基于多尺度的气液两相流图像信号质量指数谱分析

针对多重分形质量指数谱曲率和面积的概念,应用分形特征最为典型的LORENZ信号与HENON信号以及Cantor集对两个参数进行验证,结果证明两个参数在探测混沌时间序列复杂性方面是完全有效的。引入了多尺度的分析方法,详细研究了气液两相流图像灰度波动信号在不同尺度频率下的两个参数的变化情况,确定采用尺度5信号进行分析。在不同流型的5尺度灰度波动信号下获得了非线性特征值,通过分类指标和单因素方差分析(ANOVA)检验结果,并同其余文献中应用的混沌吸引子形态周界特征(吸引子面积、长轴、短轴)相对比,发现此特征值对不同流型具有较好的分类效果,且在不同液相流速下随气相速度的增加能够揭示出气液两相流型机理结构变化情况,是一种有效的研究气液两相流的方法。

Mass exponent spectrum analysis of image signals of gas-liquid two-phase flow based on multiple scale factors

To the concept of multi-fractal mass exponent spectrum's curve-rate and area,the two parameters are verified by LORENZ signal,HENON signal and Cantor set,which are most typical signals of fractal characteristics.The results show that the two parameters are completely effective in the detection of the complexity of chaotic time series.The change of gray-scale fluctuations of the images in a different scale and frequency is studied by the multi-scale method.Then determine that the scale of 5-signal is used.The nonlinear character is acquired in the gray-scale fluctuations of 5 of different flow patterns.It is better classified results for different flow patterns through the test results of classification index and single-factor variance(ANOVA),the comparison with the form and perimeter characteristics of chaotic attractor(attractors area,long axis,short axis).And the change of the mechanism of gas-liquid two-phase flow patterns with the increase of the gas-phase velocity in different liquid-phase flowing velocity.It is an effective approach for gas-liquid two-phase flows.