AFPM Q&A-1 Refiners address gasoline processing issues

Aug. 7, 2017
During the 2016 American Fuel and Petrochemical Manufacturers Q&A and Technology Forum (Sept. 25-28, Baltimore), US domestic and international refiners discussed trends in gasoline processing operations, paying special attention to topics of safety, isomerization, reforming, and meeting the US Environmental Protection Agency's more stringent Tier 3 gasoline standards that took effect Jan. 1, 2017.

During the 2016 American Fuel and Petrochemical Manufacturers Q&A and Technology Forum (Sept. 25-28, Baltimore), US domestic and international refiners discussed trends in gasoline processing operations, paying special attention to topics of safety, isomerization, reforming, and meeting the US Environmental Protection Agency's more stringent Tier 3 gasoline standards that took effect Jan. 1, 2017.

This annual meeting addresses real problems and issues refiners face in their plants and provides an opportunity for members to sort through potential solutions in a discussion with panelists and other attendees.

This is the first of three installments based on edited transcripts from the 2016 event. Part 2 in the series (OGJ, Sept. 4, 2017) will highlight discussion surrounding processes associated with hydrotreating, while the final installment (OGJ, Oct. 2, 2017) will focus on fluid catalytic cracking (FCC).

MPC is executing a project to unify its 459,000-b/d Galveston Bay refinery with its nearby 86,000-b/d refinery at Texas City, Tex. (shown), to form the fully integrated 585,000 b/d Galveston Bay-Texas City refining complex. First announced in fourth-quarter 2015, the South Texas Asset Repositioning (STAR) program is scheduled to be completed in 2021 at a cost of $1.5-2 billion. Designed to improve operational efficiency, the STAR project will enable the two refineries to achieve the US Environmental Protection Agency's updated Tier 3 gasoline sulfur standards. Photo from MPC.

The session included a four-person panel of industry experts from both refining companies and other technology providers 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 by which to address specific issues.


What process safety management (PSM) factors do you consider when contemplating a reformer unit rate increase?

Hutchinson Any change should be handled through the management of change (MOC) process, according to OSHA 1910.119 (PSM of Highly Hazardous Chemicals). This standard includes requirements for preventing or minimizing the consequences of catastrophic releases that may result in toxin, fire, or explosion hazards. For refiners, the MOC process should include considerations for a number of factors mentioned in this standard, including a rate change. The relief system design and basis must be updated to consider the new unit rate, along with the impact on the flare system. The relief valves will need to be rerated or replaced based on the appropriate cases, including settling-out pressure, blocked outlet, and loss of reflux. Changes in reboiler duties will impact the relieving cases as well.

A hazard and operability (HAZOP) study should be performed and unit PSM information updated, including material and energy balances. The impact on the unit hydraulics also should be evaluated to determine if vibration issues might occur in equipment such as piping, heat exchangers, and heater tubes, especially in cases with two-phase flow.

The impact on corrosion rates must be reviewed, and the inspection schedule should be updated. Safe upper and lower-operating limits for temperature, pressure, flow, or composition should be updated, along with safety systems such as interlocks and alarms. Additionally, operating procedures must be updated to reflect the new operating limits and any required changes resulting from the increased rate.

Rhodes For reformers, the reactor heater is typically one of the limits. As the unit increases rates, heater limits such as duty, tube-wall temperature, or fuel-gas pressure will potentially force the unit to operate at lower reactor temperatures, resulting in lower octane. The limits for the heater should be identified and monitored. The debutanizer reboiler will also need to have heater-duty, firebox, and fuel-gas pressure limits established before any rate increase.

An increased feed rate will also impact coke formation on catalyst. The regenerator coke-burn capacity should be reviewed to ensure carbon can be maintained at a reasonable level. Ability to recycle hydrogen may become limited, which will reduce the hydrogen-to-hydrocarbon ratio. This hydrogen-to-hydrocarbon ratio will impact the amount of coke made but will also push the unit toward heater tube inner-diameter (ID) carburization, and ultimately, to metal dusting. Sulfur injection can be used to mitigate the situation, but heater tubes should be checked for carburization during turnarounds to prevent it.

While not a PSM concern, catalyst pinning in the lead reactor can also be an issue as rate is increased in stacked-reactor designs. The licensor pinning curves should be reviewed as rates are increased. Pinning can lead to shorter run lengths and require unit outages to remove the plugging from the reactor's center screen. Pinning occurs when the gas flow rate across a radial bed stops catalyst from flowing down the reactor by gravity and presses the catalyst against the center screen, causing the catalyst to get stuck. Pinning causes high-pressure drop, leading to restricted rates or reactor damage until a unit outage is scheduled to clean the screen.

Chloride guard bed performance will need to be monitored. At higher rates there will be more hydrogen flow carrying more pounds of chloride for the beds to absorb. For the net-gas chloride guard, the upstream separator's ability to prevent liquid carryover will also decrease. Any liquid carryover will impact chloride guard bed performance.

For continuous catalyst regeneration (CCR) units, increasing feed rate doesn't necessarily mean catalyst circulation will increase beyond the regenerator's original design. It will mean, however, that average catalyst circulation rates potentially could increase, and this increased catalyst flow could cause more erosion on lift lines. Proper lift-gas flow and frequency of inspection monitoring should be reviewed at the time of rate increase.

Establishing PSM rate limits for all process units and documenting procedures for increasing these limits are good practices. Before any rate increase, the unit should be reviewed to guarantee that all relief valves are adequately sized for the projected rate increase. It's also recommended that the environmental team review and approve the PSM rate increase to confirm that any emissions from the increase are properly permitted. All affected heater limits should be identified as well.

A test run can help ensure all operational or reliability issues are identified. A robust test-run plan assists in determining the new PSM rate and should include the following processes:

• Identify existing equipment limits that must be honored during the test.

• Review critical alarm settings and instrumentation to address any instruments that may need to be rearranged.

• Identify any control valve that may require bypassing, special sample collection, and unique equipment monitoring, as well as any product specifications that may need to be waived.

• Conduct an MOC review.

• Prior to the test run, gather a baseline, including a complete mass balance, survey of vibration data for rotating equipment, and review of spring-can positions throughout the unit.

• Schedule additional operations staff to assist during the actual test run.

• Slowly increase rate within the unit until reaching the equipment limit or obtaining the desired PSM rate limit.

• Allow the unit to come to steady-state and conduct a pressure-and-temperature survey.

• Obtain mass balance samples.

• Assign reliability department personnel to check the vibration levels on the rotating equipment and all spring-can positions in the unit to ensure that no piping circuits are placed in a strained condition.

• Capture or document all control valve outputs, along with a list of any control valves that were bypassed during the test run.

• Following the test run, conduct a comprehensive review of the unit to analyze areas that can lead to operating or reliability issues.

• Calculate line velocities on key piping circuits, and establish velocity guidelines for the various piping systems and flow regimes. Address any circuits with high velocity before increasing the PSM rate.

• Check rho velocity (ρV2) for all vessel nozzles. A limit of 10,000 lb/sq fps is an absolute limit while any nozzles with ρV2 above 4,000 lb/sq fps are selected for routine inspection.

• Check vessel capacity for residence time to ensure its adequacy in providing operator response time.

• Check all pumps to ensure that available net-positive suction head (NPSH) meets pump requirements at pump projected flow rates.

• Issue a report that includes all of the reviewed information as well as a list of recommendations to be completed before increasing a unit's PSM rate limit. As those recommendations are completed, the original MOC is completed, and the unit is allowed to operate at the new PSM rate limit.


Do you have experience starting up an alumina chloride catalyst-type isomerization unit without first acidizing the reactor loop? What was the impact on catalyst activity?

Philoon Honeywell UOP strongly recommends that acidizing of the reactor circuit be included in the commissioning or startup of grassroots and revamped UOP Penex or Butamer isomerization units, both of which use chloride aluminum catalyst. If not, the consequences could include major catalyst deactivation as well as possible corrosion in the unit's reactor section. A recent experience with a customer electing not to follow UOP's recommendation of performing a dryout and acidization of the new reactor section resulted in an estimated deactivation of between 60-85% of the new catalyst load.

For units undergoing a turnaround, the UOP guideline is that equipment in the reactor circuit should be acidized anytime it is opened and exposed to air. Determining if acidizing is required and the extent of the procedure, however, involves a few considerations such as the actual scope of work, how it will be executed, and how long it will take to complete. We are aware of units that have suffered noticeable catalyst deactivation after restarting from a turnaround where acidizing was not performed. The preliminary 2016 AFPM Q&A and Technology Answer Book contains a discussion of the mechanism for the generation of water during normal unit operation if acidizing has not been completed (see accompanying box).

Hutchinson Steve did a great job covering the impacts of acidizing. Axens's standards are similar. We require, or strongly recommend, that the dryout and acidizing steps are completed following maintenance work or at any time before unit startup. We do have experience when acidizing of the reactor circuit was not optimum, and we noticed deactivation of the isomerization function at the frontend of the unit once a chlorinating agent was injected. Deactivation was noted based on a decline in temperature differential (∆T) in the first reactor. After some time, ∆T in the first reactor stabilized at the anticipated value based on saturation of benzene reactions occurring where the platinum function is still active but the catalyst has lost its acid function.

Mueller I've been a part of one of these procedures that took a very long time because we didn't check the condition of the equipment we had installed. It involved some exchangers left out and exposed to various weather conditions for a while. My one comment is that failure to continuously monitor the condition of your equipment can prolong your startup. Make sure you are not putting in something that will require a ton of hydrochloric acid, or HCl, to acidize it.

Bullen Related to the last comment, if you have large amounts of rust, some refiners will do a wet acidization step before performing the dry acidization to remove excessive rust. Where practical, physically removing rust from exchangers-like when you pull them apart-is helpful. We've had experiences where refiners left cubic feet of rust in the channel heads and then tried to acidize. That process takes forever because you're impacting a large mass that doesn't have a lot of surface area. Eventually, these refiners pulled out the exchangers, cleaned them again, then successfully reacidized. So it's very important to keep track of those details.


What range of sulfur targets for hydrotreated FCC gasoline do you anticipate for Tier 3 operation?

Rhodes The FCC naphtha stream is typically the highest-sulfur gasoline component. Sulfur targets for FCC gasoline depend on your gasoline-pool sulfur balance and the volume of low-sulfur blending components available. Butane blending in the winter may impact FCC naphtha sulfur targets depending on the sulfur level in purchased butane. We anticipate targeting between 10-20 ppmw sulfur on our units with post-FCC naphtha hydrotreaters.

Kleiss In our new gasoline processing units, we're targeting 7-10 ppm sulfur on the finished product.

Hoekstra My question is: Have you done testing on your FCC gasoline desulfurizers to find out what severity is required to make the lower sulfur level and then how much octane is lost at the deeper desulfurization?

Kleiss I do not know the testing, per se. We do have some generalized curves. In the next presentation, we will be talking about Tier 3 processing in naphtha capability, and I have a curve that shows desulfurization vs. octane loss. You may be interested in seeing that curve. When we get to that presentation, we can discuss it. I think that is the question you are asking.

Tier 3 octane loss could be as much as twice the loss with Tier 2. US FCC charge is in the range of 4.9 million b/d. With a typical gasoline yield of 50%, the impact of Tier 3 on the gasoline pool is huge. The curve shows octane loss is in the 3.7-4.9 million octane-bbl range.

Moreland Where possible, we have done test runs on our existing gasoline desulfurization units to reach a sulfur content of 10 ppm and even 7 ppm or lower. We then monitor the octane loss across those test runs and with different targets for gasoline splitter operation to see where we find octane loss minimized. In some of our units, then, we're going to do a catalyst change only and be able to meet Tier 3; in other cases, however, we'll need to make a capital investment to meet the regulation. We use the test-run results as part of our planning.

Bray The sulfur target for hydrotreated FCC gasoline is very site-dependent. But where possible, it is desirable to hydrotreat all other gasoline streams fully so that the FCC naphtha can be treated as mildly as possible. Deeper desulfurization for FCC naphtha results in increased olefin saturation with the resultant octane loss. Since the other gasoline streams can be hydrotreated without this reduction in stream value, hydrotreating of FCC naphtha should be done to the lowest level possible.

Within the structure of the Tier 3 rules, the flexibility to manage average annual sulfur content provides some flexibility for operating the FCC hydrotreater at the constant degree of desulfurization needed to drive the yearly average to 10 ppm and to ensure meeting the batch limit of 80 ppm. In practicality, it's probably better to target a sulfur level, as long as conditions required to achieve that do not become too severe. Target levels will vary depending on the amount of FCC naphtha in gasoline and the sulfur levels of other blend streams, but these sulfur levels are expected to be in the range of 15-35 ppm.

Reduction of the sulfur target can require different strategies for the hydrotreating of FCC streams. Many sites separate the lighter, higher-olefin FCC gasoline from the heavier, more aromatic cut. The heavier cut goes through a more intensive hydrotreating, while the lighter stream is either blended or processed with technologies like extractive Merox treating to avoid olefin saturation. This treatment may not be sufficient to meet Tier 3 specifications in all cases. If not, further processing of the light material will be required, resulting in more olefin saturation. This saturation, in turn, may drive the need for higher octane generation in other gasoline streams to replace the octane loss from FCC naphtha hydrotreating.

Naphtha hydrotreating

What strategies do you employ to meet cycle-length targets in naphtha hydrotreaters that are reaching catalyst-activity limits due to capacity increases or feedstock quality decreases?

Hutchinson It's not uncommon for naphtha hydrotreaters to be stretched due to capacity increases or decreases in feed quality. Most commonly, the decrease in cycle length is associated with coprocessing of cracked materials, particularly in conventional naphtha hydrotreaters that were designed with low-pressure, high-space velocity and only sufficient for hydrotreating straight-run naphtha. While we talked about some challenges of processing coker naphtha earlier, in this case, the severity required to process cracked feedstocks is increased due to high sulfur and high nitrogen contents. The overall capacity is often increased as well, making processing particularly difficult. Impurities found in the cracked materials-and more frequently in straight-run materials, particularly silicon-can be debilitating to hydrotreaters. In these scenarios, high-performance active metal and silicon guards are critical. Modern active silicon-guard materials are available to both initiate the reactions on olefins (which quickly increases bed temperatures) and trap the bulk of the silicon before impurities migrate to the main hydrotreating catalyst bed. With fixed-bed reactor dimensions, maximization or utilization of the catalyst inventory is necessary. Next-generation silicon guards, developed with the chemistry of silicon adsorption or trapping at their core, are available to provide the maximum silicon retention per volume of reactor.

The improved metal impurity traps are often combined with active grading materials used to address diolefin and other fouling materials found in more challenging and diverse feedstocks charged to the naphtha hydrotreaters. Activity of these grading levels is balanced to prevent polymerization of diolefins and other fouling compounds, which can lead to early shutdown and large opportunity costs, or lost-profit opportunities (LPOs), while still maximizing the saturation of fouling compounds before reaching most active catalysts.

Rhodes I want to add a few comments about some poisons. Silicon is a big concern for our naphtha hydrotreating units. For units that process coker naphtha, working with the coker operation to minimize the use of silica-based antifoams is key to improving cycle length. The silica-based antifoams used in a coker will break down and end up in the naphtha fraction, leaving the unit and affecting your hydrotreater. You can also get silica by processing crudes from locations where silica-based antifoams are used. Understand that the amount of silicon on the beds after you dump the catalyst will allow you to minimize the use of silica trap and maximize the use of active catalysts.

Arsenic is another poison that can be a problem. If a spent-catalyst analysis shows arsenic, you can use an arsenic trap to protect your catalyst. Using your post-audits, then, is very important for maximizing your bed's utilization of active catalysts. We like to use high-surface area catalysts as our active catalysts. Very active catalysts or regenerated catalysts have very low surface area and tend to be greatly effective on poisons.

Finally, coking can be a concern for units operating at low pressures and low hydrogen-to-hydrocarbon ratios and that treat outside feeds or cracked naphthas. In those units, we like to maximize our hydrogen-to-hydrocarbon ratio wherever possible to minimize coking issues.

Shah I want to understand whether you have seen any issues of iron sulfide (FeS) fouling on the naphtha hydrotreater catalyst beds, and if so, whether you have any solutions to those issues?

Philoon Yes. Not very long ago, one of the units I support in Southeast Asia suffered a significant FeS-scale event. Basically, the refinery had a turnaround, during which FeS scale that had developed on the inside of heater tubes for the naphtha hydrotreater charge heater delaminated and peeled off from the tubes' insides. The refiner put a fresh load of hydrotreating catalysts in the reactor, and after startup, experienced a rapid increase in pressure drop across the reactor over the course of just a few hours. This basically caused a loss of differential pressure control across the reactor, and as a result, forced a shutdown because chips of FeS scale covered the top of the reactor bed.

The normal feed to this unit is a fairly high-sulfur naphtha, and the FeS scale had developed quickly over the heater's 5 to 7-year run. In this case, the resolution involved skimming the chips from the top of the bed and cleaning the heater tubes to remove any additional sulfide chips. At a turnaround, we generally recommend that you inspect for the presence of FeS scale on the inside of heater tubes. If you find any evidence of it, you need to clean the tubes either by physically pigging them or by using a procedure to burn off the FeS scale from the tube insides.

Joshi Regarding the feedstock quality, Mr. Hutchinson, you mentioned that by putting this activity in the top bed, it can prevent or lower this polymerization of diolefins. What is the temperature effect on this polymerization? Even at the naphtha hydrotreating template, can this also work if activity on the top bed is just lowered? How about the polymerization of the diolefins?

Hutchinson I think what you are asking is what happens to the diolefins at the top of the bed. They would be saturated so they could not form gums downstream.

Joshi Okay, yes. Actually, one of our customers is having an issue with some diolefins ending up in the fuel. When the customer operates the naphtha hydrotreater, it experiences a lot of polymerization at higher temperatures. The customer asked this question because it doesn't experience the diolefin saturation upstream of the naphtha hydrotreater. So whatever the feedstock, it processes everything in the single reactor.

Hutchinson Yes. I think that probably needs to be looked at on a case-by-case basis. It depends on how your unit is designed as to whether you can put in some kind of a selective hydrogenation or grading to help out.

Joshi Thank you.

Srivatsan I think Kenneth mentioned arsenic as a contaminant. Have you measured arsenic in naphtha, and if so, what were the levels?

Rhodes We haven't measured arsenic in the feeds. In our naphtha hydrotreating units, arsenic is usually very low. We've seen it once or twice in our spent-catalyst results but at very low levels. So, for us, it hasn't been a big issue to date. It's something we are watching, however, because we've heard what it's like in other refineries' process units.

Hutchinson Yes, I think a lot of these contaminants can be very difficult to measure in the feed. Often, the best place to find them is on the catalyst after you dump the bed and do the analysis to see what has collected on your catalyst. This includes arsenic as well as other contaminants such as sodium.

Wagner Feedstock source and chemistry can have significant impacts on hydrotreater catalyst performance. Phosphorous, mercury, and arsenic can reduce catalysts activity. Corrosion products such as FeS can foul the catalyst in straight-run feedstocks. In cracked feedstocks, especially coker naphtha, unsaturated compounds may form polymers that can coke up on catalysts. Residual silica from silicone antifoam used in the delayed coker can also be detrimental.

Strategies to sustain or increase cycle length targets include:

• Implementation of a corrosion inhibition program-either by selection of construction materials, chemical treatment, or a combination of both-in the unit upstream can significantly reduce inorganic fouling. When using high-temperature corrosion inhibition (HTCI) to process high-total acid number (TAN) crudes, Dorf Ketal's Tanscient chemical inhibitor can reduce phosphorous added to the crude by up to 80%.

• Proactively determine metal content in the crude and develop a crude blending strategy to minimize the impact of metals. Dorf Ketal's non-acid reactive adjunct desalter chemistry can supplement emulsion breakers fed to the desalter and remove iron and calcium to increase flexibility in the selection of crudes.

• Implementation of a chemical treatment program containing antifoulant chemistry-which may include antioxidants, organic, and-or inorganic (FeS) dispersants-has proved successful in increasing unit run-length.

The panel

Matt Hutchinson, reforming technology manager, Axens North America Inc.

James Kleiss, director of strategic planning and economics,

Valero Energy Corp.

Steve Philoon, senior technology specialist, UOP LLC

Kenneth Rhodes, reforming and petrochemical technologist,

Marathon Petroleum Corp. (MPC)

The respondents

Jeffrey Mueller, Illinois refining division operations manager, MPC

Patrick Bullen, UOP

Jocelyn C. Daguio, UOP

George Hoekstra, Hoekstra Trading LLC

Andrew Moreland, Valero

Jeff Bray, UOP

Parag Shah, Dorf Ketal Chemicals (I) Pvt. Ltd.

Umakant Joshi, Crystaphase

Srini Srivatsan, Amec Foster Wheeler PLC

Ralph Wagner, Dorf Ketal Chemicals LLC

Acidizing; UOP Answer Book response

Daguio: Acidizing is used primarily to prevent water from forming and deactivating catalyst, but it also serves an important function in preventing corrosion of the piping. The following outlines the mechanism for generation of water during the normal operation of the unit. Acidizing the unit removes the source of the oxygen using the same mechanism before catalyst is loaded in the reactors.

The Penex unit is constructed of carbon steel, so the pipe and vessel walls are mostly iron. When that iron is exposed to the atmosphere, it reacts with oxygen to form iron oxide. In new units, there is often a rust scale on the piping that has been exposed to air and moisture for months. After turnarounds, there should be much less scale, more like a light rouge coating of iron oxide on the piping or vessel walls.

During normal operation, the catalyst promoter PERChloroethylene (PERC) is injected. PERC reacts with hydrogen in the presence of catalyst and heat to create hydrogen chloride (HCl), which is carried throughout the reactors as well as through the heat-exchange train and connecting piping. In the stabilizer column, HCl is carried overhead into the caustic scrubber where it is neutralized by the sodium hydroxide in solution. HCl is necessary to maintain the acid sites on the catalyst so that the reactions will take place.

HCl reacts with iron to form iron chloride and H2. This is a harmless reaction, and in a dry environment, the iron chloride provides a protective layer to prevent exposed iron surfaces from further corrosive reactions.

HCl also will react with iron oxide to form iron chloride and water. If rust is present in the reactor loop, water will form, resulting in permanent deactivation of some catalyst. Also, as iron chloride is soluble in water, when water is present, the protective iron chloride layer will breakdown and expose more iron to react with HCl. If water remains in the system, this cycle can continue until there is enough iron loss to result in a leak.

To prevent this corrosion, the piping is acidized with HCl before commissioning flow through the reactors and normal PERC injection. The iron oxide will be reduced to iron metal and water. The water generated must be drained from the system until the moisture readings indicate the system is dry.

Refer to the UOP general operating manuals for more details on the acidizing procedure.