基于LS-SVM的多传感器气体质量流量测量  

The Gas Mass Flow Measurement Based on Multi-Sensor and LS-SVM

作　　者：赵伟国[1] 郑永军[1] 孙斌[1] 郑恩辉[1] 梁国伟[1] 李文军[1] 

机构地区：[1]中国计量学院计量测试工程学院,杭州310018

出　　处：《传感技术学报》2010年第5期739-743,共5页Chinese Journal of Sensors and Actuators

基　　金：浙江省自然科学基金资助(Y1080950);国家质检总局科技计划项目资助(2006QK66)

摘　　要：针对流量测量中流速分布不规则对气体流量测量精度的影响,提出了一种多传感器气体质量流量测量新方法。该方法基于均速管测量原理,在测量管道中按照对数线性法分布了4个热式气体流量传感器,采集不同特征位置的流量。通过流量标定实验,获得不同质量流量下测量管道内4个传感器的电压。然后利用GA和LS-SVM算法,将传感器电压和气体的质量流量作为训练集,建立了气体流量模型。实际测量中由4个传感器的电压计算出气体的质量流量。不规则流场的流量实验结果表明该方法是有效的。A new multi-sensor method was proposed for the gas mass flow measurement in the disorder flow field of the pipeline.The method principle was based on the averaging pitot flowmeter.Four thermal flow sensors were distributed in the pipeline according to the Log-linear method in order to acquire the flowrate of different characteristic point.Through the calibration test,the voltage of the sensors was obtained at different mass flowrate.From these data,the gas mass flow model was established based on GA and LS-SVM.In the measurement process,the mass flow was computed by using the model.The experiment reseach showed the mesurement is valid.

关 键 词：多传感器 质量流量 对数线性法 最小二乘支持向量机 遗传算法 

分 类 号：TB937[一般工业技术—计量学]