刘超锋,柳金江,赵伟,吴学红,刘亚莉,龚毅
LIU Chao-feng,LIU Jin-jiang,ZHAO Wei,WU Xue-hong,LIU Ya-li,GONG Yi

• 1:郑州轻工业学院
• 2:杭氧股份有限公司
• 3:河南煤化集团中原大化公司

摘要(Abstract)：

为了满足现场软测量的要求,研究了某制氧机公司正在调试的深冷空分设备用增压膨胀机组的膨胀空气流量、膨胀端进口压力、膨胀端进口温度、膨胀端出口温度、增压端出口压力和膨胀机转速的因果关系,建立了RBF神经网络模型,实现了运行特性量间关系的高精度拟合和预测。获得的神经网络模型的精度至少在96.97%以上。利用学习好的RBF神经网络可以迅速方便地实现现场运行特性参量间关系的预测。
In order to meet the requirements of soft measurement for site,relationships of cause and effect between expansion air flow,expansion-side inlet pressure,expansion side inlet temperature,expansion-side outlet temperature,booster-side outlet pressure,and expansion turbine rotational speed from booster expansion turbine of cryogenic air separation unit debugged by an oxygen machine company were studied. RBF neural network model were established,precision fitting and forecastings between operating characteristics were achieved. The accuracy of the resulting neural network model is at least 96. 97%. trained RBF neural network can easily achieve the predicted relationship between field operation characteristics parameters.

关键词(KeyWords)： 深冷空分增压膨胀机;运行特性参量;软测量建模;RBF
cryogenic air separation unit;booster expansion turbine;operating parameters;soft measurement modeling;RBF

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(21076200);; 2013年地方高校国家级大学生创新创业训练计划项目(201310462103)

作者(Author): 刘超锋,柳金江,赵伟,吴学红,刘亚莉,龚毅
LIU Chao-feng,LIU Jin-jiang,ZHAO Wei,WU Xue-hong,LIU Ya-li,GONG Yi

参考文献(References)：

• [1]舒水明,陈彩霞,杨斌,等.大型低压空分流程精馏提氩过程[J].华中科技大学学报(自然科学版),2009,37(4):101-104.
• [2]脱瀚斐,厉彦忠,谭宏博.空气分离与天然气液化组合循环[J].化工学报,2008,59(10):2498-2503.
• [3]车进,刘大铭,胡钢墩,等.离心式CO2压缩机流量软测量研究[J].流体机械,2012,40(12):6-10.
• [4]黄华.基于SPSS的PCA-RBF神经网络模型的仿真试验[J].新疆农业大学学报,2010,33(6):535-538.