James R. Murphy
Consultant
El Paso, Tex.
In the 50 years since the birth of fluid catalytic cracking, there have been so many changes in the appearance and design of the FCC unit that its forefathers wouldn't recognize their offspring.
These changes have been brought about by modifications in almost every element involved, including feedstocks, catalysts, equipment, and products. Additionally, a better understanding of the technologies, particularly fluidization, brought about major changes. This development of FCC unit designs was therefore a function of the improved knowledge of many factors affecting the process.
This article follows the evolution of fluid catalytic/heavy oil cracking designs, without regard to specific chronology, and concludes with a review of current and subjective future designs. A thorough discussion of the history of the catalytic cracking processes was presented recently. 1
THE FIRST
Standard Oil Co. (New Jersey), now Exxon Corp., put the first FCC , unit on stream in 1942 at its Baton Rouge, La., refinery (p. 41). The FCC process caught on in a big hurry, driven by the needs of World War II. M.W. Kellogg Co. alone built 40 units between 1942 and 1950.
A number of innovators combined their technologies and licenses under the "Rec. 41" agreement. Kellogg and Universal Oil Products Co. (UOP) were the primary designers along with major oil companies, Jersey Standard, Sinclair Oil (later acquired by Arco), and Shell Oil Co.
The other oil companies appeared to choose a design and stick with it. Gulf Oil Corp., Texaco Inc., and Standard Oil Co. (Indiana) chose Kellogg's "side-by-side" design. It was called side-by-side because the reactor and stripper were in one major vessel with the regenerator alongside.
UOP had most of its FCC units installed in U.S. Midwest refineries. Its initial side-by-side had an elevated regenerator and a low reactor-stripper. The elevated regenerator permitted low pressure and reduced air blower driver costs.
One unit at Phillips Petroleum Co.'s, Sweeny, Tex., refinery had two reactors. The second reactor had a deep bed to re-treat the gasoline made in the first reactor for aviation gasoline. UOP subsequently marketed a stacked unit wherein the reactor was located directly over the regenerator with a long riser from the bottom of the regenerator to the reactor.
Standard Oil Co. of California (now Chevron Corp.) chose Jersey Standards' Model IV with no slide valves. This gave it a low center of gravity and made it suitable for earthquake areas.
In the late 1940s, Kellogg developed the plug valve. The plug valve was placed in the direct line of flow rather than at right angles, as were the slide valves. It permitted a design which had the three major vessels in one vertical line-the Orthoflow design.
There were two types of plug valves: one, the closed plug used at the bottom of standpipes to control flow, the other, the open plug valve which allowed oil or air to flow through its center with catalyst flowing at controlled rates on the outside of the plug.
The first unit, the Orthoflow A, was started in 1951. It was similar to most designs with an oil lift to the reactor, then to a parallel stripper and down to a regenerator. It was followed in 1953 by the Orthoflow B which had its regenerator on top to lower utility costs. Feed was injected directly into the reactor bed, with no riser.
In 1961, the Orthoflow C was introduced, this design returned to the "A" configuration with the reactor above the regenerator. The innovation being all-riser cracking of fresh feed and bed cracking of recycle streams. The Texaco design employed a side-by-side configuration with similar riser cracking of fresh feed and bed cracking of recycle.
Although many disciplines were involved, the history of FCC designs should realistically be divided into two major eras: the synthetic catalyst era and the zeolitic catalyst era.
SYNTHETIC CATALYST ERA
In the early days of catalytic cracking, the catalysts were highly temperature sensitive. Regenerator temperatures were limited to 1,100 F., controlled in some instances by catalyst coolers. Rather elaborate water spray systems were used in the dilute phase of the regenerator.
In spite of all these efforts to limit temperature, a major concern was runaway afterburning. Because the CO2-to-CO ratio at 1,100 F. was about one, considerable quantities of CO were present in the dilute phase of the regenerator.
All that was needed was a hot spot above the CO ignition temperatures of about 1,150 F. and afterburning commenced, quite often reaching temperatures of 1,800 F. Spray water, cyclone quench, and flue gas line sprays were used extensively.
The catalyst coolers in use at the time are an example of poor engineering. The catalyst/air mixture was passed up through tubes at velocities of 20-30 fps.
The entrance effect in the form of a venturi brought the catalyst back against the metal tubes at a sharp angle. This action caused the tubes to soon erode and fail. The coolers would then be blocked off, and the FCC unit had to operate at reduced severity. Most catalyst coolers today operate with steam in the tubes, surrounded by hot catalyst fluidized at low velocities to give high heat transfer coefficients. 2
Another consequence of low regenerator temperatures was the low burning rate. Holding times in the regenerator were about 10-15 min compared to the 3-4 min used today.
The synthetic catalysts had equilibrium activities of about 25-35 D+L (described in those days as distillation plus loss) and there were not major differences in quality. Large refineries used mixtures of catalysts from the various suppliers.
Because of their relatively low activity, high recycle rates were used. The preferred mode of operation for maximum gasoline was low reactor bed height and high recycle rates. And as a result, the coke yield was about twice that of the single pass operations in use today.
Most FCC units were designed with feed and recycle meeting hot (1,050-1,100 F.) regenerated (0.6% carbon) catalyst at the bottom of a sloped and curved riser. The vaporized hydrocarbon streams traveled with the catalyst upward through a plate grid fitted with about 2-in. holes.
Total open area was sized to give about a 1-3 psi pressure drop to ensure good distribution. The pressure drop required was a function of the amount and density of the catalyst in the reactor. Practically every unit design introduced feed and a catalyst through a riser and a perforated grid (Fig. 1). The one exception was the Orthoflow B which introduced feed directly to a bed.
Because there was very little conversion in the risers of early units (about 15-20% was assumed for calculation purposes), little care was given to the design of the feed injection system. In most cases, it was a single 4-in. to 8-in. pipe located in the center of the riser and extending past the intersection of the regenerated catalyst standpipe. Some units that had vertical introduction sections used multiple feed injectors entering the riser at an angle of about 30.
Steam was added to give a nozzle velocity of 50-60 fps. These were the first of the multifeed nozzles widely used today. Other units had little or no risers. Such units were prime subjects for revamps to riser cracking in subsequent years.
The preferred mode of operation in the early days of catalytic cracking was minimum reactor bed height, low bed temperatures, and high recycle rates. Temperatures as low as 960 F. and recycle rates equal to fresh feed rate were common. As catalyst activity and stability improved, it was only natural that designers would go from minimum bed height to riser cracking.
The first Orthoflow C unit with riser cracking of fresh feed was a unit designed for reduced crude at Phillips' Borger, Tex., refinery. It went on stream in 1961. Some ten more Orthoflow C units were built, the largest (47,500 b/sd) being at Amoco Oil Co.'s Texas City, Tex., refinery. This unit currently runs 100,000 b/a.
ZEOLITIC CATALYSTS
In the mid-1960s zeolitic catalysts were first introduced. As might be expected, the operators were the first to become familiar with them and their requirements.
In the late 1960s, Gulf was considering a large (80,000 b/sd) FCC unit for its Alliance, La., refinery. Gulf had been using zeolitic catalysts in its existing units and, consequently, had carried out considerable associated pilot plant work.
All possible licensors were evaluated and it was concluded that, at that stage of development, the operators understood the requirements of zeolitic catalyst better than the designers. The Gulf FCC unit design was born.
With zeolitic catalysts came some large changes in design. Because practically all refineries had units designed for the lower activity amorphous catalysts, most action was in developing revamp designs for units with varying initial configurations. These revamps primarily involved removing recycle and going to a single pass operation at the same throughput.
Because recycle has a great influence on coke yield and air rate, a number of units, particularly those that supplied all their air through a lift line, needed to make some drastic changes. These involved a decrease in feed preheat and a large increase in fresh feed rate.
Today, we see FCC units running at fresh feed rates two to three times higher than the original capacity.
The 25 years since the birth of zeolitic catalysts has seen many interesting variations in riser design. The initial changes involved extension of the existing riser, regardless of whether it was curved or straight. The Orthoflow B units with just bed cracking required an open seal well and a folded riser to fit into the confines of the existing reactor.
REACTION SYSTEM
Today's reaction system is simple; it is mostly vertical pipe. But cracking gas oils or reduced crudes requires extreme care in the design of each section of the "riser" cracker.
For discussion, the "riser" should be separated into three major sections:
1. The initial section wherein feed is introduced, contacts and cools the catalyst, and is vaporized.
2. The intermediate section where the majority of the cracking is carried out.
3. The final section where the goal is to crack the remaining convertible material without destroying the valuable products formed previously.
The regenerated catalyst reaches the feed section at very high catalyst temperatures (1,300 F.), along with flue gas and other inerts. In some unit designs, the inerts are removed and replaced with steam to make downstream processing easier.
There are two widely different concepts in designing feed introduction systems. One involves prior catalyst lift or acceleration; the other uses introduction of catalyst to the feed section in its maximum free-flowing density. Those that prefer a lift gas claim advantages of pretreating catalyst to lessen the effect of metals, lowering of oil partial pressure, and ease of distribution of feed over catalyst.
Another advantage rarely mentioned is that the lift gas moves the feed out of the high-temperature mixing zone rapidly. Pilot plant data have shown that lowering the diameter of the feed zone, at the same feed rate and conversion, increased gasoline yield by 3%.
The disadvantage of gas lift is that gas rates above minimum fluidization velocity cause bubbles to be formed. Feed entering the catalyst-free bubbles is thermally cracked. Introducing catalyst at its maximum free-flowing density virtually eliminates the possibility of thermal cracking.
Catalyst flowing at high density requires a feed introduction system of high penetration. Nozzle studies show that a continuous liquid stream flowing in a flat, fan-like pattern from a number of nozzles provides the desired action. The catalyst particles are coated with feed as they flow through the flat spray and are simultaneously cooled. The intimate contact between feed and catalyst rapidly vaporizes the feed, which enters the catalyst pores and is cracked at the active sites. The increased vapors due to reaction and vaporization raise the velocity and lower the density of the flowing system (Fig. 2).
INTERMEDIATE SECTION
The function of the intermediate section is to continue the good contact between feed and catalyst, while avoiding excessive backmixing. This is achieved by using straight vertical risers operating at maximum velocity.
Maximum velocity is that which avoids excessive erosion. Because velocity is continually rising due to increasing conversion, the velocity limitation fixes a height where the riser is expanded. This limiting velocity has gradually been increasing as refractories and riser design have been improved.
Originally, a number of risers were hot wall, i.e., metal risers with external insulation. Currently, most risers are cold wall risers with about 3 in. of refractory that acts as both insulation and erosion resistance to the external metal wall.
The limiting velocity was the same as for cyclones originally, i.e., 60 fps; however, this has gradually risen as erosion in straight-vertical risers has not been a significant problem.
Some designers have provided for quench introduction at the downstream sections of the riser. 3 This permits achieving the desirable high-temperature effect on gasoline octane ratings without excessive temperature effect on gas make.
FINAL SECTION
The final section of the riser is the most critical because this is where the operator tries to get the last 5-10% of feed conversion without recracking the already-formed desirable products, mainly gasoline. A comment on this problem was made at the 1988 National Petroleum Refiners Association Q&A session on refining technology.,
Tom McCollough of Sun Refining & Marketing Co., Tulsa, reported: "At one of our units, residence time is less than 2 sec. This unit has a much higher gasoline selectivity than our other catalytic cracking units. By maintaining 70 MAT, and a 6 catalyst-to-oil ratio, we can achieve an 82% conversion at 980 F. reactor temperature with no indication of overcracking.
"The other catalytic cracking units within our company have long residence times, 3-5 sec, and experience overcracking when riser temperatures or catalyst activity are increased to any degree.
When I first started to design riser crackers in the late 1960s, the design contact time based on outlet mols was 1.8-2.0 sec. I was amazed when I encountered designs with contact time as high as 6 sec.
Many schemes have been proposed to avoid recracking of gasoline. These involve not only riser design but also cyclone systems, which limit further cracking while making a good separation of catalyst and vapor.
Riser modifications consist primarily of those that minimize the opportunity for recracking gasoline. This involves a design of the final riser section that is downflow to virtually eliminate slip-a so-called folded riser design. The riser at Sun, Tulsa, mentioned in the preceding, is a folded riser.
A complete downflow riser is considered the riser of the future. It is only reasonable to expect that a downflow riser that minimizes catalyst backmixing would also minimize the adverse effects of metals on catalyst and therefore be the preferred reaction system for cracking reduced crudes.
There have been many cyclone modifications which lower the time for undesirable cracking that normally occurs in the dilute phase of the reactor. 5 These include the vented riser, the direct connected cyclones, and the closed cyclones. All of these decrease the time the riser effluent remains in the dilute phase of the reactor, and thereby decrease aftercracking of desirable products.
As FCC unit design has progressed, the reactor changes have been those which progressively have lowered the opportunity for recracking of desirable products. Thus, we have gone from high bed height to minimum bed height; to upflow riser cracking; to partial upflow and partial downflow, "folded risers." It is expected that the next step will be all downflow risers.
An illustration of the improvements brought about in FCC unit design is the unit at Sun in Tulsa. This unit was built in 1948 for a feed rate of 13,500 b/sd and a hoped-for 2 year run. It was revamped in 1987 for 30,000 b/sd and 80 + % conversion. It is currently in the fifth year of a planned 5-year run with no shutdowns.
STRIPPER DESIGN
The early FCC units with low-activity catalyst required high recycle rates and high catalyst-to-oil ratios. In addition, the need for a tall reactor made it necessary that the stripper be short.
Stripping steam rates were as high as 6 lb/1,000 lb of catalyst circulation. Consequently, the major effort in stripper design was to avoid bypassing of catalyst through the use of baffles. just about every possible baffle was used; these included shed decks, disk and donut baffles, inverted angles, and even compartmented strippers.
Little attention was paid to design of steam distributors, most of which were located under baffles. Such designs allow steam to accumulate in a large bubble under the baffle until the pressure builds up sufficiently to allow the steam to escape around the baffle in a large bubble.
With the advent of zeolitic catalysts, major changes were made to reactor and regenerator design, but little modifications to stripper design.
The most important aspect of stripper design is the contact of catalyst with steam. This is best accomplished by introducing the steam through small openings to give small bubbles and excellent contact with the solids.
This was the fundamental concept in developing the highly successful multinozzle air distributors This concept is now being applied to distributor design for strippers.
We then had the situation where the vapors were entering the stripper in two phases: one, an easily displaced phase between particles, and two, a phase within the particles, which would require time to be desorbed and displaced.
To handle both problems, it was necessary to use two distributors, one near the entrance to remove the easily displaced phase, the other at the bottom. Using two distributors with many holes also allows sufficient open area for good catalyst flow.
Although little attention has been devoted to baffling in this discussion, it can be an important consideration in stripper design. Any vertical member in a fluidized bed provides a path for the solids to short circuit to the bottom of the bed. Annular strippers double the problem because they provide four paths rather than two.
The baffles required with good steam and catalyst distribution are considerably smaller than used in the past; just of sufficient size to direct the flow of catalyst away from the walls.
The other important consideration in stripper design to prevent bypassing and circulation problems is to provide perfectly symmetrical entrance and exit flow.
Because fluidized catalyst moves very slowly in a horizontal direction, it is important that it start out in a well-distributed manner. Units that do not provide this distribution initially require extensive baffling to achieve the desired distribution.
Strippers designed with symmetrical flow and multinozzle steam distribution have provided efficient stripping (5-6% hydrogen in coke) at steam rates as low as 1.5 lb/1,000 lb of catalyst circulation.
REGENERATOR
The design of the initial FCC unit regenerators was limited by the capability of the original catalysts. The natural or amorphous catalysts were not capable of maintaining activity at temperatures above 1,100 F. Control was obtained primarily by limiting air rate so that the flue gas oxygen content was essentially nil.
The CO2/CO ratio could be calculated from the equilibrium of the reaction CO2 + C = 2CO and ran about 1.21.4. Additionally, spray water systems were used in the dilute phase. Steam and sometimes steam/water sprays were incorporated in the regenerator cyclones and flue gas lines.
The greatest fear was runaway afterburn, which occurred when CO and O2 accumulated in the dilute phase of the large regenerators. A spark or hot spot sufficient to raise the temperature of CO above its ignition point would occur and temperatures as high as 1,800 F. would be noted. Spray systems would be used to counter the runaway afterburn.
Another means of controlling regenerator temperature was the catalyst coolers. These circulated catalyst downward from the regenerator which was then lifted by air at pickup velocities through the tubes of a steam/water exchanger and back into the regenerator.
Their problem was that the catalyst/air mixture, with its entrance effect, cut the tubes. After a few months service the leaking tubes needed to be cut out of service, requiring regenerator temperature control at lower feed rate or severity.
There were many types of air distribution systems including Christmas trees, rings with 2-3 in. nozzles, and perforated grids. Every possible combination of introducing air and catalyst to the regenerator was employed (Fig. 3).
It should be noted that the development of FCC designs was not a concerted effort by one group, but rather involved a number of companies testing their own but different ideas until about every possible design alternative was examined.
The predominant air distribution system was perforated grids. In some of the larger units, these grids were enormous, made into segments with elaborate supports.
Grids were built to pass air alone or to pass both air and catalyst. With their many holes, grids were good distributors of air. Their major problem was with the seals, the purpose of which was to prevent air or flue gases from leaking along the walls or internals. These seals would develop leaks, which would cause shutdowns.
The air rings progressed from just holes, to ferrules on top of rings, to nozzles with an orifice at its inlet to ensure good air distribution. The nozzle was of sufficient length for the venturi formed at its inlet to dissipate (Fig. 4).
As catalyst quality improved, temperatures increased to about 1,200 F. and carbon on regenerated catalyst was lowered to about 0.2%. CO burning occurred readily and a method of regenerator control using the temperature rise developed by CO burning was initiated.
A bypass valve around the main air supply was used to control afterburning. The bypass would adjust the trim air supply as determined by the temperature rise above the bed.
Some years after the introduction of zeolitic catalyst, Amoco developed complete CO burning, which resulted in small amounts of excess oxygen in the flue gas. 6
Regenerator bed temperature increased dramatically, going from about 1,200 F. to 1,350 F. Shortly thereafter, CO promoters were introduced, which further improved the CO burning. The result is a greater ease of start-up and the elimination of devices used to withstand runaway afterburn, such as regenerator water sprays, cyclones, and flue gas line quenches.
Regenerator temperature is controlled today by the overall heat or coke load. In most gas oil operations, the coke load is such that excessive temperatures are not reached. If reduced crudes are included with the gas oil feed, and temperature becomes excessive, feed rate is lowered.
Along with complete CO burning and high temperatures have come problems of uneven afterburning and excessive catalyst deactivation. Uncontrolled afterburning is caused by uneven ratios of air-to-coke across the regenerator. In the sections of the regenerator having a low air-to-coke ratio, oxygen is depleted and CO is formed.
Conversely, in the sections having high air-to-coke ratios, the gas leaving the bed contains portions of free oxygen. This oxygen combines with the CO, usually near a cyclone entrance, liberating, heat and raising temperatures to a dangerous level.
Air distribution systems are fairly easy to design. The multinozzle, triangular air distributor (Fig. 5) has provided excellent service. 7 The same high hole count as described for steam distributors is used with air distributors.
Spent catalyst distribution is another story. In designing spent catalyst distribution systems, the most important point is that in a well fluidized bed, catalyst moves and mixes very well vertically but very slowly horizontally.
The upward flowing gas provides the vertical movement. There is very little movement horizontally within the bed; most of the horizontal movement results from the eruption of the bubbles at the bed level. 8 People who have studied catalyst mixing in small diameter plastic vessels and extrapolated the results to very large diameter regenerators have reached wrong conclusions.
Adequate spent catalyst distribution can only be obtained by positively moving catalyst to a number of locations around the regenerator. The six troughs in the design mentioned previously do a very good job (Fig. 5).
The side-by-side units with deep conical bottoms adequately distribute spent catalyst. The ever-decreasing head in the conical section allows the catalyst to move horizontally. A side-by-side unit designed with this spent catalysts and the previously mentioned air distribution, gave a regenerator in which there was no temperature more than 5 F. different than any other temperature.
Along with good spent catalyst distribution, control of the initial burning rate is essential in avoiding excessive deactivation. Combining air at its highest oxygen content with spent catalyst at its highest coke content can raise particle temperatures some 300-400 F. above the average bed temperature.
With the troughs at the top of the bed, initial burning with a gas that has been depleted of its oxygen limits the initial burning rate. In the side-by-side units, meeting the spent catalyst with only the lift air (about 10-15% of the total air) controls the initial burning rate.
In reduced crude cracking, feedstocks of about 3% carbon residue can be run in units designed for gas oils with minor changes in operating variables, primarily an increase in dispersion steam rate. With poorer feedstocks, catalyst coolers must be used.
The design of catalyst coolers has improved considerably since the early days. 9 The improvement has resulted from putting steam through the tubes rather than air/catalyst, and by using much lower velocities to increase heat transfer coefficients.
In recent years, regenerator designs have included high velocity, low density systems. Better contact between gas and solid is achieved if it is carried out in the dispersed phase rather in a fluidized bed.
This has been demonstrated in riser cracking vs. bed cracking. The same design considerations noted previously to avoid excessive deactivation in bed regeneration must be followed, notably the use of an initial partial burn.
TRANSFER SYSTEM
Catalyst transfer systems and their designs have varied considerably over the 50 years of the FCC process. These changes are a result of developments in catalyst, improved reaction and regeneration designs, and a better understanding of fluidization.
In the days of amorphous catalyst, it was believed that all that was necessary for good catalyst flow was to have sufficient carbon on catalyst. Little attention was paid to aeration in vertical or slanted standpipes. Standpipes were designed with single aeration points at the top, bottom, and middle, with no understanding of why they were necessary.
In horizontal or near-horizontal pipes, it was a different story. Aeration points were included about every 3 ft, with about three points around the pipe. Of course, they were on manual control to allow the operator to adjust aeration rate for best catalyst flow.
In some areas, it was felt that better results were obtained by excluding aeration gas. For example, introduction of air was purposely restricted for about a 2-ft diameter around the exit hopper.
It was also felt that the catalyst should be partially deaerated before it entered the hopper. A deaeration height was used below stripping steam rings so the catalyst did not entrain too much steam and affect standpipe flow. An aeration ring is now used in the bottom cone of the stripper.
The high (0.6%) carbon on catalyst in the early days of cat cracking was believed to lower the friction between particles and permit good flow. With zeolitic catalysts having higher static charges, and with high temperature regeneration lowering carbon on catalyst to less than 0.1%, a different means of promoting good flow characteristics was necessary.
It was logical that, as the friction between catalyst particles increased, designers would look for ways to keep the particles apart. And so a great deal more attention was paid to aeration.
Aeration systems were designed so that the flowing catalyst density was not allowed to increase above the minimum fluidization density. Because aeration can form a bubble which can restrict flow, aeration was inserted at about four points at each level.
In the design of standpipes, kinks or angles in the line should be avoided. In addition, horizontal or near-horizontal sections can cause serious flow problems if not thoroughly aerated.
The one exception to avoiding near-horizontal standpipes is the downward slanting standpipes. These should preferably be short and designed with a low mass velocity to give a slip flow with dense catalyst at the bottom and mostly gas at the top.
The general statement that fluidized catalyst acts like a liquid in its flow characteristics is true with one major and highly important exception. The exception is that liquids exert the same density across their cross-sectional area. Fluidized solids do not.
A flowing fluidized solid on making a turn will exhibit an increasing density towards the outside of the turn. Because there is no give on the outside, the high density builds up towards the inside, restricting flow. A solution is to provide some give or cushion in the form of a small fluidized bed (Fig. 6).
One aspect of good standpipe flow quite often neglected is the influence of the preceding bed. This has become especially true as catalyst density has increases and deaeration time has decreased. Excellent gas distribution in the preceding bed results in good flow through the subsequent standpipes.
Distributors with poor fluidization characteristics and catalysts with low fines content will allow clumps of poorly fluidized catalyst to enter the hopper, restricting flow. This usually occurs spasmodically, with periodic recovery making the problem extremely difficult to troubleshoot.
FUTURE FCC/HOC UNIT
It is expected that the future FCC/HOC (heavy oil cracker) unit will involve high-temperature, short residence time cracking, dispersed phase regeneration, and complete removal of strippable hydrocarbons at minimum steam rates. A proposed design which incorporates these features is shown in Fig. 7.
Practically every article or patent on catalytic cracking mentions high activity, high temperature, and short residence time as the reaction system for future FCC units. The upflow risers currently in vogue lead to backmixing and recracking.
As mentioned previously, this is particularly a problem near the end of risers, where it is desired to crack the remaining gas oil while refraining from additional gasoline cracking. It is also reasonable to assume that the adverse effect of metal poisons would be less in reactors that avoid catalyst backmixing.
The proposed reactor for the future FCC/HOC unit is a downflow reactor. Downflow, as compared to upflow, poses a problem in that no minimum velocity is needed to lift catalyst. Thus, it is possible to have a downflow reactor that is very short and very wide, compared to upflow risers.
The problem then becomes achieving good catalyst distribution in a wide downflow reactor. The proposed solution is to provide multiple low-diameter downflow reactors. This not only permits good solids distribution but also provides for segregated cracking of different quality feedstocks.
Referring to the drawing, which shows just two downflow reactors, catalyst flows from the hopper in a partially settled state free of gas bubbles. Its rate is controlled by a plug valve to give a desired outlet temperature.
Steam is introduced to lower oil partial pressure and to provide a high velocity, short time in the high-temperature sections of the downflow reactor. Feed is then introduced through a specially designed nozzle which provides a high impact force, ensuring complete coverage of the downflow reactor.
The downflow reactors have lengths necessary to give the desired reaction time. The catalyst is separated from vapors at the bottom of the vessel by a simple inertia separator.
The vapors, with a small amount of catalyst, then pass through high-efficiency long cyclones for final separation. The catalyst flows symmetrically downward into the stripper.
The stripper is the high efficiency, two-stage, minimum-baffled design described previously. The regenerator design includes controlled initial burning and dilute phase regeneration.
In the proposed design, about 1015% of the air required for complete combustion is used to lift the coked catalyst to the main regenerator. With the sparsity of air, the particle temperatures do not become excessive.
Air and catalyst meet at the bottom of the stripper and flow through a horizontal line at velocities exceeding the saltation velocity. They then flow through a right angle turn upward into the main regenerator where additional air sufficient for complete combustion is provided. A right angle turn is used because it results in less pressure drop than a gradual turn.
A catalyst cooler, if necessary, is located near the main regenerator. Regenerated catalyst is then separated from flue gas and enters a hopper stage, from which it flows to the reactor hopper. Flue gas passes through two-stage cyclones to an external plenum chamber, then to final separation equipment and power recovery.
REFERENCES
1. Avidan, A.A., Edwards, M., and Owen, H., "Innovative improvements highlight FCCs past and future," OGJ, Jan. 8, 1990, pp. 33-58.
2. Murphy, J.R., "An External Catalyst Cooler For HOC Operations," Ketjen Cat. Symp. '86, pp. 58-62.
3. Bryson, M.C., and Murphy, J.R., U.S. Patent 617,497, Nov. 2, 1971.
4. NPRA Q&A 1988, p. 62.
5. Heldman, J.D., et al., Proc. API (111), Vol. 36, 1956, pp. 258-264.
6. Horecky, C.J., et al., U.S. Patent 3,909,392,1975.
7. Glasgow, P.E., and Murcia, A.A., "Process and Mechanical Considerations for FCC Regeneration Air Distributor," Katalistiks 5th FCC Symp., May 23-24, 1984.
8. Milne, L.P., Nienew, A.N., and Patel, K., "Lateral Mixing in Batch Beds of One or Two Components," Fluidization VI, May 7012, 1989.
9. Johnson, T.E., "Improve regenerator heat removal," Hyd. Proc. 70, Nov. 11, 1991, pp. 55-57.
Copyright 1992 Oil & Gas Journal. All Rights Reserved.
