模拟颗粒流体系统的混合动态多尺度方法
发布时间：2019-05-29 00:19
【摘要】：颗粒流体系统广泛存在于自然界并应用于制药、冶金、能源、化工等工业过程。近年来,由于计算机的快速发展,基于计算流体力学对颗粒流和气固两相流的模拟方法受到了越来越多的关注,并广泛地应用于工业反应器的放大、设计和优化过程中。目前已有的模拟颗粒流和气固两相流的模型主要分为两大类：基于欧拉方法的连续模型和基于拉格朗日追踪法的离散模型。连续模型一般采用连续介质假设并结合颗粒动理论,其优点是计算量较小,因此可以用来模拟大型工业问题。但是,由于连续模型把固相处理成“拟流体”,其精度和普适性都低于基于拉格朗日追踪的离散模型。然而,在离散模型中,每个颗粒的运动轨迹都需要用单独的方程追踪,所以对工业装置的模拟将会产生庞大的计算量,目前的计算能力远无法满足。为解决连续模型精度有限及离散模型效率低的矛盾,本论文开发了一种混合动态多尺度模型(离散-连续跨尺度耦合模型)用以解决连续模型的精度和离散模型的计算效率问题。该模型的核心是在一个模拟体系中同时使用连续和离散两种模型,即只在连续模型不适用的关键区域使用离散模型来捕捉准确的流动细节,而其余大部分区域仍然使用连续模型以降低计算量。离散和连续两种模型在空间尺度和时间尺度上通过合理的物理模型实现动态耦合,从而兼顾离散模型的精度和连续模型的计算经济性。本论文围绕混合多尺度模型的模型建立、算法实现和实际应用展开研究,主要内容安排如下：(1)论文第一章首先阐述了研究背景与意义并对颗粒流体两相流的模拟方法进行了概述；紧接着对混合多尺度模型在单相流中应用的文献进行了综述；最后提出了本论文的研究思路。(2)混合多尺度模型需要同时使用连续模型和离散模型,而目前已有的商业软件难以实现两种模型的扩展和实时交互。因此本论文在第二章叙述了两种模型各自的算法实现过程,并使用程序模拟了一个典型的提升管装置。计算结果显示连续模型和离散模型都与实验结果定性吻合,但由于实验测量及物理模型参数存在不确定因素,模拟的定量分析结果与实验测量仍有一定差别。(3)论文第三章建立了适用于颗粒流的混合多尺度模型框架,并通过模拟颗粒管流初步验证了所构建的混合多尺度模型的合理性。在该模型中,本文构造了一个重叠区域以使不同模型在此处交换边界条件。重叠区域又可进一步分为三个不同用途的子区域,即连续模型向离散模型提供边界的区域、缓冲区域和离散模型向连续模型提供边界的区域。在进口固相速度为抛物线速度分布的算例中,纯离散模型、纯连续模型及混合多尺度模型三者得到了定性一致的径向固相体积分数分布及速度分布。然而,在进行定量对比时发现纯连续模型同纯离散模型的模拟结果在壁面附近的偏差较大,其中原因是在壁面附近努森数较大使得连续介质假设失效。因此,本文在构造混合多尺度模型模拟区域时,在壁面附近使用离散模型,而在管中心仍然使用连续模型以减少计算量。混合多尺度模型模拟的定量结果同纯离散模型吻合很好。(4)论文第四章进一步将混合多尺度模型扩展到气固两相流系统,并使用建立的混合多尺度模型模拟了无重力条件下的气体-颗粒管流及Geldart D类粗颗粒的循环流态化。在气体-颗粒管流计算中,双流体模型预测的管壁附近处的固含率、固相速度及气相速度与颗粒轨道模型相差较大,而混合多尺度模型则与颗粒轨道模型吻合较好。在粗颗粒的循环流态化算例中,由于在壁面及进出口附近处的颗粒努森数较大,混合多尺度模型的结果与颗粒轨道模型的一致性比双流体模型更好。(5)工业流化床反应器常使用Geldart A类及B类细颗粒,并且在流化床内存在着团聚物等非均匀结构,因此在模拟这些细颗粒的循环流态化时需要考虑介尺度结构的影响。第五章建立了考虑介尺度结构的混合多尺度模型,其核心是使用了能量最小多尺度(EMMS)模型将气固系统处理为稀密两相结构的思路。该模型将基于EMMS的介尺度双峰速度分布函数应用在连续模型和离散模型的相互边界映射过程中,并引入EMMS稳定性条件来约束新插入颗粒位置的选择。考虑介尺度结构的混合多尺度模型得到的轴向时均固相体积分数分布及径向的时均流场变量都与颗粒轨道模型的模拟结果吻合良好。模拟结果显示本文所开发的混合多尺度模型充分利用了连续模型和离散模型各自的优势,为实现颗粒流体系统的高效、准确地实时模拟及虚拟过程工程提供了相应的理论基础。
[Abstract]:The particle fluid system is widely used in the natural world and applied to the industrial processes such as pharmacy, metallurgy, energy, chemical industry and the like. In recent years, due to the rapid development of the computer, the simulation method of the particle flow and the gas-solid two-phase flow based on the computational fluid dynamics has attracted more and more attention, and is widely used in the process of amplification, design and optimization of the industrial reactor. At present, the models of the simulated particle flow and the gas-solid two-phase flow are mainly divided into two types: the continuous model based on the Euler method and the discrete model based on the Lagrange tracking method. The continuous model generally adopts a continuous medium assumption and combines the particle motion theory, and has the advantages that the calculation amount is small, and therefore, the continuous model can be used for simulating large-scale industrial problems. However, due to the continuous model, the solid phase is treated as a "pseudofluid", and the accuracy and universality of the solid phase are lower than that of the discrete model based on the Lagrange tracking. However, in the discrete model, the motion track of each particle needs to be tracked by a separate equation, so the simulation of the industrial device will produce a large amount of calculation, and the present computing power is far from being met. In order to solve the contradiction between the precision of the continuous model and the low efficiency of the discrete model, a mixed dynamic multi-scale model (discrete-continuous cross-scale coupling model) is developed to solve the problem of the precision of the continuous model and the calculation efficiency of the discrete model. The core of this model is to use both continuous and discrete models in an analog system, that is, using discrete models only in critical areas where the continuous model is not applicable to capture accurate flow details, while the rest of the region still uses a continuous model to reduce the amount of computation. The discrete and continuous models achieve dynamic coupling through a reasonable physical model on the spatial scale and time scale, thus taking into account the precision of the discrete model and the computational economy of the continuous model. In this paper, the model establishment, algorithm realization and practical application of the mixed multi-scale model are studied. The main contents are as follows: (1) The first chapter of the thesis first expounds the background and significance of the research and gives an overview of the simulation method of the particle-fluid two-phase flow; The paper reviews the application of the mixed multi-scale model in the single-phase flow, and finally puts forward the research ideas of the paper. (2) The mixed multi-scale model needs to use the continuous model and the discrete model at the same time, and the existing commercial software is difficult to realize the expansion and real-time interaction of the two models. Therefore, in the second chapter, the paper describes the implementation process of the two models, and uses the program to simulate a typical riser device. The results show that the continuous model and the discrete model are in good agreement with the experimental results, but due to the uncertain factors in the experimental measurement and the physical model parameters, the results of the simulation and the experimental measurement still have a certain difference. (3) The third chapter of the thesis establishes a mixed multi-scale model framework for particle flow, and verifies the rationality of the constructed mixed multi-scale model by simulating the particle pipe flow. In this model, an overlapping region is constructed to allow different models to exchange boundary conditions here. The overlap region may in turn be further divided into three sub-regions of different uses, i.e., the region where the continuous model provides the boundary to the discrete model, the buffer region, and the region where the discrete model provides the boundary to the continuous model. In an example of the distribution of the velocity of the inlet solid phase as a parabola, a qualitative and consistent radial solid volume fraction distribution and a velocity distribution are obtained from the pure discrete model, the pure continuous model and the mixed multi-scale model. However, when the quantitative comparison is carried out, it is found that the simulation results of the pure continuous model and the pure discrete model are larger in the vicinity of the wall surface, and the reason is that the number of the Nousin near the wall surface is large, so that the continuous medium is assumed to be invalid. Thus, in constructing a mixed multi-scale model model, a discrete model is used near the wall surface, while a continuous model is still used at the center of the tube to reduce the amount of calculation. The quantitative results of the mixed multi-scale model are in good agreement with the pure discrete model. (4) The fourth chapter of the paper further extends the mixed multi-scale model to the gas-solid two-phase flow system, and uses the established mixed multi-scale model to model the circulation fluidization of the gas-particle pipe flow and the Geldart D-type coarse particles under the gravity-free condition. In the gas-particle pipe flow calculation, the solid-solid content, the solid-phase velocity and the gas-phase velocity in the vicinity of the tube wall predicted by the two-fluid model are quite different from that of the particle track model, and the mixed multi-scale model is in good agreement with the particle track model. The results of the mixed multi-scale model are better than that of the two-fluid model due to the large number of particles in the vicinity of the wall and the inlet and outlet, and the results of the mixed multi-scale model are better than that of the two-fluid model. (5) The industrial fluidized bed reactor often uses the Geldart A and B fine particles, and there are non-uniform structures such as agglomerates in the fluidized bed, so the influence of the medium-scale structure needs to be taken into account in the simulation of the circulation fluidization of these fine particles. In the fifth chapter, a mixed multi-scale model considering the medium-scale structure is established, the core of which is the idea of using the energy minimum multi-scale (EMMS) model to treat the gas-solid system as a dilute-dense two-phase structure. The model is applied to the boundary mapping process of the continuous model and the discrete model based on the medium-scale double-peak velocity distribution function of the EMMS, and introduces the EMMS stability condition to restrain the selection of the new inserted particle position. The axial time-averaged solid volume fraction distribution and the radial flow field variable obtained by the mixed multi-scale model of the medium-scale structure are in good agreement with the simulation results of the particle track model. The simulation results show that the hybrid multi-scale model developed in this paper makes full use of the advantages of the continuous model and the discrete model, and provides a corresponding theoretical basis for realizing the efficient and accurate real-time simulation of the particle fluid system and the virtual process engineering.
【学位授予单位】：中国科学院研究生院(过程工程研究所)
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TQ021
，
本文编号：2487487
[Abstract]:The particle fluid system is widely used in the natural world and applied to the industrial processes such as pharmacy, metallurgy, energy, chemical industry and the like. In recent years, due to the rapid development of the computer, the simulation method of the particle flow and the gas-solid two-phase flow based on the computational fluid dynamics has attracted more and more attention, and is widely used in the process of amplification, design and optimization of the industrial reactor. At present, the models of the simulated particle flow and the gas-solid two-phase flow are mainly divided into two types: the continuous model based on the Euler method and the discrete model based on the Lagrange tracking method. The continuous model generally adopts a continuous medium assumption and combines the particle motion theory, and has the advantages that the calculation amount is small, and therefore, the continuous model can be used for simulating large-scale industrial problems. However, due to the continuous model, the solid phase is treated as a "pseudofluid", and the accuracy and universality of the solid phase are lower than that of the discrete model based on the Lagrange tracking. However, in the discrete model, the motion track of each particle needs to be tracked by a separate equation, so the simulation of the industrial device will produce a large amount of calculation, and the present computing power is far from being met. In order to solve the contradiction between the precision of the continuous model and the low efficiency of the discrete model, a mixed dynamic multi-scale model (discrete-continuous cross-scale coupling model) is developed to solve the problem of the precision of the continuous model and the calculation efficiency of the discrete model. The core of this model is to use both continuous and discrete models in an analog system, that is, using discrete models only in critical areas where the continuous model is not applicable to capture accurate flow details, while the rest of the region still uses a continuous model to reduce the amount of computation. The discrete and continuous models achieve dynamic coupling through a reasonable physical model on the spatial scale and time scale, thus taking into account the precision of the discrete model and the computational economy of the continuous model. In this paper, the model establishment, algorithm realization and practical application of the mixed multi-scale model are studied. The main contents are as follows: (1) The first chapter of the thesis first expounds the background and significance of the research and gives an overview of the simulation method of the particle-fluid two-phase flow; The paper reviews the application of the mixed multi-scale model in the single-phase flow, and finally puts forward the research ideas of the paper. (2) The mixed multi-scale model needs to use the continuous model and the discrete model at the same time, and the existing commercial software is difficult to realize the expansion and real-time interaction of the two models. Therefore, in the second chapter, the paper describes the implementation process of the two models, and uses the program to simulate a typical riser device. The results show that the continuous model and the discrete model are in good agreement with the experimental results, but due to the uncertain factors in the experimental measurement and the physical model parameters, the results of the simulation and the experimental measurement still have a certain difference. (3) The third chapter of the thesis establishes a mixed multi-scale model framework for particle flow, and verifies the rationality of the constructed mixed multi-scale model by simulating the particle pipe flow. In this model, an overlapping region is constructed to allow different models to exchange boundary conditions here. The overlap region may in turn be further divided into three sub-regions of different uses, i.e., the region where the continuous model provides the boundary to the discrete model, the buffer region, and the region where the discrete model provides the boundary to the continuous model. In an example of the distribution of the velocity of the inlet solid phase as a parabola, a qualitative and consistent radial solid volume fraction distribution and a velocity distribution are obtained from the pure discrete model, the pure continuous model and the mixed multi-scale model. However, when the quantitative comparison is carried out, it is found that the simulation results of the pure continuous model and the pure discrete model are larger in the vicinity of the wall surface, and the reason is that the number of the Nousin near the wall surface is large, so that the continuous medium is assumed to be invalid. Thus, in constructing a mixed multi-scale model model, a discrete model is used near the wall surface, while a continuous model is still used at the center of the tube to reduce the amount of calculation. The quantitative results of the mixed multi-scale model are in good agreement with the pure discrete model. (4) The fourth chapter of the paper further extends the mixed multi-scale model to the gas-solid two-phase flow system, and uses the established mixed multi-scale model to model the circulation fluidization of the gas-particle pipe flow and the Geldart D-type coarse particles under the gravity-free condition. In the gas-particle pipe flow calculation, the solid-solid content, the solid-phase velocity and the gas-phase velocity in the vicinity of the tube wall predicted by the two-fluid model are quite different from that of the particle track model, and the mixed multi-scale model is in good agreement with the particle track model. The results of the mixed multi-scale model are better than that of the two-fluid model due to the large number of particles in the vicinity of the wall and the inlet and outlet, and the results of the mixed multi-scale model are better than that of the two-fluid model. (5) The industrial fluidized bed reactor often uses the Geldart A and B fine particles, and there are non-uniform structures such as agglomerates in the fluidized bed, so the influence of the medium-scale structure needs to be taken into account in the simulation of the circulation fluidization of these fine particles. In the fifth chapter, a mixed multi-scale model considering the medium-scale structure is established, the core of which is the idea of using the energy minimum multi-scale (EMMS) model to treat the gas-solid system as a dilute-dense two-phase structure. The model is applied to the boundary mapping process of the continuous model and the discrete model based on the medium-scale double-peak velocity distribution function of the EMMS, and introduces the EMMS stability condition to restrain the selection of the new inserted particle position. The axial time-averaged solid volume fraction distribution and the radial flow field variable obtained by the mixed multi-scale model of the medium-scale structure are in good agreement with the simulation results of the particle track model. The simulation results show that the hybrid multi-scale model developed in this paper makes full use of the advantages of the continuous model and the discrete model, and provides a corresponding theoretical basis for realizing the efficient and accurate real-time simulation of the particle fluid system and the virtual process engineering.
【学位授予单位】：中国科学院研究生院(过程工程研究所)
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TQ021
，
本文编号：2487487
本文链接：https://www.wllwen.com/kejilunwen/huaxuehuagong/2487487.html
