MAINTENANCE AND RELIABILITY How FCCU trickle valves affect catalyst losses Here are the fundamental forces that affect their operation and the pros and cons of various types D. F. SHAW and R. E. WALTER, Carma: catalyst losses from fluid catalytic cracking unit plugged diplegs, often believed to he caused by one oF ‘more malfunctioning trickle valves. Sometimes units are sha down and catalyst unloaded to “clear” a plugged dipleg. This, article reviews che fundamental forces that affect trickle valve opening and closing for both positive- and negative-pressure cyclones. These forces are caused by the particular eikle valve design desils, weight of the flapper or counterweight and catalyse level inthe dipleg. The featutes of various trickle valve configurations will also be discussed, including the advantages and disadvantages for various applications, as well as measures taken to improve theie performance and erosion resistance. Background. Fluid solids processes crew late huge quantities of solids to achieve feed stock conversion to desired produ vacuum gasols to gasoline in an FCCU). Catalyse circulation rates can range from 10 to more than 100 tons/minute between the reactorand the regenerator vessels. Cyclone systems typically consist of pri- smary- and secondary-stage cyclones in both the eactorand regenerator to recover the cata- lyst solids ftom the reactor products and che regenerator flue gases. As more catalyst is col> lected with the eyclone systems, fewer process ing problems occur the product recovery circuit, and he potential for solids emisions inthe regenerator fue gas seem is lowered The role ofthe trickle valve has become more important as FCC technology has advanced from primarily bed cracking to shore contact-time (SCT) riser reactor. Inc tal concems have also driven refiners to minimize catalyst losses during normal operation, startup and upset situations. FCCUs, did not originally use trickle values on the cyclone diplegs. This, ‘worked fine once a bed level was established above the level of che dipleg discharges. The bed acts as a seal for the diplegs and low loses ean be achieved. However, high catalyst losses were observed, during unit startups before bed was established, or during upsets, ‘where the bed levels were reduced below the level ofthe dipleg Addition ofa trickle valve ro the end of che dipleg was su fal in maintaining low catalyst loses during unit starup before bed levels are esablished and during upset sicuations. The erickle @ The purpose of trickle s. valves that are inserted into the bed is primarily to minimize catalyst losses during startup and other periods when the bed level is below the valve. Engineering, inc., Rockaway, New Jersey valves also help mainain the dipleg seals when negative-pressure reactor cyclones ate utilized in SCT or other operations that dis- charge the catalyst ino a dilute phase ‘While overall experience with trickle valves has been good, now there was a mechanical device chat adds to the possible list ‘of reasons for increased catalyst losses from either the reactor or regenerator. While trickle valves can cause high catalyst loses, che following describes how a properly designed trickle valve should not bea primary cause. Trickle valve purpose. Simply puts the purpose of the trickle valve on the discharge end of che dipleg is similar to that ofa check, valve in a process line in that che flapper prevents catalyst or gas flow up the dipleg. A cyclone will not work properly with gas flow up the dipleg. Upllow is prevented ether by pressure differential, a mechanical device or bya massive downflow of solids. Fig 1 shows how a typical cyclone dipleg works. Cyclones ‘ean operate a positive oF negative pressures relative wo che vesel Also, che diplegs ean dis- charge either into a bed or into due phase. All chese factors affect how the trickle valve functions and what itis expected to do. ‘The purpose of trickle valves that are inserted into the bed is primarily o mi see catalyse losses during startup and other periods when the bed level is below the ‘valve. Typically, rickle valves that operate in the bed have a shroud arrangement to prevent a gas bubble from ‘entering the dipleg, Few problems are typically experienced with these trickle valves if they are immersed in a wel-luidized bed. Also, posiive-pressure cyclones are not expected to experi- ‘ence flow up the diplegs except during stareup or other low-rate periods. The diplegs are inserted in a bed with positive-pressure cyclones in an attempt co create a backpressure and minimize gas flow down the dipleg. Excepr for certain coupled cyclone systems in reactors, eyelones in FCCUs ypically operate ata negative pressure relative to the vessel. Fig. I shows a rypical negative-pressure cyclone system with presure inthe vessel as Pv and the pressure inthe eyclone as Pe Typically Pris less than Pp by the amoune of presure drop through the cyclones. This differential pressure tends to push the flapper closed. To balance this closure force, a catalyst evel must build up, inthe dipleg, creating a“hydrostatic” force 0 open the flapper. TheSPECIALREPORT weight of the flapper (or counterseights) creates an additional closure force thae must also be overcome by the catalyst level Trickle valve types. Ac leas five dffer- enc types of trickle valves are widely used in FCCUs, Typical considerations for select ing valve type include: “Is the eyelone in negative- or positive- presse service? + What isthe eyelone design pressure or required catalyst level inthe dipleg? + Will the dipleg discharge inthe dilute phase orin che bed? * Whar is the mass flux rate down the dipleg? + How large should che dipleg diameter * Whar catalyst level is required wo oper- ate the valve? Table | summarizes some ofthe pros and cons for the various trickle valve types. MAINTENANCE AND RELIABILITY TABLE |. Pros and cons of trickle valve types Conventional iter body tickle valve Pror + Relatively simple design, inexpensive + Depends on rai to ose + No tight dearance to seize o bind thus apper operation nat affected by coke formation or refractory spals Free dining ‘+ Can operate in bed anit phase ‘+ Support components subject to minimum wear + Flaper blockage requires a large chunk of react o coke ons “Large vertial opening approximately 20) » When flopper rotates. ges can emer the cle vain elute hase peat » Catalyst eight in peg may not roach the top ofthe opening for large valves andor law cyclone differential pressure + Relatively sal seating forces due to sal hanging angle Miter trickle valve with refractory lined flapper and blankouts Pros + Sim to conventional miter tickle valve + Developed to address vapor leakage by larkng portion and acing eractry 0 help contol + Increased flapper weight ceases seating ores and stably. ons + Similar to conventional miter trickle valve + Provides aide seating surface + Excesive blakeout can cause operating problems ‘Conventional tickle valve with elbow Pros + Siilr to comentonal ikl vale except: Reduced height ofvricl opening beter controls gs leakage atop of appr (5 D) > Erosion estat reacties apie to ete surfaces can contol erosion » Increased flapper weight increases seating ‘ences and stably cons « Simlrto conventional miter tickle ve. Us patent 4,996,028 tion on the horizoncal flapper, or any dynamic Forces associated with Forces acting on a counterweighted trickle valve. The forces acting on a counterweighted trickle valve are relatively straightforward. The closure force is established by hanging a tem- porary weight on the center ofthe Flapper place and moving the counterweight along a moment arm to balance thae weight (includ- ing the flapper plate). Based on calculations itis relatively easy to determine the equivalent hydrostatic pressure chat is required to just ‘open the flapper (i. offset the counterweight force). I should be noted that these calculations ignore any catalyst or coke accunmula- che eaalyst momencum. Some allowance based on experiences ao included in the setings for ches considerations. The pressure due co eatalys height requited to open agaist the counterweight (plus flapper is simply: P= W Changing weightdipleg area (based on 1D) ‘The actual pressure may vary depending on the sickle valve manufacturer; however, this is typically in the 0.1 co 0.2 psibd aes range. The catalyst level in che dipleg must balance che counterweight plus the cyclone system differential pressure. For a secondary cyclone, the ypical accumulated cyclone differential pressure (Pr Pe) is on the order of 1.5 psiso thatthe coral closure pressure is on the order of 1.6 co 1.7 ps Fora catalyst density of 29 Ibi a catalyst dipleg level of almost 8 fis requiced to just ‘open the valve against the eyelone differen- tial pressure and the counterweight Forces acting on a miter body trickle valve. The forces char act on and copen a traditional miter bend tickle valve are also quite simple but are more com cated t0 calculate. The following addresses the forces and moments that control the ‘opening and operation ofa typical cyclone trickle valve installed in the dilute phase of a fluidized bed reactor oF regenerator using negative-pressure primary or second- ary cyclones. For diplegs chat operace in a bed, there are additional dynamic forces due to the eurbulene nature ofthe bed thar are not considered here. Also, for diplegs submerged in a bed, the hydroseatic head MAINTENANCE AND RELIABILITY ‘TABLE 1. Pros and cons of trickle valve types (continued) Horizontal counterweighted trickle valve Pror ‘ Adeeses miter ck valves lage vera opening concer by efisinating and making seat heizontal vith a postive seating force «Intended pray fr ut phase serie cons More complex design wth counterweight, beatings, tight earances, et, and greater fabrication contol required + May have ations in bed sevice + May be vulnerable to catajstocoke buildup on Amina) ‘ie sia) horizontal pat in srvice «+ Refactor or coke chunks may hang up onthe seat resulting in flapper hanging open * Cok can form on the counteraight increasing the seating forces or on bearing surfaces binding ‘operation—eithar open or sed. Fig.5 Courerwoignt ‘Angled, counterweighted wickle valve Pros : + Developed to adress the cancers with ahoizontl rr Counterweighted plate holding up debris—angle of flapper and { seat tend 0 clear ou any debs t ‘cons «Sina to hovieortalcounterasghted wiki valve except catalyst, and debris are not expected to bull up on Napper plat ‘of the catalyst bed height above the tickle valve discharge must be accounted for. This calculation basis covers typical trickle valves as shown in Fig, 2. ‘The calculations are based primarily on the forces acting on the trickle valve flapper plate as it begins wo open. The model for the cquationsis based on the fact that the flapper plate is supported a the top with rwo hinge rings, and thar che place tends to rotate around dhe hinge rings as it opens. As opening forces act on the flapper plate, the rotation around the support hinges causes the opening atthe bortom of che seat to increase more than the opening a che top of the plate. Of course the hinge ring is not an absolute pivor point and there can bbe some small lateral translation ofthe plate This small ransation, however does not havea significant affect onthe calculations. Also, these calculations do nor include any dynamic force that will tend to further open the trickle valve once it has opened! and catalysis flowing. See Fig 7 forthe terminology used in the equations fora miter tickle valve. Trickle valve forces. An opening force arises from the pressure acting on the flapper plate (area atthe valve opening) due to the catalyst hydrostatic head inthe dipleg. This pressure cts horizon- tally on the flapper plate surface. For simplification, we take the calculated pressure atthe middle of the trickle valve opening and spread this over the entre area ofthe elliptical trickle valve open- ing. This assumption is valid as longas there isa reasonable catalyst level inthe diple, ‘A net closing force occurs due to the differential gas pressure that acts across the lapper plate (i. the external pressure inside the vessel, Po, minus the pressure inside the cyclone hody, P). This net differential pressure only acts on che valve discharge opening area, and tends to close the valve in a negative-pressure cyclone system since the vessel pressure is higher chan the pressure inside 78 | May 2007 HyoROCARGOWFROCESSING the eyclone, Since the tickle valve flapper plate hangs from ies support ring, ava small angle (approximately 3° from vertical), the weight of the flapper plate creates a seating force against the valve opening, his seating force is determined by che moment created by the horizontal offset, y, of the flapper place mass from the vertical centerline chrough the hinge ring support. Dynamic forces are caused by the momencum of the filling sgasfeatalys stream in the dipleg impinging on the flapper plate and have not been included. Instead, the calculations determine the condition where the trickle valve just opens and any dynamic forces will tend to furcher open the valve. As the valve opens further, y increases, thus inereasing the closure forces until che ‘opening forces! moments are in balance with the closing forces! Calculating trickle valve moments. To calculate the effec of the trickle vale flapper weight, eralys head in the dipleg, and the differential pressure beoween the vessel and inside the eyelone, ‘one takes moments around the flapper plate vertical support point (Point Cin Fig. 7) For static equilibrium che summation of moments, Me, around Point C muse equal zero. Lme=0 o ‘Ac the point of opening, the moment due to pressure on theSPECIALREPORT inside ofthe flapper (Pe + atc catb head inthe dpleg) muse equal the closure moment duc to the weight ofthe trickle vale flapper and the pressure in the vessel. Using the moment arm lengths showin in Fig 7, the following ae determined: ‘M, — opening moment ~ moment due to catalyst head inthe dipleg (counterclockwise moment), in-Ib: bx x A = xP x bx. Q 1a aa a” inlA) where: p= catalyst density in che dipleg bifi? talyse height in dipleg, fe trickle valve opening area, in: moment arm length, in, (se Fig. 7) = dipleg ID, i. ‘= miter bend angle sickle valve lapper seat angle angle of llapper Seat on miter = ¢ — B) (Fig. 7) ‘Mz ~ Moment due to Pe (counterclockwise moment), in Ib: Mp7, =P 4° sin(A) oy My ~ Moment duet Pe(okvse momen in Myx Lied, = bev ® 4” Sn(A) ‘My ~ Moment due o wickle valve flapper weight (clockwise moment, inclb: =} Pressure ibe ineyelone ‘of Flapper i + Py= pressure eed plate oo |) angie A= ange) minus angle 6) = | EN oct Satie “=a ee MAINTENANCE AND RELIABILITY M,=yW, 6) where: W= flapper weighs, Ib {y= horizontal oflet for wickle valve flapper concer of gravity (CG) from the hinge ring due to angle, &, in. (y= stan 8) 22= distance from Mapper support hinge to flapper CG. Considering clockwise moments 2s postive and counterelock- ‘vise moments as negats -M,~M,+M,+M, © therefore, | xP xB xO || trex Bx PR | 144" 4" Gin(A) 4 an A) 2,0! wet P| ww, 2 [» 4 Za) ee ve Combining the pressure moments to get differential pressure [Bh F fot Zialim-n+ yW,=0 @) Simplifying this equation we get: By PVP a(n pe) ay (xd Fo mol ») Ww, 8) Solving for the catalyst head in the dipleg required 10 open the tice valve “(A yaw, Jenna 0) ® | | agate rte aor [Sra asin patterns on flappers esemble an inverted peed 0 | MAY 2007 #OROCARBON HROCESSING horseshoe ContinuedSPECIALREPORT For an elbow-type valve shown in Fig. 3 the equation would be moe a Jorn] ° ay ‘The following is an example of the moments calculated for a typical 6-in. miter bend trickle valve. Fora typical 6-in. valve on a secondary dipleg (without refrac- tory lining on the flapper): ee Wy= 17 Ib (4 in, thick X 8 in, X 15 in, no refractory) Po Pen 15 pal p= 29 bffe “The overall moment balance is determined using Eq. 6, —M, + (MyM) + My vidual moments forthe example one deter- mines the moments as follows: Starting with Eq 2 for the trickle valve opening moment: 2 my 6 M,=9.9bx 2x2 x i= ag 32) ai M, =106.45 Combining Eqs. 3 and 4 co get the differential pressure M-M, nD A xex 4” sina) (Po Pe)=99%5 5) ans: M,—M, = 7923 inclb due to dilferential presure. To calculate the trickle valve flapper weight moment, start with Eq. 5: M,= 3, 392x17 6.7 inclb due w flapper weight. MAINTENANCE AND RELIABILITY Solving the equation for M). the opening moment: MM, =(M,—M,)+ My =7923+6.7 = 799 incl nce M, = 106.44 ‘Therefore, #=7.5 Effect of flapper weight not significant on closure forces. I: is important ro note that for normal operation, the catalyst height in the dipleg required for the pressure balance is ‘not materially aecred by the flapper weight forthe small 3° hang- ing angle. The weight effect of che flapper is ess than 196 of che total opening moment for the trickle valve flapper. If the flapper ‘weight is doubled in the previous example, the opening moment Increases (0 13.3 in.-lb and the catalyst height in the dipleg is increased by approximately 0.06 ft vo approximately 7.6 f. “Another impact considered is that during normal operation, the flapper is opened wider than 3°; therefore, che “weight effec!” is increased as the y distance increases. If we assume that the Aapper is operating a a 20" angle (large for a secondary cyclone, ‘but possibly typical fora primary) with a double-weigh« flapper, the corresponding “weight moment” increases to 87 in-—lb, and the required catalyst heighe increases to approximately 8.2 fi In this case itis seen that the weight effect ofthe flapper on total opening moment is still only approximately 10% of the pressure moment for a secondary cyclone. ‘The previous calculations demonstrate that flapper weight does not havea significant effect when compared to differential pressure on a trickle valve opening where the normal hang- ing angle is 3° from vertical. Ie should be noted, however, that increasing the weight has a poscive effect on flapper stability and the dipleg level since small pressure fluctuations are dampened ‘out by the Bapper weight ot changes in che catalyst level in the dipleg ‘The weight stability effect has also been reported in cold flow studies of light plastic flappers where the catalyst level in the dlipleg can fluctuate significantly because small pressure fluctua- sions cause erratic lapper movement. Cold flow studies with steel flapper plates simulating an actual tickle valve also report ‘hat increasing the flapper weight can stabilize the catalyst level, in the dipleg, especially as the flapper plate weight is increased due to adding refractory. 25; , , , 25 , : ‘Based on Gn miter tickle vane Basod on Gin. mir tickle valve with 0 in plat Rapper 20 | PrPo Real = | [Pere teat 025 12 Zistlos 12 | Os 25 Bos 2s 075 37 2 07s 37 10 50 Zit | 19 50 1574 2 1574 T ol oo a | El | Tan Tiseet tegre is ereZ dan ao tse Tape Fate ‘apper flapper 1 in. refractory 2 in. refractory ‘flapper flapper 1 in. refractory 2 in. refractory. ‘Flapper type Flapper type : 2 | May 2007 yyDROCHABON FROCESSNGSPECIALREPORT Ieshould be noted that cold flow studies that atificially increase flapper weight by adding weights on pegs should be considered suspect since the weights are generally offset From che lapper center ‘of mass and the moment arm effect is exaggerated. Simple mechan- ies demonstrate chat the flapper does not simply rorate around the Jhinges with the eccentric weight, bu rather the top of the flapper tends 0 open more than the bottom thus allowing more gases to ener through the opening a the top ofthe valve. Elbow-type trickle valves. Fig. 4 shows a patented clbow-ype trickle valve used in dilute phase service (US patent 4,996,028). The principal benefit ofthe elbow is thatthe vertical opening (O in Fig. ) is reduced by approximately 25%. Also, this design inereases the closure force since it adds weight co the Aapper with refractory on both sides. Typically the vertical open- ing fora 35° miter valve is approximately 2 D whereas the vertical opening for an elbow valve is 1.5 D. Based on observing erosion pateerns on trickle valve flapper places, ithas been determined that asthe flapper opens, gases from the vesel rend to enter the dipleg atthe top ofthe lapper. A prin- cipal advantage ofthe elbow valve and crickle valve with blank-out isha they reduce the quantity of gas tac enters the dipleg. Typical crosion patterns on flappers resemble an inverted horseshoe (Fig 8), Asthese gases enter the cop, they tend to cause flow disruption inthe dipleg and erratic dipleg levels Ie should be noted that the vertical opening for a typical 24-in, dipleg is 48-in. high with a miter valve. This height opening can be significant compared to a ‘ypical dipleg level ina primary cyclone. poeea ines BORGER Not for all. One for you: The Boerger Rotary Lobe Pump. ‘The compact design combined with the variety of sizes and ‘otor options gives the best optimum solution for almost ‘any targeted application of slid aden liquid. Thanks to ‘the elaborated engineering combined with oustanding ‘onstruction quality and the maintenance-in-place principles we achieve bes if cycle cost effectiveness. Makes our pump, your first choice. Big Gb Beings 2414625 hen Masta /Garmany woe punes.com Selec 17 at wovwydrocarbontrocessng comms MAINTENANCE AND RELIABILITY Effect of flapper weight on catalyst level. Fig. 9 illustrates the effect of trickle valve flapper weight on che dipleg catalyst level. Note thatthe eatalyst level required due to flapper ‘weight is given in inches of catalyst, therefore a 2-in, steel plate with I-in, refractory on each side requires only about 2 in. of catalyse 0 open Effect of seat angle on catalyst level. Fig. 10 shows how the seating force is increased with increasing seat angle. For ‘example, increasing che seat angle from 3° to 6° approximately doubles the seating forces. These forces are small relative ro the closure force due to the differential pressure as shown in the table. Based on this, itis suggested that consideration be increasing the seat angle to control leakage up to the secondary dipleg during startup, Possible causes of trickle valve malfunction. Possible ‘causes of trickle valve malfunction are: as leaking up the diplegs += Binding of bearings on councerweighted trickle valves + Chunks of refracvory/coke hanging flapper open Coke, refractory or debris between flapper and stop bars keeping flapper closed + Erosion of seating surfaces + Impingement from discharging solids in close proximity “These should be checked if itis suspected chat high catalyse losses are due to the trickle valve. HP ACKNOWLEDGEMENT evn and updated from a presntaton a the 2006 Spring AICHE westing (Onan, Florida (per 28170), Donald F. Shaw rete fem finan eserch ae Enh rng Coin 2000 after 37 years woting 3 mechanical 20 nee nthe design, startup, rep and troubleshooting oi mechanical equipment sod png syeers ured n refining an6 rant plans wou, Foo Febement. he was primal ‘espansbe for developing and implementing mecharical engneting technology for CCUs FLUID CORING and REXCOKING process, and for emerging processes ‘ach as gisoriquis. tr Shaw was coneveloper of CCU Sro-Centact-Time [SCP techosogy and averse numer SCT revamps nthe Key epanshiy wa arrow heii ac serce factors of FCUs. Mavis a 3 constant thih Ccnmagen Engen ie pray with idl process and other nee proces tedloges, roudgmechancal engineering Peat and conezptul den foe process emanation unis, commercitization of emerging techlogkes, un upgrades and voubeshootng, Ne Shaw has several patents eats o mechanical cemporantsotalas nFCCUs\e9,ScT rer ernnation arsvar, Cierbutr gril. regenerator vera sho vale aed econ ys Suppo" He has 8 degre in mechanical engineering fom Diesel Uniesty and an WS degre fehl engineering fom Steere ete of Tachlogy, Richard E, Walter ois for Carmagentngnearng, ne ara fis ade process engineer Snce aningCarmagen, he has bee Inve in developing the camera unt eeegne freer emerging fide sold chereate processes n alton to ding ‘onaltng on exeing FCCU opeatons fora mumber of mer ot refines Pr tojoring Carmagen Mr Waker teed fom Bann Reseac ae Engineering Cooter 37 yeas, marin theFCCU eld. While at ExorMebil his work nied rearing and deweping New innouatrs, eevee ropes fer ne unts or moseatios fr exstg unis, voudeshocting operation prob Inne wri, agretr training, risblty mproverant, se atesiment Uni ‘urnaround nspactons, unt statis, proces optimisation inane vashetins, tnd develoing FEU design eters, recedes and anaars We Wate heise {BS degree n chemical engnesing from Lafayette Colege and an MS degree {Granicaengneeng om NIT