Search

Olivier Le-Coz (Axens)

More severe regulation in term of Benzene in the gasoline pool can lead to increase the Benzene content to the C5/C6 isomerization unit. This can happen in two different ways.

The refinery process operation can be modified to decrease the benzene precursors content in the heavy naphtha to the Reformer. This is achieved by increasing the light naphtha end point in the topping lights ends naphtha splitter, light naphtha being the Isom unit feed. At the same time C7+ in the Isom feed must be limited to 2 – 3 vol% as those products will undergo undesired hydrocracking reactions in the Isom reactors. With such a scheme, straight run naphtha Benzene (native Benzene) is basically treated in the Isom. This approach can typically be applied in the frame of a new project.

When the “pre-fractionation” scheme cannot be implemented or if it cannot allow reaching the overall pool Benzene specification, a “post-fractionation” option can be implemented. It consists in splitting the reformer product and recover a light Benzene rich reformate which will be treated in the Isom unit in blend with the light straight run naphtha. Depending on the Isom unit existing configuration, some modification to the hardware may be required or not. As a matter of fact, Benzene concentration at the Isom reactors inlet should better not exceed certain value to ensure proper operation and performances of the Isom catalyst (about 4 vol%).

-If the Isom unit is equipped with a recirculation, the recirculated stream acting as diluent may allow maintaining Benzene below the desirable value at the reactor inlet. The extra Benzene amount in the feed will be hydrotreated by the Isom catalyst without disturbing too much the operating conditions and without preventing suitable isomerization rate to be achieved.

-If the Benzene content at the inlet of the reactors cannot be maintained low enough (too low or no recirculation), a dedicated Benzene saturation reactor must be added.

In the case of new units implementation, those schemes have proved to work very well. In the case of revamp projects, existing equipment modifications or idle equipment reuse, a through basic design study upfront including the catalytic aspects is strongly recommended.

Brad Palmer (ConocoPhillips)

In general, refineries with C5/C6 Isomerization units or Aromatic Extraction units have increased feed rate and/or benzene content to these units. Reformer octane has gone down due to ethanol blending but, in most cases, Isom octane demand has remained strong. The primary successes include implementing these projects safely and achieving our benzene reduction requirements. Additionally, heavy reformate blend qualities have improved which has made blending premium gasoline easier and has provided additional opportunities for blend component sales.

A number of technology options were chosen by ConocoPhillips refineries to meet benzene regulations according to the existing configuration and site economics. These options include revamping Aromatic Extraction Units (AEUs) to increase feed and benzene production capacity, sending light reformate or heart cut to other AEUs, modifying C5/C6 Isom units to include benzene saturation reactors, new benzene saturation unit construction, reducing benzene production through prefractionation and use of credits.

All completed projects are working, some with very few operating problems and a few with requiring design modifications and/or operating changes. Operating, design and reliability issues continue to be worked to improve unit performance; a few specific examples are provided below.

Isomerization Unit Challenges

-When all benzene saturation reactors are complete, two will have Pt catalyst and four will have nickel catalyst. One of the reactors will have changed from Pt to Ni.

-The units that added a benzene saturation reactor in front of their Isom reactors have had challenges controlling temperatures profiles of all three reactors especially when liquid recycle is added or removed.

-Isom units have heavier feeds (increased X-Factor). One unit has and XF of 30 lv% average (35 lv% highest) and 9 lv% Benzene Average (10 lv% highest). Another unit has an XF of 25 lv% average (27 lv% highest) and 5 lv% Benzene Average (10 lv% highest).

-Determining when and how much liquid recycle is necessary for safe operation while maximizing fresh feed throughput has taken time. Vendors advertise an upper benzene level of 5 lv% to the inlet of a benzene saturation reactor. While we have gone a little higher by lowering the inlet temperature to accommodate the exotherm, this is a good rule of thumb.

-Increased unit rate can impact dryer operation by fluffing up-flow desiccant beds. Higher rates increase HCl loading to existing caustic scrubbers; less than adequate neutralization can lead to corrosion problems.

-Benzene saturation catalyst has been deactivated or poisoned by feed (organic sulfur, H2S, FeS, Chlorides) or hydrogen purity (CO and CO2) problems.

Aromatic Extraction Unit Challenges

- Changes in feed quality have required operations to find new equilibrium; one unit has reported bigger swings in aromatic content with new feed streams.

- Stripper foaming has occurred in one unit.

Ujjal Roy (Indian Oil Corporation)

In India, the benzene specification in gasoline is 1 vol.% max. In order to meet this specification, number of changes in the refinery configurations have been done. (a) Light Naphtha splitter has been introduced to produce C-5 & C-6 isomerization feed. (b) Naphtha splitter modified to reduce Benzene precursor from Reformer feed Naphtha. (c) Reformate splitter has been installed to separate Benzene from the Reformate. Over and above FCC gasoline being a component of Gasoline, for reduction of Benzene, a FCC gasoline splitter has been put to take away the Benzene rich cut called Heart Cut from Gasoline. For meeting benzene regulation in the Gasoline, Isomerisation unit has been designed with catalysts having dual functions – Isomerisation and complete saturation of Benzene. The metal sites are used for saturation of benzene and acid sites are used for isomerisation of C-5/C-6. Up to 9.8 vol.% benzene in feed, catalyst is able to saturate to nil level of benzene in isomerate.

Erik Myers (Valero)

The Valero approach has been to consolidate the benzene rich streams from various refineries and capture benzene as a product stream. This has been accomplished through use of a side draw stream from the reformate splitters and then feeding these streams through a centralized benzene extraction unit.

Olivier Le-Coz (Axens)

The countermeasures to pressure drop build-up in naphtha hydrotreaters units obviously depend on the cause of the fouling. The two main causes that we know in Naphtha HDT units of are corrosion particles usually coming from outside the battery limit and gums or coke. Axens addresses those two issues at design stage.

Corrosion particles are a potential problem and we indeed address that by implementing both feed filters and grading beds. Feed filters are specified as mechanical filtration devices, usually using metallic cartridges. We don’t have experience with magnetic filters.

As regards grading, Axens uses in its new unit's design or in catalyst replacement loading diagrams a wide range of products from inactive and high void fraction materials to lower void fraction and active products which can also address the removal of specific contaminants. Grading materials have proven to be efficient against particles in many cases. Grading arrangement can be studied on case-by-case basis and when relevant. Axens can propose arrangements of newer generation high void fractions and various pore size materials called CatTrap.

The gums and coke is a problem with units treating olefin and especially diolefins rich feeds from FCC or Coker Naphtha. The important considerations when designing a unit to treat such feedstocks are:

- avoid storing these materials but treat them directly from fractionation columns.

- foresee the injection of an antioxidant chemical if storing cannot be avoided.

- avoid hot spots in the heating system that would cause diolefins to polymerize, optimize the feed preheating scheme to avoid the use of a fired heater, or to reach the full vaporization point ahead of the heater.

But the most efficient way to stay away from gums pressure drop issues is to implement a selective hydrotreating reactor upfront the main HDT reactor. At low temperature and using a dedicated selective hydrotreating catalyst this pre-treatment reactors eliminates the diolefins without giving them a chance to coke further downstream in the process in the heater or at the top of the main and more active HDS catalyst which operates at higher temperature. Axens has been successfully applying this philosophy in many Coker Naphtha and FCC Gasoline (PrimeG) units. We have successfully revamped diolefins rich Naphtha HDT units with addition of a selective pre-treatment reactor, achieving a dramatic decrease of the downstream equipment fouling rate.

Ujjal Roy (Indian Oil Corporation)

We have much naphtha hydrotreaters (NHT) in our ten refineries, some operate with total straight run naphtha and others with cracked gasoline varying from 10% to 40% in feed. In many of the hydrotreaters, we have experienced run length limitations due to high-pressure drop-in reactors or pre-heat circuits while the catalyst was still active for continued operation. Depending upon the basic design, source of feedstock and its composition, we have feed filters (magnetic or cartridge or candle) in all NHT along with graded bed in few of them.

In one of our NHT processing 40% FCC gasoline, we have both magnetic feed filter and graded bed. But even in this unit, we have experienced pressure drop problem in CFEs due to caustic carry over from up-stream FCC unit. In another unit, where we have added graded bed 3 years back in addition to cartridge filter in feed, we have observed increased run length after addition of graded bed. In another unit, in which we have basket and cartridge filters in series but no graded bed, we had to do three skimmings in four years’ operation. The feedstock for this unit is straight run naphtha comprising about 30% material transported from other refinery by tank wagon. We have, off late, replaced the transportation from tank wagon to pipeline and directionally there is improvement in pressure drop, perhaps due to reduction in oxygen and iron pick-up from tank wagon. In another unit, processing about 10% FCC gasoline, having both cartridge and magnetic filter but no graded bed, skimming had to be done five times in six years. The reasons for the high pressure drop as observed after opening of the reactor bed is found to be central hip created at catalyst top bed. This may be due to some design deficiency. In all our hydrotreaters where we are processing cracked gasoline, we directly route the gasoline to hydrotreater with provision of intermediate balancing tank. All these tanks are nitrogen blanketed. Also, in one of the refineries, we inject antioxidant stored in the tank.

General causes contributing to pressure drop in hydrotreaters are either iron scales or coke/polymers. Iron scales are carried with feed from up-stream equipment like tanks and piping. Magnetic and other filters would be helpful in arresting foulant coming from up-stream units and tanks. Coke and polymers come from CFEs and charge heaters. High olefin content in feed than design and dissolved oxygen picked up during storage will aggravate pressure drop. In some case, we have observed high sodium content in the crusts formed on top bed. The source of sodium is likely carry over from up-stream unit. We have taken additional operating measures in up-stream unit to arrest sodium carry over.

In many instances in hydrotreaters, we have observed spikes in Delta P after restart of compressor subsequent to its tripping. It is suggested by our licensor that this might be due to two phase flow at the start of the compressor carrying coke from CFEs and charge heater to reactor. To avoid two phase flow, they recommended to reduce reactor pressure considerably when starting the compressor and to increase the reactor pressure only when reactor attains 260°C and above.

Where we have coker naphtha in feed, in one of the hydrotreaters, multi-layer grading beds have been used. The selective hydrogenation upfront also acts as guard to hydrotreater.

Most of the pressure drop problems in hydrotreaters are unit specific and might have been overlooked in design stage. Preventive measures can only be determined through careful studying the problem over run length.

Brad Palmer (ConocoPhillips)

ConocoPhillips generally uses graded beds on all our hydrotreating units. Several units also have feed filters. The graded beds are usually adequate for all but the worst cases, in spite of precautions. We have experienced extreme cases of upstream corrosion that have forced us to occasionally skim reactors and clean preheat exchangers, in spite of precautions. The upstream problems were eventually corrected by alloy or chemicals, although we prefer to avoid too many chemicals in the naphtha feeds. The difficulty with iron sulfide in units is that the particles can be extremely small (< 1 micron), so filtration is not always effective. Filtration for the naphtha units is usually cartridge filters.

Another more frequent cause for fouling in our system is polymerization of cracked feed stocks. This is promoted by exposure of the feed (or any feed component) to oxygen in tankage but can also be caused by numerous other polymers initiating factors. Filtration is not often effective at removing the polymers, except for those gums already formed in tankage. Additional polymers form rapidly during preheat, downstream of any filtration. The polymers or gums will foul the preheat exchangers, fired heater and the reactors, if they make it that far. To manage polymers in a naphtha hydrotreater, we prefer to add antioxidant to the cracked stocks as they rundown to tankage and add anti-polymerants to stocks as they are feed to a unit. The chemicals help mitigate polymerization, but do not completely prevent it. We also make sure that the dry point of the feed is reached ahead of the fired charge heater to prevent polymer lay down in the heater, subsequent coking, and potential tube failure.

Erik Myers (Valero)

Valero uses the following key approaches:

1. Filter the feed

2. Aggressively use grading material as our naphtha hydrotreaters are not activity limited

3. Utilize mechanical solutions where they look to be effective, such as trash baskets or pressure drop reducing inlet trays.

A key operating area to focus on is avoiding two phase flow in the charge heater. Liquid in the charge heater can lead to coking which when thermally cycled will transfer this coke to the reactor. Similar transfers of iron scale can take place with upsets in any upstream fractionation towers or other equipment.

This topic was also covered in detail in last year’s Q&A (Gasoline question #35). I recommend referencing the transcripts from that review for more information.

At least one of our sites has had very good success with a chemical treatment program incorporating dispersant and antioxidant components to significantly extend the run length of the feed – effluent exchangers. Feed effluent exchanger fouling was also covered in depth as question #36 from the 2010 Q&A session. The 2010 answers for gasoline and FCC naphtha hydrotreating also provide good information on the impacts of olefins and feed gum and polymerization impacts.

Brad Palmer (ConocoPhillips)

There are a number of vertical RODbaffle® exchangers in ConocoPhillips Reformer preheat service. They have very good heat transfer and excellent reliability. They rarely foul; thus, seldom need cleaning. When polling the network, it was common to hear that a bundle had never been cleaned or never needed repair. Those exchangers cleaned were after multiple turnaround cycles, 10+ years. A Reformer application is provided below along with more detailed technology information from one of our heat exchanger experts.

One of our CCRs installed two very large, vertical RODbaffle® CCR Combined Feed Exchangers in 1996. Both exchangers were 83" in diameter and 40' long. The hot exchanger had a HAT of 60 °F, a CAT of 71 °F, a surface area of 48,500 ft2 and a process duty of ~ 160 MM Btu/hr. The cold exchanger had a HAT of 71 °F, a CAT of 15 °F, a surface area of 47,400 ft2 and a process duty of ~ 145 MM Btu/hr. These exchangers have just been cleaned for the 1st time since they were originally installed in 1999, and only because we wanted to inspect the equipment, not because of fouling or problems.

Over 2500 RODbaffle® exchangers have been installed in numerous process applications since ConocoPhillips patented this technology in 1971. RODbaffle® exchangers have been designed in all common TEMA configurations, having shell diameters in excess of 180" and over 80' long. Feed/Effluent service is the most common application for vertical RODbaffle® exchangers, though there are many vertical RODbaffle® thermosiphon reboilers and gas coolers. Thermal/hydraulic performance of a RODbaffle® exchanger is similar to that of a double-segmental plate-baffle unit on maximum cut and spacing. RODbaffle® offers improved shell-side flow-field uniformity, reduced form drag and flow reversals, lower weight, and reduced shell-side fouling rates when compared to traditional plate-baffled exchangers. Tube-side cleaning for RODbaffle® units is identical to that of plate-baffle units. Shell-side cleaning for RODbaffle® exchangers is normally much easier than plate-baffled exchangers, as shell-side solids collection does not occur in RODbaffle® the way it does in the stagnant flow zones found in plate baffle exchangers. There is typically not a materials spec break inside RODbaffle® exchangers. The support rod and baffle ring materials will generally be the same as that of the tubes to prevent dissimilar material corrosion. In shell-side boiling feed-effluent applications, RODbaffle® offers far superior vibration resistance to traditional plate-baffle exchangers. This is also true in high-pressure shell-side gas cooling applications.

Both HTRI and HTFS have a RODbaffle® option in their respective shell & tube exchanger design programs. Matthew C. Gentry, P.E. wrote a paper for the 1999 NPRA Meeting in San Antonio entitled, Industrial Applications for RODbaffle® Heat Exchanger Technology which provides a great deal of technical information about this technology. The NPRA RODbaffle® paper identification number is AM9