| 학회 |
한국화학공학회 |
| 학술대회 |
2022년 봄 (04/20 ~ 04/23, 제주국제컨벤션센터) |
| 권호 |
28권 1호, p.128 |
| 발표분야 |
[주제 2] 기계학습 |
| 제목 |
Data-Driven Soft Sensor to Predict Real Time Process and Quality Variables in the Chemical Industry |
| 초록 |
Some product qualities and key variables are very difficult to measure online in industrial process control due to technical and economic limitations. Soft sensors to play an essential role in ensuring effective process control, monitoring and optimization. In order to address the issue of different sampling intervals of process and quality parameters, a deep learning technique, which allowed for exploiting the historical data of the chemical process plant was developed. Deep learning methods for high-level abstract feature extraction in pattern recognition domains have recently been developed, and they have tremendous potential for soft sensing applications. This research focused on the development of a data-driven soft sensor model utilizing autoencoders, an unsupervised learning technique.Through the deep architecture, the proposed model is able to capture the crucial information of the input data, resulting in the development of a soft sensor with high performance. The effectiveness of the developed soft senor was validated on the debutanizer column of a refinery based on available data. Deep learning provides an effective and promising method for soft sensor modeling. |
| 저자 |
APPIAH PIUS1, daewon han1, 한대원1, 박준규1, 이창하2, 오민1
|
| 소속 |
1한밭대, 2연세대 |
| 키워드 |
공정시스템(Process Systems Engineering) |
| E-Mail |
|
| 원문파일 |
초록 보기 |
