【作者】 侯琳熙；

【导师】 阳永荣；

【作者基本信息】 浙江大学 ， 化学工程， 2005， 博士

【摘要】 气固流化床聚合反应器中流化床层的传递特性参数、反应效果、以及聚合物产品均受到流化床内颗粒性质及其运动状态的影响,而颗粒的性质又足随着反应的变化而变化的。因此,研究流化床聚合反应器的流体力学行为是化学反应工程研究领域的前沿,极富有挑战性。本研究论文以声波的测量和分析为研究手段,从多尺度的角度对流化床聚合反应器中的颗粒生长及其粒径分布、结块状况、起始流化速度、流动模式和料位等参数的变化进行了研究。不仅建立了新的测量和分析方法,而且发现了新的实验现象,提出了新的概念和理论,研究结果具有重要的理论意义,对于工业反应器的安全生产、优化操作和产品开发也具有重要的实用价值。 颗粒流态化时会产生各种各样的信号,它们往往是时空多尺度结构的综合反映。由于各尺度信号相互混杂,妨碍了对过程的了解。虽然传统的流化床参数测量手段,如压力、激光、温度、射线衰减、超声、电导等技术被广泛地应用,但测量信号中所包含的丰富内涵还远未被人们所理解,更何况许多测量手段还由于对测量环境的要求苛刻,对流场的破坏以及对人体的危害,而受到限制。受古人“听音辨琴”的启发,本文将具有检测灵敏、绿色环保、非侵入性和实时监测等特点的声波测量技术应用于流化床各参数的测量。进一步,通过小波分析和R/S分形分析对声波测量信号进行多尺度的分解,论文主要开展了以下四个方面的研究工作: 1.研究比较了声波信号和压力脉动信号的多尺度结构特征。根据各小波尺度下波动信号的Hurst指数变化规律(均小于0.5、均大于0.5或两者都有的),建立了声波信号的尺度划分的标准,认为可将声波信号划分为微尺度(1-5小波尺度)、介尺度(6-7小波)和宏尺度(8-10小波尺度),分别反映颗粒行为、气泡和颗粒团聚物行为、以及平均流动行为等。与压力脉动信号多尺度结构的比较表明,声波信号主要反映微尺度的颗粒信息,而压力信号主要反映介尺度的气泡信息。至于处于介尺度的颗粒团聚物的信号,由于信号强度较弱,其变化只能在声波的介尺度频段得到较明显的体现。 2.研究了流化床聚合反应器在微尺度的特性。在直径分别为100mm和150mm流化床冷模实验装置(带加热装置)中,以及φ1000的HDPE中试流化床聚合反应器、φ3500的LLDPE、φ3500的HDPE和φ5000的Bi-mode PE工业流化床聚合反应器装置中,测量了颗粒流态化产生的声波信号。通过对声波微尺度信更多还原

【Abstract】 Hydrodynamics of fluidization, behaviour of reaction and properties of resin produced in gas phase polymerization fluidized bed reactor are strongly affected by the movement of the particles in fluidized bed. Therefore, it is the frontal and full of challenge in chemical engineering research. The thesis focuses on the growth of particles and its evolvement of particle size distribution (PSD), including occurrence of agglomeration, definition of minimum fluidization velocity, flow pattern, bed height and so on in view of the multi-scale approach based on the acoustic emission (AE) measurement. Not only the novel measurement and analytical method were established, but also discover many new phenomena and then theory established. All the results have the important significant academically and commercially.The signals emitted from the fluidized particles were the synthesis of the multi-scale information, and they were so immingle that could not easily to be understood. Traditional measurements had been devised to measure the signals, but they may not be used reliably on an industrial scale under hostile conditions of atmosphere, temperature and pressure. A novel and non-invasive AE measurement was used in this study. Further, the AE signals would be decomposed in multi-scale based on wavelet analysis and R/S analysis. The creative works can be summarized as following:(1) Compared AE signals of the multi-scale structure character with pressure signals. The criterion of dividing the information to three characteristic scales was established by Hurst analysis of the wavelet decomposed signals: micro-scale signals with all Hurst exponents less than 0.5 (lst-5th wavelet decomposing scales); macro-scale signals with all Hurst exponents more than 0.5 (6th-7th wavelet decomposing scales); meso-scale with two characteristic Hurst exponents (8th-10th wavelet decomposing scales), and the three scales reflected the action of particles, average dynamics and bubbles and agglomeration respectively. Energy profiles of the three scales components confirmed that the AE signals and pressure signals mainly reflect apart the micro-scale and meso-scale.(2) The character of the polymerization fluidized bed reactor in micro-scale was studied. The AE signals came from not only the experiment apparatus with the diameter of 100mm and 150mm, the pilot reactor with the diameter of 420mm, butalso the industrial units with diameter of 3500 mm (LLDPE), 3500 mm (HDPE) and 5000 mm (Bimodal PE). The signal main-frequency measurement Model, Hou-Yang Model and the new criterion of Umf were based.(a) A model of signal main-frequency measurementi. 1 2 v5J =- = ■r 5.7was presented for the gas-solid fluidized bed. It was found the signals main-frequency decreased with the increasing of the particles’ size, temperature and density and with the decreasing of gas velocity. The experiment in the experimental fluidized bed showed that a satisfactory agreement was obtained between calculation and experimental data with respect of the main-frequency, and the average error all less than 8.3%. When this model was applied to the industrial fluidized bed of LLDPE, HDPE and bimodal polyethylene, the error between calculation and experimental data were less than 8.8% showed the model is effective.(b) Hou-Yang Model of particle size distribution (PSD)7=1 7 = 1was devoted to combine Hou-Yang Model with AE signals measurement and the wavelet analysis to determine the PSD in the fluidized bed with three different polyethylene powders of LLDPE, HDPE and Bi-mode PE respectively. It was found the corresponding deviations between Hou-Yang method and sieves method was less than 8.8%. When using between Hou-Yang Model to predict PSD in the industrial fluidized bed with the diameter of 3500mm of LLDPE, 3500mm of HDPE and 4000mm of Bimodal HDPE, the average deviations Hou-Yang method and sieves method was less than 15.8%. Farther study found Hou-Yang Model could be applied to some measurement that there were many wavelet scales in multi-scale, and gas velocity beyond 3Umf.(c) A new method to define minimum fluidization velocity in fluidized beds was put forward with AE measurement and wavelet packet analysis. When using this method to define the Umr with the particles of ordinary PE (LLDPE) and especial PE (Bimodal PE) under normal temperature conditions and LLDPE particles on the temperature increasing process, it showed that a satisfactory agreement was obtained between this new method and traditional ways. The traditional ways could not estimate Umf accurately with the Bimodal PE particles on the temperature increasingprocess, but the new method was effective. It was found Umf of LLDPE particles could not changed with temperature and Umf of Bimodal PE particles increased with the temperature increasing. Further, Using the new method give the result that Umr arranged from low to high in the same average particle size was FCC-type, bipeak-type, flat PSD, binary PSD, Gaussian-type and narrow cut PSD.(3) The character of the polymerization fluidized bed reactor in meso-scale was studied. It was found in the experiment apparatus with the diameter of 100mm and 150mm and the pilot reactor with the diameter of 420mm that the change of PSD affected the energy percentage of d’-d5 in multi-scale, and the agglomeration size increasing brought on the energy percentage of d6-d7 increased in turn. A formula to measure the agglomeration size was:S = a(Edb-E°d6) + b(Ed1 -E°d1) + c(Ed&-￡°8).and 3=580, b= 1920, c=6800. the error between calculation and experimental data were less than 11.2%. On the other hand, the result showed there are reversible and non-reversible phases in the agglomerating process.(4) Investigate the AE signals with the different height in fluidized beds by using energy analysis, mean square deviation and complexity analysis. The flow pattern in the experimental apparatus and industrial unit (HDPE) was worked over. It was found that there are two solid circulations cells with smaller size and solid circulation with larger size (main circulation zone), and a stagnant zone located between the two circulations cells. On the other hand, the bed height could be measured in experimental apparatus and industrial unit (HDPE), and the error between the prediction by AE measurement and actual height was less than 2%. In a word, the character of macro-scale in fluidized beds could be measured availably by AE sensor array.更多还原