Search

Gopi Sivasubramanian (Foster Wheeler USA Corporation)

Middle distillates are maximized within the constraints of D86, flash point and density specifications. There is some optimization play between the pumparound heat recovery, which is specific to a unit, andmaximum distillate recovery. There are other unit constraints such as column overhead temperature, etc. that will require detailed analysis by someone who is familiar with such multi-parameter optimization.

BURTON (Motiva Enterprises LLC) 

Hydrogen needs to be dry and have less than 20 ppm of contaminants such as water, light hydrocarbons, CO, CO2, H2S, and NH3. Each of these can have different effects on the catalyst’s final activity. The purpose of the reduction step is to remove the oxygen from the catalytic metals while minimizing chloride stripping. Contaminants either compete with this step or, for example, light hydrocarbons might react with the metal oxide forming CO, which will poison the platinum and result in lower activity.  

LAMBIE (KBC Advanced Technologies, Inc.) 

Charles covered most of it. I will add that the water strips the chloride from the catalyst, and it could lead to increased downstream corrosion. CO is a poison for the unreduced catalyst and will lower the activity. CO2 is a precursor to CO. To a certain extent, hydrocarbons react with the metal oxides to form CO, which poisons the catalyst. My only other comment is that it would be ideal for the hydrogen medium to be electrolytic hydrogen, if it is available. Alternatively, hydrogen plant hydrogen is acceptable as long as the CO or CO2 is low. A minimum of 90% purity should be used in all cases. 

DUNHAM (UOP LLC, A Honeywell Company) 

All I want to add is that UOP recommends electrolytic hydrogen for promoted catalyst. 

SCOTT LAMBIE (KBC Advanced Technologies, Inc.) 

A good hydrogen reduction of the catalyst is one key to the proper operation of a reforming unit. A good reduction step dries out the catalyst, reduces the metals, and improves the metals dispersion while avoiding excessive stripping of chloride from the catalyst and low-pH corrosion of wet effluent equipment. The reduction step is done in a hydrogen atmosphere with adequate hydrogen partial pressure. 

The hydrogen supply should be dry, contain no CO, CO2, H2S, or hydrocarbons heavier than ethane to prevent poisoning of the catalyst before the reduction is complete. Ideally, the H2 purity should be greater than 90%. Electrolytic hydrogen and hydrogen from a hydrogen plant with a low CO content are recommended sources of reduction hydrogen. 

The main problems associated with different qualities of hydrogen used for the reduction step differ depending on the impurities present. Hydrocarbons, particularly propane and heavier hydrocarbons in the reduction gas, can react with residual oxygen and the metal oxides on the catalyst to form CO. CO is a catalyst poison for the unreduced catalyst, and the result is a lower activity of the regenerated catalyst. Note that any recycle gas driers should not be put online until after the initial reduction step as hydrocarbons or CO/COx may be present.  

Water present in the reduction gas will increase stripping of chlorides from the catalyst and inhibit metals reduction. This will result in increased corrosion in the regeneration system and will slow up the line-out period after oil-in. Water is generated in the reduction process, so it is important to regularly drain all low point drains to remove water that is formed in the process. Alternately, recycle gas drying can be used once initial reduction is complete. This will also speed up the line-out period.  

H2S will also inhibit reduction, so presulfiding is done after the reduction is complete. Ideally, any high H2S or SOx should be detected and eliminated before the final reduction. This is done with a desulfation step to enable good oxychlorination. H2S can still be produced during the final reduction step, so H2S levels should be monitored. If high H2S levels are seen, then a decision must be made as to whether a second desulfation step is justified, though this is rare. Typically, high H2S at this stage is a result of iron scale buildup in the system and reactors and may indicate a need to clean the equipment and/or screen the catalyst. 

FRY (Delek Refining) 

With regard to reforming economics, the key criterion to determine catalyst condition is the surface area. As the surface area decreases, you will see yield loss on your unit, and you will have to raise the temperature in order to compensate for the lower surface area. Raising the temperature results in more cracking; so you will want to monitor the surface area. You will also want to monitor the volume loss across the unit, as well as monitoring your off gas rates.  

There will come a point where you will save more in yield improvement by changing out the catalyst, and it will simply pay for itself. What that is will be unique to each facility. 

LAMBIE (KBC Advanced Technologies, Inc.) 

For semi-regeneration units, contamination, such as iron in the lead reactors, or unit upsets are more common issues. Skimming and catalyst replacement is a common practice to prevent downstream contamination. Skimming may occur every three to five years, depending on the effectiveness of the regenerations. Semi-regen units’ catalyst surface area and activity are not as much of an issue these days as reformers are typically operating at low severities. 

For CCRs, catalyst surface area and activity are typically limiting. As the surface area reduces, the ability of the chloride to be retained on the catalyst reduces, thereby requiring high chloride injection rates, which can lead to increased downstream salting and fouling issues. 

Other issues more common with CCR units that may lead to a changeout would be poisoning of the catalyst from, say, silicon contained in coker naphtha or arsenic. Poor regenerator operation can lead to sintering of the catalyst. Increased fines production in the CCR can also lead to issues, but you have the ability to replace the catalyst online for CCRs. 

From a cost standpoint, the cost of semi-regens versus yield improvements, in today’s scenario where people are running very low severity reformers and getting long run-lengths, is frequently not the deciding factor. Often new catalyst formulations will come out with improved yields that may warrant a change in catalyst. Similarly for CCRs, yield improvements and activity are the deciding factors. Chloride handling costs become an issue as the surface area is reduced and more frequent changeouts of the chloride guard beds are required, which results in increased costs. And again, new catalyst formulations would improve yields and stability. 

PATEL (Valero Energy Corporation) 

So broadly, it depends on the operating severity. Due to lower severity operations and the fact that our CCR is a kind of noncontinuous regen type of operation we had, we changed our catalyst after 13 years of service. That catalyst change occurred during a scheduled unit inspection that required removal of the catalyst. This changeout during a reactor inspection was based on an economic evaluation. We were guaranteed C5 plus yield improvement by the licensor with the new catalyst. 

KEVIN PROOPS (Koch Industries, Inc.) 

If you have a semi-regen reformer, pay attention to your catalyst activity because the cycle length can very drastically. I know when I worked with Daryl years ago, one of the Conoco refineries had gotten some fairly old catalyst in a semi-regen unit; and by changing to new catalyst, it went from a six-month cycle to over a two-year cycle. So you can eliminate many of the regenerations by having a more active catalyst.  

SCOTT LAMBIE (KBC Advanced Technologies, Inc.) 

For semi-regenerative reformers, catalyst life is often dependent on contamination or upset-related issues. Lead reactors can become contaminated with iron, which ultimately affects the product yields. Skimming or possible replacement to prevent downstream contamination is a common practice. Depending on the regeneration procedures used, skimming may take place every three to five years. 

Surface area reduction of the catalyst does not typically affect the yield loss across the unit. In extreme cases, surface area may be low and the catalyst activity affected. In some cases where good regenerations practices are followed, tail reactors may achieve up to a 10-year catalyst life. 

Contaminants such as arsenic or silicon are permanent catalyst poisons for reformers and will have a detrimental impact on the catalyst activity if they are present in high enough quantities. These poisons should be managed in upstream hydrotreaters. 

The main economic criteria for changing the catalyst are the cost and yields. In many refineries today, as a result of lower feed rates and/or lower operating severities, run-lengths are approaching two years and catalyst life is extending upwards of 10 years with only a relatively small change in yield loss. New catalyst formulations with improved yields and stability are increasingly dictating catalyst changeout decisions. 

For CCR units, catalyst surface area reduction and activity loss are the main technical contributors to the decision to change the catalyst. The inability of the catalyst to retain chloride during regeneration results from the low surface area of the catalyst. Higher chloride injection rates and the resulting downstream fouling and corrosion become prohibitive to continuous operation for extended periods.  

Contaminants such as arsenic or silicon will have a detrimental impact on the catalyst activity. These poisons should be managed in upstream hydrotreaters. 

CCR units have more mechanical equipment and are more prone to potential failures or upsets that may affect the catalyst life than semi-regenerative reformers. Poor regenerator operation could lead to catalyst sintering or increased fines production. CCR units do have the capability to replace catalyst online which helps to maintain overall activity, but there is a cost associated for the lost catalyst. In some cases, consideration of the number of regeneration cycles weighs on the decision to change the catalyst. 

Again, the main economic criteria for changing the catalyst are the cost and yields. The ability to control downstream fouling and corrosion resulting from increased chloride injection due to surface area loss may result in increased downtime and costs associated with water-washing and/or increased chloride bed changeout costs. New catalyst formulations in the market with improved yields and stability may justify a catalyst change. 

PATEL (Valero Energy Corporation) 

The major issues with some of those originally installed ball valves in CCR cyclic service are packing leaks. Hydrogen leaking into the atmosphere causes small packing fires. All too often, these failures cause process interruption and frequent ball valve changeouts. Repair and replacement create a situation that places personnel and others in danger. Our licensor dropped the old ball valve vendors from the approved vendor list. 

There are better performing walls in the cyclic service that provide longer life and better reliability: the common failures experienced in the older design. All of those styles are minimized due to their superior design in the areas of packing seal and coating of the stem, which handles the frequent cycling. The detailed specification of these walls and approved supplier list, as well as their installation and the testing procedures, can be found in the licensor process specifications. 

We changed eight of the old-design ball valves. They were replaced with the licensor-approved double-seated ball valves. Six were in hydrogen service and two in nitrogen service. The reason for the upgrade was poor MTBF (mean time before failure) due to numerous packing fires, leaks, and seating problems. Those valves were also removed from the licensor approved list.  

After initial installation, we had to make few modifications. We had to install a stronger ball valve actuator. We had to change out the packing and lubrication because they were designed for the temperature that was specified in the original valve spec. We were not reaching that temperature. After addressing those issues, we have not had any incidents with the packing fire in the last two years, and the internal leak problems are also at minimum. 

KEADY (Technip) 

A gentleman I know was an Operations engineer at a client startup. He said that one of the units he worked on had issues with some of the original valves. There was wear and leakage after only a brief number of cycles, and the client had all sorts of trials with other valves and materials, including ceramics. Eventually, the client sorted out the problem, and then it became less of an issue. We do have a client in India who did not have problems with the ball valves on the CCR platform where this has been for 10plus years. 

FRY (Delek Refining) 

All I can add is that we use a metal-seated ball valve in our service, and it is generally reliable. 

ALMA SCHURIG (Big West Oil) 

We see some extreme cycles in the valves in our reformer, which uses a unique catalyst regen system (CycleX). One of the issues that we experienced, which has not yet been discussed, is that, in addition to the valve selection, we had some fairly severe piping stresses that were put on these valves just due to the piping configuration. We installed some expansion loops in the piping, and it helped a lot. We also went through several rounds trying different valve manufacturers and valve types. We have had best success with Argus metal-seated ball valves and good success with Everlasting metal-seated disc valves. The combination of piping modifications and improved valve selection has helped us quite a bit. 

WAYNE WOODWARD (Valero) 

With respect to the maintenance, a point lost on the turnaround folks is that you commonly pull all of these lock hopper valves (20, 30, or 40 of them) and have them serviced, and then they come back to the unit. The metal balls and metal seats always leak. If you ask your instrument man if the valves passed a bubble type of test, he will say, “Yes, it did.” Well, brand-new valves leak. My point is that you need to do the right leak check in the maintenance cycle on these valves. A brand-new valve fresh from the factory leaks 10% of allowable. Talk with your licensor about the leak check and its limits and understand that all metal-seated ball valves leak. You will need to do the leak check correctly to ensure that the valves only leak an acceptable amount. You also need to know that you are installing valves fit for service after you have done your turnaround. 

GRAHAM NEWMAN (Emerson Process Management) 

From the question, it is not clear which style of valve in cyclic service is of the highest interest. There are two types of cyclic services that are specific to the CCR process, each with different requirements and potential pitfalls. There are on/off valves in vapor service with catalyst dust, such as those that vent the lock hoppers, and the on/off valves in flowing catalyst service, such as those used for catalyst flowing in and out of the lock hoppers. The on/off valves in the dust-laden vapor flow require tight shutoff and the capability to withstand the erosive effects of flowing catalyst. Early solutions to this application used a complicated plug design which involved moving parts and springs exposed to the process fluid, which often became jammed due to catalyst buildup. This additional complexity often meant that regular maintenance could not be performed onsite for these valves; they had to be sent back to the factory. Fisher’s EZ-OVT design solved these issues by using a valve plug with no moving parts, using a dual seal and a deflector ring to keep erosive flow away from the soft sealing surface of the plug that provides the tight shutoff. This extends the operating life of the valve between service intervals. The soft components that are in contact with the process fluid are entirely replaceable, including the soft material in the plug which provides the tight shutoff. Maintenance procedures are almost identical between the EZ-OVT and the standard Fisher EZ, and the majority of the valve’s components are common to the regular EZ. This ease of maintenance and lowered parts consumption reduces these valves to a standard maintenance item at turnaround time rather than a valve that needs special attention. 

The challenge for control valves in flowing catalyst service in the CCR section is to avoid damaging catalyst as much as possible when the valves close against the downward flow of catalyst. A segmented ball valve with a clearance gap between the ball and seal is specified. This clearance minimizes catalyst from being crushed between the ball and seal when the segmented ball rotates during valve closure, which would increase catalyst attrition. The segmented ball valve with clearance gap is specified for units that operate the regenerator close to atmospheric pressure. For more current designs, the regenerator operates at the fuel gas header pressure, typically around 35 psig. The same concerns about crushing catalyst apply in the pressurized units, with the additional requirement that the control valve provide bi-directional shutoff capability. For these services, a segmented ball valve with a zero-deflection seal can provide shutoff without excessively damaging catalyst during valve movement. Fisher has seen success using another variant of the Fisher Vee-Ball™ line, the SS-252B, which includes these design features. These valves are not in a high cycle service but typically only close during abnormal operation, which also minimizes catalyst damage. 

GINGER KEADY (Technip) 

The original UOP CCR used Hills McCanna ball valves, and there were issues with wear and leakage after only a brief number of cycles. There were trials using other valves and materials including ceramics. Eventually, the valves became less of an issue. 

We have an Indian client that has had no problems with ball valves on its CCR platformer which has been in service 10+ years.