Multi-Scale Multiphase Modeling of Transport Phenomena in Spray-Drying Processes

Spray drying is an extensively used technology in process engineering for receiving small particles by rapid moisture evaporation from a spray of droplets. This contribution summarizes achievements and results of the comprehensive scientific research on multi-scale multiphase modeling of transport phenomena in spray-drying processes undertaken by our research group: (1) study of particle formation on the scale of an individual droplet; (2) modeling and simulation of droplet–droplet and particle–particle collisions in a spray; (3) study of gas-spray mixing; (4) 2D and 3D study of spray drying by an innovative multi-scale simulation tool coupled to a commercial CFD software. The proposed multi-scale multiphase model of transport phenomena in a spray-drying process has been developed based on a thorough analysis of previously published experimental and theoretical works. The content of this paper will be useful for both academia and industry; e.g., pharmaceutical, biotechnology, chemical, ceramics, materials, nutrition, and other applications of spray drying.     

Experimental Study and CFD Modeling of Wall Deposition in a Spray Dryer

The wall deposition phenomenon in a pilot-scale spray dryer was investigated based on mathematical modeling and experimental trials. For this purpose, the governing equations were obtained and solved numerically by applying a mathematical modeling technique and an open-source computational fluid dynamics (CFD) software. The wall deposition, velocity distribution of the existing phases, and droplet trajectory in the drying chamber were determined. The effect of the operating parameters including the feed flow rate, inlet concentration of dissolved solid, and initial droplet diameter on the air flow pattern, droplet trajectory, and wall deposition was investigated. Through the experiments, the wall deposition of powder product in different positions of the drying chamber was measured. In modeling part of this study, we attempted to determine the effect of particle diameter on the percentage of wall deposition and the position where it occurred.

The model results obtained for wall deposition were compared with collected experimental data and good agreement was observed.     

Coupled CFD-PBE Simulation of Nucleation in Fluidized Bed Spray Granulation

Fluidized bed spray granulation is used to produce porous granular particles from suspensions, solutions, and melts. It is the general aim of our work to provide a physics-based simulation tool for this process. For this the process-relevant mechanisms such as droplet deposition, dust integration, and drying are implemented using computational fluid dynamics (CFD) to generate growth kinetics of the granules and of the dust particles. The latter gives rise to the nucleation rate. This kinetics is applied to simulate the development of the particle size distribution of granules for a continuously operated fluidized bed spray granulation process using population balance equations of the granules.     

Reduction of a model for single droplet drying and application to CFD of skim milk spray drying

In this work, a novel methodology for the development of a high-accuracy computational fluid dynamics (CFD) model for the spray-drying process is described. Starting point is an own spatially resolving model of droplet/particle drying, which was developed and validated on the basis of a series of single droplet drying (SDD) experiments. This sophisticated model is transformed to a much simpler version: the characteristic drying curve approach, after running the full SDD model in a wide range of operating conditions. Then, the obtained reduced model is implemented into the CFD solver. The CFD spray-drying model takes into account the hydrodynamics of the continuous phase, particle drying kinetics, changes in the particle diameter, and the heat loss from the drying chamber to the environment. Validation of the entire procedure is provided by data obtained from drying experiments performed in a co-current laboratory spray tower. High accuracy of the developed CFD model of skim milk spray drying has been found for both phases, for the mean outlet temperature of the continuous phase (air) and for the change in average particle moisture content along the spray tower (discrete phase).     

Conditions for Accurate CFD Modeling of Spray-Drying Process

The paper presents concluding results of extensive experimental and theoretical research on confident CFD modeling of spray drying. An earlier developed experimental method to determine spray-drying kinetics in a lab scale allowed us to find a critical material moisture content and to determine generalized spray-drying curves. The generalized drying curves, identical in shape in the laboratory and pilot plant units, were used in the CFD model of spray drying process. Extensive simulations for spray drying of 10, 30, and 50% of initial solid content of maltodextrin proved high accuracy of the predictions of discrete (particle size distribution, particle moisture content, particle velocity, spray temperature) and continuous-phase parameters (gas temperature and humidity). Maximum error of the predictions of discrete-phase parameters was on the level of 20%, which is probably close to the current capacity of the CFD technique for modeling of spray-drying process. Comparison of experimental measurements and theoretical results shows that incorporation of realistic spray-drying kinetics into the CFD model and correct definition of initial drying and atomization parameters enable reliable simulations of spray-drying process.     

甘露醇喷雾干燥过程中液滴粒度分布变化的群体粒数衡算模拟和实验研究

利用群体粒数衡算（population balance，PB）计算机模拟和实验研究了甘露醇水溶液的喷雾干燥过程中液滴的粒度分布的变化规律。液滴干燥过程中的颗粒粒度的萎缩速率，在群体粒数衡算模型中描述为液滴的逆（或负）生长项，通过单个液滴反应动力学方法（reaction engineering approach，REA）获得。基于单个液滴干燥的反应工程方法模型REA和群体粒数衡算模型PB集成建立了PBREA模型。PBREA 模型的求解是通过高分辨率数值方法。本文模拟研究了不同工况下，不同粒径液滴的干燥时间、液滴平均含湿量以及液滴粒度分布随时间的变化。结果显示，液滴粒径越大，干燥时间越长，模型预测的颗粒平均粒径为实验值的1.0~1.5倍，粒度分布跨度是实验值的0.61~0.89倍。模拟误差主要来源于液滴及颗粒粒径分布统计精度、单个静止液滴与群体运动液滴干燥的差异、热导率及扩散系数是经验值3个方面。在使用Buchi 290 小型喷雾干燥仪进行的实验中，使用了图像采集和分析方法得到了液滴及颗粒的数密度分布，并和模拟结果做了对比。结果表明该模型可以有效地预测喷雾干燥过程中干燥颗粒的平均粒度及分布跨度。     

Droplet Size Distribution Obtained by Atomization with Two‐Fluid Nozzles in a Spray Dryer

The possibility of predicting the droplet size distribution from the particle size distribution was investigated. For that purpose, suspensions of different types of materials were dried in a laboratory‐scale spray drier. Drying of suspensions was performed with different sizes of two‐fluid nozzles. Droplet size distribution was evaluated from the data obtained for spray drying of bismuth molybdate suspension. The method was validated experimentally with other tested materials. Investigated systems involve processes of drying, crystallization, and coating. The proposed methodology can be applied when nonagglomerated particles, spherical particles, or spherical agglomerates were obtained by spray drying.     

The influence of airflow on fuel spray characteristics from a slit injector

Optimization of fuel spray, airflow, and their interaction with the cylinder and piston wall is crucial to achieve stable combustion of stratified charge with minimum emissions in direct-injection spark-ignition (DISI) engines. In this study, the interaction between air and fuel spray from slit injector was investigated in a steady airflow system generated by a wind tunnel under atmospheric conditions. Both Mie scattering images and phase Doppler anemometry (PDA) measurements of the spray were analyzed for different air velocities. Three-dimensional computational fluid dynamics (3-D CFD) have been employed to further explain the mutual interaction between air and spray. It was found that increasing the airflow velocity across the spray results in a significant change in the bottom part of the spray while a slight change was observed close to the nozzle exit. The variation in spray geometry, which is mainly attributed to an aerodynamic effect and to extracted droplets from the main spray by the airflow effect, was evaluated and presented for different air velocities. The spray droplet size redistribution within the spray plume was investigated, and regions with smaller and larger droplets were identified and discussed. The results indicate that the effect of airflow pattern on droplet size distribution within the spray is a considerable factor in the optimization of airflow and spray together. This could be considered in achieving a limited ignitable region without much diffusion of smaller droplets to the non burning zone during the part load operation of DISI engines.     

Solid-liquid separation in suspension atomization

The aim of this paper is to define the conditions controlling the fragmentation process within the atomization of a suspension. Correlations for the droplet diameter of a suspension spray generated by a twin-fluid nozzle have been derived. Two separate regimes in suspension atomization have been identified with respect to the solid particle size. The atomized droplets from suspensions containing relatively fine solid particles are suspension droplets (containing liquid and solid particles). In this case a correlation for the drop size distribution in the spray of a twin-fluid nozzle has been deduced. Droplet size measurements in the suspension spray with varying solid particle sizes showed that when the suspended solid particle size exceeds a critical value, solid particles and liquid will be more and more separated. This effect is indicated by a bimodal size distribution in the suspension spray. It is shown that complete solid-liquid separation in the suspension spray may be achieved, where the pure liquid drops are significantly smaller than the separated solid particles. The critical process conditions where the solid-liquid separation process is found will be derived. Depending on the operating conditions of the atomizer, the resulting pure liquid droplet size is equal or less than the hydraulic diameter.     

Thermal analysis of the droplet solidification in the PGSS-process

In this work, a manufacturing process for powders from polybutylenterephthalate (PBT) is analysed in terms of the formation of differently shaped particles. For the micronization of the polymer the particles from gas saturated solutions (PGSS) process is used. Several applications require specific particle shapes and particle size distributions. In particular, spherical particles are required to obtain a free flowing powder. To optimise the quality, powders were produced at different spray temperatures and pressures and with varying mass flows of polymer and gas. With different spray conditions, powders with various shapes, morphologies and apparent densities were obtained. The temperatures before and after the expansion appear to have the main influence on the sphericity. To obtain further knowledge about the solidification process, the temperature distribution in the spray was analysed. For this purpose, heat images of the spray were taken. Based on the results, a simple model is introduced, which is able to estimate the temperature field in a polymer droplet during the spray process. An explanation for the formation of differently shaped particles is proposed.     

液液静电雾化特性

对水静电雾化弥散于玉米油的液-液雾化过程进行了实验研究。通过拍照对雾滴形态进行了观察。观察表明,不同电压下液-液雾化会呈现出滴状和云状雾化两种较为典型的雾化形态,在两种形态下液滴具有不同的形貌和运动特点,本文给出了两种雾化形态的出现条件及特征描述。通过Winner99颗粒图像分析仪及雾滴尺寸的分布理论,对不同静电电压下雾滴直径的分布规律进行了定量分析。研究结果表明,液-液静电雾化中雾滴的直径分布服从Rosin-Rammler分布规律。随着电压的升高,雾滴直径分布趋向均匀,雾化细度得到改善。与在空气中雾化有所不同,液-液雾化中雾滴分布的概率密度曲线峰值两边呈现出显著的不对称性,小液滴数尺寸分布较窄而大雾滴数的尺寸分布较宽。随着电压的升高,大雾滴尺寸分布有所变窄,概率密度曲线趋近对称。     

Interaction of droplets with porous structures: Pore network simulation of wetting and drying

In fluidized bed spray agglomeration, the time evolution of a liquid droplet deposited on a porous particle is of paramount importance for the success of the process. The combination of droplet penetration into the pores and evaporation, either directly from the droplet surface or from the surrounding wet pores, determines how long free liquid remains on the particle surface so that other particles can bind via liquid bridges. In this work, a two-dimensional pore network model that combines the algorithms of liquid migration and drying is developed to track the full droplet evolution, from its deposition on network surface to complete evaporation of the liquid. The influence of the pore structure for mono-modal and bi-modal networks with different spatial correlation of the pore size on the evolution of the liquid droplet is investigated. The effect of the liquid viscosity on the evolution of the droplet in the pore network is studied. Moreover, pore network simulations with multiple depositions of liquid droplets on the same network are presented as a rough approximation of spray agglomeration process.     

Dairy Milk Particles Made with a Mono-Disperse Droplet Spray Dryer (MDDSD) Investigated for the Effect of Fat

Mono-disperse droplet generation and subsequent drying in a spray-drying chamber, i.e., mono-disperse droplet spray dryer (MDDSD), provides a better-defined “flight experience” for liquid droplets. The related particle formation can be investigated more easily than that in the usual poly-disperse droplet spray dryer (PDDSD). Previously, skim milk, which is of high protein and high lactose content and is one of the two main dairy fluids that are processed into powder form for consumer markets, was subjected to this kind of investigation in Australia. Here, whole milk, which is the other main dairy fluid, has been spray-dried in a MDDSD set-up at Xiamen University (China). Because the initial droplet size is uniform, measurable, and the particle morphology after drying is consistent, it was possible to investigate the initial solids content effect upon shrinkage and inlet air temperature effect upon shrinkage. In contrast to what had been found for skim milk particles, the formation of the fat-containing (whole milk) particles does not follow the perfect shrinkage model as the skim milk does. This work has improved our quantitative understanding of the whole milk drying process. A fundamental analysis invoked with a modified one-dimensional modeling of spray drying has been given that has shown some further insight about the process.     

Effect of Maltodextrin Addition during Spray Drying of Tomato Pulp in Dehumidified Air: I. Drying Kinetics and Product Recovery

This work investigates the effect of maltodextrin addition on the drying kinetics and the stickiness during spray drying of tomato pulp in dehumidified air. A pilot-scale spray dryer was employed for the spray-drying process. The modification made to the original design consisted in connecting the spray dryer inlet air intake to an absorption air dryer. Twenty-seven different experiments were conducted varying the dextrose equivalent (DE) of the maltodextrin, the ratio (tomato pulp solids)/(maltodextrin solids), and the inlet air temperature. Data for the residue remaining on the walls were gathered. Furthermore, the effect of maltodextrin addition on the drying kinetics and the stickiness of the product was investigated using a numerical simulation of the spray-drying process modeled with the computational fluid dynamics (CFD) code Fluent. The code was used to determine the droplet moisture content and temperature profiles during the spray-drying experiments conducted in this work. The stickiness was determined by comparing the droplet temperature with its surface layer glass transition temperature (T_g ). The T_g was determined using a weighted mean rule based on the moisture content profiles calculated by the CFD code and the experimental data of T_g , which were obtained for the different tomato pulp and maltodextrin samples and fitted to the Gordon and Taylor model.     

Spray characteristics of a swirl atomiser in trigger sprayers using water–ethanol mixtures

Pressure swirl atomisers are widely used in both industry and daily life. It is critical to understand the spray transient behaviour for better design of these systems. This paper presents an experimental study of conical liquid sheets breakup from a swirl atomiser nozzle in trigger sprayers. Spray and atomisation characteristics were measured and analysed. Water–ethanol mixtures were used to simulate different fluids on the breakup and atomisation quality of the spray development process with a wide range of surface tension while maintaining relatively small changes in fluid viscosity and density. The spray images were taken by a high speed digital camera and post‐processed to analyse the global spray structure, spray cone angle, and breakup length. The droplet size and its distribution were measured using a laser diffraction technique. It was observed that the surface waves grow rapidly on the cone‐shaped liquid sheets and breakup into liquid ligaments and droplets during the initial stage of fluid dispensing. Then the spray transitions into the developed stage. Near the end of the dispensing process, the liquid cone collapses with poor atomisation (large droplets) due to momentum loss. The comparison between different fluids showed that the spray cone angle and liquid breakup length decreased with the increase of ethanol percentage ratio. The percentiles parameters, Sauter mean diameter (SMD) and particle size distribution were measured and compared for different locations. High surface tension fluids produce larger droplets than lower surface tension fluids, which have the same trend as the percentiles parameters and SMD. Results also show that droplet size and its distribution depend on the location of the measurement. Generally speaking, smaller droplet size is found for a location away from the nozzle axis in the vertical direction. In the horizontal direction, larger droplet sizes are found for a location closer to the nozzle exit. © 2013 Canadian Society for Chemical Engineering     

Nanoparticle evolution in flame spray pyrolysis—Process design via experimental and computational analysis

In flame spray pyrolysis (FSP), the evolution of metal oxide nanoparticles relies on quite a number of droplet (liquid) and vapor phase related physical mechanism as for instance precursor evaporation, oxidation, nucleation via gas-to-particle conversion mechanism, and subsequent particle (solid) growth mechanisms based on coagulation, sintering/coalescence, and agglomeration. The liquid precursor and dispersion oxygen feed rates are relevant control parameters of the FSP process for tailoring the nanoparticle size (diameter) and structure as well as the atomizer nozzle configuration. Sophisticated nonintrusive, laser-based in situ and ex situ diagnostics with multiscale spatial resolution (micrometer to meter range) are applied for analyzing droplet formation and size, gas velocity, temperature, species concentration, as well as primary and agglomerate diameters along the flow direction. Computational fluid dynamics (CFD) are coupled with population balance modeling (PBM) to elucidate the nanoparticle dynamics within the reactive spray. It is found that the CFD-PBM approach allows estimations of primary and agglomerate nanoparticle diameters within 80 and 75% accuracy compared to experimental data, suggesting that the methods presented could pave the way for designing next-generations of flame reactors.     

大型喷雾粒径分布的图像法测量

雾化技术在能源、化工等领域应用广泛,研究雾化机理和优化雾化喷嘴性能的前提是对其雾化液滴尺寸及粒度分布进行准确有效的测量和表征。目前常用的雾化液滴粒度测量技术,如基于光散射或衍射原理的激光粒度仪和相位多普勒分析仪等,能够较准确地测量粒径分布比较窄、最大粒度在2000 μm以下的喷雾,但对含特大颗粒且粒径分布很宽的喷雾,往往难以得到可靠结果甚至不可能进行测量。本文提出了用图像法测量这类大流量喷雾,构建了图像法测量系统,编写了图像处理程序,经标定实验后,采用该系统对某喷嘴喷雾液滴粒径分布及规律进行了测量研究。研究结果表明图像法可用于大型喷雾液滴粒度及分布的测量。     

Numerical simulation of heavy fuel oil atomization using a pulsed pressure-swirl injector

It is known that increasing the injection pressure reduces the breakup length and the droplet size. Adding pulses, on the other hand, helps to atomize the liquid into finer droplets, similar to air-assisted injectors but without altering the air-to-fuel concentration.To further reduce the droplet size and breakup length, a novel injector type, called ‘‘Pulsed Pressure-Swirl"(PPS), is introduced in this work, which is a combination of pressure-swirl and ultrasonic pulsed injectors. A pressure-swirl atomizer was designed and fabricated specifically for Mazut HFO(Heavy Fuel Oil). The droplet formation process and droplet size distribution have been studied experimentally(by shadowgraphy high speed imaging) and numerically(with the open-source Volume-of-Fluid code Gerris).Changing liquid injection pressure effect on the spray angle and film thickness has been quantified.These simulations have been used to study the primary breakup process and quantify the droplet size distributions, using different injection pulse frequencies and pressures.The numerical results have revealed that the new injector concept successfully produces finer droplets and results in a decrease in the breakup length, especially when applying high pulse frequencies, with no significant changes in the spray angle.     

A simplified analysis method for correlating rotary atomizer performance on droplet size and coating appearance

The dominant method of atomizing automotive paint is through the use of rotating bell sprayers. For this class of atomizer, the problem of paint thickness across the bell has been theoretically solved on a representative geometry that includes factors such as fluid flow rate, bell speed, bell cup radius, and fluid properties. It was assumed that the paint film eventually forms uniform ligaments at the bell cup edge that break due to hydrodynamic stability during the paint spray process; thus, creating a characteristic particle size distribution for the spray. These particle size distributions will vary as the spray parameters, specifically fluid flow rate, bell speed, and bell cup radius, vary. The theoretical model that has been developed strongly correlates to the literature data available for paint droplet size from rotary bell atomizers. Expansion of the correlation of the theoretical model to paint appearance wavelength measurements, Wc and Wd, in place of droplet size provides further understanding of the effect of paint spray parameters on paint appearance. Use of these correlations can help to optimize paint appearance and improve paint spray simulation results.     

MTO催化剂离心喷雾干燥制备过程的数值模拟

采用CFD模型对MTO催化剂喷雾干燥过程进行了非稳态数值模拟,得到了10s内热空气的湿含量、温度、速度、速度向量分布,以及催化剂颗粒的运动轨迹、停留时间和粒径分布.通过与试验结果的对比分析表明,模拟结果可信,模拟离心式喷雾干燥过程可行.