PDF(1491 KB)
PDF(1491 KB)
PDF(1491 KB)
基于图信号排列熵的气液两相流流动特性分析与流型识别
Flow Characteristics Analysis and Flow Pattern Identification of Gas-liquid Two-phase Flow Based on Permutation Entropy for Graph Signals
提出了一种基于图信号排列熵的垂直管道气液两相流流型辨识方法。使用数字化电阻层析成像系统采集垂直管道气液两相流实验数据,计算每个电阻层析成像(ERT)电极测量值的幅值增量序列,提取每个幅值增量序列的图信号排列熵,并分析各个流型的流动特性,将提取的图信号排列熵作为特征输入卷积神经网络(CNN)以识别流型。结果表明:该方法能够有效识别泡状流、泡状-段塞流、段塞流,平均正确辨识率可达96.67%。
Based on graph signal permutation entropy,a flow pattern identification method for gas-liquid two-phase flow in vertical pipelines is proposed. A digital electrical resistance tomography (ERT) system to collect vertical pipeline gas-liquid two-phase flow experimental data is used, the amplitude increment sequence of each ERT electrode measurement value is calculate, and the permutation entropy for graph signals of each amplitude increment sequence is extracted, the flow characteristics of each flow pattern is analyzed. The extracted graph signal permutation entropy is input into a convolutional neural network (CNN) as a feature to identify flow patterns. The results show that this method can effectively identify bubble flow, bubble-slug flow and slug flow, and the average correct recognition rate can reach 96.67%.
流量计量 / 气液两相流 / 电阻层析成像 / 图信号排列熵 / 流型辨识
flow metrology / gas-liquid two-phase flow / electrical resistance tomography / permutation entropy for graph signals / flow pattern identification
| 1 |
张立峰,王智.基于多元经验模态分解与卷积神经网络的气液两相流流型识别[J].计量学报,2023,44(1):73-79.
|
| 2 |
张立峰,王智.基于递归图的两相流流动特性分析与流型识别[J].计量学报,2022,43(11):1438-1444.
|
| 3 |
|
| 4 |
|
| 5 |
|
| 6 |
周航锐,孙坚,徐红伟,等.基于EEMD和低秩稀疏分解的超声缺陷回波检测方法[J].计量学报,2022,43(1):77-84.
|
| 7 |
|
| 8 |
|
| 9 |
|
| 10 |
|
| 11 |
|
| 12 |
|
| 13 |
|
| 14 |
|
| 15 |
|
| 16 |
张立峰,张思佳,刘帅.基于Choi-Williams分析与神经网络的两相流流型识别[J].计量学报,2023,44(12):1819-1826.
|
/
| 〈 |
|
〉 |