湿蒸汽凝结流动过程湿度软测量模型  

Soft-sensing model of humidity in wet steam condensation flow process

作　　者：黄竹青[1] 左从瑞[1] 曹小玲[1] 

机构地区：[1]长沙理工大学能源与动力工程学院,湖南长沙410076

出　　处：《中南大学学报（自然科学版）》2012年第4期1532-1537,共6页Journal of Central South University:Science and Technology

基　　金：湖南省自然科学基金资助项目(05JJ30207)

摘　　要：针对汽轮机末级湿蒸汽凝结流动过程湿度分布影响因素的复杂性和非线性,采用灰色关联分析方法对实验测得的质量中间半径、水滴数密度、蒸汽湿度与初压、进汽口温度、背压、尾部温度、水比容、汽比容的相互关系进行关联分析处理,确定影响因子之间的主次关系;并以影响湿蒸汽凝结流动过程湿度分布的主要因素的初压、进汽口温度、背压、尾部温度、水比容、汽比容等作为支持向量机网络的输入,建立湿蒸汽凝结流动过程支持向量机网络预测模型,结果表明:支持向量机网络预测湿度的最大相对误差小于9%,以初压、入口温度、背压、尾部温度、水比容、汽比容为输入的支持向量机网络具有较高的精度,可以很好地预测汽轮机末级湿蒸汽凝结流动过程的湿度分布。The humidity distribution in wet steam condensation flow process is affected by many factors,and their relationship is complex and nonlinear.A grey relational analysis method replacing the traditional mathematical statistic was used to deal with the relationship between droplet distribution parameter,droplet number density,steam humidity and initial pressure,initial temperature,back pressure,outlet temperature,specific volume of water,specific volume of steam,define relationship between primary and secondary impacting factors.Moreover,the forecasting model was built with the primary factors affecting humidity distribution in wet steam condensation flow process-initial pressure,initial temperature,back pressure,outlet temperature,specific volume of water,specific volume of steam used as support vector machine network input.The result shows that the maximum relative error of predicted and actual values is less than 9%,which illustrates this model is reliable.

关 键 词：汽轮机 蒸汽湿度 灰色关联分析 支持向量机 

分 类 号：TK33[动力工程及工程热物理—热能工程] O436[机械工程—光学工程]