乙烯气相聚合工艺研究与技术进展

聚乙烯已成为全球产量最大、用途最广的合成材料。由于生产流程短和设备投资低的优势,流化床气相聚合工艺是聚乙烯的主要生产路线,得到了广泛而深入的研究和开发。本文综述了国内外乙烯气相聚合工艺技术的研发历程,特别是中国在气相法聚乙烯技术方面做出的重要贡献。总结了气相聚合工艺的七种操作模式,即干法气相聚合、超干法气相聚合、气相冷凝聚合、气相超冷凝聚合、气液固云雾聚合、露点聚合以及交替交变聚合。分析了各操作模式所涉及的关键科学问题和工程技术解决方案,例如反应器时空产率模型、冷凝态聚合操作点优化、溶剂的作用、黏性流化颗粒熔融温度预测、流化质量监控与声发射、多流型复合以及一器多用等,并重点介绍了气-液-固云雾聚合流化床反应器中颗粒团聚、气泡运动的新特征,云雾聚合流化床运行的新稳定机制,以及实现聚合物产品结构调控的新途径——乙烯的露点聚合和交替交变聚合技术。指出聚合反应工程学科正从聚焦传统的过程强化,发展到过程强化与产品多样化相互兼顾的变化趋势。     

线性低密度聚乙烯气相聚合工艺模拟与优化

基于流化床反应器内聚合反应的复杂性，导致聚乙烯聚合工艺的能耗通常较大，同时设计生产也存在一定困难。因此本文通过建立聚乙烯气相聚合的反应器模型，对聚合物特性及流动形式进行模拟，通过模型模拟来优化生产工艺，降低能耗。     

气相法聚乙烯工艺冷凝态操作模式的稳定性和动态行为

气相法聚乙烯工艺冷凝态操作模式由于显著提高了循环气移热能力和反应器时空产率,已成为流化床乙烯聚合工艺的主流操作模式。建立了气相法聚乙烯工艺冷凝态操作模式的数学模型,包括流化床反应器模型,多级换热器模型和反应温度、压力以及循环气组成的控制模型。基于此,采用流程模拟方法,计算了系统在反应器温度采用闭环控制时的稳态解;根据系统对小扰动的动态响应特点,定性判断了反应器温度采用开环控制和闭环控制时聚合反应系统的稳定性;考察了系统对1-己烯分压和催化剂进料速率的阶跃响应特性。结果表明,反应器温度采用闭环控制时,聚合反应系统在所考察操作条件下均是稳定的,而采用开环控制时,解曲线被分叉点分割为稳定区域和不稳定区域。反应器温度对1-己烯分压阶跃变化的动态响应表明聚合反应系统存在长、短周期两类振荡,表明冷凝态操作模式下乙烯聚合反应过程是一个多控制回路耦合的复杂过程。     

气相法聚乙烯工艺冷凝态操作模式的稳定性和动态行为

气相法聚乙烯工艺冷凝态操作模式由于显著提高了循环气移热能力和反应器时空产率,已成为流化床乙烯聚合工艺的主流操作模式。建立了气相法聚乙烯工艺冷凝态操作模式的数学模型,包括流化床反应器模型,多级换热器模型和反应温度、压力以及循环气组成的控制模型。基于此,采用流程模拟方法,计算了系统在反应器温度采用闭环控制时的稳态解;根据系统对小扰动的动态响应特点,定性判断了反应器温度采用开环控制和闭环控制时聚合反应系统的稳定性;考察了系统对1-己烯分压和催化剂进料速率的阶跃响应特性。结果表明,反应器温度采用闭环控制时,聚合反应系统在所考察操作条件下均是稳定的,而采用开环控制时,解曲线被分叉点分割为稳定区域和不稳定区域。反应器温度对1-己烯分压阶跃变化的动态响应表明聚合反应系统存在长、短周期两类振荡,表明冷凝态操作模式下乙烯聚合反应过程是一个多控制回路耦合的复杂过程。     

D类颗粒节涌流态化的实验和数值模拟

应用双流体气固两相流模型,对均一粒径分布的Geldart D类颗粒的节涌流态化过程进行了模拟. 研究了流化过程中流化状态转变、床层膨胀比以及压力脉动功率谱,从Froude准数和颗粒动能的角度分析了模拟结果中的流化特性,并与实验数据进行了对比. 模拟结果准确预测了Geldart D类颗粒的节涌流态化特性,对床层膨胀高度和功率谱的预测与实验基本吻合. Froude准数和颗粒动能随时间的周期性脉动较好地反映了流化床内节涌流态化的气泡行为.     

氧化铬催化剂气相法制高密度聚乙烯

前言七十年代初,美国联合碳化物公司创建了世界上第一套气相法高密度聚乙烯工业装置。这种方法,采用高效催化剂,以流化床为反应器,完全革除了后处理过程,聚合生成的聚乙烯,颗粒粗,堆积比重大,从反应器中输送出来后,可以直接造粒或加工成制品。我院于1976年建立了第一套小型流化床装置,围绕气体净化,超微量杂质分析,聚合工艺,高效催化剂,关键单元设     

环流流化床反应器中乙烯气相聚合反应研究

研究开发了１种新型的用于乙烯气相聚合反应的环流流化床反应器。通过对工业Ａ催化剂和实验室自制的ＱＣＰ－０１催化剂的乙烯气相聚合反应评价及聚合物产品的颗粒形态等方面的研究,认为环流流化床反应器具有聚合反应平稳、催化剂的聚合活性高、产品粒径分布均匀等特点,是１种具有良好应用前景的新型反应器。     

气相流化床聚乙烯颗粒粒径分布模型的研究

气相法流化床聚乙烯生产过程中,聚乙烯颗粒粒径分布影响聚合速率、脱挥速率、后处理工序生产成本和最终的聚乙烯物性。为预测流化床中聚乙烯的粒径分布,考察各操作变量和动力学参数对粒径分布的影响,运用了质量衡算的方法,在综合分析聚合过程中聚乙烯颗粒的生长、磨损和扬析等行为的基础上,建立了稳态操作时聚乙烯颗粒粒径分布预测模型。模型计算结果表明,当单一粒径催化剂进料时,随着催化剂粒径或是操作气速的增大,聚乙烯粒径分布向大粒径方向偏移,且分布变宽;相比于操作气速,催化剂粒径作用更为显著。同时发现,床层压力和温度对聚乙烯粒径分布的影响很小。在催化剂粒径有分布的情况下,不同粒径催化剂的质量比不但影响树脂的平均粒径,而且也影响树脂的粒径分布曲线,并在一定的配比下会出现双峰颗粒粒径分布曲线。最后,对某聚乙烯厂家的流化床反应器的聚乙烯颗粒粒径分布进行了模拟计算,计算结果与实验数据基本一致,说明该模型较好反映了气相流化床反应器的反应历程和规律,具有良好的工业应用前景。     

气相流化床聚乙烯颗粒粒径分布的定制

基于气相法聚乙烯流化床反应器颗粒粒径分布的预测[1],提出了颗粒粒径分布定制模型.通过模型的优化计算,可得到催化剂粒径及其分布、操作气速、反应温度、乙烯浓度和丁烯浓度等生产操作参数,由此进行生产可获得具有良好流态化特性的聚乙烯颗粒粒径分布,能为生产具有特定粒径分布的聚乙烯颗粒提供理论指导.模型由工业装置的生产数据分析了计算结果的合理性.最后,以三种粒径分布的聚乙烯颗粒为例讨论了模型的可行性.同时,运用粒子群优化算法求解模型的非线性规划问题,算法具有调整参数少、收敛速度快和全局优化等优点.     

工业流化床聚乙烯树脂性能模型的研究

本文首先分析了聚乙烯树脂熔融指数和密度变化的影响因素,建立了流化床动态过程中聚合物性能指标的递椎关系和工业流化床乙烯气相聚合过程产品质量的预测模型,利用乙烯气相（共）聚合过程中不同工艺条件下树脂的质量数据,对模型参数进行了优化,提出由于影响因素的相互关联,熔融指数和密度的控制应该联合进行;熔融指数和密度的计算值与测量值之间的平均误差分别为8.19%和0.13％。预测精度满足要求,模型可以应用于具体的工业生产过程。     

引入气相对D类颗粒间歇卸料过程特性的影响

通过测量并分析引入气相前后D类颗粒间歇卸料过程中示踪颗粒的流动状态、压力分布和颗粒流率，发现床体内气固两相流动特性不仅随时间变化，还受到气速（正负压差）的影响。据此将卸料过程按时间分为3个阶段：初始蓄压（PS）阶段、稳定卸料（SD）阶段和非满管流（PP）阶段；并给出各卸料阶段不同正、负压差和重力条件下的气固流动特性。在时间较长、流场较为稳定的SD阶段，发现卸料口颗粒阻力是影响颗粒流率的关键参数，通过修正De Jong公式、Beverloo公式，依次建立卸料口颗粒阻力、D类颗粒卸料流率预测模型，与实验值吻合较好，有望为引入气相调控D类颗粒卸料流率的方法提供参考。     

Investigation of Drying Geldart D and B Particles in Different Fluidization Regimes

Drying of nylon (Geldart D) and expanded polystyrene (Geldart B) particles in fixed and fluidized beds were studied experimentally and theoretically. Fluidized bed dryers are sometimes operated at velocities beyond bubbling fluidization to mitigate against de‐fluidization of surface wet particles. It was found that theoretical analysis using three different drying methods could predict the constant‐drying rate at such velocities and also across the entire fluidization regimes (fixed bed, bubbling, slugging and turbulent fluidization) as long as the bed remains completely fluidized. Results also showed that the theoretical predictions were accurate beyond previously reported velocity limits in a laboratory scale dryer. During bubbling fluidization, the cross flow factor method was used effectively to predict the influence of bubble phase on drying rates. In the falling‐rate period, it is demonstrated that the drying behaviour of nylon at different gas velocities can be characterised by a single normalized drying curve.     

EFFECTS OF HYDROCARBON LIQUID FEED IN POLYETHYLENE POLYMERIZATION PROCESS ON PARTICLE SURFACE TEMPERATURE

In the fluidized bed gas phase polymerization of polyethylene (PE), the heat generated by the exothermic polymerization process is dissipated into the gas mixture flowing past the polymer particles. The polymer particle temperature is determined by the extent of convective heat transfer and other mechanisms of heat removal. In addition to the heat removal by convective heat transfer, liquid hydrocarbon (HC) is often injected into the reactor to further remove heat by evaporation but without partaking in the reaction. The effects of adding this liquid HC on the particle surface temperature have been investigated numerically by means of a one-dimensional polar model. Results indicate that the primary mechanism for removal of the heat of polymerization from the particles is by means of convective heat transfer to the bulk gas, which amounts to 99.5% removal of total heat of polymerization. The PE particle temperature rises only by 1–2°C above the surrounding bed gas mixture. The addition of liquid HC to the feed, however, has a pronounced effect on controlling the reactor gas temperature as most of this liquid is evaporated to the gaseous phase before it reaches the polymer particles. To state it clearly, heat of polymerization is transferred from the particles to the reactor bulk gas predominantly by convection, and part of this heat is subsequently absorbed by evaporation of the fresh liquid HC in the feed. Comparison with a detailed computational fluid dynamic (CFD) model of polymerization in a generic gas phase reactor has also been conducted. The results confirm that the particle temperature rise above the reactor gas temperature is consistent with the one-dimensional model. However, local gas temperature variations are present in the reactor due to the unsteady gas-solid hydrodynamics. Hence, there are some zones that are a few degrees hotter/colder than the bulk reactor temperature with corresponding increase/decrease in particle temperature in these zones.     

EFFECTS OF HYDROCARBON LIQUID FEED IN POLYETHYLENE POLYMERIZATION PROCESS ON PARTICLE SURFACE TEMPERATURE

In the fluidized bed gas phase polymerization of polyethylene (PE), the heat generated by the exothermic polymerization process is dissipated into the gas mixture flowing past the polymer particles. The polymer particle temperature is determined by the extent of convective heat transfer and other mechanisms of heat removal. In addition to the heat removal by convective heat transfer, liquid hydrocarbon (HC) is often injected into the reactor to further remove heat by evaporation but without partaking in the reaction. The effects of adding this liquid HC on the particle surface temperature have been investigated numerically by means of a one-dimensional polar model. Results indicate that the primary mechanism for removal of the heat of polymerization from the particles is by means of convective heat transfer to the bulk gas, which amounts to 99.5% removal of total heat of polymerization. The PE particle temperature rises only by 1-2°C above the surrounding bed gas mixture. The addition of liquid HC to the feed, however, has a pronounced effect on controlling the reactor gas temperature as most of this liquid is evaporated to the gaseous phase before it reaches the polymer particles. To state it clearly, heat of polymerization is transferred from the particles to the reactor bulk gas predominantly by convection, and part of this heat is subsequently absorbed by evaporation of the fresh liquid HC in the feed. Comparison with a detailed computational fluid dynamic (CFD) model of polymerization in a generic gas phase reactor has also been conducted. The results confirm that the particle temperature rise above the reactor gas temperature is consistent with the one-dimensional model. However, local gas temperature variations are present in the reactor due to the unsteady gas-solid hydrodynamics. Hence, there are some zones that are a few degrees hotter/colder than the bulk reactor temperature with corresponding increase/decrease in particle temperature in these zones.     

一个基于双流体理论预测三维流化床内流动特性的数学模型(Ⅰ)气固体系散式或鼓泡流态化特性

将考虑拟平衡状态下颗粒与流体相互作用的附加力添加到基于双流体理论动量方程的数学模型中,用于Geldart A类物料散式流态化和B类物料鼓泡/床层塌落特性的三维数值模拟。该模型主要特点是将表征颗粒离散属性的特征长度视为颗粒直径的同一数量级且只需曳力系数一个关联式来封闭控制方程。在商业软件CFX4.4平台上,通过增加用户自定义子程序模拟了长0.2 m、宽0.2 m和高0.5 m流化床内瞬态流动特性。为了检验数学模型的实用性和数值模拟的可靠性,首先考察了两种A类物料在表观气速为u_mf和1.5u_mf下的散式流态化特性,结果展示出床层均匀膨胀的固有属性。随后,考察了扰动对A类物料在网格尺度上的局部空隙率和固体速度分布以及在设备尺度上床层压降的影响,探索了B类物料在网格尺度上鼓泡和床层塌落以及在设备尺度上鼓泡过程中床层压降和塌落过程中平均床层高度和相界面标准偏差的动态特性。上述模拟结果与经典的Geldart理论、前人的实验或模拟结果相吻合,说明该模型可以用来预报三维气固流化床内A类物料散式流态化和B类物料鼓泡及塌落的时空特性。     

Hydrodynamics of Geldart group A particles in gas-solid fluidized beds

Geldart group A particles were fluidized in a 10 cm i.d.×1.8 m high Plexiglas-made bed with ambient air to determine the hydrodynamic properties in a gas-solid fluidized bed. The effects of static bed heights, position of pressure measuring points, differential and absolute pressure fluctuations on the hydrodynamic behavior of a Geldart group A particles in a gas-solid fluidized bed were investigated. The particles used in this study were 80 micrometer FCC powders and 60 micrometer glass beads. The variance of pressure fluctuations was used to find the minimum bubbling velocity. The obtained minimum bubbling velocity was compared with the other methods available in the literature. This method was found to be much easier and had better data reproducibility than the classical visual method or sedimentation method. The variance of pressure fluctuations increased due to the increase of superficial gas velocity and static bed height. The obtained minimum bubbling velocity and pressure fluctuations were found to depend on the measuring position along the axial direction. The effect of measuring position was discussed. Cross-correlation of two pressure signals was used to find the delay time, then the bubble rising velocity.     

Bed density and circulation mass flowrate in a novel annulus-lifted gas–solid air loop reactor

Air loop reactors (ALR) have been widely used as promising and high-efficiency gas–liquid and gas–liquid–solid reactors. Extensive research on ALR has been conducted, but mostly limited to gas–liquid and gas–liquid–solid systems. Work associated with gas–solid systems has been rare and mainly focused on draft tube-lifted spouted bed treating coarse Geldart B, D particles. The present paper proposed a novel gas–solid air-loop reactor treating fine Geldart A particles and operating in a new annulus-lifted mode, with bubbling or turbulent bed upward flow in the annulus in parallel with bubbling bed downward flow in the draft tube. In view of these differences, distinct hydrodynamic behaviour can be anticipated for the gas–solid annulus-lifted air-loop reactor. The influence of operating conditions and geometric configuration on the distribution of bed density is discussed in a cold model annulus-lifted air loop reactor. A mechanistic model for the circulation mass flowrate is established based on an energy balance and resistance analysis. Nearly 50% and 30% of the energy dissipation rate occurs in the bottom and top regions, respectively. With increasing draft tube height, the energy dissipation rate increases in the annulus and draft tube regions, while it decreases in the top and bottom regions. The circulation mass flowrate decreases with increasing draft tube height. Analysis of the distribution of bed density and energy dissipation rate leads to suggestions regarding optimization of the design and axial location of the ring distributor and gap height.     

Why the two-fluid model fails to predict the bed expansion characteristics of Geldart A particles in gas-fluidized beds: A tentative answer

It is well known that two-fluid models (TFMs) can successfully predict the hydrodynamics of Geldart B and D particles. However, up to now, TFM have failed to accurately describe the hydrodynamics of Geldart A particles inside bubbling gas-fluidized beds: Researchers have reported that bed expansions are over-predicted by as much as 70%. In this work we show—for the first time—that TFM can predict the correct bed expansion, without any artificial modifications, provided that a sufficiently fine grid size and small time step is used. This suggests that the previously reported failure of TFM is mainly due to the lack of scale resolution, and that from a modeling point of view there is no fundamental difference between Geldart A particles and Geldart B and D particles.     

Axial segregation in bubbling gas‐fluidized beds with Gaussian and lognormal distributions of Geldart Group B particles

Bubbling, gas‐fluidized bed experiments involving Geldart Group B particles with continuous‐size distributions have been carried out. Sand of various widths of Gaussian or lognormal distributions were completely fluidized, then axial concentration profiles were obtained from frozen‐bed sectioning. Similar to previous works on binary systems, results show that mean particle diameter decreases with increasing bed height, and that wider Gaussian distributions show increased segregation extents. Surprisingly, however, lognormal distributions exhibit a nonmonotonic segregation trend with respect to distribution widths. In addition, the shape of the local‐size distribution is largely preserved with respect to that of the overall distribution. These findings on the nature of local‐size distribution provide experimental confirmation of previous results for granular and gas‐solid simulations. Lastly, an interesting observation is that although monodisperse Geldart Group D particles cannot be completely fluidized, their presence in lognormal distributions investigated still results in complete fluidization of all particles. © 2010 American Institute of Chemical Engineers AIChE J, 2010