DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kim, Heeyoung | - |
dc.contributor.advisor | 김희영 | - |
dc.contributor.author | Ko, Taeyoung | - |
dc.date.accessioned | 2019-09-03T02:41:57Z | - |
dc.date.available | 2019-09-03T02:41:57Z | - |
dc.date.issued | 2019 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=843192&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/266236 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 산업및시스템공학과, 2019.2,[iii, 24 p. :] | - |
dc.description.abstract | In complex industrial processes, process fault detection and diagnosis is an important task in reducing production cost and improving product quality, and it can be done by monitoring and analyzing a relationship between a large amount of data collected from various sensors and process working conditions. Most existing methods for fault detection and diagnosis assume sufficient labeled data available for training. However, label acquisition is costly and laborious in practice, whereas abundant unlabeled data are often available. To make effective use of a large amount of unlabeled data for fault detection and diagnosis, we propose a new approach using semi-supervised deep generative models. In particular, to consider temporal correlation and inter-variable correlation in multivariate time series of process data collected from multiple sensors and to model the complex relationship between high-dimensional process data and process status, we propose two kinds of semi-supervised deep generative models with convolutional neural networks incorporated, namely semi-supervised convolutional deep generative model (SS-CDGM) and semi-supervised convolutional auxiliary deep generative model (SS-CADGM). The proposed models are assessed on data from the Tennessee Eastman benchmark process. The results demonstrate more effective performance of the proposed models than competing methods. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | convolutional auxiliary deep generative model▼amultivariate time series | - |
dc.subject | semi-supervised convolutional variational autoencoder▼aTennessee Eastman process▼aunlabeled data | - |
dc.subject | 합성곱 보조변수 심층 생성 모델▼a다변수 시계열▼a준지도학습 합성곱 변분 오토인코더▼a테네시 이스트만 공정▼a레이블이 없는 데이터 | - |
dc.title | Fault diagnosis for high-dimensional complex processes using semi-supervised deep convolutional generative models | - |
dc.title.alternative | 준지도학습 심층 생성 모델을 활용한 공정 고장 분류 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :산업및시스템공학과, | - |
dc.contributor.alternativeauthor | 고태영 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.