Search

KOEBEL (Grace Catalysts Technologies)

]I am going to take the question in a few parts. I will cover the reactor side first. In the reactor side, dipleg plugging will generally be due to coke formation that can be subdivided into two categories: the coke formation that occurs either internal to the cyclone or externally. On the gas outlet tube of the cyclone, you will see the stereotypical coke formation on the backside of the gas outlet tube, perhaps from incomplete feed vaporization. It could also be a result of running a very low conversion operation for some time and leaving crackable material still in the slurry. You end up with a sheet of coke that forms on the back of the gas outlet tubes. Then, during a thermal cycle, that sheet of coke will fall off to the bottom of the dipleg; and since it is bigger than the dipleg diameter, it blocks the flow of the dipleg.

Alternatively, you can get coke formation external to the disengager at the diplegs. In these cases, the coke formation could be due to inadequate purge steam or dome steam. The dipleg valve can become ensconced in coke, as you see in the photograph. Physically, it just will not move. Effectively, you end up with a plugged dipleg. So, the question asks: What are the successes with clearing these types of blockages online? Because these are just physical obstructions, there is little chance that you will be able to fix them with the unit online.

On the regenerator side, one common method of dipleg plugging is due to low melting point eutectics that you associate with iron and calcium contaminants in FCC catalyst. These contaminants tend to form a crust on the outside diameter of the catalyst particle. As the catalyst circulates through the unit, the fines chip off the outside and become extremely enriched in iron and calcium, which exaggerates the formation of eutectics. The formation of these eutectics is why iron, and calcium close off the porosity of the catalyst and become a problem with conversion. That same phenomena can cause the fines to melt locally, especially in high velocity areas where you can get some friction. Combine that with the temperatures in the regenerator, and you could develop a very solid cement-like deposit. I have seen it in the tops of cyclone diplegs, perhaps where the vortex extends down into the top of the dipleg. You get a disc of this sintered catalyst at the top of the dipleg. Again, those are very cement-like deposits, and they are hard to remove online.

The data on the slide is an example of an analysis of such a deposit. You can see that it has rare earth, as well as all of the other components you associate with FCC catalyst. Relative to the e-cat, the deposit is essentially double the contaminant level of iron and calcium. So, one way to avoid this type of deposit is just to keep iron and calcium under control to the maximum extent possible.

Finally, one of the other causes of dipleg pluggage is fluidization issues around diplegs, particularly submerged diplegs. If you have diplegs that are submerged in the catalyst bed – either in the reactor or the regenerator – and if you have fluidization issues, either you are losing catalyst, and your average particle size is getting high, or you have a damaged distributor or just poor distribution in general in the bed. You can get localized defluidization around the dipleg opening, which results in the effective blocking of the flow out of the dipleg. You then get catalyst carryover from that type of event. For this type of blockage, you will have some moderate degree of success clearing these online. One option is to lower the reactor or the regenerator level down to below the bottom of the dipleg opening, raise it back up, and then re-seal the dipleg in an attempt to reboot the fluidization of the bed. Some refiners have success doing pressure bump procedures to help with fluidization in these cases.

KEVIN KUNZ [Shell Global Solutions (US) Inc.]

Shell has had quite a bit of experience with occasional dipleg plugging from either coke formation or catalyst hang-up. I very much agree with a lot of what Jeff Koebel of Grace said. The coke formation in the reactor side is most often found to be caused by the lack of feed vaporization, particularly in resid cat crackers, and typically toward the end of the run when the feed nozzles begin to wear. To counteract that, obviously, good feed distribution and minimum droplet size are required to minimize the time required to vaporize the feed as much as possible. There are not too many FCCUs with cyclones external to the reactor, but Shell does have one. In this case, it is critical to make sure you have good, uniform heat to the diplegs. Checking for cold spots is a good practice.

As Jeff also pointed out, one of the challenges – online cleaning of a coked-up reactor dipleg – is tough to do. However, sometimes increasing riser temperature or decreasing the sour water use has helped improve the feed vaporization, particularly toward the end-of-run. On the regenerator side, catalyst hang-up can have multiple causes. Uncovering the dipleg termination devices or pressure bumps, and even cooldowns, have been successful at times.

Diplegs which contain obstructions that are operating with faulty termination systems are less likely to be successfully cleared online. Keep in mind that when planning and executing pressure bumps, the key is to do it very safely. You must consider everything that could go wrong and make sure that all of your proper safeguards are incorporated to prevent the plant or other unit shutdowns or upsets. This is very critical.

LEE WELLS (LyondellBasell Industries)

In the past, we have had issues with catalyst becoming defluidized in the reactor cyclone diplegs and plugging up the cyclones. Since we have improved our riser termination devices and lowered the catalyst flux in those diplegs, it has gotten even worse. But we have had success doing pressure bumps on the reactor side to free up that catalyst. The trick is recognizing that you are losing the catalyst before you carryover too much of it into your main fractionator.

MICHAEL WARDINSKY (Phillips 66)

One time we plugged up the diplegs in a third-stage separator due to what we believe was a combination of about three factors. The first was the existence of a lot of circulating fines in the inventory due to an attrition source that had developed during a startup. The second was relatively high concentrations of alkaline metals being present in the circulating inventory. We had been using a SOx reduction additive that contained a high level of magnesium. The third factor was that the unit experienced an exotherm due to a reversal scenario. The combination of those three factors led to the formation of a very hard ceramic like deposit in the TSS diplegs. I think we had to use a diamond-tipped drill bit to get them out.

JEFF KOEBEL (Grace Catalysts Technologies)

Dipleg deposits or blockages that lead to excessive catalyst losses happen as a result of a variety of causes. On the reactor side, dipleg plugging or blockages are often the result of coke formation. Coke can form on the internals of the cyclone gas outlet tubes, opposite from the gas inlet duct. This coke formation is generally stable and can build up to a thickness of several inches during normal operation. These sheets of coke can become dislodged during a thermal cycle and fall into the cyclone body where they come to rest at the top of the cyclone dipleg. This physical blockage restricts the flow out of the cyclone. The key to preventing these deposits is to prevent the coke formation in the first place. Make sure that feed is adequately atomized in the riser feed distributors because unvaporized feed can travel up to the top of the riser and lay down as coke in the cyclones. Additionally, a catalyst with adequate active alumina for bottoms cracking is critical to assure proper feed vaporization and pre-cracking, particularly when running a feed with any resid content.

Coke formation can also come in the form of generalized reactor coking throughout the reactor vessel. In this scenario, coke can form on dipleg outlet valves. This coking can restrict the proper movement of these valves, leading to catalyst losses.

Generally speaking, with either of these types of dipleg deposits caused by coke formation, there is little experience with successfully clearing the deposits with the unit operating.

On the regenerator side, dipleg deposits are often the result of the formation of low melting point eutectics. These eutectics occur when specific contaminant metals – such as Fe (iron), Ca (calcium), and Na (sodium) – are elevated. These deposits form a crust on the outside surface of the catalyst that can chip off during normal operation and cause the FCC catalyst fines in the inventory to be very highly enriched with these contaminants. When the fines are circulating in high velocity areas such as the bottom of a cyclone dustbowl, at the top of a cyclone dipleg, or perhaps orifice chamber plates, the combination of the heat from the regenerator and local friction heating is enough to cause the fines to melt and form hard deposits at the point of first contact with the internals. These deposits, when analyzed, will typically show Fe and Ca enrichment about two times the level of contamination in the circulating catalyst inventory. Table 1 is an example of a typical analysis of these deposits compared to e-cat. The deposits also often show that the zeolite is compromised, as if exposed to high temperatures which would not be present according to the unit data. These types of regenerators dipleg deposits are very hard and must often be physically chipped out of the cyclone. They cannot be cleared online.

Finally, fluidization issues can also cause dipleg blockages. During times of catalyst losses, the FCC catalyst inventory can get very coarse with inadequate fines content. Fluidization issues can develop in a dipleg or around the dipleg openings for submerged diplegs, even if the catalyst circulation is unaffected by the change in the catalyst fluidization characteristics. This can be a problem, particularly after a turnaround when the catalyst inventory is reloaded into the unit. There has been some success with clearing these problems during unit operation. Pressure bumps or draining of the vessel level below the dipleg discharge level are solutions that have been successful in some circumstances and should be attempted as a last resort measure only, as the risk for a significant catalyst loss event is high. These types of blockages are particularly frustrating because they will often disappear during a unit shutdown when the catalyst is deinventoried, leaving behind no evidence as to what caused the blockage in the first place.

CATHERINE INKIM (PETROTRIN)

Though we have not experienced dipleg plugging or fouling, dipleg plugging can be a result of coke, refractory, or catalyst plugging. Typically, the diplegs in the reactor can be plugged by spalled coke after a thermal cycle or by defluidized catalyst in both the reactor and regenerator diplegs. Pressure bumps may help the latter.

CHRIS STEVES (Norton Engineering)

In reactor diplegs, plugs are often caused by coke, catalyst, and refractory. With the decreasing 6 oil market, the push to convert more of the bottom of the barrel is increasing. FCCs share in this effort, and some have been referred to as catalytic cokers. The heavy resids and tars are difficult to atomize and convert. Even if the conversion looks good, these high molecular weight molecules slowly build in all parts of the reactor, especially in areas of low velocity or adjacent to exposed “cold” surfaces. Thermal cycles from startups and shutdowns will eventually spall coke from the reactor internals, and this coke can end up plugging cyclone outlets and diplegs. Prevention of coke buildup is the key to avoiding this type of issue. Maintaining high enough riser temperatures to ensure vaporization of the feeds being processed and ensuring that cold wall reactors are well insulated to prevent coke formation is essential. Anti-coking baffles have successfully been installed in some reactors to allow for a steam purge of low velocity areas (typically above cyclones) so that coke will not form in these areas.

Catalyst can also plug reactor diplegs. Experience has shown the plugging is most frequent on unit restarts and likely from wet catalyst. The catalyst is wet either from oil or from steam/condensate from the restart or shutdown. To prevent wet catalyst from plugging diplegs, use of superheated steam in the stripper during startup, as well as early circulation of hot catalyst from the regenerator to the reactor, can help prevent wet catalyst from plugging diplegs.

In regenerator cyclone diplegs, refractory and defluidized catalyst can lead to dipleg plugging. Proper refractory inspection and repair during unit shutdowns is essential to prevent refractory damage and potential dipleg plugging. Defluidized catalyst “plugging” of diplegs may also occur, especially if the air distributors are damaged and zones of defluidized catalyst exist in the regenerator dense bed.

If dipleg plugging is observed during operation of the FCC unit (as diagnosed by high losses and sampling of the catalyst fines for PSD analysis), some online techniques have been successfully executed. Changes in bed level (for units with submerged diplegs) may help in changing the pressure balance in the cyclone dipleg enough so that the plug is dislodged. Rapid pressure swings on the unit can also be used to dislodge a blockage. With external diplegs, diagnosis of plugged diplegs can be easier (looking for cold spots). The use of vibrators, external heat, or a hot gas injection can be used to clear the blockage.

ROBERT TORGERSON and SYDNEY GARRETT (Gayesco International)

It is possible for diplegs to foul during the startup process if the dipleg is too cold. The resulting condensation can cause the catalyst to bridge and impede catalyst circulation. We have many customers who now require the attachment of removable skin thermocouples to the diplegs to monitor that temperature during startup. Removable thermocouples allow for easy replacement, simplified dipleg maintenance, and easier installation as the attachment hardware can be installed by the dipleg/cyclone manufacture during initial fabrication.

LARSON (KBC Advanced Technologies, Inc.)

This answer will be very similar to what was already discussed about how to treat the resids. The example shown on the slide is a Maya blend, a typical tight oil, and then a tight oil with resid. Again, we are seeing significant reductions in sulfur and Conradson carbon metals and also a much higher hydrogen content. So just by comparison, let us look at this at constant reactor temperature. What we would expect to see is low coke, which will raise our cat oil and pull a lot more heat out of the regenerator. We will get much more liquid yield, which is what is reported by those who are running more and more tight oil in the system.

The question is: What can we do to operate these units? You are going to have some open capacity. So operationally, what can you do to hydraulically increase the feed to the unit? Given our economics is more feed, more feed, more feed, how far can we push the “more feed”? On the heat balance side, consider first getting all of the diesel out of the feed so you do not have an additional cooling effect with material going through that will not crack very well. Look at processes like deep cutting your VGO. How far are you going into the material? Are you going to 1000°F to 1050°F? Can you go to 1075°F? There are some people who have shut down the vacuum columns and are actually just running pure catalyst from the tight oil into the unit. And, we have already had a question on processing whole crude.

So, there are actions to consider. Again, look at catalyst change. We have been taking recycle out of the riser for years. We might put it back in now as a way to manage the coke balance. You should consider these questions: 1) What is your standpipe flux? and 2) What is the maximum catalyst flux rate that you can tolerate in your standpipe? You will not get to choke flow. You will get to a point where it will not flow any faster. Look at the slide valve opening. Most units have been designed with a slide valve that will run approximately 50 to 60% open under normal operation. Now with tight oil’s higher circulation rates, you might be looking at a slide valve that wants to operate at a much higher percentage opening. That may be your first point. It is not the pressure drop. It is the feeling of running at 80 or 90% open. Resize the slide valve. It may be a very easy modification to get you where you can run much higher cat to oil rates.

Make sure that you have the right pressure balance. Now that you may not have as much air demand, maybe you can change the pressure balance of the unit to raise coke make in the reactor and improve the ∆P across the system. You will have to be aware of additional catalyst erosion if you go to higher pressure drop on your slide valves. Look at the velocity profile in the regenerator itself. Make sure that you can push the unit or that the unit is down to where you can operate in a comfortable range. Again, I look at this as a perspective of something that occurs today and is not happening tomorrow, so what should we do? We might back some steam off the feed nozzles to make a little more additive coke. Perhaps we will lower the stripping steam again to manage it with more coke in the extreme conditions; and I do say ‘extreme.’ Adding torch oil to the regenerator would be an extreme if you are just trying to keep the unit on. We do not recommend that because it substantially chews up the catalyst. Keeping the unit on would be a last resort.

We have heard of people wanting to start the air heater in the bottom of the regenerator to add enough heat to the system. That would be a control issue. There are a lot of issues around the air heater with temperature profiles and the reliability of the air grid that the metallurgy would have to be checked out for a continuous long-term operation. But fundamentally, check your catalyst type. What are you using? What can you change in the catalyst system? What can you do to adjust your feed quality? You add carbon to the system. In the U.S., it has to be moving to a higher and higher percentage of tight oils because we are finding it; and by law, we have to process it. So this would be a new challenge. I think the catalyst manufacturers have done a great job, and they will continue to find ways to give us a catalyst that makes a little more coke and gives us the yields we want.

GIM (Technip Stone & Webster)

Many of the available crude oil analyses of the tight oil show that the coke precursors in both the gas oil and the resid portions are very low. Those FCCs that were designed for heavy sour crudes may have some detrimental heat balance issues on the converter side, as well as light ends constraints on the gas plant side. I am going to discuss what you can do in terms of both operational and hardware changes. I will not repeat the redundant points that Mel Larson described previously.

First is riser outlet temperature. The obvious solution is to raise the riser outlet temperature not only to heat up the system but also to gain back some of the octane loss resulting from processing paraffinic feeds. This may be problematic because your gas plant will probably already be overloaded with the lighter ends stemming from the processing of tight oil.

Using higher feed preheat, if you have room, will certainly alleviate the higher catalyst circulation rates expected from the low delta coke feeds, such as tight oil.

Step jumps in the conversion level from the improved feed qualities need to be counterbalanced. Raising the reactor vessel pressure will certainly help in terms of increasing the unit delta coke directionally. That would add to the delta coke because of the higher hydrocarbon partial pressure. It will also help alleviate the light ends circuit in the gas plant. Reduction in dispersion steam and the addition of torch oil were already discussed by Mel.

Hardware Changes: I know it is a big-ticket item, but certainly the fired feed preheater could be an option. Again, that is with the installed TIC cost and permitting issues. We spent a great deal of our resources on conducting pilot plant testing of recycle streams in terms of both HCO (Heavy Cycle Oil) and slurry. Being derived from very paraffinic feeds, these high-quality feeds do not have much HCO and slurry yields to begin with. The amount of recycle stream that is available to recycle back to the riser is not there. So that is the problem one.

Second is the coke precursor for these HCO and slurry streams. Even if you are able to recycle them, you will not get as much coke out of these recycle streams as you would have in your conventional FCC feed.

Problem number three: We found that the actual conversion level for these streams – converting it into something other than the cycle oils, dry gas, LPG, gasoline, or coke – was also low. It was quite nonreactive in that sense and is another issue of recycling these streams.

Enlargement of catalyst standpipes and port openings of slide valves may be necessary for catalyst circulation step-jumps, even after making all other operational and hardware changes.

Continuous usage of air heater may also be evaluated, but this may require redesigning the air distribution system to ensure that the increase in the exit tip velocities will not result in unsustainable catalyst attrition.

PAUL DIDDAMS (Johnson Matthey INTERCAT, Inc.)

One additional suggestion might be to consider oxygen enrichment and reducing your air rate to decrease the heat losses a little bit. Of course, you might also consider running a residue stream in directly with your feed.

KEN BRUNO (Albemarle Corporation)

The panel did an excellent job reviewing the operational and hardware changes. I want to point out that to help manage these changes, it is critical to use the right catalyst. To that end, Albemarle has developed a tight oil family of catalysts, in particular Upgrader T and Amber T. Again, please consult the Answer Book response for more information on the application of these products, as well as all of the commercial experience we have with those products.

LARSON (KBC Advanced Technologies, Inc.)

I want to highlight the need for you to have a good baseline of where you are currently operating. Jeff talked about unit monitoring earlier. When you know that you will be processing a higher percentage of tight oil, to the extent that you can either ratably, put it in or put some material in storage and regularly bring it into the unit. Controllable rates of tight oil will give you the chance to learn what the effects are as opposed to jumping out and saying, “We can handle 50%.” Get yourself a really good baseline in advance; so that when you do make changes, you will have a better educated starting point.

PARAG KANADE (Lummus Technology)

What do you think will be the major challenges in the fractionator and the VRU (vapor recovery unit) section? You mentioned that there will be excess capacity in the compressor and the overhead circuit; but at the same time, the debutanizer stripper bottoms will be loaded. All of these are very heat-integrated with the pumparounds on the reboilers. To utilize the compressor, I think the incentive will be to push more fill through the unit. However, you will also have problems in the stripper debutanizer in the downstream unit. So, my comment is that a small study will be helpful to determine the right throughput to your unit which will avoid the need for a major revamp.

LARSON (KBC Advanced Technologies, Inc.)

The answer is that on a hydraulic basis, you will have to find your pinch point. You would like to be able to analyze it with kinetic modeling and seeing where your pinch points may come up. Also, I appreciate the heat balance comment. I think some of our panelists have actually had experience with this issue. As you process more tight oils, you physically cannot run the riser top temperature; i.e., the heat going to the main column is less because you cannot tolerate 90% plus conversion in the unit. You have a different balance when you run more tight oil than we have been running in the last 20 years. So, there will be different pinch points.

MICHAEL WARDINSKY (Phillips 66)

I cringe every time I hear people talk about using torch oil to maintain regenerator temperature above some minimum for an extended period of time. Make sure you understand the environmental and process safety implications of trying to run long-term on torch oil. In addition, the catalyst will suffer accelerated attrition and activity loss. There are environmental concerns about firing torch oil. Any time you put torch oil in the regenerator, your CO emissions will go up. In some units, I think you will be challenged to stay under a 500-ppm limit. You also have process safety issues that deal with the consequences of the blower tripping. If there is no interlock on your torch oil supply system, you will be injecting a lot of fuel into your regenerator and could have a real problem. So, make sure you understand those consequences before you use torch oil outside of startup.

LARSON (KBC Advanced Technologies, Inc.)

Just as a follow-up, my comment on that was generally not regarding refineries in the U.S. We have people outside the U.S. who are, in fact, using torch oil and who do not have the same EPA and environmental constraints, or even some of the other safety issues you mentioned. I totally agree with what you are saying from our position here in the U.S. The window of operation is much tighter here than it is in other countries.

MEL LARSON (KBC Advanced Technologies, Inc.)

For the last 20 years, the industry has been driven by tighter gasoline and diesel sulfur specifications, as well as increased demands for propylene. The catalyst manufacturers and licensors have done a great job making improvements that focus on maximizing the profit from these operations.

Now with a greater percentage of tight oil in the refinery, what are the changes? The tight oil might be similar to a severely hydrotreated virgin gas oil FCC feed. The typical shift from typical WTI (West Texas Intermediate) to tight oil (not all tight oils are equal) will be as a result of:

• Increased hydrogen content;

• Lower contaminants, such as Conradson carbon, nitrogen, and sulfur; and/or,

• Low aromatic content.

As previously mentioned, the diesel content of the FCC feed should be reduced as much as possible since light boiling feed will reduce regenerator temperature. One point: Diesel quality from the tight oil has greater value as diesel than as FCC feed. Secondly, consider substantially increasing the HVGO/resid cutpoint or checking the quality of the overflash from the vacuum column as a potential feed to the unit.

As feed quality improves, the regenerator bed temperatures reduce, and catalyst circulation rates go up to a “limit”. There are a number of steps recommended to define the real limits in processing tight oils:

1. Define a base line of the circulation rate. This step is crucial for understanding the current or anticipated changes with tight oil. The base line should be consistent with typical or routine operation. Included in this step would be a single-gauge pressure survey to aid in system definition and base line.

2. Identify the limiting issue, i.e., the flux in the standpipe, the residence time in the stripper or regenerator, the slide valve (SV) position(s), or ΔP across the SV. From a flux rate perspective, there are a number of units operating successfully at mass flux rates in the standpipe well over 300 lbs/sec/ft2 (pounds per second per square foot). Historically, FCCs used to operate successfully with lower regenerator bed temperatures in the range of 1200°F to 1250°F and residual carbon on catalyst at levels of as high as 0.30 wt%.

Operational considerations are similar to those discussed previously:

• Consider different feed sources to add carbon, such as vacuum overflash and a Slip-Strip™ of vacuum residua.

• Change the pressure balance.

   – Move from partial to total CO combustion in the bed, not just in the regenerator proper. Add (more) promoter to the regenerator.

• Change the catalyst.

• Recycle a cycle oil to add carbon.

• Increase reactor temperature.

• Use air heater continuously.

• Add external heat oil direct into the regenerator.

Hardware changes to consider for accommodating higher catalyst flux rates would include, but not be limited to, the size of the standpipe and slide valve and capacity expansion, given that the air blower is less constrained.

STEVE GIM (Technip Stone & Webster)

Cracking of paraffinic FCC feed derived from tight oil crude behaves similarly to severely hydrotreated feeds or cat feeds from other sweet crudes like West Texas Intermediate. As many of the available crude assays show, these crudes do not have many coke precursors in the gas oil or even in resid fractions. FCCs designed for sour heavy feeds may face heat balance constraints in the converter (reactor/regenerator) and light ends circuit constraints in the gas plant. Sometimes, too much of a good thing can be a bad thing.

Operational Changes

Higher reactor outlet temperature may be necessary not only to heat up the overall system, but also to make up the gasoline octane loss stemming from the paraffinic feed. This may not be easy, however, given the fact that the lighter feed would be already taxing for the light products circuit in the gas plant. Higher catalyst activity accompanied by more matrix content will help increase the catalytic coke.

Talking about catalyst, for partial-burn units, the addition rate of CO promoter can be increased to preferentially raise the regenerator bed temperature.

Higher feed pre-heat (if there is room in the pre-heat train) can help alleviate the expected increase in the catalyst circulation rate and also counterbalance the step jump-in conversion due to improvement in the feedstock qualities and higher operating severity I just described.

Raising the reactor pressure will directionally increase the delta coke due to the higher hydrocarbon partial pressure. It will also help alleviate the increased burden in the lighter ends of the recovery section.

I do not like this idea as much, but lowering feed dispersion can help increase the delta coke. There will be some debits in selectivity.

I also do not like the idea of injecting torch oil into regenerator, but that is always an option.

There is a possibility of a compatibility issue of highly paraffinic tight oil and asphaltenes. It could form two phases and may have to be injected separately into the riser.

Hardware Changes

A big-ticket item, in terms of cost and permitting issue, is to install a fire feed pre-heater to counteract the lower system-wide Btu (British thermal unit) posed by the lighter fresh feed.

Recycling HCO and slurry streams (preferably upper section, otherwise they will be entrained to regenerator) into stripper may also be an option. We have done some pilot tests of these recycles from light feeds. There are a few not-so-obvious issues. Issue 1: Quantities of these recycle streams are lower than those from heavy feeds since conversion levels are expected to be higher. Issue 2: Coke precursor level is low for even for these bottoms. Issue 3: Conversion level (i.e., converting into something other than the cycle oils such dry gas, LPG, or gasoline) is low. Depending on the starting point of delta coke for the operation without the recycles, only recycling the heaviest cut may be a delta coke additive.

Enlargement of catalyst standpipes and port openings of slide valves may be necessary if catalyst circulation step-jumps even after making all other operational and hardware changes.

Continuous usage of air heater may also be evaluated, but this may require redesigning the air distribution system to ensure that the increase in the exit tip velocities will not result in unsustainable catalyst attrition.

RAUL ARRIAGA and KEN BRUNO (Albemarle Corporation)

Tight oil (TO)-derived feeds, in general, show high API gravity and a high amount of paraffins compared to traditional VGOs being processed in FCCUs. On the other hand, TOs tend to come with higher amounts of some metals, particularly iron and calcium. This combination of properties makes TOs more crackable than traditional VGOs due to their molecular distribution, but it also results in a lower contribution towards delta coke, lower regenerator temperature, and increased catalyst circulation provided the iron and calcium are low. On the other hand, if the metals (usually iron and/or calcium) are high, TOs are conducive to increased deactivation rates, especially pore mouth blocking, which tends to increase catalyst mass transfer limitations and FCCU slurry yields. The increased slurry production can push a refiner to reduce feed rate in order to stay within limits, resulting in economic losses. To overcome mass transfer limitations and optimize delta coke, high accessibility catalysts such as AMBER™ and UPGRADER™ have been developed and proven successful for tight oil applications.

Based on the above, operational changes that can be employed to address heat balance issues include increasing activity via higher catalyst addition rates and HCO or slurry recycling. Another option is the processing of bi-modal types of feed; meaning that as a refiner increases the amount of TOs in the feed, the refiner could compensate for the TOs by lowering the API gravity of the rest of the material in the blend. However, care needs to be taken regarding the compatibility of the different components or asphaltene precipitation may occur.

It is not recommended to reduce stripping residence time as valuable product is burned in the regenerator. It is also not recommended to use torch oil to increase regenerator bed temperature due to the possibility of drastically higher catalyst deactivation rates and lower e-cat activity, which would create a negative contribution to delta coke and likely result in increased catalyst addition.

KOEBEL (Grace Catalysts Technologies)

The schematic on the slide shows the representation of the coke yield and the coke balance from the FCC. Of course, the total overall weight percent coke yield is set by heat balance, but the sources of the coke vary significantly from one feed to the next. Everyone talked thoroughly about how the coke precursors are just not there in these lighter feeds. The contaminant metals will be drastically lower, so your FCC catalyst will be called upon to provide a much larger percentage of the overall coke yield than it does in a normal operation. That can call on the unit to run a much higher catalyst circulation rate than it is physically able to do, so certainly you change the catalyst as warranted in these instances.

The next slide is a quick pilot plant representation of what happens in these cases. The darker blue represents a base VGO feed and the lighter blue: a shale oil type of feed. I will just pick a number here and say 2.5% coke yield in the pilot plant. In order to do that, we need about 5.5 cat-to-oil, which results in about 74% conversion in this operation on this catalyst. If we do nothing to the FCC catalyst and run the lighter feed, you can see that the required cat-to-oil to generate even that same 2.5% coke yield jumps up to 8%. So a full 50% increase in the catalyst circulation rate is required to keep the unit operating. We are talking about having to actually increase the coke yield, not hold it constant. That coke yield results in a little over 76% conversion. Clearly, a catalyst reformulation to a higher activity is in order here.

BULL (Valero Energy Corporation)

I am probably going to commit sacrilege here, from a catalyst vendor standpoint; but in many cases, a shift to a less-coke-selective catalyst can help process these increased amounts of tight oil-derived feeds. A catalyst with less zeolite and a higher quantity of non-active, non-coke-selective matrix can help you maintain a minimum regenerator temperature. We have actually done this at one or two sites.

JOE McLEAN (BASF Corporation)

I will go back to the previous question that we had before the break about the benefits of catalyst porosity. One thing you can do is reduce the porosity to reverse the delta coke benefits. We have a number of customers who have gone back to more old-style, lower porosity catalysts that are less coke-selective but just as active. These catalysts work quite well in this application. Because of the increase in conversion and, specifically, a big boost in LPG, we have seen at least one client take out ZSM-5 completely and still maintain the same light olefin yields that were in the previous operation with a fair amount of ZSM-5. So, in that case, the side benefit was the ability to save on the additive cost.

KEN BRUNO (Albemarle Corporation)

We agree, Jeff. There are cases where a lower coke selectivity is beneficial. But related to accessibility or the diffusion character with this kind of feed, quite often there is overcracking or secondary reactions that you do not want. So, it remains critical to have the right accessibility and porosity to minimize those secondary reactions.

WARREN LETZSCH (Technip USA)

I want to remind people that the pilot plant data is quite accurate. But when you start increasing catalyst circulation rate like this, chances are that the stripper performance may well deteriorate. You will have a much shorter residence time; and basically, you may end up pulling a lot more hydrocarbons through. It will be different for every unit, depending upon where you are operating and the type of equipment you have in it. I think if you have what I would call conventional a disk-and-donut type of stripper, then these types of circulation rates will almost guarantee that you will need to have much higher hydrogen on cokes because the flux rate will be very, very high. So, every situation is really different.

JEFF KOEBEL (Grace Catalysts Technologies)

Delta coke is the difference between carbon on spent catalyst as it leaves the stripper and carbon on regenerated catalyst. Delta coke is the primary variable that determines the regenerator bed temperature.

There are four primary sources of coke in the FCC process. They are: feed Conradson carbon, contaminant metals, stripping coke, and the catalytic coke produced by the FCC catalyst. The sum total of these four components adds up to the total coke yield in the FCC. Since the FCC heat balance determines the overall weight percent coke yield in the FCC process, a reduction in the contribution of one of these four coke sources must typically be offset by an increase in one of the other sources. For example, let us consider an extreme case, which is represented in Figure 1, of a unit that experiences a change in feedstock from resid feed to a light shale oil feed. When the unit shifts from heavy or resid feed to lighter feed, the total weight percent coke requirement does not necessarily change; however, the contribution of coke from each source will shift. Assuming stripping coke stays relatively constant, the feed contributes less to the required coke (feed carbon and contaminant); thus, the catalyst must make up the difference.

That means that the FCC catalyst will contribute a larger percentage of the overall heat required for the process. If the catalyst is not active enough, the catalyst circulation rate must increase so that conversion, and thus the coke yield from the catalyst, can increase to satisfy the FCC heat balance. This will lower the delta coke and the regenerator temperature. For a set riser outlet temperature, the lighter feed will require much higher catalyst circulation rates to satisfy the FCC heat balance. If the FCC catalyst section cannot physically circulate enough catalyst, it will be necessary to either reduce the unit charge rate or the reaction severity to stay within the FCC catalyst circulation limit.

In the pilot plant example below (Figure 2), an FCC unit operating on standard VGO is contemplating a move to lighter shale oil feed type. The base case catalytic coke of 2.5 wt% requires a cat/oil ratio of about 5.5 and results in 74% conversion. In order to keep the 2.5% coke yield with the lighter shale oil feed, a cat/oil ratio of over 8.0 is necessary with an increase in conversion to about 77%. Most FCC units are not capable of this dramatic increase in the catalyst circulation rate, and the catalyst circulation hydraulics will likely limit the unit severity or throughput.

In this same example, we consider a catalyst reformulation to a more active catalyst with a different coke to conversion relationship (Figure 3). Here, Catalyst A is applied, and a much more modest cat/oil ratio of 6.5 is necessary to satisfy the coke yield. This is due to the inherent catalyst activity of Catalyst A. Because of the coke to conversion relationship of Catalyst A, higher conversion is achieved.

Grace has had multiple experiences with reformulations such as these for processing lighter feeds from shale oil or traditional hydrotreated FCC feed. Using a high activity catalyst is required to counter the effects of low delta coke, but it is important to select a catalyst with the proper coke selectivity (coke-to-conversion relationship).

Lastly, another issue with processing shale oil is the possibility of Fe and Ca contamination. To minimize the effect of Fe and Ca poisoning, a high porosity, high diffusivity catalyst should be considered. Since processing shale oils will often result in both issues (lower delta coke and high Fe/Ca), a high activity and Fe/Ca tolerant catalyst should be considered.

RAUL ARRIAGA and KEN BRUNO (Albemarle Corporation)

From a catalyst formulation point of view, it is recommended to increase catalyst activity and tune the selectivity's, including delta coke, to the desired targets while keeping the FCCU against its constraints. When optimizing the catalyst formulation, it is important to maintain or improve the mass transfer character of the catalyst. The objective is to prevent increased pore mouth blockage rates due to the higher amount of iron and calcium often observed in tight oil (TO)-derived feeds.

It is better to achieve higher catalyst activity with the use of high accessibility technology than with additional active ingredients, particularly zeolite, in order to maximize catalyst tolerance to iron and calcium. If the catalyst applied does not have the optimal pore and surface architecture, the result could be increased slurry yields and additional bottlenecks. Albemarle’s AMBER™ and UPGRADER™ are proven catalysts for use with tight oil.

One tactic to increase activity is to raise the amount of rare earth on zeolite for increased delta coke. While case-dependent, another approach for consideration is to reduce the use of vanadium traps or nickel-selective matrices which would enhance the metals contribution to delta coke. By the same token, if a refiner is consuming a flushing catalyst of any kind, it is recommended to re-think that strategy and evaluate reducing its utilization because lower use of a flushing medium may be desired to let metals concentrate on the catalyst. It is recommended to consult with various catalyst suppliers to compare the merits of each manufacturing technology and for commercial references with this new type of feedstock.