| 학회 |
한국화학공학회 |
| 학술대회 |
2022년 봄 (04/20 ~ 04/23, 제주국제컨벤션센터) |
| 권호 |
28권 1호, p.1132 |
| 발표분야 |
[주제 12] 화학공학일반(부문위원회 발표) |
| 제목 |
하이브리드 인공지능 모델기반 수질 총질소/총인 버츄얼 측정 Metrology 개발-InformerVM |
| 초록 |
Monitoring total nitrogen (TN) and total phosphorus (TP) in stream water is essential to assess the water quality. However, TN and TP are always difficult to measure in an online fashion, due to the high costs, uncertainties, and time lag associated with real-time estimation. This study aims to develop a novel artificial intelligence (AI)-virtual metrology model for estimating long/short-term TN and TP concentrations in river stream based on Informer deep network (InformerVM) with hybrid statistical-AI feature selection (HFS). The proposed AI-InformerVM can effectively capture complex dynamic relationships among water quality parameters using the self-attention mechanisms. The inputs of InformerVM consisted of eleven biological-chemical variables in which the HFS is utilized to select the important factors for TN and TP estimation. Acknowledgments This research is funded by the BK21 FOUR program of NRF and by the Korean government (MSIT) (No. 2021R1A2C2007838), project for Collabo R&D between Industry, Academy, Research Institute funded by Korea Ministry of SMEs and Startups in 2021(Project No .S3105519). |
| 저자 |
Ba Alawi Abdulrahman1, Tayerani Charmchi Amir Saman2, Mohammad Moosazadeh1, 허성구1, 우태용1, 김상윤1, 김민한3, 원동찬3, 유창규1
|
| 소속 |
1경희대, 2KyungHeeUniv., 3에이치코비 |
| 키워드 |
공정시스템(Process Systems Engineering) |
| E-Mail |
|
| 원문파일 |
초록 보기 |
