Enhancement of fine particle filtration with efficient humidification☆

Yumei Zhang,Weidong Zhang*,Zhengyu Yang,Junteng Liu,Fushen Yang,Ning Li,Le Du*

The State Key Laboratory of Chemical Resource Engineering,Beijing Key Laboratory of Membrane Science and Technology,Beijing University of Chemical Technology,Beijing 100029,China

1.Introduction

Fine particles as a major air pollutanthave received worldwide attention in recent years[1].Filtration offers many advantages over fine particles removal,in which bag filters have played an important role due to their excellent dust collection capabilities[2,3].In filtration processes,pressure drop of dust cake is dominant in efficiency and economy.

The dust removal arises as the dust-gas mixture passes through the filter media,whereby a dust cake is built up on the filter media.The pressure drop caused by the deposition of dust particles on the filter medium increases with the increasing thickness of dust cake.The increase of pressure drop during filtration is of significant importance because it determines the regeneration frequency of the filter and consequently affects the lifetime of the filter bag.As soon as a defined maximumpressure drop is reached,a regeneration ofthe filtermedium is indispensable.Dust cakes should be detached using widely employed mechanical shock with pulse jet air,and thus the filter could get recovered to its initial state[4,5].Therefore,as the frequency of cleaning increases due to the fast increase in pressure drop,replacement cost is increased due to the higher probability of mechanical damages to the filter bag.

The pressure drop on the surface of the filter medium depends on the structure of dust cake,of which a high porosity is favorable for low pressure drops.Schmidt[6,7]found that the structure depended on the particle size distribution,particle charge,adhesive and cohesive properties of the particles composing dust cakes.St?cklmayer and H? flinger[8]used a 2-dimensional computer simulation model for filtration and showed that the pressure drop increased with a compressible dust cake.Earlier studies have shown that the adhesion characteristics of particles have a great in fluence on the dust cake structure[9-11].For instance,Jeon and Jung[12]used a simulation to investigate the compression behavior of dust cakes and results clearly showed that adhesion,friction,and compression forces acting on a particle played a major role in determining the porosity ofdustcakes.A dustcake composed ofsticky particles is likely to have a loose structure and consequently a high porosity,whereas less sticky particles result in a dust cake with a dense structure and a low porosity.

Several attempts have been made to increase the adhesion characteristics of particles for the loose-structured dust cake including adjustments of electrostatic force and van der Waals force of the particles[13,14].Butmostofthem are based on simulation,and the experimentalintensification was also insufficient due to the long-range forces ranging from only 10-9N to 10-7N for fine particles[15],so that they have little effects on the structures of dust cake and thus the filtration process.

Liquid bridge force,forming from capillary condensation when particles are in a dynamic equilibrium with vapor,has recently gained attention and was investigated in particle packing processes[16].For fine particles,the liquid bridge force can be adjusted in a wide range from 0 to 10-6N[17].However,it remains challenging to uniformly introduce into the particles and most studies were performed in experiment[18,19].Herein,we present one of the first attempts to experimentally introduce liquid bridge into particle removal systems to reduce the pressure drop and intensify filtration process.Upon effective humidification,ultra fine water droplets are uniformly dispersed into air to increase the relative humidity and the vapor pressure,so that the ultra fine water droplets could be combined with particles.

In this work,a novel technique for introducing liquid bridge to enhance the efficiency of filtration based on effective humidification was developed.Two strategies for introducing liquid bridge were investigated,including:(1)ultrasonic atomization,a state-of-the-art technique thatcan generate ultra fine droplets;and(2)steamhumidification(controlling of ambient humidity),a pretreatment technique that makes particles sticky.The in fluences of operating conditions on the pressure drop and the dust cakes structure were experimentally investigated.The structures of the dust cakes were characterized,and the pressure drop and mass loading during filtration were also investigated.

2.Experimental

2.1.Experimental material

Talcum powder used as the dust particles was provided by GUOLI Talcpowder Co.,Beijing,China.The key parameters and the size distributions of the talcum powder are shown in Table 1 and Fig.1.

Table 1 Key parameters of talcum particles

Fig.1.Size distributions of the talcum particles.

2.2.Experimental procedure

The experimental setup mainly composing of a dust generator,a filtrating chamber,an electric ultrasonic atomizer,a water bath,a pressure drop gauge,and a particle counter is schematically illustrated in Fig.2.The gas was dried after drawn into the pipeline.Dust-laden gas was generated when the dry gas passed through a vessel containing a mass of talcum powder.Then the gas mixed with dust that entered into the filtrating chamber,where a hydrophobic polytetra fluoroethylene(PTFE)membrane with nonwoven fiber as the support was employed as the filter medium.For humidification,two different strategies were used,including:(1)ultrasonic atomization,and(2)steam humidification.

(1)In the procedure for ultrasonic atomization,the water bath(No.10 in Fig.2)and the confined cabinet(No.10 in Fig.2)were notused.The fluidized particles passed through the electric ultrasonic atomizer where the water droplets were generated.The droplets were injected at 90 s after the filtration started,to prevent the filter medium being wetted by these droplets at the beginning of filtration.

(2)In the procedure for steam humidification,the ultrasonic atomizer(No.6 in Fig.2)was notused.Before filtration,the particles were in contactwith wetted gas under varying ambienthumidity for6-8 h and then were fluidized for filtration experiment.

Both the aims were to introduce liquid bridge force.After the steps that were mentioned previously,dust was retained on the surface of the medium and the dust cake was formed,as soon as the fluidized particles passed through the filter medium.The pressure drop across the dustcake and the filter medium was measured.At the outlet,the particle counter was used to detect fine particles that passed through the filter medium.

The concentration of particles,which increased with the increasing of the gas flow rate,was determined by the volume flux of the gas passing through the vessel.The mass concentration of the droplets in the filter could be adjusted by the power of electric ultrasonic atomizer or the water bath.The pressure drop across the dust cake and the filter medium were measured and recorded.All the experiments were performed at a gas flow rate of about 0.03 m·s-1,which was a typical high flow rate for gas filtration.

To observe the microstructure of dust cakes using scanning electron microscopy(SEM),fastening of dust cake is necessary due to the crumbly feature.Zhang et al.[20]have successfully fastened the structure of dust cake with adhesive compound.The adhesive compound was injected into the dust cake,and then the dust cake was held on for 5 h to be made fastened.Then the dust cake should be cut and polished so that the structure could be observed.But this method required long time and plenty of adhesive compound.In this study,sodium carboxymethylcellulose was mixed with the particles in the vessel at a dosage of 5 wt%.As the dust cake formed,the sodium carboxymethylcellulose was also retained in the cake on the surface of the filter medium.After the dust cake was wetted by the droplets,the sodium carboxymethylcellulose could act as the adhesive,and then the fastened cake was suitable for observation.

2.3.Analytical methods

A laser particle counter(CLJ-E,SZSYJHSB Co.,China)was used to measure the particle size distributions and the concentration of particles at the outlet of the filtrating chamber.The laser particle counter has 6 channels,and the size range was 0.3-10μmin aerodynamic diameter.The dustcake samples were taken from the surface ofthe filtermedium,and were observed using a scanning electron microscopy.

2.3.1.Concentration of particles

The concentration of particles was given by weighing method,and can be calculated by

where Cpis the concentration ofparticles,mpis the mass of the particles retained on the surface ofthe filter medium,Q is the totalvolume flux of the gas in the filtrating chamber,and ttfis the total filtration time.Because the retaining efficiency of particles could be almost 100%as mentioned previously,the mass of the emitted particles was then neglected.In this study,the concentration of particles was controlled at about 5.36 g·m-3.

Fig.2.Schematic diagram of experimental system.1—Air compressor;2—Gas flowmeter-1#;3—Dust generator;4—Mixing chamber;5—Gas flowmeter-2#;6—Ultrasonic atomizer;7—Dust collector;8—PTFE membrane;9—Digital differential-pressure device;10—Water bath;11—Particle counter;12—Computer;13—Con fined cabinet.

2.3.2.Mass loading

Mass loading is one of the vital parameters indicating the properties of the filter.During filtration,mass loading can be calculated by

where MLis the mass loading,tfis the filtration time,and A is the filtration area.The increase of pressure drop is the result of mass loading,which also increases with filtration time.The maximum mass loading is not only related to filtration time and concentration of particles,but also concerned with the properties of particles.

3.Results and Discussion

3.1.Exploratory experiment of humidification

A series of experiments was performed to measure the pressure drop during filtration.The variations of pressure drop with filtration time and mass loading were determined.A slow increase of pressure drop could be achieved upon conditions of humidification.

Fig.3 showsthe corresponding pressure drop asthe dustcake forming on the surface ofthe filtermedium.In the case ofhumidification,the pressure drop was significantly decreased,as shown in Fig.3(a).At the same pressure drop,the mass loading is obviously larger than that without humidification,as shown in Fig.3(b).Filtration time is then extended until the pressure drop reached 1.0 kPa,which is normally the critical point for regeneration of the filter.The results indicate that liquid bridge force was successfully introduced by humidifying.The collisions between particles and droplets occurred,and then the collisions resulted in the increase of between particles,which was also equivalent to the increase of adhesion force.Then the dust cake composed of larger particles could have a loose structure,which resulted in low pressure drop through the dust cake.

According to L.Bocquet's study[21],the liquid force could be controlled by adjusting humidity and then be accurately quantified by the measurement of the repose angle of the particle systems.In this study,a rapid increase of liquid bridge occurred when the relative humidity was about 40%(water content of 10.3 g·m-2),at which liquid bridges were formed at low water content and thus affecting the process.

The frequency of regeneration of the filter could be reduced by 55%with the ultrasonic atomization,while the controlling of ambient humidity could lead to a 78%reduction in regeneration frequency.Upon the steam humidification,a more obvious reduction was achieved.This is because steam humidification could uniformly introduce the liquid bridge and could increase the porosity of filtercake.Butfor ultrasonic atomization,it is dificult to uniformly disperse the droplets in to fluidized particles,and some droplets were agglomerated into large ones.Therefore,not all the particles were wetted and less liquid bridges formed between the particles,resulting in the less loose structure.

Fig.3.Effects of humidification on pressure drop and mass loading during filtration.

In addition,the pressure drop increases linearly with the filtration time and mass loading,which agrees well with generally known experimental results[8].Because the number of the particles depositing on the surface of the filter medium was constant,and the average porosity of the dust cake formed on the surface of the filter medium is nearly constant.

3.2.Structure characterization of dust cakes

3.2.1.Top surface of dust cakes

SEM results of dust cakes forming without humidification and with humidification are respectively illustrated in Figs.4 and 5.Obviously,the top surface of dust cake formed without humidification is smooth and compact(Fig.4).As the liquid bridge introduced,there were dendrites growing on the top surface of the dust cake,and then the cake is loose and has a high porosity.This is because the improvement of adhesion force made the dust cake loose and incompressible,and thus dendrites could efficiently grow on top surface of dust cakes.

Compared with the ultrasonic atomization,the dendriteswere higher and the dust cake has a looser structure upon the steam humidification.Because when the agglomerated droplets mixed with particles,the excessive water would cause a “muddy”effect,the formation of dendritic structure on the surface was hindered.With the more compacted structure,pressure drop was thus rapidly increased,as mentioned previously(Fig.3).

This phenomenon ofdendrites on the top surface was also noticed in some simulations,in which it is explained as a result of the increasing adhesion force[22-25].Without the great adhesion force,dust cakes would be compressible,and the dendrites would also be destroyed.Upon the formation of porous and loose structure,the growth of dendrites could efficiently increase the permeability across the dust cake,in which case the pressure drop was also deceased.

The dendrites on the surface of dust cakes can be also regarded as a network that is insensitive to compression with increased adhesion force between particles.When the dust-laden gas passes through dust cake,a great quantity of particles could be trapped by the dendrites.In the meantime,the dendrites will grow with the trapping of particles.

3.2.2.Fracture surface of dust cakes

Fig.6 shows SEM results of the fracture surfaces of dust cakes.Similar to the results of the top surface morphology,a dust cake with dendrites was formed in the case of humidification and has a loose structure with a high porosity.The relation between dust cake structure and pressure drop can be explained by the Kozeny equation

where K is the Kozeny constant;εis the cake porosity;μis the gas viscosity;u is the filtration super ficial velocity;Δp is the pressure drop;L is the average thickness ofthe cake,and dpis the dustparticulate diameter.According to Eq.(3),a thick dust cake is considered to have a structure of higher porosity at the same pressure drop.In the case of humidification,the dust cake has a loose structure and is rugged top layer of which is incompressible.The results could adequately explain the decrease of pressure drop with increasing of the cake porosity(Fig.3).

During filtration,compression movements,which are dominant in determining the filtration property,often take place due to the pressure drop across dust cake.To demonstrate the compression phenomena in dust filtration,St?cklmayer and H? flinger[8]proposed to divide the dust cake into three layers of different structures from top to bottom:a loose,incompressible layer on top,a compressible layer,and a compressed,incompressible layer.

In our experiments,the dust cake could be divided into two layers:including a dendritic layer,where the compression movements never take place;and a compressed layer,which is nearly incompressible.In general,the dust cake formed without humidification consists of only one layer,the compressed layer.The major resistance of dust-laden gas passing through the dust cake comes from the compressed layer,while the pressure on the dendritic layer is negligible.This is because the porosity of dendritic layer is much higher than compressed layer.In anotherword,the pressure drop acrossthe dustcake is mainly determined by the compressed layer.

The difference between dust cakes formed by ultrasonic and steam humidification could also be explained with this compression theory.At the beginning of filtration,there are no droplets injected in the filtration chamber before a thin compressed cake layer forming on the filter medium.Withoutthe thin cake layer,some fine droplets would cohere to form large droplets and clog the hole in the filtermedium atthe beginning.And then,with droplet addition,the dendritic layer on the compressed layer arises from cohesion and deposition of fine particles.When the dustladen gas passes through the dust cake,some particles are trapped by the dendrites on top surface of dust cake,and the other particles deposit on the compressed layer.When agglomerated droplets are introduced by ultrasonic atomization and thus large particles collide with dendrites,the dendritic layer is damaged and the compression occurs.Then the thickness of dust cake increases correspondingly,resulting a faster increase of pressure drop than that in the case of the steam humidification.

3.3.In fluence of humidification on filtration

3.3.1.Effects on pressure drop and mass loading

In filtration processes with humidification,droplet-to-particle ratio is an important operating parameter that can in fluence pressure drop.For ultrasonic atomization,droplet-to-particle ratio is calculated based on the amount of atomized water;for steam humidification,the ratio is based on the calculation of the ambient humidity and the volume of the confined cabinet.

Fig.4.Top surface of dust cake formed without humidification.Filtration time:600 s.

Fig.5.Top surface of dust cake formed with humidification.(a,b)ultrasonic atomization, filtration time:600 s;(c,d)steam humidification, filtration time:600 s.

Fig.6.Fracture surface of dust cake with humidification.(a,b)Ultrasonic atomization, filtration time:600 s;(c,d)steam humidification, filtration time:600 s.

Fig.7.Pressure drop with different droplet-to-particle ratios upon ultrasonic atomization.(a)Time dependency of pressure drop;(b)mass loading dependency of pressure drop.

In ultrasonic atomization experiment,the pressure drop on the surface ofthe filtermediumvarying with filtration time and mass loading is shown in Fig.7.The pressure drop increases with the increase of filtration time,and droplet-to-particle ratio has a great in fluence on the pressure drop.Upon injection of water droplets at a droplet-toparticle ratio of 1.13,the pressure drop increases more slowly than thatgenerated withouthumidification.Likewise,the pressure drop generated with humidification at a droplet-to-particle ratio of 1.53 increases much more slowly.However,at a droplet-to-particle ratio up to 3.86,the pressure drop increases most rapidly.Some droplets are combined with the particles or volatilized into ambient air,and the excessive droplets passed through the dust cake.Owing to the hydrophobicity of the PTFE filter medium,the excessive droplets were retained between the surface of medium and the bottom of dust cake.As the accumulation of droplets,liquid layers formed,preventing the gas flow passing through the filter medium.Moreover,the bag sticking occurred,in which case the dust cake is full of excessive liquid.As soon as the liquid layerformed,the pressure drop increased rapidly.Because when the increase of pressure drop is effectively slowed down,the frequency of the regeneration can also be decreased.

In steam humidification experiment,the pressure drop was also measured underdifferentconditions of filtration time and mass loading,asshown in Fig.8.The pressure drop decreasesand the mass loading increases with the increase of the ambient humidity.Compared with the results in Fig.7,under the same pressure drop,a filtration cycle of 2880 s and a mass loading of 720 g·m-2could be achieved at the droplet-to-particle ratio of only 0.0282(water content of the dust cake:0.4%);while the filtration cycle was 1600 s and the mass loading was 552 g·m-2at the ratio of 1.53(water content of the dust cake:26%).Obviously,the steam humidification is more efficient than the ultrasonic atomization.The amount of water was two orders of magnitude less,and the filtration cycle was also longer.Because of the uniform humidifying by controlling ofthe ambient humidity,numerous liquid bridges between particles are produced,increasing the filter cake porosity and leading to the formation of dust cakes with a looser structure.

3.3.2.Optimization of the droplet-to-particle ratio

Serial experiments were carried out to investigate the effects of humidification on mass loading.From Fig.(9),one can note that the maximum mass loading varies with the droplet-to-particle ratio.The maximum mass loading is corresponding to the defined maximum pressure drop(1.0 kPa),which means that the filter can retain more dust particles and filtration cycle will become longer.

In ultrasonic atomization experiment,the maximum mass loading increases gradually from 105 g·m-2to 143 g·m-2at the droplet-toparticle ratio of 0-1.13,as shown in Fig.9(a).With the continuous increase of the droplet-to-particle ratio,the maximum mass loading suddenly increases to 552 g·m-2at the ratio of 1.53,indicating that the water content is about 333.8 g·m-2.This might be caused by the monodispersed fine droplets in the gas and thus the maximum liquid bridges in the particles.After that,the maximum mass loading decreases rapidly to 206 g·m-2atthe ratio of1.53-3.86,and then decreases gradually to 11 g·m-2.Atthis stage,there are excessive droplets accumulating and liquid layerforming on the surface of filtermedium,resulting in bag sticking.

In steam humidification experiment,the maximum mass loading is larger on the whole,compared with ultrasonic atomization experiments.The maximum value is approximately 720 g·m-2at the droplet-to-particle ratio of 0.0282,indicating that the water content is about19.7 g·m-2,as shown in Fig.9(b).Even at this ratio,which corresponds to the relative humidity larger than 100%,uniform humidifying could also be achieved by controlling of the ambient humidity.Not only are numerous liquid bridges produced,but also they are uniformly distributed in most of the particles.The dust cakes then have a loose structure due to the adhesion effects of particles.

Fig.8.Pressure drop with different droplet-to-particle ratios upon steam humidification.(a)Time dependency of pressure drop;(b)mass loading dependency of pressure drop.

Fig.9.The in fluence of droplet-to-particle ratio on the maximum mass loading:(a)ultrasonic atomization;(b)steam humidification.

In brief,both humidification methods can successfully introduce liquid bridge to the particles and significantly reduce pressure drop during filtration.Upon the control of adding water,the interparticle force was adjusted through the liquid bridge.Uniform liquid bridges can effectively preventparticle sliding which resulta loose cake structure.The steam humidification exhibits better performance due to the uniform humidifying.However,the relation between the packing structure ofdustcake and the liquid bridge still remains unclear,which requires further theoretical research.

4.Conclusions

The aim of this study is to develop a novel technique for enhancing the efficiency of filtration based on effective humidification.The experimental results show that the introduction of liquid bridge could ef ficiently slow down the increase of pressure drop during filtration,and the regeneration frequency could be reduced by 50%-78%compared to that of the filtration without humidification.Two strategies,ultrasonic atomization and steam humidification,were used to introduce liquid bridge.The pressure drop and the mass loading during filtration were measured,and the microstructures ofthe dustcakes were characterized.Because of the introduced liquid bridge force,there were dendrites growing on the cake surface and the dustcake also had a loose structure.The steam humidification exhibits better performance due to the uniform humidifying,while the ultrasonic atomization is slightly inferior for the agglomeration ofdroplets and thus hindrance ofthe liquid bridge formation.Furthermore,the amount of water was experimentally optimized.With the droplet-to-particle ratio of 1.53 and 0.0282,the maximum mass loading was 552 g·m-2for the ultrasonic atomization and 720 g·m-2for the steam humidification.Future research is required to reveal the relation between liquid force and particle packing behaviors.

Nomenclature

A filtration area,m2

Cpconcentration of particles,g·m-3

D talcum particle size,μm

dpdust particulate diameter,μm

K Kozeny constant

L average thickness of dust cake,m

MLmass loading,g·m-2

mpmass of the retained particles,g

Δp pressure drop,kPa

Q total volume flux of the gas,ml·min-1

tffiltration time,s

ttftotal filtration time,s

u filtration super ficial velocity,m·s-1

μ gas viscosity,mPa·s

References

[1]J.Tao,L.M.Zhang,G.Engling,R.J.Zhang,Y.H.Yang,J.Q.Cao,C.S.Zhu,Q.Y.Wang,L.Luo,Chemical composition of PM2.5in an urban environment in Chengdu,China:Importance of springtime dust storms and biomass burning,Atmos.Res.122(2013)270-283.

[2]M.Saleem,G.Krammer,Optical in-situ measurement of filter cake height during bag filter plant operation,Powder Technol.173(2007)93-106.

[3]Y.C.Zhao,Extraction of zinc from electric arc furnace dust by alkaline leaching followed by fusion of the leaching residue with caustic soda,Chin.J.Chem.Eng.12(2004)174-178.

[4]D.Koch,J.Seville,R.Clift,Dust cake detachment from gas filters,Powder Technol.86(1996)21-29.

[5]Z.L.Ji,H.Q.Jiao,H.H.Chen,Image analysis on detachmentprocess of dust cake on ceramic candle filter,Chin.J.Chem.Eng.13(2005)178-183.

[6]E.Schmidt,Experiment Investigation into the compression of dust cakes deposited on filter media,Filtr.Sep.32(1995)789-793.

[7]E.Schmidt,Theoretical investigations into the compression of dust cakes deposited on filter media,Filtr.Sep.34(1997)365-368.

[8]C.St?cklmayer,W.H? flinger,Simulation of the filtration behavior of dust filters,Simul.Pract.Theory 6(1998)281-296.

[9]E.Schmidt,T.Pilz,Raw gas conditioning and other additional techniques for improving surface filter performance,Filtr.Sep.33(1996)409-415.

[10]K.Morris,R.W.K.Allen,Experiment investigation into the compression of dust cakes deposited on filter media,Filtr.Sep.33(1996)339-342.

[11]H.C.Chi,Z.L.Ji,D.M.Sun,L.S.Cui,Experimental investigation of dust deposit within ceramic filter medium during filtration-cleaning cycles,Chin.J.Chem.Eng.17(2009)219-225.

[12]K.J.Jeon,Y.W.Jung,A simulation study on the compression behavior dust cakes,Powder Technol.141(2004)1-11.

[13]J.P.K.Seville,C.D.Willett,P.C.Knight,Interparticle forces in fluidisation:A review,Powder Technol.113(2000)261-268.

[14]C.J.Van Oss,M.K.Chaudhury,R.J.Good,Interfacial Lifshitz-van der Waals and polar interactions in macroscopic systems,Chem.Rev.88(1988)927-941.

[15]N.G.Van Kampen,B.R.A.Nijboer,K.Schram,On the macroscopic theory of van der Waals forces,Phys.Lett.A 26(1968)307-308.

[16]J.Crassous,L.Bocquet,S.Ciliberto,C.Laroche,Humidity effect on static aging of dry friction,Europhys.Lett.47(1999)562.

[17]P.A.Kralchevsky,N.D.Denkov,Capillary forces and structuring in layers of colloid particles,Curr.Opin.Colloid Interface 6(2001)383-401.

[18]N.Stevens,J.Ralston,R.Sedev,The uniform capillary model for packed beds and particle wettability,J.Colloid Interface Sci.337(2009)162-169.

[19]A.F.Payam,M.Fathipour,A capillary force model for interactions between two spheres,Particuology 9(2011)381-386.

[20]W.D.Zhang,G.Wang,Y.Q.Ge,J.T.Liu,Z.Y.Yang,Z.Q.Ren,B.Ding,Method for fixing filter cake in bag type dust collector,CN Patent 101912705 B,6 June,2012.

[21]L.Bocquet,E.Charlaix,S.Ciliberto,J.Crassous,Moisture-induced ageing in granular media and the kinetics of capillary condensation,Nature 396(1998)735-737.

[22]S.J.R.Simons,Modelling of agglomerating systems:From spheres to fractals,Powder Technol.87(1996)29-41.

[23]F.X.Fan,L.J.Yang,J.P.Yan,J.J.Bao,X.L.Shen,Experimental investigation on removal of coal- fired fine particles by a condensation scrubber,Chem.Eng.Prog.48(2009)1353-1360.

[24]L.J.Yang,J.J.Bao,J.P.Yan,J.H.Liu,S.J.Song,F.X.Fan,Removal of fine particles in wet flue gas desulfurization system by heterogeneous condensation,Chem.Eng.J.156(2010)25-32.

[25]C.F.Yang,Aerosol filtration application using fibrous media—An industrial perspective,Chin.J.Chem.Eng.20(2012)1-9.