Claus Plant Combustor Survey

Dan Banks, P.E.

Banks Engineering Inc.

Tulsa, Oklahoma

Introduction / Summary

Every Claus sulfur recovery plant includes a Reaction Furnace as the first processing step and a Tail Gas Incinerator as its last. Although these units are relatively simple, the conditions under which they operate make them more prone to problems than other parts of the plant. Careful preparation of specifications, attention to detail in fabrication and proper maintenance procedures can make combustor problems a minor issue in Claus plant operation.

Claus Fundamentals

In its simplest form, the activity of any Claus sulfur plant can be summarized as:

H₂S + O₂ => H₂O + S

In the 1800’s researchers found that this reaction occurred best on the solid surfaces of their test equipment. About 1890 the Claus process was commercialized, with H₂S being oxidized with air over a bauxite catalyst in a single reactor. In 1937, I. G. Farbenindustrie improved the process by adding a specialized combustor (Reaction Furnace) and heat recovery boiler upstream of the bauxite reactor. This sets the date for the first Claus plant combustor problem.

It was found that several catalytic reactors in series, each followed by a sulfur condenser, would increase overall sulfur recovery. After each condenser the process gas must be reheated before entering the next reactor. Reheating with steam is usually preferred due to the complexity of combustion devices, but some plants use combustors (Inline Reheaters) firing acid gas or natural gas for this purpose.

Since even a very efficient Claus plant reactor train leaves some residual H₂S in the final "tail gas", soon special Tail Gas Incinerators were introduced to assure that any sulfur leaving the system was in the form of SO₂.

As a further means to treat the tail gas stream, in recent years Tail Gas Units were added to the Claus plant upstream of the Tail Gas Incinerator. The Tail Gas Unit employs another combustor, the Reducing Gas Generator, to assure proper conditions for the special catalyst bed. In operation, residual SO₂ is reduced to H₂S, removed with a TGU amine unit, and recirculated to the Reaction Furnace.

The purposes of the four combustors in Claus plants are:

· Reaction Furnace: oxidize part of the sulfur bearing compounds in the plant feed gas under reducing conditions to form sulfur molecules.

· Inline Reheater: reheat the cooled process gas for entry into the next in a series of reactors.

· Reducing Gas Generator: reheat the tail gas from the last catalyst bed, adjusting the content of H₂ so that the special Tail Gas catalyst performs best at converting SO₂ back to H₂S for recycle to the Reaction Furnace.

· Tail Gas Incinerator: completely burn any residual sulfur bearing compounds to SO₂ in the presence of excess oxygen and disperse the resulting combustion products at altitude.

Each of these units maintains a flame in a refractory lined chamber. In each case the flame is suspended within the chamber and monitored with instruments to shut down the Claus plant should problems occur. Materials of construction are selected to resist the high operating temperature, potential flame impingement and the temperature swings associated with startup activities. Sensors and automatic controls are provided to assure safe and controlled combustor operation.

Combustion Fundamentals

Each combustor is equipped with a burner to manage mixing and ignition of the process gas, fuel gas and air streams. Each burner must bring the streams together at the right time to achieve quick lightoff. The burner must be shaped to hold the ignition point at the intended position, rather than allowing it to wander into the furnace or back into an area of the burner not designed to handle flame temperatures. The mixing of air and combustibles must be thorough enough to avoid creating pockets of fuel-rich gas, which can result in formation of carbon due to incomplete combustion. Improper mixing can also lead to noise caused by flame pulsations, as fuel-rich gas pockets combine intermittently with oxygen rich gas to create small, repetitive explosions.

For Reaction Furnaces, Inline Reheaters and Reducing Gas Generators, combustion must always be "substoichiometric" once the various catalyst beds have been placed in operation. This is because any residual oxygen in the product gas will react with the catalyst bed, deactivating it and even damaging it if the oxygen level is high enough. The amount of air used must be controlled accurately to maximize overall plant efficiency. Likewise, combustion product temperatures must be controlled to avoid damaging the combustor or downstream equipment. The Tail Gas Incinerator, on the other hand, must be operated with excess air in order to completely oxidize all residual sulfur compounds to SO₂. Here the amount of excess air must be held as low as practical so that fuel gas usage is minimized – extra air will not aid incineration but must be heated to stack temperature nonetheless.

Examples of each of the combustor systems are shown in the figures:

Figure 1 shows a typical Reaction Furnace, including internal mixing enhancements achieved with a "choke ring" and "checker wall." A number of different types of burners may be fitted to this design.

Figure 2 shows a common type of Acid Gas Burner, with discrete air injectors arranged to induce swirl for enhanced mixing at the refractory capped acid gas gun.

Figure 3 shows a typical Acid Gas Burner fitted for oxygen enrichment.

Figure 4 shows a furnace design commonly used for Inline Reheaters and Reducing Gas Generators. The version shown incorporates a swirling air injection system, so the burner is simply a pilot and gun for introduction of the acid gas or fuel gas at the narrow "throat" of the refractory lined combustion chamber. Process gas flows around the combustion chamber, mixing with the hot flue gas before exiting the vessel.

Figure 5 shows a simple Tail Gas Incinerator furnace, designed for natural draft operation. The stack height is 210 ft. to insure dispersion of the SO2 in the exhaust and guy wires are used for support. For heat recovery, the bottom section of the furnace would be mounted horizontally and the flue gas would flow through a boiler before entering a stack to atmosphere.

Figure 6 shows a common Tail Gas Incinerator Burner design, with tail gas introduced around the fuel gas flame burst. This design may be forced or natural draft.

Problem Reports:

Interviews were conducted with personnel who operate Claus plants, those who supply combustors and those who specify and handle startups of Claus plant combustors. The purpose of the interviews was to gather problem reports from as many sources as possible. The diverse interview pool helped assure that problems at each level of supply and operation were included. Questions about feed preparation, system control, combustor operability and maintenance needs were covered. Observations and suggestions for improvement were noted in each case. The interviews were grouped by type of combustor.

Reaction Furnaces

Sighting Problems: Many problems were noted with connections designed to allow monitoring of the interior of the reaction furnace. Sometimes the locations and orientations of sight ports or flame scanner mounts were incorrectly designated on design drawings. In other cases shop personnel fail to aim the external connections at the intended spot, or refractory installers fail to maintain the sight line established by the external hardware. In several instances the external connections were blocked by external process piping and rendered useless. Since correcting of these problems requires shutting down the furnace, if the problem is not fatal, the unit operator often chooses to "make do." If the problem is more serious the offending hardware is modified as needed. Close attention should be paid to sight lines at all stages of supply.

Rainshield Problems: Most Reaction Furnaces are located outdoors and include a covering of corrugated metal or other system to prevent localized cooling, as may occur during a rainstorm. The rainshield is built with a standoff frame, air gaps and sometimes dampers to assure appropriate flow of air between the shield and the Reaction Furnace shell. The refractory lining of the furnace is designed to accommodate this heated air surrounding the shell. Trapped air creates extra insulation, raising the vessel shell above target levels. No reports of structural damage were mentioned in the interviews. A missing shield reduces the intended shell temperature and allows localized cooling, which has resulted in refractory spalling in at least one case. Correct design, installation and maintenance of the rainshield is important.

Control Problems: Control of fuel gas and air flows during startups has proven tricky in numerous cases where the hardware is designed for normal furnace operating rates without consideration for startup conditions. The designer should consider using dedicated flow circuits to allow controlled operation at all times. Optical pyrometers are often used to monitor Reaction Furnace temperature – they are reported to work well when a preventative maintenance program is in place. Loss of temperature measurement may be tolerated unless substantial ammonia is being fed or O₂ enrichment is being used. In one case, loss of the pyrometer in a Reaction Furnace using O₂ enrichment led to serious refractory damage when the operating temperature inadvertently increased to 3000^oF for an extended period. In other cases these combustors have been operated for years with no temperature monitoring at all except during the initial refractory cure. A relatively constant acid gas feed composition permits this type of operation. One user observed that ceramic thermowells often were broken as a result of sulfur deposition where the well passes through the refractory hotface. He has had good luck with "double purged" thermocouples (purging with nitrogen both inside and outside the thermowell.)

Process Gas Problems and Burner Deposits: Proper control of upstream processing equipment is important to avoid problems in the Claus plant. In one case, offgas from a sour water treatment system causes periodic solids plugging of the acid gas gun in the Reaction Furnace burner. The gun requires periodic cleaning of deposits, which are thought to be ammonia bisulfide by the plant operator. The same operator observes that NH₃ in the reflux water in the amine treatment system should be kept below 1% (resulting in no more than 1000 ppmv NH₃ in the gas stream) to avoid solid deposits in the burner gun. One report of solid sulfur deposits in the air plenum of an oxygen enriched Reaction Furnace burner was noted. The sulfur was contained in the recycle stream from the first sulfur condenser, and is thought to collect due to poor insulation of the sulfur drain area of the burner. The operator believes switching from a recycle blower to a steam eductor may solve the problem.

Materials of Construction: Since the Reaction Furnace operates at high temperatures, refractory selection and installation quality are very important. Careful verification of these factors during design and installation can reveal and eliminate startup and operating problems. Published refractory temperature ratings must be reduced because the hot face environment is reducing. Typically a 300 – 400^oF derating is recommended. The actual operating temperature depends on the heating value of the process gases, as well as whether oxygen enrichment is used. If O₂ enrichment is added to an existing Reaction Furnace, the loss of nitrogen in the reacting mix results in an increase in operating temperature unless recycle gas or other cooling medium is used. One operator reports problems with the refractory protecting the O₂ injector nozzle in his burner – the material and nozzle design are being changed to handle the higher temperature. In some Reaction Furnace designs a brick "checker wall" is installed across the furnace to insure mixing downstream of the burner. If careful temperature control is not practiced the bricks can actually sag, requiring shutdown to restore proper operation. The acid gas burner must tolerate very high temperatures, even without oxygen enrichment. Reports of thermal damage to the acid gas feed tip and pilot hardware are common, leading to special attention being paid to tip construction and materials. It is not unusual for the natural gas pilot burner to be built to allow partial or complete removal from the combustion zone once the main flame is established.

Inline Reheaters

Control Problems: Many of these units fire acid gas bypassed around the Reaction Furnace, while others, especially small ones, fire fuel gas. Reducing (starved air) operation is required so that no excess O₂ is present in the product gas flowing to the downstream catalyst bed. Excessively reducing conditions can result in soot formation in the combustor, leading to fouling of the catalyst, which means loss of catalyst activity and excessive pressure drop across the bed. Unfortunately, soot formation is fairly common with these units, since instrumentation is sometimes skimpy, and air control based on flame color is a learned art. Steam is sometimes added to the burner air stream to minimize the potential for soot formation, especially in units that fire fuel gas. During Claus plant startup, fuel gas and air are burned to bring the associated catalyst bed up to operating temperature. If the catalyst is new, this step can be done with excess air, but on subsequent startups reducing operation is essential to avoid catalyst damage and steam is normally injected to achieve the needed combustor outlet temperature. Flow rates during bed heat-up are much lower than the rates during normal plant operation. Both the combustor and the control system are optimized for normal operation. This means that during startup, both burner mixing and gas flow measurement accuracy are compromised, which can make holding stoichiometry and temperature difficult. The ability to precisely control the flow of air, steam and acid gas or fuel gas is therefore very important. The lack of proper flow control hardware is a problem voiced by a number of users. As a result, many prefer to reheat using a high pressure steam exchanger rather than a combustor.

Retrofit Problems: One user had been firing the burner on natural gas, but switched to burning H2S instead. The natural gas gun tip was left in place for use during startups, but was not purged with nitrogen or any other gas. Radiant heating in this area gradually burned the tip off during normal operation, so flame stability problems were encountered during the next startup attempt. The solution was to shorten the natural gas gun by several inches, drawing it back into the combustion air stream where it was cooled enough to avoid future problems. The same user reported problems with new "improved" burners installed to replace existing worn out models. No one recognized that there was a "sweet spot" where the new burners operated well – operation below this point (during startup) resulted in excessive soot problems and catalyst plugging. Recommendation: have the operations personnel "calibrate their eyes" during normal operation, so that changes in flame characteristics will be noted and dealt with quickly.

Other Problems: Inline Reheaters can have the same problems with sighting, rainshields and materials of construction as noted for Reaction Furnaces.

Reducing Gas Generators (RGG)

Reducing Gas Generators are very similar in design to Inline Reheaters, and share many of the same problems.

Fabrication Errors: One operator found that the low flows required by initial catalyst bed heating were controllable during initial plant startup, when excess oxygen in the combustor exhaust was permissible. During subsequent heat-up, when reducing operation was required, no adjustment to the fuel gas or air flows could reduce the effluent O₂ reading to zero. Finally the unit was shut down for inspection. It was discovered that an internal plate separating the combustion air plenum from the process gas plenum was not seal welded in place, allowing part of the start-up air to bypass the burner throat directly into the discharge plenum. Once this fabrication error was corrected, reducing operation was achieved and a normal startup was possible. Unfortunately, the inexperienced startup crew allowed heavy sooting of the catalyst, which finally was replaced. The operator observed that quality control during design and fabrication is essential, and that experienced Claus plant start-up should be present at critical times.

Control Problems: It is common to design RGG controls for normal plant operation without considering the requirement for much smaller gas flows during startup. This often makes accurate measurements impossible during startup and operating personnel are reduced to qualitative judgments about flame quality (yellow = oxidizing, slightly orange = just right, more orange = too reducing and making soot.) The use of analyzers in the Tail Gas Cleanup Unit provide critical feedback about RGG stoichiometry, but separate flow measurement loops for startups is recommended by several plant operators despite the added cost. One operator has units operating with refinery fuel gas in one location and with natural gas in other locations. Due to the variable oxygen demand encountered from day to day with refinery fuel gas, this operator strongly recommends "always use natural gas to avoid sooting problems."

Operating Errors: One respondent mentioned an experience with SCOT catalyst bed sooting which was caused by off-target operation of the reaction furnace. The hydrogen analyzer was drawing a sample downstream of the SCOT catalyst bed – a non-zero hydrogen reading at that point would indicate that the RGG was running substoichiometric enough to generate the H2 required to drive the tail gas SO2 to H2S, as desired. The operator continued adjusting the air / fuel ratio lower and lower, but never saw the expected hydrogen. Fortunately, before excessive soot was produced, he noticed that the Reaction Furnace air flow was too high, resulting in a much greater conversion of H2S to SO2 than desired. The extra SO2 was consuming the H2 generated in the RGG at a rapid rate. Once the Reaction Furnace chemistry was fixed, hydrogen consumption in the SCOT unit returned to design levels and he began seeing the hydrogen he was expecting at the analyzer.

Tail Gas Incinerators

These units are often designed for natural draft operation, but may be forced draft, particularly when equipped with a waste heat boiler. Stack height is specified to limit ground level concentration of the SO₂ in the flue gas. In recent years air permit limits have resulted in the addition of Tail Gas Cleanup Units between the last Claus sulfur condenser and the Tail Gas Incinerator, as well as increased incinerator operating temperature to limit CO levels in the flue gas.

Control Problems: Several users report air flow control problems. For good operation a flue gas O₂ level of about 3% volume is needed, which sets the flow of combustion air needed. At least one user plans to switch to forced draft operation after living with a natural draft arrangement incapable of holding O₂ levels below about 10%.

Air Supply: One user reports provision of an undersized air blower on an incinerator with a waste heat boiler. This design error limited overall Claus plant capacity and was resolved by changing to a blower with more pressure capability. Another user replaced the natural draft pilot burner on his incinerator with one supplied with plant air – a response to pilot stability problems. He recommends use of a "robust" pilot, operating continuously rather than only at startup.

Design Sizing: Most combustors are purchased on a competitive bid basis, driving the suppliers to minimize fat in their design. One user encountered excessive carryover from the SCOT absorber tower, sending extra combustibles to the incinerator. Although the combustion volume was adequate, the unit ran low on air. An extra air register was added to solve the problem.

Materials of Construction: One user reported failure of the fuel gas injection hardware in the burner. The damage was caused by a combination of incorrect alloy specification and the relatively high level of reflected heat from the flame. This burner used a ring burner to distribute the fuel gas. A number of users have suffered corrosion problems in the tail gas injection plenum area. In one case the plenum was corroded to the extent that tail gas was mixing with the incoming combustion air upstream of the fuel gas injection point and causing flame stability problems. Refractory problems are less common with Tail Gas Incinerators than with the other Claus plant combustors, but proper refractory and refractory anchor installation are critical. This type of problem is often discovered soon after startup if missed during fabrication inspection. As with any refractory lined vessel, quick startups and shutdowns can shorten refractory life through spalling due to differential thermal expansion.

Tail Gas Handling: Tail gas is sent to the Incinerator through a pipe sized to minimize capital cost. The tail gas velocity can be quite high and combustion stability problems can result if the velocity is not slowed upstream of the burner mixing zone. In an incinerator using a plenum type tail gas injection arrangement, problems can be avoided by designing for a constantly increasing tail gas velocity up to the point of injection. Incorrect design has resulted in heating of the tail gas plenum due to back flow of combustion products. In older units the tail gas is saturated with sulfur vapor, so the tail gas line and injection plenum must be insulated to avoid condensing out liquid sulfur. Even so, sulfur can collect in the burner unless a sulfur drain arrangement is provided to allow the liquid to flow into the incinerator for burning. Any solid deposits can affect tail gas distribution and therefore combustion efficiency. One user in the northern US operated tail gas transfer piping with insufficient external insulation. The line became plugged, as solid sulfur and iron sulfide corrosion products formed in the lines, eventually requiring early shutdown for removal.

Fuel Gas Handling: A number of reports of fuel gas capacity problems or flame shape issues at start-up have been noted. While some of these have been due to hardware design or fabrication errors, most have been solved by removing construction debris from the burner fuel gas gun. In one large Tail Gas Incinerator employing two identical burners, both fuel gas guns were found to contain two beverage cans inserted by unconcerned installation personnel. A similar problem with fuel gas pilot operation is even more common, probably because the fuel gas injector in many pilots requires a very small fuel gas passage. One user reported that Tail Gas Incinerator temperature could be controlled at target level during initial heat-up, but upon introduction of tail gas only a reduced temperature could be maintained despite increased fuel flow. No carbon monoxide monitoring was available, but stack O₂ levels were adequate. The problem was resolved when the fuel gas guns were redrilled to "collapse" the fuel gas fan pattern. The jets of fuel gas had been entering the tail gas stream prematurely, quenching the primary combustion reactions, which never recovered due to the low furnace temperature (1100^oF in this case.)

Burner Problems: One operator reports problems with flame stability and relight capability in his early 90’s tail gas incinerator. The furnace is horizontal and fired with three forced draft fuel gas burners mounted in a triangle on the furnace end wall. The pilot burners were designed for natural draft operation and draw their air from outside the burners via fuel gas jet venturis. The furnace exhaust passes through a fire tube boiler, so furnace pressure could sometimes be greater than the small pressure generated by the pilot venturis. The operator solved this problem by adding "purge air" to the pilot assembly and closing the venturi air doors fully, converting to forced draft pilots. The tail gas is introduced tangentially through the cylindrical shell of the furnace a few feed from the end wall where the burners are mounted, resulting in very turbulent conditions in the flame zone. The burners manage to operate, but occasional plant upsets can spike the tail gas flow, extinguishing a burner and shutting down the incinerator. The operator prefers to relight the furnace without diverting the tail gas, but the turbulence causes problems – sometimes removing the tail gas is necessary.

Emission Problems: CO emissions have been a problem in a number of Tail Gas Incinerators. Once proper tail gas mixing has been achieved, the only fix has been to increase the furnace operating temperature. In a number of cases a CO destruction "step" has been observed in the area of 1430 – 1450^oF. Operation at 1500^oF is typically specified today (vs. the 1100 – 1200^oF needed to handle the sulfur compounds alone.)

SO₃ emissions are not as easy to solve. The goal of the Tail Gas Incinerator is to completely oxidize sulfur compounds to SO₂, but in the presence of excess oxygen part of the SO₂ is further oxidized to SO₃. For design purposes a 3% conversion of SO₂ to SO₃ is often assumed. Once the flue gas exits the stack it is cooled and the SO₃ combines with water vapor to form sulfuric acid droplets. This aerosol is visible as a blue – gray plume, usually separated from the stack tip. In many units the plume intensity varies seasonally and even from day to day. Although considerable attention has been paid to eliminating plume formation by reducing furnace O₂ content, adjusting furnace temperature, and designing for reduced furnace residence time, the plume problem remains largely unsolved.

Corrosion Problems: One plant reported operating a tail gas incinerator waste heat recovery boiler at 50 psig steam pressure. With the level of sulfur in the tail gas, too close an approach to the SO₃ dew point was suspected. On inspection, a dark gray-black wet paste was extracted from the blocked stack drain connection, indicating potential corrosion problems. The guyed stack is now being inspected for shell thickness, and steps are planned to increase the flue gas temperature to avoid future problems.

Resonance Problems: One operator reported the presence of a very low-pitched noise problem in a forced draft Tail Gas Incinerator with waste heat boiler. It is thought that slight combustion instabilities (burner noise) triggered the problem, but apparently the specific geometry of the system amplified the noise to unacceptable levels. The problem was finally solved after a noise expert recommended installation of a perforated screen between the boiler outlet and the stack. This interrupted the resonance pattern and solved the problem.

Stack Design Problems: The presence of SO₃ in the Tail Gas Incinerator flue gas results in an elevated dewpoint. In many units the dewpoint can be as high as 340 – 350^oF and the liquid that condenses is sulfuric acid. Since the stack shell is carbon steel and the coolest areas in the stack are between the refractory and the shell, attention must be paid to controlling the shell temperature. A target of 400^oF is often used. In colder climates, air temperature can vary from minus 40^oF to +90^oF over the year. Wind and rain can substantially increase cooling. Rainshields are always used on these stacks to control corrosion, as well as to the refractory spalling problems also found with the other combustors. A number of stack corrosion problems were reported when rainshield design, installation or maintenance were handled poorly.

Thermal Damage: One user operated a tail gas incinerator with 300’ tall stack supported with guy wires. Periodically, higher levels of hydrocarbons were sent to the unit, resulting in sudden temperature increases and thermal spalling of the refractory lining at about 50’ elevation. The external rainshield prevented detection of the problem, despite periodic thermal scanning. Eventually the stack began to bend, forcing immediate shutdown and installation of a derrick support to the 250’ level (along with repair of the refractory lining.) This user recommends the use of "skin" thermocouples to detect refractory loss under a rainshield.

Acknowledgements:

My thanks to the Claus plant operators, plant design engineers, hardware suppliers and consultants who agreed to help with this work. In particular, thanks to Callidus Technologies and Goar, Allison & Associates for permission to use the equipment drawings.

Return to Banks Engineering home page