抗生素发酵过程溶氧浓度模糊PID-神经网络控制器响应特性分析  

Response characteristics analysis of a fuzzy-PID neural networks controller used for DO concentration control in antibiotic fermentation

作　　者：李聪颖[1] 李泳宁[1] 郭养浩[1,2] 孟春[2] 石贤爱[1,2] 

机构地区：[1]福州大学药物生物技术与工程研究所,福州350002 [2]福州大学生物科学与工程学院,福州350108

出　　处：《中国抗生素杂志》2009年第6期333-336,371,共5页Chinese Journal of Antibiotics

基　　金：福建省科技重大专项(编号:2004HZ03-1)

摘　　要：针对抗生素发酵过程溶氧浓度自动控制的需要,建立了三输入双输出模糊PID神-经网络控制器,通过计算机仿真和计算机实时控制,对标度因子取值不同时控制器的响应特性进行了分析,确定了标度因子的初始值(GS=10、GQ=20、Ke=10.0、Kc=2.0、Kh=2.0)及其在线自调整策略。该控制器在对金霉素和卡那霉素发酵过程溶氧浓度自动控制中取得了良好的应用效果。A fuzzy PID-neural networks controller with 3 inputs and2 outputs was built for DO concentration control in antibiotic fermentation. Through computer simulation and real time control with computer, the analysis of response characteristics of the controller and the determination of the parameters of fuzzy neural networks were performed. The initial values and the policy for self-regulating of those parameters were determined as well. When used in the control of dissolved oxygen in aureomycin and kanamycin fermentation process, the controller ran with a satisfactory result.

关 键 词：模糊PID-神经网络 标度因子 响应特性 计算机仿真 

分 类 号：TQ465.2[化学工程—制药化工] TP273[自动化与计算机技术—检测技术与自动化装置]