基于差分进化算法的在线自适应控制及其在高密度酵母培养中的应用研究

发布时间：2018-03-06 04:23

  本文选题：酿酒酵母　切入点：毕赤酵母　出处：《江南大学》2016年硕士论文　论文类型：学位论文

【摘要】：酵母分批补料培养中,碳源添加过量会导致副产物乙醇的大量积累,破坏细胞结构及功能,降低碳源利用效率;碳源添加不足会限制细胞生长。为解决这一矛盾,提出了一种基于差分进化算法(Differential evolution algorithm,DE)的在线自适应底物流加策略(DE-PID)。该策略以传统比例-积分-微分(Proportional-Integral-Differential,PID)控制为基础,利用自回归移动平均模型(Autoregressive moving average model,ARMA)辨识酵母培养过程的动力学特性,再根据ARMA系统辨识模型预测被控变量,以被控变量的预测值与设定值之间的误差为目标函数。利用DE算法求解目标函数达到最小值时的PID控制参数,实现碳源流加的在线自适应控制。主要研究内容与结论总结如下:(1)以乙醇浓度为被控变量,利用已有的酿酒酵母培养模型,对传统PID控制与DE-PID控制策略的性能进行计算机仿真比较研究。结果表明,使用DE-PID策略时,发酵液中的乙醇浓度能够被稳定地控制在1 g·L-1的低水平,细胞浓度达到34.45 g·L-1,比采用传统PID策略的批次提高29%。(2)在实际的酿酒酵母分批补料培养中,利用DE-PID控制策略分别对乙醇浓度和呼吸商(RQ)实施定值控制。培养30 h,乙醇浓度均可以始终控制于低水平(0.02~2.35g·L-1)。乙醇浓度定值控制条件下的细胞密度仅为23.25 g-DCW·L-1;RQ定值控制条件下的细胞密度则达到47.75 g-DCW·L-1,比采用DO-Stat流加策略的对照批次提高了85.44%。(3)在重组毕赤酵母高密度培养过程中,利用DE-PID策略分别对乙醇浓度和溶解氧浓度(DO)实施定值控制,培养34 h后细胞密度分别达到112.25 g-DCW·L-1和113.25g-DCW·L-1,乙醇浓度也始终可以被限制在低水平(0.09~1.75 g·L-1)。与前期构建的改良型DO-Stat甘油流加策略相比,DE-PID策略同样能够在抑制乙醇积累的同时获得更高密度的细胞,DO控制水平稳定,过程控制和细胞培养性能明显改善。
[Abstract]:In yeast batch feeding culture, excessive addition of carbon sources will lead to a large amount of byproduct ethanol accumulation, destroy cell structure and function, reduce carbon source utilization efficiency, and carbon source deficiency will limit cell growth. An online adaptive bottom logistics strategy based on differential evolution algorithm (DE-PIDD) is proposed, which is based on the traditional proportional-integral-differential-integral-integral-integral-differential differential algorithm (PIDs) control. The autoregressive moving average model ARMA was used to identify the dynamic characteristics of yeast culture process, and then the controlled variables were predicted according to the ARMA system identification model. Taking the error between the predicted value and the set value of the controlled variable as the objective function, the DE algorithm is used to solve the PID control parameters when the objective function reaches the minimum value. The main research contents and conclusions are summarized as follows: (1) ethanol concentration is taken as the controlled variable, and the existing Saccharomyces cerevisiae culture model is used. The performance of traditional PID control strategy and DE-PID control strategy is compared by computer simulation. The results show that the ethanol concentration in fermentation broth can be stabilized at a low level of 1 g 路L ~ (-1) when using DE-PID strategy. The cell concentration reached 34.45 g 路L ~ (-1), 29.3% higher than that of the traditional PID strategy) in the actual batch feeding culture of Saccharomyces cerevisiae. DE-PID control strategy was used to control ethanol concentration and respiratory quotient (RQ) respectively. After 30 hours of culture, ethanol concentration could always be controlled at a low level of 0.02 ~ 2.35 g 路L ~ (-1). The cell density was only 23.25 g-DCW 路L ~ (-1) RQ under the control of alcohol concentration. The cell density reached 47.75g-DCW 路L -1, which was 85.44.4% higher than that of the control batch using DO-Stat flow-adding strategy in the process of high density culture of recombinant Pichia pastoris. The DE-PID strategy was used to control the concentration of ethanol and dissolved oxygen, respectively. The cell density reached 112.25 g-DCW 路L -1 and 113.25 g-DCW 路L -1 after 34 h culture, and ethanol concentration could always be limited to 0.09 ~ 1.75 g 路L ~ (-1) 路L ~ (-1). Compared with the modified DO-Stat glycerol infusion strategy, DE-PID strategy could also inhibit ethanol accumulation at the same time. A higher density of cells with stable control levels of do was obtained. Process control and cell culture performance were significantly improved.
【学位授予单位】：江南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TQ920.1;TQ223.122
，

本文编号：1573316