| 학회 | 한국화학공학회 |
| 학술대회 | 2017년 봄 (04/26 ~ 04/28, ICC 제주) |
| 권호 | 23권 1호, p.215 |
| 발표분야 | 공정시스템 |
| 제목 | 데이터 마이닝과 머신러닝 이용한 MBR공정의 막 오염 진행 정도 예측 |
| 초록 | Membrane bioreactor (MBR) has several advantage such as a reduced footprint, an improved effluent quality compared with a conventional biological process. However, MBR has a disadvantage as a fouling which increases operating cost. This study predicts and assesses a fouling progress using a data-mining technique with machine learning techniques. First, the fouling progress is assessed by data-mining technique. Then, Trans membrane pressure (TMP) is predicted using recurrent neural network (RNN) and deep neural network (DNN). Finally, the fouling progress is classified into clean, transitional and fouled conditions from deep belief network (DBN) based on the data-mining’s information. Fouling progress prediction and assessment abilities of RNN with DBN were better than those of DNN with DBN. This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIP) (No.2015R1A2A2A11001120) |
| 저자 | 남기전1, 이승철1, Usman Safder2, 유창규1 |
| 소속 | 1경희대, 2kyunghee |
| 키워드 | 공정모델링; 공정모사; 공정제어; 이상진단 |
| 원문파일 | 초록 보기 |
