AFPM Q&A - 3 Discussion turns to fluid catalytic cracking

Oct. 5, 2015
This is the final of three articles that present selections from the 2014 American Fuel and Petrochemical Manufacturers Q&A and Technology Forum.

This is the final of three articles that present selections from the 2014 American Fuel and Petrochemical Manufacturers Q&A and Technology Forum (Oct. 6-8; Denver). It highlights fluid catalytic cracking (FCC) processes, including issues related to feed and profitability.

The first installment, based on edited transcripts from the 2014 event (OGJ, Aug. 3, 2015, p. 52), addressed gasoline processing operations, with a focus on safety, blending, and reforming issues. The second (OGJ, Sept. 7, 2015, p. 88) continued the discussion of safety and added discussions of mechanical integrity and profitability related to hydroprocessing.

Brazil's Petroleo Brasileiro SA (Petrobras) completed an expansion and modernization project in 2014 at its 415,000-b/d Replan refinery in Paulinia, Sao Paulo (OGJ Online, Aug. 4, 2014). The $5 billion overhaul included the addition of installations designed to produce cleaner derivaties as well as integration of processing units to help boost output of diesel and gasoline from the refinery, which is the country's largest. Photograph from Petrobras.

The session included four panelists comprised of industry experts from refining companies and other technology specialists responding to selected questions and then engaging attendees (see accompanying box) in discussion of the relevant issues.

The only disclaimer for panelists and attendees was that they discuss their own experiences, their own views, and the views of their companies. What has worked for them in their plants or refineries might not be applicable to every situation, but it can provide sound guidelines for what would work to address specific issues.

Process

What are the options for removing catalyst fines from the main fractionator-bottoms product? Which, if any, can reduce ash content to 50 ppm or less?

Wojtowicz: There are four main methods to reduce the ash content of slurry oil:

• Hydrocyclones (also known as slurry settlers).
• Tank settling.
• Mechanical filters.
• Electrostatic filtration.

Slurry settlers are less efficient. They are a safety hazard and being phased out of the industry. These are not recommended by UOP. Tank settling, mechanical filtration, and electrostatic precipitation can all achieve the 50-ppm target. Tank settling uses additives. It could be a maintenance headache as the sludge that collects at the bottom of the tank has be to removed and disposed.

Mechanical filtration uses sintered metal cartridges with very fine pore size. Filter cake builds up on the filter elements to filter the slurry oil. After the pressure drop reaches a certain value, they must be backflushed. Usually light-cycle oil (LCO) or heavy-cycle oil (HCO) is used. These are sensitive to temperature and composition, especially asphaltenes. UOP only recommends mechanical filtration for units processing vacuum gas oil (VGO). One vendor, Filtrex, has a rotating mechanical filter that uses a fine-mesh screen. This is relatively new to the FCC industry, although it has been used in hydroprocessing quite a bit. We have only one unit that utilizes this filter.

Electrostatic filtration uses beads and an electrical current to filter the slurry product. This type of filter has a larger backflush requirement, so raw oil is typically used. Any of the options that have backflush will increase the fines in the circulating slurry, typically, two to three times. The backflush will also increase your stack emissions out of the regenerator, which needs to be considered.

Mason: I am glad to hear that the settling aids are allowed and used. With regard to the tank configuration, it plays a big part in getting the levels down and reducing your chemical usage. Cascading tanks in series help a lot, rather than just having a live tank. One way to address the real issue, if the product works, is to settle out cat fines, solids, ash, and sludge that will build up over time. What have we done when it was time for a tank turnaround? In the past, we worked with a refiner to ensure that there was upgrade value when taking that decant oil to a #6 fuel oil, for example. The refiner took a portion of that benefit that we gave up. They put it in a fund, like a rainy day fund. When that tank did come up for turnaround, they were prepared.

Nikitczuk: In Phillips 66, we actually have three refineries with sintered-element filtration systems. All of those consistently filter to less than 50 ppm. The filtration system, however, has created operational headaches, particularly with the programmable logic controller during backwash.

Diddams: Another consideration: We do not like fines in the first place, right? Operationally, many of the fines that end up in your slurry are formed in the feed-injection zone of the riser. So you want to be looking at your superficial velocity at your feed nozzles. Feed nozzles will come with some design criteria. What is the maximum superficial velocity? Some people say 107 m/sec; some say 120 m/sec. Just respect your licensor's value for the feed nozzle tips with superficial velocity to avoid excessive catalyst attrition.

Wardinsky: If you choose to recycle a backwash stream from one of these systems back to your riser, you will need to develop some type of monitoring tool to look at erosion of your reactor overhead line. We eroded through a reactor overhead line on a unit that backwashes a fines-rich stream to the riser when an attrition source developed from a damaged steam distributor in the reactor stripper. The increased catalyst losses from the attrition, along with normal fines -load across the overhead line due to the backwash, accelerated erosion in the overhead line and caused a leak that forced the unit down for repairs.

Going back to a comment I made earlier about preparing for turnarounds and cyclone life, you can also develop a simple monitoring tool that combines catalyst loading to your reactor overhead line, along with line velocity, to estimate the erosion rate of the overhead line.

Ludolph: To add to Mike's comments, I suggest that you review the nozzle design for the return material to the riser. If the design does not atomize or distribute the material well, then coke deposition in your riser could occur. This would create a source of backpressure and maldistribution in the riser, which would lead to problems over the course of your run.

Haynes: Slurry oil catalyst fines settling-aid chemistries have been used for many years in this type of application. Electrostatic precipitators and filtration equipment are also available for minimization of ash content in slurry oil. These technologies have demonstrated cases of reduction to below 50-ppm ash for FCC slurry oil product.

Savage: Ash is a particular problem for slurry oils, especially those that are heavy and viscous and which need long residence times to allow for catalyst settling. The source of ash in the slurry oil stream is composed of catalyst fines carried over into the fractionator from the reactor section of the FCC. The reactor cyclones are the first point where catalyst fines are removed from the reactor vapors going to the fractionator. Although highly efficient at separation of the vapors and catalyst, a certain level of fines is always present in the slurry oil stream. FCC hardware manufacturers are continuing to make improvements in cyclone design to remove a greater fraction of the fines.

To obtain low ash, special techniques (such as heating, chemical additives, filters, electrostatic precipitators, centrifuges, and cyclones) might be used. Catalyst selection may help reduce attrition to a great extent.

Many refiners make use of slurry filtration devices to remove as much of the fines as possible and recycle the fines back to the FCC reactor by backwashing the filters. These devices can reduce ash in slurry but may require frequent maintenance and a significant capital investment for the equipment and installation.

Tank settling is the most common means of reducing the ash content of slurry oil. Often, limited tank capacity reduces the residence time available for settling of the ash. Many refiners "de-ash" with chemical settling aids which speed up ash settling in storage. These chemicals are polymeric compounds that adhere to the catalyst surface, causing agglomeration of the fine particles in order to accelerate separation.

In order to consistently meet very low ash content, a combination of the options described above may be required.

How do nitrogen compounds distribute in the product streams of FCC units? What effect do riser severity and feed properties have on this distribution?

Devine: The way I approached this was that I did a little literature review and then checked that against some of our operating data from our units. In terms of the last part, we have to deal with nitrogen compounds. Obviously, the distribution is a function of crude selection and operations. Really, hydrotreating is the best method for reducing your nitrogen in the feed.

Feed nitrogen is basic and nonbasic. The basic nitrogen accounts for 30-50% of the feed. Almost all of that feed ends up on your coke in the catalyst, and then it goes to the regenerator. Increasing conversion will directionally put more of the nitrogen on the coke. In the regenerator, of the nitrogen that goes over the coke, 90-95% goes to nitrogen and the rest goes to nitrogen oxide (NOx) or intermediates, ammonias, or hydrogen cyanides, depending on if you are in partial or full-burn.

On the product streams, the literature suggests that about 50% ends up in your slurry of either HCO and LCO and 10-20% goes into your LCO. I do not have sample data for the slurry in the HCO; but on our LCO, I will stick with 13-16%. So that was consistent. Increasing conversion should reduce the amount of nitrogen in the LCO. And then on the gasoline cut, it is usually less than 5%, which is consistent with our lab data. You also get a certain amount going in as ammonia or hydrogen cyanide into your dry gas streams, and that will increase your conversion.

Wojtowicz: I do not have much to add. The nitrogen mainly concentrates in the coke and the cycle oils, especially the heavy portion. We did some reaction mix sampling on a unit, and it did not seem to change much with operating conditions. Changing your cut points, such as increasing LCO ASTM 90%-point, will have a much larger impact on the nitrogen in the various products.

De Graaf: I fully agree with the comments by Matt and Tyner. Nitrogen from the feed ends up with about 10% as cyanide ammonia in the gases, about 2% in the gasoline, about 50% in the LCO, and about one third in the slurry. The remainder, the basic nitrogen, ends up with coke. If you increase conversion, you will have a higher amount of nitrogen going to coke and a higher amount of nitrogen going to cyanide. Basic nitrogen is the main source for nitrogen coke. It is obvious because you have a lot of acid sites on your catalyst that will be titrated with basic nitrogen. So if you have high-nitrogen feeds, you will suffer in conversion. About 100 ppm of nitrogen build gives you about a 1 vol% conversion loss. So there is a lot of nitrogen just clinging to the acid sites. Basic nitrogen is also typically an aromatic type of nitrogen, so it has a higher tendency to form a sort of "chicken-wire"-type coke on the catalyst.

Letzsch: I also want to point out that nitrogen burns slowly off of the catalyst. It is left on the catalyst if you do not do a complete combustion. So some of it, then, is carried over. For instance, if you have a 10% coke, that material carries over to the reactor. Some of that nitrogen can come off in the reactor, say cyanides or ammonia; so some of it is reduced. You may actually see that. Nitrogen also really slows down the coke burning process, as far as the regenerator is concerned.

Profitability

What are the most profitable dispositions for slurry oil and what issues do you consider for each option?

Santos: The most profitable applications for FCC slurry oil are, in decreasing order:

• Carbon fiber feedstock.
• Carbon black feedstock.
• Delayed coker feedstock.
• Fuel oil diluent.
• FCC recycle.

Petrobras has three refineries that have dedicated pipelines from them to customers' carbon black feedstock storage tanks. What is most important to us is the quality of slurry. For those three refineries, you have a specific crude that gives us a correct specification of Bureau of Mines Correlation Index.

For ash content in those refineries, we have used antifouling additives to control the ash. We make a lot of money in this refinery, and we are trying to do the same for the other two refineries. On the other hand, we are trying to find a solution, such as a filter or other technology, to minimize ash content.

Savage: There are several possible uses for slurry oil. It can be recycled to the FCCU feed for destruction. It is, however, quite resistant to cracking and does not give good gasoline yields.

Today, some major uses for FCC slurry oil are:

• Blending stock for heavy fuel oil.
• Carbon black oil (CBO).
• Feedstock for production of needle coke.
• Feedstock for the delayed coker to reduce furnace fouling.
• Feedstock for fixed-bed hydrocracker or ebullating-bed (H-oil) conversion processes.

Heavy fuel oil is normally the lowest value disposition for slurry oil. Depending on the ash content and the viscosity, significant quantities of low-viscosity blending stock must often be used to meet fuel oil specifications.

If the quantity of catalyst fines in the slurry oil can be reduced to 0.05 wt% and other specifications met, the slurry oil can be upgraded to CBO. FCC slurry oil is an important source for CBO. The upgrade value of slurry oil from heavy fuel oil to CBO can be significant.

For the production of needle-grade coke, slurry oils can be used to increase the aromatics content of the feed. Again, however, the ash must be reduced to avoid adversely affecting the coke quality. Decant oil can reduce fouling at the coker furnace; however, it can reduce overall coker liquid yield.

When used as hydrocracker feedstock, the ash contained in the slurry oil often causes plugging of the catalyst pores in fixed-bed processes and reduces the degree of conversion achieved. In ebullating-bed processes, the ash contained in the slurry can cause erosion problems in the ebullating pumps used in the process.

Devlin: Slurry oil is typically sold into the fuel oil and carbon black feedstock markets. Both fuel oil and carbon black purchasers will impose a high percentage-ash maximum. Exceeding the maximum will reduce the value of the slurry and, in some contracts, will trigger a price penalty. Reducing ash is accomplished by allowing the ash to settle in tankage before sale. If the process is too time-consuming or does not achieve the desired ash reduction, an ash-settling chemical may be recommended.

The increase in light tight oil as a percentage of the North American crude slate has resulted in lower FCC feed production and, consequently, a reduced FCC feed rate at several refineries. To address these issues, what strategies have you implemented operationally and catalytically? Are you considering sending new streams to the FCC or increasing the proportion of existing streams, such as resid? If so, what steps do you take to evaluate these potential new feedstocks, and what steps do you take to minimize uncertainty and reduce risk?

Foskett: According to several reports, tight oil will displace basically all of the light crude imports; some say already, and others think within the next 12 months. We expect light and medium crude supply to increase by 3-4.5 million b/d by 2020. The blends of light and heavy crude are possibly limited by dumbbell crude and asphaltene precipitation issues. Given the current restrictions on crude exports, medium crudes will be increasingly displaced as we move forward. So the amount of spare FCC capacity is likely to continue to increase. This situation may be at least partly offset by the installation of new topping units and subsequent export of the products.

We do see a lot of refiners looking to increase the amount of resid in the feed, and it is fortunate that tight-oil residue feed quality is quite good. One of the problems with processing resid from tight oil is that in many refinery configurations, it is difficult to segregate the good quality resid from a heavy resid that would be less suitable to put into the FCC. Another issue is that the availability of imported FCC feeds is limited.

Many refiners are using spare capacity in their cat units to increase the reactor severity and run higher catalyst activity in the unit to maximize conversion. Also, a lot of refineries have access to low-cost isobutane coming from the recent surge in natural gas liquids production. Many of these are using Zeolite Socony Mobil (ZSM-5) to maximize butylenes to fill out the alkylation unit as another profitable way to use the spare capacity that is coming as a result of tight oil. There is currently no trend of refiners sending new or different streams to the FCC unit, although this is something that may change in the future as the production of light tight oils continues to grow.

Devine: I do not have a lot of data. I will say that there is a Principles & Practices (P&P) topic tomorrow that will touch on this a little. So I just have a couple of comments regarding the second part of the question about the evaluation. You start evaluating feeds that are way outside of your normal window. You may need to do some recalibration of your statistical or kinetic models to get some different results, but typically that would be your first pass.

Well, let me back up. I guess you should first validate and define your base case, verify with your linear programs to understand what makes sense, define what you are going to benchmark everything against, and then start doing your evaluations. If you get to the point of doing a test run, follow your change management processes and keep good communication with operations and your exploration and production group for the post-evaluation to make sure that the change matched what you expected out of your model when you are hitting the same constraints. Catalyst vendors are usually up to speed on all this, so keep them in the loop, too, on all phases of the evaluation.

Diddams: I will just reiterate what Tyner said. There is a P&P session in the morning. I am going to be giving a brief presentation which covers some of the aspects that we touched on here. I do not really want to go into all of the answers right now; but if anyone wants to follow up, tomorrow morning we will be going through some of these questions in a bit more detail.

Bryden: The increase in the quantity of tight oil as a percentage of the North American crude slate has resulted in numerous changes at refineries. Tight oils, like other light sweet crudes, have a much higher ratio of 650° F.:650° F+ material when compared with conventional crudes. Bakken tight oil has a nearly 2:1 ratio while typical crudes, such as Arabian Light, have ratios near 1:1. A refinery running high percentages of tight oil could become overloaded with light cuts, including reformer feed and isomerization feed while being, at the same time, short on feed for the FCC unit and the coker. Some refiners have balanced the use of larger amounts of tight oil with increased use of heavier crudes such as Canadian Syncrude. Others have charged a portion of whole tight oil to the FCC to keep the FCC full.

As new feedstocks are considered, testing is a valuable tool to reduce risk. Testing provides an understanding of feed properties and potential yield changes.

Testing of feed metals levels is especially important since tight oil-derived feeds often contain varying levels of conventional contaminants such as sodium, nickel, and vanadium, and unconventional contaminants such as iron and calcium. Understanding the expected metals levels of a new feed allows refiners to work with their catalyst vendor to choose catalyst options that mitigate the challenges of these metals. Grace's newest catalyst family, Achieve catalyst, is designed to address the unique challenges associated with tight oils. Achieve catalyst formulations are flexible, enabling Grace to design a custom solution for refiners proactively increasing the amount of tight oil in their crude diet.

Feed properties such as API, Concarbon wt%, and hydrocarbon types can provide insight into the expected crackability of a feed but may not tell the whole story. A fuller understanding of how a feed will crack in a unit can be obtained through testing. Either bench-scale testing [advanced cracking evaluation (ACE) or microactivity test (MAT)] or pilot-scale testing, such as Grace's Davison circulating riser (DCR) pilot plant, can be done. MAT and ACE testing have the advantages that they are easy to set up and require small amounts of material. These units, however, cannot provide the detailed product analysis or feedback on extended operation that pilot-scale units can. Larger-scale test equipment, such as a pilot unit, can deliver sufficient liquid product for distillation and detailed analysis (i.e., API gravity and aniline point of LCO produced, viscosity of bottoms, octane engine testing of gasoline, etc.) and can give information on continuous operation. Compared to bench-scale units, the DCR pilot plant also can mimic all of the processes present in commercial operation, and it can operate at the same hydrocarbon partial pressure as a full-scale commercial unit.

Grace's technical service and research-development teams help refiners assess potential challenges from feedstock shifts before they occur via feed characterization, feed- component modeling, and pilot-plant studies. Understanding feed impacts earlier provides an opportunity to optimize the operating parameters and catalyst management strategies, enabling a more stable and profitable operation.

The panelists

Tyner Devine, process engineering lead, Flint Hills Resources LLC (FHR)
Dan Elling, technology service manager, Marathon Petroleum Corp.
Stuart Foskett, global technical services manager, BASF Corp. (Refining Catalysts)
Geraldo Santos, technical manager for FCC, alkylation, and isomerization, Petroleo Brasileiro SA (Petrobras)
Matt Wojtowicz, FCC technical service specialist, UOP LLC

The respondents

Brad Mason, NALCO Energy Services LP
Whitney Nikitczuk, Phillips 66
Paul Diddams, Johnson Matthey Process Technologies Inc.
Michael Wardinsky, Phillips 66
Robert Ludolph, Shell Global Solutions (US) Inc.
Dennis Haynes, NALCO Energy Services LP
Greg Savage, NALCO Energy Services LP
Elbert "Bart" De Graaf, Johnson Matthey Process Technologies Inc.
Warren Letzsch, Technip Stone & Webster Process
Technology Inc.
Brian Devlin, NALCO Energy Services LP
Kenneth Bryden, W.R. Grace & Co. (Catalyst Technologies)