【作者】 刘正军；

【导师】 刘明言；

【作者基本信息】 天津大学 ， 制药工程， 2007， 硕士

【摘要】 目前运用大孔树脂吸附分离银杏黄酮主要是在固定床和膨胀床中进行,固定床操作稳定性高、易于控制,但是当固体杂质存在时,易阻塞流道。膨胀床在进行吸附操作时,降低了床层阻力,使颗粒实现了稳定分级,但是其吸附、清洗、洗脱、再生均要进行切换操作,比较繁琐。液固循环流化床(LSCFB)作为一种吸附分离装置,与固定床、膨胀床相比,具有操作连续化、传质速度快、液固间接触良好以及轴向返混程度小的优点。本文旨在探讨LSCFB银杏黄酮吸附作用方式及其影响因素的研究,主要考察了LSCFB的颗粒轴向固含率分布、流体流动状况、液相混合行为以及银杏黄酮大孔树脂吸附的特性及规律。本文首先运用电导法和CCD(电荷耦合器件)摄像法研究了LSCFB下行床和上行床的轴向固含率分布,研究表明,下行床中密相区颗粒轴向固含率在0.35～0.55之间;稀相区在0.1以下,自由区为0。上行床中轴向固含率在0.01～0.04之间,基本上呈均匀分布。然后考察了LSCFB流体流动状况的影响因素,结果表明,在同样的表观液速下,颗粒粒径越大,密相区高度越小。主水流和辅助水流单独变化时,颗粒循环速度先增大后减小。以KCl为示踪剂,采用电导率仪分别研究了LSCFB下行床和上行床的停留时间分布(RTD)。实验得出下行床RTD无因次方差在0.1左右,液相返混系数Daxl在1.85×10-6～1.81×10-5之间,Bo准数在10.4～16.9之间,床层比较稳定,接近平推流。上行床无因次方差在0.2～0.3之间,Daxl在1.48×10-3～2.96×10-3之间,Bo准数在5.65～7.44之间,说明上行床由于滞留区、循环流等原因存在一定程度的返混。对LSCFB吸附影响因素进行了考察,其结果表明,下行床表观液速越小、颗粒装柱量越大、颗粒循环速度越大对LSCFB吸附越有利。从LSCFB四因素三水平吸附正交实验得出最佳的实验条件为:颗粒粒径范围为0.45～0.6mm、装柱高度为50cm、颗粒循环速度为0.285kg/m2·s、下行床表观液速为0.15mm/s。最后进行了LSCFB上行床脱附模型和下行床吸附模型的研究,其结果表明,上行床模型能够预测床体的脱附情况,实验值与计算值之间吻合的比较好;下行床模型能够预测床体的吸附情况,由于密相区大孔树脂颗粒流化时会有少量的返混,所以实验值和模型计算值之间会有一定的偏差。更多还原

【Abstract】 Separating gingko flavones with macroporous resins is now mainly operated in fixed bed or expanded bed. It has higher stability and is easy to be controlled of the fixed bed operation. When solid impurity existing in the bed, it is easy to be jammed. Expanded bed operation can decrease bed resistance. However, it has a disadvantage, which is that the adsorption, washing, elution and regeneration must be in switching operation. Comparing with fixed bed and expanded bed, liquid-solid circulating fluidized bed (LSCFB) has some advantages. For example, it is continuous in operation, fast in mass transfer, well in the contacting between liquid and solid and small in the axial backmixing extent. Main research content in this paper is as follows: particle axial holdup distribution, residence time distribution (RTD), liquid mixing behavior and the performance of absorption and separation of the macroporous resin for ginkgo Flavonoids.The axial solid holdup of the downcomer and the riser in the LSCFB had been measured firstly with conductivity meter and CCD. Experimental results show that the particle axial solid holdup of the dense area in the downcomer is between 0.35 and 0.55,and it is below 0.1 in the dilute area; The axial solid holdup in the riser is between 0.01 and 0.04 and the distribution is uniform. Investigation on the factors influencing the LSCFB fluid flow indicates that the larger the particle diameter is, the lower the dense area height will be. During the respective change of main fluid and assistant fluid, the particle circulating rate (Gs) increased first, and then decreased.In the residence time distribution (RTD) experiment of the downcomer, the dimensionless variance is about 0.1, the liquid backmixing coefficient is between 1.85×10-6 and 1.81×10-5, and the Bo value is between 10.4 and 16.9, so the bed is very stable, the flow pattern was similar to that of plug flow reactor (PFR). And in the RTD of the riser, the dimensionless variance is between 0.2 and 0.3, the liquid backmixing coefficient is between 1.48×10-3 and 2.96×10-3, the Bo value is between 5.65 and 7.44, so the results indicate that the riser is backmixing in an extent because of the existence of stagnant area and circulating flow.Investigations on the factors influencing the LSCFB adsorption indicate that it is favorable for LSCFB adsorption of a lower Uld, a higher Gs and a bigger charge. The results of LSCFB adsorption with the orthogonal test of four factors at three different levels gave the optimum experimental conditions: the particle size with 0.45～0.6mm diameter, the particle column height with 50 centimeter, Gs with 0.285 kg/m2·s and Uld with 0.15 mm/s.Finally, the riser desorption model and downcomer adsorption model were carried out. The results indicate that the riser desorption model can predict desorption process and the experimental value is inosculate to the calculated value; the downcomer adsorption model can predict adsorption process and the experimental value has some deviations from the calculated value because there will be a little backmixing when the macroporous resins fluidized in dense area.更多还原