`
`ELSEVIER
`
`Powder Technology 200 (2010) 224-233
`
`.
`-
`5
`5
`Contents lists available at ScienceDirect
`
`Powder Technology
`journal homepage: www.elsevier.com/locate/powtec
`
`ime
`|
`|" POWDER
`
`
`
`
`
`Jet cup attrition testing
`
`R. Cocco*, Y. Arrington, R. Hays, J. Findlay, S.B.R. Karri, T.M. Knowlton
`Particulate Solid Research, Inc. Chicago, IL, United States
`
`ARTICLE INFO
`
`ABSTRACT
`
`so10
`
`ruary
`
`Article history:
`Received 28 May ane is Feb
`Received
`in revised form
`19 Fel
`3010
`Available online 1 March 2010
`
`Keywords:
`Particle attrition
`Jet cup
`
`Jet cup attrition testing is a common method for evaluating particle attrition in fixed fluidized beds and
`circulating fluidized beds. An attrition index, calculated from jet cup data, is used to compare with one or
`:
`.
`.
`ont
`.
`:
`more reference materials. However, this method is far from perfect despite its popularity. Results obtained at
`Particulate Solid Research, Inc. (PSRI) in different-sized jet cups and a 29-cm (11.5-in,) diameterfluidized
`bed test unit did not provide the same ranking of catalyst with respect to particle attrition, To obtain a better
`understanding ofattrition in a jet cup, both computational fluid dynamics (CFD) and cold flow studies were
`performed with a 2.5-cm (1-in.) diameter Davison-type jet cup and PSRI's cylindrical 7.6-cm (3-in.) diameter
`jet cup. Results showed that a significant amount of material in the Davison and PSRI jet cup remained
`stagnant. Based on these results and additional CFD modeling, PSRI designed a new jet cup, where most of
`the material was hydrodynamically active. The new jet cup showed a 25% increase in attrition compared to
`PSRI's cylindrical jet cup under similar conditions and run times, Results were also compared to cyclone
`attrition data for several materials at PSRI. The new jet cup provided data that correlated with attrition
`results from the 29-cm (11,5-in,.) diameter fluidized bed unit.
`© 2010 Elsevier B.V.All rights reserved.
`
`
`1. Introduction
`
`The Davisonjet cup attrition method is the most common method
`of ranking particle attrition for fluidized bed, riser, and cyclone
`applications. The Davisonjet cup consists of a 2.5-cm (1-in.) diameter
`cup [1]. The cup has a tangential gas inlet and is attached to a large
`disengagement chamber, as shown in Fig. 1. Approximately 5 or 10g
`ofthe test material are placed into the Davisonjet cup. Gas is added at
`high velocities through the tangential inlet. Fines generated in the cup
`dueto attrition enter the disengagementsection, where they refluxed
`back into the jet cup. This process continues until the particles become
`too small and escape through the outlet. The materialloss is related to
`attrition loss.
`PSRI has expanded onthis concept by capturing thelost material in
`a filter and using a 7.6-cm (3-in.) diameter jet cup with 100g of
`material [2]. Capturing the entrained material allows for a complete
`material balance. The larger sample size reduces measuring errors and
`enables material balances in excess of95% to be achieved. This is often
`not the case for the smaller sample sizes in the 2.5-cm (1-in.)
`diameter Davison jet cup. The larger jet cup size also allows attrition
`testing of larger Geldart Group B or D materials, such as pelletized
`high density polyethylene (HDPE).
`Both jet cup methodsuse a tangential gas inletin a cylindrical cup
`to produce tangential or swirling flow that mimics the particle-wall
`
`* Corresponding author. Tel.: +1 773 523 7227.
`E-mail address: ray.cocco@psrichicago.com (R. Cocco).
`
`0032-5910/$ — see front matter © 2010 Elsevier B.V. All rights reserved.
`doi:10.1016/j.powtec.2010.02.029
`
`impacts in cyclones, fluidized beds and risers. The jet cup method is
`primarily used to rank catalyst attrition in termsof an attrition index
`(Al), where the weightfraction of particles smaller than a specific size
`is compared before and after the attrition testing. Typically, the
`particle size used to denote “fines” is at sizes less than 20 or 44 um.
`The Davisonor standard cylindrical jet cup method has been found
`to provide only relative comparisons among different catalysts. It
`cannotpredict the quantitative extentoftheattritionorattrition rate
`that will occur in a commercial process. This limitation was proposed
`because the prevailing stress mechanismsin thejet cup test differ from
`thosein large-scale processes [3]. PSRI has found similar results. The
`attrition index data were not useful for predicting absolute attrition
`values in commercial units. Yet, jet cups are commonly being used to
`compare ranking ofvarious materials catalyst using the attrition index.
`In other words, the attrition rate of new material is based on some
`reference material, perhaps a predecessorof the new material.
`However,PSRI has found that the standardjet cup method may not
`even be suitable for ranking catalyst and other materialattritionrates.
`PSRI used CFD and cold flow experimental studies to discern the
`underlying hydrodynamics responsible forparticle attrition in the jet
`cup device. Results showed that the standard jet cup design was
`ineffective in causing all particles to be in motion whether a 2.5-cm
`(1-in.) or 7.6-cm (3-in.) diameter cup was being used. Based on these
`results, PSRI designed a new conical jet cup that was able to achieve
`better particle mobility and higherattrition rates. The relative cyclone
`attrition ranking from the conical cup also agreed well with the
`attrition ranking from a 29.2-cm (11.5-in.) ID fluidized bed cyclone
`attrition test unit.
`
`WRG-1015
`
`1
`
`WRG-1015
`
`

`

`R. Cocco et al. / Powder Technology 200 (2010) 224-233
`
`130 cm |
`
`225
`
`Filter
`Housing
`
`
`
`
`Disengagement
`
`
`Jet Cup \_.-GasInlet
`
`Fig. 1. The overall Davison jet cup (left) and close up of the jet cup (right). Based on
`Weeks [1].
`
`Fig. 3. PSRI's cylindrical jet cup apparatus.
`
`chamberthroughthe outlet port. A pressure gauge is also located on
`the chamber. The PSRI jet cup is equipped to reach temperatures of
`815 °C; however,all measurements conducted in this study were at
`room temperature.
`Gas flow rates were controlled with two Dwyer 50 and 400 SCFH
`rotameters. Gas flow rates were checked after the first 15 min and
`then at each hourafter that. Flow rates were corrected for error due
`increasing back-pressure (ifany) from the filtering media. Magnehelic
`and Marshpressure gauges were used to measure the pressure in the
`test unit.
`Prior to testing, the powder was dried in an oven at 425°C
`overnight. The powder was then riffled into equal 10 and 100g
`samples. The particle size was analyzed on one ofthe dried riffled
`material samples using an electrical zone sensing Coulter Counter
`(model Multisizer II). The other sample was poured into the jet cup,
`and the cup was fastened to the attrition test unit shownin Fig. 3.
`The rotameter was set such that jet velocities were either 76.2,
`137.2 or 182.9 m/s (250, 450 and 600 ft/s). The test was conducted for
`1h and operated below the threshold velocity where fragmentation
`dominates.
`After testing, the jet cup was carefully removed and the contents
`were weighed. Material collected on the filter media and inside the
`disengagementsection of the jet cup unit was removed and weighed.
`Particle size analysis was conducted on material left in the jet cup and
`the material collected from the filter media. A material balance was
`conducted to ensure that at least 95% ofthe material was accounted for.
`Jet cup results were presented in terms ofan attrition index (Al).
`The attrition index is determined by comparing the cumulative
`weightpercentatthe sizeofinterest after the test to theinitial weight
`percentof that size fraction. Particles smaller than this particle size
`were considered as fines. For this study, fines were defined as particles
`smaller than 20 and 44 um.
`
`2.3, 29.2-cm (11.5-in.) diameterfluidized bed cyclone attrition test unit
`
`The attrition indices from the jet cup studies were compared to
`attrition indices obtained from studies in a 29.3-cm (11.5-in.) ID
`fluidized bed with primary, secondary, and tertiary cyclones, as
`shown in Fig. 4. The solids loading and inlet gas velocity to the primary
`cyclone were held constantfor each test at 3.2 kg/m?(0.2 Ib/ft?) and
`12.2 m/s (40 ft/s), respectively. The superficial gas velocities in the
`bed and in the freeboard were varied independently to preserve the
`loading and gas velocity restrictions on the primary cyclone. Collected
`particles from the primary cyclone were returned to the fluidized bed.
`
`2. Experimental
`
`2.1, Powder material
`
`Equilibrium FCC catalyst powderwasused forthe cold flow studies
`and modeled in the CFD studies. The particle density was to be
`1492 kg/m? (93 lb/ft?). Fig. 2 showsthe particle size distribution of
`the FCC catalyst material, which had median particle diameter (d,50)
`of 78 um and Sauter mean diameter of 81 pm.
`Proprietary catalyst used in the 29.2-cm (11.5-in.) ID fluidized bed
`cycloneattrition study hada d,s9of53 um and a Sauter mean diameter of
`55 pum. The corresponding particlesize distribution is also shown in Fig. 2.
`The proprietary catalyst particle density was 1458 kg/m? (80 lb/ft*).
`
`2.2, Jet cup attrition measurements
`
`Jet cupattrition studies using the 2.5-cm (1-in.) and 7.6-cm (3-in.)
`diameterstandard jet cup were performed in the sametest unit. Fig. 3
`provides a schematic drawing of the test unit. The 2.5-cm (1-in.)
`diameter jet cup was typical of a Davison jet cup design. The test
`procedureswere also similar for each cup size except 100 g of sample
`was used for the 7.6-cm (3-in.) diameter cup instead of the 5 or 10g
`used in 2.5-cm (1-in.) diameter cup. The larger sample size reduces
`Measurement error compared to the 2.5-cm (1-in.) diameter or
`Davison jet cup.
`As shownin Fig. 3, the cup is attached to a 130-cm (51-in.) high
`disengagementsection, where the diameteris increased to 30.5-cm
`(12-in.). A 5pm sintered metal filter is inserted in the expansion
`
`—— FCC Catalyst
`—— Proprietary Catalyst
`
`25
`we
`
`OTC
`
`\ 1
`
`0
`
`50
`
`100
`
`50
`
`200
`
`250
`
`300
`
`dp, um
`
`Fig. 2. Differential particle size distribution of the FCC catalyst and proprietary catalyst
`powder used in the CFD simulations and jet cup experiments.
`
`2
`
`2
`
`

`

`é) Ball Valves Recycle Compressor
`
`Fig. 4. The PSRI 29.2-cm (11.5-in.) diameter fluidized bed cyclone attrition test unit.
`
`The particles collected from the secondary cyclone were used for the
`attrition measurements and notreturned to the fluidized bed.
`The unit was operated for an extended period oftime to ensure that
`the equilibrium attrition rate for each sample was being measured.
`Samples were collected periodically from a side port on the bed as well
`from the secondary cyclone dipleg. Particle size analysis was
`conducted in a similar method as with the jet cup samples.
`
`2.4, Jet cup cold flow study
`
`Several Plexiglas™ jet cup configurations and test conditions were
`examined. The Plexiglas jet cups were used for visualization and
`matched, in design, their stainless steel counterparts used forattrition
`studies. PSRI uses a 7.6-cm (3-in.) diameterjet cup containing 100 g of
`
`R. Cocco et al. / Powder Technology 200 (2010) 224-233
`
`material, Jet velocities used in this study were at 76, 137, 183, and
`274 m/s (250, 450, 600, and 900 ft/s). The axial length from the
`bottom ofthe cupto the bottom ofthe disengagementsection was 15-
`cm (6-in.). The jet orifice inner diameters were 0.24 or 0.48 cm
`(0.0938 or 0.1875in.).
`The second jet cup was a 2.5-cm (1-in.) diameter cup that
`represented jet cups typically used in accordance with the Davison
`methodology [1]. The 2.5-cm (1-in.) diameter Davison jet cup was
`tested atjet velocities of 76, 137, 183, and 274 m/s (250, 450, 600, and
`900 ft/s). The jet orifice inner diameter was 0.24 cm (0.0938in.). The
`jet cup height was 8.25 cm (3.25in.). The axial length from the
`bottom of the cup to the bottom of the disengagement section was
`18cm (7in.). A 9.75-cm (3.8-in.) long conical spool piece was
`inserted between the Davison jet cup and the PSRI jet cup unit to
`ensure a smooth transition between the cup and disengagement
`section at the same open angle as the disengagementsection. The
`Davisonjet cup was tested with 5 and 10g of equilibrium FCCcatalyst.
`New cup designs were based on improving the standard PSRI jet
`cup's performance. Plexiglas cups were constructed to test various
`concepts including displacing the stagnantregion, adding morejets to
`reduce the stagnant region and/orincrease the axialor lifting velocity
`in the cup.This resulted in the following alternative cup designs: the
`angled jet cup, the dual jet cup, the dual jet with conical insert jet cup
`and the conical jet cup. All of these concepts are illustrated in Fig. 5.
`All jet cup concepts were designed with a 7.6-cm (3-in.) diameter
`outlet. The conical jet cup diameter was reduced to 3.8-cm (1.5-in.) in
`diameter at the bottom ofthe cup. The inlet jet diameter was either
`0.24- or 0.48-cm (0.0938- or 0.1875-in.) ID and fastened tangentially
`to the bottom portion of the cup. The Plexiglas™ cups were attached
`to the same attrition unit as that used for the attrition measurements
`
`shown inFig. 3.
`
`2.5. CFD simulations ofjet cup hydrodynamics
`
`A CFD modelusing Barracuda™ version 10.0 from CPFD-Software,
`LLC. was used to explore gas and solid hydrodynamics in the jet cup
`attrition test units. Barracuda™ is a Lagrangian-Eulerian hybrid code
`employing the multiphase particle-in-cell (MP-PIC) numerical meth-
`od, which has been formulated for dense particle flows [4,5].
`Thecarrier gas is treated as a continuum in an Eulerian framework
`using the Reynolds Averaged Navier Stokes (RANS) equations. The
`particles are treated as discrete entities. In order to track a large
`number ofparticles (millions to billions), particles of similar sizes are
`treated as parcels or “clouds.” The drag force between each phase is
`coupled (two-way). Parcels of N particles are assumed to have a drag
`force of one similar particle times N. Particle drag within eachparticle
`parcelor cloud is assumed to interact with the gas phase independent
`of other clouds. For this study, both phases were assumed to be
`isothermal.
`The interparticle, normalstresses for collisions between particles
`were defined by the relationship of Harris and Crighton [6]. The
`particle-fluid drag was expressed using the drag law proposed by
`Gidaspow [7] where drag is described by Wen-Yu and Ergun.A linear
`
`
`
`Standard Cylindrical
`
`Angled
`
`Dual Jet
`
`Conical
`Dual Jet with Conical Insert
`
`Fig. 5. Five configurations used for evaluating jet cup attrition.
`
`3
`
`3
`
`

`

`R. Cocco et al. / Powder Technology 200 (2010) 224-233
`
`227
`
`(93 Ib/ft?). The entire particle size distribution was modeled using
`Barracuda™,
`The boundary conditions for the simulation were a pressure
`boundary condition at the top of the disengagement region and a
`velocity boundary condition at the tangential jet. The pressure
`boundary condition was set at 104,771 Pa (15.3 psia). The velocity
`boundary condition was set at 137 m/s (450ft/s) corresponding to
`one of the experimental jet cup conditions.
`
`3. Results
`
`3.1. Cold flow study
`
`Fig. 6. Grid domain used to simulate the cylindrical jet cup.
`
`transition from one modelto the other was used between 75 and 85%
`
`Fig. 7 shows severalstill shots from the video taken of PSRI's 7.6-
`cm (3-in.) diameter cylindrical jet cup with a 0.48-cm (0.1875-in.)
`nozzle diameterat the three jet velocities tested. A significant portion
`of the material in the jet cup remained stagnanteven at a jet velocity
`of 183 m/s (600 ft/s). Only at a jet velocity of 274 m/s (900 ft/s) were
`all the particles observed to be in motion (not shown in Fig. 7).
`However, this was not an ideal case as the material was blown from
`the cup into the disengagement section after a few seconds of
`of close packed loadings. Details of this can be found in Snider[8].
`operationat this jet velocity.
`Similar particle stagnation was found with the PSRI jet cup having
`Barracuda™ 10,0 uses an algebraic gas phase turbulence model.
`the smaller 0.24-cm (0.0938-in.) jet inlet diameter, as shown in Fig. 8.
`Although this is a rudimentary method for modeling gas phase
`turbulence, Derksen et al. [9] have shown that for high loadings, the
`Here, the extent ofthe stagnation region was observed to be higher. At
`gas phase turbulence is dampened and is not a major contributorto the
`76 m/s (250 ft/s) only a few particles were observed to be in motion.
`Fig. 9 shows thestill shots of video taken of the Davison jet cup
`general hydrodynamics.
`with 5 g of FCC catalyst material. The Davison Jet Cup did better in
`Fig. 6 shows the grid domain used to study the PSRI jet cup. Only a
`portion of the disengagementsection was modeled. Any particle that
`terms of the amountofparticles in motion. At 76 m/s (250ft/s) jet
`reached the edge ofthe disengagementsection was considered as lost
`velocity, most of theparticles were still stagnant, similar to that found
`with the PSRI jet cup having the smaller nozzle size (the same size as
`to the domain. Jet cup designs were modeled at near ambient
`the Davisonjet cup). At 137 m/s (450 ft/s), only 30% of the particles
`conditions with a temperature of 25°C and an initial pressure of
`104,771 Pa (15.3 psia) and a feed pressure of 172,368 Pa (25 psia). At
`were stagnant. At 183 m/s (600 ft/s), most of the material was
`these conditions,
`the air density and viscosity were 1.18 kg/m?
`observed to be in motion in the Davison jet cup whereas not all the
`(0.07 lb/ft?) and 0.000018 kg/m/s (0.000012Ib/ft/s), respectively.
`particles were in motion in the larger PSRI Jet Cup.
`Particle properties were based onatypical equilibrium FCC catalyst
`Using 10 g of FCC catalyst powderin the Davisonjet cup did reduce
`powder discussed above and with the particle size distribution
`the size of the stagnant region, as shown in Fig. 10. With a jet velocity
`shown in Fig. 2. The particle density was assumed to be 1492 kg/m*
`of 74 m/s (250 ft/s), 56% of the material was stagnant. At 183 m/s
`
`Stagnant Region
`
`
`Stagnant Region
`
`76 m/sec (250 ft/sec)
`
`137 m/sec (450 ft/sec)
`
`183 m/sec (600 ft/sec)
`
`Fig. 7. Still shots from a video of the 7.6-cm (3-in.) ID cylindrical jet cup with a 0.48-cm (0.1875-in.) diameter nozzle.
`
`——_
`
`76 m/sec (250 ft/sec)
`
`137 m/sec (450 ft/sec)
`
`183 m/sec (600 ft/sec)
`
`Fig. 8. Still shots from a video of 7.6-cm (3-in.) ID cylindrical jet cup with a 0.24-cm (0.09-in.) diameter nozzle.
`
`4
`
`4
`
`

`

`R. Cocco et al. / Powder Technology 200 (2010) 224-233
`
`Stagnant Region
`
`
`
`
`
`76 m/sec (250 ft/sec)
`
`137 m/sec (450 ft/sec)
`
`183 m/sec (600 ft/sec)
`
`Fig. 9. Still shots from a video of the 2.54-cm (1-in.)ID Davison jet cup with 5 g of material.
`
`76 m/sec (250 ft/sec)
`
`Fig. 10. Still shots from a video of the 2.54-cm (1-in.) ID Davison jet cup with 10 g of material.
`
`Based on the performance of the standard and PSRIjet cups, the
`new cup designs shown in Fig. 5 were tested. The first new design
`tested was the angled jet cup. The angle essentially removed the
`stagnant region observed in the cylindrical jet cup. From the results
`with the cylindrical jet cup, a mean jet width of 1.5-cm (0.6-in.) with
`an angle of repose of ~ 45° was observed. Thus, the angled jet cup was
`designed with a similar wedge(angle of “repose” of 45° and jet width
`or gap of 1.5 cm (0.6 in.)) removed from the PSRIjet cup.
`The results of the angled jet cup design are shown in Fig. 13. As
`with the standard PSRI cylindrical jet cup, 100 g of material was tested
`atjet velocities of 76, 137, and 183 (250, 450, and 600 ft/s). Although a
`significant improvement in particle mobility was observed, there
`were still regions of stagnant particles at the lowergas velocities. At
`
`whereRis the radiusofthe jet cup, h is the heightof the highest point 76 m/s (250ft/s), 14% of the material was observed to be stagnantat
`of the wedge and @ is the angle of repose of the wedge, as shown in
`the bottom of the angled jet cup. At 137 m/s (450 ft/s), less than 10%
`Fig. 11.
`ofthe material was stagnant. At highervelocities, most ofthe material
`was observed to be in motion.
`With Eq. 1, the amount of material observed in the cup was
`calculated and compared ona relative basis. As shown in Fig. 12, the
`stagnant material was significant at 76 m/s (250 ft/s) and 137 m/s
`(450 ft/s) for all cups investigated. This wasalso true for the PSRI jet
`cups even at 183 m/s (600 ft/s) jet velocities. More than 50% of the
`bed was stagnant. The Davison jet cup did better at higher velocities,
`but to ensure thatall the particles were in motion duringtesting, a jet
`velocity of 183 m/s (600 ft/s) or higher needs to be employed.
`However,often this is too high of a jet velocity for realistic attrition
`testing. Such high velocities may not provide significant attrition
`results as particles are more likely to be in the disengagementregion
`than in the jet cup.
`
`
`
`40
`
`20
`
`*,
`
`
`
`%ofStagnantMaterial
`
`~~
`
`50
`100
`150
`200
`250
`Jet Velocity, m/sec
`
`7.6 cm ID Cup, 0.47 cm ID Nozzle, 100g Material
`7.6 cm ID Cup, 0.24 cm ID Nozzle, 100g Material
`2.54 cm ID Cup, 0.24 em ID Nozzle, 5g Material
`
`2.54 cm ID Cup, 0.24 cm ID Nozzle, 10g Material
`
`Fig. 12. Amount of stagnant material as a function of gas velocity measured in PSRI's
`cylindrical jet cups and the Davison jet cups.
`
`(600 ft/s), most of the material was moving, as observed with the 5g
`sample. At 274 m/s (900 ft/s), the material was lost from the jet cup
`after only a few seconds ofoperation.
`By assumingthat the stagnant material in the jet cups resembles a
`cylindrical wedge, as shown in Fig. 11, the volume of stagnant
`material visually observed was calculated by measuring the height
`and angle of the wedge with the expression,
`
`(3Sini6]—36Cos||—Sin*(6
`y — DR:
`> T—Cosid)
`
`1
`)
`
`
`
`Fig. 11. Estimating the amount of stagnant material in the cylindrical jet cup.
`
`5
`
`

`

`R. Coccoet al, / Powder Technology 200 (2010) 224-233
`
`229
`
`Small Stagnant Region
`
`183 m/sec (600 ft/sec)
`
`
`
`No Apparent Stagnant Region
`v a
`
`76 m/sec (250ft/sec)
`
`137 m/sec (450 ft/sec)
`Fig. 13. Still shots from a video of the 7.6-cm (3-in.) ID angle jet cup with 100 g of material.
`
`76 m/sec (250 ft/sec)
`Fig. 14. Still shots from a video of the 7.6-cm (3-in.) ID dual entry jet cup with 100 g of material.
`
`183 m/sec (600 ft/sec)
`
`The dual entry jet cup added anextrajet to the PSRI cylindrical jet
`cup. To use the samegasflow rate as with the cylindrical jet cup, the
`nozzle diameters for the dual entry jet cup were reduced to 0.24-cm
`(0.09375in.), which corresponded to one ofthe standard PSRI jet cups
`using the similar smaller nozzle diameter(Fig. 8).
`Fig. 14 shows thestill photos from the video of the dual entry jet
`cup with nozzle gas velocities of 76, 137, and 183 m/s (250, 450, and
`600 ft/s). The photographfor the 274 m/s (900 ft/s) case is not shown.
`At 76 m/s (250ft/s), more than 20% of the material was observed to be
`stagnant. However,unlike the cylindrical jet cup where the stagnant
`material formed a cylindrical wedge on one sideof the cup, the dual
`entry cup resulted in a conical-shaped stagnantarea in the center of
`the cup, as shown in Fig. 15. At higher velocities, this conical region
`reduced in size but was still present even at 274 m/s (900 ft/s). Forall
`jet velocities, the dual entry jet cup performed worse than the angled
`jet cup, but was still significantly better than the standardcylindrical
`jet cup.
`The conical stagnant region observed at jet velocities of 76 and
`137 m/s (250 and 450ft/s) was measured in terms of angle of repose
`and the gap betweenthe stagnant material and the jet cup wall. The
`
`Small Stagnant Region
`
`Fig. 15. Bottom view of dual entry jet cup at 137 m/s (450 ft/s) gas jet velocity.
`
`average gap width was determined to be 1.25-cm (0.5 in.) with an
`angle of repose of 22°. From these data, the dual entry jet cup was
`modified with a conical insert similar to that measured as the stagnant
`region noted above. This modified dual entry jet cup is referred to as
`the dual entry with insert jet cup.
`Fig. 16 shows still photos from the video taken for the dual entry
`with insertjet Cup study. This modified cup performed better than the
`original dual entry jet cup. However,its performance mirrored that of
`the angled jet cup.
`The last jet cup design investigated was the conical jet cup. The
`conical jet cup expanded from a 3.8-cm (1.5-in.) diameter at the
`bottom to a 7.6-cm (3-in.) diameterat the top. Thegas jet still entered
`tangentially at the bottom of the cup with the same inlet diameter of
`0.47 cm (0.1875in.) as that used for the 7.6-cm (3-in.) cylindricaljet
`cup. Nearly all the particles were in motionin the conical jet cup at gas
`velocities of 76, 137, 183, and 274 m/s (250, 450, 600, and 900 ft/s)
`without entraining into the disengagement zone, as shown in Fig. 17.
`At 76 m/s (250 ft/s), 95% of the material was in motion.
`Fig. 18 compares the percentage of stagnant material for each of
`the jet cups examined. All the new jet cups examined showed
`significantly smaller stagnant regions than thatof the cylindrical jet
`cup with the 0.48-cm (0.1875-in.) diameter nozzle. However, the
`conical jet cup measurements were significantly better than the other
`alternative jet cups tested.
`
`3.2. CFD study
`
`Fig. 19 shows the CFD results from simulating the standard PSRI
`cylindrical jet cup with 100 g of FCC powder. The simulation was for
`the operating conditions of 137 m/s (450 ft/s) jet velocity at 25 °C and
`104,771 Pa (15.2 psia). The jet cup diameter was 7.6 cm (3 in.) witha
`nozzle diameter of 0.48 cm (0.1875in.). As shown in Fig. 19, most of
`the particles in the cup remained stationary. Simulation times were
`limited to the first 5 s of operation. Little change was measured in the
`hydrodynamics after the first second of operation.
`Fig. 20 showsthe particle velocities in the bottom regionof the jet
`cup for the same simulation shown in Fig. 19. The gray-scale bar
`indicates the particle velocity where blue represents low particle
`velocities and red represents high particle velocities. From Fig. 20,
`
`6
`
`

`

`230
`
`R. Cocco et al, / Powder Technology 200 (2010) 224-233
`
`76 m/sec (250 ft/sec)
`
`137 m/sec (450 ft/sec)
`
`183 m/sec (600ft/sec)
`
`Small Stagnant Region e
`
`
`76 m/sec (250 ft/sec)
`Small Stagnant Region
`
`Fig. 16. Still shots from a video of the 7.6-cm (3-in.) ID dual entry jet cup with insert with 100 g of material.
`
`Fig. 17. Still shots from a video of the 3.8-cm (1.5-in.) bottom ID by 7.6-cm (3-in.) top ID conical jet cup with 100 g of material.
`
`several key features are apparent. First, the “stagnant” region is
`prominent with most of the particles being shown at the near zero
`velocity magnitude. The few particles that were moving had a
`trajectory of an inclined, ellipsoidal path from the bottom to the top
`of the jet cup and back down again. Second,particle velocities were
`only ~ 12% of that for the entering gas jet velocity of 137 m/s. At room
`temperature, the maximum particle velocity was 17 m/s (55.8 ft/s).
`Barracuda™ models particle trauma with a wall segment via the
`expression
`
`wv =m(ven)?
`
`(2)
`
`where m is the massofthe particle, v is the particle velocity crossed
`with the outward normal vectorn of the wall. Parameters a and b are
`user-defined. For this study, both were set to one so that trauma
`corresponded to particle momentum.Particle trauma in the cylindri-
`caljet cup was limited to a few select particles, as shown in Fig. 21. The
`70
`
`Material
`%ofStagnant
`
`Jat Velocity, m/sec
`
`Standard Cylinrical
`Angled
`Dual Jet
`Duat Jet with Conical Insert
`
`Conical
`
`Fig. 18. Amount of stagnant material measured in all the 7.6-cm (3-in.) diameter jet
`cups tested including the new conical jet cup.
`
`majority of the particles did not appear to have any stress with the
`wall during the 5 s simulation time. Maximum trauma was measured
`to be 0.23.
`Barracuda™ provides wall impact measurements as the numberof
`hits in the cell domain bordering the wall boundary condition. A 5-s
`simulation resulted in 48,100 hits (with clouds not particles). In
`addition, a secondary impact with the wall was detected further down
`the particles path. Fig. 22 showsthe wall impact for the cylindrical jet
`cup simulations. These results suggest that in a jet cup, particles tend
`to collide (or impact) with the wall instead of grazing the wall.
`As with the cold flow observations, the CFD model predicted the
`stagnant regions in the PSRI jet cup. This suggested that particle
`concentration,particle velocity, wall impact, and trauma might be good
`metrics forjet cup design. Thus, a Barracuda™ model was employed also
`for the new conical jet cup design using the same metrics.
`Fig. 23 shows the results of the CFD modeling for the conical jet
`cup. Nearly all the material was in motion at a gas inlet velocity of
`
`137 m/s (450ft/s). Simulations also showed that with the exception
`
`Fig. 19. Simulations of the standard cylindrical jet cup with FCC catalyst powder.
`
`7
`
`

`

`R. Cocco et al, / Powder Technology 200 (2010) 224-233
`
`231
`
`3.0 Seconds
`
`aay
`
`Fig. 20. Simulation results for particle velocities for the standard cylindrical jet cup with
`FCC catalyst powder at a gas jet velocity of 137 m/s (450 ft/s).
`
`of absent stagnantmaterial, the solids hydrodynamics were similar to
`the few particles that were moving in the cylindrical PSRI jet cup. The
`particle trajectory followed an inclined, ellipsoidal path from the
`bottom to the top of the jet cup and back down to the bottom.
`Similarly, the maximum particle velocity was approximately 12% of
`the gas inlet velocity. The maximum particle velocity was measured at
`17 m/s (56ft/s), similar to that observed for the cylindrical PSRI jet
`cup. Fig. 24 shows the particle velocity at the bottom of the conical jet
`cup.
`The most revealing of the CFD simulations was the particle trauma
`results. Unlike the particle trauma shown in Fig. 21 for the cylindrical
`PSRI jet cup, most of the particles in the conical jet cup experienced
`higher trauma, as shown in Fig. 25. Maximum particle trauma was
`measured at 1.4 for the conical jet cup compared to 0.23 for the
`cylindrical PSRI jet cup. Thus, not only were more particles exposed to
`trauma, the magnitude ofthe trauma was higherin the conical jet cup
`for at least a few of the particles. Fig. 26 shows the particle trauma for
`the bottom portion ofthe conical jet cup.
`Hence, CFD simulations supported the observations of the cold
`flow study. Most of the particles were stagnantin the cylindrical jet
`cup. Simulations of the conical jet cup showed that more particles
`were in motion but with similar maximum particle velocities. The
`
`
`
`Maximum Trauma = 0.23
`
`3.0 Seconds
`Wall a
`Impact y
`1720
`
`i
`
`NN
`
`1550,/
`
`
`-172
`
`-0
`
`Fig. 22. Simulation results ofwall impact in terms ofcumulative number of hits with the
`wall for the standard cylindrical jet cup with FCC catalyst powder at a gas jet velocity of
`137 m/s (450 ft/s).
`
`particle trauma indicated more particles may be experiencing more
`wall attrition and a higher magnitudeofthatstress.
`
`3.3. Attrition study
`
`Attrition loss rates collected from the secondary cyclone of the
`fluidized bed cyclone attrition unit shown in Fig. 4 were compared to
`the attrition indices obtained from the cylindrical 7.6-cm (3-in.)
`diameter jet cup for several lots of a catalyst material labeled Lots A
`through F. The only difference in the catalyst lots was particle
`strength. Thus, different attrition rates or attrition index numbers
`were expected. As shown in Fig. 27, the loss rates for the materials due
`
`10 Seconds
`Solids. Ji ERiy
`Volume ee
`sexe
`i
`;
`Fraction
`0.627
`
`0.584
`
`0.501
`
`0.428
`
`0.376
`
`0.313
`
`0.251
`
`Fig. 21. Simulation results of particle trauma with wall segments for the standard
`cylindrical jet cup with FCC catalyst powder at a gas jet velocity of 137 m/s (450 ft/s).
`
`Fig. 23. Simulation results of solids volume fraction (side view) in the conical jet cup
`with FCC catalyst powderat a gas jet velocity of 137 m/s (450 ft/s).
`
`8
`
`

`

`R. Cocco et al. / Powder Technology 200 (2010) 224-233
`
`Zaz
`
`10 Seconds
`
`
`
`20.8
`18.7
`16.6
`14.5
`12.5
`10.4
`8.3
`6.2
`4.2
`
`-24
`-0
`
`Maximum Velocity = 17 + 11.2 m/sec (56 ft/sec)
`
`Fig. 24. Simulation results of particle velocity (bottom view) in the conical jet cup with
`100 g FCC catalyst powder at a gas jet velocity of 137 m/s (450 ft/s).
`
`to attrition in the cyclones did not correlate with the corresponding
`cylindrical PSRI jet cup results. Thus, even qualitatively, the attrition
`results from the standard jet cup may notbe applicable to a large
`fluidized bed unit despite the popularity of this method.
`In contrast, the results from the conical jet cup were found to be
`comparableto the attrition loss rates from the fluidized bed cyclone
`attrition test unit. The Al<20um data from the conical jet cup
`correlated well with the attrition loss rate data. However, complete
`agreement was not obser