• 디지털융복합연구
• /
• 제11권6호
• /
• pp.169-174
• /
• 2013

음성 인식 성능 저하의 문제는 모델 훈련 환경과 인식 환경의 차이이다. 이러한 환경의 불일치를 줄이기 위한 방법으로 다양한 묵음 특징 정규화 방법을 사용하고 있다. 기존의 묵음 특징 정규화 방법은 낮은 신호 대 잡음비에서 묵음 구간의 에너지 레벨이 증가하여 음성과 비음성에 대한 분류의 정확도가 떨어짐으로 인해 인식 성능이 저하되는 문제점이 있다. 본 논문에서는 음성 에너지 최대화와 묵음 특징 정규화를 이용한 잡음 환경에 강인한 음성 검출 방법을 제안하였다. 제안한 방법은 높은 신호 대 잡음비에서는 음성 에너지를 최대화시켜 특징이 잡음의 영향을 적게 받는 특성을 이용하였고 낮은 신호 대 잡음비에서는 음성/비음성의 켑스트럼 특징 분포 특성을 이용하여 인식 성능을 향상시켰다. 인식 실험 결과 기존 방법에 비해 향상된 인식 성능을 확인할 수 있었다.

• 한국지진공학회:학술대회논문집
• /
• 한국지진공학회 1997년도 추계 학술발표회 논문집 Proceedings of EESK Conference-Fall 1997
• /
• pp.217-224
• /
• 1997

One way in which a comparison can be mode between various reinforced concrete structures is to compare the rate of energy absorption capacity. It is useful to use a well-known standard hysteretic rule as a benchmark for comparisons. The concept of energy absorption efficiency with respect to an elaste-perfectly plastic (EPP) system has been adopted. The normalized cumulative energy, cumulative plastic drift and energy response spectra are used for the method. The previous study using the energy spectra developed by Chang and Mander (1994) indicates the cumulative plastic demand for most earthquakes to be 0.1 rad., but a conservative upper bound of 0.2 rad. could be expected for a maximum credible earthquake. From the present study, the energy absorption efficiency in R.C. structures with respect to the EPP system may range in 30%-45% for most cases.

• 한국태양에너지학회 논문집
• /
• 제33권5호
• /
• pp.18-23
• /
• 2013

This paper presents a simple but valid loss calculation method of grid-connected photovoltaic system based on normalized yield model. The proposed method can be represented as a quantitative value for five losses and performance of grid-connected photovoltaic system with three years monitored data. These results will indicate that it is useful to investigate various loss factors causing the performance obstruction, enhance the lifetime yield for changing meteorological conditions, and determine the optimal design and performance improvement of grid-connected photovoltaic system.

• Journal of Electrical Engineering and Technology
• /
• 제11권2호
• /
• pp.519-528
• /
• 2016

In order to achieve fast computational speed with good visual quality of output video, we propose a frequency domain based new fractal video compression scheme. Normalized cross correlation is used to find the structural self similar domain block for the input range block. To increase the searching speed, cross correlation is implemented in the frequency domain using FFT with one computational operation for all the domain blocks instead of individual block wise calculations. The encoding time is further minimized by applying rotation and reflection DFT properties to the IFFT of zero padded range blocks. The energy of overlap small size domain blocks is pre-computed for the entire reference frame and retaining the energies of the overlapped search window portion of previous adjacent block. Quadtree decompositions are obtained by using domain block motion compensated prediction error as a threshold to control the further partitions of the block. It provides a better level of adaption to the scene contents than fixed block size approach. The result shows that, on average, the proposed method can raise the encoding speed by 48.8 % and 90 % higher than NHEXS and CPM/NCIM algorithms respectively. The compression ratio and PSNR of the proposed method is increased by 15.41 and 0.89 dB higher than that of NHEXS on average. For low bit rate videos, the proposed algorithm achieve the high compression ratio above 120 with more than 31 dB PSNR.

• 한국수산과학회지
• /
• 제53권2호
• /
• pp.218-230
• /
• 2020

The single anatomical attribute that has the greatest influence on acoustic scattering from fish is the presence or absence of a swimbladder. This study examined the effect of removing the gas from the swimbladder on the broadband backscattering characteristics of six species of swimbladdered fish: striped beakperch Oplegnathus fasciatus, black scraper Thamnaconus modestus, dark-banded rockfish Sebastes inermis, goldeye rockfish Sebastes thompsoni, black rockfish Sebastes schlegeli and panther puffer Takifugu pardalis. Before and after removing the gas from the swimbladder, the species-specific, frequency-dependent backscattered echo signals from anaesthetized individuals of each fish species were measured at approximately 1° intervals spanning a 90° aspect angle range from -45° (head down) to +45° (head up) using a broadband echo sounder operating at 100-200 kHz. The relationship between the wavelength-normalized backscattering cross section (σ/λ²) and fish length (L/λ) was calculated for each species. The average σ/λ² value for the six fish species at a L/λ range of 19.79-25.85, with a mean of 22.89, was reduced by approximately 52.3% when the gas was removed.

• 한국농공학회지
• /
• 제44권1호
• /
• pp.103-115
• /
• 2002

During this study, various stress path tests in previous isotropic and anisotropic (compression and tension) stress histories are performed on weathered granite soil sampled at Iksan, Jeonbuk. Yielding points are determined from various stress-strain curves(stress ratio-shear strain, volumetric strain, normalized energy and dissipated total energy curves). The shape and characteristics of isotropic and anisotropic yielding curves are examined. The main results are summarized as follows . 1) Yielding curries defined from stress ratio - normarized energy and dissipated total energy curves show almost perfect ellipse. 2) Directions of plastic strain incremental vector are not perpendicular to yielding curve. 3) Normarized energy and dissipated total energy spread with similar tendency with respect to yielding currie in stress space.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2004년도 학술대회 논문집 정보 및 제어부문
• /
• pp.319-321
• /
• 2004

The average glandular dose (AGD) is determined by the breast entrance skin exposure, x-ray tube target material, beam quality (half-value layer), breast thickness, and breast composition. Almost breast cancer always arises in glandular breast tissue. As a result, the average radiation absorbed dose to glandular tissue is the preferred measure of the radiation risk associated with mammography. If the normalized average glandular dose is known, the average glandular dose can be computed from the product of the normalized average glandular dose and breast entrance skin exposure. In this study, AGD was calculated by the breast thickness and various x-ray energy (HVL) in 50% glandular 50% adipose breast by Mo.-Rh. assembly. AGD is 84 mrad in compressed 5 cm breast. These results show that as increasing the breast thickness, dose also increases. But as increasing the x-ray tube voltage, dose decreases because of high penetrating ratio through the object. But high tube voltage is reducing the subject contrast. From this result, we have to consider the trade-off between subject contrast of image and dose to the patient and choose proper x-ray energy range.

• Nuclear Engineering and Technology
• /
• 제55권4호
• /
• pp.1571-1584
• /
• 2023

In this study, a numerical analysis was performed as part of an international joint research project to reproduce a real cable tray fire that occurred in the heater bay area of the turbine building of a nuclear power plant. A sensitivity analysis was performed on various input parameters to derive results consistent with the sprinkler activation time obtained from the fire event analysis. For all sensitive parameters, the normalized sprinkler activation time correlated well with the power function of the normalized sprinkler height. A correlation equation was developed to identify the sprinkler activation time at any location when determining the slope or fire growth rate under the conditions assuming a linear or t-squared heat release rate (HRR) time curve. Various cable fire growth assumptions were used to determine which assumption was better to provide the prediction coincident with the information given from the fire event analysis in terms of the sprinkler activation time and total energy generated from cables damaged by fire. In the comprehensive analysis of all the sensitive parameters, the standard deviation of the input parameters increased as the sprinkler height decreased. Within the range of the sensitivity parameter values given in this study, when considering all sprinkler heights, the standard deviation of the cable model change was the largest and that of the overhang position change was the smallest.

• Journal of Power Electronics
• /
• 제17권4호
• /
• pp.849-859
• /
• 2017

This paper illustrates the analysis and design of a multi-resonant converter applied to an electric vehicle (EV) charger. Thanks to the notch resonant characteristic, the multi-resonant converter achieve soft switching and operate with a narrowed switching frequency range even with a wide output voltage range. These advantages make it suitable for battery charging applications. With two more resonant elements, the design of the chosen converter is more complex than the conventional LLC resonant converter. However, there is not a distinct design outline for the multi-resonant converters in existing articles. According to the analysis in this paper, the normalized notch frequency $f_{r2n}$ and the second series resonant frequency $f_{r3n}$ are more sensitive to the notch capacitor ratio q than the notch inductor ratio k. Then resonant capacitors should be well-designed before the other resonant elements. The peak gain of the converter depends mainly on the magnetizing inductor ratio $L_n$ and the normalized load Q. And it requires a smaller $L_n$ and Q to provide a sufficient voltage gain $M_{max}$ at ($V_{o\_max}$, $P_{o\_max}$). However, the primary current increases with $(L_nQ)^{-1}$, and results in a low efficiency. Then a detailed design procedure for the multi-resonant converter has been provided. A 3.3kW prototype with an output voltage range of 50V to 500V dc and a peak efficiency of 97.3 % is built to verify the design and effectiveness of the converter.

• 전기학회논문지
• /
• 제61권3호
• /
• pp.394-400
• /
• 2012

In this paper, charging modes of series resonant converter for a high voltage energy storage capacitor are compared in terms of charging time, peak resonant current, normalized peak resonant current and voltage in each operation mode. Operating principles of the full bridge series resonant converter with capacitor load are explained and analyzed in discontinuous and continuous operation mode. Based on the analysis and simulation result, $0.6{\omega}_r$ < ${\omega}_s$ < $0.75{\omega}_r$ and $1.3{\omega}_r$ < ${\omega}_s$ < $1.4{\omega}_r$ are evaluated to the best range of switching frequency for charging of an high voltage energy storage capacitor. 1.8 kJ/s SRC prototype is assembled with TI 28335 DSP controller and 40 kJ, 7 kV energy storage capacitor. Design rules based on the comparative analysis are verified by experiment.