http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
초희박 직접분사식 가솔린엔진에서 분사 및 점화ㆍ전략에 따른 배기 특성
오진우(Jinwoo Oh),김홍석(Hongsuk Kim),김성대(Sungdea Kim),박철웅(Cheolwoong Park),정용일(Youngil Jeong) 한국자동차공학회 2011 한국자동차공학회 부문종합 학술대회 Vol.2011 No.5
Particulate matter are became a concern and will be limited by future emissions regulation in gasoline direct injection(GDI) engines. Such as the upcoming Euro 6 legislation BMW N53 3.0L In this research, GDI engine was used to universal ECU. Universal ECU was operated under a mototron program is based on MATLAB simulink program. The size distributions are measured using a scanning mobility particle sizer (SMPS). The size distributions are measured using a scanning mobility particle sizer. THC, NO<SUB>x</SUB> and emitted particle are reported for ignition timing injection timing and number of injection at 1500rpm BMEP 1.5bar condition. The result shows that increased THC and decreased particle size at advanced injection timing. However THC, NOx and particle size was measured to same value at advanced ignition timing.
오진우(Jinwoo Oh),김홍석(Hongsuk Kim),이석환(Seokhwan Lee),정용일(Yongil Jeong) 한국자동차공학회 2010 한국자동차공학회 학술대회 및 전시회 Vol.2010 No.11
Recently, a few ultra lean burn GDI(Gasoline direct injection) engines showing good fuel economy have been introduced in the market. The BMW N53 GDI engine is one of these engines, and its ultra lean burn strategy was investigated in this study. The BMW N53 engine uses a spray-guided lean burn concept. In this engine, piezo injectors are located at the center of combustion chamber to form stratified mixture in the vicinity of spark plug. This study showed that brake specific fuel consumption rate is reduced about 11% to 25% by lean combustion mode compared with the homogeneous combustion mode. Spark timing was retarded with the increase of BMEP at the homogeneous combustion mode, but it was fixed at BTDC 15° CAD at the lean combustion mode. This study also showed the fuel injection and valve train strategies for lean combustion of the engine.
23㏄ 가솔린 소형원동기의 카뷰레터 전ㆍ후단에서의 흡기 맥동 특성
오진우(Jinwoo Oh),최영하(Youngha Choi),김병국(Byeongguk Kim),이동근(Dongkeun Lee),윤석주(Suckju Yoon),김동선(Dongsun Kim),한종규(Jongkyu Han) 한국자동차공학회 2009 한국자동차공학회 부문종합 학술대회 Vol.2009 No.4
This paper presents the pulsation of carburetor inlet and outlet pressure of a small spark-ignited gasoline engine at non-load condition and increasing dynamometer load. The engine used in this paper is a single cylinder, diaphragm carburetor, two-stroke, air-cooled 23㏄ SI engine for brush cutter. We measured the rpm, air pressure at carburetor inlet and outlet at 4779,6378,7473, WOT rpm and dynamo load The results showed that the pressure variation at carburetor inlet is not changed. and the pressure variation at carburetor outlet is each occurred to 0.73㎏/㎠ at 4779rpm and 6378rpm, 0.71㎏/㎠ at 7473rpm, and 0.84㎏/㎠ at WOT. According to increasing dynamometer load, carburettor outlet pressure is occurred to 0.84 ~ 0.86㎏/㎠ similarly at WOT.
오진우(JinWoo Oh),이장현(JangHyun Lee),김학선(HakSun Kim),이정호(JungHo Lee),이수혁(SooHyuk Lee),김한샘(HanSam Kim) 한국소음진동공학회 2015 한국소음진동공학회 학술대회논문집 Vol.2015 No.10
Various absorber is used in the noise reduction equipments and acoustic construction site. Representative absorbers are Glass-wool and Poly-ester. Both of them has a high absorption coefficient. When the absorber is installed, absorber is needed finishing, perforated plate and film, to maintain the shape, prevent scatter and keep from water. These finishing layer cause shift of the absorption coefficient. Experiment was performed at acoustic facility(reverberant room) by referring KS I ISO 2805 standard.
이종 프로세서 환경에서의 복수의 딥 러닝 어플리케이션 스케줄링 기법
오진우(Jinwoo Oh),하순회(Soonhoi Ha) 한국정보과학회 2020 정보과학회 컴퓨팅의 실제 논문지 Vol.26 No.7
이종 프로세서 환경에서의 복수의 딥 러닝 어플리케이션 스케줄링은 기본적으로 NP-난해 (NP-Hard) 문제에 속하여 매우 큰 문제 공간을 가진다. 그래서 일반적으로 유전 알고리즘(GA, Genetic Algorithm)과 같은 메타 휴리스틱이 적용될 수 있지만 이는 수행 시간이 길어서 런타임에 적용하기 어렵다는 단점을 지닌다. 따라서 본 연구에서는 이러한 단점을 보완하면서 성능 또한 크게 떨어지지 않는 새로운 기법의 스케줄링 휴리스틱을 제안하였다. 제안하는 휴리스틱은 복수 응용 스케줄링의 스케줄 가능성 문제를 고려하지 못 하는 기존의 리스트 스케줄링 방식 휴리스틱들의 한계를 극복하여 ‘합성과 반복 개선’ 이라는 새로운 방식을 도입하였다. 그리하여 CPU, GPU, NPU로 구성되는 이종 프로세서 환경에서 여러 딥 러닝 네트워크들을 대상으로 하는 성능 비교 실험을 통해 제안하는 휴리스틱이 빠른 시간 내에 좋은 스케줄링을 생성함을 확인하였다. The scheduling of multiple deep learning applications on heterogeneous processors is basically an NP-hard problem with a very large problem space. Meta-heuristics such as GAs (Genetic Algorithms) may be applied, but these have the disadvantage of having too long an execution time to be applied at run time. Therefore, this study proposes a new scheduling heuristic, which complements this shortcoming and does not significantly degrade scheduling performance. The proposed heuristic overcomes the limitations of traditional list scheduling techniques that fail to take into account the schedulability issue in the scheduling of multiple applications and introduces a new approach called ‘synthesis and iterative improvement’. It is confirmed through experiments with different deep learning networks on heterogeneous processors (including CPUs, GPUs, and NPUs) that the proposed heuristic produces good scheduling results that are sufficiently fast to apply at run time.