Refining & Petrochemical : Questions & Answers

+ QUESTION 1

What is the highest reforming pressure (catalyst-tube outlet pressure) in commercial practice today for hydrogen manufacture at 95% purity hydrogen? What is the corresponding heater outlet temperature? How are tube temperatures measured? What operating philosophy is used to balance individual cell outlet temperatures?

SMITH: Riyadh Refinery produces 97% pure hydrogen by processing naphtha reformer offgas containing 79% hydrogen. Processing occurs at 312 pounds, 878 °F inlet, and a 1508 °F outlet.

In general, the steam reforming reaction is favored by low pressure and high temperature. Operating pressure for the hydrogen plant is normally an economic decision. The equilibrium temperatures are also governed by steam to ¬gas ratio for a given operating pressure. To produce a 95% product hydrogen purity requires a steam-to-gas ratio of 3.5. The Riyadh refinery believes they can produce a minimum 95% purity hydrogen with up to 420 pounds pressure with an outlet temperature between 1580 °F and 1590 °F.

On the other part of the question, the infrared base optical pyrometers are used to accurately measure the tube skin temperatures. To balance the cell outlet temperatures, furnace manifolds are fitted with thermocouples. These are regularly monitored to ensure that they are all within 10 °F of their average and bear a constant differential with the combined reformer outlet temperature. Furnace firing and burners should be adjusted to balance the outlet temperature. It is advisable to run a constant reformer outlet temperature; this way, the methane content of the reformer effluent indicates the aging of the catalyst.

HUNKUS: The highest reformer outlet pressure typically used for commercial plants that we know of is about 400 psig. The corresponding outlet temperatures are in the 1600° range. Tubes in this range of temperature are operating in creep failure modes of a limited service life. I wanted to recom¬mend that people newly assigned in this area become familiar with all their heater details because they are a bit different from the rest of the industry.

For tube temperature measurements, a portable optical pyrometer is probably the best tool. I have also done a lot of work with thermocouples welded onto the tubes, shielded, and unshielded, with and without cooling purges, and mounted in the firebox for many different types of services. These can provide valuable information and help develop both instrument calibration and a cali¬brated eye for weekly visual inspections after dark when you can really see. The burner type, geometry, tips arrange¬ment, angle of inclusion of gas ports, their primary angle, and induced swirl angles are some of the most important facts to understand.

I have an old pamphlet that I cherish and I feel this is another area of lost knowledge among plant process engi¬neers and a lost opportunity for many designs especially for smaller revamp projects. The pamphlet is out of print, but I will see if NPRA can make it available.

KODIMAN: Our reformers run about 240 psig. We target 1460°F at start-of-run and 1500 °F at end-of-run. We use a hand held laser pyrometer calibrated in the 1370 °F to 2800 °F range to measure tube wall temperatures. We will then adjust cell outlet temperatures by adding or removing burners as needed.

ARNDT: We operate six large, well-aged hydrogen plants. Their tube outlet conditions are roughly 300 psi and 1550°F. As Mr. Smith said, high pressure is unfavorable for the re¬forming reaction equilibrium. In addition the higher the pressure, the thicker, and more expensive the tubes will be. That is an expense that will occur roughly every 10 years (or whatever tube life is). So high pressure is not necessarily good.

The second part of the question deals with temperature measurement. Like everyone else, our prime indicator is the portable infrared pyrometers. However, we are experi¬menting with skin temperatures again. There have been quite a few improvements in skin temperature indication (TI’s). We put in 5 skin TI’s 3 years ago in one of our reformers and they did well. We recently retubed that furnace and went back in with 32 skin TI’s. They are type K thermocouples. They have beefy leads, housed in Hoskins 2300 sheathing. We are using GAYESCO Retractopads for the tube attachment. These pads allow you to pull the thermocouple out for replacement. All 32 TIs have been reading accurately since mid-summer.

BARKER: We have several steam-methane reformers within our company; they operate in the 250 psig to 400-psig range, not a particularly high pressure, but the guidelines for temperature measurement and control are the same.

We use a few different types of pyrometers at our various units and locations: Minolta Cyclops, Land Cyclops, and Quantum Logic. The frequency of readings ranges from once a shift to once a week. Operators manually adjust burners to balance individual cell outlet tem¬peratures to within 50°F to 750F of each other.

In one of our Selas side fired heaters, we have installed thermocouples to assist in finding more subtle problems that could result in hot tubes. We have found a direct and relatively constant correlation between the tube tempera¬ture and the sub header temperature directly below the tube. Since the sub header is in an enclosed trough out of the fire zone, thermocouples were installed without worry about them being burned off. The thermocouples are on 2 ft centers along all of the sub headers and are tied into the DCS with alarms. Each of these thermocouples and alarms allow the operators to know when a group of three tubes is deviating from the norm (from a variety of possible causes) and alerts them to conduct a detailed investigation of the process problem.

In a different Selas furnace, which has two cells, there are four outlet header temperatures per cell. The highest temperature is selected for control. The controller output is fed to a duty controller, which outputs to the fuel gas control valve for that cell. The four temperatures in each cell are manually balanced within the cell by the burner-firing pattern, specifically, the number and location of firing burners.

In our more modern, downfired reformer, the com¬bined outlet temperature at the waste heat boiler inlet is measured and this controls the flue gas temperature at the outlet of the firebox, which in turn, controls the fuel flow. There is little balancing needed with this style of furnace — all burners fire all the time.

ANDERS NIELSEN (Haldor Topsoe NS): Haldor Topsoe built two hydrogen plants with reform¬ers operating at 40 kg/cm2, 570 psig, as early as 1966. One of these plants is still in operation, producing 98% hydro¬gen. The original tube material was HK4O. With the tube materials available today such as microalloy HP, much higher reforming pressures are technically possible. Only in unusual circumstances would it be economical to use such pressures.

The tube outlet temperature was 820°C (1508 °F). We have measured tube skin temperatures in a number of different ways in order to find a useful method for easy industrial use. In our full size reforming pilot plant, we have used infrared pyrometers correlating the readings with measurements obtained from thermocouples embed¬ded into the tube material and from a PYROLASER®. We have found that reliable measurements can be obtained with an infrared pyrometer provided corrections are made to reflect radiation from the furnace walls. This may be a little more difficult in some configurations.

To balance outlet temperatures, it is essential that the catalyst loading has been performed properly, ensuring equal pressure drop in all of the tubes, which mean that in many cases you may have to empty out tubes after filling and reload them to get the pressure drop in line.

ROSS BRUNSON (United Catalysts, Inc.): On the comment on the pressures for the hydrogen plants, I think we have to also recognize the newer generation plants that have PSA for cleanup. Where the conven¬tional plants will run at a lower pressure, as has been stated around 400 pounds, with a much higher steam-to-carbon or steam-to-gas ratio, the new PSA plants run up in the 500 pound range and steam-to-carbon ratio is down to 2.5 to 3 and with outlet temperatures up to 1600 °F. There are really two generational types of plants when you look at the allowable pressures.

RAGHU MENON (BOC Group): In terms of the infrared temperature measurements, there are new methods that minimize interference by mounting blackbody type objects on the tubes. A Cali¬fornia company that provides this type of technology is E2 Technology Corporation in Ventura. Of course, for outer walls of secondary reformers, there are also fiber-optic cable-based options, which assist one to avoid hot spots by providing a complete temperature scan.

STEVE CATCHPOLE (101 Katalco): As with all chemical plants, part of the economics of both the capital and operating cost of the plant can depend very much on the chosen operating pressure. Much of the refinery hydrogen need is for a hydrogen supply at a considerably higher pressure than that normally encoun¬tered in a typical hydrogen-manufacturing unit (HMU). If the HMU is feeding directly to units operating at higher pressures, then there does exist an option to consider a higher HMU process pressure and thus decrease the size and operating cost of the hydrogen compressor. Alternatively, the HMU may feed the product into collective hydrogen main that may be at a pressure, which avoids the need for compression of the product hydrogen.

Traditionally, HMUs have pressures in the 145 psig to 290-psig range or 10 kg/cm2g to 20 kg/cm2g. The choice of this modest pressure is based on a number of compli¬cated and interactive variables. The key issue is the effect that pressure has on the methane-steam equilibrium exiting the steam reformer on the HMU. This needs to be reviewed for both types of HMU commonly encountered as follows:

  1. C02 liquid wash removal/methanation hydrogen plants at higher pressure, due to methane-steam equilibrium considerations; the hydrocarbon conversion is reduced (increased methane slip). This can result in an unsatisfactory product purity (i.e. less than 95% H2) unless the HMU significantly increases the steam-to-carbon ratio in the steam reformer. However, if the steam ratio is increased then this will reduce the plant efficiency and reduce the amount of steam export.
  2. PSA type hydrogen plants At higher pressure, the PSA system will need sizing to cope with increased methane slip from the steam reformer as mentioned in 1), but the key point is that the HMU will then be operating with an excess of fuel (recycled PSA offgas) and so will be forced to export low-grade fuel which is an inefficient use of higher value fuel, off gases and hydrocarbons used as the HMU feedstock as in 1), higher steam-to-carbon ratios will offset this problem but with the corresponding decrease in steam export and overall plant efficiency.

The obvious solution would be to increase the steam reforming exit temperature for either plant type. Historically, this idea has been limited by the metallurgy of steam reformer tube materials, leading to a trade-off between process pressure and steam ratio. Modern tube materials have challenged this limitation and now, HMUs can be constructed with a higher pressure and higher operating temperature.

In reality, many steam reformers have been built with pressures as high as 522 psig (36 kg/cm2g). These units have supplied hydrogen rich gas for manufacture of am¬monia and methanol. Both these processes have a stage of compression required for synthesis reaction and thus the front-end pressure is kept high to minimize compressor size and operating cost. In the case of ammonia, the drawback of higher methane slip is of little consequence with a secondary reforming stage converting most of the slipped methane. In methanol manufacture, the exit methane is more critical, and steam reformer exit temperatures (15440F to 15800F or 8400C to 8600C) have generally been higher than those seen in the CO2 liquid wash/methanation type of HMU. Modern HMUs are now being built with steam reformer exit temperatures as high as 17240F (9400C), although 15800F (8600C) would appear to be the more typical. Pressures have increased to over 290 psig (20 kg/cm2g) but this is not the limit. Advances in tube metal mean that the historical limit of high pressure and temperatures has been removed and that true optimization of the process conditions will take place. Thus it is highly likely that newly constructed hydrogen plants that are to be truly integrated to a high pressure hydrogen consumer will be designed with higher operating pressures, possibly to catch-up or overtake those traditionally adopted by the ammonia and methanol producers.

As HMU steam reformer operation severity increases, the detailed observation of the steam reformer tube tem¬peratures becomes essential. By far the most common technique is to use an infrared pyrometer. Modern designs help ensure that the accuracy is improving, but fundamen¬tally, these devices read high (up to +50 °C (or) + 90 °F) in most cases. This error is due to:

  1. The infrared pyrometer reading being not just the thermal emission from the steam reformer target tube but also the background radiation;
  2. The incorrect or estimated emissivity of the steam reformer target tube.

There is a variety of techniques that can be used to overcome these problems. ICI Katalco has developed a practical technique for measuring absolute tube tempera¬tures using a “Gold-Cup” device. This is then used to calibrate the infrared pyrometer. In addition, there are some simpler, but less accurate, techniques that can be used to “correct” the infrared pyrometer reading. These methods, as well as the theory of infrared pyrometry, are described in ICI’s paper presented at the 1992 AIChE Ammonia Safety Symposium “Tube Wall Temperature Measurement in Steam Reformers.

With respect to twin or multiple cell reformers, ICI Katalco would like to comment on the principles of why we should achieve similar cell outlet temperatures. The hydrocarbon conversion that occurs at one single tempera¬ture (i.e., all cells the same) is more than that achieved by two cells with the operating temperature in one cell higher than in the other, for example, even though the “average” temperature is the same. The same principle applies with efficiency (fuel use) with a single temperature having a better efficiency compared with the average. As furnace monitoring and catalyst performance assessment can also be important in the HMU daily operation, a consistency of operation from one cell to the other allows for simple comparison and diagnostics of problems.

DAVID. L. KING (Howe-Baker Engineers, Inc.): The highest reformer outlet pressure typically used for commercial plants is 410 psig, corresponding to a product pressure of 375 psig from the PSA unit. The correspond¬ing outlet temperature is 1550 °F. However, Howe-Baker has the capability to design for up to 490-psig-outlet pressure, corresponding to a product pressure of 450 psig. The outlet temperature can be as high as 1600 °F.

For tube temperature measurements, we recommend a portable optical pyrometer. We use the Minolta/Land Cyclops 152 model manufactured by Land Infrared, Bris¬tol, Pennsylvania.

Our box reformer is a single-cell design. The burners are arranged such that uniform firing can be achieved across the furnace. We believe achievement of uniform tube wall temperatures is a more accurate indication of good distribution than measurement of individual header or cell outlet temperatures.

+ QUESTION 2

Compare the performance of the latest reformer tube materials with HP modified tubes.

ARNDT: We have retubed 3 of our 6 reformers with the new micro-alloy HP material. They replaced HP modified tubes. Our first micro-alloy HP tubes were installed in 1992. The principle justification for 2 of the 3 retubes was increased tube life. We feel we are going to get 20% more life. The tube dimensions were kept the same. On the third reformer, we reduced the tube wall to get more catalyst volume and less delta P at the same tube life. The materials we used were Manaurite XM and Paralloy H39WM.

BARKER: I agree with Mr. Arndt’s comments and note that thinner tubes also allow operating at a lower tube wall temperature, which reduces fuel consumption.

KOOIMAN: The HP 40 modified centrifugally cast reformer tube is a blend of 25 chrome/35 nickel plus niobium. In general, HP 40 modified tubes have an 18% to 19% strength increase in the critical temperature region over the HK 40 tubes. Average lifetimes of 100,000 hours or 11 years are common. The newest micro-alloys are blends of 25 chrome, 35 nickel, niobium, and titanium. They are roughly twice as strong as the original HK 40 tubes. They are cost-effective, in other words not quite the cost of HK 40, and offer many options including higher heat fluxes, increased catalyst volume, fewer tubes, and longer tube life. Fabrication of micro-alloys is much more difficult, which has limited the attraction of these tubes. Most recently built plants are using HP 40 modified metallurgy, except in Europe where the HP-BST metallurgy is preferred, due to better fabrication techniques. At Koch, we use the HP 40 modified tubes.

SMITH: We have similar experience. Previous experience with the HK 40 materials in a reformer heater in the Riyadh refinery was poor. As a result, there were frequent creep rupture problems. The tube metallurgy was upgraded with the HP modified materials in 1994. To date, the perform¬ance has been excellent. The furnace has operated at conditions similar to those for HK 40 service; yet, no significant creep was observed when the tubes were checked during this year’s turnaround.

ANDERS NIELSEN (Haldor Topsoe A/S): Haldor Topsoe has well over 10 years of experience in specifying the micro-alloyed HP reformer tubes. The ex¬perience has been an unqualified success in a large number of plants, both ammonia and hydrogen units, presently in operation. Improved strength of the material has allowed installation of tubes with thinner tube walls making the tubes less susceptible to thermal shock and permitting higher heat fluxes without adversely affecting tube life expectancy. Of course, with thinner tubes more catalyst can be installed.

VINAY KHURANA (Kinetics Technology International Corp. (KTI)): The use of micro-alloys in steam reformer service has been widely accepted in the last decade. KTI has used micro-alloys in several applications with very good success. High temperature operation favors the use of micro-alloy due to its higher strength, about 15% to 25% higher stress values at same design life.

In retrofit situations in steam reformers, the upgrading to micro-alloys does benefit. It increases capacity while retaining the same number of tubes by using larger internal diameters and thinner wall thickness. I think it is fair to conclude that use of thinner tubes, as was mentioned by the panel, reduces the effect of thermocycling, which in¬creases the life of the tubes and reduces maintenance costs of the steam reformers.

DAVID. L. KING (Howe-Baker Engineers, Inc.): We have used HP modified micro-alloy tubes (Manaurite XM; Schmidt & Clemens 4852 MOD; Kubota KHR35CT) for the past 7 years. This micro-alloy material is 10% to 15% stronger than the older HP modified material (Manaurite 36X; Thermax 63; Kubota KHR35C). The micro-alloy material allows design for thinner tube walls in the reformer.

Thinner walls increase tube life since the effect of secondary stresses associated with thermal cycling (start¬ups, shutdowns, and rate changes) is much less with a thinner tube. Thinner tubes also operate at a lower tube wall temperature, thereby reducing fuel consumption.

+ QUESTION 3

What are the current NOx emission limits from new reformers and the processes used to attain them?

BARKER: In the California San Francisco Bay area, new reformers have a NOx emission limit of 9 ppmw. Flue gas NOx is reduced by selective catalytic reduction (SCR) catalyst installed in the flue gas stack. Vaporized aqueous ammonia (greater than 25 wt% NH3) is injected upstream of the SCR catalyst. The reformers in the California Los Angeles Basin are limited to 7 ppmw NOx. They also use SCR to achieve the required NOx levels.

DAVIS: CENEX started up a Foster Wheeler hydrogen unit in 1993. A guarantee on the burners was 0.05 lb/MMBtu and that is what we based our permit on.

GENTRY: We are aware of recently permitted reformers that have been equipped with selective catalyst reduction (SCR) to control NOx. Emission rates as low as 0.0112 lb NOx / MMBtu have been achieved. In existing reformers, low NOx burner technology is achieving emission rates of between 0.22 lb and 0.034 lb NOx / MMBtu.

HUNKUS: NOx limits vary by the attainment status and PSD status of each site and facility, with new installations typically requiring low NOx burners. Burners can be designed for induced recycle of flue gas to control the heat of combustion, or can use multiple fuel and/or combus¬tion air injection points to both control the excess oxygen available to form NOx, and to limit the flame temperature to retard the shift towards NOx production. This is a staged combustion design. Other technologies incorpo¬rate the addition of a chemical agent to control NOx, which is an assisted design. Some burners rely on a com¬bination of both assisted and staged technologies. These burners can often achieve NOx values as low as 0.020 lb to 0.025 lb NOx/MMBtu, sometimes even less. The typical guarantee range for any of these systems is in the 0.05 lb range.

In our experience, low NO~ burners do not typically have the operational range that “normal” burners have, and may go unstable at higher rates of firing or excessive draft. They may experience bumping where the flame seems to jump off the burner and then relight, rapid shuddering or actually go out.

So, if you are looking at pushing the original design numbers on your low NOx burners, review your system carefully as operators may not foresee this potential.

RONALD BREDEHOFT (Kinetics Technology International Corp. (KTI)): KTI has installed a number of hydrogen plants in the U.S. and each hydrogen plant has its own specific NOx requirements. In California, we have installed SCRs on these hydrogen plants and they make, as the panel sug¬gested, less than 10 ppm NOx. They meet all the Best Achievable Control Technology (BACT) guidelines in California.

Elsewhere in the U.S., the requirements are not as strict and we have used various technologies. In general, our experience with burners firing PSA gas at 2% oxygen and no air preheat is as follows:

• Conventional burners produce around 50 ppm of NOx • Staged fuel low NOx burners produce 35 ppm to 45 ppm NOx. • Ultra low NOx burners produce about 25 ppm to 35 ppm NOx. • SCRs produce less than 10 ppm NOx.

DAVID L. KING (Howe-Baker Engineers, Inc.): NOx limits vary with the sire. We typically use low NOx burners. These burners are designed for induced recycle of flue gas. For hydrogen plants burning PSA offgas, these burners can achieve NOx values as low as 0.03 lb NOx/MM Btu LHV heat release. For such burners, the typical guarantee is 0.05 lb NOx/MM Btu LHV.

+ QUESTION 4

We have two furnaces with air preheaters equipped with glass tubes. Although we follow the manufacturer’s recommended washing procedure, we repeat¬edly experience breakage of these tubes. What have other refiners’ experiences been with these types of air preheaters?

JACKSON: Yes, we have experience with air preheaters with the glass tubes, and we too have suffered from the tubes cracking. We have put the problem down to one or more of a number of issues:

• washing the tubes too hot; • Vibration of the tubes in operation; • Breakage at the point of contact of the supports; • Hardening of the PTFE tubesheets; • Aging and cycling of the glass; and • Poorly designed tubes that break through thermal expansion.

When the glass cracks, the inlet air bypasses the tubes and exits via the stack. We would typically install sampling points on the flue gas side upstream and downstream of the preheater. Once a month, the flue gas would be sampled and the difference in oxygen content checked. If the tubes are broken, then there will be more oxygen in the downstream flue gas section.

One company has introduced a glass lined metal tube and it claims that the supporting of the tubes and the tube sheet system give a better and more reliable design. BP has no experience with these designs to date.

BP is also looking into Open Channel Air preheaters. These are plate type constructions and are being consid¬ered for service in one of our refineries.

ARNDT: One of our ex-Chevron refineries uses glass tubes in a furnace air preheater. By and large, they have performed well. Some breakage has occurred on-line. It is believed that the cause is due to sucking in cold rainwater into their fan and the water runs down the ducting shocking the glass tubes. The tube failures have been mainly on the outer ring of tubes. Another problem is the gradual hardening of the support material, reducing the flexibility of the tubes to withstand stresses.

+ QUESTION 5

We have experienced persistent mechanical failure of our soot blowers in our vacuum furnace. The components, which normally fail, are the limit switches and line steam admission valve, which fail during soot blowing operation. What is other refiners’ experience with soot blowers?

DAVIS: A survey of the CENEX maintenance staff finds the limit switches to be the most frequent failure. They do feel, however, that the 2-year life expectancy they are experienc¬ing is satisfactory.

LEMMON: One Tosco refinery has experienced persistent me¬chanical failure of soot blowers in several furnaces. The problem is related to insufficient air pressure to the air motors that drive the lances in and out of the furnace. They are going to upgrade the control system from pneu¬matic to electrical to get more consistent air pressure to the drive motors.

SMITH: For commiseration, we have additional problems to offer. We have packing leaks on the lance. We have break¬ing chains on sprockets, drives, and carriage in one of our refineries. We have erosion of the extended tube services (studs) and surrounding tube damage, and damage to the convection section refractory near the soot blower. Of course, the intrinsic design and the severe operating con¬ditions make these problems likely. The only control parameter we have found to help ensure a better operation is to make sure that your steam is as dry as possible. Wet steam will cause erosion problems in the lance as well as in the nearby furnace components.

RONALD E. MARRELLI (Phillips Petroleum): This question does not pertain to the previous two process heaters you have talked about, but it pertains to heaters in general and it is a development that has come up recently. Does anyone in the audience or on the panel have any experience with emissivity coatings inside fireboxes to improve efficiency and reduce fuel costs on process heaters?

JACKSON: We considered this at one of our refineries, but instead we focused our attention on better, more efficient burners. We are not convinced that it would pay back for all of the effort it requires.

CHARLES LeROY (CITGO Petroleum Corporation): We have a refinery down in Corpus Christi that is getting ready install some electromagnetic devices on the piping right near the burner tips. The firm supplying this claim that it can increase burner efficiencies by 10%. I have not heard of this before, and am curious if anybody has any knowledge of successful applications. I have no idea how it works. It seems a little bit like snake oil to me. I was just curious if anybody in the audience or the panel had heard of this before or could comment on this.

JACKSON: First and foremost, I have heard of it, but I do not have the derails. It does do something and it has something to do with the effect on the molecules. I do not know the specific details.

+ QUESTION 6

How do you maximize heater run length while maintaining efficiency? Do coker operators tend to run their heaters at higher than normal O2 and less preheat, thereby sacrificing energy efficiency for potential run length?

BILLS: The short answer is yes; some efficiency has to be sacrificed to maximize heater life. I think this is particularly true for older heaters. Other factors are whether the heater is at maximum or minimum firing, the design of the heater, and whether or not you are now making shot coke; e.g. operating at higher heater outlet temperatures than original design. Generally speaking, most methods used to increase time between decokes (or spalls) will reduce the fuel efficiency of the heater. Higher excess O2 in the heater box helps to maintain good circulation around the backside of the heater tube, preventing coking of the tubes by radiant heat from the wall but uses more fuel to heat the air. Maintaining correct tube velocities with BFW or steam injection helps to mitigate coke laydown, but also consumes fuel energy. Less preheat will not gain much; the charge heater does not coke up (or should not). I presume that the desire to extend time between decokes is driven by economics. Usually the value of throughput is greater than the cost of fuel; therefore, it is advantageous to sacrifice some energy efficiency to buy additional run length. A little planning and analysis will enable the engineer to strike the proper balance.

KRISHNA: We run our heaters for normal oxygen levels, as limited by our burners’ design and operation, and have not found it effective to run at higher oxygen, sacrificing efficiency for run length.

STEFANI: Our experience has been that there does not appear to be a strong correlation between heater run length and heater efficiency. A well designed heater will give long run lengths regardless of efficiency, as long as the critical design parameters such as velocity, residence time, pressure drop, film temperature etc. are appropriately set. Conoco does not run their heaters at a high O2 to save fuel.

REZA SADEGHBEIGI (RMS Engineering): I was going to poll the panel and maybe the audience to see how many of you process coker naphtha into the FCC riser, and when you do, have you seen any problem with the alky C3 product being corrosive (i.e., failing the copper strip test)?

DARDEN: We have run coker naphtha into our fluid unit. We did not see any issues with the propylene stream off the FCCU, but we did see about a 400 ppm increase in sulfur off of the fluid gasoline, especially in the heavy cat naphtha.

REZA SADEGHBEIGI (RMS Engineering): How about anything on a propane product off the alky failing the corrosion test?

DARDEN: No. We saw no issues with the alky.

KRISHNA: We have had coker naphtha to our FCC’s, and I do not recall that particular issue with the C3 product.

REZA SADEGHBEIGI (RMS Engineering): We designed a system, to put the coker naphtha in the middle of a riser for a client, because they had a problem with fouling the reformer preheat exchangers. So when they commissioned it, they told us that they have to replace the KOH treaters several times, and they do not understand why, what caused the propane. They sell the propane to the utilities, and it becomes corrosive. 1999 NPRA Q &A Session on Refining and Petrochemical Technology 103 Is it silica that is causing that to be corrosive, or what is really happening?

TARIQ MALIK (CITGO Refinery & Chemical): As a follow-up to this question of running the heaters at higher than normal oxygen, I would like to poll the panel to see what kind of heater run lengths they are getting between two successive decokes?

BILLS: We run about three to six months between spalls, and we do an actual decoke after about three or four spalls.

DALY: Six to nine months and with this new heater we are hoping for longer.

DARDEN: We usually have enough bobbles along the way that get us to decoke or online spall a tube that we really cannot give you a good answer on that.

KRISHNA: We operate on about 12 month cycles with on-line spalling two to three times during that cycle.

PROOPS: Three to six months, recently higher as we experiment with pigging and steam spalling.

STEFANI: We recommend to our clients that predicting how often you on-line spall is directly dependent upon the feed. Some feeds require more frequent on-line spalls, because they have a greater propensity to coking the heater than others. We recommend that the heater be spalled as often as possible to avoid a thick layer of coke buildup on the tubes. With most crudes a one year run between steam air decokes is a reasonable expectation.

+ QUESTION 7

What do you do during decoking to save on cycle time?

DARDEN: Between the two cokers at LCR, we have a couple of items that are common between the two for saving on decoking time. The first of those is making sure the drilling equipment is reliable. If we have to slow down rates because of a problem with a cable, with a problem with a drill hose, leaking fitting, that seems to be where we lose a lot of time. We have also been looking into the blowdown systems, which have been limiting on both of the units, and we have actually paralleled the Blowdown line on the new system, so that we can increase the quench water rates. I think individually, it comes down to looking at each unit to see where you have available time to trim. But you have to remember to do it safely, that is the only thing I would caution.

KRISHNA: I echo Mr. Darden’s sentiments. It is a case by case situation looking at each individual coker to see what all the individual steps and where there can be savings in each part of the step. For us right now, we are focusing on the steam strip and the drum preheat parts of the cycle.

PROOPS: I would agree with those comments. We have steamed for as little as twenty minutes when we are running eleven hour cycles, and we have reduced warm-up to as low as an hour and a half. Obviously, we try to push cooling to pressure limits, and we use combination cutting bits. We also restrict vapor flow to the fractionator to force vapor back to the warming drum and speed preheat.

STEFANI: Conoco, at their Lake Charles and Ponca City refineries have drums that operate on ten hour cycles. Most of the cycle time reduction is accomplished through shortening the heat-up cycle. For larger drums a rapid heat-up can be detrimental to the skirt weld attachment. To compensate for this potential problem, the Conoco/Bechtel design incorporates a special design for the shell skirt attachment that makes allowances for a shortened heat-up cycle time. In addition, as some of the panelists have mentioned, a properly sized jet pump, the use of combination drill bits, and automatic unheading are beneficial to maintain the decoking cycle time. We also recommend that procedures be established to avoid and/or handle cave-ins and blowouts. This will not save you time directly, but will avoid having additional time tacked onto your decoking cycle to remedy any potential upset condition.

JAMES JONES (Turner Mason Company): This is for Mr. Stefani. I guess there has been a lot of reduction in cycle times over the last four or five years from what we see in Solomon surveys and what we have heard from the panel. What is the lowest cycle time that you are actually designing units for?

STEFANI: The choice of the initial cycle time is, in general, reflective of the client’s comfort level. For an initial design, we see 16-18 hour cycles as being a standard, with perhaps 15 hours at the extreme. Further reductions in cycle times usually are a matter of necessity rather than premeditated design and are employed when a capacity increase or processing a heavier feed is required. The Conoco-Bechtel standard coke drum design does allow for reduction of the cycle time down to ten hours with no impact on drum life.

ROY KESSLER (Fluor Daniel): We have a delayed coker that came on stream this year that can be unheaded with the drum still full of quenchwater. The way it works is the unheading area is structurally enclosed on three sides, and it uses a slide deck in combination with our automated unheading system. The operation has been very successful and can eliminate more than an hour from the cycle time by eliminating the drain step. In addition, we have our new generation of boltless unheading devices that can reduce the unhead and rehead times down to five to ten minutes from the thirty minutes or so that has been typical for a bolted system.

TARIQ MALIK (CITGO Refinery & Chemical): Since 1997, we have consistently operated on eleven hour cycles, and have demonstrated ten to ten and a half hour cycles as well. We run out of feed, so we cannot maintain short cycles. Ours is a containment system that we call a shot coke and water containment system—it is operated remotely. We were able to cut the drain time to about fifteen to thirty minutes and we normally complete our drain through the bottom head. We are protected from bottom blowouts and cave-ins with this system. Everyone on the switch deck is well protected and shielded from any hazardous operation. We have a two-drum Coker, and can run 43,000-44,000 bbl through it. This system of ours is protected by US patents and is licensed through Hahn & Clay, and I understand six units have been licensed thus far.

+ QUESTION 8

We currently on-line spall our coker heaters between heater decoking. Has anyone done this on their crude or vacuum heaters? How successful were they and what procedures were used? Specifically, were any special precautions used to prevent plugging of bottoms piping, pumps and towers?

STEFANI: Most refiners decoke their atmospheric and vacuum heaters during turnarounds. If decoking is required more often due to excessive heater firing to achieve greater capacity or when processing heavier feedstocks, on-stream spalling can be considered. However, the ability to on-line spall a heater requires four conditions. Namely, the unit must be capable of operating continuously at about 50% capacity while making specification products; a means to continuously remove the spalled coke is required; the heater must have at least two cells and both radiant and convection sections must be capable of reduced firing on one cell while the second cell is at maximum firing and the heater must have a horizontal, downflow configuration. Of all of the requirements necessary for on-stream spalling, removal of the spalled coke from the unit is the biggest challenge. Options are to pipe up a heater cell to the decoking system during operation or to add provisions for coke removal to the atmospheric and vacuum towers. If on-stream spalling is to be employed, our recommendation is to add provisions to allow for a safe redirection of the heater effluent from the corresponding fractionation tower to the heater decoking system.

BILLS: I concur with Mr. Stefani. There are no Equilon Enterprises or Motiva refineries that perform on spalls of vacuum heaters.

ELLIOTT: We, too, have not observed on-line spalling of crude or vacuum heaters, and Mr. Stefani covered all of the concerns that we have about this practice.

+ QUESTION 9

How do you maximize heater run length while maintaining efficiency? Do coker operators tend to run their heaters at higher than normal O2 and less preheat, thereby sacrificing energy efficiency for potential run length? BILLS: The short answer is yes; some efficiency has to be sacrificed to maximize heater life. I think this is particularly true for older heaters. Other factors are whether the heater is at maximum or minimum firing, the design of the heater, and whether or not you are now making shot coke; e.g. operating at higher heater outlet temperatures than original design. Generally speaking, most methods used to increase time between decokes (or spalls) will reduce the fuel efficiency of the heater. Higher excess O2 in the heater box helps to maintain good circulation around the backside of the heater tube, preventing coking of the tubes by radiant heat from the wall but uses more fuel to heat the air. Maintaining correct tube velocities with BFW or steam injection helps to mitigate coke laydown, but also consumes fuel energy. Less preheat will not gain much; the charge heater does not coke up (or should not). I presume that the desire to extend time between decokes is driven by economics. Usually the value of throughput is greater than the cost of fuel; therefore, it is advantageous to sacrifice some energy efficiency to buy additional run length. A little planning and analysis will enable the engineer to strike the proper balance.

KRISHNA: We run our heaters for normal oxygen levels, as limited by our burners’ design and operation, and have not found it effective to run at higher oxygen, sacrificing efficiency for run length.

STEFANI: Our experience has been that there does not appear to be a strong correlation between heater run length and heater efficiency. A well designed heater will give long run lengths regardless of efficiency, as long as the critical design parameters such as velocity, residence time, pressure drop, film temperature etc. are appropriately set. Conoco does not run their heaters at a high O2 to save fuel.

REZA SADEGHBEIGI (RMS Engineering): I was going to poll the panel and maybe the audience to see how many of you process coker naphtha into the FCC riser, and when you do, have you seen any problem with the alky C3 product being corrosive (i.e., failing the copper strip test)?

DARDEN: We have run coker naphtha into our fluid unit. We did not see any issues with the propylene stream off the FCCU, but we did see about a 400 ppm increase in sulfur off of the fluid gasoline, especially in the heavy cat naphtha.

REZA SADEGHBEIGI (RMS Engineering): How about anything on a propane product off the alky failing the corrosion test?

DARDEN: No. We saw no issues with the alky.

KRISHNA: We have had coker naphtha to our FCC’s, and I do not recall that particular issue with the C3 product.

REZA SADEGHBEIGI (RMS Engineering): We designed a system, to put the coker naphtha in the middle of a riser for a client, because they had a problem with fouling the reformer preheat exchangers. So when they commissioned it, they told us that they have to replace the KOH treaters several times, and they do not understand why, what caused the propane. They sell the propane to the utilities, and it becomes corrosive. Is it silica that is causing that to be corrosive, or what is really happening?

TARIQ MALIK (CITGO Refinery & Chemical): As a follow-up to this question of running the heaters at higher than normal oxygen, I would like to poll the panel to see what kind of heater run lengths they are getting between two successive decokes?

BILLS: We run about three to six months between spalls, and we do an actual decoke after about three or four spalls.

DALY: Six to nine months and with this new heater we are hoping for longer.

DARDEN: We usually have enough bobbles along the way that get us to decoke or online spall a tube that we really cannot give you a good answer on that.

KRISHNA: We operate on about 12 month cycles with on-line spalling two to three times during that cycle.

PROOPS: Three to six months, recently higher as we experiment with pigging and steam spalling.

STEFANI: We recommend to our clients that predicting how often you on-line spall is directly dependent upon the feed. Some feeds require more frequent on-line spalls, because they have a greater propensity to coking the heater than others. We recommend that the heater be spalled as often as possible to avoid a thick layer of coke buildup on the tubes. With most crudes a one year run between steam air decokes is a reasonable expectation.

+ QUESTION 10

What strategies, both short and long term, are refiners formulating to meet the proposed ultra-low sulfur specifications for gasoline? If a banking and trading system were established today, whereby credits could be earned for sulfur reduction, what methods are likely to be employed?

VERRENKAMP: I will answer this in terms of BP’s generic overall strategy and other panelists may talk about the specifics required to achieve those things, specific processes. Our short term strategy, in addition to meeting the new gasoline and diesel specifications, is to make clean fuels available within the next 18 months in a network of 40 of the world’s cities troubled by pollution and smog. This was alluded to in the opening address to the conference. As an example, we now only sell fuels in Great Britain that are lead free and low sulfur and are in the process of converting service stations in Paris and Istanbul. A further immediate strategy of the company is the launching of the world’s first CO2 emission trading system to point the way to cost effective reductions. In the medium term, we are proposing to phase our investments ahead of the regulatory dates, as I talked about in the last question, and avoid peak demand periods on contractor and workshop capacity. We also established teams with the express objective of investigating and proposing advantages and novel technology levers to try and minimize clean fuel capital requirements. Areas we are looking at, and we are not alone in oil companies doing this are: low cost hydroprocessing, naphtha desulphurization, oxygenating diesel, the injection of novel diesel fuel additives, low cost benzene removal, novel clean gasoline. Long terms, there appears to be quite a few varied options. There is a lot of excitement and hype but the jury is still out on whether fuel cells are going to be the key technology vehicle of the future. They have already been around for 150 years. An alternative view is that developments in existing engine technology such as gasoline direct injection engines and common-rail diesel coupled with sophisticated catalyst post combustion treatments and ultra low sulfur fuels will remain the dominant technology at least as far as 2015. Fuel synthesized entirely from gas rather than derived from the products of oil refinery are a possibility. We have actually signed a cooperative agreement with General Motors to investigate these alternatives. Finally the rules of the U.S. Sulfur Banking and Trading System will be set by the EPA. A draft is already available.

DAVIDSON: This is really a two-part question. What is the gasoline sulfur spec going to be? And then what impact credits might have on this. The gasoline sulfur spec is going to be 30 ppm, but the real question is how long that is going to stay in effect and what the next step is going to be. Engine manufacturers are really pushing for zero, but that is probably unattainable. So the future target is probably some compromise around 5 to 10 ppm. Not too many people are really anxious to spend a whole bunch of money to get to 30 when 5 or 10 is the next step down the road. So most people, I understand, are asking for technologies that will allow them to do both. At the very low sulfur levels we are talking about, refiners need to look at the impact that desulphurizing the FCC gasoline will have on other pool properties. In particular, I am thinking about octane. Since most of the FCC gasoline is going to need severe hydroprocessing in some form to achieve the required level of desulphurization, octane loss arising from olefins saturation will become significant. It will be even more difficult to make up this octane if the aromatics content of the gasoline is limited and the ability to blend high octane components, such as MTBE, is restricted. Consequently, the use of technologies, which have the capability to maintain octane, or minimize the octane loss, will be very attractive. Turning to the issue of credits, if credits for implementing low sulfur gasoline ahead of the legislated date were available, this may help refiners offset the cost of implementing gasoline desulfurization technologies. Encouragement to produce low sulfur gasoline early will also help offset the potential problems with procurement and construction we were discussing in the previous question. The EPA has been designing banking and trading schemes for different pollutants since the early eighties. The popularity of these programs has been increasing ever since for two reasons. First, it is relatively cheap for the regulatory authority and second, these programs minimize the cost of meeting an ambient standard across an industry. At least in economic theory, there is no cheaper way for the industry, as a whole, to reach the given level of pollution control. This is the key reason for implementing this type of program. How does it work? There are a few possibilities depending on the physical attributes of the pollution you are trying to control and how the government handles the initial endowment of the sulfur credits. First let me mention a bit more on why a trading scheme achieves the least cost of cleaning up. Most pollution laws force the worst polluters to clean up first. A trading scheme is the exact opposite. The companies that can clean up their emissions for the cheapest are the ones that have the incentive to do so under a trading scheme, while the companies that would be the hardest and most expensive to clean up are the last to have the incentive to upgrade their refineries. Does this make sense? Absolutely, the companies that have newer equipment can sell their permits to the companies with older equipment. The value of each permit or credit will eventually be worth exactly how much it would cost the worst polluter in the industry to upgrade its equipment. All the policy makers are concerned about is having an ambient standard met, it makes no difference to them as long as at the end of the day we do not have more than X amount of pollution in the air. So once the level of acceptable amount of pollution is determined, say 100,000 tons of SO2, you can simply divide that into 100,000 permits each allowing the owner to produce one ton of SO2.

Possibility #1: They give each stakeholder the same amount of sulfur credits. In this case the total amount of sulfur allowable is divided between existing stakeholders. This can be allotted based on historical share of the market or simply divide the amount of permits by the number of stakeholders. If a new company wishes to enter the industry after the program has been instituted they must first obtain enough credits for what they will produce. Existing companies can finance new high tech, low sulfur equipment by selling credits. After the initial endowment, companies who have a newer more friendly process will have an excess of credits. Older technology firms will then be faced with the choice of either upgrading and not buying the reduction credits, or keeping the existing old technology and paying other companies for enough credits to operate. In this case, the cost of the credits should eventually be the same as the cost of upgrading (or at least very close). In the end, this approach punishes older technology firms and rewards new technology.

Possibility #2: They require each stakeholder to meet a minimum required level, and if you can reduce even further, you get rewarded with credits. This operates much the same way but the value of the credits is unclear. In the previous example we know what the economic choice is and therefore that the price should eventually reflect that of upgrading. In this scheme it is unclear what further reduction is worth, i.e., who is it sold to? Why would you buy it, unless they allow firms unable to meet the minimum required level to purchase these credits from companies that go a bit further? How do they know how long they will be allowed to operate above the level? Also, new companies are allowed to start up as long as they meet the minimum standard, this avoids problems from the previous scheme of 1 or 2 companies trying to corner the market. So, ultimately this type of program is designed to force the following incentive. Those that can upgrade for the cheapest are the first ones to do so. This ensures that the standard will be met at least cost to the industry as a whole, however, those companies for whom it will be very expensive to upgrade, will first purchase other companies’ credits until the value of those credits is fully capitalized. There are some other critical elements in order for this type of program to work. There has to be enough participants that the market is not too thin for the price of the credit to be fully capitalized. Any sort of little distortion in the market affects the price of the credit to other than just the cost of upgrading and will distort the price of credits and hence the efficiency of the entire pro-gram. In the end, the policy rewards investment in new environmentally friendly equipment and punishes older technology. So the short term strategy is to just purchase enough credits to cover what you need from companies that have newer equipment, but eventually your long term strategy has to be to upgrade the equipment in order to avoid paying the high cost of credits to your competitors. So alluding to Mr. Hawthorne’s remarks, which are probably true, this says that some of us are not going to be around here in the next decade. These regulations are just going to force some operators out of business.

GATTE: This question has been discussed frequently during the past several NPRA Q&A Sessions and Annual Meetings. There are several options available, as everyone knows, but all of them tend to revolve around some changes to the FCC process since the FCC typically contributes about 90% of the sulfur to the gasoline pool. Therefore, the available sulfur-reducing options can best be broken down into three categories: pre-FCC, in-situ and post-FCC. Of course, the proper strategy, both long and short term, is going to depend on the specific refinery configuration. Grace Davison recently conducted a survey of refineries in North America asking about their probable short term and long term strategies. This will be discussed during the Heavy Oil section. Also, the post-FCC options will be discussed in the Hydrogen Processing section. Pre-FCC, there are two basic options. First, a refinery can choose to switch their crude diet, purchasing sweeter crude, which reduces the resulting product sulfur levels. Use of sweeter crude also typically reduces the hydrogen demand on the various Hydrotreating operations that run to a sulfur specification. We know of a few refineries that have chosen this path as part of their strategy. A second pre-FCC option is increased FCC feed hydrotreating, and many refineries are already using, installing or planning hydrotreating facilities to treat FCC feed. This option is attractive because there are additional benefits to the FCC operation that result from feed hydroprocessing, which we’ll discuss in detail later. In-situ options are very attractive short-term solutions, because there usually no up-front capital costs. There are several FCC operational changes that can be used to help reduce gasoline sulfur. These include reducing riser outlet temperature, dropping naphtha endpoint, improving fractionator efficiency, increasing the amount of naphtha recycle and changing the catalyst properties. Changing the FCC catalyst technology to help reduce the amount of feed sulfur that finds its way into the FCC gasoline stream is a very viable alternative. Of course, this is the one that the catalyst suppliers would advocate. We will discuss the details of the types of catalyst technologies that are available and their impact on the product sulfur distribution during the Heavy Oil Session. In short, reductions in naphtha sulfur content of up to 25% have been observed commercially in FCC operations by using new catalyst technology specifically designed to cut gasoline sulfur levels. Grace Davison’s product family designed for this purpose is called GSR. This technology is proven, available today, and could be used immediately as part of a banking and trading system to obtain credits toward future product sales. As I stated earlier, the advantages of the FCC catalyst approach are the lack of any up-front capital costs, and the ability to quickly make changes to the FCC catalyst inventory to see immediate impacts of gasoline pool sulfur. We have many customers who are currently looking at this option in the short-term.

STYNES: Here are the short term strategies for building credits that we are looking at: maximize available feed hydrotreating, reduce the crude sulfur that goes to the FCCU’s unhydrotreated, investigate FCC additives, and reduce cat gasoline 90% point. Long term, we are looking strictly at post treatment options. Of course, we like the new Phillips S-sorb technology. Where we have unused hydro-treating capacity, we are also looking at splitting out the heavy portion of the cat gasoline, then hydrotreating it. We did an informal survey of four other refiners that are not represented on the panel. Here are their concerns: -Everyone’s number one concern, and ours too, is loss of octane. -Concern about installing a solution that is flexible enough to handle higher sulfur crudes in the future. -Concern about not being able to meet a future, even lower, sulfur level in gasoline.

+ QUESTION 11

Have refiners experienced fouling of crude preheat exchangers when processing low sulfur, waxy crudes mixed with condensates or other non-similar crudes? Has anyone identified a fouling mechanism and a means of mitigating this fouling?

VERRENKAMP: Our Asia Pacific refiners used to routinely process a diet of low sulfur waxy residue mixed with condensates when the condensates straight run/margins were greater 3 to 5 years ago. All the sites increased experienced fouling in the crude preheat exchangers. We developed a plot back then using data sourced from all the sites to show the typical fouling rates to expect when processing this type of cocktail as a function of crude temperature, as shown in Figure I-2. These fouling factors were calculated assuming a 6 month exchanger cleaning cycle. As you can see, there are two very definite regions of high fouling, one from about 190 to 250°F, the other from 320 to 390°F. The fouling mechanism is essentially ashphaltene deposition. Unfortunately, we were not able to identify how to dramatically retard the fouling rate, although we did have some success with dispersants. The graph was useful in assisting us to better plan exchanger cleaning programs. Note the extremely high fouling factor values around 0.07 on a six month cycle. This meant that exchangers operating in the temperature regions of asphaltene deposition were cleaned perhaps every three months to maintain adequate preheat to the furnace. Other exchangers were not cleaned for up to four years essentially during unit turnarounds.

DAVIDSON: Low sulfur waxy crudes are primarily paraffinic in natures. Nonsimilar crudes such as condensates or asphaltic type crudes are more aromatic in nature. The asphaltenes in these crudes are stabilized in a micelle by resins and aromatics. There is an equilibrium between the oil medium and the outer part, i.e., stabilizing portion, of the asphaltene micelle. When there is a large swing in the paraffin to aromatic ratio in the medium, it can upset the stability of the asphaltene micelle, especially when another destabilization force is present, as is temperature, in the hot preheat train. Mitigation of this fouling mechanism can be addressed through appropriate blending and the use of additives.

ELLIOTT: We have had recent experience with a refiner that is processing low sulfur waxy crude and an asphaltic, Middle Eastern crude. This refiner did not expect to have any problems, since he was processing these crudes in block operation. However, the refiner began to experience severe fouling of the two atmospheric resid versus crude exchangers. Fouling developed on the resid side of the hottest exchanger and then progressed to the cooler exchanger. The exchangers needed to be cleaned within 90 days of start of run. As Mr. Verrenkamp mentioned, we suspected asphaltene deposition as the mechanism. We did in fact collect samples of the crude, took them to the laboratory and performed an asphaltene stability test. The individual crudes were tested and both were found to exhibit a moderate tendency for asphaltenes to precipitate with the addition of a nonpolar solvent. However, when the crudes were blended asphaltenes precipitated much more readily. Therefore, we concluded that asphaltene precipitation was the primary mechanism of fouling and proposed application of an asphaltene dispersant. Initially a moderate dosage of asphaltene dispersant was applied during both crude runs. Results were encouraging, so the dispersant dosage was optimized. The refiner observed optimum performance with a low dosage on the asphaltic run, moderate dosage during the waxy run and a high dosage during changeover and slop processing. The chemical provides benefit during the waxy run by dispersing inorganic elements in the crude and preventing agglomeration with high molecular weight paraffins. After application of this program, the exchangers ran for 11 months without cleaning.

PARK: We handle more than 50 kinds of crude; therefore, we have also same experience when processing low sulfur.crude with a condensate. Based on our experience, after and during crude blending, the solvency effect of condensate can cause loss instability of crude tank bottom sludge, which contains inorganic salts, asphaltene and mud, etc. The sludge dissolved in the crude mix and crude carried the sludge into the crude unit and then eventually the sludge can be precipitating the desalter and the crude preheat exchanger circuit. The other hand, in our plant, the yield variation of high sulfur crude mix is relatively small in order to supply the feed stock to the upgrading facilities like RFCC plant, but that of low sulfur crude mix is very high. Therefore, the fouling tendency of low sulfur crude is more severe than that of high sulfur crude.

STYNES: We also had similar experience to Mr. Verrenkamp’s, but not in the 200°F region. The problem we had was in a sweet crude unit where we experienced fouling when we got above 350°F in the desalted crude side of the heat exchanger. For us, the non-similar crude was Forcados and was usually in a blend with Brent, Quai Iboe, and Cusiana. We injected a dispersant chemical and got rid of the problem.

S. P. KHANDARE (Bharat Petroleum Corporation Limited): We have a refinery in Bombay where we are processing nonsimilar crudes. The crude was then changed to low sulfur, high waxy crude of paraffinic nature. The percent-age of the sweet crude was increased gradually. There was no experience of fouling. However, when the percentage was increased to more than about 60 to 70%, we observed that the heat exchangers, heavy boilers in primary crude and vacuum residue started showing serious fouling. The problem was affecting the production. A study has been done to determine the principal reason. We started track-ing back from the crude into the atmospheric bottom collecting various samples. We sought help from the national laboratories in India and carried out studies from the raw crude, the crude from desalting, crude during transportation with additives, crude in storage and atmospheric residue during the process. We learned that whenever there were exchangers having a particular type of tube oil temperature, the higher molecular additives, which were there were depositing out in the exchangers leading to the problem. The solution was to go to the block operation of processing sweet crude followed by the sour crude. With the help of various additive manufacturers and the national laboratory we developed a chemical to remedy this problem. Presently this problem is not being faced. In a nutshell, the additives, which were being used to reduce the pour point of the waxy crude was precipitating out at a certain tube.

DARIN RICE (ARCO Products): For the panel members that answered, are there any observed threshold, as far as the mixture of low sulfur, waxy crudes with condensates or other non-similar crudes, that when crossed would initiate the phenomenon of asphaltene precipitation?

STYNES: I do not know.

DARIN RICE (ARCO Products): I am curious about the order of magnitude (10/90 mix, 50/50 mix, 30/70 mix?) for the ratio of sweet low sulfur waxy crude to condensate crude that may initiate asphaltene precipitation. When blending different crudes, will the degree of asphaltene precipitation change depend-ing on the order that the crude are blended?

STYNES: Forcados was the waxy crude and runs about 70% of the mix when we have precipitation problems. A warning: we would not consider percent waxy crude to be sufficient alone to predict fouling potential.

+ QUESTION 12

What is the level of heat integration across units in a refinery (such as integrating power plant boiler feed water with hot hydrocarbon products? What safety pre-cautions were adopted to avoid oil ingress to process boilers)?

VERRENKAMP: Heat integration between crude and vacuum units is quite common in a lot of our refineries, and occasionally, the FCC transfers excess heat directly to other units. A lot of sites transfer heat using the steam system, steam generated on one unit and condensed on another. Regarding transferring heat between hydrocarbon and utility systems, we have one site in particular that does this quite often. For example, boiler feedwater is heated by FCC Fractionator bottoms and HCCO products. The design requirement here is that we guarantee that the boiler feedwater side is always maintained at a higher pressure so in the event of a tube leak, the water flows into the oil, making it impossible to contaminate the utility system with hydrocarbon. This, of course, necessi-tates protecting the exchanger against overpressure from water vaporization and leads to a larger size relief valve or rupture disk. There are no issues with water in the fuel oil product regarding tank foamovers, etc., provided the tank temperature is maintained at less than 100°C, so we feel quite comfortable in doing this process. The same site also has demineralized water exchanging heat against shifted syngas exiting the hydrogen unit, except here the syngas is the higher operating pressure. The tube leak scenario in this case is protected by the installation of a tube rupture relief valve on both the inlet and outlet water side of the exchanger and the downstream deaerator has a CO detector to warm of the presence of a syngas leak.

JACOB: Integration of BFW with process units can be accomplished if the BFW is softened before the heat is added to the system. (It is important to reduce the conductivity to prevent fouling on the process heat exchanger tubes.) The system we have developed takes BFW after softening and before deaeration to increase the feed temperature to the deaerators. A careful review of the metallurgy is required due to the reverse solubility of oxygen in water. As heat is added to softened water, oxygen will be liberated out of solution and the corrosion potential increases substantially. This problem was solved in our application by increasing the metallurgy of the BFW return lines from CS to 304SS. Additionally, a temperature controlled by-pass was installed to prevent the heated BFW from vaporizing in the process unit heat exchanger. The process unit that supplies the heat for the BFW was designed to provide enough cooling with trim cooling water exchangers for emergency situations when BFW may not be available or a tube leak were to occur. No tube leaks have been experienced in the process described above.

PARK: We are integrating the crude unit and secondary units. In the original design concept there was no integration, but in the actual operation we are adopting integration, such as hot feed supplying to secondary unit with bypassing the water cooler. To do this, we are taking special precaution in terms of safety operation because there is some distinction from actual design condition. Also we have to consider a few points such as thermal expansion and reinforcing the supporting device, and have to consider the fouling program over the cooler that may result from each mode of resuming service or bypassing these cooler depending on the situation of related process. Regarding condensate recovery, the condensate recovered from the process is being recycled as the boiler feedwater of the power plant. To avoid oil contamination of this condensate, we are adapting some safety devices, such as oil and water analyzer and a shut off valve system for condensate dump out. If any oil exists in the condensate, then the shut off valve operates automatically and the condensate dumps out to the wastewaters facility.

PROOPS: Our Pine Bend refinery routes hot feeds between process units with minimal intermediate tankage. This results in large quantities of byproduct steam generation in process units. As a result, our boiler house steam generation is typically 10% to 12% of summertime steam requirements. Heat integration of steam or boiler feedwater versus process is high in our Hydrogen Plants, Sulfur Recovery Units, FCC, Cokers and Hydrotreaters (excluding the steam production in fired heaters). We do not take any extraordinary safety precautions outside of industry standard practices. We have had occasional instances of hydrocarbon contamination of conden-sate.

STYNES: We have a pretty good size solvents plant at one of our refineries, and we do a lot of heat integration where we take the hot overheads off of one tower and use them to reboil other towers. As far as boiler feedwater goes, we have numerous instances where we use process streams to heat up boiler feedwater. And in most cases, we try to keep the boiler feedwater pressure a lot higher than the hydrocarbon, so if it is going to leak, it leaks into the hydrocarbon system. In some cases, you do not want water in the hydrocarbon either. For example, we cool top pumparound from the FCC main fractionator to heat boiler water, using a glycol stream as a heat transfer medium.

GERRY HERZOG (Clark Refining & Manufacturing): Regarding the heat integration of boiler feedwater with refinery hydrocarbon streams, there are a couple more issues to consider. I heard the oxygen issue mentioned in that cold boiler feedwater contains a substantial amount of oxygen, which will come out of solution as the water temperature increases. The combination of the oxygen and higher temperature create a corrosive environment for carbon steel. The temperature of boiler feed water can easily exceed 180°F when it is being exchanged with hot hydrocarbon streams, so attention needs to be given to selection of the metallurgy used in the exchanger and outlet piping on the water side. Another issue involves the process used to treat boiler feedwater. Boiler feedwater is either de-mineralized (positive and negative ions removed) or sodium zeolite softened (positive sodium ions replace calcium and magnesium ions). If zeolite softening is used, a substantial amount of soluble chloride salts will remain in the water. If stainless steel is used, then the chlorides could be an issue. If anybody happens to have the idea of taking deaerated boiler feedwater and heating it further with waste heat, then there you need to consider the effect on the boiler economizer. Boiler economizers are not designed for two-phase flow, and you can get very undesirable effects if the boiler feedwater enters the economizer at temperatures much higher than its design. If the outlet temperature on the economizer gets too close to the saturation temperature, then local areas will generate steam resulting in two-phase flow in a piece of equipment designed for only the liquid phase. Two-phase flow in an economizer can hammer and result in a failed tube.

JAHN SVENDSEN (Shell Global Solutions U.S.): All Shell refineries are very heat integrated. It is part of our energy conservation effort and we do use all low level heat sources including boiler feedwater and streams like that. My experience is, as the previous speaker mentioned, getting the system clean, oxygen free and ensure all the small black particulates are removed prior to start up of such systems. I would like to share another item. We have four refineries in Shell, which are even further heat integrated. All the overhead condensers have been converted to hot water generators, and supply the hot water to a large pressurized water distribution grid, which subsequently delivers heat to the showers and the house heating to the communities around these refineries. At one of the sites we have about 45,000 people who everyday get their hot shower and the home heating during the winter from the refinery. They are conserving energy, and as a matter of fact in a very elegant way by not sending any sort of calories or BTUs directly from the overhead condensers (aircoolers) out to the birds.

JAMES WEITH (Fluor Daniel): At Unocal’s Parachute Creek Shale Oil Project we had one system where we were heating a high pressure oil stream with low pressure steam. We had an oil analyzer on the condensate stream that sent a signal to a diversion valve. If the oil analyzer ever went off, the condensate got diverted to the sewer instead of back to the condensate system. It worked fine for nine years.

+ QUESTION 13

What types of accountability systems are refiners using for energy management? Who is responsible for ensuring that the units are running at their optimal energy?

Based on our experience in interacting with a number of refineries, most of the refineries have created an energy conservation group which takes care of the energy usage in the refinery. Energy groups are mostly setting targets, and then monitoring the performance of the equipment. As far as fired heaters are concerned, which are the major energy consumers in the refinery, the set point is 2-3% oxygen in the stack. In a number of refineries, the energy group engineers or technicians will go around the heaters and set these values on stack oxygen content. This is not very effective, as the operators change the settings frequently. In several refineries, Furnace Improvements have trained the operators. The onus of operating the furnaces efficiently, then shifts to the operators and the shift supervisors. It is more effective as all the heaters cannot work at the same oxygen level in the flue gas at all times. The operating conditions of furnaces vary and it may not be possible to get 2-3% oxygen. It is very important to set target O2 realistically, after reviewing the performance of each furnace.

+ QUESTION 14

What experiences have refiners had with the ceramic coating of Naphtha reformer tubes? Have there been any recent failures or is the technology developed such that the risk is relatively low? What quantitative improvements have refiners noted in capacity, flux, reduction of convection inlet temperature, or fuel gas consumption etc.? How do ceramic coatings of heater tubes and refractory affect NOx emissions ? Does burner type matter?

Ceramic coating in the furnace tubes was tried on the Naphtha Reformer heater tubes by one of our clients. In this heater, Furnace Improvements modified the burners and eliminated flame impingement and reduced the oxygen level to 2-3%. At the same time tubes were coated with ceramic also. Initially, the improvement in furnace performance was credited to the ceramic coating. Within 4-6 weeks, the coating started peeling off, as the vendor had used the wrong paint. Still, the furnace was performing much better, without the paint. Furnace Improvements did the modeling of the heater, to prove that the improvement in the heater performance was only because of burner modification and removal of external scale from the tubes, and not because of the coating. We have modeled the heater with coated tubes (emissivity of 1.0) and found very little improvement in the furnace performance. We do not think that the coating has any impact on NOx emissions.

+ QUESTION 15

CRUDE / VACUUM DISTILLATION - FOULING:

What are the dominant factors affecting coke in crude and vacuum heaters – outlet temperature, peak flux, percent vaporization, velocity, residence time. Or other? Are there low cost effective ways to reduce coking, such as steam injection, oil recycle, additives or tube coatings?

Three of the five factors mentioned above are directly responsible for coking – ie. The outlet temperature, velocity and peak flux. We have seen in several heaters, higher peak heat flux leads to coking, as it affects the peak film temperature. A typical industry accepted practice is that every 20-25°F increase in film temperature, doubles the coking rate. Higher mass velocity through the tubes always helps in reducing the film temperature, and removing layers of coke built inside the tube. Steam injection helps to a certain extent. Furnace Improvements recently modeled a heater with and without steam injection. There was a 16% reduction in the coking rate. Please refer to the attached graphs showing the change in relative coking rates with steam injection and the film temperature profile with steam injection. Downstream of the steam injection point, there shall be a rapid increase in the pressure drop, depending on where the steam is injected. One of the other factors is the flame shape and size with respect to the tube and burner size. The coking rate can be reduced significantly by increasing the number of burners and reducing the flame size. Oil recycle will help, as it increases the velocity through the tubes. We have not seen any successful application of additives or tube coatings in crude and vacuum heaters. Reducing the heat duty back to the design levels will also help in reducing the coking rates. Any refiner having problems of coking in their crude and vacuum heaters should contact Furnace Improvements.

+ QUESTION 16

DELAYED COKING :

How do you safely spall a multi-pass heater with a common convection section?

It is very difficult to safely spall one pass in a multi-pass heater with a common convection section. If one or more radiant passes share a radiant section, then it is even more difficult, unless the heater is on a turndown and enough steam can be introduced into the coil to limit the tube metal temperature to the maximum allowable limits.

+ QUESTION 17

HYDROGEN SYSTEMS :

What have been refiners’ experiences with cold end corrosion in plate type air pre-heaters in the convection sections of hydrogen reformers? What life has been achieved? Has porcelain coating of the plates provided long term protection against acid dew point corrosion?

Cold end corrosion in a plate type air heater in the convection section is a common problem, especially if your fuel gas has some H2S or other Sulphur components. In plate type air pre-heater, there is no way to control the metal temperature of the plate. We recommend installing a steam air pre-heater ahead of the plate type air pre-heater, to eliminate the root cause of the problem. Low pressure steam can be used to preheat the combustion air and as a result, the metal temperature in the plate air pre-heater will always be above the dew point.

+ QUESTION 18

INSTRUMENTS, CONTROL AND OPTIMIZATION:

What percentage of the fired heaters in your plant use APC techniques to control excess air? Were the heaters designed with automated excess air control or were these heaters retrofitted with quired instrumentation?

We have seen very few natural draft heaters, automatically controlling excess air. The problem with most of these heaters is the poor quality of stack dampers and the number of burners. Several years ago, a leading instrumentation company started the oxygen control program, but it could not be executed successfully. Similarly, a leading refinery tried to install combustion control system in most of their heaters, but again the success rate was very low. Most of these heaters, ended up working on manual controls. While burners can be enclosed in a plenum and the air inlet can be controlled, dampers have remained the stumbling block. FIS has started designing reliable stack dampers which can be used with pneumatic controls to control the draft in the heaters. Five such dampers are being supplied to one of the refineries.

+ QUESTION 19

ENERGY :

What types of accountability systems are refiners using for energy management? Who is responsible for ensuring that the units are running at their optimal energy?

Based on our experience in interacting with a number of refineries, most of the refineries have created an energy conservation group which takes care of the energy usage in the refinery. Energy groups are mostly setting targets, and then monitoring the performance of the equipment. As far as fired heaters are concerned, which are the major energy consumers in the refinery, the set point is 2-3% oxygen in the stack. In a number of refineries, the energy group engineers or technicians will go around the heaters and set these values on stack oxygen content. This is not very effective, as the operators change the settings frequently. In several refineries, Furnace Improvements have trained the operators. The onus of operating the furnaces efficiently, then shifts to the operators and the shift supervisors. It is more effective as all the heaters cannot work at the same oxygen level in the flue gas at all times. The operating conditions of furnaces vary and it may not be possible to get 2-3% oxygen. It is very important to set target O2 realistically, after reviewing the performance of each furnace.

+ QUESTION 20

What experiences have refiners had with the ceramic coating of Naphtha reformer tubes? Have there been any recent failures or is the technology developed such that the risk is relatively low? What quantitative improvements have refiners noted in capacity, flux, reduction of convection inlet temperature, or fuel gas consumption etc.? How do ceramic coatings of heater tubes and refractory affect NOx emissions ? Does burner type matter?

Ceramic coating in the furnace tubes was tried on the Naphtha Reformer heater tubes by one of our clients. In this heater, Furnace Improvements modified the burners and eliminated flame impingement and reduced the oxygen level to 2-3%. At the same time tubes were coated with ceramic also. Initially, the improvement in furnace performance was credited to the ceramic coating. Within 4-6 weeks, the coating started peeling off, as the vendor had used the wrong paint. Still, the furnace was performing much better, without the paint. Furnace Improvements did the modeling of the heater, to prove that the improvement in the heater performance was only because of burner modification and removal of external scale from the tubes, and not because of the coating. We have modeled the heater with coated tubes (emissivity of 1.0) and found very little improvement in the furnace performance. We do not think that the coating has any impact on NOx emissions.

+ QUESTION 21

CRUDE / VACUUM DISTILLATION - FOULING:

What are the dominant factors affecting coke in crude and vacuum heaters – outlet temperature, peak flux, percent vaporization, velocity, residence time. Or other? Are there low cost effective ways to reduce coking, such as steam injection, oil recycle, additives or tube coatings?

Three of the five factors mentioned above are directly responsible for coking – ie. The outlet temperature, velocity and peak flux. We have seen in several heaters, higher peak heat flux leads to coking, as it affects the peak film temperature. A typical industry accepted practice is that every 20-25°F increase in film temperature, doubles the coking rate. Higher mass velocity through the tubes always helps in reducing the film temperature, and removing layers of coke built inside the tube. Steam injection helps to a certain extent. Furnace Improvements recently modeled a heater with and without steam injection. There was a 16% reduction in the coking rate. Please refer to the attached graphs showing the change in relative coking rates with steam injection and the film temperature profile with steam injection. Downstream of the steam injection point, there shall be a rapid increase in the pressure drop, depending on where the steam is injected. One of the other factors is the flame shape and size with respect to the tube and burner size. The coking rate can be reduced significantly by increasing the number of burners and reducing the flame size. Oil recycle will help, as it increases the velocity through the tubes. We have not seen any successful application of additives or tube coatings in crude and vacuum heaters. Reducing the heat duty back to the design levels will also help in reducing the coking rates. Any refiner having problems of coking in their crude and vacuum heaters should contact Furnace Improvements.

+ QUESTION 22

DELAYED COKING :

How do you safely spall a multi-pass heater with a common convection section?

It is very difficult to safely spall one pass in a multi-pass heater with a common convection section. If one or more radiant passes share a radiant section, then it is even more difficult, unless the heater is on a turndown and enough steam can be introduced into the coil to limit the tube metal temperature to the maximum allowable limits.

+ QUESTION 23

HYDROGEN SYSTEMS :

What have been refiners’ experiences with cold end corrosion in plate type air pre-heaters in the convection sections of hydrogen reformers? What life has been achieved? Has porcelain coating of the plates provided long term protection against acid dew point corrosion?

Cold end corrosion in a plate type air heater in the convection section is a common problem, especially if your fuel gas has some H2S or other Sulphur components. In plate type air pre-heater, there is no way to control the metal temperature of the plate. We recommend installing a steam air pre-heater ahead of the plate type air pre-heater, to eliminate the root cause of the problem. Low pressure steam can be used to preheat the combustion air and as a result, the metal temperature in the plate air pre-heater will always be above the dew point.

+ QUESTION 24

INSTRUMENTS, CONTROL AND OPTIMIZATION:

What percentage of the fired heaters in your plant use APC techniques to control excess air? Were the heaters designed with automated excess air control or were these heaters retrofitted with the required instrumentation?

We have seen very few natural draft heaters, automatically controlling excess air. The problem with most of these heaters is the poor quality of stack dampers and the number of burners. Several years ago, a leading instrumentation company started the oxygen control program, but it could not be executed successfully. Similarly, a leading refinery tried to install combustion control system in most of their heaters, but again the success rate was very low. Most of these heaters, ended up working on manual controls. While burners can be enclosed in a plenum and the air inlet can be controlled, dampers have remained the stumbling block. FIS has started designing reliable stack dampers which can be used with pneumatic controls to control the draft in the heaters. Five such dampers are being supplied to one of the refineries.

+ QUESTION 25

Has anyone had success with dual firing (fuel oil/fuel gas) of low NOx burners? We continually need to repair and replace burners due to heat deterioration of the metal parts and coking of gas tips. If yes, who is the manufacturer of the burners and how are the burners typically operated – several burners in just oil service, some firing all fuel gas or all burners firing part gas/part oil?

TRAEGER: Unlike many refiners, we still do burn liquid fuel in our refinery. In some cases, we have dual fired with fuel gas. However, we have not duel fired low NOx design burners. Relative to the issue of the tip problem, our fuel system can range from straight FCC clarified oil to as much as 50% PDA derived pitch. Coking and deterioration can be a problem for these fuel oil tips, especially as fuel viscosity is increased. Generally speaking, maintaining the viscosity of the fuel oil to the burner tip at or below about 15 centistokes by adjusting the temperature of the circulating fuel oil loop has resulted in good performance of these fuel tips for us.

I did speak to one manufacturer who does design and sell air stage dual fired low NOx burners. This company is Zeeco, Inc. Their air stage design can burn just fuel or a combination of fuel oil and fuel gas. Generally for combined operation, where they fire oil and gas, they would fire all burners in that service in dual service.

ALLEN: The feedback I have had from our friends and partners in Syn Technolgies at Fluor Daniel is that they have not heard of any successful commercial applications. However one of the burner manufacturers has recommended that one way round this problem would be to segregate the type of fuel burned in each individual burner. One of the fuels could be used to base load the duty and the total firing of the heater adjusted with the other feed. Typically the fuel gas is used to adjust the total firing.

BENNETT: Coking of the burner tips in these type burners could in fact be related to the burner tip placement relative to the burner tile. If the burner tip was inserted too high up in the tile, it could result in coke build up on the oil tip and oil dripping down due to loss of the primary air velocity across the tip. If the tip is too low in the tile, the flame will impinge on the tile on the exit. This will then cause sooting of the edges of the outlet and coke buildup on the tile, oil splashing, oil spills and coking of the gas tip.

I agree with Mr. Traeger on the viscosity and you also want to take a look at the oil pressure and steam pressure differential. We think the steam pressure should be at least 30 pounds above the oil pressure.

We also do not think that you should fire both fuels in the same burner at the same time. This is partly because the gas tips will foul unevenly over time and the gas flame burns faster than the oil flame, so it can tend to rob the air from the oil flame.

Regarding temperature control, you want to base load with one fuel and temperature control with another. If you do dual fire, we expect that the overall flame lengths will be longer by 20-50%.

JAMES D. WEITH (Fluor Daniel, Inc.): Mr. Traeger, were you filtering the slurry oil? What kind of particulates did you have in the slurry oil (part per million)?

TRAEGER: Part per million, I cannot answer you. We generally run about a .2 to .3 BS&W on our clarified stream. We do not have any separation on that stream.

ROGER WITTE (John Zink Company): I agree with the panelists in regard to the control of the fuel on the gas and baseload the other one. The oil temperature and the steam differential is very important. The tolerances and the positioning of the oil tip, tile and the gas tip is very important to ensure good combustion. Probably the most important thing on low NOx type burners is the control of the air through three separate dampers and the air distribution inside of the pea box is very important. We have found that in some applications that if air distribution is poor, it will cause coking on the gas tips and problems with burning the fuel oil.

ASHUTOSH GARG (Furnace Improvements): Fuel oil and fuel gas have been successfully fired using staged air burners in a number of installations. These burners are essentially low NOx oil burners with the fuel gas firing capability built into them. The air staging works better with oil but it does reduce NOx with fuel gas. A lot depends upon how this burners are operated as staged air burners have three air registers per burner and it is normally very difficult to adjust all the three dampers in a multiple burner furnace.

The other type of burner that has been used in this service is the ultra low NOx burner, which comes with flue gas internal recirculation. These burners are essentially gas burners and reduce NOx based on fuel staging and flue gas recirculation. Sometimes an oil gun is installed in the center to fire oil. Typically these burners used to have a metal bluff body where the primary fuel gas firing would take place. In some cases these bluffs started getting deformed after a period of 3-4 years. Since these burners are still being developed the burner companies have come out with models that have tangential firing of primary gas tips and even the metal body has been reduced to a cylinder. These burners have been in service for at least 3-4 years and no problems have been reported.

Good operation of the burners is very important in ensuring long life of these burners. It is recommended that all the burners should be fired whenever possible. If the burners are shutdown a little airflow will help cool the internals in the burners. The best ways to fire combination fuels in a heater is to base load one of the fuels (mostly gas) and supplement it with the other fuel (generally oil). The best performance of the burners will be achieved if all the burners are fired uniformly. There is no way to control combustion airflow in individual burners if some burners are fired on gas and other on oil. This is because there is no way to know how much excess air should be fed to the burners firing gas and to the burners firing oil. It may lead to sub-stoichiometric combustion in some burners and super stoichimetric combustion in the other burners. Shutdown some of the burners at the turndown if they are not needed.

Good Ultra Low NOx gas and oil burners are made by Callidus Technologies and John Zink Company in US, by Air Oil (now a part of Peabody group) in UK and LD Duicker in the Netherlands. Our recommendation would be to use staged air oil burners for combination firing. We also recommend that you test the burner at the vendors test facility before you buy and install them in the heater. Burner specifications also play a great part in the performance of burners. Correct sizing and specification of the burners is also very important. Furnace Improvements has been helping a number of clients in sizing, specifying burners correctly and also witnessing the burner tests.

+ QUESTION 26

What has been the panel's experience with regard to pigging of heater coils? How does it compare with steam air decoking?

MASTRACCI: Steam/air decoking of course will burn off any carbon or hydrocarbon found in the heater coils. Any non-combustible solids, which are small and loose may be carried away by the steam and nitrogen. However, gas velocities may not be high enough to carry away any metal scale adhered to the inner wall of the coil. Pigging is intended to solve this problem. By passing pigs sized for the inside diameter of the coils, any remaining metal scale should be removed.

We have used the pigging procedure at 3 of our 4 refineries. The procedure was quite successful, and in one application, mechanical damage did not allow steam/air decoking to be used. Pigging was the only option. We had one case on a coker heater where we successfully pigged 2 of 4 passes. As it turns out, we may have ended the pigging procedure prematurely on the other 2.

Some practical advise. Whether you steam/air decoke or pig, the heater tubes must be inspected for damage like bulges - by strapping or other method, residual coke, that is ringing the tubes and any other routine turnaround inspection normally done. If not, you run the risk of restarting with a potential problem still in the heater.

RADCLIFFE: I have a couple of examples of use of the pigging technique. PCK in Germany used pigging to clean their vacuum visbreaker, high severity visbreaker furnaces. The vacuum furnace was the first where this technique was applied in 1994. As a result, they observed lower off gas temperature and a 1.8% improvement in fuel efficiency.

The Visbreaker furnace is cleaned every 18 months. Cleaning the four pass furnace takes about 18 hours. Similarly, the two pass high severity visbreaker takes about 18 hours. The pigging procedures takes about one day less than the steam air decoke and results in less stress to the furnace tubes.

The cost of the procedure has dropped significantly since they first started using it.

While I was at Milford Haven, we switched from steam air decokes to pigging for the vacuum furnace in 1992. Again, this proved to be extremely successful. We had never really successfully got a good burn on the steam air decokes we had previously attempted.

We saved about one day going from about five days to four days, and we solved one of the major problems we had with steam air decokes namely noise complaints from neighbors. Also, the gas flows for the furnace were dropped by about 2% after the decoke.

The other thing that was interesting for us is that we were able to determine the location of most of the coke. The predominant amount of coke was in the roof tubes of the cabin furnace, as a result of flame impingement. The most heavily coked passes were the third and fourth out of five. The overall cost is almost certainly lower if increased tube life, reduced interference with the turnaround schedules due to live utilities, and fuel costs are taken into account.

RASBOLD: At our Tulsa refinery, we pigged two crude unit vacuum heaters for the first time during our March 1998 turnaround. One heater is four pass while the second heater is two pass. We compared the heat transfer rate in the heaters once they were online after the pigging. These rates were very similar to those results after a typical steam air decoking process. Inspection of the tubes yielded no significant loss from potential metal scoring from the pigs. One of our incentives of pigging the furnaces in Tulsa was that the process was essentially a turnkey operation from our supplier, with very little Sun operator or engineer involvement during the process itself. With steam air decoking, monitoring of the process required one engineer and one operator during the entire event. Pigging these heaters allowed Sun personnel to concentrate on other turnaround activities. Our largest incentive to pigging these heaters was timing. Heater work for our turnaround was one of the critical paths. Pigging of both heaters took 36 and 24 hours respectively, as compared to over 96 hours total for the steam air decoking.

Our Yabucoa refineries have had similar experiences with a crude vacuum heater. We reduced decoking time there from five days for steam air decoking to two days with pigging and had almost identical recovered heat transfer coefficients for both processes.

As far as pigging furnaces in different types of operating units, we have also pigged a two pass coker charge heater in our Tulsa refinery. We typically steam air decoke this furnace and will continue to do so due to the furnace needing to be decoked every six months. However, during one operating cycle, we severely coked the tubes and were unable to get enough flow through the furnace for a steam/air decoke. So we hydroblasted one pass and pigged the second pass. Once complete, we steam air decoked the entire furnace as a proof burn. The amount of proof burn between the two passes was insignificant in either case.

Finally, as a side comment concerning an experience we had pigging our Philadelphia refinery crude unit heater. We pigged one pass only to remove scale. During subsequent inspection of the outlet headers for unrelated naphthenic acid corrosion, pigs were found in several locations within the header after the pigging process had already been completed. Needless to say, accurate inventory of the pigs being utilized by the supplier must be done.

SMITH: At Valero in Corpus Christi, we have pigged both our HDS heater and our vacuum furnace with very good success. We used a Canadian company (Pinnacle Pigging Systems, Inc., Red Deer, Alberta). I believe the last one we did cost approximately $30,000. No additional thinning of the U-bends were noticed afterwards. We do a thorough inspection after we pig. The advantages of pigging have been stated as; any restrictions will be found, reduction in pipe damage, improved safety, cost effectiveness, significant reduction in shutdown time and environmental friendly.

TRAEGER: We have seen similar experiences. We have done one heater with the pigging procedure. One other advantage I would mention is that while the pigging procedure is going on, you can go into the heater and perform other maintenance such as burner repair and things of that nature.

I have a couple of minor problems that we did run into that I will mention. On the outlet radiant key fitting or the last two in the radiant section, we missed the fact that its internal shape was oblong instead of round. The result was that we tore up a whole bunch of pigs that we had not intended to tear up. But we did get through the heater and got it clean.

The second thing we ran into is this heater has key fittings on both ends of the tubes. We went through great efforts to make sure that before we told the contractor he was finished and could leave, that it was clean, and checked that there were not any pieces of coke left. However, after startup, we plugged some feed nozzles on the riser to the FCCU. I think what was going on was the non-uniformed shapes on the key fittings were holding some small pieces of coke that did not get moved out of the system until we got up to full charge rates.

The other final area of discussion in regards to pigging has to do with non-uniform or deformed tubes. There has been some discussion in industry that the pigging process is unable to remove coke from buldged sections of tubes, for example. The result could be localized higher tube wall temperatures because of the failure to remove the coke in that section. Some of the different pigging contractors claim to be able to utilize varying degrees of over sized pigs to address this problem and should be considered when you are looking at decoking a heater.

ASHUTOSH GARG (Furnace Improvements): Based on the feedback received from our clients, the results are mixed. Some clients have had very good experience with pigging but others are not happy. It appears that the pigging process is still improving and is yet to reach maturity. At the same time, due to infrared thermography becoming more and more common and affordable, steam air decoking is also gaining ground.

In the pigging process, the client needs to do a little homework before the pigging process can start. The coke in the tubes needs to be estimated to determine the initial pig specifications. The coke thickness in the tubes is not uniform and also the coke thermal conductivity varies widely anywhere from 1.2 to 30 btu/hr ft2. If the coke thickness is overestimated the price of pigging as well as time required goes up. On the other hand if it is underestimated, there is a chance of pig getting stuck in the tube. If the pig is stuck in the tube then the process of locating pig, cutting the tube and putting a new piece back is a very long and time consuming affair. It is a good idea to do a field survey of the heater and then do heater modeling to estimate the coke thickness. This can save a lot of hassles and time. Furnace Improvements has estimated coke thickness prior to pigging for several clients.

If the coke deposits are of normal thickness like 3 or 2 inches and the provision of steam air decoking is already existing in the plant, steam air decoking often turns out to be more economical. Use of infrared thermography can cut down the time by monitoring the tube temperatures and hot spots, if any. The time required by steam air decoking is generally longer than pigging. It is normally not controlling most of the time.

Pigging is useful and economical if the coke layers are thicker than 2 inches as steam air decoking will take a much longer time to burn out the coke inside the tubes. Pigging is also faster and does not pollute the atmosphere. Pigging can be a problem if heater tubes have plug headers on one side or both sides of tubes. Some clients have reported that the pigging was not able to remove coke completely and some saw scorching of tubes inside surface. A number of companies are coming out now with inspection instruments which can check the tube inner surface very accurately and ensure that the tubes are completely cleaned.

+ QUESTION 27

Our refinery crude heater is a box type fired heater. We have noticed significant hard scale deposits on the convection bank tubes. The convection bank is comprised of a bare and stud tube combination. We would like to know of an effective on-line cleaning method other than conventional soot blowing.

RADCLIFFE: At Elf Milford Haven refinery, we have a very similar design of furnace that was firing both oil and gas. Generally, the more oil we fired, the bigger the problem. This oil was fairly low vanadium though which made it rather easier. We had lots of major problems with the conventional steam soot blowers that the furnace was built with. We replaced them with ultrasonic soot blowers and found these were very effective in terms of reducing the fouling level from what had previously essentially filled the whole of the convection section with fine ash, to just having small piles on top of each tube. The important thing we found was that you have to keep the soot blowers working all the time. Once the fouling builds up, they will not dislodge it. They will only avoid the deposition in the first place.

MASTRACCI: The originator of this question may have a sulfur and, possibly, a dew point problem, which should be tackled first if you are burning refinery fuel gas. If you are burning fuel oil, check the Na/V ratio. You may have a sodium meta-vanadate problem. With regard to on-line cleaning of a convection bank, we have experience with a related procedure, which may not work with a vertically fired heater with low flue gas velocities. We have injected 'Black Beauty' - a type of sandblasting sand - into the combustion air stream of a horizontally fired CO Boiler. The Black Beauty was added for 3, twenty-minute periods as a way to remove FCCU catalyst adhered to the tubes. We achieved only minor success the few times we used this procedure and we no longer use it. The only reason we think we had any success is likely due to the fact that this boiler has a high superficial velocity (more than 50 ft/sec). WALDRON: The only comments I have focuses on a heater that was mainly oil fired. The refiner we are talking about is a European refinery. They have had success injecting an alkalide nitrate solution. They injected through the inspection doors with lances. The mist that forms reacts with deposits on all surfaces, tubes, ducts and air preheater elements to loosen deposits and improve heat transfer. Again, the chemicals are most effective on fuel oil deposits containing vanadium. The two known chemicals that I have information on are called Thermochem SF12 and Akzo SDR. P.P. UPADHYA (Mangalore Refining & Petrochemical Company): Is the ultrasonic instrument to be kept on all the time? RADCLIFFE: Yes, it is important that you keep the ultrasonic surplus on all the time you are running the furnace. If the fouling starts to buildup, it will not dislodge. It is prevention rather than cure.

+ QUESTION 28

What experience have refiners had with ceramic coating in furnaces? Have they been used in coker furnaces, and have intervals between decokes increased? Have they been used in reformer furnaces? Review the service life, application techniques, benefits analysis, and pitfalls.

PETERSON: We ceramic coated two heaters in one of our refineries. The first one we did was a naphtha hydrotreater charge heater. We did this in 1996, and it was not noticeably successful. It was probably the wrong application for this type of technology, as the heat transfer was largely convective and we do not fire this heater very hard. It was an opportunity to see what the process looked like. We may also have put the coating on too long before we started the heater up, as it was quite humid, so some of the coating may have left the tubes. In February of this year, we went on to a heater that was probably a better application, a naphtha reformer charge heater which is all radiant heat transfer. I wish I could tell you that I had skin thermocouples before and after and tell you what the difference is, but we came close to this. We installed skin thermocouples and we have a sister heater in the same cabin and the side that is coated is noticeably dark, and the side that is not coated is bright from all the material on the tubes, the old ash and scale. So, it appears as if the coating has done some good from emissivity and heat transfer. Even if it has not done some good that way, the exercise of sandblasting the tubes and putting a ceramic coating on probably is going to keep them from oxidizing again. So, in that standpoint, we view that as being successful. Skin thermocouples between the coated heater and the uncoated heater show a difference of about 100 F between the two, both on the peak temperature and average temperature with the coated heater being colder. We are considering doing some additional heaters in the refinery to help even out heat flux, as well as improve efficiency in the heaters. So far it looks like it has been successful for us. ASHUTOSH GARG (Furnace Improvements): As a designer of fired heaters, refractory emissivity does not figure anywhere in the radiant section calculations. We have changed the refractory lining of several heaters from castable to ceramic fiber, and the radiant section performance was never affected. But, coating on ceramic tubes will definitely make some minor differences, because emissivity of the tubes does come into play. LLOYD WINGER (Conoco Inc.): We have installed ceramic coatings on four heaters: reformer feed hydrotreater reboiler crude feed and coker feed. We have observed more heat transfer in the radiant section. However, the net benefit is difficult to quantify due to numerous other changes to the process units involved.

+ QUESTION 29

What are refiners doing to control or eliminate fouling of burner tips? We suspect that most of the fouling results from amine system carryover.

HENKE: Complete separation of amine from fuel gas is a critical element in elimination of burner tip fouling. Most of our fouling events can be traced back to poor amine contactor operation. Much can be done to improve the separation of amine from fuel gas in the sweet fuel gas knockout drum. Proper distribution of the inlet stream and installation of coalescing devices are just a few. In cold weather climates, the proximity of the sweet fuel gas knockout drum to the furnaces must also be considered. Other areas for consideration in reducing burner tip fouling include particulate filters and optimized burner tip design.

BINFORD: I think Mr. Henke covered most of the points. Most of this fouling, when we get involved in a system like this, is either coming from upsets in the amine system or from iron sulfide scale sloughage that is not properly caught in dip legs, knockout pipes, or trash pipes. It is very important that these "catch" systems are operating properly and that they are periodically maintained. Good housekeeping practices in the amine unit, which is covered in detail in the amine section, are a real key to keeping your fuel gas systems clean. If these measures are still not effective, chemical antifoulant programs can effectively mitigate the deposition of trace contaminants at the burner tips. This is one area of the refinery where antifoulants have been used successfully for clean-up purposes. For a program to be successful, it needs to be tailored to address the fouling mechanism, specifically based on deposit analysis. We usually use deposit analysis rather than fuel gas analysis, because, most of the time, fouling is incidental in nature, so that a spot fuel gas analysis will not give a representative indication of the true fouling components.

CAIN: We are in the process of installing fuel gas filters at each heater down stream of the fuel gas knockout drum. We are addressing our amine carryover problems as well.

HAHN: Very similar to Mr. Cain’s answer, we feel that most of our fuel gas plugging problems were related to amine problems, and we are addressing those. But, we also put in filters on our fuel gas system, and that took care of the problem very nicely for us in most cases.

RONALD BREDEHOFT (KTI Corporation): In addition to what the panel has addressed, I would add that some of our clients have insisted on a 3/32 in. minimum drilling on their burner tips to avoid or minimize this plugging.

CHARLES T. ORLANDO (Pall Corporation): I would just like to add to the panel’s discussion that a high efficiency coalescer will take either the hydrocarbon or the amine, or both, down to levels less than 0.01 ppm, and the particulate levels down to 0.3 micron absolute. It is very important to protect the burner tips.

+ QUESTION 30

Has anyone tried controlling fuel gas flow to process heaters on a mass basis (software CEMS or other technologies) to compensate for rapid fluctuations in BTU content to the furnace, or to better measure actual fired/absorbed duty, or emissions impact ?

FOSTER: In some services, we do use a calorimeter value as a feed forward into a temperature controller for duty control application. We do not use CEMS, but we are using PEMS, predictive emissions monitoring system. In fact, we are implementing that even as we speak on our new crude unit and on our new coking unit heaters to comply with new permit requirements. This technology can accurately predict emissions and is recognized as a viable technology by our regulators in Texas. The step testing is expensive because we were required, as part of the step testing, to run at less than optimum rates on the heaters in order to gather the emissions data that were required for predicting the emissions. But, I expect that the predictive system is going to be a cheaper, more cost-effective, and more accurate system in the long run, because it will require less maintenance and it will be more reliable.

HAHN: At one time, we had a relatively sophisticated control system on our fuel on one of our new heaters. It had BTU on-line analyzers which are based on chromatograph results, but we found that the on-line analyzer loop feedback was too slow to keep up with the changes in the fuel gas composition. So, the feed forward was not effective. We could never get it tuned properly, so we finally abandoned that whole system. We do, however, understand that lately some people are now putting mass meters on their fuel gas system, and are using them, apparently successfully, to control their fuel gas. You may want to talk with some of the suppliers of those types of instruments for specific examples.

CAIN: We have feed forwards on our fuel gas gravity and feed gas flow, as well as the steam flow. These feed forwards work well for us.

+ QUESTION 31

Are there any new developments in soot blowing? What has been the success with continuous sonic soot blowers?

SAMUELS: There is a distinction between sonic soot blowers and infrasonic soot blowers in that infrasonic soot blowers operate at a frequency below audible sound, while sonic soot blowers can be loud during their cycle. The soot blower vibration frequency should be selected to fit the application. Two infrasonic soot blowers were recently installed in the superheat section and economizer section of a 200,000 lb/hr 600 lb steam CO boiler on Robinson’s catalytic cracker. Actually, this was done in 1994. Low frequency soot blowers were selected because they do a better job at removing particulates. Previously, we had two retractable 600 lb steam soot blowers in the superheat section, and 12 smaller fixed 150 lb steam soot blowers in the economizer section. The steam soot blowers caused problems with opacity spikes, steam drum level upsets, and occasional tube failures due to tube impingement.

The infrasonic soot blowers generate a low frequency 22 hz wave for 60 seconds every 2 to 3 minutes. This provides a multidirectional effect knocking catalyst off the tubes, even upstream of the fire box section of the boiler. Tube cleaning is more effective; opacity spikes and drum upsets have subsided, and required equipment maintenance is less compared to the old soot blowers. Vibration has not been a problem, even though the boiler skin is very thin in the economizer section. One hundred psig plant air in a surge tank drives the soot blowers. The blowers themselves consume 1000 scf/min in the incinating mode and 200 scf/min in the idling mode. One lesson we learned in the installation of these soot blowers is that you do want, particularly in the CO boiler, to put a knife valve between the horn of the blower and the boiler itself for maintenance, because you have positive draft there.

ROMAN: We have installed sonic soot blowers in both the air preheater and the radiant section of one of our large crude heaters. A small high frequency sonic soot blower operated at 150 hz to 250 hz in the air preheater and it was successful, but its effective cleaning area was limited. The large low frequency sonic soot blower operated at 20 hz in the radiant section and dislodged and damaged refractory in that heater. We have also tried sonic soot blowers on the dirty side of a pulse jet bag house module, and severe bag damage occurred.