摘要
针对青霉素发酵过程中的基质浓度、菌体浓度、产物浓度等关键生物参数难以在线实时测量的问题,提出了一种基于粒子群模糊神经网络的软测量建模方法。采用模糊径向基函数-神经网络(RBF-NN)构建青霉素发酵的软测量模型,同时,结合改进粒子群优化训练算法(PSO),建立了青霉素反应过程的软测量模型,并对发酵工艺进行了仿真试验研究。仿真试验结果表明,所建立的软测量模型测量精度高、效果好,能够满足工程实际的要求。
Normally,it is difficult to realize realtime and online measurement of the critical biological parameters for penicillin fermentation process,such as the concentrations of the matrix,biomass,and products.Aiming at this problem,the soft-sensing method based on particle swarm fuzzy neural network is proposed.The soft-sensing model of penicillin fermentation is established by using fuzzy radial basis function neural network（RBF-NN）,and combining with the improved particle swarm optimization（PSO） training algorithm,the soft-sensing model of penicillin reaction process is established.The simulation experiment of the fermentation process is researched.The result of simulation experiment indicates that this soft-sensing model is precise enough for getting good effect,and meets practical engineering requirements.
出处
《自动化仪表》
CAS
北大核心
2011年第5期46-48,52,共4页
Process Automation Instrumentation
关键词
粒子群优化算法
模糊神经网络
径向基函数
软测量
建模
Particle swarm optimization（PSO） Fuzzy neural network Radial basis function（RBF） Soft-sensing Modeling
