Reactor Containment and Heat Loss
Gasification is a process that is carried out under harsh conditions. Even in a steam methane reforming furnace where the feedstock, natural gas, is clean and the use of catalysts allows syngas generation at 850-900°C, special alloys are required to work at their limit. In a gasification reactor the temperatures can be much higher—up to 1500-1600°C—making the environment subject to chemical attack from slag and the pressures are often higher.
The simplest and lowest cost design of a reactor wall is a lining with a refractory capable of withstanding the prevailing temperature and chemical conditions.
Refractory lining is used universally for partial oxidation of petroleum residues and natural gas. Typically, the design consists of three layers, as shown in Figure 6-5. The inner “hot face” layer is a high-quality corundum brick (>99% A1203) suitable for temperatures up to about 1600°C. The intermediate layer is a castable bubble alumina, and the outer “cold face” is a silica firebrick with good insulation properties. This three-layer design combines the properties of high temperature resistance and good insulation. At the same time, it hinders the propagation of cracks, which may arise in the hot face through to the vessel shell. The design is selected to ensure that no condensation takes place on the inner wall of the steel shell, which would for a pressurized reactor typically have an operating temperature somewhere between 200 and 300°C. In locations with extremes of temperature, this requires care to ensure that the wall temperature does not fall below the dewpoint of the synthesis gas in a cold winter wind, but the maximum allowable wall temperature is not exceeded in summer.
In oil service the principle source of chemical attack is from the vanadium content in the feedstock. In normal operation this is not a major concern, since in the reducing atmosphere the vanadium is present as V203. Care must be exercised, however, during start-up or shutdown to avoid significant quantities of V205 being formed, since at temperatures above 700°C this is liquid and penetrates the refractory very quickly, causing a breakdown of the bonding matrix. With suitable operating procedures, refractory lifetimes of between 25,000 and 40,000 hours are possible.
For coal, however, the nature of the ash creates a very different situation. The large quantities of silica in most coal ashes would break down an A1203 hot face in a very short time (weeks rather than months). In addition to chemical attack, the refractory is also subject to erosion by the liquid slag flowing down the wall, although this only makes a minor contribution to the refractory wear. The solution currently employed is to use chromium oxide and/or zirconium oxide-based refractories, which have a better chemical resistance to the specific atmosphere prevailing. Nonetheless, this still cannot be considered satisfactory. Lifetimes are reported of 6-18 months (Clayton, Stiegel, and Wimer 2002), and considering that replacement of a refractory lining requires three to four weeks offline, a 25,000-hour life as with oil gasifiers must be achieved. A problem with all hot-face bricks is that for corrosion
and erosion reasons these fusion cast materials should ideally be mono-crystalline, but because of thermal requirements during start-up and shutdown, poly-crystalline materials have to be used. The latter materials are less corrosion — and erosion-resistant at the crystal boundaries but have the advantage that they are more resistant to spalling.
Research into improved linings is being conducted as described by Dogan et al. (2002). In her paper Dogan describes the mechanism of liquid and vapor-phase penetration of silica, calcium oxide, and alumina into the matrix of the refractory. Subsurface swelling occurs and subsequent cracks develop parallel to and about 1 to 2 centimeters below the surface of the brick. While these cracks are developing, wear rates of about 0.003-0.005 mm/h can be expected, but when the cracks reach the edge of a brick, then there is a sudden loss of the whole material in front of the crack. Dogan then describes the development of a phosphated chromium oxide refractory with better resistance to liquid penetration. It is anticipated that test panels of this modified refractory may be installed into a commercial reactor late in 2003.
Fluid-bed coal gasifiers also have an insulating brick wall comprising dense erosion-resistant bricks and insulating bricks. Temperatures can be as high as 1150°C. Although there is no liquid slag present, there is mechanical erosion from the ash, limestone (for sulfur removal), and sometimes the heat carrier, that are circulated at high velocities in these gasifiers. The shape of the CFB and transport gasifiers is more complex (see Section 5.2); the construction in general and the domed and vaulted “roofs” in particular must be designed so as to keep their integrity over the whole temperature range, from ambient to the gasification temperature.
Biomass gasification, which is virtually always carried out in a fluid-bed, often in the presence of sand as a heat carrier, is performed at the lowest temperatures. This low temperature of 900-1050°C is often determined by the ash quality of the biomass rather than by the intrinsic reactivity towards gasification per se. Biomass ashes have relatively low softening and melting points and when molten are extremely aggressive in terms of corrosion owing to the high salt content.
The alternative to refractory linings is a water-cooled membrane wall construction such as that shown in Figure 6-6. The design shown is that of the Noell reactor, but it is typical also of other entrained-flow slagging gasifiers such as SCGP and Prenflo.
The membrane wall consists essentially of high-pressure tubes, in which steam is generated, connected by flat steel bridges of which the width is about equal to the outer diameter of the tubes. Tubes and bridges are welded together into a gas-tight wall. The tubes are provided with studs that act as anchors for a thin layer of castable refractory, usually silicon carbide. During operation of the gasifier the castable will ideally be covered by a layer of solid slag, over which the liquid slag will run to the bottom of the reactor (see Figure 6-6). In principle, the castable is not required, as it mainly acts as a “primer” to which the slag can adhere. There is a chance, though, that without this “primer” the coverage of the tube wall with slag may be erratic.
|
Figure 6-6. Membrane Wall (Source: Lorson, Schingnitz, and Leipnitz 1995) |
In the ideal situation, the first liquid slag hitting the wall solidifies and forms a layer of solid slag on the castable. The elegance of a membrane wall is that the liquid slag then only comes in contact with solid slag, and hence no corrosion or erosion of the reactor wall takes place. Small temperature excursions will cause the boundary between solid and liquid slag will move a little, but in principle the solid slag layer seals the whole wall. Such a wall is self-repairing. Hence the membrane wall is very robust and has a long life. A service life of eight years or more can be achieved.
One drawback is that the heat loss through the reactor amounts to 2-4% of the heating value of the coal, whereas with an insulating brick wall this heat loss is less than 1 %. In the case of a membrane wall, the heat loss is mainly determined by the radiant heat of the reactants and the total surface area of the gasifier reactor.
Another drawback is the high cost of a membrane wall. The wall itself is already expensive, but for constructional and maintenance reasons, these reactors are built with a space between the membrane and the steel outer shell of the reactor, and therefore the cost of penetrations for burners and instruments is also high. This space must also have an open connection with the gasification space, as the membrane wall is not built to stand pressure differences across the wall.
For control purposes, it is a great help if the steam production of the membrane wall can be accurately measured, as the heat loss through the wall will then be known. This heat loss is an important variable in reactor simulations and therefore for reactor control reasons. Moreover, the steam make is an important indicator for the reactor temperature. Its accuracy is however, heavily influenced by the state of the refractory/slag covering of the wall.
The use of a steam-generating water jacket is a well-proven solution in the context of the Lurgi dry-bottom gasifier and the Koppers-Totzek gasifier. The Noell technology also uses a jacket for applications with low ash feedstocks (Schingnitz et al. 2000). Internal jackets are an elegant and low-cost solution for protecting the pressure shell from high temperatures. The space within the jacket should be in open communication with the gasifier proper, as the internal wall of the jacket cannot withstand pressure differences. This (steam) connection may be located well downstream of the gasifier. Advantageously, the connection is, for example, made before a CO shift in case synthesis gas has to be produced or after the gas cleaning section in case the gasifier is part of an IGCC. The steam from the jacket can also be used as (part of) the quench gas. For slagging gasifiers, the hot inner wall of the jacket has to be protected on the gas side with a castable that must be anchored with studs.
The quality of the steam produced in a jacket is rather low, as the pressure of the saturated steam has to be equal to the pressure in the reactor, whereas in the tubes of a membrane wall, saturated steam of 100 bar can be produced. However, the jacket construction is not only lower in cost (not least because the vessel diameter can be up to a meter smaller), but also wall penetrations for such things as burners and instruments are simpler than for reactors with a membrane—or an insulating brick wall, because the wall is only 10-15 cm thick instead of 60-70 cm.
Heat Loss Calculations at a Membrane Wall
Heat losses through a brick lined reactor wall can be calculated easily. This is not the case for a membrane or jacket wall where liquid and solid slag layers cover the wall (Reid and Cohen 1944). The companion website includes a program to calculate the heat loss through all types of vertical reactor wall.
In the calculations, steady-state conditions have been assumed where the heat is flowing in a horizontal direction. The latter assumption is approximately correct, as the vertical heat flow is virtually limited to the flow of sensible heat contained in the liquid slag flowing down the vertical wall. In the calculations it has been further assumed that both the slag and the castable have a fixed melting point rather than a melting range. When the proper input data are used, good approximations can be obtained. For details about the calculation methods used in the program, the reader is referred to the help files associated with the program on the website.
Six different situations on a membrane wall may be distinguished, which are illustrated in Figure 6-7. Some of the conclusions of the calculations made with the computer program are discussed in the following.
1. Conditions where the refractory is at its melting point or can react with gaseous components should be avoided at all times, as the solid slag will then not adhere to it and its function as a “primer” for the subsequent slag coat is lost. Most important for a membrane wall is that the whole wall of the reactor is covered with solid slag. Without such a slag layer the wall will have a large heat loss, as the membrane is only protected by a thin layer of castable that has a relatively high heat conductivity owing to the use of SiC and the steel studs by which it is anchored to the membrane wall.
|
Figure 6-7. Conditions of a Membrane Wall |
2. The above has consequences for the gasifier start-up procedure and operation if coal with a very high ash-melting point of, for example, 1650°C is used. The reactor must be started with a temperature well above the 1500°C required for gasification until the wall has been covered with a slag layer. This results temporarily in a less than favorable set of reaction conditions: a higher oxygen consumption and a lower CGE and gas make. Once the slag layer has built up, the temperature may be lowered and the operation will become nonslagging. Although this may result in a lower heat loss, the advantage of a slagging gasifier is that most of the ash in the coal turned into inert slag is lost. For this reason it is more attractive to add flux to the coal in order to lower the ash melting point and ensure a slagging operation.
The use of a thicker layer of castable on the membrane wall is not an alternative to using a slag layer for insulation under these circumstances, as the refractory may either melt or react with the gaseous reactants.
3. The design heat loss of the reactor will always be based on the ideal wall where a layer of solid slag covers the refractory, and this in turn is covered by a layer of liquid slag. Decreasing the melting point of the slag by adding fluxing material to the coal will always result in a lower heat loss and will make it possible to run with the ideal wall condition with almost any coal.
When processing feedstocks with a low ash content, the heat loss through the wall will only increase marginally although it will take longer to build up the layer of solid slag. The same holds for slags with a low viscosity. Low viscosities will result in a thicker layer of solid slag and a thinner layer of liquid slag.
4. The layer of liquid slag depends on how much slag reaches the wall. An increase in this slag flux can be accomplished by introducing the reactants in such a way into the cylindrically shaped reactor that the slag will preferentially be deposited on the wall, for example, by giving the reactants some swirl upon leaving the burner. Care should be taken to ensure that this cyclonic motion does not result in a countercurrent flow in the center of the reactor, as this may have undesirable side-effects at the reactor outlets. Some swirl is also favorable for a good carbon conversion, since this will also increase the residence time (see Section 5.3.8).
5. The conditions in the reactor often make it ideal for reducing iron compounds present in the ash to liquid iron. For the vertical wall of the reactor this is not much of a problem, but in the bottom of the reactor this may lead to a layer of molten slag floating on top of molten iron. This situation is similar to that encountered in the bottom of a blast furnace. Generally, the geometry of the bottom of the reactor is such that the iron flows out of the reactor together with the molten slag. The iron is then found in the slag as small lumps that gives the slag particles a rusty appearance.