基于压力特性的管内蒸汽射流凝结流型识别
|
摘要
本文实验研究了蒸汽质量流速0～800 kg·m~(-2)·s~(-1)和管内流动水温度30～70℃的条件下蒸汽喷射进入流动水中的凝结波动特性。运用小波多分辨率变换工具,提取与四种射流凝结流型相映射的凝结压力波动信号在时-频域不同尺度上的特征参数,结果表明尺度2～5能反映凝结压力样本的主要信息特征。经过对比,选取尺度2～5的相对能量和绝对值平均数作为特征量,通过聚类分析构建四维样本空间(前四个主成分的累计贡献率ACR> 85%),并结合神经网络模型提出一种多参数融合的流型在线自动识别方法,得到管内蒸汽射流凝结中的间歇振荡、界面振荡、气泡振荡、稳定凝结四种流型识别率均超过91.7%。
A series of experiments on steam injected into flowing water in a vertical pipe were carried out. The inlet steam flow rate and the flowing water temperature were in the range of 0～800 kg·m~(-2)·s~(-1)and 30～70℃, respectively. The features of condensation pressure signals at different scales over a time-frequency plane were revealed by the method of wavelet multiresolution analysis.Absolute mean and relative energy of condensation pressure signals at levels 2～5 were selected as the characteristic parameters. Principal component analysis was applied to extract four principal component parameters, and the accumulative contribution rate exceeds 85%. A regime intelligent recognition system is constructed based on the neural network model. The recognition rates of Chugging, Ocil-I, Ocil-II and stable condensation in the pipe flow system were all over 91.7%.
A series of experiments on steam injected into flowing water in a vertical pipe were carried out. The inlet steam flow rate and the flowing water temperature were in the range of 0～800 kg·m~(-2)·s~(-1)and 30～70℃, respectively. The features of condensation pressure signals at different scales over a time-frequency plane were revealed by the method of wavelet multiresolution analysis.Absolute mean and relative energy of condensation pressure signals at levels 2～5 were selected as the characteristic parameters. Principal component analysis was applied to extract four principal component parameters, and the accumulative contribution rate exceeds 85%. A regime intelligent recognition system is constructed based on the neural network model. The recognition rates of Chugging, Ocil-I, Ocil-II and stable condensation in the pipe flow system were all over 91.7%.
引文
[1] Chun M H, Kim Y S, Park J W. An Investigation of Direct Condensation of Steam Jet in Subcooled Water[J].International Communications in Heat and Mass Transfer,1996, 23(7):947-958
[2] Xu Q, Guo L, Chang L. Mechanisms of Pressure Oscillation in Steam jet Condensation in Water Flow in a Vertical Pipe[J]. International Journal of Heat and Mass Transfer,2017, 110:643-656
[3] Chan C K, Lee C K B. A Regime Map for Direct Contact Condensation[J]. International Journal of Multiphase Flow, 1982, 8(1):11-20
[4] Xu Q, Guo L, Zou S, et al. Experimental Study on Direct Contact Condensation of Stable Steam Jet in Water Flow in a Vertical Pipe[J]. International Journal of Heat and Mass Transfer, 2013, 66:808-817
[5] Webb A R. Statistical Pattern Recognition[M]. John Wiley&Sons, 2003
[6] Ye J, Guo L. Multiphase Flow Pattern Recognition in Pipeline-riser System by Statistical Feature Clustering of Pressure Fluctuations[J]. Chemical Engineering Science,2013,102:486-501
[2] Xu Q, Guo L, Chang L. Mechanisms of Pressure Oscillation in Steam jet Condensation in Water Flow in a Vertical Pipe[J]. International Journal of Heat and Mass Transfer,2017, 110:643-656
[3] Chan C K, Lee C K B. A Regime Map for Direct Contact Condensation[J]. International Journal of Multiphase Flow, 1982, 8(1):11-20
[4] Xu Q, Guo L, Zou S, et al. Experimental Study on Direct Contact Condensation of Stable Steam Jet in Water Flow in a Vertical Pipe[J]. International Journal of Heat and Mass Transfer, 2013, 66:808-817
[5] Webb A R. Statistical Pattern Recognition[M]. John Wiley&Sons, 2003
[6] Ye J, Guo L. Multiphase Flow Pattern Recognition in Pipeline-riser System by Statistical Feature Clustering of Pressure Fluctuations[J]. Chemical Engineering Science,2013,102:486-501