文献类型：期刊文献

英文题名：Data-Augmented Deep Learning Algorithm for Accurate Control of Bioethanol Fermentation Using an Online Raman Analyzer

作者：Ji, Kaidi[1];Yu, Xiaofei[2];Chen, Lifan[2];Wang, Yongbo[2];Guo, Zhiqiang[3];Chen, Biao[4];Li, Qingyang[5];Li, Zhen[6];Zhang, Hu[6];Wang, Guan[2,7];Zhuang, Yingping[2,7];Ruan, Yinlan[3,8]

机构：[1]Guilin Univ Elect Technol, Sch Elect Engn & Automat, Guilin, Peoples R China;[2]Chinese Acad Sci, Qingdao Inst Bioenergy & Bioproc Technol, Qingdao New Energy Shandong Lab, Qingdao, Peoples R China;[3]Guilin Univ Elect Technol, Sch Optoelect Engn, Guilin, Peoples R China;[4]Guilin Univ Elect Technol, Sch Life & Environm Sci, Guilin, Peoples R China;[5]Guilin Univ Elect Technol, Guangxi Key Lab Intelligent Proc Comp Images & G, Guilin, Peoples R China;[6]Subphoton Technol Co Ltd Shenzhen, Shenzhen, Peoples R China;[7]East China Univ Sci & Technol ECUST, State Key Lab Bioreactor Engn, Shanghai, Peoples R China;[8]Guilin Univ Elect Technol, Guangxi Key Lab Optoelect Informat Proc, Guilin, Peoples R China

年份：2025

卷号：122

期号：9

起止页码：2366

外文期刊名：BIOTECHNOLOGY AND BIOENGINEERING

收录：;EI(收录号：20252318568283);WOS:【SCI-EXPANDED(收录号:WOS:001507993100001)】；

基金：This study was funded by the National Key Research and Development program of China (Grant no. 2021YFC2101000), the Taishan Scholars Program of Shandong Province (Grant no. tspn202408281), Shanghai Rising-Star Program (Grant no. 21QA1402400), Innovation Project of GUET Graduate Education (Grant no. 2023YCXS228), National Natural Science Foundation of China (Grant no. 62275058), Natural Science Foundation of Guangxi, China (Grant no. 2023GXNSFAA026259) and Guangxi Science and Technology Program Project (Grant no. AD25069073). The authors would also like to thank Haodi Zhao from the School of Optoelectronic Engineering, Guilin University of Electronic Technology, for his contribution to the data collection of Raman spectra.

语种：英文

外文关键词：data augmentation; deep learning; feedback control; fermentation; raman analyzer

摘要：Fed-batch fermentation has become the preferred strategy in many industrial biomanufacturing processes. However, a key challenge remains in optimizing the feeding strategy to achieve stable maximum yields. In this study, we present an online Raman spectroscopy-based monitoring and control system, using bioethanol production by Saccharomyces cerevisiae as a case study. To address the issue of limited labeled data, a pseudo-labeling approach based on semi-supervised learning was employed, expanding the available training data set by 100-fold compared to conventional labeling methods. In addition, we developed a spectral-temporal concatenation convolutional neural network (STC-CNN) that incorporates sequential spectral features. Comparative evaluations with multiple machine learning algorithms demonstrated the superior performance of STC-CNN, achieving a root mean square error (RMSE) of 3.63 g/L for glucose prediction. The system enabled rapid and automated glucose feeding to maintain various target concentrations. Notably, a glucose setpoint of 30 g/L yielded the highest ethanol concentration of 140.68 g/L-an increase of 3.85% over traditional Fed-batch fermentation-while reducing glycerol by 6.67%. These results highlight the significant potential of Raman spectroscopy combined with deep learning for automated bioprocess optimization and discovery of optimal operating strategies.