Abstract:

The components of microbial fermentation process are extremely complex. The physicochemical and biochemical reactions are mixed and overlapped. Many metabolic states cannot be measured directly. The hybrid cybernetic model combines the elementary mode analysis and the cybernetic model control. It not only involves both intracellular product information and metabolic behavior, but also introduces control variables to regulate the activity and synthesis of catalytic enzymes. Based on the kinetic equation of microbial metabolism, the real-time state monitoring of microbial metabolism could implement. In view of the fermentation process of poly-β-hydroxybutyrate, by means of the metabolic pathway analysis, the nonlinear dynamics of the existing hybrid cybernetic model is first linearized. Then the extended Kalman filter is used to estimate the metabolic state of fermentation process. Simulation results show that it has a good estimation effect on some important state variables in microbial metabolism.

Key words: hybrid cybernetic model, elementary mode, extended Kalman filter, metabolism, enzyme, fermentation

摘要：

微生物发酵过程中组分异常复杂，物化和生化反应混合交叠，多数状态不能直接测量。混合赛博模型结合了基元模式分析和赛博(cybernetic)模型控制作用，既考虑细胞内产物信息和代谢行为，又引入控制变量来调节催化酶的活性和合成，从而建立了微生物代谢过程的动力学方程，基于该模型可以实现对微生物体代谢状态的实时监测。针对聚羟基丁酸发酵过程，在代谢途径和基元模式分析基础上，对其混合赛博模型中非线性动态进行线性化，再采用扩展卡尔曼滤波对代谢过程状态进行估计。仿真结果表明，对于微生物代谢中一些重要状态变量具有良好的估计效果。

关键词: 混合赛博模型, 基元模式, 扩展卡尔曼滤波, 代谢, 酶, 发酵