A sulfur recovery unit is usually a downstream secondary plant and operates to remove pollutants so that a primary plant, oil refinery, gas treatment plant, etc. can be profitably operated. It must operate reliably, efficiently and not suffer failures that slow or shut down the primary plant. The sulfur reaction furnace is the heart of an SRU and must be carefully monitored. It is partially burning poison gas at over 2000°F, is potentially very hazardous and can be the cause of large economic losses. The SRU (and the thermal reactor) must run reliably and efficiently, without shutdowns.
The sulfur reaction furnace must be operated hot enough to efficiently carry the reactions forward, yet not so hot that its refractory lining or the associated waste heat boiler are damaged. The primary reaction rate in a Claus thermal reactor is very fast and its temperature can rapidly increase hundreds of degrees during an upset. This balancing act is made even more precarious by the addition of supplemental oxygen. The reaction gases (at 2000 to 3000°F) are turbulent, extremely active, and highly corrosive.
The simplest and most reliable way to measure temperature in a Claus thermal reactor is with a thermocouple. This is because a thermocouple output signal is an intrinsic value directly representing absolute temperature. It is not one inferred from other physical changes, which are themselves variable and must be interpreted. The signal does not change with ambient temperature, or the composition of the reactor gases and it does not have to be re-calibrated during service.
A thermocouple output signal is generally accepted as being a true and correct temperature reading. This output signal is a proven standard and is so acknowledged by all international standards organizations. In the U.S., this organization is the "NIST" (National Institute of Standards and Technology).
There is little or no maintenance required for a thermocouple, after proper installation. The thermocouple itself can be removed, inspected, tested, replaced, and verified without shutting down the thermal reactor.
Most people believe that an accurate and reliable temperature measurement using a simple thermocouple certainly would be a very desirable method of measuring reactor temperature.
But everyone knows that a thermocouple will soon be destroyed by the gas stream and conditions in an operating Claus thermal reactor. Thermocouples are commonly used short term to get an accurate reference reading. These become contaminated and are only useful for a very short period of time. They are either sacrificed or discarded after a few uses.
The idea of short-term accuracy and a quick failure are absolutely correct if the thermocouple used is one of the basic designs commonly used for general-purpose high temperature industrial measurements.
Research and many years of field test experience have taught us what occurrences cause basic thermocouples to read in error and to rapidly fail in this application. Noble metal platinum thermocouples are the only recognized basic types that can be successfully used over the full Claus reactor range of 1800 to 3100° F. They also have many decades of proven reliability in Claus Thermal Reactor service.
Platinum thermocouples are destroyed if they are inserted directly into a Claus thermal reactor. Failure is not due to the temperature, but to corrosion, elemental absorption and general contamination. Logically then, if the thermocouple can be protected from the internal reactor environment, then it will accurately and reliably measure the reactor temperature for long periods of time. That is the goal and here’s how it is achieved.
The first step is to place a physical barrier between the thermocouple wires and the reactor gases. This is accomplished by enclosing the wires inside a custom grade ceramic tube which does not itself contaminate the wires. This prevents direct contact between the gases and the wires.
The barrier tube and T/C must be supported in the reactor. The reactor is under pressure, so the tube must be attached to a steel pressure flange, which is mounted gas-tight to the pressure containing steel outer shell. Claus thermal reactor design conditions are typically 14 PSIG (94 PSIG in Europe) at 650°F.
The tube + T/C must be inserted through a hole in the refractory liner and into the reaction chamber. A hole is drilled through an existing refractory liner. The hot junction is inserted to the point where temperature measurement is required (normally face of the refractory).
The ceramic refractory liner has a significantly different expansion coefficient from that of the steel shell. A hole to admit the tube, drilled through the refractory, on the centerline of the flanged nozzle will move due to differential expansion. It will no longer be in alignment with the nozzle centerline after a large temperature change has occurred, e.g. cold start to operating temperature. Assume that the reactor is cold. The barrier tube is sticking through the drilled hole and is in contact with the refractory. It is also attached to the steel shell by a pressure flange. As the reactor heats up, differential expansion will cause increasing misalignment of the hole with the centerline of the nozzle and the tube will soon be broken. Reactor gases will then contact the thermocouple wires and it will fail due to corrosion and contamination.
Making the hole through the refractory large enough to prevent contact between the barrier tube and the refractory wall solves this potential problem of breaking the barrier tube. This is done by boring out the above drilled hole. A mandrel is used on an unlined shell. The tube cannot now be broken due to the differential movement.
But, an oversize hole creates another problem; the large annulus will channel hot gases up to the shell and overheat it. Cooled sulphur gases will eventually precipitate into a solid. Clearances needed to allow for differential expansion will be filled by the solid and the protective tube will be broken.
To prevent this, a second larger tube is fitted into the refractory hole plugging off the hot gases. This second (outer) tube is supported only by the refractory and therefore, moves with it. The inner barrier tube and T/C is now free to move around in a large hole through the refractory without hitting the wall and being broken by thermal and mechanical shifting between the shell and the refractory. The flanged nozzle on the shell is packed with insulation to improve the measuring accuracy, limit nozzle heating and to prevent sulphur deposits from forming in the cooler nozzle area. The thermocouple is now isolated from direct contact with the reactor gases. The refractory and shell will not interfere with the barrier tube/thermocouple and an accurate temperature measurement can be made.
Reactor gases include H2S, H+, S-, N+ and a host of other elemental species. The high operating temperature of the reactor causes these species to be extremely active and to move around vigorously. Some will move through seemingly solid ceramics, particularly at seal edge boundaries. Very small amounts of these species will find their way through the tube and will come into contact with the thermocouple wires.
These atoms and molecules move into the platinum wire and take up residence inside its metallic grain structure. The contaminated thermocouple is now an alloy of platinum and other atoms. It is no longer a standard Type "R" or "B" thermocouple. Its output signal begins to decrease and it reads lower than the actual temperature inside the reactor. Erroneous low readings of over 400°F have been reported due to this phenomenon. Additional contaminants will continue to be alloyed in until the strength of the wire gets so low that it breaks due to vibration and a complete failure occurs.
Such contamination can be prevented by sweeping the infusing intruders out before they can contact and enter the thermocouple wire. This is done with a clean air or nitrogen purge. Purge gas is forced down the annular space between the thermocouple wire and the ceramic supporting it. A very small volume of purge gas flows down the support at a very slow rate and exits around the hot junction. It is heated up to reactor temperature before reaching the hot junction and does not noticeably cool it. A second larger purge flow moves down through the support, exits just above the hot junction, and does not contact or cool the hot T/C junction. The purge gas then flows upward between the thermocouple support and the protective barrier tube. The double barriers and purge flows keep any incoming contaminants swept out before they can reach and contaminate the thermocouple wire.
The internal body of the thermocouple unit is kept pressurized above that of the reactor by maintaining a purge gas supply of 3 to 5 PSI above the maximum reactor pressure. This raises the partial pressure differential and also prevents any leakage through seal boundaries. The purge gas flow rate is regulated at 0.4 to 0.5 SCFH.
There are also various compounds and elements, even in trace amounts, that will adversely effect the thermocouple and its ability to accurately measure temperature. Custom ceramics prevent contamination from this source. Various well materials are also required for certain temperature ranges, insertion positions, and insertion depths.
After proper installation, the thermocouple unit will require little or no maintenance. However, if the purge gas pressure or flow is lost for an extended period of time, then the thermocouple element may become contaminated and may need replacement.
In the event of a
mechanical failure due to improper installation, shifting refractory or thermal
shock severe enough to break the element protective well, the primary thermocouple
element would be exposed to the reactor gases. For a period of time, the operator
of the reactor may not be aware that thermocouple corrosion has begun. Thermocouple
failure can be a slow process. As the corrosion progresses, the T/C output signal
will gradually drop relative to the actual temperature. This "low"
error may allow the reactor to be operated at undetected higher and higher temperatures.
This potential error problem has been solved by a thermocouple design enhancement
that provides a "self-monitoring" feature that warns of a failing
T/C element. This design incorporates two separate independent thermocouples,
an operating thermocouple and a reference thermocouple. The operating one is
installed in the conventional position inside the element well and is constantly
flushed with N2 or clean, dry instrument air and provides the fast response
needed for operations. The reference T/C is installed adjacent to the operating
T/C, but it is embedded in a solid ceramic-blend sealant material. It is not
exposed directly to the flush gas or to the reactor gases if the thermowell
becomes damaged. Because the reference T/C is embedded in solid ceramic material,
it is slower to respond than the operating T/C element and therefore it is not
suitable for use as the primary measurement.
As the exposed operating element becomes corroded, its output signal wil slowly begin to drop relative to the actual reactor temperature. On the other hand, the reference T/C, which is encased in solid ceramic, is shielded and therefore becomes contaminated at a far slower rate. Therefore, in the event of a thermowell failure, the two output signals would show an increasing deviation from each other. This deviation in output signals provides a warning that corrosion has begun on the primary thermocouple element and that it is necessary to plan to use alternate temperature monitoring methods, such as other thermocouples, infrared or even analyzers.
Another new thermocouple development is the addition of a third thermocouple element, which is used as a "refractory dryout" temperature sensor. (Fig. 3) This very accurate thermocouple element, usually a "T" type, is used to measure the critical low temperatures (usually 658°C to 300°C) required to properly cure the newly installed refractory material. This additional thermocouple element provides an easy means to accurately monitor this temperatures without having to install additional thermocouples or to change wiring as the reactor is brought up to operating temperature. As the "dryout" is completed and the reactor operating temperatures are reached, this thermocouple element burns up without damaging the other elements. The advantage of this third thermocouple is that it permits accurate control of the refractory dryout cycle and reduces costs during the startup period. With the new thermocouple, all three, the primary, secondary and the dryout T/C's are all permanently wired through the assembly's conduit head, eliminating the need to install temporary wiring for "dryout". This simplifies the installation and avoids the need to later remove the exposed wiring under operating conditions.
INSTALLATION ISSUES
Improper installation practices have been identified as the primary reason that the otherwise suitable thermocouple designs sometimes suffer a relatively short life. The main problem with thermocouple installations has been the creation of the "bore-hole" through the refractory below the mounting nozzle. An installation "mandrel" can be used to create the proper bore size, centering the hole in the nozzle and maintaining the alignment of the hole relative to the mounting flange. The mandrel is bolted onto the mounting flange prior to installing the refractory; the refractory is then installed in the normal manner around the protruding tube that extends down into the reactor. An adjustable "Stop-Disc" that is used to keep refractory castable and mortar from entering the base of the nozzle and upsetting the critical dimensions that are required for successful thermocouple installation. As soon as the refractory sets, the mandrel is removed, leaving a proper bore hole for the thermocouple installation
In recent years, refractory installers have preferred to drill the bore hole through the refractory instead of mounting the mandrel, which requires that the refractory brick be cut and fitted around the mandrel pipe.
But, it has been found that creating a "drilled" hole that is on-center, straight, uniform and of the proper bore size, is very difficult to achieve. Boring an incorrect hole through the refractory is very likely to cause the thermocouple assembly to become broken by shifting refractory, which will result in early failure. Also, once an improper hole is cut in the refractory, it is very difficult, expensive and time consuming to correct it.
To assure proper drilling of the refractory, a new "Drill-Guide" is now available. The "drill-guide" is mounted to the nozzle flange. The refractory core drill is inserted down inside the guide pipe, which extends down into the nozzle to within approximately 25-mm of the refractory. The guide pipe is slightly larger than the drill size, which is typically 2.25 inches (approximately 60-mm). At the top, a drill shaft centering guide is adjusted to help keep the drill bit aligned so that a straight perpendicular bore hole, located in the exact center of the nozzle is created. Proper positioning of the hole insures that shifting refractory will not place high stress loads on the protective well.
When the installation is to be made in new refractory, the "stop-disc" is used to keep the castable refractory or mortar from entering the nozzle.
After proper installation, the only operating requirement is that N2 or clean dry instrument air flush gas pressure and flow is maintained through the thermocouple. To assure that the necessary flush gas hardware is provided and installed in the proper manner, a pre-piped flush gas panel is available. This panel not only provides the necessary components to measure and control the pressure and flow of the flush gas, but also assures that the proper piping and connections are in place.
SUMMATION
The interior of a Claus thermal reactor is an extremely hot, corrosive and unfriendly place. An accurate temperature measurement of its interior is vitally important for efficient and safe operation. It is also imperative that the temperature measurement be continuously reliable. A thermocouple, properly designed and installed, will reliably do this job and at a relatively low total cost. This has been proven in hundreds of installations, worldwide, for over 30 years.
Go to Model HTP Thermocouple Page
Go to Delta Controls Home Page
Go to Delta Controls Literature Page