딥러닝을 이용한 정삼투 막모듈의 플럭스 예측

다국어 초록 (Multilingual Abstract)

Forward osmosis (FO) process is a chemical potential driven process, where highly concentrated draw solution (DS) is used to take water through semi-permeable membrane from feed solution (FS) with lower concentration. Recently, commercial FO membrane modules have been developed so that full-scale FO process can be applied to seawater desalination or water reuse. In order to design a real-scale FO plant, the performance prediction of FO membrane modules installed in the plant is essential. Especially, the flux prediction is the most important task because the amount of diluted draw solution and concentrate solution flowing out of FO modules can be expected from the flux. Through a previous study, a theoretical based FO module model to predict flux was developed. However it needs an intensive numerical calculation work and a fitting process to reflect a complex module geometry. The idea of this work is to introduce deep learning to predict flux of FO membrane modules using 116 experimental data set, which include six input variables (flow rate, pressure, and ion concentration of DS and FS) and one output variable (flux). The procedure of optimizing a deep learning model to minimize prediction error and overfitting problem was developed and tested. The optimized deep learning model (error of 3.87%) was found to predict flux better than the theoretical based FO module model (error of 10.13%) in the data set which were not used in machine learning.

목차 (Table of Contents)

• ABSTRACT
• 1. 서 론
• 2. 연구 방법
• 2.1. 딥러닝에 사용된 FO 막모듈 실험 데이터
• 2.2 딥러닝 모델 구축
• 2.3 딥러닝 모델의 변동성 및 과적합 분석
• 3. 결과 및 토의
• 3.1 딥러닝 모델의 변동성 평가 및 최소화 방안
• 3.2 검증 오차를 최소화하기 위한 딥러닝 모델 최적화
• 3.3 딥러닝 모델의 과적합 분석
• 4. 결 론
• References

참고문헌 (Reference)

1 지상민, "학습률(Step-Size)변화에 따른 디지털 신호의 기계학습 방법 개선" 한국디지털정책학회 18 (18): 261-268, 2020

2 전종민, "통계기반 정삼투 모듈 모델의 적용성 검토" 대한환경공학회 41 (41): 611-618, 2019

3 배경태, "인공신경망의 은닉층 최적화를 통한 농산물 가격예측 모델" 한국정보기술학회 14 (14): 161-169, 2016

4 전종민, "와권형 정삼투 막모듈의 성능 분석" 대한환경공학회 40 (40): 481-486, 2018

5 조준모, "빅데이터의 정규화 전처리과정이 기계학습의 성능에 미치는 영향" 한국전자통신학회 14 (14): 547-552, 2019

6 이상민, "머신러닝 기법을 활용한 낙동강 중류 지역의 Chl-a 예측 알고리즘 비교 연구(수질인자 및 수량 중심으로)" 대한상하수도학회 34 (34): 275-286, 2020

7 이도연, "딥러닝을 이용한 음악흥행 예측모델 개발 연구" 한국콘텐츠학회 20 (20): 10-18, 2020

8 주기훈, "기계학습 옵티마이저 성능 평가" 한국전기전자학회 24 (24): 766-776, 2020

9 Jang, I. D., "The analysis telecommunication service market with data mining" 28 (28): 1-3, 2001

10 Cho, Y. H., "Study on the activation functions for efficient learning in DNN" The Institute of Electronics and Information Engineers 800-803, 2016

11 Kim, J. E., "Practical considerations for operability of an 8“ spiral wound forward osmosis module:Hydrodynamics, fouling behaviour and cleaning strategy" 404 : 249-258, 2017

12 Caruana, R., "Overfitting in neural nets: backpropagation, conjugate gradient, and early stopping" 402-408, 2002

13 Ciresan, D. C., "Multi-Column Deep Neural Network for Traffic Sign Classification" 32 : 333-338, 2012

14 "Google Tensorflow"

15 Kim, J. E., "Forward osmosis membrane modular configurations for osmotic dilution of seawater by forward osmosis and reverse osmosis hybrid system" 128 : 183-192, 2018

16 김수빈, "Forward osmosis membrane filtration for microalgae harvesting cultivated in sewage effluent" 대한환경공학회 20 (20): 99-104, 2015

17 Kum, D., "Development and Assessment for extended daily streamflow regression equation of TMDL station using Machine Learning" 289-290, 2017

18 Cireşan, D. C., "Deep neural networks segment neuronal membranes in electron microscopy images" 2843-2851, 2012

19 Lee, J., "A statistics-based forward osmosis membrane characterization method without pressurized reverse osmosis experiment" 403 : 36-45, 2017

20 Jeon, J., "A simple modeling approach for a forward osmosis system with a spiral wound module" 433 : 120-131, 2018