Discussion expands to crude, vacuum distillation, and coking

Houston Refining LP, a wholly owned subsidiary of LyondellBasell Industries NV, completed planned maintenance earlier in the year at the crude and coking units of its 268,000-b/d full-conversion Houston refinery, which is configured to process heavy, sour crudes into low-sulfur gasoline, diesel, and jet fuel (OGJ Online, Apr. 4, 2016).

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

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 can provide sound guidelines for what would work to address specific issues.

Crude distillation

Please describe your experience with the occurrence of phosphorus and barium fouling in the distillate section of the crude tower. What steps have you taken to identify and mitigate the problem?

Watts I know this topic generated a lot of discussion in our hydrotreating session yesterday. I am mainly going to focus on the crude side of it. Basically, it starts with analyzing processed crudes for impurities to manage and minimize fouling coming from salts, asphaltene precipitation, and other impurities.

Regarding corrosion, we obviously look at total acid number (TAN) of the crude, organic chlorides, amines, and ammonia. We do not always do a full analysis on the crudes we are processing day in and day out, however, since we have a long history on those crudes. When we are looking at new crudes, we do a more detailed analysis.

We also look for catalyst poisons, mainly to make sure we do not have high levels of silica, phosphorus, arsenic, and other materials that would cause accelerated loss of catalyst life. We then look at downstream impacts on those units. Overall, our goal is to manage equipment reliability and catalysts between turnaround cycles.

Last year, we took a short outage to replace a piece of equipment on one of our crude units. We were a little over 5½ years into the run since the last turnaround cycle. When we started up after the outage, we saw that the top section of the crude tower differential pressure had increased to 5 psi. Before the outage, it was 2 psi, and that was for the top 20 trays of the crude tower.

We had experience with trays fouling prior to this outage. When we would shut down the unit for turnarounds, we would see that the top six to eight trays had some fouling from salts that deposited on the trays.

Next we did a tower scan and a more detailed pressure survey which showed that liquid had started to back up in the tower just below the kerosine pumparound section around Tray 20. As we were scanning the tower, we basically adjusted the liquid loading in the tower. We dropped the pumparound and the top reflux. Once we did that, we saw that the pressure drop returned to a normal range. The tower was no longer flooding.

Before the outage, we actually had noticed an increase in fouling in our kerosine pumparound exchangers. We talked to some of the operators after we saw the pressure drop increase in the tower top. We also had some issues with valves closing when they were isolating the kerosine pumparound exchangers. We did an analysis of that stream, and basically, we saw a combination of corrosion products and hydrocarbons. We also discovered trace levels of phosphorus as high as 800 ppm. As I said, this occurred April 2014, but we have been able to continue running the crude tower.

To manage this increased pressure drop and fouling, we have adjusted our kerosine pumparound. Before April 2014, we typically ran that pumparound at a rate of 1,300 bbl/hr or higher. Currently, we are running the pumparound at the minimum (750 bbl/hr), so obviously our kerosine production has dropped off. We have also lowered our top reflux in the tower and adjusted the heater-outlet temperature and stripping steam to the tower to reduce the vapor load for certain crudes.

We have a planned outage in first-quarter 2016, as it will have been a little over 7 years since the previous turnaround. We have worked with a company to redesign the trays, which we plan to replace. We will also be able to verify the fouling during the scheduled maintenance outage.

Braden We have conducted deposit analysis on samples in the jet kerosine trays and found the phosphorus component, along with iron and sulfur. The phosphorus usually comes from an upstream additive that is used in the fracing aspect of the water-sensitive clay formations. They use a mono- and diphosphate ester to help with the fracing process that can be complexed with an inorganic material and removed from the crude. During the manufacturing of the mono- and diphosphate esters, a triphosphate ester is also formed. This triphosphate ester is oil-soluble and cannot be complexed.

The triphosphate ester is the material coming in with the crude once it passes through the desalter, but once it gets into the heat exchangers and into the towers, it starts decomposing. Specifically, it starts hydrolyzing to phosphoric acid, so essentially, phosphoric acid distills up-tower, deposits onto the tray, and then precipitates at the jet kerosine trays. When the phosphorus-containing deposits increase in size, they begin closing the holes in the trays.

Some refiners will replace the trays with trays containing larger holes, therefore allowing you to get more flow in the jet because the jet fuel has a phosphorus spec. We try to mitigate the phosphorus from distilling up the tower by injecting a chemical additive that complexes the phosphorus to keep it in the resid fraction. If you want to know a little more about that approach, contact your chemical vendor.

Weber We have had some practical experience in the kerosine section. The first time we experienced it in 2011, we went into a tower for turnaround and found fouling in the kerosine section, with no other fouling occurring above or below. It didn't cause an operational problem, and we actually didn't analyze foulant that was on the tray at that time.

At another MPC refinery, the crude unit ran at reduced rates for economic reasons for about a month. When we attempted to return the unit to full rates, flooding was observed in the kerosine section. We were referred to the Canadian Crude Quality Technical Association (CCQTA), which immediately said, "Oh yes, you have phosphorus issues." After further discussion, it was discovered that CCQTA has extensive experience with phosphorus contamination. Its website indicates that, since 1995, they have had a project on phosphorus fouling and have worked with some Canadian producers to help minimize it.

A similar incident occurred at a third refinery. One of the remedies we have implemented is installation of fouling-resistant, fixed-valve trays in the kerosine section. In refineries that still have floating-valve trays, we try to keep vapor traffic up in the kerosine section. We have found occasions where the valves were found stuck open but, fortunately, had not caused an operational issue. It's when we ran at lower rates and velocities that the valves became stuck closed and caused flooding problems. We have also had experience with barium fouling in the wash section of the crude tower.

Appalla Is there an industry-accepted limit regarding monitoring the phosphorus content of crude? If so, does anyone distinguish the differences between volatile and nonvolatile phosphorus?

Watts I want to echo some of what John said. I did quite a bit of research on this and know that it is not a new issue. Information dating back to the mid-1990s indicates specification levels of 1-1.5 ppm. Based on work with our lab, we have not seen a detectable level of phosphorus on the crudes we have analyzed. We have not traced it back to a crude source.

Braden In our experience, the phosphorus compounds enter the refinery via the crude and normally come in slugs because of the fracing issue, so phosphorous content can be up to 50 ppm when you get a slug of highly phosphorous material. But our measurements normally show a phosphorous content of 1-3 ppm. We have run that by inductively coupled plasma (ICP) analysis. So phosphorous presence in crude is normally very, very low. Sometimes, however, you do get slugs. So when do you catch a slug? When do you see that? You just see the results of the slug coming through. It is hard to pick up.

Appalla What is the industry's experience on the cause of this fouling when using high-temperature corrosion inhibitors in the atmospheric tower?

Lordo The inhibitors that are used for high-temperature treatment are not really part of that particular description. As Mike indicated, those are triester compounds which come in with some of the fracing gels that are used in Western Canada or in clay formation-type crudes where they are water sensitive. In the US Lower 48, though, there is no fracing with phosphorus-based gels. Now we are targeting to look at phosphate esters being used for scale control and corrosion inhibitors.

I looked at crude-tracking data from one refinery that analyzed the phosphorous content of each crude batch that came in over the course of an entire year. Most of it was in the 1-3 ppm range, except in November, when it shot up to about 10-15 ppm for a 1-month period, after which it came right back down. During that timeframe, that refinery's crude tower had some issues which have yet to be resolved, so the tower will have to be taken offline for cleaning. But as far as the inhibitors go, most of them do not cause a problem. If you overtreat, then yes, you can have some issues.

Price Just a comment to Mr. Appalla from Reliance: If you have not visited the CCQTA website, I recommend that you do because it contains is a lot of information about lab methods that might be helpful.

Appalla Yes, but the CCQTA only talks about the phosphorus coming from the Canadian crudes. Suppose you are not processing Canadian crudes. What are the other sources of it?

Eggert It is an odd question when you say barium and phosphorus in the same sentence. Why would you pick those two particular elements to examine? Like the panel has mentioned, not all phosphorus is going to end up fouling your trays. It is always coming in at a low level. Some of it is benign phosphorus, and some of it will contribute to tower fouling. What is currently being used-and the reason the barium came up-is that the phosphorus causing some of these problems is injected to prevent barium-sulfate scale upstream. It is a scale inhibitor. If you see barium and phosphorus, chances are it is from the production chemicals. The barium used to be called normally occurring radioactive material, or NORM. That is why we are looking for phosphorus and barium. It is the combination of the two that sends up a red flag.

Zurlo Just to clarify, you are right. The CCQTA talks about phosphate esters from Canadian crudes. One of the test methods it has on the website specifically identifies volatile phosphates. The CCQTA test separates the volatile phosphorus components from the general phosphate components by performing a distillation on the whole crude and analyzing the middle distillate cuts. Although the CCQTA work is done on Canadian crudes, the crude source really does not matter. What is important is if phosphate is in the crude; if it is volatile; and, if it can hydrolyze, complex, and form this fouling material. So it is not necessarily the source but more the effect of what type of phosphate is in the system.

Cates If I understood the question, what Mr. Appalla is really asking is, other than light, tight oils, where are they finding crudes containing phosphorus?

Watts I mentioned earlier that we have not traced it back to its source, at least in our refinery.

Braden The phosphate esters for the fracing are mainly coming from the Canadian side, the north coast of North America, and the south coast of Canada.

Naquin I have heard rumblings that the fouling would have less affinity for metallurgies. For example, if you have carbon steel or different variations of stainless trays or parts in the tower, sometimes the phosphorus fouling will not have the same affinity; so then, you will not have the same accumulation of the fouling in the tower. Does anyone have experience with those observations and any best practices?

Watts I do not have experience on the metallurgy. I would say that what we are doing is similar to what John said in that we redesign the trays to go to a fixed valve. So we avoid the potential - if you turn down, increase, or shut down a unit-of your valve getting locked in place; that is, if you have a floating valve. Obviously, that changes your flexibility on how you can run the crude unit when you reduce rates.

Price Once again, I am not an expert on this subject. Based on the experience of our clients, I think the jury is still out on the subject of phosphate fouling. People are finding, in some cases, that when they solve the fouling problem in their crude tower, the fouling problem migrates downstream to the exchanger. While exchanger fouling is not a good problem, it is solvable. If you can install a bypass and have the ability to do online cleaning of that exchanger, it will help to mitigate the impact on your unit throughput.

Vacuum distillation, coking

What type of facilities have you used to cool hot vacuum residue going to storage to avoid plugging problems and facilitate reprocessing?

Watts I am going to focus on a system we have and also some of the issues we have experienced. The majority of our resid produced off of the crude unit vacuum towers is sent through the hot resid system straight to the cokers. We have two crude unit trains. We process 120,000-140,000 b/d of crude on each unit. We have two cokers. Each coker has four drums.

One coker can process up to 40,000 b/d of resid on a four-drum operation. The other can process 60,000 b/d of resid on a four-drum operation.

Our cold resid system is where we send excess resid, but we also maintain that system during normal operation. So what we do is take uncut resid at about 400° F.+ and add cutter. A base cutter for us is heavy cycle oil from the FCC. That is typically the only place we send it. Then we make up, as needed, with distillate-range material. The majority of the time it is light cycle oil from the FCC. Basically, we target a maximum tank temperature. We have a temperature limit of 210° F., so we try to keep it around 200° F. When we are stacking a lot of resid, we can hit a viscosity spec. We have found that we need to add about 30-40% cutter to hit that spec.

We have three main modes of operation. The first is what I call resid-system balanced, or the mode in which we typically operate, where the coker is pacing the cold resid. Normally, we are sending 6,000-10,000 b/d to tank and pulling that equivalent amount back to the cokers.

Next, we have what I call resid-stacking mode. With the current system, we can stack up to 30,000 b/d of resid with a minimum of 30% cutter. That leaves about 21,000 b/d of resid you can stack.

The last mode, which we do not do very frequently, I call resid pull. This typically occurs after a short outage on the cokers. After the outage, we will pull back resid that was stacked.

Occasionally, the economics support going out and purchasing resid, but the way our refinery operates is that we are typically close to limits on the cokers when both crude units are at full rates because the resid yield is typically above 30%.

The last piece I want to talk about is how we minimize the potential for resid line plugging; basically, it is managing the temperature. As I said, we have limits on what we can send to the tank, but we operate in a relatively tight window. We try to keep the temperature above 190° F. Basically, we have added in orders for console operators to heat up the cold systems one time per shift, so they do that about twice per day. Resid is hard to meter. What we have done, based on operating experience and historical data, is set minimum-valve output limits where we are all alarmed if the resid rate gets too low. So for us, the biggest challenge for managing the hot and cold resid system is during major upsets, when we lose production of our main cutter source and significant coking capacity. This is typically where one or more of the cokers go to two-drum operation and where we are most likely to have issues with plugging in the sections of the hot or cold systems.

In fact, back in 2010, we had a refinery-wide emergency shutdown. We actually plugged up the hot system between one of the crude units and the cokers. It took a lot of money and time to unplug that system. In 2014, we also plugged up a small part of the cold system. The highest risk for us is when we have a major refinery outage because multiple asset or operating teams have to communicate with each other to make sure we get cutter in the lines.

Price I want to second what Ed said: The challenge of storing hot resid is very difficult, although folks who make asphalt and store it have more chances to do this than others.

To overcome the difficulty in storing hot resid, many refiners in southern California will use a box cooler, diversion air coolers, or a tempered water bath. Normal cooling water cannot be used because it will cause extreme fouling on the water side, as well as plugging on the hydrocarbon side. The diversion air coolers incorporate special design features to ensure that the approach to pour point is adequate to prevent plugging on the hydrocarbon side. On an emergency basis, quenching with cold gas oil product is about the best option you have.

Lucke When you process cold vacuum residue in the coker unit, because of heat integration and the amount of cutter stock you need, do you have a limit, such as a percentage of total fresh feed, up to which you can process?

Watts A typical operation is where, I would say, you'd pull less than 10,000 b/d out of 90,000 bbl. I know that we pulled quite a bit more than that in the past-up to 20,000 bbl, or somewhere in that range-out of 90,000 bbl; so, a little over 20%.

Doherty To add to the point and answer the question, I work on the cokers with Ed. We would hit that limit. Sometimes, right before we got to that limit, we would see increased foaming on the cokers. That would limit our pullback instead of a heat limit.

Watts I want to comment on what Maureen mentioned. I did not really add detail to how we cool off our resid. We add cutter. We have cooling water exchanges on the crude units. But a couple of years ago, we installed a temporary cooling system just before we sent the resid to the tank. We actually cool off the resid with glycol. That is a much better system in terms of reducing exchanger fouling. It allowed us to reduce our cutter significantly and also to stack more resid. With that system, we were able to stack as much as 40,000 b/d of resid. Without that system in place, we were limited to somewhere around 20,000 b/d.

Desalting, wastewater treatment

What strategies do you employ to purge solids from recovered oil at the wastewater treatment plant (WWTP) to avoid recycling solids back to the crude unit?

Braden Oh, yes, solids: the bane of the refinery. The crude unit wants them out of the crude oil, and the WWTP would prefer not to have them. Essentially, we are talking about wastewater treatment solids. Some people call it recovered oil; some people call it slop oil; and, some people call it skimmed oil. So if I interchange these definitions, what I mean is recovered oil from the WWTP.

A little background here: this oil will be chemically stabilized due to the chemical treatment of the primary wastewater treatment program. It is a water-in-oil emulsion. The solids and the water itself are chemically emulsified in the oil. If you do nothing and send it back to the crude unit, you will get massive upset. One barrel of untreated slop oil or recovered oil will give you back two barrels. So essentially, the refiners have to ask the question, "What are you going to do with the skimmed oil?"

First, if the refinery has a coker unit and the coker is not making anode-grade coke, then you could send that material as a coolant to the coker. Again, everyone has to be on board with that. The second option is to sell the skimmed oil to a remediator or have them buy it from you.

Typically, this recovered oil has a good dollar value for the refinery, so the refinery will want to recover the oil by removing the water and the solids.

There are two methods for recovering the oil that are viable for the refinery. One is tank treating. For tank treatment, the tank should be heated and have agitation. You will need a chemical additive to resolve the chemically stabilized emulsion. Do not be shocked if the chemicals needed to resolve the emulsion may be up to 3,000-5,000 ppm. You will need to add the chemical, thoroughly mix the chemical with the oily emulsion, and then let it stand at a minimum of 140° F. (but no greater than 180° F.) for about 72 hr. As a result, some solids will fall out. You will have a water layer and an oil layer. In between the oil and water layers, you will always have a rag layer. You will have to make a decision about what to do with that rag layer. It will contain a lot of emulsion and solids. Some refineries will put the rag layer into a different tank and wait another day, even though the oil itself will have less than 2% basic sediment and water (BS&W), just delaying the decision about what to do with the solids.

The other way is to remove the solids from the system entirely, which is your goal, by using a centrifuge method. This method is very similar to a tank method, although you do not wait for the separation. You use heat, chemical, and agitation. You need to let the treated slop oil stand for 1-3 hr, and then send it to a two-phase or a three-phase horizontal bowl centrifuge. A three-phase centrifuge is better.

In the waste-oil stream going to the centrifuge, you will have to add a high molecular-weight cationic polymer, usually emulsion polymer made down to a 0.5-1.0% solution. This chemical will take out solids in the rag layer as well as solids that fall out in the centrifuge. Because the centrifuge separates by specific gravity, that rag layer has its own specific gravity vs. the other three. You will have solids, water, rag, and oil. The emulsion polymer will mix the solids and the rag layer together, resulting in a specific gravity which is higher than water. You will then have solids plus the rag layer, water, and oil. The solids go out one end, and the liquid goes out the other.

In a three-phase centrifuge, the operator controls the flow rate of the waste oil stream and the chemical polymer injection. The BS&W on the recovered oil is typically less than 1%, and you can send the solids to a nonhazardous landfill if the oil content meets the required specifications. If it has too much oil on the recovered solids, the refinery alternatively could send it to a hazardous landfill.

Allred We have a full-time, three-phase separator on site in the refinery. It is owned and operated by a third party and it is in operation Monday-Friday during daytime hours. It is used primarily for our wastewater treatment system, collection of solids, and recovery of the oil in the water. We have found that just keeping the solids out of our system is well worth that expense. We have polymer and heat that helps remove those solids. There are times when we do a tank cleaning and keep this unit in operation around the clock to handle the solids coming out of the tank. By working through the weekends and extra hours, we are able to remove the solids and keep up with the daily needs of the wastewater treatment system. We have found that having that three-phase separator on site all the time has been well worth it for us.

Weber We also have an outside third party who manages the solids. I do not really have anything else to add.

Appalla I have two questions. The first is: Has any refinery deployed the system of skimming the rag layer online, and if so, what is the destination for the skimmed rag layer? Secondly, you have talked about deploying centrifuge separation, but a centrifuge works on the principle of the particle size. If the particles are very small, say less than 10 μm, does chemistry really work for coagulating and dropping the particles in a cyclone?

Braden To clarify your second question first, you are asking about removing the 10 μm-sized particles? Typically, wastewater treatment can remove down to 4 μm. Filters can get below that if you have the right filter system. But as a centrifuge, it really depends on the polymer you select and the agitation. Because you feed the emulsion polymer based on the flow, it is tough to say the degree of particle size that is removed. We have not really measured the particle size that is recovered in the crude. We do know that the filterable solids are very low using a 0.45 μwm filter, but we have not looked at particle size. The primary goal for a refinery is to look at their specification of less than 2% BS&W. That is a normal specification for consideration as recovered. For some refineries, the minimum BS&W could be a little higher or a little lower. Could you repeat the first question please?

Appalla If the solids are bound to create the problem at the interface by forming a rag layer, it does not allow the emulsion to separate. So what is the best way is to skim out the rag layer online? Has anyone deployed continuous skimming of the rag layer online and routing it to some destination? Where is the best destination to route the rag layer?

Braden Some refiners will have a rag-layer draw on their desalters to remove an increasing rag-layer volume in the desalter. At Nalco Champion, we recommend using the correct desalter emulsion breaker that will allow the rag-layer volume to remain at steady state. For Nalco Champion, an increasing rag-layer volume is a sign of a poor emulsion-breaker selection. Our goal is to have no oil undercarry and a steady-state rag layer.

Watts For our refinery, we manage the rag layer by adjusting the mix valves, washwater, and desalter chemistry to minimize the potential for oil undercarry. We don't skim the interface to remove the rag layer. The question he is asking is: If you have a rag layer at the desalter, rather than trying to bleed it all out the bottom, do you have a skimmer at the interface to continue to remove the rag layer? And if you do that, what is the destination for that material?

Braden Good question. Usually the rag layer is put into a different tank, and we wait for it another day. This rag layer is normally not chemically stabilized and can be fairly easy to separate; however, it may take a secondary emulsion breaker to drop the solids and release the oil.

Lordo We are referring to what is commonly called cuff draw, which is a pipe located at the interface. You pull off that pipe and actually send it to a different tank, as Mike indicated. Typically, you can treat it as it goes to the tank, which is preferable to building inventory in a slop tank, which would make it more of a challenge to handle. You would pull the rag out of the desalter and send it to a separate tank. Sometimes it is a brine tank, so you can mix the emulsion with water and flip the emulsion a little. So there are a couple of strategies you can utilize with the cuff draw, but they are gaining popularity with the high-solid crudes.

Braden Usually that is not chemically stabilized, so it is easily separated.

Allred I just have a quick comment about the second question. Our trigger point is 1% BS&W. So when we run it through the centrifuge to recover the oil, we test that oil. If it is greater than 1%, we will send it back through the centrifuge. If it is less than 1% BS&W, we will send it on to a crude tank.

Prorok Nalco treats our desalters, and we do a draw of the cuff layer to remove iron. We still make petroleum-grade coke for anodes. We take the cuff, which is about 10% oil and 90% brine, and send it to the free-water separator, which breaks the oil and water mixture, leaving us with a high-solids oil. The oil goes to a storage tank, which feeds the centrifuge. The oil from the centrifuge still has a significant number of fines. We have pilot-tested a ceramic membrane for a final polishing step to get the fine solids out of the oil.

The panel

Bruce Allred, production engineering manager, Suncor Energy Inc.

Michael Braden, desalting expert, Nalco Champion

Maureen Price, director of downstream process engineering, Fluor Corp.

Ed Watts, refinery process engineering technical team leader, LyondellBasell Industries NV

John Weber, refinery planning manager, Marathon

Petroleum Corp. (MPC)

The respondents

Nagashyam Appalla, Reliance Industries Ltd.

Bill Cates, Hunt Refining Co.

James Doherty, LyondellBasell Industries NV

Harold Eggert, Athlon Solutions LLC

Samuel Lordo, Nalco Champion

Eberhard Lucke, CH2M Hill Inc.

Jessica Naquin, Valero Energy Corp.

James Prorok, Husky Energy Inc.