SHELL'S RESID FCC TECHNOLOGY REFLECTS EVOLUTIONARY DEVELOPMENT
Mart J.P.C. Nieskens, Frank H.H. Khouw
Shell Internationale
Petroleum Maatschappij B.V.
The Hague
Martin J.H. Borley
Shell U.K. Ltd.
Ellesmere Port, U.K.
Karl-Heinz W. Roebschiaeger
Koninklijke/Shell-Laboratorium
Amsterdam
Commercial operating experience has been gained on conversion of long residues by fluid catalytic cracking (FCC) in a long-residue catalytic cracking unit at Shell U.K. Ltd.'s Stanlow refinery in Ellesmere Port, U.K.
The unit design is the result of many evolutionary developments in Shell's FCC technology since its beginnings in the 1940s.
Shell's first long residue FCC unit at the Stanlow refinery was started up in May 1988. The unit has operated at its design capacity of 9,500 metric tons/day (66,000 b/d), processing feedstocks with a Conradson carbon content of up to 5 wt %. With these heavy feeds, the gasoline yield is about 45 wt % (55 vol
The produced butane/butylene stream goes to C4 Solvents and a new alkylation unit. About 5 wt % (9 vol %) of the production consists of high-grade propylene, whereas the ethylene will be used to produce ethylbenzene.
The unit has proven to be extremely flexible. It includes Europe's largest power recovery train, capable of generating 21,000 kw.
Specific design features include a proprietary liftpot and feed injection system, a compact reactor design, a staged catalyst stripper, a unique two-stage, but single-vessel regenerator, controllable catalyst coolers on the regenerator vessel, and a catalyst-handling system for continuous addition and withdrawal.
A high priority is given to energy efficiency of the units. Most new Shell FCC designs and revamps incorporate the Shell swirl tube separator and power recovery technology.
Not only does this proven technology greatly reduce plant operating costs by the recovery of energy from the flue gas through an expansion turbine, but it also makes long-residue FCC's net power exporters.
New developments will enable processing of increasingly heavier residual feedstocks.
EARLY SHELL WORK
Today, some 30 Shell or Shell-advised FCC units are in operation worldwide, with capacities ranging from 1,000 to 11,000 metric tons/sd (7,000 to 80,000 b/sd). A wide variety of feedstock types is processed, ranging from clean vacuum gas oil to straight-run atmospheric residues.
Coprocessing of many other unconventional feed types is practiced, such as flashed distillates from visbreaking/thermal cracking and lube oil extracts.
Since the early 1940s, the technology of catalytic cracking has made very significant progress with respect to catalyst as well as hardware efficiency.
Fig. 1 shows the trends in the catalyst developments. The effects of the improvements are reflected in terms of yield breakdowns in Fig. 2 showing, mainly as a result of the catalyst developments, a pronounced shift from residual components to gasoline.
The simultaneous hardware developments have enabled the FCC technology to maintain its prominent role in conversion. Gross margins of $30-70 US/metric ton of feed have resulted in a primary emphasis on maintaining high throughput and onstream time.
The high differential values between distillates and residual material (fuel) have led to the inclusion of more and more residue in FCC feedstocks. The problems to be solved, from initial experience, include high coke-forming tendency, catalyst deactivation by metals and consequently poorer yields, particularly high gas yields, and coke depositon in the reactor section.
Much progress has been made, including that by catalyst manufacturers, to overcome, or at least contain, the harmful effects of residue in feed. As a consequence, in the past 10 years Shell units have seen a gradual increase in feed residue content.
Currently, feedstocks with about 5 wt % Conradson carbon and about 10 wt ppm nickel (Ni) + vanadium (V) are being processed in Shell units. The latest Shell long-residue FCC units, in Singapore and Geelong, Australia, are designed for feedstocks containing 6-7 wt % Conradson carbon and about 20 wt ppm Ni+ Y.
The Singapore unit was started in the first quarter of 1990. Progress in feed heaviness is shown in Fig. 3.
The developments in FCC catalysts have led to significant adaptations in the designs of the reactor/regenerator/stripper sections. Fig. 4 shows units which typify Shell's FCC technology over the decades since the 1940s. All units are drawn on a same intake/same scale" basis to make comparisons realistic.
The design of the first FCC units was largely dictated by the quality of catalyst available at the time. Amorphous, silica-alumina catalysts, by today's standards, displayed very low activity and very poor product selectivity.
Thus the designs included very large, dense-bed inventories (often over 100 metric tons), very high catalyst-to-oil (C/O) ratios (often up to 12:1), and low temperatures around 500 C. (to suppress dry gas make).
These reactor conditions resulted in very modest regenerator bed temperatures, usually in the range of only 600-650 C. At the same time, coke yields were high, typically around 8 wt %, even for clean, flashed distillate feeds, due to the very poor selectivity of amorphous FCC catalyst.
With relatively cheap, lowtemperature metallurgy for internals, the early regenerators ran in partial CO combustion to maintain overall unit heat balance. This always required a relatively complex, fired CO boiler.
Catalyst stripping in the 1940s FCC was usually carried out in several distinct stages, including dilute-phase stripping in a steam-lift riser, and dense-phase stripping in at least one stripper dense bed. A simple dollar plate was originally used as a primitive vapor/catalyst separation device at the end of the stripper riser.
The introduction of zeolite catalyst, with its vastly improved activity and selectivity, led to a number of very significant design changes: pure riser reactor (short catalyst/feed contact time), integral dense-bed stripper (as part of reactor vessel), and high-temperature regenerator (in complete CO combustion).
Gas/solid separation atop the riser was improved with various curved inertial separation devices replacing the simple flat dollar plate. Reduced catalyst entrainment offered the possibility of using only a single stage of cyclones for the now combined reactor/stripper vapors.
In search of more octane and more C3-C4 light olefins, reactor temperatures have risen because the better yield selectivity of zeolites could maintain dry gas make at acceptable levels. Simultaneously, C/O ratios declined (typically in the range 5-8) due to higher catalyst activity.
These effects pushed regenerator temperatures to the 700 C. level, demanding higher-alloy stainless steel for exposed internals and pipes. Due to the much lower coke yields (4-5%) found when flashed distillates were cracked over the new zeolite catalysts, overall unit heat balance could best be satisfied by operating the regenerator in complete CO combustion.
The CO boiler could thus be replaced by a much more compact, simple waste heat boiler. It was also found that units operated much more smoothly in complete combustion (without coke-buildup problems) and that catalyst was burned much cleaner.
CRC levels (coke on regenerated catalyst) of below 0.1 wt % could be easily achieved. This, in turn, further enhanced product yields on the reactor side.
To minimize entrained coke, the popular side-by-side, two-vessel unit of the 1970s saw the introduction of baffles in the stripper bed.
This restored the staging otherwise lost by the demise of riser stripping.
With the steady drop in fuel oil demand, the FCCU assumed a new role. The job of "fuel oil reducer" by residcracking became economically crucial. Henceforth, .,gasoline producer" would be too narrow a definition for the FCC process.
The thrust in the direction of the ability to process heavier feedstocks was met in the first by revamps. In practical terms, this mostly meant upgrading of the regenerator for higher maximum operating temperature (usually 750 C.), modification to the reactor/stripping section, addition of catalyst coolers, and addition of power recovery facilities, with or without additional air capacity.
Shell's first residue FCC design was developed in the early 1980s. A very important factor in the design of residue FCCU's is the coke-making propensity of residue feedstocks.
Many different views have been expressed in the literature on the relation between Conradson carbon residue (CCR) and rapid coke deposition.' It is therefore important to recognize the effects of the various factors involved from commercial plant and laboratory feedback.
These factors include CCR content, nature of feedstock, type of catalyst, and type of plant and operating conditions. The combined effect of these influences may produce rapid coke deposition levels of anywhere between 40% and 70% of CCR content, governing the size of the major equipment items, and determining to a large extent the nature and size of utility integration schemes.
Reactor design changes dramatically lowered the high-temperature hold-up time of product vapor. This minimizes thermal overcracking, a phenomenon which inevitably downgrades gasoline/light cycle oil (LCO) to dry gas/slurry and produces coke.
The exposure period from riser exit to fractionator should be as short as possible to accommodate the trend to ever-hotter reactors to improve coke-conversion selectivity from heavy feeds. The typical 30-sec quench time of traditional big-vessel reactors was cut by a factor of more than three with external cyclones, also designed for easy access.
Reactor yields are greatly enhanced by Shell's proprietary liftpot/feed injection system, where optimally dispersed heavy feed is rapidly and very uniformly mixed with prepassivated, prefluidized, regenerated catalyst.
To cope with the relatively high coke yields of heavy resid cracking, the design minimizes heat release in the regenerator by returning to partial CO combustion. A degree of staging in the single-vessel regenerator achieves both high CO in the flue gas and low CRC. Side-mounted catalyst coolers (licensed from UOP) are used to trim the final unit heat balance.
The two latest resid RCC designs, also shown in Fig. 5, are scheduled to come on stream in 1990-1992. They evolved from on-going development efforts to minimize unit size, complexity, and capital costs, without compromising process performance.
The latest designs include: a more compact reactor, with only a few seconds residence time between riser exit and fractionator quench; a novel, proprietary high-efficiency, low pressure drop riser-end-separator; an integral STS (swirl-tube separator) inside the regenerator vessel; a single stage of conventional regenerator cyclones; complete or partial CO combustion in the regenerator; larger capacity catalyst coolers (without external piping); and reduced unit catalyst inventory (to speed change-overs).
Inside the small, hot-wall reactor (now a simple, cylindrical vessel) a newly designed horizontal riser-end cyclone is connected to easily retractable swirl-tube cyclones. This combination has low pressure drop and high efficiency, and it virtually eliminates all internal dead spaces.
This last factor, plus minimal heat losses (from the extra small vessel surface area), significantly cuts the risk of coke-lump formation while processing heavy resids.
Catalyst stripping is now enhanced by counterflow and staging. By using new internals, the required degree of staging can be achieved in a very simple, single vessel.
The regenerator can be designed for either complete or partial CO combustion, with or without catalyst cooling.
The regenerator return to complete CO combustion in the Geelong long-residue I FCCU was prompted by the catalyst cooler reliability.
With reliable catalyst coolers there is no need to restrict heat release in the regenerator.
Therefore, the unit can operate at complete CO combustion, with the benefit of using sulfur oxides reduction catalysts to minimize SOx emissions when processing high-sulfur resids.
STANLOW RESIDUE FCC
unit The Stanlow long-residue FCCU is shown in Fig. 6. The unit has an efficient feed/catalyst mixing system with a short contact time riser reactor to minimize coke and gas make, including minimum space for coke deposition, short vapor residence time, and application of antimony passivation.
The unit uses stripper technology to minimize coke make. This required a severe, but practical, stripping system with multiple stages, including quick prestripping of hydrocarbons from the segregated catalyst and a second-stage stripper designed to efficiently desorb and displace remaining hydrocarbons from the catalyst.
A high-efficiency regenerator with high-temperature capability is incorporated that restricts heat release in the regenerator, as a result of partial CO combustion, allowing heat in the form of CO to pass to the CO boiler. A reliable and uncontrolled catalyst cooler is included to assist unit heat balance with ample coke-burning capacity.
The unit is capable of operating under a range of temperature and pressure conditions with relatively short, inclined standpipes to ensure stable catalyst flow. Modern catalyst addition and withdrawal facilities are also provided.
Heat integration was carefully considered in the design. This is particularly important for residue FCCU's due to the energy export from the unit. High-pressure steam (110 bar gauge) is produced in the CO boiler, while super-heated medium-pressure steam is recovered from other waste-heat sources.
A power recovery system (PRS) is included with a high net power generation. Fig. 7 shows the overall flow scheme. The Shell-patented swirl-tube separator removes all particles above 20 u, plus a substantial number in the 4-20 u range.
Expander power output closely matches power demand of the direct-coupled FCC air blower. A small power deficit can be handled by a motor/generator in the same train, which also fixes the speed of the whole system and prevents varying regenerator pressures.
The Stanlow system size also allows the integeration of a steam turbine feeding into the machinery train (to consume net FCC steam output), so that the PRS train then becomes a significant exporter of electricity (up to 21,000 kw).
Besides its attractive energy economics, the Shell PRS also has a positive environmental impact. The swirl-tube separator significantly reduces particulates emissions (to below 1 00 mg/NM3) in the FCC flue gas.
The STS is a multi-tube, swirl-vane type of centrifugal separator. As shown in Fig. 8, gas enters the vessel axially and flows via trash screens into the space between two tube sheets. The gas flows down through the swirl-tube assemblies and is given a helical rotation by vanes in the assembly.
The heavier catalyst particles are centrifuged from the gas and spiral down the sides of the tubes. After leaving the swirl-vanes, the clean gas turns 180 and flows up through the gas tubes into the plenum at the top of the separator.
Particle sizes from the STS are in the range of 4-5 u, and the STS produces an underflow stream with a high solids concentration (approximately 20,000 mg/Nm3), which makes subsequent cyclonic separation of the solids relatively easy. The combined STS and STS underflow separator system is usually able to comply with local emission regulations.
The unit is designed to process 9,500 metric tons/day of feed containing up to 5 wt % Conradson carbon and about 12 wt ppm Ni + Y. Reactor and regenerator temperatures are designed for 525-540 C. and 680-750 C., respectively. The whole plant consumes less than 5 wt % on feed of standard refinery fuel and exports 21,000 kw of electricity.
OPERATING EXPERIENCE
Various sections of the plant were brought on-line successively and systematically with on-specification products to storage in mid-May 1988. Through May 1989, the throughput was constrained by the capacity of the gas recovery section.
In June 1989, the new gas recovery section was brought on stream, allowing operation at design throughputs. The unit has operated since then in excess of the design capacity of 9,500 metric tons/day with a feed containing up to approximately 5 wt % Conradson carbon.
Some typical performance data are included in Table 1. The stripper performance is excellent as is evident from the low content of hydrogen in coke (approximately 6 wt %).
Some fouling problems have been encountered in the slurry system and have been sorted out. There are no signs of coke buildup in the reactor system.
The power recovery train has given flawless service and has contributed significantly to the profitability of the utility system at Stanlow.
The regenerator has the flexibility to operate in both partial and complete CO-combustion mode. It is worth mentioning that the CRC is found to be in the order of 0.08-0.10 wt % from the new regenerator which operated in partial CO combustion at moderate temperature (approximately 680 C.). The flexibility of the unit to operate with a moderate regenerator temperature, even when processing long residues, gives the additional advantage of a low dry gas make (Table 1).
Although the initial performance of the unit has exceeded design expectations, further improvements will likely include: optimization of the catalyst choice, further optimization of Shell's latest system for catalyst/feed contacting, and maximization of the production of light olefins through the application of additives and a judicious choice of reactor/regenerator operating conditions.
CATALYST PERFORMANCE
During the first period of operation, the catalyst gradually became less active in converting heavy feeds (Fig. 9). During the same period, the metals content increased gradually (Fig. 1 0). The vanadium content increased from 2,200 to about 4,300 wt ppm, which resulted in a yield deterioration (Table 2).
The loss of catalyst activity and selectivity was not detected by the microactivity test (MAT) performed on a routine basis by the manufacturers of FCC catalyst.
The MAT activity is usually measured using a relatively light flashed distillate type of feedstock. It appears, therefore, that the catalyst with the high vanadium content is still capable of converting light feeds, but not heavier feeds.
The above theory is indeed confirmed in the commercial unit and Shell's FCC riser pilot plant. When processing a relatively light feed, only a slightly lower conversion level was obtained with the catalyst containing 4,300 wt ppm Y than with the catalyst containing only 2,200 wt ppm Y (Fig. 11). Obviously, conversion decreases significantly when processing a relatively heavy feed.
It is clear that metals resistance in FCC catalysts is becoming a very important aspect when processing heavy feeds. To date, discussions with catalyst manufacturers have revealed that many aspects of vanadium resistance are not yet fully understood. In the near future, FCC catalysts must be capable of tolerating high vanadium contents because catalyst replenishment rates can become very high when processing heavy feeds.
ON-LINE OPTIMIZATION
The FCCU is a dynamic and vital contributor to the refinery's gross margin. However, the process is also complicated in its interactions of feedstock and operating conditions. The situation may change substantially with time. For instance, catalyst properties may change or product values may vary.
The process is very flexible with respect to throughput, feedstocks, and product yield and quality. It is evident, therefore, that for best sustained results, the performance has to be monitored accurately. The unit's flexibility provides large potential for optimization, either on-line or off-line.
For the preceding reasons, Stanlow's LR-FCC unit is equipped with an on-line optimization system to be able to operate the unit as close as possible to the economic optimum, within all hardware constraints and market requirements.
Before an on-line optimization system can be implemented in a commercial unit, the following requirements must be fulfilled:
- A heat-balanced process model capable of predicting realistic yields and product properties for different operating conditions and feeds.
- Monitoring in the operational and technological sense, including adequate feed and products OMI'S.
- Control of the unit so that the unit is held close to the relevant constraints, or at the desired targets. To this end, the unit is equipped with an advanced control system.
- An optimizer which can be run within a short time frame to generate new targets for the plant controllers.
The Shell-developed system is able to optimize appropriately rigorous, nonlinear models developed for on-line optimization. Full advantage is taken of available on-line measurements, which are used to fit the plant model to current plant operations.
The models are updated to reflect the actual unit hardware/catalyst specifics. For updating/tuning of the models, dedicated tools have been developed by Shell's laboratories to sample the relevant process streams accurately. These samples are subsequently analyzed in a mobile laboratory.3
The optimization program adjusts the control targets based on the current plant conditions and constraints, while maximizing the objective function (profit).
RESIDUE FCC'S FUTURE
Residue conversion by FCC will continue to co-exist with other carbon-rejection processes (delayed coking and other thermal processes) as well as with hydrogen-addition processes. Each process has its particular field of application.
For resids with high metals or Conradson carbon levels, the applicability of FCC can be expanded by adding a feed pretreatment unit in front of the FCCU. Yet without such a unit, a large number of benign resids (some 30% of the world's crudes) can be processed directly. Fig. 12 illustrates the potential of residue FCC.
More metals-tolerant FCC catalysts, the prospect of demetallization/rejuvenation of equilibrium catalyst, bigger and better catalyst coolers in order to cope with higher coke yields (from feeds with more Conradson carbon), and developments in hardware will result in a steady increase in the heaviness of resids being processed by FCC units.
Resid cat. feed hydrotreating (RCFH) is capable of substantially improving the yields and quality of all FCC products. However, the high capital cost of RCFH makes this option difficult to justify with the presently foreseen economics.
Compared to feed treatment, product hydrotreating requires a smaller capital investment because smaller distillate streams (rather than the residue feed stream) must be processed. To meet increasingly stringent sulfur-in-gasoline specifications, hydrotreating/reforming of product boiling above 100 C. becomes inevitable.
To improve diesel-pool quality, LCO hydrogenation looks equally promising. Shell laboratories have recently developed a sulfur-resistant, zeolite-based hydrogenation catalyst.
The yields, quality, and value of nearly all FCC products can be significantly improved by product separation. A capable, flexible distillation scheme is vital for maximum FCC profitability.
Examples include:
- Ethylene recovery from dry gas. By appropriate treating and cryogenic distillation, polymerization-grade ethylene can be upgraded from fuel gas. Alternately, ethylene might in the future be absorbed in situ by direct alkylation of benzene/toluene (from light reformate) over an appropriate catalyst.
- Light LPG separation. Higher recoveries and sharper separation of propylene and butylene components can increase the profitability of associated chemical complexes, such as polypropylene and aklylation units. There may be wider application of new light ends distillation equipment, such as heat pumps, lower-pressure/temperature operation, and closer/more efficient trays.
- Gasoline/light cycle oil separation. Cutpoint variation here is a key source of yield flexibility from maximum gasoline mode to maximum LCO mode.
- Slurry oil separation. To make carbon black feedstock quality, a very heavy slurry is required. If components for carbon black feedstock are present in the slurry, a slurry flasher may be considered rather than running a higher fractionator bottom temperature, with the risk of faster heat exchanger fouling rates.
FCCU's are often the major single source contributors to refinery emissions. The process is consequently facing more and more stringent limits worldwide on SOx, NOx, and particulate emissions. Table 3 gives some data on emission standards.
SOx can be substantially reduced by feed hydrodesulfurization.
An SOx-reducing (deSOx) additive circulating with the FCC catalyst is a convenient way of reducing SOx emissions.
In this way, 70% sulfur removal can be obtained in complete CO-combustion, with about 35% removal efficiency in the case of partial CO combustion. If a higher degree of desulfurization is required, wet/dry-gas scrubbing must be used.
Several processes including catalytic/noncatalytic reduction or absorption have been developed for removal of NOx from the flue gases, with selective catalyst reduction being the most widely used.
In the selective catalytic reduction (SCR) process, NO, is reduced to N2 and H2O by ammonia, thereby reducing NOx by 60-85%.
Three SCR units have been installed downstream of Shell FCCU'S, all of them located in Japan. The units are performing well and they meet local NOx emission standards.
Both wet and dry emission control systems can be applied to keep the particulate emission from FCCU's at acceptable levels. Swirl-tube separators are used to protect the flue gas expander by removal of all the 20 u+ catalyst particles, thereby reducing the particulate levels in the regenerator flue gas to values ranging from 50-100 mg/NM3 , depending on type of catalyst and on the design of the air grid in the regenerator to minimize attrition and catalyst addition/withdrawal rates.
With an electrostatic precipitator, the dust concentration in the flue gas can be reduced to below 50 mg/Nm3.
REFERENCES
- Papers 7, 15, and 23, Katalistik's annual symposium, Budapest, 1987.
- Maxwell, I.E., "The dynamic and vital role of catalyst in the present and future oil & gas refining industry," KNCV Congress, Delft, The Netherlands, Aug. 25, 1988.
- Akbar, M., Nieskens, M.J.P.C., Parker, W.A., and Wesdorp, H.A., "Testing of FCC catalysts in commercial units," Paper No. F-10, Ketjen Catalysts Symposium, Scheveningen, 1988.
Copyright 1990 Oil & Gas Journal. All Rights Reserved.