文献类型：期刊文献

中文题名：基于PSO-BP神经网络的催化裂化C3含量软测量模型

英文题名：Soft Sensing Model of C_3 Concentration of FCCU Based on PSO-BP Neural Network

作者：王学武[1];顾幸生[1];刘卓倩[1];商雨青[1,2]

机构：[1]华东理工大学自动化研究所,上海200237;[2]上海电机学院电气工程系,上海200240

年份：2009

卷号：21

期号：4

起止页码：973

中文期刊名：系统仿真学报

外文期刊名：Journal of System Simulation

收录：CSTPCD;;Scopus;北大核心:【北大核心2008】；CSCD:【CSCD2011_2012】；

基金：上海市重点学科建设项目(B504)

语种：中文

中文关键词：C3含量;软测量;催化裂化;PSO-BP;神经网络

外文关键词：C3 concentration; so.sensing; FCCU; PSO-BP; neural network

摘要：C3含量的在线估计对催化裂化生产过程操作具有重要意义,但是其在线测量难以实现,针对这一问题,利用软测量技术来实现C3含量的在线估计。建立了基于主元分析的神经网络模型和PSO-BP神经网络模型,并对其仿真结果进行分析和比较。结果显示,基于PSO-BP神经网络的C3含量软测量模型具有较高的精度和较好的性能,满足实际生产过程操作的要求。
Real-time measuring the C3 Concentration is important for the process of fluid catalytic cracking unit （FCCU）, but it＇s difficult to measure it directly, so soft-sensing method was applied to solve this question. The neural network model based on PCA-BP and PSO-BP neural network model were established, and the simulation results were analyzed and compared. The results show that soft sensing model of C3 Concentration based on PSO-BP neural network has good precision and quality, and it can meet the demands of process in chemical plant.