Xiaolai Zhang ,Weixin Qian ,Haitao Zhang ,Qiwen Sun ,Weiyong Ying ,*

1 Engineering Research Centre of Large Scale Reactor Engineering and Technology,Ministry of Education,State Key Laboratory of Chemical Engineering,East China University of Science and Technology,Shanghai 200237,China

2 State Key Laboratory of Coal Liquefaction and Coal Chemical Technology,Shanghai 201203,China

Keywords:Fischer-Tropsch synthesis Laser Doppler velocimetry Fluidized bed

ABSTRACT In this paper,the distributions of particle velocity in a gas-solid fluidized bed with branched pipe distributor or circle distributor were measured by using a laser Doppler velocimetry.Our results show that,within a certain range of super ficial gas velocity,when using circle distributor,the particle velocity is large and the distribution of the particle velocity is even more compared with the branched pipe distributor.On the basis of the amplitude oftangentialmovementstatistics,the amplitude oftangential movementstatistics(AVATMS)decreases with increasing the axial height under the appropriate super ficial gas velocity.

1.Introduction

Fischer-Tropsch synthesis(FTS)has been known as a feasible and environmentally friendly technology for the production of liquid from synthesis gas derived from the relatively more abundant resources,such as natural gas,coal and biomass.The fuels produced by FTS are of high quality[1]because they do not contain sulfur and nitrogen compounds as well as aromatics.Due to the increased concern about the problems of remote gas utilization and environmental pollution,and the implementation of more stringent environmental legislation on liquid fuels,the production ofgasoline fromthe FTS process becomes more suitable[2].

There has been increased attention in the development of FTS technology in the last couple of years.In general,there are five major types of commercial FT reactors that commonly implemented in the open literatures,including the multi tubular fixed-bed reactor(MTFBR)[3,4],the slurry phase reactor(SPR)[5-7],the circulating fluidized-bed reactor(CFBR)[8,9],the fluidized-bed reactor(FBR)and micro channel or micro structured reactors(MR)[10,11].

FBR and CFBRare only used forthe FT processes athigh temperature,which produce gasoline or linear ole fins with the Fe-based catalysts in the range of 300-350°C[8,12].Due to the successful application of the CFBR system for catalytic cracking(FCC process)during the World War II by Standard Oil,several organizations have developed versions of this type reactor for FTS.The developing history of FBR has been well documented by Davis[13].The circulation of large amount of catalyst using CFBRs leads to considerable compression of recycle gas with associated capital and operation costs[14].Sasol R&D have developed a fluidized-bed system,which has similarselectivity and higher conversion rate than the commercial CFBR for the FTS pilot plant[15].Dry and Steynberg enumerated some advantages about FBR(called SAS reactor)relative to FBR and CFBR,such as the higher pass per conversion,lower catalyst consumption,less maintenance and easier construction[15,16].Steynberg reported that a total of 16 CFBR in the Sasol plants located in Secunda and South Africa have been replaced by FBR,and the FBRs(SAS reactors)have been successfully used over 30 years at Sasol[16].

Although the laser Doppler velocimetry(LDV)measurement is limited by many factors,such asthe near-walleffectand dilute flows,many researchers still use it owing to its accuracy,simplicity and capability in measuring the individual particle velocities.Sayeed-Bin-Asad analyzed the turbulent wake formed behind a semi-circular step cylinder at a constant flow rate by using LDV measurements[17].Mychkovsky utilized the LDV measurements to study the gas and particle velocity profiles in a jetting fluidized bed[18,19].Liused LDV to discuss the effectof the operating parameters on the particle velocity in turbulent fluidized beds[20].By using a PV6D optical probe,Wu measured the solid concentration and the distribution of particle velocity in the transition section of a Φ 200mmturbulent fluidized bed(TFB)and a Φ 200 mm annulus turbulent fluidized bed(A-TFB)[21].

To develop the advantages of fixed fluidized bed in F-T synthesis,one of the key factors for determining the law of particle motion is to optimize the gas distributor.However,due to the limitation of the testing level,it is difficult to obtain the accurate data from the fixed bed fluidization,which affects the understanding of the flow behavior.At present,the hot-wire film anemometer(HWFA)and the laser Doppler velocimetry are two commonly used methods for measuring the particle flow.The HWFA belongs to the contact measurement,however,the probe goes through into the fluidized bed and willaffectthe flow behavior,in particular,the particles directly impact the measurement probe.Moreover,the surface of the catalyst is easy to coke in the F-T synthesis,making the impact more serious.The LDV is a non-contact measurement,thus,the flow field in the fluidized bed would not be affected.In this paper,in order to discuss the velocity distribution of the particles in the fixed fluidized bed with both the branched pipe distributor and the circle distributor,we employed the laser Doppler velocimetry to determine the motion of particles.

2.Experimental

The fluidized bed reactor is made by a transparent plexiglass material.The entire system is sustained by a steel stent.The size of the fluidized bed is Φ324 mm×12 mm×5000 mm.The discharge port is set at the bottom ofthe reactor and the inlet is installed at1.8 m from the bottom of the reactor.With the operation of the fluidized bed,some particles run into the cyclone with the gas,and then go back to the fluidized bed after the gas-solid separation.It should be noted that the distributorscan be separately installed atthe bottom ofthe fluidized bed reactor(Fig.1).

Fig.1.Fluidized bed cold modelexperimental apparatus for Fischer-Tropsch synthesis.1—lf ow controlvalve,2—rotameter,3—pressure gauge,4—discharge valve,5—gas distributor,6—feed inlet,7—cyclone,8—exhaustport,9—particle circulation controlvalve,10—particle circulation discharge valve.

Both the branched pipe distributor(BPD)and circle distributor(CD)are designed for FTS,and their structures are shown in Fig.2.The branched pipe distributor is centrosymmetric.The diameter and the length of the main tube are 0.032 m and 0.24 m,and the diameter of the side tube is 0.015 m.Overall,56 nozzles were setup in the branched pipe distributor and the diameter of each nozzle is 0.002 m.For the circle distributor,the diameterofthe main tube is 0.025 m and the internal diameter is 0.21 m.The external diameter is 0.26 m.The nozzles were also set up in the circle distributor,and the diameter of each nozzle was 0.002 m.The 29 nozzles are distributed on a circle ring and 9 nozzles were in the diameter,and each string had 7.The total number was 56.The opening rate of two distributors was 2.5‰.The design of circle distributor was based on the pores in the centre of the ring near the wall,while the design of branched pipe distributor was based on a uniform pore distribution.

Fig.2.The structural representation of the distributors.

This experiment was the cold model test.The results show that the particle motion was only dominated by the main features of the gas phase flow.For matching the experimental results with the actual conditions of the catalyst,the particles applied in the fluidized bed were all spherical glass particles with diameters ranging from 150 to 180 μm.

The gaseous medium is air.The air goes through the gas compressor and buffer tank,which remove water,oil and other impurities by passing gas filters and freeze dryer.The flow rate was regulated by a set of rotameter.The operation was carried out at room temperature and the normal pressure.Initially,a certain amount of glass particles were filled in the fluidized bed with a given bed height.After supplying air forabout15 min,the particles in the fluidized bed were in complete fluidization situation without involving other undesirable flow phenomena,such as large bubbles,slugging and channeling,and this fluidized state was regarded as a fine condition at a certain super ficial gas velocity.After the fluidization situation is stable,the particle velocity under different experimental conditions was measured using laser Doppler velocimetry system.The test height of 2.05-2.65 m was selected.The super ficial gas velocity ranged from 0.39-0.59 m·s-1and the static bed height ranged from 0.3-1.1 m.

The LDVsystem(TSIInc.)was made up by a 5 Wargon ion laser(Coherent,LA70-5E).The optical arrangementused a double probe system,which has a two-dimensional fibreoptic probe(TR-260)and a one-dimensional fibreoptic probe(TR-160).The TR-260 provides the green(514.5 nm)and blue(488 nm)laser and the TR-160 affords the purple(476.5 nm)laser.The LDV system also has Fibre Light(FBL-3),Flow and Size Analyzer(FSA3500)and Photo Detector Module(PDM1000).The software of FlowSizer64(TSI Inc.)was used to realize some functions,such as hardware control,data acquisition,detailed tabular and graphic displays,coordinate frame control and data output.

3.Results and Discussion

3.1.Effect of the super ficial gas velocity on the axial particle velocity

The super ficial gas velocity in the fluidized bed was found to affect the state of the gas-solid behavior.Fig.3 shows how the super ficial gas velocity ranging from 0.43 to 0.59 m·s-1affects the axial particle velocity at a height of 2.5 m with the circle distributor(CD)and branched pipe distributor(BPD),respectively.In the abscissa of Fig.3,r means the distance from the centre of the circle,and R is the radius of the circle.

Fig.3.Radial profiles of axial particle velocities versus super ficial gas velocity(test bed height is 2.5 m,static bed height is 0.9 m,(a)with CD(b)with BPD).

The common ground of the two figures is that the particle velocities rise with the increase ofsuper ficialgas velocity for both distributors and the particle velocities near the wall are almost zero under all conditions,which means that the particles tend to gather near the wall.Away from the wall,the particle velocitiesincrease rapidly.For BPD,there isa velocity hump between the dimensionless radial coordinate from 0.2 to 0.8,which indicates that larger bubble forms at this zone,and there is a larger gas drag force acting on the particles.For CD,the particle velocities only changed little when particles are away from the wall.The reason may be that the circle distributor is based on the pores in the centre of the ring and closes the wall that decreases the wall effect.

3.2.Particle velocity distribution in FBR with CD and BPD

Fig.4 shows the axialparticle velocities atdifferentbed heights ranging from 2.2 to 2.65 m in FBRwith CDand PBD,where Uisthe super ficial gas velocity.

Fig.4.Radial profiles of axial particle velocities at different bed heights((a)with CD,(b)with BPD,static bed height is 0.9 m,super ficial gas velocity is 0.51 m·s-1).

As shown in Fig.4(a),the particle velocities increase with the increased axial coordinate within this height range with CD.For the considered kind of particles,the drag and gravity are the most relevant forces.These results mean that the drag force acting on the particles was greater than their own gravity,therefore,the solid particles obtained an upward acceleration within this height range.Fig.4(b)also shows thatthe particle velocities increase with the increased axialcoordinate ranging from 2.2 to 2.5 m with BPD.Under the same conditions as CD,the particle velocity is larger at the height of 2.65 m than that of the height of 2.5 m,which means the gravity force is dominant above 2.5 m with BPD.The fluctuation of the particle velocity with CD is less than with BPD.

Fig.5 summarizes the axialparticle velocities atdifferentbed heights ranging from 2.2 m to 2.65 m in FBR with CD and BPD.

Fig.5.Radial profiles of axial particle velocities at different bed height with BPD and CD(red scatter:BPD,black scatter:CD,static bed height is 0.9 m,super ficial gas velocity is 0.55 m·s-1).

Fig.5 shows that the increasing rate of particle velocities near the wall with CD is higher(dimensionless radial coordinate range 0.8-1)than that with BPD.The average opening rates of BPD and CD are the same.The uneven opening may account for the above results.The wall effectis considered as one major reason to affect the particle distribution in fluidized bed.The result with the test height(Z)of 2.5 m and 2.65 m indicates that the CD with uneven open pores could effectively overcome the resistance on the particle caused by the wall effect.

3.3.Effect of the static bed height on the particle velocity distribution

Fig.6 shows the effectofstatic bed heightrange from 0.7 to 1.1 m on the axial particle velocity.

One can see that,when the static bed height increases from 0.7 to 1.1 m,the changing trends along with the radial coordinates of the axial particle velocity are almost the same.At the same test height,the particle velocity increases with the static bed height.When using CD,the trends are almost the same.The distribution trend of particle velocity does not change with increasing the static bed height.

3.4.The absolute value of the amplitude of tangential movement statistics in a FBR

The moving direction of a single particle in the fluidized bed is random,but we can obtain some statistical properties in the same horizontal plane.In the fluidized status,the particles have an overall axial motion except the random motion.Considering the axial symmetry of the reactor,to simplify the analysis,only one-dimensional tangential movement was considered which is shown in Fig.7.

Fig.6.Axial particle velocity versus static bed height with BPD(super ficial gas velocity is 0.51 m·s-1,test height is 2.5 m).

The test of 300 s was used as the statistical unit and the data were measured by the laser Doppler velocimetry.After being processed by the software named Flowsizer 64,the data for the moving particles along the three directions were obtained.Among the three directions,the main direction is the axial direction,which represents the particle velocity running along with an axial motion.But the axial motion cannot show the level of turbulence,so a new idea should be given to describe the tangential movement of the particle.On the basis of the tangentialdata ofthe testpoint,the horizontaltangentialpositive direction is shown in Fig.7,in which the symbol of the tangentialdata represents the moving direction of the particles.By adding up the absolute values of the two directions together,the sum was the amplitude at thatpointwas defined as the absolute value of the amplitude of tangential movement statistics(AVATMS),which is given by

Fig.8 shows the relationship between AVATMS and r/R at different test heights with super ficial gas velocity of 0.55 m·s-1and the static bed height of 0.9 m using branched pipe distributor.Fig.9 shows the contact between AVATMS and r/R at different super ficial gas velocity under the condition of test height of 2.5 m and the static bed height of 0.9 m using branched pipe distributor.The other experimental conditions are the same as before.

From Fig.8 one can see that the AVATMS decreases with increasing the axial height under an appropriate super ficial gas velocity,which is coincident with the actual situation and reflects the violent movement of the circulating fluidized in the fluidized bed.Fig.9 shows that the AVATMS on each cross section tends to be the same under an appropriate super ficialgas velocity.The curve with the super ficial gas velocity of 0.39 m·s-1corresponds to the initialboundary of a developed fluidized condition,therefore,the AVATMS shows a significant volatility.These experimentalresults may provide a new effective method foronline detection of the fluidized state in the production process.The experimental results in circle distributor show the similar case.

Aiming at the difficulty in the determination and quantification of the concentration of the fluidized bed,the AVATMS experimental results can be used to accurately measure the fluidized bed at a certain height.The range of the corresponding AVATMS can be determined on the basis of the optimal reaction conditions of the Fischer-Tropsch synthesis.Using AVATMS,it is possible to accurately and continuously control the Fischer-Tropsch synthesis reaction process,which provides an effective and continuous on-line test control criterion.In industrial production,since the reactor is running,it is difficult to know the internal fluidization state.In order to detect whether the fluidized bed is in a homogeneous fluidization state,the AVATMS can be used at this momentsince the LDVdoes notneed to enter the reactor interior.In the research of high-temperature Fischer-Tropsch synthesis,the AVATMS method was studied by laser Doppler velocimetry,which can improve the current research mode by seeking the combination of the continuous experimental measurement with the computer simulation.

Fig.7.Schematic diagram of tangential movement of particles.

Fig.8.The relationship between AVATMS～r/R(static bed height is 0.9 m,super ficial gas velocity is 0.55 m·s-1).

Fig.9.The relationship between AVATMS～r/R at Z is 2.5 m(static bed height is 0.9 m).

4.Conclusions

Within a certain range of super ficialgas velocity,compared with the branched pipe distributor,the particle velocity is greater and the distribution of the particle velocity is more even when using circle distributor.This might relate to the uneven distribution of the circle distributor.The particle velocities increase with increasing the super ficial gas velocity for both distributors.Under the same conditions,the fluctuation of the particle velocity with CD is less than with BPD.On the basis of defining the AVATMS,the experimental results suggest that under the appropriate super ficial gas velocity,the AVATMS decrease with the increase ofaxialheight.Underan appropriate super ficial gas velocity,the AVATMS on each cross section tends to be the same.AVATMS directly reflect the initial boundary shape of fluidized bed.