Jones D.S.J., Pujado P.R. Handbook of Petroleum Processing

Подождите немного. Документ загружается.

252 CHAPTER 6

Stone & Webster Kellogg ATOMAX II

Dual Slot Impact

Lummas MicroJet UOP OPTIMIX

Oil Inlet

Steam Inlet

Target BoltBolt

Figure 6.8. Commercial feed injection systems.

RISER SEPARATOR STRIPPER

VORTEX SEPARATOR

Figure 6.9. Riser separators.

FLUID CATALYTIC CRACKING 253

Tee Rough Cut

Cyclones

Vented

Riser

Linear

Disengaging

Device

Close-Coupled

Cyclones

Figure 6.9. (Cont.)

The dilute phase of the reactor-stripper is a place where secondary thermal reactions

can occur and the extent of these reactions can be estimated from equation (3).

Delta Dry Gas = Ke

E/RT

t = (8.3 × 10

)



−93,000/(1.987

∗

T (R))



t (3)

A plot of this equation (Figure 6.10) shows that high reactor temperatures and long

dilute phase residence times lead to high dry gas yields. At low reactor temperatures,

0.5

1.5

2.5

500 510 520 530 540 550 560 570

1 s

5 s

10 s

20 s

Reactor Temperature, °C

Dry Gas, wt%

Dry Gas Production v T & t

Figure 6.10. Post riser dry gas production.

254 CHAPTER 6

External

Roughcut Cyclone

Internal

Roughcut Cyclone

Axial Cyclone

QuenchQuench

QuenchQuench QuenchQuench

Figure 6.11. Reaction systems with post riser quench.

below 505

◦

C (941

◦

F), the impact of the dilute phase is small and the use of advanced

termination technology is difﬁcult to justify especially if there is a chance of downtime

due to equipment malfunction. The value of K can be determined for an individual

unit by measuring the amount of dry gas produced at two different temperatures and

measuring the vapor residence time with tracers. A number of factors can add to the

dry gas make such as catalyst type, feed properties, the amount of dilute phase catalyst

and carbon on regenerated catalyst.

Another approach to reducing the post gas dry gas make is to reduce the temperature at

the end of the riser. This has been done at the end of the riser before the separator and

just downstream of the catalyst/vapor separator at the end of the riser. Quenching in

the riser will reduce gas make and may be needed if there is a metallurgical limit in the

reactor-stripper vessel. However, the amount of quench is high due to the catalyst that is

cooled and the stripper must operate at a lower temperature. The coke make and air re-

quirements are increased since this is a recycle stream from a heat balance perspective.

Downstream quenching only cools the vapor and does not add to the coke yield. It

can be applied to any unit that has a rapid/effective catalyst/vapor separator as shown

in Figure 6.11. The quench material is taken from the main fractionator and from

a heat balance perspective acts the same as a pump around on the main column. It

can be adjusted to accommodate changes in the reactor temperature and discontinued

at low temperatures. Beneﬁts from three different units are listed in Table 6.8. As

equation (3) implies and Figure 6.10 illustrates, the effects of time and temperature

are equivalent. However, the vapor quench also reduces the cracking that occurs in the

secondary cyclones, plenum changer and overhead vapor line. Care must be exercised

when using quench to avoid coke formation. These precautions include the choice of

the quench medium, operating above minimum operating temperatures and automatic

shut-offs and purges.

FLUID CATALYTIC CRACKING 255

Table 6.8. Impact of vapor quench on FCC yields

Unit A B C

Temperature (

◦

Riser outlet 513 549 532

After quench 484 519 494

Yield shifts (wt%)

Dry gas −0.23 −0.80 −0.66

Gasoline +0.43 +1.80 +2.89

LPG – – −1.58

LCO – – +0.25

DO – – −0.86

Stripping technology

After the cracking reactions are completed, the spent catalyst needs to be stripped of

the hydrocarbons that would accompany it to the regenerator. This is done in a staged-

ﬂuidized bed where steam enters from the bottom and pushes the hydrocarbons in

the gas phase out of the top of the bed. Design parameters for new units are given in

Figure 6.12, which is a common disk and donut design.

The bafﬂes improve contacting between the steam and catalyst and increases the

number of contacting stages. As shown in Figure 6.13, seven stages of stripping is

sufﬁcient to remove at least 95% of the hydrocarbons. Each of the design parameters

are important to the proper operation of the stripper. A minimum amount of steam is

necessary to displace the hydrocarbons in the emulsion and bubble phases of the ﬂuid

bed. The ﬂux rate determines the catalyst velocity through the bed. If the downward

velocity of the catalyst gets too high, it will sweep hydrocarbons and steam with it

and adversely affect the stripper performance. The residence time is a function of the

stripper’s catalyst inventory and the catalyst circulation rate and relates to the number

of stages and their efﬁciency that the stripper can obtain.

Disc and Donut Trays

Steam

2-3 lb/1000 lb

Catalyst Circulated

Design Catalyst Flux

600-900 lb/ft

2 Min

Cat. Residence Time

60-90 Seconds

Figure 6.12. Disk and donut stripper.

256 CHAPTER 6

100

Stripping Gas

Overall Stripping Efficiency (%)

N=1

= T

E (n =1)

N=2

N=3

N=4

N=5

N=6

N=7

(lb./1000 lb Catalyst)

Figure 6.13. Disk and donut fray efﬁciency/stage efﬁciency.

While the disk and donut design shown has proven to be both reliable and effective,

there are other variations on this design. Holes can be placed in the bafﬂes to improve

contacting, the skirts can be lengthened to provide a larger gas P, vent tubes have

been used to allow the gas from the bottom of the bafﬂes to pass to the next stage and

rods or shed decks can be substituted as the contacting devices. The use of structured

packing has been reported recently with excellent results. Lower steam usage, better

contacting, and utilization of almost the entire cross-sectional area of the stripper are

beneﬁts claimed with the new design. Horizontal trays with small holes conﬁgured as

distillation trays have been tested in the lab and will also be implemented in the ﬁeld.

Regeneration technology

The object of the regenerator is to remove the coke that builds up on the catalyst in the

reactor without damaging the catalyst. Many studies have been made on the burning

rates of coke in a ﬂuidized bed of cracking catalyst. Equation (4) describes the major

regeneration operating variables

= K × C

× L

× e

A/ RT

(4)

The contacting between the oxygen and catalyst is improved signiﬁcantly as the air

rate or superﬁcial velocity is increased in the regenerator. As the velocity increases

the bed goes through three stages. A bubbling bed occurs at low superﬁcial velocities

(up to about 1.5 ft/sec or 46 cm/sec). Here relatively distinct bubbles are formed and

pass through the bed. A turbulent bed (1.5–4.0 ft/sec or 46–122 cm/sec) exists at

higher superﬁcial velocities in which an emulsion is formed and the diffusion rate of

oxygen is signiﬁcantly increased. At higher velocities a fast-ﬂuidized bed (4–8 ft/sec

or 122–244 cm/sec) exists in which turbulence is maximized. A return line from the

recovered catalyst to the combustor is required to provide enough residence time and a

FLUID CATALYTIC CRACKING 257

100

1000

1150 1200 1250 1300 1350

Regeneration Temp.°F

700

300

500

Plug Flow

Bubbling

Bed

Carbon 0.85 0.05 wt%

Constant Pressure & XS Oxygen

Figure 6.14. Comparison plug ﬂow and Backmix regenerators (ideal).

sufﬁcient mix temperature for the coke to completely burn. The difference in burning

rates between these idealized regenerator designs is shown in Figure 6.14.

A few observations can be made from Figure 6.14 and equation (4). At high tem-

peratures (>1,300

◦

For704

◦

C) the burning rate is very high. Oxygen availability

limits the burning rate though the small size of the ﬂuid cracking catalyst eliminates

or minimizes diffusion as a reaction barrier at normal regenerator temperatures. The

plug ﬂow curve moves up, as catalyst is recycled in the combustor since the effect is

to increase the residence time. Regenerators with larger inventories will reduce the

carbon satisfactorily even if there is a partial malfunction of the air distributor while

smaller inventories allow faster change outs of the catalyst inventory. Over the years,

the regenerator temperatures have increased due to better metallurgy and the need

to burn all of the carbon off the catalyst to restore the zeolite catalysts’ activity and

product selectivity.

At 1,100

◦

F (593

◦

C) the regenerator inventories had to be large to provide enough time

to burn the coke. Full CO combustion raised these temperatures to 1,300–1,350

◦

(704–732

◦

C). CO promoters are also frequently employed to assist the carbon burn

and prevent the afterburning of CO in the dilute phase or downstream hardware

where serious equipment damage can occur. The differences in catalyst inventory are

illustrated in Table 6.9 for a 50,000 B/D unit that circulates between 30 and 40 tons/

min of catalyst.

258 CHAPTER 6

Table 6.9. Regenerator parameters (50,000 B/D)

Residence time Regenerator inventory

Regenerator type (Min) Temperature (

◦

F) (tons)

Bubbling bed 5–20 1,100 300–800

Turbulent bed 3–5 1,250–1,350 200

Fast-ﬂuidized bed 1–3 1,275–1,350 120

Shorter contact times and smaller catalyst inventories limit operable regeneration

conditions and need higher internal catalyst recycle rates for increased throughput or

coke burn. Since catalyst is frequently added on a pound per barrel basis each of these

units would use about 5 tons/day of fresh catalyst. Equation (5) relates the catalyst

activity to S, the daily fractional replacement rate or age of the catalyst:

A =

+ S

. (5)

This equation implies the smaller inventory would give the highest equilibrium or unit

activity. However, K is also a function of the contacting between the spent catalyst

and air, the mix temperature, the catalyst type and activity and the number of cycles

the catalyst makes through the system. This latter fact implies there is an optimum

unit inventory for a given processing capacity.

Commercial regenerator designs are shown in Figure 6.15. These utilize either turbu-

lent beds or fast-ﬂuidized beds. Cocurrent or countercurrent contacting of the catalyst

and air is practiced and care is taken to prevent short-circuiting of the catalyst from

the regenerator inlet to the outlet to ensure an even, low carbon distribution on the

regenerated catalyst.

Resid catalytic cracking

Processing heavier feeds poses challenges to the normal FCC design due to the higher

coke laydown on the catalyst during the cracking reactions. The coke layed down in

the cracking process has been shown to come from four main sources as shown in

Table 6.10.

The catalytic coke comes from the secondary cracking reactions and are caused by

polymerization and condensation of hydrocarbons. Strippable coke are the hydro-

carbons that are entrained with the spent catalyst that enters the regenerator. Heavy

metals that lay down on the catalyst surface promote dehydrogenation and lead to

extra coke and hydrogen. Nickel, vanadium, and Iron are the main contaminates

though occasionally copper, zinc, and lead have been known to cause problems. Feed

coke has been associated with the carbon residue in the feed as measured in the

Conradson Carbon Test (ASTM). This has been also referred to as additive coke.

FLUID CATALYTIC CRACKING 259

EXXON FLEXICRACKER

UOP HIGH EFFICIENCY DESIGN

M. W. KELLOGG

S&W SINGLE STAGE DESIGN

Figure 6.15. Commercial FCC regenerator designs.

As Table 6.10 shows the sources of coke shift dramatically when resid is in the

feed. The percentages given in Table 6.10 are not ﬁxed and shift as the composition

of the feed and operating parameters change. The total coke make is different in each

case. There are other factors that lead to coke formation that are included in the four

categories. Basic nitrogen is known to cause coke since these molecules are strongly

adsorbed on the acid sites in the reactor and are burned off in the regenerator.

Table 6.10. Sources of coke production

Feedstock Gas Oil Residue

Coke categories

Catalytic 65 45

Strippable 25 5

Contaminant 5 20

Feed Coke 5 20

260 CHAPTER 6

Contaminant Coke from Con. Carbon

Far East Brent Arab Lt.

Con. Carbon to Coke, %

FCC Feed Resid Factor

Increased

Reactor Temp

Figure 6.16. Percentage of conradson carbon going to coke. Reprinted with permission from HYDRO-

Some of the very heavy hydrocarbons in resid may not be vaporized and be laid

down on the catalyst surface where they eventually coke. Figure 6.16 shows that the

percentage of Conradson Carbon that goes to coke is a function of the feedstock and

reactor temperature.

Both desorption of basic compounds and feedstock vaporization would be improved

by raising the reactor temperature (and consequently the stripper temperature) so the

relationship shown in Figure 6.16 is directionally correct. The carbon laydown or delta

coke can be represented by equation (6). The ﬁrst term in is the Voohries relationship

for carbon laydown for gas oil feedstocks while the second term reﬂects the feed coke

contribution. A, B, and C are constants that depend on operating conditions, feed

properties and catalyst tested.

Delta Coke = A



Catalyst Res. time

in the reactor



C ∗ Conradson Carbon

Catalyst/oil ratio

. (6)

The overall heat balance equation for any cracking unit is:

Wt% Coke = Delta Coke × Catalyst/Oil ratio. (7)

The dependent variable is the catalyst/oil ratio, which needs to be high enough to

give the desired conversion. Weight percent coke is strictly a function of the operating

variables (i.e. fresh feed and recycle rates, feed temperature, reactor temperature,

steam rates, heat of cracking, air rates, and carbon burning mode). Since higher delta

FLUID CATALYTIC CRACKING 261

Table 6.11. Methods of increasing coke make for resid processing

Modiﬁcation Objective Consequence

Partial burn Lower regen temp.

Burn more coke

CO emissions rise, carbon on

catalyst increases

Mix temp. control Cool regen

Aid vaporization feed

Higher coke

Water injection Cool regen More coke

CAT deactivation?

Oxygen enrichment Increase coke burn

Maintain superﬁcial

Air velocity

Higher regen temperature

Auxiliary air Burn more coke Inc. regen superﬁcial velocity

Catalyst cooler Reduce regen

Temperature

Higher coke

More air

Second regenerator Increase throughput

and resid

Cost

cokes are caused by heavier feeds, the resulting cat/oil ratio becomes too low for

medium to high conversion levels. The coke make must be increased to raise the

catalyst/oil ratio and this can be done by any of the means shown in Table 6.11. For

the heaviest feeds a catalyst cooler will be required.

In Figures 6.17–6.21 the commercially offered resid FCC units are pictured.

Much of the reactor-stripper design is the same for resid crackers as it is for gas oil

designs. However, there are feed injectors designed speciﬁcally to process residual

feeds that require more dispersion steam than the normal gas oil models.

Figure 6.17. S&W/Axens R2R.