基于小波包分析和K-OPLS的集成方法在颗粒粒径分布检测中的应用

颗粒粒径分布的实时在线检测对于调控气固流化床中颗粒的流化特性具有重要意义。基于混料均匀设计法安排实验,以声发射(AE)技术为检测手段,结合小波包分析,提出采用K-OPLS方法构建颗粒粒径分布的声信号预测模型,定量描述小波包能量特征与颗粒粒径分布的非线性变化规律。实验结果显示,留一交叉验证法的均方根误差(RMSE)仅为0.063,表明基于K-OPLS的颗粒粒径分布声信号预测模型能准确测量气固流化床中颗粒的粒径分布,具有良好的工业应用前景。     

Non‐intrusive characterization of particle size changes in fluidized beds using recurrence plots

An on‐line method is developed for monitoring of mean particle size in fluidized beds using pressure fluctuations (PFs) and acoustic emissions (AE) signal by recurrence plot (RP) and recurrence quantification analysis (RQA). PFs and AE signals of a lab‐scale fluidized bed were measured simultaneously at various superficial gas velocities and mean particle sizes. Although the AE signals are often very complicated due to many different acoustic sources in the bed, applying RP analyses showed that small changes in mean particle size can be detected by visual comparison of AE‐RP structures, while this cannot be distinguished by graphical RP analysis of PFs. Moreover, the hydrodynamics of the bed was inspected through RQA analysis of both signals. For this purpose, recurrence rate, determinism, laminarity, average length of diagonal and vertical lines were extracted from RPs showing the effect of an increase in the mean particle size. © 2016 American Institute of Chemical Engineers AIChE J, 62: 3547–3561, 2016     

Measuring Average Particle Size for Fluidized Bed Reactors by Employing Acoustic Emission Signals and Neural Networks

Average particle size is one of the key parameters of fluidized bed reactors. A novel method to measure the average particle size in fluidized bed reactors by acoustic emission (AE) signal is proposed. The measurement of the average particle size by AE is superior to other methods since it is inherently safe to use and it is also a non‐invasive method. AE signals originating from different particle sizes were 8‐level decomposed by a Daubechies wavelet of order 3 after being denoised with a sym8 wavelet filter combined with the rigrsure threshold method. Principal component analysis (PCA) was applied to overcome the complex collinearity and reduce the number of input variables of the neural network. A feed‐forward back propagation neural network with two hidden layers was used to predict the average particle size according to the principal components. The results show that this soft‐measuring model is suitable for measuring the average particle size in the fluidized bed reactors by online AE. It achieved high accuracy when applied to a model‐scale fluidized bed.     

Characterization of flow regime transition and particle motion using acoustic emission measurement in a gas‐solid fluidized bed

Particle motion is a major determinant of the dynamical performance of a fluidized bed. It plays an important role in determining and optimizing the complex correlation of fluidization condition between particle‐particle and particle‐environment in a system. A passive acoustic emission (AE) technique is applied to monitor, characterize, and control the fluidization condition of polyethylene particles in a gas‐solid fluidized bed. Experimental results show that AE signals are very sensitive to the particle movements by analyzing energy distribution, which can help to understand the status of the system. The AE energy temporal analysis is further used to identify the transition of flow regimes. Moreover, the activity of particle motion can be quantitatively determined by using a combination of granular temperature and AE spatial energy analysis. This work provides valuable insights into the dynamic behavior of particles in a gas‐solid fluidized bed based on AE technique. © 2009 American Institute of Chemical Engineers AIChE J, 2010     

Suspension of solid particles in multi‐impeller three‐phase stirred tank reactors

Solids suspension characteristics in gas—liquid–solid three‐phase stirred tanks with multi‐impellers were experimentally examined. Minimum impeller speeds for ultimately homogeneous solid suspension have been measured stirred tank reactors. Three impellers were installed: two four‐pitched blade downflow disk turbines and one Pfaudler type impeller chosen to provide good gas dispersion and to accomplish off‐bottom suspension of solid particles, respectively. Gas dispersion causes an increase in particle sedimentation associated with a decrease in power consumption and as a result, minimum impeller speeds for ultimately homogeneous solid suspension increase with increasing gas flow rates. A correlation was developed to predict minimum impeller speeds for ultimately homogeneous solid suspension. The proposed correlation, which agrees satisfactorily with the experimental results, is expected to be useful in design and scale‐up.     

利用声发射技术测量搅拌釜的淤浆悬浮高度

根据颗粒运动碰撞搅拌釜壁面产生声波的机理,结合声信号的频谱分析、小波分解和R/S分析,获得了代表颗粒运动的特征信号频段（d1、d2频段）。同时,基于声波特征信号频段能量沿搅拌釜轴向的规律性变化,提出了声波法测量搅拌釜淤浆悬浮高度的判据,即当声波特征信号频段能量或声波特征信号频段能量比出现阶跃性变化时的高度为淤浆悬浮高度。以水-玻璃珠体系为例,研究发现,无论是盘式涡轮还是桨式叶轮的搅拌桨,基于声信号测定淤浆悬浮高度的判据都能较好地得到验证,与目测法相比,其平均相对误差小于10 %,具有较高的精度。由此,获得了一种简单快捷、灵敏准确、非侵入式的搅拌釜淤浆悬浮高度测量技术,能够实现淤浆悬浮高度的实时监控。     

A CFD‐PBM‐PMLM integrated model for the gas–solid flow fields in fluidized bed polymerization reactors

Although the use of computational fluid dynamics (CFD) model coupled with population balance (CFD‐PBM) is becoming a common approach for simulating gas–solid flows in polydisperse fluidized bed polymerization reactors, a number of issues still remain. One major issue is the absence of modeling the growth of a single polymeric particle. In this work a polymeric multilayer model (PMLM) was applied to describe the growth of a single particle under the intraparticle transfer limitations. The PMLM was solved together with a PBM (i.e. PBM‐PMLM) to predict the dynamic evolution of particle size distribution (PSD). In addition, a CFD model based on the Eulerian‐Eulerian two‐fluid model, coupled with PBM‐PMLM (CFD‐PBM‐PMLM), has been implemented to describe the gas–solid flow field in fluidized bed polymerization reactors. The CFD‐PBM‐PMLM model has been validated by comparing simulation results with some classical experimental data. Five cases including fluid dynamics coupled purely continuous PSD, pure particle growth, pure particle aggregation, pure particle breakage, and flow dynamics coupled with all the above factors were carried out to examine the model. The results showed that the CFD‐PBM‐PMLM model describes well the behavior of the gas–solid flow fields in polydisperse fluidized bed polymerization reactors. The results also showed that the intraparticle mass transfer limitation is an important factor in affecting the reactor flow fields. © 2011 American Institute of Chemical Engineers AIChE J, 58: 1717–1732, 2012     

流化床声发射机理及其在故障诊断中的应用

引　言流化床反应器能否长期稳定运行是一个十分严重的挑战 ,如果不彻底掌握流化质量的控制规律 ,对诸如强放热聚合反应过程 ,床层中流化粒子易于聚集、结块 ,甚至导致流化床崩溃而被迫停车 .因此 ,流化床故障诊断一直是工业界和学术界共同面临的难题 .流化系统一旦出现故障就会有反映流动规律本质的特征物理量出现较大变化以至突变 ,只是有些物理量对这些变化存在空间或时间上的不敏感性 .如压力信号就存在时间上的不敏感性 ,即往往压力信号出现显著变化时 ,床层流化质量已无法通过改变操作条件来改善 ;而光纤测量则存在空间上的不敏感性 ,…     

气固流化床中声发生机理及在工业装置中的应用

利用声测量技术,结合频谱分析,建立了颗粒碰撞的声波频率模型,可定量描述声波主频随颗粒粒径、弹性模量和密度的变化规律.通过改变流化颗粒的粒径、弹性模量参数和密度,发现声波主频与频率模型计算值之间的最大偏差为8.3%,说明声波主频可以代表颗粒在壁面的碰撞频率.讨论了热态和冷态条件下声波主频之间变化规律,通过对弹性模量参数的校正,声测量技术可以用于预测工业装置中物料的平均粒径变化,并将该模型应用于线性低密度聚乙烯、高密度聚乙烯和双峰聚乙烯工业生产装置中的平均粒径测量,发现与传统的取样筛分方法所得测量结果十分接近.同时,发现当系统产生聚合物颗粒结块时,声波主频将急剧降低,声波频谱的能量分布将明显集中增大,这可作为判断流化床稳定运行的一个判据.     

声波测量在搅拌釜中固体颗粒临界悬浮转速测定的应用

基于声发射测量技术,结合声信号的频谱分析、小波分解和R/S分析,获得了代表颗粒运动的特征信号频段（d₁、d₂频段）。同时,根据声波特征信号频段能量随搅拌转速的规律性变化,以固体颗粒碰撞壁面产生信号高频区域的声能量分率值为特征参数,提出了搅拌釜临界悬浮转速的测量判据,即声能量分率快速减少并开始趋于稳定时所对应的搅拌转速为临界悬浮转速。以水-玻璃珠体系为例,与目测法相比,声波法测量值的平均相对误差为3.51%,具有较高的精度。利用经典的Zweitering临界悬浮转速计算公式对声波法测得的实验数据进行拟合,计算值与测量值之间的平均相对误差为3.17%,表明公式对于临界悬浮转速的计算具有较高的准确性。由此获得了一种快速、准确、安全的搅拌釜反应器临界悬浮转速测量技术,有利于工业生产流程的优化和控制。     

Experimental and computational study on the bubble behavior in a 3-D fluidized bed

The results from a two-fluid Eulerian–Eulerian three-dimensional (3-D) simulation of a cylindrical bed, filled with Geldart-B particles and fluidized with air in the bubbling regime, are compared with experimental data obtained from pressure and optical probe measurements in a real bed of similar dimensions and operative conditions. The main objectives of this comparison are to test the validity of the simulation results and to characterize the bubble behavior and bed dynamics. The fluidized bed is 0.193 m internal diameter and 0.8 m height, and it is filled with silica sand particles, reaching a settle height of 0.22 m. A frequency domain analysis of absolute and differential pressure signals in both the measured and the simulated cases shows that the same principal phenomena are reproduced with similar distributions of peak frequencies in the power spectral density (PSD) and width of the spectrum. The local dynamic behavior is also studied in the present work by means of the PSD of the simulated particle fraction and the PSD of the measured optical signal, which reveals as well good agreement between both the spectra. This work also presents, for the first time, comparative results of the measured and the simulated bubble size and velocity in a fully 3-D bed configuration. The values of bubble pierced length and velocity retrieved from the experimental optical signals and from the simulated particle fraction compare fairly well in different radial and axial positions. Very similar values are obtained when these bubble parameters are deduced from either simulated pressure signals or simulated particle volume fraction. In addition, applying the maximum entropy method technique, bubble size probability density functions are also calculated. All these results indicate that the two-fluid model is able to reproduce the essential dynamics and interaction between bubbles and dense phase in the 3-D bed studied.     

Hysteresis in cloud heights during solid suspension in stirred tank reactor: Experiments and CFD simulations

Solid suspension in stirred tank reactor is widely used in process industries for catalytic reactions, dissolution of solids, crystallization, and so on. Suspension quality is a key issue in design and operation of stirred reactor and its determination is not straight forward. Cloud height measurements of solid suspension provide a relatively simple way to quantify quality of suspension. In this work, experiments were carried out to quantify variation of cloud heights with impeller speed and particle characteristics. These measurements were carried out using visual observations, image analysis, and ultrasound velocity profiler techniques. The obtained data demonstrated the existence of hysteresis in cloud heights with respect to impeller speed. Apart from possible applications in reducing power required for achieving desired solid suspension quality, the existence of hysteresis also provides a new way to evaluate computational fluid dynamics (CFD) simulations of solid–liquid flows in stirred vessels. An attempt was made to capture observed hysteresis in cloud heights in CFD simulations. The simulated results were compared with the experimental data. The presented models and results (experimental and computational) will be useful for simulating complex solid–liquid flows in stirred reactors. © 2010 American Institute of Chemical Engineers AIChE J, 2010     

Particle suspension in top-covered unbaffled tanks

Unbaffled stirred tanks are seldom employed in the process industry as they are considered poorer mixers than baffled vessels. However, they may be expected to provide significant advantages in a wide range of applications (e.g. crystallization, food and pharmaceutical processes, etc.), where the presence of baffles is often undesirable. In the present work solid–liquid suspension in an unbaffled stirred tank is investigated. The tank was equipped with a top-cover in order to avoid vortex formation. A novel experimental method (the “steady cone radius method”, SCRM) is proposed to determine experimentally the minimum impeller speed at which solids are completely suspended. Experimental N_js and power consumption data are provided over fairly wide ranges for particle size, density and concentration. Dependence of N_js on particle density and concentration is similar to that observed in baffled tank. Conversely, a negligible dependence of particle diameter on N_js is observed in the unbaffled tank, a difference from baffled vessels with important practical implications.Finally, the mechanical power required to achieve complete suspension in unbaffled tanks is shown to be much smaller than in baffled vessels. This, in conjunction with the previously ascertained excellent particle-fluid mass-transfer promotion, could make unbaffled tanks a best choice for many solid–liquid operations, where mass transfer is the main limiting factor.     

混合-反应结晶过程(Ⅰ)实验

实验研究了半连续结晶器中混合特性与化学特性对反应结晶过程的影响,实验发现,混合特性对产物粒度分布的影响是非单调的,存在转捩点(criticl point),在该点平均粒径或方差为极小。分析表明,转捩点的出现是微观混合与宏观混合共同作用达到平衡的结果。     

The Use of Positron Emission Particle Tracking in the Study of Multiphase Stirred Tank Reactor Hydrodynamics

This paper describes the use of positron emission particle Tracking (PEPT) in the analysis of local particle and fluid velocities in solid‐liquid stirred tank reactors agitated with a Rushton turbine and an upward‐pumping pitched blade turbine. PEPT captures the full three‐dimensional characteristics of hydrodynamics and mixing in stirred vessels, allowing the analysis of the two‐phase flow fields. Furthermore, by comparing the liquid and particle velocities, the spatial and temporal variation of the relative particle‐liquid velocity can be estimated. Such information reveals considerable heterogeneity in the vessel and facilitates the evaluation of impeller design, particularly with the aim of minimizing mass transfer limitations.     

Liquid‐Solid Mass Transfer Behavior of a Stirred‐Tank Reactor with a Fixed Bed at Its Bottom

The mass transfer behavior of a new batch stirred tank with a fixed bed of Raschig rings at the bottom was studied using diffusion‐controlled dissolution of copper in acidified dichromate. Variables studied, amongst others, were the impeller rotation speed, Raschig ring diameter, fixed‐bed height, and impeller geometry. The rate of mass transfer from the fixed bed to the solution increased with increasing impeller rotation speed, decreasing particle size, and decreasing bed height. The axial‐flow turbine is more efficient in increasing the rate of mass transfer than the radial‐flow turbine. The presented reactor is especially useful for conducting diffusion‐controlled liquid‐solid catalytic reactions involving reactants that need to be dispersed first, such as sparingly soluble solid particles.     

Efficient production of uniform nanometer‐sized polymer vesicles in stirred‐tank reactors

Polymer vesicles, so‐called polymersomes, gain more and more attention as potential carriers for medical and biotechnological applications. To put the production of these nanocompartments into action at an industrial scale, an efficient and scalable process has to be established. Moreover, being able to control the resulting particle size distribution (PSD) is vital. In this work, the amphiphilic triblock copolymer poly(2‐methyloxazoline)₁₅–poly(dimethylsiloxane)₆₈–poly(2‐methyloxazoline)₁₅ is formed into polymersomes in miniaturized stirred‐tank reactors. Varying flow conditions have a huge impact on the resulting PSD. Dynamic light scattering measurements show that driving a S‐shaped stirrer at 4000 rpm in unbaffled reactors leads to a monomodal PSD with a low polydispersity index (PDI<0.2). Vesicles with a mean diameter of 200 nm are achieved within less than 1 h in a single production step. The robustness of the established process is shown by producing uniform polymersomes at different temperatures and varying pH and buffer molarities. © 2015 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2016 , 133, 43274.     

Experimental profiles of lateral mixing of feed particles in a three‐dimensional fluidized bed

Solids mixing data of high quality is one of the most crucial steps for quantitative studies, but it is a difficult task to obtain in a fluidized bed especially with a 3D configuration. Therefore a novel sampling technique is developed with bed collapse method, for measuring lateral mixing of feed particles in a 3D fluidized bed. The sampling tool is designed using a “bottom‐to‐top sampling” idea. Its development, configuration and measurement repetition are discussed in detail. The effects of mixing time, fluidizing gas velocity, and particle size of bed material on the tracer distribution are investigated. A quantitative comparison of lateral dispersion coefficient shows that our results agree fairly well with measurements and predictions of correlations for lab‐scale fluidized systems in previous studies. The presented 2D profiles of the lateral mixing can be used to validate fundamental solids mixing models or verifying convenient measurement techniques. © 2010 American Institute of Chemical Engineers AIChE J, 2011