【作者】 田凤国；

【导师】 章明川；

【作者基本信息】 上海交通大学 ， 热能工程， 2007， 博士

【摘要】 内循环流化床在城市固体废弃物焚烧领域具有独特的优势。本文采用离散单元法(DEM)数值模拟与台架试验相结合的方法,系统研究了流化床内的气体、颗粒流动特性。基于对颗粒相的离散处理,本文利用气固速度场、颗粒加速度场、压力场、压力波动等特征信息量化分析了流化过程机理。可视化观测、物料分层及其停留时间分布等试验研究则是正确认知流化现象、检验数学模型合理与否的第一手段、合理实施工业应用的依据。本课题研究主要包括:流化过程的CCD (Charge Couple Device)可视化观测与DEM数值预报;气体通过床层的流动行为与流量分配、颗粒的微观运动特征;非均匀布风内循环流化床内气泡运动的可视化分析、颗粒流动规律及其动态混合过程的定量评价、物料换热过程的数值模拟;多组分内循环流化床内的分层现象与停留时间分布的试验研究。采用CCD可视化试验详细验证了DEM模拟结果。对比分析显示,数值模拟成功预报了气泡的形成、分离、长大、爆炸等过程。颗粒受力分析表明:在扩散气流曳力和压力梯度力作用下,射流点处颗粒被外推,初始气泡空穴形成,并且逐步长大。随着时间的推进,底部颗粒所受压力梯度力方向逐渐由向外扩张转变为向里收缩,颗粒涌入空穴底部;空穴最终以气泡的形式脱离布风板进入床层。模拟所得气泡周期与试验结果十分接近。压力信号频谱快速傅立叶变换(FFT)分析发现,入口射流速度越快,气泡的产生和通过频率也越高;高射流风速下,高频小幅波动也有所增加。DEM计算过程中,空隙率直接依赖于当地颗粒密度,尾迹的有无则随气泡的进展而变。因此,模拟所得气泡周围压力分布与文献试验结果更为一致:气泡上下两端等压线并不对称,并且内部存在一定的压力梯度。气相速度场直观表明,气泡为低阻空间,具有短路效应,气泡相和乳化相之间存在强烈的气体交换。DEM模拟直观描述了气泡内外的流线特征;流线基本与等压线呈垂直交叉分布,合理反映了流体选择最小阻力途径行进的本质特征;气体流线整体排布较为规则,床内气体表现为层流流动。依据DEM模拟结果,量化考察了床内的流量分配特征。统计显示,在超过临界流化状态的过量气体流量中,可见气泡流量、穿流流量、乳化相中的过量流量分别占7%、36%、57%。穿流作用下,超过临界流化风量的过量流量气流通过气泡捷径导入上部乳化相,而不是生成更多的气泡。流化风速增加,更多的过量气流进入床内乳化相,而两相理论则假定乳化相流量保持在临界流化状不变。可见气泡流动、穿透气流与乳化相气体流动彼此密切相关,将超过临界状态的过量流量完全划归其中某一项则有失恰当。利用颗粒速度场、加速度场、可视化观测图片等数据信息,本文还研究了床内的颗粒运动特征。(1)在床层内部,颗粒围绕上升气泡向下运动;它们在气泡底部“迎面”碰撞,颗粒上抛,并保持一定的容积,形成尾迹;另外,气泡下方颗粒受射流作用进入气泡,也是尾涡来源之一。(2)在床层上部空间,气泡携带颗粒行至床面时呈现不同的弹射现象,主要包括:穹弹射、尾涡弹射、射流弹射。穹爆炸为单气泡爆炸的基本特征;此时,曳力约为压差作用的6倍。风速较高时,尾迹离开床层主体,进入自由空间,出现尾迹喷射。射流风速进一步加快,气泡连续生成,形状拉长、穿透床层,出现高频、高能颗粒射流喷射过程。(3) (u-umf)矢量场和流线显示,气泡爆炸前床层上部空间存在两个局部旋流,左侧涡流呈顺时针方向,右侧为逆时针;而在爆炸过程中,自由空间内的局部气体旋流方向出现逆转。上述局部循环的转换周期与气泡相同。可视化观测发现,由于床层厚度不均,非均匀布风、倾斜布风板流化床中的气泡在上升过程中还发生一定的横向偏移;DEM颗粒速度矢量场直观表明床内物料存在定向循环流动。分析认为,高风速区为气泡活跃区,气泡在上升过程中将向阻力较小的低风速区运动,许多颗粒被带到低风速区。低风速区的颗粒为填补高风速区的气泡离开后留下的空隙将向下移动。另外,采用倾斜布风板设计,低风速区下部的稠密颗粒相在自重作用下具有更为显著的下移趋势。这样就形成了流化床内颗粒的大尺度横向循环流动。颗粒循环流量统计表明,高风速区风速加快,气泡携带颗粒的作用得到加强,内部循环增强;低风速区风速的影响作用与此相反,合理匹配高低风速具有重要的意义。引入Ashton混合指数,并结合示踪颗粒场变化信息,详细讨论了流化床内的微观混合过程。研究发现,床内物料混合过程大致可归纳为快速对流混合、缓慢扩散混合、局部剪切混合三种微观机理。在很大程度上,颗粒循环尺度决定了物料的混合方式。与文献试验对比分析表明,Ashton混合指数评价所得混合时间尺度具有实际意义。对于水平布风板、均匀供风流化床,床内气泡携带颗粒上升运动显著,重力作用下颗粒轴向返混强烈,内部颗粒循环沿轴向贯穿整个床层,轴向混合具有明显的对流特征。均匀布风时,径向气泡特性较为均匀,其横向运动受到限制,内部颗粒循环在径向方向限于气泡周围区域,横向区域间的颗粒交换能力较弱,径向混合则以扩散方式为主。相同条件下,轴向混合速度约为径向混合速度的3.5倍。倾斜布风板、非均匀布风时,左右两侧间的时均颗粒流量远高于水平布风板、均匀布风工况。由于大尺度横向颗粒循环流动,左右两侧区域内的物料存在强烈的对流交换,物料横向混合速度显著加快,与轴向混合几乎相同。建立了DEM流化床传热模型,其中考虑到了气固对流换热与颗粒碰撞传热过程。均匀布风时,受活动范围限制,颗粒局限于同当地气相进行对流换热,其温度分布在较大程度上受当地风温的影响。内循环流化床内的定向颗粒循环运动,增强了物料的横向扩散能力,颗粒在高低风速区交替换热,温度分布有较强的整体平衡恢复能力。颗粒温度场分布对气体温度场分布的依赖程度显著降低。文章最后还针对垃圾成分复杂多变的特点,进行了多组分内循环流化床的混合、分离特性试验研究。分层取样试验结果表明,较大高风速区风速范围内,大块示踪物轴向浓度分布基本保持不变。低风速区风速的增加,该区主体颗粒相运动状态由下移为主转变为流化混合为主,大块物浓度分布发生显著变化。与其他物性相比较而言,大块示踪物密度是影响其轴向浓度分布特性的主要因素。大块物料的平均停留时间先是随高风速区流化速度增加而减小、随后显著回升;对于排渣过程而言,存在一个最有利的流化速度梯度的控制范围。增加低风速区的流化速度,局部返混现象明显,示踪物平均停留时间延长。就示踪物物性而言,对于密度较大的示踪物,若其尺寸增大,则它在床内的停留时间变短。对于密度较小的示踪颗粒,若其尺寸增大,相应在床内的停留时间变长。此外,示踪物形状对停留时间分布起到重要的影响作用。示踪物若趋于球形、表面光滑,则其停留时间变短,反之亦然。研究表明,可燃物与不可燃物处于强分离状态,能够保证充分燃烧与迅速排渣。更多还原

【Abstract】 Internally circulating fluidized beds (ICFB) own their unique advantages in the field of municipal solid waste incineration. Gas-particles flow behaviors in fluidized beds are systemically studied in this dissertation, using discrete element method (DEM) and experimental approaches. Based on the discrete treatment of the particle phase, various kinds of characteristic information are employed to quantitatively analyze fluidization mechanism, including gas-particle velocity fields, particle acceleration fields, pressure fields and pressure fluctuations, and et al. Also, detailed experimental investigations are carried out in this paper, aiming at visual measurements of the bubbles, spatial segregation of the tracer particles, and their residence time distribution (RTD). All these benefit us with more accurate understanding of the fluidization phenomena, provide the first-hand evidence for the mathematical models, and lend bases of favorable industrial applications.Present research mainly consists of following sections: (i) visual measurements of fluidization phenomena using CCD camera and DEM simulations; (ii) gas flow behavior through the bed and its division, microscopic characters of particle movement; (iii) visual analysis of the bubble movement in the ICFB with uneven gas feed, and thereof, investigations on particle streaming, quantitative evaluations of the dynamic mixing process, and numerical predictions of the heat transfer concerning materials in the beds; (iv) experimental researches on the segregating phenomena and RTD of tracers in multi-component ICFBs.CCD experiments are carried out to validate the numerical predictions of DEM under the same condition. Comparisons indicate that numerical simulations successfully reproduce the complicated bubble phenomena, e.g., the arising, escaping, enlarging and bursting of a bubble, et al. Force analysis show that, at the initial stage of a bubble, drag and pressure gradient forces push particles outward and build a small cave near the jetting point. As the time proceeds on, such caves expand. Meanwhile, the directions of the pressure gradient force, exerted on particles at the root of the cave, shift from outward to inward gradually. Consequently, particles fill the bottom of the cave,which finally escapes from the distributor in the shape of a round bubble. Moreover, the predicted bubbles’periods are nearly identical with the experimental results. Fast Fourier transfer (FFT) analysis of the pressure fluctuations manifest that the faster the jetting velocity at the inlet, the higher the frequency of the bubble. Further more, under the faster air jetting velocities, fluctuations of high frequency and small amplitude increase.During DEM calculations, voidage directly counts on local particle densities, and the existence of the wake varies with the bubble development. Therefore, the simulated pressure fields around a bubble agree more with the reported experimental results. Isobars on the top and at the bottom of the bubble are not symmetrical. And, there is pressure gradient within the bubble. In gas velocity fields, it can be vividly found that, being regions of low flow resistance, bubbles serve as a short cut for gas flow, and there is intensive gas exchange between the emulsion phase and the bubble phase. Furthermore, DEM results also give the visual description the patterns of streamlines in the fluidized column. Basically, gas streamlines are normal to isobars, which rationally reflect the flow’s instinct of traveling through paths of the lowest resistance. Also, the regular streamline pictures implicate that fluidizing air embodies laminar flow patterns.Using results of DEM simulations, the gas flux division in the fluidized columns is quantitatively investigated. It is found that, within the excess flux, the proportion of the visible flow, the through-flow, and the interstitial flow in emulsion phase is 7%, 36%, 57%, respectively. Under the effect of through-flow, excess gas flows through the bubble into the above emulsion, in stead of in the form of more visible bubbles. The higher fluidizing air velocity, the more proportion of the excess gas flows into the emulsion. Whereas, it is assumed in two-phase theory that emulsion maintains the status of critical fluidizing condition. In fact, there is a close relationship among the visible bubble flow, through-flow and the interstitial flow, and it is inappropriate to attribute excess flow to one of theses three components.By the means of particle velocity fields, acceleration fields, and CCD snapshots, particle motions are also discussed in detail in this paper. (i) particles moving downward along the boundary of the rising bubble, and collide against each other and extrude at the bottom of the bubble, forming the wake of certain volume; dragged by the jet flow, particles below the bubble tend to penetrate into it, which partially provide the source of the bubble wake, as well. (ii) when bubbles approaching the bed surface, their explosions result in three kinds of particle ejection, including bulge-burst, wake-spike eruption, and jet-spray mechanism. The most common is the bulge burst mechanism due to single bubble bursts, where the drag force on particles is 6 times of the pressure gradient force. As the gas velocity increases, the wake escape from the bed into the freeboard, which is the so-called wake-spike mechanism. At even higher gas supply, the jet-spray mechanism occurs when the trailing bubble elongates, and entrains the solids in the bulge of the leading bubble up above the bed surface. (iii) Consecutive vector fields and streamline pictures of (u-umf) indicate that there are two local swing circles in the freeboard before the bubble eruption, with the left one clockwise and the right anticlockwise. During the eruption, such local circle streams have the opposite direction. Periods of such switch equal to that of the bubble.Visual snapshots of fluidized beds with uneven gas supply and inclined air distributor show that rising bubbles earns lateral displacement, due to different bed depth along the bed width. Particle velocity fields directly illustrate that there are regular granule circulating streams in such beds. With higher gas velocity, the higher velocity side is a bubble active zone, where there is an ascending solids flow carried by the bubbles, consequently. Because of the gradient in particle concentration between the two sides, particles in the right side move to the left. The design of inclined distributor enhances such a descending flow over the distributor as a result of gravity. Hence, an overall convective particle circulation is set up. When gas flow rate in the higher velocity zone increase, the carriage of the bubble are enhanced and results in a more intensive particle circulation. Whereas, change in the lower velocity zone causes reverse effect. The inner circulating flow rate is sensitive to the gas velocities in both zones, suggesting that it is important to configure the ratio of their flux.A statistical mixing index, Ashton index, is firstly introduced to assess the mixing process in fluidized beds in detail, with the aid of tracer particle fields. The mixing process mainly involves the three following microscopic mechanism: fast convective mixing, slow diffusive mixing, and local shear mixing. It is concluded that the scale of inner particle circulation is crucial in determining the mixing patterns within the beds. Comparisons with reported experiment prove that the mixing time scale, evaluated with Ashton mixing index, deserves its practical meaning.In the case of fluidized beds with horizontal distributor and even gas supply, due to bubbles’pronounced ascending movement and the inherent direction of gravitation, particle circulation penetrates through the whole bed axially, which induces a faster convective blending mode along the bed height. On the other hand, bubbles’lateral motion is restrained by their neighbors, and particle circulations are localized. Therefore, particles’horizontal activities are reduced, then the diffusive mixing is the prominent mode in the lateral direction. Under the same condition, the axial mixing rate is 3.5 times of the lateral. For the case of fluidized beds with uneven gas supply and inclined air distributor, the time averaged particle flow rate between the two lateral part of the bed is far larger than the case of horizontal distributor and even gas injection. For the large scale particle circulating stream in ICFBs, materials in different lateral parts enjoys a rapid convective exchange, and the mixing rate is obviously accelerated, with the magnitude nearly equal to the axial mixing rate.DEM heat-transfer models are established for fluidized beds, taking the gas- particle and particle-particle heat exchange into account. Under condition of even gas supply, particles are confined to undergoing convective heat transfer with local gas, and their temperature distribution depends more on the gas phase temperature field in the bed. The regular particle streams improve transverse diffusion performance of the solid phase. As a result, particles are transferred between the high temperature zone and the cold of the gas phase frequently, carrying the heat from the left side to the right. Additionally, turbulent particle flow helps to increase the particle-particle heat transfer. All these benefit the fluidizing system with a better ability to recover homogenous temperature field of particles from external disturbing effects.In terms of the changeable physical properties of the municipal solid waste, segregation and mixing characteristics are also experimentally researched in a multi-component ICFB setup.Results of layer sampling experiments show that there are no significant variations of the distributions of bulky materials within a wide range of air velocities in the higher gas flow rate zone. On the other hand, the distributions of bulky tracers are closely related to the gas velocity in lower gas flow rate zone, since such increase reduces the particle circulation and enhances local mixing performance. Compared with other physical parameters, density of tracer particles plays an important role in determining their axial segregation.The averaged residence time of the bulky materials initially decreases with the increasing flow rate in higher velocity zones, and then gains in value. For the deslagging operations, there is a most favorable lateral velocity gradient configuration of the fluidizing air. The averaged residence time increases positively with the air velocities in the lower gas flow rate zone, due the strengthening recirculation of the particles. With respect to the tracers’physical characters, larger particles of higher density tend to leave the bed quickly. On the other hand, larger particles of lighter materials have a longer stay in the fluidized bed. Additionally, the shape has a prominent effect on the tracers’RTD. Being more spheric and smooth, tracers retain in the bed for a shorter averaged residence time, vice versa. It is found that incombustible materials are relatively separable from the combustible ones and the bed materials, which ensures the complete combustion and desirable deslagging performance.更多还原