节点文献
基于实验数据的湍流扩散模态分析
Modal Analysis of Turbulent Diffusion Based on Experimental Data
【作者】 王芳;
【作者基本信息】 天津大学 , 工程硕士(专业学位), 2021, 硕士
【摘要】 湍流条件下的物质扩散现象在化工过程中十分常见,因此深入研究湍流传质扩散的规律,对精馏、吸收、反应等单元操作具有重要意义。本文采用粒子成像测速(PIV)和激光诱导荧光(LIF)技术同时测量了水平通道中湍流传质扩散过程的速度和荧光剂浓度瞬时分布。此外,还测量了不同切面位置的瞬时浓度分布。通过对实验测量结果的分析,得到了湍流传质扩散过程平均浓度分布的三维结构,并分析了不同切面位置平均浓度分布的变化规律,且下游平均速度分布呈现中心区域速度比壁面附近速度大的趋势。为了从复杂的实验数据中提取湍流传质扩散的特征,并进一步实现瞬时浓度分布的数值重构和预测,本文通过本征正交分解(Proper orthogonal decomposition,POD)和动力学模态分解(Dynamic mode decomposition,DMD)两种模态分析方法,对湍流扩散中荧光剂的瞬时浓度分布实验数据进行了模态分解。首先将POD模态分析方法应用于瞬时浓度分布的特征分析中,得到了瞬时浓度分布的各阶POD空间模态,并分析了各阶空间模态的分布特征及其所含的能量。结果表明,瞬时浓度分布在距离荧光剂靠近入口的区域具有较强的周期性,且低阶POD空间模态的能量占主导地位。应用POD分析得到的空间模态可以较准确地重构荧光剂入口附近区域的瞬时浓度分布,重构的相对误差在远离荧光剂入口的区域较大。这表明,本文提出的基于POD模态分析的湍流瞬时浓度分布重构方法,更适用于周期性较强的系统和区域,为预测未知时刻的湍流扩散瞬时浓度分布提供了基础。随后,本文应用DMD对瞬时浓度分布进行模态分析,得到了DMD空间模态和特征值。结果表明,DMD空间模态与POD空间模态十分相似,但DMD特征值实部和虚部均在单位圆内或单位圆上,表明了空间模态的稳定性。基于DMD空间模态方法,得到的瞬时浓度分布重构结果比POD重构结果的的误差相对更小。最后,本文基于DMD方法对未知时刻的湍流传质扩散瞬时浓度分布进行了预测。
【Abstract】 The diffusion phenomenon of substances under turbulent flow conditions is very common in chemical processes.Therefore,in-depth study of the law of turbulent mass transfer and diffusion is of great significance to unit operations such as distillation,absorption,reaction,etc.In this paper,velocity and concentration distributions of the fluorescent agent in turbulent liquid flow in a horizontal channel were experimentally measured by simultaneous use of particle image velocimetry(PIV)and laser induced fluorescence(LIF)technologies.In addition,the instantaneous concentration distribution at different sections of the diffusion process is also measured.The three-dimensional structure of the mean concentration distribution in the turbulent mass transfer and the variation rule of the mean concentration distribution at different sections are obtained through the analysis of the experimental results.The average velocity distribution in the downstream region tends to be larger than that near the wall.In order to extract the characteristics of turbulent mass transfer and diffusion from complex experimental data and further realize the numerical reconstruction and prediction of instantaneous concentration distribution,in this paper,two modal analysis methods,proper orthogonal decomposition(POD)and dynamic mode decomposition(DMD),were used to decompose the instantaneous concentration distribution in the process of fluorescent agent diffusion under turbulent conditions.Firstly,the POD mode analysis method was applied to analyze instantaneous concentration distribution characteristics and obtain different modes of concentration distribution.The spatial characteristics of the modes and their energy were analyzed.The results show that the concentration distribution is periodic in the region closer to the fluorescent agent injection nozzle.The low-order POD modes have higher energy and dominate the turbulent mass transfer process.It can be concluded that the POD analysis can be used to reconstruct the characteristics of concentration distribution near the fluorescent agent injection nozzle numerically,and the relative error is mainly concentrated in the area which is far from the injection nozzle.This result suggests that the POD analysis approach proposed in the present paper is suitable for numerical reconstruction of turbulent concentration distribution with periodicity and thus lays a foundation for effective prediction.Then the DMD was applied to the instantaneous concentration distribution modal analysis,and the DMD space mode and eigenvalues were obtained.The results show that the DMD spatial modes are very similar to the POD modes.The real and imaginary parts of the DMD eigenvalues are within or on the unit circle,which indicates the stability of the spatial modes.Based on the DMD spatial modal method,the error of the instantaneous concentration distribution reconstruction results is smaller than that of POD reconstruction results.Finally,the DMD method predicts the instantaneous concentration distribution of turbulent mass transfer diffusion.
【Key words】 turbulent diffusion; POD analysis; DMD analysis; PIV; LIF;
- 【网络出版投稿人】 天津大学 【网络出版年期】2024年 11期
- 【分类号】TQ021.4