AFPM Q&A-1: Safety, gasoline processing questions addressed at annual conference

Aug. 3, 2015
Gasoline processing operations, with a focus on safety, blending, and reforming issues, garnered considerable attention during the 2014 American Fuel and Petrochemical Manufacturers Q&A and Technology Forum (Oct. 6-8; Denver).

Gasoline processing operations, with a focus on safety, blending, and reforming issues, garnered considerable attention during the 2014 American Fuel and Petrochemical Manufacturers Q&A and Technology Forum (Oct. 6-8; Denver).

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 2014 event. Part 2 in the series (OGJ, Sept. 7, 2015) will focus on hydroprocessing, while the final installment (OGJ, Oct. 5, 2015) will highlight discussion surrounding processes associated with fluid catalytic cracking.

The session included five panelists comprised of industry experts 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 it can provide sound guidelines for what would work to address specific issues.

Safety

What are the recent safety improvements in the procedures or equipment for sampling sulfuric acid?

PRESLEY: For this response, I am going to focus on three things. The first is the location in the process from which the sulfuric acid is sampled. Second, I will provide some information on the apparatus used for sampling; and then lastly, I will touch a little on personal protective equipment (PPE).

I will start with the location of the sample. This is a somewhat recent change for many refiners. Traditionally, you would take the sample on the transfer or spent-acid line. The problem is that there is a significant lag time when sampling from that location: more than 2 hours' lag from where you really care about understanding the acid strength. What really matters is the acid strength inside the reactor.

So you may ask what this has to do with safety. It matters because it helps prevent acid runaway. From the standpoint of an acid runaway, 2 hrs may be too late, particularly if you are in a situation where you are spiraling downward on acid strength. So for that reason, we now generally recommend that for most of the new units, you sample directly from the Contactor reactor. One of the challenges associated with sampling here, though, is that the mixture is roughly 50% hydrocarbon and 50% acid; so you have to figure out a way to safely sample.

At DuPont, we have developed what we call the Reaction Zone Monitoring System (RZMS). This is a sampling and instrumentation package that allows the operator to better understand what is going on inside the reactor. What this system does is pull a stream from the high pressure side of the impeller and return it back to the suction side in order to create a constant circulating flow.

We have a Coriolis meter that measures the density continuously, which will give you an indication of the acid-to-hydrocarbon ratio. We also have a refractometer that measures the refractive index, which can be correlated to the acid strength inside the reactor. Clearly, continuous monitoring of acid strength is ideal for optimization and acid runaway prevention. We also have the ratio glass as part of the RZMS. The ratio glass is used for acid-to-hydrocarbon monitoring, but it also provides a way to look at the emulsion inside the reactor. In other words, is it frothy? Is it bubbly? It is like a window into the reactor.

The way the system operates is that the emulsion is normally flowing through the ratio glass; and then when you want to take a reading, you basically block the outlet valve and let it settle out for 15-25 min. We use the fact that you are already settling out the acid, so we have installed a sampling system under the bottom of the glass.

The sampling system allows you to first settle the acid and hydrocarbon and then vent off any hydrocarbon before collecting the acid sample. We catch the sample in a capped bottle with a needle that is used to pierce the septum on the top of the bottle. We usually catch a very small sample, which is typically all that is needed, especially if you are using an autotitrator. We have found these systems to work very well.

Now I will add a little about PPE. Other than the basic refinery PPE used when you are sampling acid, we generally recommend chemical goggles and face shields. The reason we recommend both is because we see a lot of refiners who just have face shields and safety glasses but not goggles. If you are getting splashed with acid, your instinct-if something is coming toward your face-will be to turn away. The problem is that if you do not have goggles behind the face shield, then the acid may get into your eyes; so we generally recommend both. We also recommend acid-resistant gloves and an acid-resistant apron or a splash suit.

HUDE: The goals of sulfuric acid sampling: You are trying to obtain a representative sample for process monitoring so you can know the current condition of your unit. As an operator, you want to minimize the chance of any employees or contract workers coming in physical contact with the acid. Also, the acid you are draining, if you have a long deadleg type of system, all has to go down to a sewer where it will be treated; so you want to minimize acid to the sewer where possible.

We had an opportunity to redesign our sample system in 2010. Our primary consideration was trying to minimize equipment operator exposure. We installed a closed-loop sample system that uses needle valves with a sample enclosure. With this type of enclosure, the goal is to keep any unexpected splattering from that sample contained within the enclosure, which reduces the potential for the test operator being exposed. This also allows him to downgrade the PPE required to do the task. The closed loop is tubed down the enclosure box with small legs, which runs down under the installed plexiglas shield; essentially a custom design. There was nothing particularly fancy about it, but the whole point was to reduce the operator's exposure to acid when performing his job duties.

The second type of sample station of which we are now aware is an in-line sample valve. With this device, the sample collection system is completely enclosed. Obtaining the sample requires no draining or flushing, but it is specialty manufactured and requires manufacturer-specific sample bottles.

Here are a couple of pictures (see accompanying photos). On the left is a sample system we built at Houston. Basically, the operator puts the bottles inside, opens this lid, puts the bottles on a tray, and then obtains a sample. He controls the flow rate using the needle valves. The other kind we have seen is the in-line sampling shown on the right. We do not have one at our plant, but one of our neighboring acid plants has it.

We went through our procedures using digital cameras to help provide clarifications and did, basically, a step-by-step photo-op through the procedure. Afterwards, we retrained all the operators to make sure they were all aware of the upgrades and requirements. The goal is to keep them protected and safe.

As a final comment, at Houston we went through and labeled the unit with simple "GOT PPE" memory joggers like advertising. The whole point was to make the operators and maintenance personnel aware and thinking about PPE requirements for the job tasks in the area.

Valero Energy Corp.'s sample station for sulfuric acid at the Houston refinery (a) features closed-loop tubing with an enclosed bottle-filling station, where all valves are operated outside the enclosure to protect employees, while a nearby acid plant uses an in-line sample valve (b). Both systems reduce operator exposure to sulfuric acid.

Blending

In recent years, the gasoline blend pool has shifted due to increased ethanol blending, larger volumes of high-RVP material from processing lighter crudes, and other specifications changes. How are you taking advantage of these changes to optimize gasoline processing units?

MITZNER: The impact from ethanol and shale crudes has created unique challenges in the overall gasoline pool. The first big change came with the Renewable Fuel Standards (RFS) that required up to 10% ethanol in the gasoline pool, depending on local regulations. Other blends of ethanol exist at 15% and 85% ethanol, but the bulk of the gasoline is sold at the 10% blend.

Ethanol, due to its high RVP and high octane rating, changed the gasoline pool quickly. Using a basis of 10% ethanol in the gasoline pool, ethanol allowed refiners to decrease the octane of the traditional gasoline components by around three road-octane points. The primary means of adjusting the octane of the traditional gasoline pool was adjusting naphtha reformer severity. For refiners who run their reformer for the majority, if not all of the refinery hydrogen, this ethanol blend requirement could lead to significant octane giveaway as high severity is maintained simply to produce hydrogen. Those refiners not solely dependent on their naphtha reformer had more flexibility to adjust, but reduced reforming severity has led to very long cycles on the fixed-bed reformers or low-coke operation in the moving-bed reformers.

The RVP implications from ethanol are also pertinent. Butanes and other light paraffins were forced out of the pool, to varying degrees, at each site by ethanol. The RFS did allow an RVP waiver that helped mitigate the impact of ethanol, but the impact of ethanol is becoming more critical now as the naphtha-rich shale crudes are being processed by more refiners at increasing quantities.

There is an increasing glut of light naphtha that must be handled within the refinery. Often refiners would sell off their excess butanes and pentanes, even with moderate sulfur content still in the stream. As refiners move to ultralow-sulfur gasoline, and with naphtha hydrotreaters that do not frequently have excess capacity, purchasing outside naphthas that cannot be directly sent to the gasoline pool has become less palatable. As refiners move towards retaining this excess capacity of light naphthas in-house, reformer severity may go back up to account for the lower octane rating of the extra straight-run naphtha that must be processed.

Yet, shale naphthas, more paraffinic than traditional crude-derived naphthas, require higher-severity operations in the naphtha reformer to meet current target octanes. At several refineries, the economics of processing shale crude have led to older reformers being restarted. The reduced severity of naphtha reforming that was seen during the start of the RFS is being reversed, to a degree, by the paraffinic shale crudes now being produced.

The cracked naphthas derived from shale crudes also bring with them some benefits. As shale crudes tend to be sweeter than traditional crudes, the shale-derived cracked naphthas, similar to the shale-derived straight-run naphthas, tend to be sweeter than their conventional crude counterparts. These sweeter shale-cracked naphthas are helping to minimize the impacts-both capital and operational costs-of moving existing post-treaters into ultralow-sulfur gasoline for Tier 3.

AHERN: When it came to ethanol being added to the blend pool, one of the first challenges was its unpredictability, in terms of how it affects octane and volatility when added to the base oil blend. So before you can start to go out and look at alternative blending opportunities to reduce giveaway, the blending models will have to improve. We found that we had to move away from the traditional volume-based blend models to the more detailed molecular-based blend models, and we have had some success with that. Once you have better blending modeling, you can then start looking at alternative blending strategies.

One of the problems with ethanol is that it increases the RVP of the blend to which it is added but it has the advantage of increasing octane and diluting benzene and aromatics, which allows for alternative processing and blending options.

With regard to C4s and C5s, we have normally used the strategy of sending them to the isomerization unit to improve octane. This octane increase is accompanied by an RVP increase due to the higher RVPs associated with the isoform of a molecule. So now, we bypass the unit and send the C4s-C5s directly to the blending pool. This also has the benefit of cheaper processing costs.

GILMORE: You have described our current issue to a tee. We are long on C5 molecules. Currently, we will sell these into the market for other refiners as a gasoline blendstock or petrochemical feedstock. Are there technologies out there that could convert these C5 molecules into a larger molecule that could be used for gasoline or distillate blending?

MITZNER: I am not an expert in this area of technology. Geoff Dubin is our naphtha block technology manager and may know what would convert a C5 into a larger molecule.

DUBIN: The big question with C5s is the type with which you are starting: saturates or olefins. If you are working with a straight paraffin, you will find a problem that has been going on for 10-20 years; that is, how do we find a way to get pentanes out of the pool? For olefinic streams, you can look at oligomerization which will take those paraffinic C5s and turn them into gasoline or diesel. Those technologies exist. Axens has licensed units that have been in operation for 20 years now for oligomerizing olefins.

Reforming

What are the sources of platinum loss in precious-metals catalysts? What role can your refinery engineers play in minimizing this loss?

MITZNER: Platinum loss results primarily from catalyst losses and poor bookkeeping practices. Actual loss of platinum from the catalyst is less likely and limited to extreme regeneration conditions.

Many years will pass and several unit engineers will rotate through during the lifetime of a reforming catalyst. Having a well-thought-out and executed protocol for handling the catalyst, accounting for its whereabouts, and accurately knowing its mass and platinum content at all times during its lifecycle will enable more accurate closure on the platinum balance post-reclamation.

Catalyst losses primarily occur during handling. Unrecovered spillage often occurs during transportation, reactor loading and unloading, screening operations, etc. Occasionally, spent or fresh catalyst is inadvertently disposed of from the warehouse or at the unit, particularly if drums are not clearly labeled. Catalyst that is never loaded and left in storage can be overlooked even though it remains onsite. Catalyst bookkeeping is complicated by intermittent skimming of catalyst, dump and screens, reloading, and transferring catalyst from one reactor or unit to another.

On oil, catalyst losses result from lost containment either as whole pieces, fragments, or dust. Sometimes this material is lost to the liquid product and winds up distributed throughout the unit and downstream equipment all the way to product storage. This contamination only occurs when there is a mechanical failure that allows catalyst loss from the reactor or circulation system.

Inaccurate bookkeeping will often account for a fair portion of the platinum loss. The alumina used to make reforming catalyst is very hygroscopic. In little time, its weight can change by several percent if it is left exposed to humid air. Contaminants accumulated during the catalyst's lifetime, as well as coke and adsorbed hydrocarbons remaining on the catalyst being sent for reclamation, can also affect its weight. Accurate determination of the loss on ignition (LOI) is critical to accurate bookkeeping. The accurate determination of the platinum content of both the fresh and spent catalyst is also extremely critical.

Samples collected for assay must accurately represent the material from which they are taken. Disproportionate inclusion of inerts, scale, etc. in these samples will impact the accuracy of the platinum assay. Comingling catalysts that are different in size, shape, or platinum content increases the probability of non-representative assay results. Formation of alpha alumina, as a result of severe temperature excursion during regeneration, will encapsulate some of the platinum and render it unrecoverable.

VICE: Catalyst losses occur through the sampling process, fine generation, and continuous catalyst regeneration (CCR) while circulating the catalyst, or as a result of mechanical issues with fixed-bed support systems and during the catalyst change-out process. Relative to the sampling process, it is critical for the lab to have procedures in place to capture all of the catalysts for reclamation. Relative to the CCR engineers, they should be watching catalyst-fines generation on a routine basis, as well as the catalyst makeup. It is really critical to determine if you have a problem earlier to minimize your losses.

During the change-out process, the management of the platinum catalyst is everyone's responsibility, so that goes from the catalyst-handling company to operations and your tech-service personnel. In our system, the engineer would be tracking the catalyst from the moment it arrives in the warehouse until it goes in the unit and is then loaded in the reactor. The same is true about the spent catalyst when it is removed from the reactor to the storage area and then eventually sent out for reclamation.

As highlighted before, housekeeping is really critical during the loading and unloading processes. All spillage needs to be picked up and captured for reclamation. Any catalyst remaining in the reactor needs to be vacuumed up as well.

HUDE: I want to add that you can see losses in catalyst manufacturing. Engineers need to be aware of that if they are paying a first catalyst surcharge for platinum or a surcharge on the reclaim. We have seen losses in this area up to 1%. Also on the CCR, if you do not have a dust collector on Reactor 1, you will be losing platinum from that dust which is going into your reformate tank. So an engineer should work with maintenance to consider any recovery options when you are cleaning a reformate tank.

KAMINSKY: I just want to elucidate a bit about the chemistry going on with the platinum loss from my perspective as a catalyst chemist. There is a volatile phase of platinum, platinum oxide (PtO2), that forms under hot, oxidizing conditions such as in your regenerator. So it is not just fines but actual volatilization of Pt as PtO2, which causes movement of Pt from the catalyst in the regenerator.

This situation has been documented previously in automobile catalytic converters. Ford and GM have published papers on Pt volatility. Such volatility causes Pt to move out of the oxidation catalytic converter (that oxidizes CO to CO2), but then the PtO2 adsorbs onto the downstream selective catalytic reduction (SCR) catalyst and permanently poisons it.

It is a big problem for car manufacturers. They are solving it by alloying the platinum with palladium or other metals to reduce the propensity for platinum to form the superoxide. Maybe such a solution would help reduce Pt loss in FCC units also.

SMALL: Typical operating conditions in a reformer do not result in volatile platinum. However, it is possible for platinum to become volatile and come off the catalyst at very high temperatures. One place this can occur is in the chlorination zone of the CCR regeneration tower where slipping coked catalyst into an oxygen-rich atmosphere can result in very high temperatures. To prevent this, the refiner should make sure that the regeneration tower is operated according to design.

In UOP's experience, these questions often arise as the result of an assay difference with the reclaimer rather than with volatilizing platinum. The assay differences can be due to unrepresentative sampling or poor or biased analyses.

oyekan In order to fully answer the questions, it is relevant to separate the precious metal catalyst platinum management into five distinct stages to cover a platinum catalyst manufacture to spent-catalyst platinum reclamation lifecycle. The stages that are pertinent for our review are:

• Reforming catalyst manufacture by the catalyst supplier and platinum settlement.

• Reforming catalyst storage and catalyst loading.

• Catalyst as used in the reforming units.

• Catalyst dumping and transfer to platinum reclamation company.

• Platinum settlement with the platinum-reclamation company.

It must be clearly understood that platinum losses can occur at any of the stages of the catalyst cycle. Some of the losses are due to contractual agreements as agreed upon in the first and fifth stages as a consequence of platinum settlement. The platinum or precious-metals manager for an oil refining company should have the necessary expertise to aid in minimization of platinum losses for the oil refiner for the first and fifth stages.

In the fresh-catalyst manufacture stage, the agreement with the catalyst- manufacturing company for platinum settlement could stipulate 98-99.5% platinum return for the settlements. The platinum settlement requires that the oil refiner and catalyst manufacturer or platinum reclamation companies for the platinum settlement have appropriate analytical data (platinum assay, LOI for solid content) to permit effective conducting of the platinum settlement.

Some oil refiners conduct platinum settlement with the catalyst suppliers, and some do not. I recommend conducting fresh-catalyst platinum settlements to establish a reference for initial platinum in use in a specific process unit that would be utilizing the fresh- catalyst load, and that the nominal platinum concentrations not be relied on as indicative of the reference fresh- catalyst platinum.

In the years I managed precious metals for two oil refiners as a refinery technologist, several excess platinum troy ounces were returned to my companies' platinum-pool accounts after fresh-catalyst platinum settlements with the catalyst manufacturers. In addition, the fresh-catalyst platinum settlement data provided a good reference basis for the subsequent platinum inventory in the reactors after the catalyst loading.

In the spent-catalyst platinum reclamation, a similar legal agreement could stipulate another 98-99% platinum settlement, with some additional platinum-percent penalties for coke, catalyst alumina state (alpha or delta), and metals impurities. Thus, based on the two platinum settlements for fresh and spent catalyst for a catalyst lifecycle, platinum losses due to contractual agreements and lack of the appropriate level of platinum management expertise by the oil refiner could lead to platinum losses in the range of 3-5 wt% for the oil refiner.

Major additional losses could occur in Stages 2-4. These combined areas of catalyst loading, in-unit catalyst usage, catalyst dumping, and precious metals management are so intertwined and extensive that I strongly recommend securing the services of experienced technical experts who understand clearly the three major catalytic reforming technologies-semi-regeneration, cyclic, and continuous catalytic regeneration reformers-and how their operations could greatly contribute to significant platinum losses.

If you also own paraffin isomerization units and other process units that use platinum catalysts, seek the assistance of a technical expert who fully understands platinum or precious metals management, as well as the operations of the relevant oil refining process units that utilize platinum catalysts.

An excellent oil refining expert could also work with your engineers and other relevant oil refinery personal on proactive steps for cost-efficient catalyst management, process monitoring, and optimization and equipment management to minimize platinum losses.

The panelists

John Ahern, gasoline and catalyst specialist, Phillips 66
Jeff Hude, process engineering manager, Valero Energy Corp.
Michael Mitzner, senior regional sales manager, Axens North America
Shane Presley, technical service leader, DuPont Clean Technologies
Rick Vice, alkylation and Merox technologist, Marathon Petroleum Corp.

The respondents

Christopher Gilmore, Irving Oil Ltd.
Geoff Dubin, Axens North America
Mark Kaminsky, Aramco Services Co.
Troy Small, UOP LLC
Ka Lok, UOP LLC
Soni Oyekan, Prafis Energy Solutions