基于改进串联混合建模方法的发酵过程菌体浓度软测量.pdfVIP

学兔兔 第32卷 第l期 仪 器 仪 表 学 报 VoL 32 No．1 2011年1月 Chinese Joumal of Scientific Instrument Jan．20ll 基于改进串联混合建模方法的发酵过程 菌体浓度软测量木 杨强大 (东北大学材料与冶金学院 沈阳 1 10004) 摘 要：发酵过程中菌体浓度难以在线实时测量，给控制与优化带来困难。针对这一问题，本文利用软测量技术来实现菌体浓 度的在线估计，并提出了一种改进的串联混合建模方法用以建立菌体浓度软测量模型。改进的串联混合建模方法，克服了现有 方法需要利用插值所得的数据进行软测量模型构建的不足，从而保证了建模数据的可靠性。利用诺西肽发酵过程生产数据进 行仿真研究，仿真结果表明，基于改进串联混合建模方法的软测量模型是有效的，比基于现有方法的软测量模型具有更好的估 计性能。 关键词：发酵；软测量；算法改进；串联混合建模 中图分类号：TP274 文献标识码：A 国家标准学科分类代码：120．30 Soft sensor of biomass in fermentation process based on improved series hybrid modeling method Yang Qiangda (School ofMaterials and Metallurgy，Northeastern University，Shenyang 110004，China) Abstract：Biomass is hard to be measured on．1ine in real—ti me in fermentation process，which brings difficultv in process control and optimization．To solve this problem，soft sensor technique is applied to implement the on—line es— timation of biomass，and an improved series hybrid modeling method is presented to develop the soft sensor model of biomass in fermentation process，The improved series hybrid modeling method overcomes the shortcoming that exist· ing method requires using the data obtained through interpolation to develop the soft sensor model，and guarantees the reliability of the modeling data．Simulation is performed using the production data from Nosiheptide fermentation process．Simulation results show that the soft sensor model based on the improved series hybrid modeling method is effective and has better est