摘要
针对谷氨酸发酵产物浓度难以在线检测问题,提出一种基于核岭回归(KRR)的谷氨酸浓度软测量方法。该方法利用一致关联度(URD)对输入辅助变量进行相关性分析,降低模型复杂性,提高模型性能。然后,结合KRR软测量模型,实现谷氨酸发酵过程中产物浓度的预测。发酵数据仿真结果验证了所提方法的有效性,有助于对谷氨酸发酵过程优化控制提供及时指导。
As product concentration in the glutamate fermentation process is difficult to measure in real time, a soft sensor modeling method based on kernel ridge regression is proposed. Firstly, the correlation analysis of input secondary variables and product concentration are carried out by uniform relational degree (URD) to reduce the model complexity and improve the model performance. Then, the uniform relational degree is combined with kernel ridge regression (KRR) to predict the product concentration of glutamate fermentation. Simulation results show that the proposed method is effective, and can provide effective operation guidances for the control and optimization of the glutamate fermentation process.
出处
《控制工程》
CSCD
北大核心
2017年第11期2195-2200,共6页
Control Engineering of China
基金
国家自然科学基金资助项目(61273131)
江苏省普通高校研究生科研创新计划项目(CXZZ12_0741)
中央高校基本科研业务费专项资金(JUDCF12034)
关键词
谷氨酸
核岭回归
一致关联度
软测量
发酵建模
Glutamate
kernel ridge regression
uniform relational degree
soft sensor
fermentation modeling