提出了基于过程机理和神经网络修正的混合建模方法。利用差分进化算法对诺西肽分批发酵过程机理模型的参数进行辨识,用神经网络建模方法对机理模型进行修正。模型的训练与验证数据都取自实际的实验过程一诺西肽分批发酵。验证结果表明,混合模型比单纯的机理模型具有更高的精度。
A hybrid modeling method based on process mechanism and neural network was proposed, which used differential evolution (DE) algorithm to identify parameters of the mechanical model and neural network to modify the model, Training and validation data are both from experimental process-Nosiheptide batch fermentation process. Validating results show that the hybrid model has higher precision than the mechanical model.