• 제목/요약/키워드: Long-Input Transformer

검색결과 14건 처리시간 0.027초

전력용 변압기 내부고장시 전압-차전류의 변화에 관한 연구 (Transition of voltage-differential current under internal fault on power transformer)

  • 박재세
    • 대한전기학회:학술대회논문집
    • /
    • 대한전기학회 2004년도 학술대회 논문집 전문대학교육위원
    • /
    • pp.92-95
    • /
    • 2004
  • Power transformer is an important apparatus in transforming and delivering the power in a power system. It shows less accident ratio than other system apparatus, but once the accident occurs, it causes long-term operation stoppage and economic loss. It brings high bad spillover effects. Therefore, the role of protective relaying, which is to prevent internal fault a power transformer is highly important. This study proposed advanced algorithm that can clearly determine internal fault of the power transformer and magnetizing inrush, through numerical analysis by using the terminal voltage and input output current.

  • PDF

대화 맥락에 기반한 한국어 휴지 예측 모델 (Korean Pause Prediction Model based on Dialogue Context)

  • 이정;나정호;정정범;최맹식;이충희;나승훈
    • 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
    • /
    • 한국정보과학회언어공학연구회 2023년도 제35회 한글 및 한국어 정보처리 학술대회
    • /
    • pp.404-408
    • /
    • 2023
  • 음성 사용자 인터페이스(Voice User Interface)에 대한 수요가 증가함에 따라 음성 합성(Speech Synthesis) 시스템에서 자연스러운 음성 발화를 모방하기 위해 적절한 위치에 휴지를 삽입하는 것이 주된 과업으로 자리잡았다. 대화의 연속성을 고려했을 때, 자연스러운 음성 기반 인터페이스를 구성하기 위해서는 대화의 맥락을 이해하고 적절한 위치에 휴지를 삽입하는 것이 필수적이다. 이에 따라 본 연구는 대화 맥락에 기반하여 적절한 위치에 휴지를 삽입하는 Long-Input Transformer 기반 휴지 예측 모델을 제안하고 한국어 대화 데이터셋에서 검증한 결과를 보인다.

  • PDF

Improving Transformer with Dynamic Convolution and Shortcut for Video-Text Retrieval

  • Liu, Zhi;Cai, Jincen;Zhang, Mengmeng
    • KSII Transactions on Internet and Information Systems (TIIS)
    • /
    • 제16권7호
    • /
    • pp.2407-2424
    • /
    • 2022
  • Recently, Transformer has made great progress in video retrieval tasks due to its high representation capability. For the structure of a Transformer, the cascaded self-attention modules are capable of capturing long-distance feature dependencies. However, the local feature details are likely to have deteriorated. In addition, increasing the depth of the structure is likely to produce learning bias in the learned features. In this paper, an improved Transformer structure named TransDCS (Transformer with Dynamic Convolution and Shortcut) is proposed. A Multi-head Conv-Self-Attention module is introduced to model the local dependencies and improve the efficiency of local features extraction. Meanwhile, the augmented shortcuts module based on a dual identity matrix is applied to enhance the conduction of input features, and mitigate the learning bias. The proposed model is tested on MSRVTT, LSMDC and Activity-Net benchmarks, and it surpasses all previous solutions for the video-text retrieval task. For example, on the LSMDC benchmark, a gain of about 2.3% MdR and 6.1% MnR is obtained over recently proposed multimodal-based methods.

Improved Deep Learning-based Approach for Spatial-Temporal Trajectory Planning via Predictive Modeling of Future Location

  • Zain Ul Abideen;Xiaodong Sun;Chao Sun;Hafiz Shafiq Ur Rehman Khalil
    • KSII Transactions on Internet and Information Systems (TIIS)
    • /
    • 제18권7호
    • /
    • pp.1726-1748
    • /
    • 2024
  • Trajectory planning is vital for autonomous systems like robotics and UAVs, as it determines optimal, safe paths considering physical limitations, environmental factors, and agent interactions. Recent advancements in trajectory planning and future location prediction stem from rapid progress in machine learning and optimization algorithms. In this paper, we proposed a novel framework for Spatial-temporal transformer-based feed-forward neural networks (STTFFNs). From the traffic flow local area point of view, skip-gram model is trained on trajectory data to generate embeddings that capture the high-level features of different trajectories. These embeddings can then be used as input to a transformer-based trajectory planning model, which can generate trajectories for new objects based on the embeddings of similar trajectories in the training data. In the next step, distant regions, we embedded feedforward network is responsible for generating the distant trajectories by taking as input a set of features that represent the object's current state and historical data. One advantage of using feedforward networks for distant trajectory planning is their ability to capture long-term dependencies in the data. In the final step of forecasting for future locations, the encoder and decoder are crucial parts of the proposed technique. Spatial destinations are encoded utilizing location-based social networks(LBSN) based on visiting semantic locations. The model has been specially trained to forecast future locations using precise longitude and latitude values. Following rigorous testing on two real-world datasets, Porto and Manhattan, it was discovered that the model outperformed a prediction accuracy of 8.7% previous state-of-the-art methods.

Study on the Transfer Functions for Detecting Windings Displacement of Power Transformers with Impulse Method

  • Shon, Chae-Hwa;Yi, Sang-Hwa;Lee, Heun-Jin;Kang, Dong-Sik
    • Journal of Electrical Engineering and Technology
    • /
    • 제7권6호
    • /
    • pp.876-883
    • /
    • 2012
  • The paper investigates three types of transfer function methods for detecting displacements of winding in a model transformer. To acquire these transfer functions, the measuring method of input voltage, current and its response is used in impulse method. The applied impulse voltages had three rising times, which were short rising time (less than 0.6 ${\mu}s$), medium rising time (about 0.8 ${\mu}s$) and long rising time (about 1 ${\mu}s$) in front waves. Every 10 measurements of voltage and current waves were averaged from 50 measurements of voltage and current waves. These transfer functions were tested in normal, 24mm elevated and 48mm elevated windings conditions and were analyzed with correlation coefficients and spectrum deviations. In the analysis, the results depend on the types of transfer functions and the rising times of input voltages.

IGBT 인버터제어 TIG용접기 성능개선에 관한 연구 (A Study on the Performance-Improvement of TIG Welder with IGBT Inverter System.)

  • 이정호;김재문;김연충;원충연;김규식;김영렬
    • 대한전기학회:학술대회논문집
    • /
    • 대한전기학회 1997년도 하계학술대회 논문집 F
    • /
    • pp.2155-2158
    • /
    • 1997
  • This paper describes the performance improvement of dc TIG (Tungsten Inert Gas) welder. The TIG welder consists of single phase full bridge IGBT inverter which incorporates the high frequency transformer, the boost converter for improving the input power factor, and the arc start system. The arc will be generated without fail even when the extension cable between the torch and the power source is 100m long. In addition, the arc start system with a short dc output voltage will generate less EMI noise than the traditional arc start system with a high frequency output voltage. To demonstrate the practical significance of the proposed methods, some simulation studies are presented.

  • PDF

A neural network solver for differential equations

  • Wang, Qianyi;Aoyama, Tomoo;Nagashima, Umpei;Kang, Eui-Sung
    • 제어로봇시스템학회:학술대회논문집
    • /
    • 제어로봇시스템학회 2001년도 ICCAS
    • /
    • pp.88.4-88
    • /
    • 2001
  • In this paper, we propose a solver for differential equations, using a multi-layer neural network. The multi-layer neural network is a transformer function originally where the function is differential and the explicit representation has been developed. The learning determines the response of neural networks; however, the response is not equal to the output values. The differential relations are also the response. The differential conditions can be also set as teaching data; therefore, there is a possibility to reach a new solver for the differential equations. Since it is unknown how to define the input data for the neural network solver during long terms, we could not derive the expressions. Recently, the analogue type neural network is known and it transforms any vector to another The "any" must be...

  • PDF

CMOS Linear Power Amplifier with Envelope Tracking Operation (Invited Paper)

  • Park, Byungjoon;Kim, Jooseung;Cho, Yunsung;Jin, Sangsu;Kang, Daehyun;Kim, Bumman
    • Journal of electromagnetic engineering and science
    • /
    • 제14권1호
    • /
    • pp.1-8
    • /
    • 2014
  • A differential-cascode CMOS power amplifier (PA) with a supply modulator for envelope tracking (ET) has been implemented by 0.18 ${\mu}m$ RF CMOS technology. The loss at the output is minimized by implementing the output transformer on a FR-4 printed circuit board (PCB). The CMOS PA utilizes the $2^{nd}$ harmonic short at the input to enhance the linearity. The measurement was done by the 10MHz bandwidth 16QAM 6.88 dB peak-to-average power ratio long-term evolution (LTE) signal at 1.85 GHz. The ET operation of the CMOS PA with the supply modulator enhances the power-added efficiency (PAE) by 2.5, to 10% over the stand-alone CMOS PA for the LTE signal. The ET PA achieves a PAE of 36.5% and an $ACLR_{E-UTRA}$ of -32.7 dBc at an average output power of 27 dBm.

선박용 변압기 없는 40W LED 조명 구동회로의 설계 및 구현 (Design and Implementation of Transformerless 40W LED Light Driver Circuit for Ships)

  • 송종관;박장식;윤병우
    • 한국전자통신학회논문지
    • /
    • 제7권3호
    • /
    • pp.485-490
    • /
    • 2012
  • 본 논문에서는 선박의 진동에 의하여 수명이 짧은 필라멘트를 사용하는 조명기구를 대체하기 위하여 선박용 LED 조명등 구동회로를 설계 및 구현하였다. 구동회로는 부피를 줄이고 비용을 절감하기 위하여 변압기가 없는 스위칭 회로로 설계되었다. 스위칭 회로는 입력 교류전압 220 V에 PWM 제어를 함으로써 안정적으로 LED를 구동할 수 있도록 설계하였다. 스위칭 회로의 펄스성 전류에 대하여 Valley-fill 방식의 역률 보상회로를 채용함으로써 역률를 개선하였다. 장기간 운항하는 선박에서의 조명등 교체 주기를 줄여 관리를 효율적으로 할 수 있도록 직병렬 배열로 LED 모듈 회로를 설계하여 LED의 일부가 손상되더라도 LED 모듈이 조명등 기능을 할 수 있도록 하였다. 개발한 구동회로를 포함한 조명기구는 전력소모와 역률이 각각 39 W, 0.925 로 선박 조명에 적합함을 확인하였다.

플라즈마 발생용 전원장치의 LCD 패널 세정효과에 관한 연구 (A Study of LCD Panel Cleaning Effect of Plasma Generation Power Source)

  • 김규식
    • 전자공학회논문지SC
    • /
    • 제45권5호
    • /
    • pp.44-51
    • /
    • 2008
  • UV 램프 시스템은 오랫동안 TFT LCD 나 PDP 의 패널 세정에 사용되어 왔으나, 저렴한 가격의 고성능 세정에 대한 필요성 때문에 고전압 플라즈마 세정에 대한 기술이 개발되고 그 성능이 향상되어 왔다. 장벽방전 (barrier discharge) 혹은 무성방전 (silent discharge)으로 불리는 유전체 장벽 방전 (Dielectric-Barrier Discharges, DBDs) 는 오존 발생기에 주로 이용되어 왔다. 본 논문에서는 LCD 세정용으로 6kW 급 고전압 플라즈마 발생장치를 구현하였다. 3상 입력전압을 직류로 정류한 뒤, 인버터 시스템에 의해 고주파 펄스로 바꾸고 고압 트랜스퍼머를 거쳐 다이오드로 정류한다. 마지막으로, 고압 플라즈마를 발생시키기 위해 양방향 고전압 펄스 스위칭회로가 사용되었다. 실험을 통해 상압 플라즈마가 LCD 패널 세정에 크게 유용함을 보였다.