摘要
提出了一种两相流离散相浓度软测量的新方法 ,该方法以 12电极电容传感器为信息获取手段 ,应用神经网络技术进行信息处理以获得离散相浓度 .利用电极间的对称性 ,将网络结构进行了化简 .实验结果表明 ,浓度测量最大误差小于 5%
electrode capacitance systems for the concentration measurement of two-phase flow have inherently non-uniform sensitivity distributions over the cross section of the pipe and therefore have different response to different flow regimes. The change of flow regimes can lead to significant measurement errors. 12-electrode capacitance sensor based on tomography technique can give a measurement less flow-regime dependent but the algorithm is difficult to implement in microprocessor based instrument system. A new algorithm, based on artificial neural network, for the concentration measurement of two-phase flow is proposed. Considering the symmetry of the plates the structure of the neural network is simplified. Training set and test set are used to train the network. Test result shows the measurement error is less than 5%.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2001年第1期62-66,共5页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金重大资助项目! (5 9995 46 0 - 5 )
国家自然科学青年基金资助项目!(2 970 6 0 0 8)
关键词
电容传感器
神经网络
浓度
两相流
capacitance sensor
artificial neural network
concentration
two-phase flow
