QQ客服 官方微博 反馈留言

题名	基于介尺度结构的颗粒流体两相流离散模拟方法
作者	宋飞飞
学位类别	博士
答辩日期	2014-05
授予单位	中国科学院研究生院
导师	李静海
关键词	离散颗粒方法   多尺度   颗粒流体两相流   介尺度   LBM   EMMS   MP-PIC
其他题名	Towards Meso-Scale Structure Based Discrete Particle Method for Particle-Fluid Two-Phase Flow
学位专业	化学工程
中文摘要	颗粒流体两相流具有非线性非平衡的特征，其中的多尺度结构对于系统的流动、传质、反应等都具有显著的影响。最近随着计算机技术的发展，离散颗粒方法因其能准确的捕捉颗粒流体两相流中的两相结构而受到越来越广泛的关注。然而现有离散化方法在流固耦合方面往往采用连续介质方法的处理方式，忽略了离散化方法给出的对应颗粒位置的非均匀信息，使得颗粒与流体在不同的尺度上进行计算，影响了计算的准确度；而对于流体计算方面，传统TFM方法受到自身并行性能的限制，计算速度较难满足大规模的计算需求，需要寻求具有更佳并行性能的方法。同时，虽然介尺度结构及其相关能量关系对于系统有显著的影响，但是目前局部区域上基于介尺度结构的能量关系仍然不明，系统局部所遵循的稳定性条件仍然欠缺。 鉴于此种情况，为了实现高效、准确的颗粒流体两相流模拟这一长远目标，本文建立了应用于离散化方法的基于结构的曳力模型，结合并改进了MP-PIC方法，对于流体相计算建立了基于LBM算法的并行效率优异的TF-LBM方法，并对于系统能量关系进行了初步的尝试与探索。 论文第二章首先在开源软件MFIX的基础之上实现了MP-PIC的程序计算，为了验证程序，对于一些鼓泡床和循环流化床提升管算例进行了模拟，结果显示与实验结果吻合，但采用传统均匀曳力会产生较大的偏差。而直接采用TFM层次的EMMS曳力如前所述会影响计算的准确度。 为解决这一准确性问题，论文第三章提出了面向DPM的EMMS曳力模型，模型引入了颗粒位置所代表的非均匀信息，并因此将控制单元内气固相划分为一个稀相及多个密相，采用压降平衡条件对各相中流体进行分配，以此回归拟合，得到改进的曳力系数。结果显示，相比于均匀的曳力系数，面向DPM的EMMS曳力模型的计算结果准确度有了显著的改善。 为实现快速模拟的目标，论文第四章提出了TF-LBM算法，它能够在保留LBM并行性能的前提下，在TFM的尺度上对于流体相进行计算，从而可望显著加快流体相的计算速度。通过一些简单的算例，对于本方法也进行了初步的验证。 最后，为探索介尺度结构对系统的影响，论文第五章对于气固流态化系统中的能量关系进行了分析。本工作加深了对于系统能耗的组成和EMMS模型中各变量的物理意义及关系的理解，通过具体的分析讨论了EMMS稳定性条件不能在局部成立的可能原因，对于未来建立局部稳定性条件与介尺度结构的关联提供了一定的基础。 论文最后总结了所得到的主要结论，并对于模型的前景以及进一步研究的方向进行了展望。
英文摘要	Particle-fluid two-phase flow is characterized with nonlinear, non-equilibrium phenomena, and its multi-scale dynamic structure has significant effects on flow, mass and heat transfer and reaction of the system. Recently, along with the improvements of computing technology, the discrete particle method (DPM) has gotten more and more attentions because it can capture the detail of two-phase structure in particle-fluid two-phase flow. However, the drag used in DPM is normally the same as in continuum methods, and researchers ignore the heterogeneous information given in DPM, which can be calculated with particle positions. This leads to the incompatibility between resolutions of particle and fluid phases, and affects the accuracy of the simulations. In addition, in DPM the fluid solver is normally at the level of TFM, which is heavily restricted in parallel performance, and not suitable for large-scale simulations. In view of these problems, to realize efficient and accurate simulation of particle-fluid two-phase flow, a structure-based drag model was proposed to achieve better accuracy. And a preliminary TF-LBM model was proposed towards fast simulation of the fluid phase at the TFM level. In chapter 2, to realize simulations with DPM, based on the open source software MFIX, we realized the program of MP-PIC. Some cases of bubbling fluidized bed and fast fluidized beds are tested. And results showed good agreements with experimental data, which is much better than using homogeneous drags. In chapter 3, to improve the accuracy, we proposed an EMMS drag model for DPM with consideration of particle position and coupled it with DPM. In this model, the system is divided into a dilute phase and several dense phases, for which the fluid distributions are determined by the pressure balance. The simulation results showed significant improvement as compared with using homogeneous drag. In chapter 4, in order to make the simulation more efficient, the TF-LBM scheme was presented, which could retain the benefit of parallel performance in LBM and at the same time was solved at the level of TFM. Some typical cases in particle-fluid two-phase flows were tested to preliminarily validate this new scheme. In chapter 5, to discuss the effects of meso-scale structures, we analyzed the energy relations in gas-solid fluidization system, which shed light to the understanding of energy composition of system and physical meaning of variables in EMMS. Finally, we summarized the main conclusions and presented the perspective of this work.
语种	中文
公开日期	2015-07-08
内容类型	学位论文
源URL	[http://ir.ipe.ac.cn/handle/122111/15522]
专题	过程工程研究所_研究所（批量导入）
推荐引用方式 GB/T 7714	宋飞飞. 基于介尺度结构的颗粒流体两相流离散模拟方法[D]. 中国科学院研究生院. 2014.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。