基于新型C~4 D的小管道气液两相流流型辨识方法

发布时间：2018-08-02 11:33

【摘要】：基于径向结构的电容耦合式非接触电阻抗检测传感器,结合小波包分析技术和K-均值聚类算法,提出一种小管道气液两相流流型辨识方法。首先,利用径向结构的电容耦合式非接触电阻抗检测传感器,获取反映被测流体信息的电阻抗测量信号实部信息和虚部信息。然后,采用小波包分解的信号处理技术将实部信息和虚部信息分别分为4个频率段,提取不同频率范围的能量分布情况,并与各自的均值、方差构成特征向量。最后,利用K-均值聚类算法进行模式分类,建立流型辨识模型。在内径为3.5 mm和5.5 mm的玻璃管道内进行验证实验,实验结果表明,所获得的传感器测量信号能反映流体流动信息,提出的流型辨识技术路线是有效的,流型辨识精度可达88%以上。
[Abstract]:Based on the radial structure of capacitance-coupled non-contact impedance detection sensor, a flow pattern identification method for gas-liquid two-phase flow in a small pipeline is proposed by combining the wavelet packet analysis technique and the K-means clustering algorithm. Firstly, the capacitive coupled non-contact impedance detection sensor with radial structure is used to obtain the real and imaginary part information of the impedance measurement signal which reflects the information of the fluid under test. Then, the real part information and the imaginary part information are divided into four frequency segments by the wavelet packet decomposition technique. The energy distribution in different frequency ranges is extracted, and the eigenvector is constructed with their mean and variance. Finally, the K-means clustering algorithm is used for pattern classification and flow pattern identification model is established. The experimental results in a glass pipe with an inner diameter of 3.5 mm and 5.5 mm show that the measured signals of the sensor can reflect the fluid flow information, and the proposed flow pattern identification technique is effective. The accuracy of flow pattern identification can be over 88%.
【作者单位】：浙江大学控制科学与工程学院工业控制技术国家重点实验室;
【基金】：国家自然科学基金(51476139,61573312)~~
【分类号】：O359;TP212

【相似文献】