http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
밀링머신 모니터링을 위한 라즈베리파이 및 센서 기반 시스템
양성모(Seongmo Yang),박승환(Seung Hwan Park) 대한기계학회 2020 大韓機械學會論文集A Vol.44 No.9
4차 산업혁명은 스마트 공장이 주도한다. 스마트 공장은 자동화 공장과 정보 통신 기술이 융합한 공장으로 데이터를 통해 생산비용과 시간을 스스로 최적화한다. 따라서 생산에 관여하는 모든 기계시스템의 물리량을 센서로 수집 및 분석하는 모니터링이 필수적이기 때문에 이에 대한 연구와 산업이 성장하고 있다. 특히 생산시스템에서 가공 기계는 품질에 직접적인 영향을 주기 때문에 모니터링이 더욱 중요하다. 하지만, 가공 기계 모니터링을 위해서는 고품질의 데이터 수집 및 분석이 필수적이지만 이를 위해서 고가의 센서와 데이터 수집 시스템 구축이 필요하다. 따라서 본 연구는 비용적 문제를 해결하고, 시스템의 자유도를 높이기 위해 MEMS 가속도 센서와 라즈베리파이 기반 데이터 수집 장치를 개발한다. 또한, 이 장치를 두 대의 밀링머신에 설치하여 진동 신호를 수집하고, 밀링머신 모니터링을 위한 지표를 개발한다. The 4th industrial revolution is led by smart factories in which the cost and time of production processes are optimized through big data by converging information and communication technology through automation. Therefore, industrial research on monitoring technology is growing because collection and analysis of physical quantities of all mechanical systems involved in production are essential. Monitoring of processing machines is especially important because processing machines have a direct effect on quality. However, expensive sensors and data collection systems are needed for high-quality data collection and analysis. Therefore, micro-electro-mechanical acceleration sensors are developed and Raspberry-Pi-based data collection devices are employed to solve the cost problem and increase the degree of freedom of the system. In addition, this device is installed on two milling machines to collect vibration signals and develop indicators for monitoring milling machines.
김근형(Geun-hyung Kim),양성모(Seongmo Yang),강지훈(Jihoon Kang),정진은(Jin-eun Jeong),박승환(Seung Hwan Park) 한국신뢰성학회 2020 신뢰성응용연구 Vol.20 No.4
Purpose: The advent of the fourth industrial revolution has led to increased interest in military defense systems, and demand for new weapon systems involving artificial intelligence techniques has also increased. In particular, data from field operations, which are collected during post- logistics support, can be used for the reliability analysis of weapon systems. The existing reliability analysis method for weapon systems has a limitation in that it cannot reflect the maintenance history during the operation of the weapons systems, practical methods for predicting actual reliability via field operation analysis. Furthermore, typical data from field operations are collected manually, and therefore, they contatin atypical features introduced by operators’ personal decisions. Methods: In this research, a text mining approach is proposed to extract meaningful features, and some visualization techniques are presented for enhancing the interpretability of operational behavior. The Doc2Vec algorithm, which can measure the similarity between feature vectors, is used to extract a feature vector, and the t-SNE algorithm is used visualization. Results: The proposed algorithm represents the availability of unstructured data through feature extraction and visualization on the basis of post-logistics data. Feature vector based visualization shows that a new classification system for causes of defects can be established through manually wrote data.
토크 제한 및 회생 제동 제어 로직 검증을 위한 전기이륜차 시뮬레이션 모델 연구
백남철(NamChul Paik),양성모(SeongMo Yang),김호경(HoKyung Kim),이준우(JunWoo Lee) 한국자동차공학회 2023 한국자동차공학회 학술대회 및 전시회 Vol.2023 No.11
This study focuses on developing agile processes for electric two-wheeler R&D to adapt to changing eco-friendly vehicle policies and diverse market requirements. By applying Model-Based Development (MBD) like A-SPICE, the research aims to promptly validate the impact of subsidy policy changes and performance requirements on component specifications and software. Various vehicle models, driving modes, and battery scenarios are implemented for real-time validation. The study models efficient motor and MCU operation, battery pack performance, and drive system characteristics. Torque and regenerative braking logics are introduced to predict test results and avoid repetitive real-world tests. The results demonstrate consistent application of torque limit logic during maximum acceleration tests, with similar trends in battery and motor temperatures showing minimal differences of less than 3 degrees. The regenerative braking logic is validated through predictive tests in CVS40 driving mode, closely resembling actual test consumption and accurately predicting energy accumulation through regenerative braking within a 2% margin.
김광섭(Kwangseub Kim),Wang Maosen,정낙탁(Naktak Jung),양성모(Seongmo Yang),유세훈(Sehoon Yoo),기대성(Daeseong Gi),서명원(Myungwon Suh) 한국자동차공학회 2015 한국자동차공학회 부문종합 학술대회 Vol.2015 No.5
Drowsy behavior is more likely to occur in sleep-deprived drivers. The accident rate is higher driver of drowsiness. Individuals’ drowsy behavior detection technology should be developed to prevent drowsiness related crashes. Driving information such as accelerations, steering angles and velocity, and physiological signals of drivers such as electroencephalogram (EEG), and eye tracking were adopted in present drowsy behavior detection technologies. However, it is difficult to measure physiological signal, as a result, driving information becomes more popular for drowsy driving detection. In this paper, vehicle information including lateral accelerations, longitudinal acceleration and steering angles, was combined into various cases to detect drowsy driving behavior. In order to increase the accuracy, it is defined the data set and predicted drowsy driving prediction using the Random Forest algorithm. As the result, the case of lateral and longitudinal of acceleration showed the best.