Articles & Issues

Language: korean

Conflict of Interest: In relation to this article, we declare that there is no conflict of interest.

Publication history: Received October 14, 2020
Accepted January 4, 2021

This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/bync/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

All issues

머신 러닝과 데이터 전처리를 활용한 증류탑 온도 예측

Prediction of Distillation Column Temperature Using Machine Learning and Data Preprocessing

이예찬^{1 2} 최영렬^{1 3} 조형태¹ 김정환^1†

Yechan Lee^{1 2} Yeongryeol Choi^{1 3} Hyungtae Cho¹ Junghwan Kim^1†

¹한국생산기술연구원 친환경재료공정연구그룹, 44413 울산광역시 중구 종가로 55 ²서울과학기술대학교 화공생명공학과, 01811 서울특별시 노원구 공릉로 232 ³연세대학교 화공생명공학과, 03722 서울특별시 서대문구 연세로 50

¹Green Materials and Processes R&D Group, Korea Institute of Industrial Technology, 55, Jongga-ro, Ulsan, 44413, Korea ²Department of Chemical and Biomolecular Engineering, Seoul National University of Science and Technology, 232, Gongneung-ro, Seoul, 01811, Korea ³Department of Chemical and Biomolecular Engineering, Yonsei University, 50, Yensei-ro, Seoul, 03722, Korea

kjh31@kitech.re.kr

Korean Chemical Engineering Research, May 2021, 59(2), 191-199(9), 10.9713/kcer.2021.59.2.191 Epub 3 May 2021

Download PDF

Abstract

화학 공정의 주요 설비 중 하나인 증류탑은 물질들의 끓는점 차이를 이용하여 혼합물에서 원하는 생산물을 분리하는 설비이며 증류 공정은 많은 에너지가 소비되기 때문에 최적화 및 운전 예측이 필요하다. 본 연구의 대상 공정은 공급처에 따라 원료의 조성이 일정하지 않아 정상 상태로 운전이 어려워 효율적인 운전이 어렵다. 이를 해결하기 위해 데이터 기반 예측 모델을 이용하여 운전 조건을 예측 할 수 있다. 하지만 미가공 공정 데이터에는 이상치 및 노이즈가 포함되어 있어 예측 성능을 향상시키기 위해 데이터 전처리가 필요하다. 본 연구에서는 인공 신경망 모델인 Long shortterm memory (LSTM)과 Random forest (RF)를 사용하여 모델을 최적화한 후, 데이터 전처리 방법으로 Low-pass filter 와 One-class support vector machine을 사용하여 데이터 전처리 방법 및 범위에 따른 예측 성능을 비교하였다. 각 모델의 예측 성능과 데이터 전처리의 영향은 R2과 RMSE를 사용하여 비교하였다. 본 연구의 결과, 전처리를 통해 LSTM의 경우 R2은 0.791에서 0.977으로 RMSE는 0.132에서 0.029로 각각 23.5%, 78.0% 향상되었고, RF의 경우 R2은 0.767 에서 0.938으로 RMSE는 0.140에서 0.050으로 각각 22.3%, 64.3% 향상되었다.

A distillation column, which is a main facility of the chemical process, separates the desired product from a mixture by using the difference of boiling points. The distillation process requires the optimization and the prediction of operation because it consumes much energy. The target process of this study is difficult to operate efficiently because the composition of feed flow is not steady according to the supplier. To deal with this problem, we could develop a data-driven model to predict operating conditions. However, data preprocessing is essential to improve the predictive performance of the model because the raw data contains outlier and noise. In this study, after optimizing the predictive model based long-short term memory (LSTM) and Random forest (RF), we used a low-pass filter and one-class support vector machine for data preprocessing and compared predictive performance according to the method and range of the preprocessing. The performance of the predictive model and the effect of the preprocessing is compared by using R2 and RMSE. In the case of LSTM, R2 increased from 0.791 to 0.977 by 23.5%, and RMSE decreased from 0.132 to 0.029 by 78.0%. In the case of RF, R2 increased from 0.767 to 0.938 by 22.3%, and RMSE decreased from 0.140 to 0.050 by 64.3%.

Keywords

Data preprocessing low-pass filter one-class support vector machine distillation column machine learning random forests Long-short term memory

References

Kartal F, Ozveren U, Energy, 209, 118457 (2020)
Sneesby MG, Tade MO, Datta R, Smith TN, Ind. Eng. Chem. Res., 36(5), 185 (1997)
Sharma N, Singh K, Syst. Sci. Control Eng., 2(1), 512 (2014)
Lee DW, Lee SW, Trans. Softw. Data Eng., 9(4), 129 (2020)
Raghavan VSR, Radhakrishnan TK, Srinivasan K, ISA Trans., 50(1), 61 (2011)
Devi HRD, Bai A, Nagarajan N, Obes. Med., 17, 100152 (2020)
Erkus EC, Purutcuoglu V, Eur. J. Oper. Res, 291(2), 560 (2020)
Zhang R, Zhang S, Muthuraman SJJ, 5th WSEAS Int. Conf. Appl. Electromagn. Wirel. Opt. Commun. Tenerife, Spain, December 14-16 (2007).
Kim J, Park NS, Yun S, Chae SH, Yoon S, J. Korean Soc. Environ. Eng., 40(12), 473 (2018)
Guinon JL, Ortega E, Garcia-Anton J, Perez-herranz V, International Conference on Engineering Education, July, Coimbra, 1-4(2007).
Guo H, Yu M, Liu J, Ning J, J. Surv. Eng., 130(4), 175 (2004)
Panigrahi S, Karali YS, Behera H, Int. J. Comput. Appl., 75(10), 13 (2013)
Zhang Z, Wu Z, Rincon D, Christofides PD, Mathematics, 7(10), 1 (2019)
Oh K, Kwon H, Roh J, Choi Y, Park H, Cho H, Kim J, Korean. Chem. Eng. Res., 23(4), 565 (2020)
Kwon HW, OhKC, Chung YG, Cho HT, Kim JH, Appl. Chem. Eng., 31(5), 520 (2020)
Kwon H, Oh K, Choi Y, Chung YG, Kim J, Int. J. Intell. Syst., 36, 1970 (2021)
Kim TJ, Hong JS, J. Soc. E-bus. Stud., 23(1), 1 (2018)
Kim D, Lee S, Kim M, Lee E, Yoo CK, Korean Chem. Eng. Res., 54(5), 621 (2016)
Giap V, Pineda IT, Lee JY, Lee DK, Kim YS, Ahn KY, Lee YD, Korean Hydrog. Energy Soc., 31(5), 436 (2020)