Optimal Operation of Fed-Batch Antibiotic Fermentations

生物过程的优化操作是极其重要的，因为这些过程使用昂贵的原材料，需要大型资本密集型工厂和低浓度的产品。由于生物系统固有的复杂性，优化和控制发酵所需的可靠数学模型很少可用。在环境中发生的任何变化都会改变模型，因此离线计算的最佳进料策略可能不是精确过程的最佳策略。因此，尽管有关于动态优化和控制的优雅理论，但在发酵实验中实施这些理论的成功非常有限。本研究将开发一种间歇和补料间歇生物过程在线优化的概念，该概念考虑到不断变化的环境条件和未建模的细胞内事件，使用具有时变参数的数学过程描述。在发酵过程中，将开发一种在线估计模型参数并计算和更新最优饲喂策略的方案。这项工作将具体解决所提出的优化概念直接应用于工业上重要的过程，即头孢菌素发酵，并将尝试通过实验验证结果。

Optimal operation of bioprocesses is extremely important because these processes utilize expensive raw materials, require large capital-intensive plants and yield products that are low in concentration. Due to the inherent complexity of biological systems, reliable mathematical models needed for optimization and control of the fermentations are seldom available. Any change occurring in the environment would alter the model and thus the optimal feeding strategy calculated off-line may not be optimal for the exact process. Thus, in spite of the availability of elegant theories on dynamic optimization and control, there has been very limited success in implementing these theories experimentally in fermentations. This study will develop a concept of on-line optimization for batch and fed-batch bioprocesses that takes into account the continually changing environmental conditions and unmodeled intracellular events using a mathematical process description with time-varying parameters. A scheme will be developed which estimates the model parameters on-line and computes and updates the optimal feeding policy while the fermentation is in progress. The work will address specifically the direct application of the proposed optimization concept to an industrially significant process, namely cephalosporin fermentation, and will try to validate the results through experimentation.