G. W. G. McDonald
IONA Ltd.
Sarasota, Fla.
Simple changes in operating conditions can improve fluid catalytic cracking vapor recovery unit (VRU) operation.
Debottlenecking the VRU can release unused cracking capacity and defer a major VRU revamp.
Process and catalyst advances permit older fluid catalytic cracking units (FCCUS) to operate at capacities and severities much higher than original design parameters. Operations are therefore often constrained by the capacities of the downstream fractionation systems and other process plants that have not benefitted from similar technological advances.
Sometimes product recoveries and specifications are relaxed to mitigate the impact.
Any improvement in VRU capacity and recovery is probably economically justifiable, especially when propylene is routed to alkylation or petrochemicals, However, when an improvement in VRU performance permits a proportionate increase in FCC capacity or severity, the economics can be very persuasive.
These economics are favored by the low cost of an increase in VRU capacity, which can be a fraction of the cost of the corresponding increase in FCCU cracking capacity.
The operations of some high-severity refineries are constrained by the excessive production of fuel gas, which must be flared. In those cases, improved light-ends recovery can permit an increase in the throughput or operating severity of the entire refinery.
Often significant performance improvements are the cumulative effect of several closer approaches to equipment constraints. It may be necessary to employ advanced controls to maintain those optimum conditions.
The data presented here are taken from studies of various VRUs. Other units of different designs operating at other conditions will exhibit different responses, but the trends will be similar.
WET-GAS COMPRESSION
While the heart of the VRU is the stripper-absorber de-ethanizer system, its operation involves interactions between all the equipment shown in Fig. 1, from the FCC fractionator to the downstream debutanizer.
Wet-gas compression is a common operating bottleneck, and a costly one to relieve.
SUCTION CONDITIONS
The temperature and pressure of the FCC main fractionator receiver have many effects on the VRU. Those conditions determine the relative wet gas and distillate production and therefore the liquid-to-volume ratio in the absorber.
The required gas-compressor capacity is doubly affected-by both the wet-gas production rate and the compressor suction conditions.
Fig. 2 shows that a 1 psi increase in overhead receiver pressure will increase the capacity of an existing compressor and the potential FCCU capacity by as much as 4%. The magnitude of the effect depends on whether the equipment is limited by driver power or by gas flow-rate considerations.
Likewise, Fig. 3 shows that a 10 F. change in overhead receiver temperature produces an 8% change in compressor and VRU capacity.
Changes in fractionator receiver conditions have little effect on propylene recovery, but have a significant effect on absorber load, as shown in Figs. 2 and 3.
The relative tower loadings used in this article are the ratios of the areas of two fractionators of equivalent tray design. The capacity ratios will be different if the tray design does not have balanced liquid and vapor areas, This explains the sensitivity of some plants to changes in ambient conditions.
The FCC fractionator receiver pressure is primarily determined by the FCC reactor-regenerator pressure balance. Any increase in the receiver pressure will increase the FCC air-blower discharge pressure and horsepower, as well as the regenerated catalyst slide-valve differential pressure. Both effects are undesirable.
Although constrained by FCC considerations, the astute operator can still optimize fractionator receiver conditions by:
- Increasing the receiver pressure until the FCC operation is simultaneously limited by the air blower and the gas compressor. (Any control gas spill-backs and air vents should be closed.)
- Lowering the receiver temperature by increasing cooling and by reducing top reflux. (The price differential between FCC gasoline and light cycle oil is often too low to justify the improvement in fractionation from a high reflux rate.)
- Reducing the condenser pressure drop and increasing the potential receiver pressure by lowering the top reflux rate and reducing the mass flow rate through the condenser. (The top reflux can be lowered by judiciously increasing heat removal lower in the fractionator. Where there is a top pumparound heat-removal system, the external reflux can be reduced to zero.)
DISCHARGE CONDITIONS
The gas-compression load can also be reduced by lowering the compressor discharge pressure. This is achieved by lowering the absorber pressure.
Fig. 4 shows that this is a less-desirable move. A 20 psi decrease in absorber pressure is required to attain a 5% increase in compressor capacity. This is accompanied by a 1.5% reduction in propylene recovery.
Increasing the de-ethanizer pressure has a negligible effect on the absorber load but demands a significant increase in the stripper tower and reboiler load. The increase in reboiler temperature also reduces the heat-transfer rate.
Variation of the compressor interstage pressure has a minor effect on compressor capacity but the temperature of that system should always be minimized.
DE-ETHANIZER OPERATION
The stripper-absorber constitutes a de-ethanizer, the function of which is to control the ethane content of the C3 in the stripper bottoms, with minimum loss of propylene in the absorber offgas. The permissible ethane content of the stripper bottoms is a function of the destination of the propane product.
PRODUCT SPECIFICATIONS
If the propane/propylene will be processed through a polymerization unit or a sulfuric acid alkylation unit to produce propane LPG, the C2 content of the stripper bottoms should be about 4.0% of the propane content, to meet the propane LPG vapor-pressure specification.
Propane/propylene can also be processed through a propane/propylene splitter, or an HF alkylation unit, which yields its propane product from the bottom of an HF stripper. In those cases the ethane content must be 1% of the propane content, or even less.
For a typical FCCU, an ethane content of 1 mol % of the propane is equivalent to about 0.04 mol % ethane in the total stripper bottoms. This presents both analytical and control problems.
Fig. 5 shows the strong incentive to control the ethane content-of the stripper bottoms close to the maximum.
One VRU attained 88% propylene recovery with ethane equal to 0.5% of propane. After better instrumentation permitted the ethane content to be controlled at 4.0%, 93% of the propylene was recovered.
The economics depend upon the price of propylene. At 1990 Gulf Coast prices, the incentive for the above change was about 8.0/bbl FCC charge when producing chemical grade propylene, and 4.0/bbl when the incremental propylene was routed to polymerization.
The advantages of better control are not limited to the economics of product recovery. Fig. 6 shows that at high stripping rates (low C2 in stripper bottoms), the stripper and absorber loads are up to 50% greater. Thus the VRU capacity, and consequently the FCCU capacity, is correspondingly reduced.
STRIPPING CONTROL
Frequently, because absorber conditions are dependent variables, the only practical control over de-ethanizer performance is variation of the stripping rate.
Fractionators can be controlled by inferential controls such as top or bottoms temperature, which are indicative of product composition. But such controls are notably unreliable on VRUs. In fact, because they may give misleading signals to operators, the author often recommends their removal.
The C2 content of the stripper bottoms can be controlled by varying the stripper reboiler heat input. This is often done by controlling the stripper overhead vapor rate.
The success of that control strategy depends on the source of the most common process disturbances. The probable disturbances-FCC severity or charge quality, extraneous streams, and ambient temperature-should be analyzed to determine the best control system. Some type of ratio or multivariable control is often justified.
Stripping can be influenced by the high-pressure (HP) receiver temperature, which is normally determined by cooling water availability. The low HP cooler inlet temperature of about 150 F. reduces the danger of water-side scaling. Therefore, it is sometimes permissible to control the HP receiver temperature by cautiously throttling the cooling water.
Fig. 7 shows how this can reduce the stripper load while increasing the absorber load and decreasing propylene recovery.
ABSORBER CONTROL
Control of absorption, except when lean oil is a variable, is limited to minor temperature adjustments.
Lean oil and intercooler temperatures are the principal absorber variables and both are normally minimized. Fig. 8 shows an incentive of about 1% improvement in propylene recovery per 10 F. reduction in absorber temperature.
Any lean-oil subcooler and the absorber intercooler(s) should be kept clean. Cooling water can be routed in series through those items and the HP cooler. This not only provides a more favorable water-outlet temperature, but the higher velocity also reduces fouling and improves heat transfer.
It is noteworthy that the increase in propylene recovery gained by lowering absorber temperature incurs a smaller increase in tower loadings than any other means of improving performance.
SPONGE ABSORBER
The gas leaving the primary absorber is close to equilibrium with the distillate from the FCC fractionator overhead receiver. But this distillate contains C3s, C4s, and C5s, so the primary gas must also contain significant concentrations of those components. A secondary, or sponge, absorber is used for their recovery.
Fig. 9 illustrates the effect of variation of the sponge-oil rate on one VRU. Naturally, a low sponge-oil rate is adequate to control the loss of the heavier C5+ (gasoline), but higher rates are required to recover the C3s and C4s. The sponge-oil rate is conventionally adjusted to control the C4+ in the offgas at less than 0.5%.
The lean sponge oil is normally unstrapped light cycle oil or naphtha from the FCC fractionator. The rich sponge oil is returned to the fractionator. The rich sponge oil is in equilibrium with the primary absorber gas and contains not only the C3+ fraction, but also the C2- fraction from that gas.
In fact the "sponge cycle" returning to the FCC fractionator is typically about 40 mol % C2-. Fig. 9 shows that this parasitic compressor load can be 10% or more of total compressor capacity.
Sponge absorber economics cannot be ignored. The sponge absorber recovers gasoline and C4S of relatively high value.
Fig. 10 illustrates the economics observed at one Gulf Coast refinery in 1990. But if compressor capacity is limiting, a refiner must carefully adjust the sponge-oil rate to optimize recovery against the potentially large loss of FCCU capacity or severity.
Sponge-oil rate may not be a completely independent variable because it can have a significant effect on the FCC fractionator heat balance and is sometimes used to control that fractionator. In addition, sponge absorbers are prone to foaming.
This problem can be mitigated by operating at a slightly higher temperature, or by injecting an antifoaming agent.
A more permanent solution is the use of a lighter stream, such as naphtha or heavy gasoline, as lean sponge oil.
EXTRANEOUS STREAMS
Refiners, especially those without a saturate gas plant, are tempted to process drips and wet gas streams from cokers, reformers, and other processes, through the VRU. While the economics can be attractive, this must be done with caution because VRU recovery will decrease with increasing dry gas production.
The effects vary greatly with stream compositions but, as a rule of thumb, a 10% increase in dry gas production can cause a 20% increase in propylene loss. The dilution of the olefinic C3/C4 stream with saturates can also have an adverse effect on downstream alkylation and other processes.
Processing varying quantities of extraneous streams can pose control problems for operators. For example, an increase in the amount of extraneous C3s and C4s demands an increase in stripping rate, but causes a decrease in the reboiler temperature.
The diligent operator who holds the reboiler temperature constant will experience an increase in C3 losses, which can easily be greater than the amount charged for recovery.
Problems caused by the variability of extraneous streams are generally easily observed and diagnosed. However, smaller variations in FCC yields are less obvious but have a similar effect because the stream is larger.
DEBOTTLENECKING THE VRU
Debottlenecking a VRU can inexpensively release unused FCC capacity. It can also defer the capital expenditure of a major VRU revamp, which may involve packing towers or changing their service, and may also result in extended downtime.
VRUs often limit FCC operations and profitability. Before embarking on a major revamp it is always necessary to investigate what can be gained from modification of the auxiliary equipment in the unit.
A typical VRU that was constrained by stripper flooding was studied (Fig. 11). The refiner had bypassed about 20% of the FCC gasoline around the VRU by drawing a heavy naphtha sidecut from the FCC fractionator, but propylene recovery had fallen to 87.5%.
STRIPPER CHARGE PREHEAT
Preheating the stripper charge will reduce the stripper tray loadings. The installation of a preheater to increase the charge temperature from 100 F. to 140 F. achieved a 16% reduction in vapor load and a 7% reduction in liquid load.
The preheater duty was about 30% of the former reboiler duty, which was reduced by 15%. The operation is therefore less thermally efficient, but preheat would be obtained by exchange from debutanized gasoline at zero cost. The 15% reduction in FCC fractionator heat removal would have to be compensated for on that heavily loaded tower.
The less-selective stripping lowered propylene recovery to 85.5% and increased primary absorber loadings by 1-2%, both of which are undesirable.
Table 1 summarizes the effect of this sequence of process changes.
HP RECEIVER COOLING
High-pressure receiver cooling will tend to offset the effect of stripper charge preheat. The addition of a stripper charge preheater will increase the HP cooler duty by about 50% of the preheater duty.
In a loaded plant this will probably demand an increased HP cooling area. The refinery can make a generous addition to its cooling area and lower the HP receiver temperature. In this case the receiver temperature was lowered from 100 F. to 90 F.
The primary absorber loads were lowered to close to the initial values and the propylene recovery returned to 87.1%. On mann, loaded VRUS, the HP receiver operates at over 100 F., and increasing the area of the cooler will produce an even greater benefit.
ABSORBER CHILLING
Absorber chilling will improve propylene recovery and allow 50 F. to be achieved with a simple propane-refrigeration cycle.
This was found to increase the propylene recovery to the target 92.8%. The stripper load was relatively unaffected but the higher recovery increased the absorber-vapor load by 5% and the liquid load by 8%.
Chilling is not inexpensive. Typically, this application demands about 0.3 MBTU or 30 hp/1,000 b/sd of gasoline.
Chilling also increases the possibility of a free-water phase on the absorber trays, and appropriate provision must be made for water rejection in the distillate and intercooler streams.
LEAN-OIL PRESATURATION
This can be used to lower the primary-absorber liquid load. If the quantity of naphtha bypassed is further increased and replaced by a similar volume of lower-molecular-weight debutanizer bottoms recycle, then the "moles/hr" of absorber lean oil are increased, with beneficial effect.
Because the debutanizer bottoms is (or should be) debutanized and the FCC fractionator overhead distillate contains as much as 20% of butane and lighter components, it is desirable to contact the primary absorber offgas with this "leaner" lean oil in a presaturator before it is mingled with the principal lean-oil stream.
The addition of a presaturator chilled to the same 50 F. increased the propylene recovery to 93.4% while the primary absorber liquid load was also reduced.
SPONGE CYCLE
The sponge cycle of absorbed C5- back to the FCC fractionator is reduced when the primary absorber offgas is contacted with the cold debutanized stream. The lean sponge-oil rate can therefore be lowered.
Propylene and butylene recovery will be reduced somewhat. However, the parasitic compressor load from the C2- absorbed in the fat sponge oil will be less.
Often a reduction in compressor load can be translated directly and profitably into FCC capacity or severity.
Reducing the lean sponge oil to 20% of the original rate lowered propylene recovery by 2.1% and released 5% of the compressor capacity to salable products.
A fat-oil flash can also be used to reduce the parasite compressor load. The arrangement shown in Fig. 11, Item 5, rejects some of the coabsorbed C2- and releases about 2% of compressor capacity while the propylene recovery drops by 0.3%.
This can be a useful device when propylene recovery can be traded for compressor capacity. The depth of the flash can be adjusted by varying the pressure and temperature.
It is not necessary that the entire debottlenecking sequence of this study be followed.
The various steps are freestanding and can be executed independently.
These cases assumed that all the fractionators were fully and equally loaded. This is rarely true and advantage may be taken of available capacity in one piece of equipment.
Advanced control of some kind will probably be necessary to consistently achieve the benefits of VRU debottlenecking. To achieve a payback it is necessary to ensure that the process is consistently operated against the remaining equipment constraints.
It is very difficult to do this manually.
It is impossible to calculate the predicted flood point of any fractionator in a dirty system processing streams with varying physical properties.
It is essential to provide the instrumentation to physically monitor equipment loadings and to initiate the response necessary when those constraints are approached.
Copyright 1992 Oil & Gas Journal. All Rights Reserved.
