http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Vibration-based Health Index Extraction of Power Transformers
Joung Taek Yoon(윤정택),Kyung Min Park(박경민),Byeng Dong Youn(윤병동),Wook-Ryun Lee(이욱륜) 대한기계학회 2012 대한기계학회 춘추학술대회 Vol.2012 No.11
Power transformer is the one of key components of power plants as well as the most frequently failed components. As the failure of power transformers can cause whole plant shut-down and substantial capital loss, failure prediction techniques, i.e., health diagnostics and prognostics, are vital. This study aims at developing a health diagnostics method for oil-filled power transformers using mechanical vibration signal against two mechanical failure modes. The vibration signals are acquired from on-site 36 transformers and two health metrics-mean and standard deviation of spectral responses-are proposed based upon a sensitivity analysis. From the fact that transformers with greater health metrics have long operation years than those with smaller it can be concluded that the proposed health metrics are suitable for health diagnostics and fault identification of power transformers.
전력용 변압기 기계적 고장 진단을 위한 센서 네크워크 최적화 방법론
윤정택(Joung Taek Yoon),윤병동(Byeng D. Youn),박경민(Kyung Min Park),이욱륜(Wook-Ryun Lee) 대한기계학회 2013 대한기계학회 춘추학술대회 Vol.2013 No.12
Power transformer is one of the key components in power plants as well as the most frequently failed components. As the failure of power transformers can cause plant shut-down and substantial capital loss, failure prediction techniques, i.e., fault diagnostics and prognostics, are vital. To prevent the mechanical failures of power transformers, vibrations sensors are installed on the surface of oil-filled power transformers. Due to high randomness in vibration and large scale of the transformers, an excessively large number of sensors are generally installed. This study aims at developing the framework of sensor network (SN) optimization to maximize the diagnostic and prognostic capability of the transformers’ potential faults. The vibrations of 36 on-site power transformers were measured. The analysis of the vibrations shows transformer tank vibration characteristics and the importance of high vibration signals in fault detection. The proposed SN optimization framework optimizes the number of sensors and their locations to measure the high vibration signals robustly against system uncertainty. Comparing evaluated health status with maintenance history demonstrates that the proposed framework is capable of estimating transformer health status with the significantly
박경민(Kyung Min Park),윤정택(Joung Taek Yoon),윤병동(Byeng D. Youn),배용채(Yong Chae Bae) 대한기계학회 2012 대한기계학회 춘추학술대회 Vol.2012 No.11
The power generator, one of the most critical components in power plant, has usually time- or usage-based maintenance strategy to avoid serious accidents such as unexpected shutdown. However, these maintenance strategies cannot prevent accidents effectively or bring about considerable waste of remaining useful life, thus resulting in high maintenance cost. Prognostics and health management (PHM) technology has recently been developed to predict potential faults with an aim at minimizing the deficiencies of the maintenance strategies. This paper proposes a new health diagnostic method of power generator windings against moisture absorption with a Mahalanobis distance. This health indicator extracted from moisture absorption data shows the health degradation of power generator windings by water leakage. A supervised health classification rule is also suggested by modeling the Mahalanobis distance in a probabilistic manner. This study is demonstrated with 42 windings over three years.
OTDR을 이용한 LNG 플랜트 설비의 배관계 누설탐지 기술 개발
김태진(Tae Jin Kim),윤정택(Joung Taek Yoon),최승혁(Seung Hyuk Choi),김현재(Hyeon Jae Kim),윤병동(Byeng Dong Youn) 대한기계학회 2012 대한기계학회 춘추학술대회 Vol.2012 No.11
The liquefying technique of natural gas is spotlighted for easy storage and transport. However, due to flammability, extremely low temperature, and toxicity of the liquefied natural gas (LNG), it is essential to maintain high reliability of an LNG plant system including compressor, container, valve and pipeline. Among them, the pipeline has highest failure frequency mainly resulting from joint loosening or poor finishing. This paper aims at developing a real time monitoring system to detect LNG pipe leakage using an optical time domain reflectometer (OTDR). The failure analysis of LNG pipes revealed that a pipe flange fails most frequently. When LNG leaks in the flange, a simulation model predicts a local pressure drop around the leakage area. Through experimental validation with an optical fiber wrapped around the pipeline and flange, it was found that an optical signal using an OTDR and signal processing can detect a local pressure drop near the leakage area in the flange.