{"title":"Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case","authors":"","doi":"10.1016/j.cjche.2024.04.018","DOIUrl":null,"url":null,"abstract":"<div><p>Neural networks are often viewed as pure ‘black box’ models, lacking interpretability and extrapolation capabilities of pure mechanistic models. This work proposes a new approach that, with the help of neural networks, improves the conformity of the first-principal model to the actual plant. The final result is still a first-principal model rather than a hybrid model, which maintains the advantage of the high interpretability of first-principal model. This work better simulates industrial batch distillation which separates four components: water, ethylene glycol, diethylene glycol, and triethylene glycol. GRU (gated recurrent neural network) and LSTM (long short-term memory) were used to obtain empirical parameters of mechanistic model that are difficult to measure directly. These were used to improve the empirical processes in mechanistic model, thus correcting unreasonable model assumptions and achieving better predictability for batch distillation. The proposed method was verified using a case study from one industrial plant case, and the results show its advancement in improving model predictions and the potential to extend to other similar systems.</p></div>","PeriodicalId":9966,"journal":{"name":"Chinese Journal of Chemical Engineering","volume":null,"pages":null},"PeriodicalIF":3.7000,"publicationDate":"2024-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chinese Journal of Chemical Engineering","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1004954124001733","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, CHEMICAL","Score":null,"Total":0}
引用次数: 0
Abstract
Neural networks are often viewed as pure ‘black box’ models, lacking interpretability and extrapolation capabilities of pure mechanistic models. This work proposes a new approach that, with the help of neural networks, improves the conformity of the first-principal model to the actual plant. The final result is still a first-principal model rather than a hybrid model, which maintains the advantage of the high interpretability of first-principal model. This work better simulates industrial batch distillation which separates four components: water, ethylene glycol, diethylene glycol, and triethylene glycol. GRU (gated recurrent neural network) and LSTM (long short-term memory) were used to obtain empirical parameters of mechanistic model that are difficult to measure directly. These were used to improve the empirical processes in mechanistic model, thus correcting unreasonable model assumptions and achieving better predictability for batch distillation. The proposed method was verified using a case study from one industrial plant case, and the results show its advancement in improving model predictions and the potential to extend to other similar systems.
期刊介绍:
The Chinese Journal of Chemical Engineering (Monthly, started in 1982) is the official journal of the Chemical Industry and Engineering Society of China and published by the Chemical Industry Press Co. Ltd. The aim of the journal is to develop the international exchange of scientific and technical information in the field of chemical engineering. It publishes original research papers that cover the major advancements and achievements in chemical engineering in China as well as some articles from overseas contributors.
The topics of journal include chemical engineering, chemical technology, biochemical engineering, energy and environmental engineering and other relevant fields. Papers are published on the basis of their relevance to theoretical research, practical application or potential uses in the industry as Research Papers, Communications, Reviews and Perspectives. Prominent domestic and overseas chemical experts and scholars have been invited to form an International Advisory Board and the Editorial Committee. It enjoys recognition among Chinese academia and industry as a reliable source of information of what is going on in chemical engineering research, both domestic and abroad.