神经元网络在谷氨酸发酵过程中的应用

针对谷氨酸发酵难以建立精确的数学模型这一问题,提出一种基于人工神经网络（ANN）的菌体浓度预测方法.首先对历史数据进行主元分析,提取特征信息,然后用实际生产过程中的训练数据训练神经网络,建立主元之间内部非线性关系,得到关于菌体浓度的非参数模型.仿真结果表明：该非参数模型泛化能力较强,并能达到很高的预测精度.

Application of Neural Network on Process of Glutamic Acid Fermentation

To solve the difficulty of establishing the accurate mathematical model of the glutamic acid fermentation,we put forward a new prediction method for mycelium density based on artificial neural network(ANN).First,we do principal component analysis to the history data and take attribute information,then we train the neural network by training data from the practical production process and build the internal non-linearity relation among the principal components,we can get the distribution free model about mycelium density.The simulations show that the distribution free model has a good generalization ability and can reach a high precision.