메뉴 건너뛰기
.. 내서재 .. 알림
소속 기관/학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
로그인 회원가입 고객센터 ENG
주제분류

추천
검색
질문

논문 기본 정보

자료유형
학술대회자료
저자정보
장재영 (서울대학교) 조재경 (서울대학교) 김성우 (서울대학교)
저널정보
대한전자공학회 대한전자공학회 학술대회 2021년도 대한전자공학회 하계종합학술대회 논문집
발행연도
2021.6
수록면
1,242 - 1,246 (5page)

이용수

표지
📌
연구주제
📖
연구배경
🔬
연구방법
🏆
연구결과
AI에게 요청하기
추천
검색
질문

초록· 키워드

오류제보하기
The Molecules of thin polymer film can be rearranged in the stretching axial direction through a stretching process to obtain optical properties as a display material. When the film is stretched, intermittent film breakages occur in the process of achieving the target draw ratio while passing through a water bath containing a continuous roller and a solution. Process variables such as temperature, solution concentration, tension, speed, and torque of each facility unit are stored and controlled in real time through facility sensors to achieve product quality specifications and stable film running. However, due to the complex non-linear relationships of multivariate data, prediction of the actual film breakages by experts is closely impossible.
In this paper, we propose an unsupervised learning method that can predict the phenomena of film breakages that occur very rarely in the film stretching process using multivariate time series data collected in real time at the manufacturing site.
We used two models based on LSTM which is useful for long time series data processing. They are Generative Adversarial Network (GAN) model capable of detecting abnormalities through adversarial training of normal and fake data, and Autoencoder model capable of discriminating abnormal data after feature learning using only normal data.
This study, aimed at the failure alarm function in the actual industrial field, was compared and evaluated with Recall, Precision and F1-Score. Recall is the ratio of predicted breakages among actual breakages. Precision is the ratio of actual breakages among predicted breakages. F1-Score is the harmonic average of the Recall and Precision. The F1-Score of GAN and Autoencoder showed a result of 0.16 and 0.14, respectively. It shows the possibility of predicting defects even with little data.

목차

Abstract
I. 서론
II. 본론
Ⅲ. 구현
Ⅳ. 결론 및 향후 연구 방향
참고문헌

참고문헌 (0)

참고문헌 신청

함께 읽어보면 좋을 논문

논문 유사도에 따라 DBpia 가 추천하는 논문입니다. 함께 보면 좋을 연관 논문을 확인해보세요!

이 논문의 저자 정보

이 논문과 함께 이용한 논문

최근 본 자료

전체보기

댓글(0)

0