• Journal of Korea Water Resources Association
• /
• v.46 no.3
• /
• pp.219-227
• /
• 2013

Time series data including change, jump, trend and periodicity generally have nonstationarity. Especially, various methods have been proposed to identify the trend about hydrological time series data. However, among various methods, evaluation about capability of each trend test has not been done a lot. Even for the same data, each method may show the different result. In this study, the simulation was performed for identification about the changes in trend analysis according to the statistical characteristics and the capability in the trend analysis. For this purpose, power test for the trend analysis is conducted using Men-Kendall test, Hotelling-Pabst test, t test and Sen test according to the slope, sample size, standard deviation and significance level. As a result, t test has higher statistical power than the others, while Mann-Kendall, Hotelling-Pabst, and Sen tests were similar results.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.3 no.4
• /
• pp.307-314
• /
• 1998

A novel high power factor Dual Half Bridge Series Resonant Inverter (DHB-SRI) for an induction heating appliance with multiple loads is proposed to remove the interferential acoustic noise caused by the difference between operating frequencies of adjacent loads. The circuit enables independent full power range control of two induction heating elements by one inverter circuit and has minimum switching losses due to the zero voltage switching characteristic. According to the mode analysis, I will explain the operation of the proposed circuit. To evaluate the required cooling capacity, loss analysis is performed through deriving some loss equations. In order to obtain the power factor correction capability and to lessen the system size, suitable design guides are given. Using the designed values, the proto-type circuit with 2.8kW power consumption for each induction heating element is built and tested to verify the operation of the proposed circuit.

• Journal of Digital Convergence
• /
• v.11 no.10
• /
• pp.305-311
• /
• 2013

The size of the market as the economy continues to evolve, in order to make the right decisions to accurately predict the economic problems the market has emerged as an important issues. To express the modern economic system, the largest of the various economic indicators, pillars stock indicators analysis and decision-making with a proper understanding of the problem for the application of the model is suitable for time-series data concealment HMM. Based on this time series model and the calculation of the time and cost savings dimension decrease techniques for the estimation and prediction of the model was applied to the problem was to verify the validity. As a result, the model predictions in both the short term rather than long-term predictions of the model estimates the optimal predictive value similar pattern very similar to both the actual data and was able to confirm that.

• Journal of Families and Better Life
• /
• v.10 no.2
• /
• pp.19-36
• /
• 1992

The purpose of this paper is to analyze empirically the tendency of household consumption expenditure according to the change of social and economical condition, and the factor which influences consumption expenditure of urban household. The data used in analysis are time-series. The data are statistic form Urban Household Economy Survey published by the Economic Planning Board, dating form the first quarter of 1970 to the fourth quarter of 1989. The income of household and consumption expenditure materials were deflated as consumer price index to exclude the influence of prices and the influence of household composition are considered to deflated as the size of the household under assumption of homogeneity. The consumption expenditure items were categorized to 12 relatively large range items. The time-series data were analyzed by using the Two Stage Least Squares and the Ordinary Least Squares. The following is the result of analysis. 1) Rather than the income increase of previous years. the average income increase for two years influences more significantly on consumption expenditure of household. In the case of influence on consumption expenditure for each item by increase in disposable income, such categories as furniture and utensils. clothing and footwear, housing, medical care, culture and recreation, and transportation and communication have significant influence. 2) Among consumption expenditure categories, the increasing factors were furniture and utensils, and clothing and footwear. And the decreasing factors were housing, medical care, culture and recreation ,and transportation and communication. The relative prices, however, had significant influence on categories such as housing, furniture and utensils, medical care , culture and recreation, and transportation and communication and all of them were the decreation factors. 3) Among with changes of social and economical conditions, miscellaneous showed the highest increase in marginal propensity to consume and foods was the lowest. Also culture and recreation and housing brought up a great change of the income elasticity of demand.

• Journal of Power Electronics
• /
• v.7 no.3
• /
• pp.181-190
• /
• 2007

This paper presents a novel circuit topology of a DC bus line series switch and parallel snubbing capacitor-assisted soft-switching PWM full-bridge inverter type DC-DC power converter with a high frequency planar transformer link, which is newly developed for high performance arc welding machines in industry. The proposed DC-DC power converter circuit is based upon a voltage source-fed H type full-bridge soft-switching PWM inverter with a high frequency transformer. This DC-DC power converter has a single power semiconductor switching device in series with an input DC low side rail and loss less snubbing capacitor in parallel with the inverter bridge legs. All the active power switches in the full-bridge arms and DC bus line can achieve ZCS turn-on and ZVS turn-off transition commutation. Consequently, the total switching power losses occurred at turn-off switching transition of these power semiconductor devices; IGBTs can be reduced even in higher switching frequency bands ranging from 20 kHz to 100 kHz. The switching frequency of this DC-DC power converter using IGBT power modules can be realized at 60 kHz. It is proved experimentally by power loss analysis that the more the switching frequency increases, the more the proposed DC-DC power converter can achieve a higher control response performance and size miniaturization. The practical and inherent effectiveness of the new DC-DC converter topology proposed here is actually confirmed for low voltage and large current DC-DC power supplies (32V, 300A) for TIG arc welding applications in industry.

• Journal of the Korea Society for Simulation
• /
• v.30 no.4
• /
• pp.1-8
• /
• 2021

Deep learning is one of the most widely accepted methods for the forecasting of time series data which have the complexity and non-linear behavior. In this paper, we investigate the modification of a state-of-art WaveNet deep learning architecture and walk forward validation (WFV) in order to forecast electric power consumption data 24-hour-ahead. WaveNet originally designed for raw audio uses 1D dilated causal convolution for long-term information. First of all, we propose a modified version of WaveNet which activates real numbers instead of coded integers. Second, this paper provides with the training process with tuning of major hyper-parameters (i.e., input length, batch size, number of WaveNet blocks, dilation rates, and learning rate scheduler). Finally, performance evaluation results show that the prediction methodology based on WFV performs better than on the traditional holdout validation.

• Journal of electromagnetic engineering and science
• /
• v.19 no.1
• /
• pp.64-70
• /
• 2019

In this study, a compact series-fed center-fed open-stub (SFCFOS) linear array antenna for K-band applications is presented. The antenna is composed of a single-line 10-element linear array. A symmetrical Chebyshev amplitude distribution (CAD) is used to obtain a low sidelobe characteristic against a uniform amplitude distribution (UAD). The amplitude is controlled by varying the width of the microstrip patch elements, and open-ended stubs are arranged next to the last antenna element to use the energy of the radiating signal more effectively. We insert a series-fed stub between two patches and obtain a low mutual coupling for a 4.28-mm center-to-center spacing ($0.7{\lambda}$ at 21 GHz). A prototype of the antenna is fabricated and tested. The overall size of the uniform linear array is $7.04{\times}1.05{\times}0.0563{\lambda}_g^3$ and that of the Chebyshev linear array is $9.92{\times}1.48{\times}0.0793{\lambda}_g^3$. The UAD array yields a ${\mid}S_{11}{\mid}$ < -10 dB bandwidth of 1.33% (20.912-21.192 GHz) and 1.45% (20.89-21.196 GHz) for the CAD. The uniform array design gives a -23 dB return loss, and the Chebyshev array achieves a -30.68 dB return loss at the center frequency with gains of 15.3 dBi and 17 dBi, respectively. The simulated and measured results are in good agreement.

• Korean Journal of Agricultural Science
• /
• v.48 no.4
• /
• pp.843-853
• /
• 2021

The objectives of this paper were to simply estimate soil loss levels as caused by wind in reclaimed tidal flat land (RTFL) and the threshold wind velocity in the RTFL. For this experiment, RTFL located at Haenam Bay was selected and a total of 150 soil samples were collected at the Ap horizon from the five soil series. The particle distribution curves, including the limit of the non-erodible particle size (D > 0.84 mm) for each Ap horizon soil, show that the proportions of non-erodible particle sizes that exceeded 0.84 mm were 4.3% (Taehan, TH), 8.9% (Geangpo, GP), 0.5% (Bokchun, BC), 1.6% (Poseung, PS) and 1.4% (Junbook, JB), indicating that the amount of non-erodible soil particles increased with an increase in the sand content. The average monthly, daily and instantaneous wind velocities were higher than the threshold friction velocity (TFV) calculated according to the dynamic velocity (V_d) by Bagnold, while the average monthly wind velocity was lower than those of the TFV suggested by the revised wind erosion equation (RWEQ) and wind erosion prediction system (WEPS). The susceptible proportions of erodible soil particles from the Ap horizon soil samples from each soil series could be significantly influenced by the proportion of sand particles between 0.025 and 0.5 mm (or 0.84 mm) in diameter regardless of the threshold wind velocity. Thus, further investigations are needed to estimate more precisely soil erosion in RTFL, which shows various soil characteristics, as these estimations of soil loss in the five soil series were obtained only when considering wind velocities and soil textures.

• Journal of Navigation and Port Research
• /
• v.47 no.6
• /
• pp.376-385
• /
• 2023

Fishing boat capsizing accidents account for more than half of all capsize accidents. These can occur for a variety of reasons, including inexperienced operation, bad weather, and poor maintenance. Due to the size and influence of the industry, technological complexity, and regional diversity, fishing ships are relatively under-researched compared to commercial ships. This study aimed to predict the rolling motion time series of fishing boats using an image-based deep learning model. Image-based deep learning can achieve high performance by learning various patterns in a time series. Three image-based deep learning models were used for this purpose: Xception, ResNet50, and CRNN. Xception and ResNet50 are composed of 177 and 184 layers, respectively, while CRNN is composed of 22 relatively thin layers. The experimental results showed that the Xception deep learning model recorded the lowest Symmetric mean absolute percentage error(sMAPE) of 0.04291 and Root Mean Squared Error(RMSE) of 0.0198. ResNet50 and CRNN recorded an RMSE of 0.0217 and 0.022, respectively. This confirms that the models with relatively deeper layers had higher accuracy.

• Smart Media Journal
• /
• v.13 no.6
• /
• pp.24-34
• /
• 2024

If the future can be predicted from network traffic data, which is a time series, it can achieve effects such as efficient resource allocation, prevention of malicious attacks, and energy saving. Many models based on statistical and deep learning techniques have been proposed, and most of these studies have focused on improving model structures and learning algorithms. Another approach to improving the prediction performance of the model is to obtain a good-quality data. With the aim of obtaining a good-quality data, this paper applies a dense sampling technique that augments time series data to the application of network traffic prediction and analyzes the performance improvement. As a dataset, UNSW-NB15, which is widely used for network traffic analysis, is used. Performance is analyzed using RMSE, MAE, and MAPE. To increase the objectivity of performance measurement, experiment is performed independently 10 times and the performance of existing sparse sampling and dense sampling is compared as a box plot. As a result of comparing the performance by changing the window size and the horizon factor, dense sampling consistently showed a better performance.