DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Seong, Poong Hyun | - |
dc.contributor.advisor | 성풍현 | - |
dc.contributor.author | Kim, Seung Geun | - |
dc.date.accessioned | 2021-05-11T19:42:30Z | - |
dc.date.available | 2021-05-11T19:42:30Z | - |
dc.date.issued | 2020 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=902011&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/283514 | - |
dc.description | 학위논문(박사) - 한국과학기술원 : 원자력및양자공학과, 2020.2,[iii, 91 p. :] | - |
dc.description.abstract | In nuclear power plants (NPPs), the reliability of instrumentation signals is crucial for making appropriate decisions. Multiple signals can become faulty simultaneously and become unavailable under harsh conditions, and this could lead to improper decisions being made, which could result in catastrophic failure. However, despite the importance of signal integrity, there is a lack of studies on the recovery of multiple missing signals in NPPs under harsh conditions. This study proposes a new signal recovery method based on a generative adversarial network (GAN), which can be applied to recover multiple missing signals and does not require prior knowledge of plant conditions. A GAN model is trained to generate realistic signal sets from given latent vectors and labels, and the model is utilized for finding the optimal latent vector and label, resulting in generation of signal set that suitable for signal recovery. The damaged signal set is recovered by replacing its missing parts with the corresponding parts of the signal set generated from the found optimal latent vector and label. To verify the applicability of the proposed method to the recovery of multiple missing signals under various NPP emergency situations, experiments were conducted based on the simulation data. Simulation data was acquired by using compact nuclear simulator (CNS) with considering four types of design basis accident (DBA) scenarios, various break sizes, and 31 kinds of signals. A GAN model was repeatedly trained to adjust the hyper-parameters and best performing model was applied for further experiments on recovering missing signals. Several kinds of signals within 1,000 randomly selected unit data were intentionally omitted and attempt to be recovered. The results shown that the proposed method is capable of recovering most of multiple missing signals under various plant conditions. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | instrumentation signal▼amissing signal recovery▼agenerative adversarial network▼anuclear power plant▼adesign basis accident | - |
dc.subject | 계측 신호▼a소실 신호 복원▼a생성적 적대 신경망▼a원자력발전소▼a설계기준사고 | - |
dc.title | Development of a signal recovery method for multiple instrumentation signal failure cases in nuclear power plants | - |
dc.title.alternative | 원전 다중 계측 신호 고장 상황을 위한 신호 복원 방법론 개발 | - |
dc.type | Thesis(Ph.D) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :원자력및양자공학과, | - |
dc.contributor.alternativeauthor | 김승근 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.