conventional methods the proposed method does not insert the transformer in between the inverter and motor, a high power rating three phase transformer is not required and no losses associated with it. In addition the proposed method is applicable to any kind of PWM motor drives regardless of their PWM methods. The effectiveness of the proposed method is proved by the experiments with a three phase induction motor (1.1kW 415V/60Hz) combined with a three phase voltage source inverter modulated by the Space Vector Modulation (SVM).

• Journal of Power Electronics
• /
• v.15 no.1
• /
• pp.246-255
• /
• 2015

This paper presents experimental results and its assessment of a variable-speed wind power generation system (VSWPGS) using permanent magnet synchronous generator (PMSG) and boost chopper circuit (BCC). Experimental results are obtained by a test bench with a wind turbine emulator (WTE). WTE reproduces the behaviors of a windmill by using servo motor drives. The mechanical torque references to drive the servo motor are calculated from the windmill wing profile, wind velocity, and windmill rotational speed. VSWPGS using PMSG and BCC has three speed control modes for the level of wind velocity to control the rotational speed of the wind turbine. The control mode for low wind velocity regulates an armature current of generator with BCC. The control mode for middle wind velocity regulates a DC link voltage with a vector-controlled inverter. The control mode for high wind velocity regulates a pitch angle of the wind turbine with a pitch angle control system. The hybrid of three control modes extends the variable-speed range. BCC simplifies the maintenance of VSWPGS while improving reliability. In addition, VSWPGS using PMSG and BCC saves cost compared with VSWPGS using a PWM converter.

• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.556-570
• /
• 2021

The existing video expression recognition methods mainly focus on the spatial feature extraction of video expression images, but tend to ignore the dynamic features of video sequences. To solve this problem, a multi-mode convolution neural network method is proposed to effectively improve the performance of facial expression recognition in video. Firstly, OpenFace 2.0 is used to detect face images in video, and two deep convolution neural networks are used to extract spatiotemporal expression features. Furthermore, spatial convolution neural network is used to extract the spatial information features of each static expression image, and the dynamic information feature is extracted from the optical flow information of multiple expression images based on temporal convolution neural network. Then, the spatiotemporal features learned by the two deep convolution neural networks are fused by multiplication. Finally, the fused features are input into support vector machine to realize the facial expression classification. Experimental results show that the recognition accuracy of the proposed method can reach 64.57% and 60.89%, respectively on RML and Baum-ls datasets. It is better than that of other contrast methods.

• Wind and Structures
• /
• v.38 no.1
• /
• pp.75-91
• /
• 2024

Reliable wind signal reconstruction can be beneficial to the operational safety of long-span bridges. Non-Gaussian characteristics of wind signals make the reconstruction process challenging. In this paper, non-Gaussian wind signals are converted into a combined prediction of two kinds of features, actual wind speeds and wind angles of attack. First, two decomposition techniques, empirical mode decomposition (EMD) and variational mode decomposition (VMD), are introduced to decompose wind signals into intrinsic mode functions (IMFs) to reduce the randomness of wind signals. Their principles and applicability are also discussed. Then, four artificial intelligence (AI) algorithms are utilized for wind signal reconstruction by combining the particle swarm optimization (PSO) algorithm with back propagation neural network (BPNN), support vector regression (SVR), long short-term memory (LSTM) and bidirectional long short-term memory (Bi-LSTM), respectively. Measured wind signals from a bridge site in a deep-cutting gorge are taken as experimental subjects. The results showed that the reconstruction error of high-frequency components of EMD is too large. On the contrary, VMD fully extracts the multiscale rules of the signal, reduces the component complexity. The combination of VMD-PSO-Bi-LSTM is demonstrated to be the most effective among all hybrid models.

• Journal of Korean Society for Geospatial Information Science
• /
• v.2 no.1 s.3
• /
• pp.99-106
• /
• 1994

Raster data mode and Vector data are the two major model in geographic information systems. These two data models are difficult to be intergrated because of their differences in structures and properties. Almost all of the current GIS systems process in one data model by converting one data type to another type. So. the loss and change of information caused by data conversiion degrades the accuracy of data. In this paper, we propose a new data model which can process two data models without conversion. We use quadtree for raster data and topological vector model for vector data. The output is formed as raster data model of quadtree. We can get more accurate overay output, and this intergrated model is more suitable for data like forest, landuses, soils that consist of classes which have small distribution changes.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.29 no.1
• /
• pp.22-31
• /
• 2001

The purpose of this study is to compare calculated pollutant loadings using pollutant load unit factors and vector type coverage, and expected mean concentration(EMC) and raster type of digital elevation model(DEM). This study is also focusing on comparison of the advantages and the disadvantages of the two methods, and seeking for a method of calculation of pollutant loadings using DEM. Estimation of pollutant inputs using pollutant load unit factors has limitations in identifying seasonal variations of pollutant loadings. Seasonal changes of runoffs should be considered in the calculation of pollutant loadings from catchments into reservoirs. Evaluation of pollutant inputs using runoff-coefficient and EMC can overcome these drawbacks. Proper EMC and runoff-coefficient values for the Koeup stream catchments of the Koheung estuarine lake were drawn from review of related papers. Arc/Info was employed to establish database of spatial and attribute data of point and non-point pollutant sources and characteristics of the catchments. ArcView was used to calculate point and non-point pollutant loadings. Pollutant loads estimated with either unit factors-coverages, i.e., pollutant load unit factors and vector coverages f point sources and land use, or EMC and digital elevation mode(DEM) were compared with stream monitoring loads. We have found that some differences were shown between monitoring results and estimated loads by Unit Factors-Coverage and EMC-DEM. Monthly variations of pollutant loads evaluated with EMC-DEM were similar to those with monitoring result. The method using EMC-DEM can calculate accumulated flows and pollutant loads and can be utilized to identify stream networks. A future research on correcting the difference between vector type stream using flow direction grid and digitalizing vector type should be conducted in order to obtain more exact calculation of pollutant loadings.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.4
• /
• pp.1730-1747
• /
• 2018

Multi-view video plus depth (MVD) is a mainstream format of 3D scene representation in free viewpoint video systems. The advanced 3D extension of the high efficiency video coding (3D-HEVC) standard introduces new prediction tools to improve the coding performance of depth video. However, the depth video in 3D-HEVC is time consuming. To reduce the complexity of the depth video inter coding, we propose a fast coding unit (CU) size and mode decision algorithm. First, an off-line trained Bayesian model is built which the feature vector contains the depth levels of the corresponding spatial, temporal, and inter-component (texture-depth) neighboring largest CUs (LCUs). Then, the model is used to predict the depth level of the current LCU, and terminate the CU recursive splitting process. Finally, the CU mode search process is early terminated by making use of the mode correlation of spatial, inter-component (texture-depth), and inter-view neighboring CUs. Compared to the 3D-HEVC reference software HTM-10.0, the proposed algorithm reduces the encoding time of depth video and the total encoding time by 65.03% and 41.04% on average, respectively, with negligible quality degradation of the synthesized virtual view.