基于DBN-DMFA的代谢通量调控模型  

Metabolic Flux Regulation Model Based on DBN-DMFA

作　　者：岳向阳 赵忠盖 刘飞[1,2] YUE Xiang-yang;ZHAO Zhong-gai;LIU Fei(MOE Key Laboratory of Advanced Process Control for Light Industry,Jiangnan University;Institute of Automation,Jiangnan University)

机构地区：[1]江南大学,轻工过程先进控制教育部重点实验室 [2]江南大学自动化研究所

出　　处：《化工自动化及仪表》2021年第5期464-468,共5页Control and Instruments in Chemical Industry

摘　　要：综合考虑实际发酵过程的非线性特征和过程变量具有多采样率的特点,提出基于DBN-DMFA的代谢通量调控模型,即先对所需的细胞外时变参数进行预测,然后使用DMFA估计代谢通量,最后用DBN建立代谢通量调控模型。通过青霉素发酵过程仿真实验,证明了该策略的有效性。Comprehensively considering the nonlinear feature of the actual fermentation process and the characteristics of the process variables with multiple sampling rates,a novel metabolic flux regulation model based on DBN-DMFA was proposed.Firstly,having the required extracellular time-varying parameters predicted,and then having DMFA adopted to estimate metabolic flux and finally,having DBN used to establish a metabolic flux regulation model.The simulation experiment of penicillin fermentation process proves the effectiveness of this strategy.

关 键 词：发酵过程 代谢通量 半监督学习 深度信念网络 建模 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]