Abstract: It is vital to improve the production of 1，3-propanediol（1，3-PD） in microbial continuous fermentations.However，3-hydroxypropionaldehyde（3-HPA） accumulation will arouse an irreversible cessation of fermentation process.Considering the inhibitions of 3-HPA to the growth of cells and the activities of enzymes，this paper proposes a novel mathematical model to describe glycerol bioconversion to 1，3-PD in continuous cultures.To identify the kinetic parameters in the above model，a parameter identification model is proposed.The identifiability is also proved.Finally，an improved differential evolution algorithm is constructed to solve the parameter identification model.Numerical results show that not only the proposed model can describe the fermentation process better，but also the algorithm is efficient.

Key words: mathematical model, parameter identification, differential evolution algorithm, continuous culture

摘要： 在甘油连续发酵生产中提高1，3-丙二醇（1，3-PD）的产量是至关重要的，然而，3-羟基丙醛（3-HPA）的积累会引起发酵过程的停止。考虑到3-HPA对细胞生长和对酶的活性的抑制作用，提出一个新的数学模型来描述利用克雷伯氏菌连续发酵生产1，3-PD的过程。为了确定模型中的动力学参数，提出了参数辨识模型，并证明了参数的可辨识性。最后构造了改进的差异演化算法求解该参数辨识模型，数值结果表明该动力学模型能更好地描述微生物连续发酵过程，并且算法也是有效的。

关键词: 数学模型, 参数辨识, 差异演化算法, 连续发酵