谷氨酸发酵过程的混合建模  

Research on the Hybrid Modeling of Glutamic Acid Fermentation Process

作　　者：黎兴宝 潘丰 

机构地区：[1]江南大学轻工过程先进控制教育部重点实验室,江苏无锡214122

出　　处：《江南大学学报（自然科学版）》2010年第6期642-646,共5页Joural of Jiangnan University (Natural Science Edition)　

基　　金：国家863计划项目(2006AA020301)

摘　　要：针对谷氨酸发酵过程时变、非线性、模型的不确定等问题,提出一种综合机理知识与改进最小二乘支持向量机的发酵过程混合建模方法。采用改进最小二乘支持向量机模型对动力学模型估计值进行校正,弥补动力学模型的不精确性。将混合模型应用到谷氨酸发酵过程的谷氨酸质量浓度、基质质量浓度与菌体质量浓度的预测估计中,结果表明混合模型提高了单纯的机理模型的预测精度。To overcome the problem of time variability,non-linearity,model uncertainty for Glutamic acid fermentation process,a hybrid model for Glutamic acid fermentation process was proposed which combined prior knowledge and improves least square support vector machines（λivi-SVM）.The λivi-SVM model was used to provide an additive correction to the simple process models,which compensates for inaccuracy in the simple process models estimate.The hybrid model was applied in predicting biomass,substrate and mycelium concentration of Glutanic acid fermentation process.The results show that thehybrid model with prior knowledge improves the accuracy of a pure prior knowledge model.

关 键 词：发酵 混合建模 最小二乘支持向量机 谷氨酸 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]