126777789 ammonia production plants

7/29/2019 126777789 Ammonia Production Plants

1/36

2 01 2 Wi le y- VC H V er la g G mb H & C o. K Ga A, W ei nh ei m

Ammonia, 3. Production Plants

MAX APPL, Dannstadt-Schauernheim, Germany

1. Complete Ammonia Production Plants . 227

1.1. Steam Reforming Ammonia Plants . . . . 228

1.1.1. The Basic Concept of Single-Train Plants. 228

1.1.2. Further Developments . . . . . . . . . . . . . . . 229

1.1.3. Minimum Energy Requirement for Steam

Reforming Processes . . . . . . . . . . . . . . . . 231

1.1.4. Commercial Steam Reforming Ammonia

Processes . . . . . . . . . . . . . . . . . . . . . . . . 2341.1.4.1. Advanced Conventional Processes . . . . . . 2341.1.4.2. Processes with Reduced Primary Reformer

Firing . . . . . . . . . . . . . . . . . . . . . . . . . . . 2371.1.4.3. Processes without a Fired Primary Reformer

(Exchanger Reformer) . . . . . . . . . . . . . . . 2401.1.4.4. Processes without a Secondary Reformer

(Nitrogen from Air Separation) . . . . . . . . 243

1.2. Ammonia Plants based on Partial

Oxidation. . . . . . . . . . . . . . . . . . . . . . . . 243

1.2.1. Ammonia Plants Based on Heavy

Hydrocarbons . . . . . . . . . . . . . . . . . . . . . 243

1.2.2. Ammonia Plants Using Coal as Feedstock 247

1.3. Waste-Heat Boilers for High-Pressure

Steam Generation . . . . . . . . . . . . . . . . . 248

1.4. Single-Train Capacity Limitations Mega-Ammonia Plants. . . . . . . . . . . . . . 250

2. Modernization of Older Plants

(Revamping) . . . . . . . . . . . . . . . . . . . . . 251

3. Material Considerations for Equipment

Fabrication . . . . . . . . . . . . . . . . . . . . . . 252

References . . . . . . . . . . . . . . . . . . . . . . . 254

1. Complete Ammonia ProductionPlants

The previous sections mainly considered the indi-vidual process steps involved in the production ofammonia and the progress made in recent years.The way in which the individual process stepsinvolved in the production of ammonia are com-bined with respect to mass and energy flow has amajor influence on efficiency and reliability. Apartfrom the feedstock, many of the differences bet-ween various commercial ammonia processes liein the way in which the process elements areintegrated. Formerly the term ammonia technolo-gy referred mostly to ammonia synthesis tech-nology (catalyst, converters, and synthesis loop),whereas today it is interpreted as the completeseries of industrial operations leading from theprimary feedstock to the final product ammonia.

The major determinant for process configura-tion is the type of feedstock, which largely

governs the mode of gas generation and purifi-cation. The other important factor is the plantcapacity, which, together with consumption and

costs of feedstock and energy, is decisive for theproduction economics. An important develop-ment was the concept of single-train plants, firstintroduced for steam reforming based productionby M.W. Kellogg in 1963 with a capacity of600 t/d. Before then maximum capacities hadmostly been about 400 t/d, with several paralleltrains in the synthesis gas preparation stage andthe synthesis loop. Today world-scale plantshave capacities of 1200 2000 t/d. The lowestcapital cost and energy consumption result when

steam reforming of natural gas is used. In addi-tion site requirements can influence the layoutconsiderably. In contrast to a stand-alone plant,ammonia production at a developed industrialsite may import and/or export steam and power,which affects the total energy consumption.

With the exception of the Koppers Totzekcoal gasification process, which operates at nearatmospheric pressure, all modern gasificationprocesses operate at elevated pressure. Steamreforming of light hydrocarbons at 30 40 barand partial oxidation of heavy hydrocarbons at40 90 bar are generally used.

DOI: 10.1002/14356007.o02_o12


2/36

1.1. Steam Reforming Ammonia

Plants

1.1.1. The Basic Concept of Single-Train

Plants

The innovative single-train concept, introducedin 1963 by Kellogg, was a technical and aneconomical breakthrough and triggered a tre-mendous increase in world ammonia capacity.No parallel lines, even for high capacity, and ahighly efficient use of energy, with process stepsin surplus supplying those in deficit, were themain features. Figure 1 shows a block flow dia-gram and the gas temperature profile for a steamreforming ammonia plant [398].

High level surplus energy is available from theflue gas of the reformer and the process gasstreams of various sections, while heat is needed,for example, for the process steam for the re-forming reaction and in the solvent regenerator of

the carbon dioxide removal system. Becauseconsiderable mechanical energy is needed todrive compressors, pumps, and blowers, it wasmost appropriate to use steam turbine drives,since sufficient steam could be generated fromthe waste-heat. As the level was high enough to

raise high-pressure steam (100 bar) it was possi-ble to use the process steam first to generatemechanical energy in the synthesis gas compres-sor turbine before extracting it at the pressurelevel of the primary reformer. Table 1 lists allsignificant energy sources and sinks within theprocess.

The earlier plants operated at deficit, andneeded an auxiliary boiler, which was integratedin the flue gas duct. This situation was partiallycaused by inadequate waste-heat recovery andlow efficiency in some energy consumers. Typi-cally, the furnace flue gas was discharged in thestack at rather high temperature because therewas no air preheating and too much of the reac-

Figure 1. Block diagram and gas temperature profile for a steam reforming ammonia plant

228 Ammonia, 3. Production Plants Vol. 3


3/36

tion heat in the synthesis loop was rejected tothe cooling media (water or air). In addition,efficiency of the mechanical drivers was low andthe heat demand for regenerating the solventfrom the CO2 removal unit (at that time aqueousMEA) was high. Maximum use was made ofdirect steam turbine drive, not only for the majormachines such as synthesis gas, process air,and refrigeration, but even for relatively smallpumps and blowers. The outcome was a rathercomplex steam system. Even after substitution

of the smaller turbines by electric motors, thesteam system in the modern plant is still acomplex system as shown in Figure 3. Ammoniaplant steam systems are described in [3, 4, 369,166169].

The first generation of the single-train steamreforming plants is discussed in [8, 164, 165, 184,185, 193, 200], and the required catalysts arereviewed in [81, 83, 204]. A survey of thedevelopment of the steam reforming conceptthrough 1972 can be found in [174]. Other re-

ferences which cover the development of thesteam reforming before the introduction of thesingle-train concept (1940 to 1960) can be foundin [3 p. 276].

The new plant concept had a triumphantsuccess story. By 1969, 30 new Kellogg largesingle-train plants with capacities of 1000 t/d ormore were in operation, and other contractorswere offering similar concepts.

The decrease in energy consumption com-

pared to the older technology was dramatic, andwith the low gas prices at that time it is under-standable that greater emphasis was placed onlow investment cost, although there was a con-siderable potential for further reducing the ener-gy consumption.

With the advent of the single-train steamreforming plants, it became standard for licen-sors and engineering contractors to express thetotal net energy consumption per tonne of am-monia in terms of the lower heating value of thefeedstock used. The total net energy consump-tion is the difference between all energy imports(mainly feedstock) and all energy exports (most-ly steam and/or electric power) expressed aslower heating value of the consumed feedstock,whereby electric power is converted with anefficiency of 25 30 % and steam is accountedfor with its caloric value.

1.1.2. Further Developments

The significant changes in energy prices from1973 onwards were a strong challenge to processlicensors, engineering contractors and plant own-ers to obtain better energy efficiency. The overallenergy consumption was reduced from around45 GJ per tonne NH3 for the first large single-train units to less than 29 GJ per tonne NH3(Table 2).

Energy saving modifications are described in

[2, 119, 175183, 205, 206, 210, 223, 230, 235,342]. For catalyst improvements see [411]. Someof the most important improvements compared tothe first generation of plants are discussed below.

Table 1. Main energy sources and sinks in the steam reforming

ammonia process

Process section Origin Contribution

Reforming primary reforming duty demand

flue gas surplus

process gas surplus

Shift conversion heat of reaction surplus

CO2 removal heat of solvent regeneration demand

Methanation heat of reaction surplus

Synthesis heat of reaction surplus

Machinery drivers demand

Unavoidable loss stack and general demand

Balance auxiliary boiler or import deficit

export surplus

Table 2. Development of thenet energyconsumptionof natural gas based steam reformingammoniaplants (real plant data) in GJ pertonneNH3

Year 1966 1973 1977 1980 1991

Plant A B C D E

Feed 23.90 23.32 23.48 23.36 22.65

Fuel, reformer 13.00 9.21 7.35 5.56 5.90

Fuel, auxiliary boiler 2.60 5.02 3.06 1.17

Export 0.55

Total 39.50 37.55 33.89 30.18 28.00

Vol. 3 Ammonia, 3. Production Plants 229


4/36

Feedstock Purification. In feedstock puri-fication, mainly desulfurization, adsorption onactive carbon was replaced by catalytic hydro-genation over cobalt molybdenum or nickel molybdenum catalyst, followed by absorption ofthe H2S on ZnO pellets with formation of ZnS.

By itself this measure has no direct influence onthe energy consumption but is a prerequisite forother energy saving measures, especially in re-forming and shift conversion.

Reforming Section. In the reforming sec-tion energy savings were achieved by several,often interrelated, measures, of which the mostimportant are the following:

. Reduction of the flue gas stack temperature toreduce heat losses to the atmosphere [82]

. Avoiding excessive heat loss by better insula-tion of the reformer furnace

. Introduction of combustion air preheating[223]

. Preheating the reformer fuel

. Increased preheat temperatures for feed, pro-cess steam and process air

. Increased operating pressure (made possibleby using improved alloys for the reformer

tubes and improved catalysts). Lowering of the steam to carbon ratio [237]. Shifting some reformer duty from primary to

secondary reformer with the use of excess air[18, 19] or oxygen-enriched air [363] in thesecondary reformer, including the possibilityof partially bypassing the primary reformer[238, 241]

. Installing a prereformer or rich-gas step is an-other possibility to reduceprimary reformer dutyand stack temperature of the flue gas [1017],

especially in LPG- and naphtha-based plants

A more recent development that breaks awayfrom the usual plant configuration is to replacethe traditional fired primary reformer with anexchanger reformer which uses the heat of theeffluent of the secondary reformer [2027, 240,241]. Also other applications have been reportedin which flue gas [240] heat from the firedreformer is used to perform a part of thereforming.

Shift Conversion. Improved LT shift cata-lysts can operate at lower temperatures to achieve

a very low residual CO content and low byprod-uct formation. A new generation of HT shiftcatalysts largely avoids hydrocarbon formationby Fischer Tropsch reaction at low vapor par-tial pressure, thus allowing lower steam to carbonratio in the reforming section (see ! Ammonia,

2. Production Processes, Section 6.1.2).

Carbon Dioxide Removal Section. In thecarbon dioxide removal section the first genera-tion of single-train plants often used MEA with arather high demand of low-grade heat for solventregeneration. With additives such as UCARAmine Guard [87, 88], solvent circulation couldbe reduced, saving heat and mechanical energy.Much greater reduction of energy consumptionwas achieved with new solvents and processes,for example BASF aMDEA or Benfield LoHeat.Other hot potash systems (Giammarco Vetro-coke, Catacarb) and physical solvents (Selexol)were introduced (! Ammonia, 2. ProductionProcesses, Section 6.1.3).

Final Make-Up Gas Purification was im-proved by removing the water and carbon diox-ide traces to a very low level by using molecularsieves. Some concepts included cryogenic pro-

cesses with the benefit of additional removal ofmethane and argon.

Ammonia Synthesis Section. In the ammo-nia synthesis section conversion was increased byimproved converter designs (see ! Ammonia, 2.Production Processes, Section 6.3.2), larger cata-lyst volumes, and to some extent with improvedcatalysts. The main advances in converter designwere the use of indirect cooling instead of quench-ing, which allowed the recovery of reaction heat

high pressure steam. Radial or cross-flow patternfor the synthesis gas instead of axial flow wasintroduced. All modern plants include installationsfor hydrogen recovery (cryogenic, membrane, orPSA technology; see ! Ammonia, 2. ProductionProcesses, Section 6.3.5).

Machinery. Developments in compressorand turbine manufacturing have led to higherefficiencies.

. Gas turbine drive for a compressor and/or anelectric generator combined with the use of thehot exhaust as combustion air for the primary



5/36

reformer or raising steam (combined cycle)[9599, 244].

. Employing electric motors instead of conden-sation turbines [245].

. Application of liquid and gas expansion tur-bines can recover mechanical work (e.g., let-

down of the CO2-laden solvent, liquid ammo-nia, purge and fuel gas).

Steam System and Waste-Heat Recovery

were improved by the following measures: in-creased pressure and superheating temperature ofhigh-pressure steam; providing a part of theprocess steam by natural gas feed saturation[19, 87, 170172]; inclusion of a steam super-heater downstream of the secondary reformer[169, 173].

Process Control and Process Optimiza-

tion. Progress in instrumentation and computertechnology has led to increased use of advancedcontrol systems and computerized plant optimi-zation. Advanced control systems [186192,338340] allow operating parameters to be keptconstant in spite of variations in external factorssuch as feedstock composition or ambient orcooling water temperatures. These systems may

be operated in open loop fashion (set valueschanged manually by the operator) or in closedfashion (set points automatically adjusted tooptimum values by using a computer model withinput of operational and economic data). Alsoplant simulation [194197] is possible by usingextensive computer models of complete plants.These models can simulate in real time thedynamic response to changes in operating para-meters, plant upsets, etc. Such systems are usedfor off-line optimization studies and for operator

training [198, 199].The above list, by no means complete, is also a

survey of options for plant revamps (Chap. 10).Quite a number of options can be found in paperspresented at the AIChE Annual SymposiumAmmonia Plants and Related Facilities givingpractical experience and presenting case stories[389].

Many of these elements are strongly interre-lated with each other and may affect differentsections of the plant concept. It is thus a demand-

ing engineering task to arrive at an optimum plantconcept, which can only defined by the condi-tions set by the feedstock price, the site influ-

ences, and the economic premises of the custom-er. An evaluation of the individual merits of thedescribed measures in terms of investment andoperational cost in a generalized form is notpossible and can be done only from case to casein real project studies.

To illustrate the forgoing discussion of theconcept of the single-train steam reforming plant,Figure 2 presents a modern low-energy ammoniaplant with flow sheet and process streams (UHDEprocess).

Figure 3 shows a simplified diagram of thesteam system. Even in such an advanced plant thequantity of steam generated from waste-heat is asmuch as 3.4 times the weight of ammoniaproduced.

1.1.3. Minimum Energy Requirement for

Steam Reforming Processes

The energy saving measures described in Sec-tion 1.1.2 have considerably reduced the demandside (e.g., CO2 removal, higher reforming pres-sure, lower steam to carbon ratio, etc.). On thesupply side, the available energy has been in-creased by greater heat recovery. The combined

effects on both sides have pushed the energybalance into surplus. Because there is no longeran auxiliary boiler which can be turned down tobring the energy situation into perfect balance,the overall savings usually could not be trans-lated into further actual reduction of the grossenergy input to the plant (mainly natural gas). Insome cases this situation can be used advanta-geously. If the possibility exists to export asubstantial amount of steam, it can be economi-cally favorable (depending on feedstock price

and value assigned to the steam) to deliberatelyincrease the steam export by using additionalfuel, because the net energy consumption of theplant is simultaneously reduced (Table 3).

A reduction in gross energy demand, that is, alower natural gas input to the plant, can only beachieved by reducing fuel consumption, becausethe actual feedstock requirement is determinedby the stoichiometry. So the only way is todecrease the firing in the primaryreformer, whichmeans the extent of reaction there is reduced.

This can be done by shifting some of the reform-ing duty to the secondary reformer with surplusair or oxygen-enriched air, although this makes



6/36

an additional step for the removal of surplusnitrogen necessary. A more radical step in thisdirection is total elimination of the fired primaryreformer by using exchanger reformers like theICI GHR and the Kellogg KRES.

Based on pure methane, it is possible toformulate a stoichiometric equation for ammoniaproduction by steam reforming:

From a mere thermodynamic point of view, inan ideal engine or fuel cell, heat and power shouldbe obtainable from this reaction. Since real pro-cesses show a high degree of irreversibility, a consi-derable amount of energy is necessary to produceammonia from methane, air and water. The stoi-chometric quantity of methane derived from theabove reaction is 583 m

3per tonne of ammonia,

corresponding to 20.9 GJ per tonne NH3 (LHV),which with some justification could be taken asminimum value. If full recovery of thereaction heat

is assumed, then the minimum would be the lowerheating value of ammonia, which is 18.6 GJ pertonne NH3. Table 4 compares the specific energy

Figure 2. Modern integrated single-train ammonia plant based on steam reforming of natural gas (Uhde process)a) Sulfur removal; b) Primary reformer; c) Steam superheater; d) Secondary reformer; e) waste-heat boiler; f) Convectionsection; g) Forced draft fan; h) Induced draft fan; i) Stack;k) HT andLT shift converters;l) Methanator; m) CO2 removal solventboiler; n) Process condensate separator; o) CO2 absorber; p) Synthesis gas compressor; q) Process air compressor; r) Ammoniaconverter; s) High-pressure ammonia separator; t) Ammonia and hydrogen recovery from purge and flash gas

1 2 3 4 5 6 7 8 9

CH4, mol % 91.24 91.24 14.13 0.60 0.53 0.65 1.16 24.34 0.12

Cn

Hm

, mol % 5.80 5.80

CO2, mol % 1.94 1.94 10.11 7.38 18.14 0.01

CO, mol % 9.91 13.53 0.33 0.40

Ar, mol % 0.28 0.25 0.30 0.37 7.21 0.01

H2, mol % 65.52 54.57 59.85 73.08 73.54 21.72 0.03

N2, mol % 1.02 1.02 0.33 23.64 20.90 25.56 24.93 46.55 0.02

NH3, mol % 0.18 99.82

Drygas, kmol/h 1713.7 534.43 5296.4 8414.2 9520.7 7764.0 8041.4 319.9 3676.6

H2O, kmol/h 3520.6 4086.1 2979.6 22.8 13.3 0.2 0.9

Total, kg/h 30213 9422 121183 199555 199555 70084 71004 6496 62626

p, MPa 5.00 0.25 3.95 3.90 3.61 3.43 3.23 0.25 2.50

T, C 25 25 808 976 229 50 35 38 20



7/36

requirement for ammonia production by steamreforming with the theoretical minimum.

Comparison of energy consumption figureswithout precise knowledge of design and evalua-tion criteria can be misleading. First of all the state

of the ammonia product should be noted. Relativeto the delivery of liquid ammonia to battery limitsat ambient temperature, the production of 3 barammonia vapor at the same temperature wouldsave 0.6 GJ per tonne NH3, while delivery asliquid at 33 C would need an extra 0.3 GJ pertonne NH3. The temperature of the availablecooling medium has a considerable influence.Increasing the cooling water temperature from20 to 30 C increases energy consumption by0.7 GJ per tonne NH3 . A detailed energy balance,

listing all imports and exports, together with thecaloric conversion factors used for steam andpower is needed for a fair comparison of plants.

The beneficial effect of energy export to the netenergy consumption is discussed above. Gascomposition is also of some importance. Nitrogencontent is marginally beneficial: 10 mol % nitro-gen leads to a saving of about 0.1 GJ per tonne

NH3, whereas a content of 10 mol % carbondioxide would add 0.2 GJ per tonne NH3 to thetotal consumption value [201 p. 263].

Energy requirements and energy saving pos-sibilities are also discussed in [2, 119, 205, 206,210, 223, 230, 235, 342].

The energy consumption figures discussed sofar represent a thermodynamic analysis based onthe first law of thermodynamics. The combina-tion of the first and second laws of thermody-namics leads to the concept of ideal work, also

called exergy. This concept can also be used toevaluate the efficiency of ammonia plants. Ex-cellent studies using this approach are presented

Figure 3. Steam system of a modern steam reforminga) Steam drum, 125 bar; b) NH3 loop; c) Turbine for syngas compressor; d) Turbine for process air compressor and alternator;e) Surface condenser; f) Condensate treatment; g) BFW pump

Table 3. Increase of plant efficiency by steam export (GJ per tonne

NH3)

Plant Difference

A B

Natural gas 27.1 32.6 5.5

Electric power 1.1 1.1

Steam export 6.4 6.4

Total energy 28.2 27.3 0.9

Table 4. Specific energy requirement for ammonia production com-

pared to the theoretical minimum

GJ per tonne

NH3 (LHV)

% theory

Classical Haber Bosch (coke) 80 90 (338 431)

Reforming, 0.5 10 bar (1953 55) 47 53 225 254

Reforming, 30 35 bar (1965 75) 33 45 139 215

Low energy concepts (1975 84) 29 33 139 158Modern concepts (since 1991) < 28 134

Stoichiometric CH4 demand 20.9 100



8/36

in [202 p. 258], [203]. Table 5 [202] comparesthe two methods. The analysis in Table 5 wasbased on pure methane, cooling water at 30 C(both with required pressure at battery limits),steam/carbon ratio 2.5, synthesis at 140 bar in anindirectly cooled radial converter.

Almost 70 % of the exergy loss in the processoccurs in the reforming section and in steam

generation. From conventional first law analysisit can be seen that almost all of the losses aretransferred to the cooling water. As the analysisassumes water in liquid state, the LHV analysesin Table 5 is not completely balanced. For aperfect balance the heat of evaporation of water(as a fictive heating value) would have to beincluded.

1.1.4. Commercial Steam Reforming

Ammonia Processes

Especially with ammonia processes based onsteam reforming technology it has become ahabit to differentiate between processes fromvarious licensors and engineering contractors.This is not so much the case for partial oxidationplants (see Section 1.2). Strong competition to-gether with increased plant size and the associ-ated financial commitment has reduced the num-ber of licensors and engineering contractors to a

few companies capable of offering world-scaleplants, often on a lump-sum turnkey basis. Insome cases these companies sub-license their

processes and special engineering know-how tocompetent engineering companies possessing noknowledge of their own in the ammonia field.There are also several smaller companies withspecific and sometimes proprietary know-howwhich specialize in revamps of existing plants or

small plant concepts.In the following, each of the commercially

most important processes is discussed in somedetail and a shorter description of economicallyless important processes is given. The processconfiguration offered and finally constructed by agiven contractor may vary considerably fromcase to case, depending on economic and siteconditions and the clients wishes. Thus plantsfrom the same contractor and vintage often differconsiderably. It is possible to categorize steamreforming plants according to their configurationin the reforming section:

1. Advanced conventional processes with highduty primary reforming and stoichiometricprocess air in the secondary reformer

2. Processes with reduced primary reformer fir-ing and surplus process air

3. Processes without a fired primary reformer(exchanger reformer)

4. Processes without a secondary reformer usingnitrogen from an air separation plant

In principle the amount of flue gas emittedshould be related to the extent of fired primaryreforming, but generalizations are questionable,because sometimes the plant layout, as dictatedby site requirements, may considerably changethe picture for the specific flue gas value.

1.1.4.1. Advanced Conventional Processes

Kellogg Low-Energy Ammonia Process

[247, 248, 342347, 364]. The Kellogg process isalong traditional lines, operating with steam/carbon ratio of about 3.3 and stoichiometricamount of process air and low methane slip fromthe secondary reformer. The synthesis pressuredepends on plant size and is between 140 and180 bar. Temperatures of the mixed feed enter-ing the primary reformer and of the process airentering the secondary reformer are raised to the

maximum extent possible with todays metallur-gy. This allows reformer firing to be reduced and,conversely, the reforming pressure to be in-

Table 5. Energyanalysis of a lowenergy ammoniaplant (GJper tonne

NH3)

HHV LHV Exergy

Input

Natural gas consumption

Reformer feed 24.66 22.27 23.28

Reformer fuel 7.49 6.78 7.08

Auxiliary boiler fuel 0.34 0.29 0.33

Total consumption 32.49 29.34 30.69

Losses

Reforming 0,38 0,38 4,94

Steam generation 0.33 0.33 2.39

Shift, CO2 removal, methanation 1.30 1.30 0.67

Synthesis 1.70 1.70 1.55

Turbines and compressors 6.50 6.50 0.54

Others (including stack) 1.30 0.68 0.46

Total losses 11.51 10.89 10.55

NH3 product 20.98 17,12 20.14

Efficiency, % 64,60 58,40 65,60



9/36

creased to some extent to save compression costs.An important contribution comes from Kelloggsproprietary cross-flow horizontal converter,which operates with catalyst of small particlesize, low inlet ammonia concentration, and highconversion. Low-energy carbon removal sys-

tems (Benfield LoHeat, aMDEA, Selexol) con-tribute to the energy optimization.

When possibilities to export steam or powerare limited, part of the secondary reformerwaste-heat is used, in addition to steam genera-tion, for steam superheating, a feature in com-mon with other modern concepts. Proprietaryitems in addition to the horizontal converter arethe traditional Kellogg reformer, transfer lineand secondary reformer arrangement, waste-heat boiler, and unitized chiller in the refrigera-tion section.

According to Kellogg 27.9 GJ per tonne NH3can be achieved in a natural gas based plant withminimum energy export, but with export of largerquantities of steam this value could probably bebrought down to about 27 GJ per tonne NH3.

Figure 4 shows a simplified flowsheet of theprocess [347] with Selexol CO2 removal systems(other options are, e.g., Benfield or BASFaMDEA).

Haldor Topse Process. In addition to tech-

nology supply, Haldor Topse also produces thefull catalyst range needed in ammonia plants.The energy consumption of a basically classicplant configuration has been reduced consider-ably by applying systematic analysis and pro-cesses engineering. Descriptions and operationalexperience are given in [82, 12, 223, 254265].

Topse offers two process versions. The firstoperates at steam/carbon ratio of 3.3 and withrather high residual methane content from thesecondary reformer. Shift conversion is conven-tional, the Benfield or Vetrokoke process is usedfor carbon dioxide removal, and the synthesispressure depends on plant size ranging between140 and 220 bar when the proprietary Topsetwo-bed radial converter S 200 is used. A sim-plified flowsheet is presented in Figure 5.

Figure 4. M.W. Kelloggs low energy processa) Feed gas compressor; b) Desulfurization; d) Primary reformer; e) Air compressor; f) Secondary reformer; g) Heat recovery;

h) High temperature shift converter; i) Low temperature shift converter; j) Condensate stripper; k) CO2 absorber; l) CO2 flashdrum; m) Recycle compressor; n) Semi-lean Pump; o) Stripper (other options are, e.g., Benfield or BASF aMDEA); p) Stripperair blower; q) CO2 lean pump; r) Methanator feed preheater; s) Methanator; t) Synthesis gas compressor; u) Dryer; v) Purge gasH2 recovery; w) Ammonia converter; x) Start-up heater; y) Refrigeration exchanger; z) Refrigeration compressor



10/36

An actual plant has reported a consumption of

29.2 GJ per tonne NH3 [263].The second version has a S/C ratio of 2.5 andshift conversion with medium- and low-temper-ature catalysts, both copper-based. For CO2removal Selexol or aMDEA is chosen. The syn-thesis is performed at 140 bar with a Topsetwo-bed S 200 radial converter, followed by asingle-bed radial S 50 converter (S 250 configu-ration). After the converters, high-pressure steamis generated. An additional proprietary item is theside-fired reformer.

For this most energy-efficient concept a figureof 27.9 GJ per tonne NH3 is claimed [262].

Uhde Process. Uhde, in the ammonia busi-ness since 1928, markets a low-energy ammoniaplant with classic process sequence and catalysts[266273, 382]. High plant reliability at compet-itive overall costs was a major objective. Aprocess flow diagram together with the mainprocess stream is presented in Figure 2.

Key features are the high reforming pressure

(up to 43 bar) to save compression energy, use ofUhdes proprietary reformer design [267] withrigid connection of the reformer tubes to the

outlet header, also well proven in many installa-

tions for hydrogen and methanol service. Steamto carbon ratio is around 3 and methane slip fromthe secondary reformer is about 0.6 mol % (drybasis). The temperature of the mixed feed wasraised to 580 C and that of the process air to600 C. Shift conversion and methanation have astandard configuration, and for CO2 removalBASFs aMDEA process (1340 kJ/Nm3 CO2) ispreferred. Synthesis is performed at about180 bar in Uhdes proprietary converter conceptwith two catalyst beds in the first pressure vessel

and the third catalyst bed in the second vessel.After each converter vessel high pressure

steam (125 130 bar, up to 1.5 t per tonne NH3)is generated (Uhde also offers its own boilerdesign in cooperation with an experienced boilermaker). Heat exchange between inlet and outletof the first bed is performed in the first vessel, andgas flow in all beds is radial. When only aminimum of energy export (steam or power) ispossible, the process heat from the secondaryreformer outlet is partly used to raise high-

pressure steam, and partly for superheatinghigh-pressure steam. Refrigeration uses screwcompressors with high operational flexibility and

Figure 5. Haldor Topses low energy processa) Desulfurization; b) Primary reformer; c) Secondary reformer; d) Shift conversion; e) CO2 removal; f) Methanation; g) Maincompressor; h) Recycle compressor; i) Heat recovery; j) Converter



11/36

efficiency. Instead of the standard synloop Uhdeoffers the Dual Pressure Process for very largecapacities (! Ammonia, 2. Production Pro-cesses, Fig. 51) (see! Ammonia, 2. ProductionProcesses, Section 6.3.2). In this variant a once-through converter at lower pressure (100 bar),

which produces about one third of the capacity,is followed by Uhdes standard loop (! Am-monia, 2. Production Processes, Fig. 50) ataround 200 bar [382, 390394].

Achieved net energy consumption is about28 GJ per tonne NH3 and Uhdes engineersexpect values of below 27 GJ per tonne NH3when a gas turbine and large steam export isincluded [271].

LEAD Process (Humphreys & Glasgow,

now Jacobs) [119, 249]. The LEAD process isa highly optimized conventional approach withsynthesis at 125 bar and two converter vessels,the first of which contains two catalyst beds withaxial-flow quenching, while the second has athird bed with small particle size catalyst andradial flow. A consumption of 29.3 GJ per tonneNH3 is claimed.

Exxon Chemical Process. The Exxon

Chemical process [275, 276], was specificallydesigned for the companys own site in Canadaand so far not built for third parties. It uses aproprietary bottom-fired primary reformer fur-nace and a proprietary hot potash carbon dioxideremoval system with a sterically hindered amineactivator. Synthesis loop and converter are li-censed by Haldor Topse A/S. Synthesis is car-ried out at 140 bar in a Topse S-200 converterand total energy consumption is reported to be29 GJ per tonne NH3.

Fluor Process. The Fluor process [2, 117,119, 246] uses the proprietary propylene carbon-ate based CO2 removal system with adsorptionrefrigeration using low level heat downstream ofthe low-temperature shift. Methanation and CO2removal are placed between the compressionstages and thus operate at higher pressure. Witha value of 32 GJ per tonne NH3 [117] this is notreally a low-energy concept.

Lummus Process. For the Lummus Processschemes [2, 119, 250, 252, 253] a consumption of29.6 [119] to 33.5 GJ per tonne NH3 [251] is

quoted. In the synthesis section either an axialflow quench converter or a radial flow converterwith indirect cooling is used. CO2 removal isperformed with a physical solvent, and there areno special features compared to other conven-tional process configurations.

Integrating the ammonia and urea process intoa single train was proposed by Snam Progetti toreduce investment and operating costs [302].

1.1.4.2. Processes with Reduced Primary

Reformer Firing

Braun Purifier Process [18, 9092, 119, 274,277, 278, 349356]. Characteristic of the low-energy Braun purifier process (Fig. 6) is thereduced primary reformer duty, which isachieved by shifting a proportion of the reform-ing reaction to the secondary reformer by usingabout 150 % of the stoichiometric air flow. Theexcess nitrogen introduced in this way is re-moved after the methanation step in a cryogenicunit known as a purifier [351], which also re-moves the methane and part of the argon. Theresult is a purer synthesis gas compared to con-ventional processes, and only minimal purgefrom the loop is required. A typical flow diagram

of this process is shown in Figure 6.Synthesis is carried out in the proprietaryBraun adiabatic hot-wall converter vessels (!Ammonia, 2. Production Processes, Fig. 49).Each catalyst bed (of which three are now usedin newer plants [350]) is accommodated in aseparate vessel with an inlet outlet heat ex-changer after the first and high-pressure steamboilers after the following. The smaller furnaceproduces less flue gas and consequently lesswaste-heat, which makes it easier to design a

balanced plant with no energy export. The lowerreforming temperature allows a reduction of thesteam/carbon ratio to about 2.75 without adverseeffects on the HT shift, because of the lessreductive character of the raw gas on account ofits higher CO2 content. In energy balanced plants,the use of waste-heat in the secondary reformereffluent is split between steam raising and steamsuperheating.

The concept shows great flexibility [352] fordesign options. It is possible, for example, to aim

for minimal natural gas consumption, even at thecost of importing some electric power. On theother hand, it is possible to improve the overall



12/36

efficiency further by exporting a greater amountof energy. In this latter case it is advantageous toincorporate a gas turbine to drive the process aircompressor. The hot exhaust (about 500 C) ofthe turbine contains 16 17 mol % of oxygenand can serve as preheated combustion air of theprimary reformer. In addition it is possible toinclude an electric generator to cover the plantdemand and export the surplus. The C. F. Braun

process can attain 28 GJ per tonne NH3 in abalanced plant, but with steam export and reali-zation of the available improvement possibilitiesa value of 27 GJ per tonne NH3 seems feasible.

ICI AMV Process. The ICI AMV process[19, 265, 279291, 415] also operates with re-duced primary reforming (steam/carbon ratio2.8) and a surplus of process air in the secondaryreformer, which has a methane leakage of around1 %. The nitrogen surplus is allowed to enter the

synthesis loop, which operates at the very lowpressure of 90 bar with an unusually large cata-lyst volume, the catalyst being a cobalt-enhanced

version of the classical iron catalyst. The proto-type was commissioned 1985 at Nitrogen Pro-ducts (formerly CIL) in Canada, followed byadditional plants in China. A flow sheet is shownin Figure 7.

In the Canadian plant, only the air compressoris driven by a steam turbine, which receives thetotal steam generated in the plant and has anelectric generator on the same shaft. All other

consumers, including synthesis gas compressor,are driven by electric motors. Separate machinesare used for make-up gas and recycle compres-sion. The make-up gas compressor is locatedupstream of the methanator to make use of thecompression heat to warm up the cold gas comingfrom the Selexol carbon dioxide scrubber.

A further key feature is that about half of theprocess steam is supplied by feed gas saturation.The synthesis converter is a three-bed designwith quench between the first two beds and an

exchanger after the second bed to raise the gastemperature of the feed to the first bed. Excessnitrogen and inerts (methane and argon) are

Figure 6. The Braun purifier ammonia processa) Sulfur removal; b) Primary reformer; c) Convection section; d) Secondary reformer; e) waste-heat boiler; f) Process aircompressor; g) Gas turbine; h) High- and low-temperature shift converters; i) CO2 removal solvent reboiler; k) CO2 absorber;l)CO2 desorber; m) CO2 stripper;n) Methanator; o) Driers; p) Purifier heat exchanger; q) Expansionturbine;r) Purifier column;s) Synthesis gas compressor; t) Synthesis converters; u) waste-heat boiler; v) High-pressure ammonia separator; w) Ammonialetdown vessel; x) Ammonia recovery from purge gas



13/36

removed by taking a purge gas stream from thecirculator delivery and treating it in a cryogenicunit operating at loop pressure. The recoveredhydrogen is returned to the circulator suction.Demonstrated efficiency is 28.5 GJ per tonne

NH3.

Foster Wheeler AM2 Process. The FosterWheeler AM2 process [238, 292], also belongs tothe group of processes that shift load from theprimary to the secondary reformer, but differsfrom the preceding concepts in that only 20 50 % of the total feedstock is treated in the tubularprimary reformer. The remaining feed is directlysent to the secondary (autothermal) reformerwhich operates with a high surplus of process

air (up to 200 %) and a rather high methane slip of2.75 % (dry basis). After conventional shift, fur-ther purification is performed by Selexol CO2removal, methanation, and molecular sieve dry-ing. A cryogenic unit operating at synthesispressure rejects the nitrogen surplus from theloop. An energy consumption of 29.3 GJ pertonne NH3 is claimed.

Humphreys & Glasgow BYAS Process.

The Humphreys & Glasgow (now Jacobs)

BYAS process [239, 289, 357] resembles theabove-described processes in its principal pro-cess features: a considerable proportion of the

feed is sent directly to the secondary reformer,bypassing the fired primary reformer; use ofexcess air in the secondary reformer; installationof a cryogenic unit as last step of make-up gasproduction to remove excess nitrogen, residual

methane, and the majority of the argon. As aconsequence the inert level in the loop can bekept rather low, with only a small withdrawal ofpurge gas. An energy consumption as low as28.7 GJ per tonne NH3 is claimed [358]. Theprocess is especially suited for revamps, where itallows plant capacity to be increased.

Jacobs Plus Ammonia Technology [242]is especially tailored for small capacities in the300 to 450 t/d range, with a load shift from

primary to secondary reformer and use of excessprocess air. To produce a stoichiometric synthe-sis gas the surplus nitrogen has to be rejected inthe final purification. This is done in a PSA unit,which receives the purge gas and part of thesynthesis gas taken ahead of the methanationstep. All nitrogen, methane, residual carbon oxi-des, and argon are adsorbed to give a stream ofpure hydrogen. Hydrogen and the remainder ofthe synthesis gas downstream of methanation aremixed to achieve a 3:1 H2:N2 gas composition,

with a lower inerts content than the synthesis gasafter methanation. The consumption figure re-ported for a totally energy-balanced plant is

Figure 7. ICI AMV processa) Desulfurization; b) Natural gas saturation; c) Process air compression; d1) Primary reformer; d2) Secondary Reformer;e) Boiler; f) High temperature shift; g) Low temperature shift; h) Selexol CO2 removal; h1) CO2 absorber; h2) Regenerator;i) Single stage compression; j) Methanation; k) Cooling and drying; l) Circulator; m) Hydrogen recovery; n) Ammoniaconverter; o) Refrigeration system



14/36

28.8 GJ per tonne NH3, and with substantialsteam export a value of 26.8 GJ per tonne NH3is claimed.

Montedison Low-Pressure Process. TheMontedison low-pressure process [119, 241,

295, 301] involves a split flow to two primaryreformers. About 65 % of the feed steam mix-ture flows conventionally through the radianttubes of a fired primary reformer followed bya secondary reformer. The balance of the feed steam mixture passes through the tubes of avertical exchanger reformer. This exchanger re-former has a tubesheet for the catalyst tubes at themixed feed inlet. There is no tubesheet at thebottom of the tubes, where the reformed gasmixes directly with the secondary reformer ef-fluent. The combined streams flow on the shellside to heat the reformer tubes in a manner similarto that described for the M. W. Kellogg KRESreformer, see Section and ! Ammonia, 2. Pro-duction Processes, Section 6.1.1). The process airflow is stoichiometric. Synthesis is performed at60 bar in a proprietary three-bed indirectlycooled converter with ammonia separation bywater, from which ammonia is then recovered bydistillation using low-grade heat. Other process

steps are conventional. As driver of the processair compressor the installation of a gas turbine issuggested with use of the hot exhaust as pre-heated combustion air for the fired primary re-former. For this process, which has been tested in

a 50 bar pilot plant, an energy consumption of28.1 GJ per tonne NH3 is claimed [241].

Kelloggs LEAP Process. In the late 1970sKellogg [2, 118, 119, 294] proposed a processwhich extends the basic idea of the concept

described above even further. The flow of thepreheated gas stream mixture is split into threestreams, with 47 % through catalyst tubes in theradiant section of the fired primary reformer,12 % through catalyst tubes in the convectionsection, and 41 % through the tubes of an ex-changer reformer heated by the effluent of asecondary reformer. It was intended to operatethe ammonia synthesis at the pressure of the frontend by using no synthesis gas compression oronly a small booster. An enormous quantity ofthe classical ammonia synthesis catalyst wouldhave been necessary, and for recovery of theammonia from the loop a water wash withsubsequent distillation was suggested, usinglow-level heat in an integrated absorption refrig-erator. A consumption below 28 GJ per tonneNH3 was calculated.

1.1.4.3. Processes without a Fired Primary

Reformer (Exchanger Reformer)

ICI LCA Process. The ICI LCA process[296300] is a radical breakaway from the designphilosophy of the highly integrated large plantused successfully for the last 25 years. Figure 8

Figure 8. ICI LCA process (core unit)a) Process air compressor; b) Start-up air heater; c) Hydrodesulfurization; d) Saturator; e) GHR; f) Secondary reformer; g) Shiftconverter; h) Desaturator; i) PSA system; j) Methanator; k) Gas dryer; l) Ammonia converter; m) Two-stage flash cooling (onestage shown); n) Chiller; o) Catchpot; p) Flash vessel q) Syngas compressor



15/36

shows a diagram of the so-called core unit whichincludes only the sections essential for ammoniaproduction (up to 450 t/d). A separate utilitysection, shown in Figure 9, supplies refrigera-tion, steam, electricity, includes cooling andwater-demineralization system, and, if needed,recovers pure carbon dioxide. Both figures showthe configuration of the first two plants built atSevernside (UK) with a capacity of 450 t/d each.

Feed gas is purified in a hydrodesulfurizationunit operating at lower than usual temperatures

and passes through a saturator to supply a part ofthe process steam, while the balance is injectedas steam. Heated in an inlet outlet exchangerto 425 C the mixed feed enters the ICI gasheated reformer (GHR) [2022] at 41 bar, pass-ing to the secondary reformer at 715 C. Theshell side entrance temperature of the GHR(secondary reformer exit) is 970 C, falling to540 C at the exit of the GHR. Methane levels atthe GHR exit and the secondary reformer are25 % and 0.67 % respectively (dry basis). Over-

all steam to carbon ratio is 2.5 2.7. The gas,cooled to 265 C in the inlet/outlet exchanger,enters a single-stage shift conversion reactor

with a special copper zinc alumina-basedcatalyst that operates in quasi-isothermal fash-ion and is equipped with cooling tubes in whichhot water circulates, whereby the absorbed heatis used for feed gas saturation, as describedabove. CO2 removal and further purification iseffected by a PSA system, followed by metha-nation and drying.

Synthesis operates at 82 bar [415] in a propri-etary tubular converter loaded with a cobalt-enhanced formulation of the classical iron cata-

lyst. Purge gas is recycled to the PSA unit, andpure CO2 is recovered from the PSA waste gas byan aMDEA wash. Very little steam is generatedin the synthesis loop and from waste gases andsome natural gas in the utility boiler in the utilitysection (60 bar), and all drivers are electric.

The original intention was to design an am-monia plant which can compete with modernlarge capacity plants in consumption and specificinvestment, and, by means of lower energy inte-gration, to achieve greater flexibility for start-up

and reduced-load operation, needing minimumstaffing. The basic plant features (GHR, isother-mal shift, and synthesis) can in principal be

Figure 9. Arrangement of core units and utility section



16/36

applied to larger capacities. The flow sheet ener-gy consumption is 29.3 GJ per tonne NH3.

Kellogg, today Kellogg Brown and Root(KBR), offers several process schemes with useof its KRES exchanger reformer and its KAAPammonia technology [2528]. With includingthe Purifier (of C.F. Braun, later Braun and Root)KBR offers now the KAAPplusProcess [371,372, 400402] shown in Figure 10.

Desulfurized gas is mixed with steam and then

split into two streams in approximate ratio 2:1.These are separately heated in a fired heater. Thesmaller of the two enters the exchanger reformerat 550 550 C, while the remainder is passeddirectly to the autothermal reformer at 600 640 C. The exchanger reformer and the auto-thermal reformer use conventional nickel-basedprimary and secondary reforming catalysts, re-spectively. To satisfy the heat balance, the auto-thermal reformer operates with surplus of air.The required heat for the endothermic reaction in

the tubes of the exchanger reformer comes fromthe gases on the shell side, comprising a mixtureof the effluent from the autothermal reformer and

the the gas emerging from the tubes. The shellside gas leaves the vessel at 40 bar. The purifierremoves the nitrogen surplus together with re-sidual methane and part of the argon.

Synthesis proceeds at about 90 bar in a four-bed radial-flow converter (! Ammonia, 2. Pro-duction Processes, Fig. 53) (hot-wall design)with interbed exchangers. The first bed is chargedwith conventional iron-based catalyst for bulkconversion and the others with Kelloggs high

activity ruthenium-based catalyst, allowing anexit ammonia concentration in excess of 20 % tobe obtained. The other process steps are morealong traditional lines. The overall energy con-sumption claimed for this process can be as lowas 27.2 GJ per tonne NH3.

The LCA and KAAPplusProcess are environ-mentally favorable because atmospheric emis-sions of both nitrogen oxides and carbon dioxideare dramatically reduced as there is no reformingfurnace.

Chiyoda Process [243]. In this process thetraditional fired primary reformer is also replaced

Figure 10. KBR KAAPplus Processa) Air compressor; b) Sulfur removal; c) Process heater; d) Automatic reformer (ATR); e) Reforming exchanger (KRES);f) Condensate stripper; g) CO2 absorber; h) Methanator; i) CO2 stripper; j) Dryer; k) Expander; l) Feed/effluent exchanger;m) Condenser; n) Rectifier column; o) Synthesis gas compressor; p) KAAP ammonia converter; q) Refrigeration compressor;r) Refrigeration exchanger



17/36

by an exchanger reformer and the heat balancerequires excess air in the secondary reformerwith the consequence of a cryogenic unit as final

step in the make-up gas preparation to remove thesurplus of nitrogen. Additionally, gas turbinesare proposed as drivers for the process air com-pressor and synthesis gas compressor with the hotexhaust being used for steam generation and feedgas preheating.

Topse has described the lay-out for an am-monia plant based on fully autothemic reforming[370].

1.1.4.4. Processes without a SecondaryReformer (Nitrogen from Air Separation)

KTI PARC Process. The KTI PARC am-monia process [87, 289, 309, 310, 312315, 348]uses the following process elements: air separa-tion unit, classical primary reformer at 29 bar,standard HT shift, power generation in a Rankinecycle with CFC to generate electric power (op-tional), carbon dioxide removal (optional, onlywhen pure CO2 product is required), PSA, nitro-

gen addition, synthesis loop. In this concept foursections of the classical process sequence (sec-ondary reforming, LT shift, CO2 removal, metha-nation) can be replaced by the fully automatichigh-efficiency PSA system, which has a propri-etary configuration (UOP) in which nitrogenflushing enhances hydrogen recovery. The over-all efficiency ranges from 29.3 to 31.8 GJ pertonne NH3.

Linde LAC Process. The Linde LAC pro-

cess [317321] consists essentially of a hydrogenplant with only a PSA unit to purify the synthesisgas, a standard cryogenic nitrogen unit, and an

ammonia synthesis loop. In principle it is similarto the PARC process, but designed for world-scale capacities. First application was for a

1350 t/d plant in India. Figure 11 compares theLAC process to a conventional one. If pure CO2is needed, it can be recovered by scrubbing theoff-gas from the PSA unit at low pressure or,probably with better energy efficiency, by instal-ling the CO2 removal unit directly in the synthe-sis gas train ahead of the PSA system. Thesynthesis converter and loop are based on ICIand Casale know-how. According to Linde theprocess should consume about 28.5 GJ per tonne

NH3 or, with inclusion of pure CO2 recovery,29.3 GJ per tonne NH3.

Humphreys & Glasgow MDF Process

(now Jacobs) [2, 93, 94, 251, 253, 308, 307,323, 324]. This concept has a configuration simi-lar to the Linde LAC process. Energy consump-tion with inclusion of pure CO2 recovery (whichis optional) is 32.8 GJ per tonne NH3.

1.2. Ammonia Plants based on PartialOxidation

1.2.1. Ammonia Plants Based on Heavy

Hydrocarbons

Although partial oxidation processes can gasifyany hydrocarbon feedstock, on account of itshigher energy consumption and investmentcosts, commercial use of this technology is re-stricted to the processing of higher hydrocarbons,

often containing as much as 7 % sulfur. Wherenatural gas is unavailable or the heavy feedstockcan be obtained at a competitive price, this

Figure 11. Comparison of Linde LAC process with a conventional process



18/36

gasification technology can be an economicchoice.

There are two commercially proven partialoxidation routes for heavy feedstocks: the Shellprocess and the Texaco process. In contrast to thesteam reforming, for which most contractors

have their own proprietary technology for theindividual process steps, the engineering firmswhich offer ammonia plants based on heavyhydrocarbons have often to rely for the individualprocess stages on different licensors. Lurgi, forexample, has built very large capacity ammoniaplants that use Shells gasification process, itsown proprietary version of the Rectisol process[412], Lindes air separation and liquid nitrogenwash, and Topses technology for synthesisconverter and loop.

Independent, experienced engineering com-panies, not directly active in ammonia plantdesign may be used as general contractors tocoordinate a number of subcontractors supplyingthe different technologies required. This is in linewith the fact that the degree of energy integrationis usually lower than in steam reforming tech-nology, because in absence of a large fired fur-

nace, there is no large amount of flue gas andconsequently less waste-heat is available. There-fore, a separate auxiliary boiler is normally nec-essary to provide steam for mechanical energyand power generation. Nevertheless, some opti-mization has successfully reduced the overall

energy consumption, for which in older installa-tions values of around 38 GJ per tonne NH3 werereported. More recent commercial bids quotevalues as low as 33.5 GJ per tonne NH3.

The arguments presented above suggest de-scribing the two principal routes, Shell and Tex-aco, which differ in the gasification process,rather than listing all individual contractor designapproaches. Figure 12 shows the classical se-quence of process steps for both cases.

Processes Using Shell Gasification (e.g.,Lurgi) [29, 3543, 50]. A cryogenic air separa-tion plant provides oxygen for the gasificationand the nitrogen for the liquid nitrogen wash andfor supplying the stoichiometric amount for thesynthesis of ammonia. Oil enters the alumina-lined gasification vessel through a central jet inthe burner nozzle. A substantial pressure drop is

Figure 12. Flow diagrams of ammonia production from fuel oil or heavy residues by the Shell (A) and the Texaco (B) process(standard configuration)



19/36

needed to ensure atomization of the oil andproper mixing with oxygen, fed through theannulus between the jet and the outer case of theburner nozzle. The temperature in the gasifica-tion vessel (generator) is between 1200 and1400 C, and the pressure between 35 and

65 bar.The hot gas contains soot, formed because of

insufficient mixing of the reactants, and fly ash. Awaste-heat boiler, a proprietary item of the Shellprocess, raises 100 bar steam and cools the gas to340 C. Soot is removed from the raw gas in atwo-stage water wash. Older installations used anelaborate technique to remove the soot from thewater by extraction with naphtha and light oil toform soot pellets which could be burnt or re-cycled to the feed oil. In newer installations thecarbon water slurry is filtered off in automaticfilters, and the moist filter cake is subjected to acontrolled oxidation in a multiple-hearth furnace.

A selective Rectisol unit with methanol ofabout 30 C as solvent is used to remove H2Sand COS (together with some CO2) to less than0.1 %. (see also ! Gas Production, 2. Process-es). The removed sulfur-rich fraction is sent to aClaus plant for recovery of elemental sulfur orconverted to sulfuric acid. The gas is heated

subsequently by heat exchange, supplied withsteam in a saturator, and then fed to shift conver-sion, which proceeds stepwise with intermediateheat removal. The gas is cooled by a direct watercooler, and the hot water is recycled to thesaturator.

A second Rectisol wash stage follows toremove CO2 by absorption at 65

C in metha-nol, which is regenerated by flashing and strip-ping. Molecular sieve adsorption then removesresidual traces of methanol and CO2. To remove

residual CO a liquid nitrogen wash is applied forfinal purification with the advantage of alsolowering the argon content in the make-up gas,which is adjusted by nitrogen addition to thestoichiometric ratio N2:H2 1 : 3. Converterand synthesis loop configuration depend on thelicensor chosen. Plant descriptions are given in[311, 316, 322, 325].

Processes Using Texaco Gasification (e.g.,Foster Wheeler, Linde, Uhde). Temperatures in

the generator are similar to those in the Shellprocess; units with operating pressures up to90 bar are in operation [9, 4449]. Some modern

installations (e.g., Linde) use pumps for liquidoxygen instead of oxygen compressors. In con-trast to the Shell arrangement, oxygen enters thegasifier through the central nozzle of the burner,and oil is fed through the annular space betweencentral nozzle and outer burner tube. Instead of a

waste-heat boiler a direct water quench is appliedfor cooling the raw synthesis gas, which is sub-sequently scrubbed first in a Venturi scrubber andthen in a packed tower to remove the soot.Texaco also offers a version operating with awaste-heat boiler instead of a water quench.Although this is preferable when producingCO-rich synthesis gases (e.g., methanol or ox-ogas), quench is thought to be more economicalwhen hydrogen-rich gases are manufactured.Soot recovery from the water is performed byextraction with naphtha. The soot naphtha sus-pension is mixed with feed oil, and the naphtha isdistilled off and recycled to the extraction stage.The shift reaction uses a cobalt molybdenum alumina catalyst [6, 7, 8486, 89] which is not

only sulfur-tolerant but also requires a minimumsulfur content in the gas for proper performance.The conversion is subdivided into stages withintermediate cooling by raising steam. The fol-lowing Rectisol process has a somewhat more

elaborate configuration than the version used inthe Shell route. The large amount of carbondioxide formed in shift conversion lowers theH2S concentration in the sour gas, and for thisreason a special concentration step is required formethanol regeneration to obtain pure CO2 and afraction sufficiently rich in H2S for a Claus plantor a sulfuric acid plant. The remaining processsteps are identical with the Shell route. Figure 13gives an example of a Linde flow diagram.Descriptions of plants using the Texaco process

can be found in [5, 7].

Topse Process. A concept using enrichedair instead of pure oxygen and methanationinstead of a liquid nitrogen wash was proposedby Topse [326, 327]. A Shell gasifier with awaste-heat boiler or a Texaco generator with aquench are equally well suited to this process.After soot removal, shift conversion is performedon a sulfur-tolerant catalyst in several bedswith intermediate cooling, leaving a residual CO

content of 0.6 mol %. An appropriate process(Rectisol or amine based) removes the sour gasesH2S, COS, and CO2, and this is followed by



20/36

methanation. Make-up gas drying, compression,and synthesis loop have no special features. Theanticipated energy consumption is 34.8 GJ pertonne NH3. A basically similar synthesis gas

preparation, but based on gasification with pureoxygen, is already used in large commercial plantin Japan [361].

Foster Wheeler Air Partial Oxidation

Process [238, 328] is a proposed modificationof the Texaco gasification process. It is intendedto operate at 70 bar with highly preheated air

(815 C) instead of pure oxygen. The synthesisgas purification train comprises soot scrubbingfollowed by shift conversion, acid gas removal

Figure 13. Ammonia production based on heavy fuel oil (Linde flow scheme with Texaco gasification)a) Air separation unit; b) Soot extraction; c) CO2 absorption; d) Methanol/H2O distillation; e) Stripper; f) Hot regenerator;g) Refrigerant; h) Dryer; i) Liquid N2 scrubber; j) Syngas compressor; k) NH3 reactorMaterial Balance



21/36

(for example Selexol), and methanation. The gasis dried and finally fed to a cryogenic unit, whichremoves the surplus nitrogen by condensationtogether with methane, argon, and residual car-bon monoxide. The rejected nitrogen is heatedand expanded in a turbine, which helps to drive

the air compressor.A major aspect in the design concept is the

separation of nitrogen and oxygen. A conven-tional air separation plant is based on the frac-tional distillation of oxygen and nitrogen, whichdiffer in boiling point by only 13 C. In thecryogenic unit for the Foster Wheeler processa lesser quantity of nitrogen is separated fromhydrogen with a much higher boiling point dif-ference (57 C). According to Foster Wheelerthis leads to considerable saving in capital in-vestment and energy consumption compared tothe traditional approach using pure oxygen froman air separation plant and a liquid nitrogen washfor gas purification. A figure of 35.6 37.6 GJper tonne NH3 is claimed for heavy oil feedstockand 31.4 32.7 for natural gas as feedstock. Asimilar variant also using air instead of pureoxygen is offered by Humphreys & Glasgow(now Jacobs) [365].

1.2.2. Ammonia Plants Using Coal as

Feedstock

In the early days the entire ammonia industry wasbased on coal feedstock. Today coal or coke(including coke oven gas) are used as feed foronly a smaller part of world-wide ammoniaproduction. In 1990, for example, only 13.5 %of the world ammonia capacity used this rawmaterial [58]. A newer statistic estimates a figure

of 19 % for 2001 [408]. But with the enormousincrease of the natural gas prices in the USA theshare of coal in the feedstock pattern of ammoniamight become larger in the future (see ! Am-monia, 1. Introduction, Chapter 8).

Apart from a few plants operating in India andSouth Africa, the majority of coal-based ammo-nia plants are found in China. Commerciallyproven coal gasification processes are Lurgi (drygasifier), British Gas/Lurgi (slagging gasifier),Winkler/HTW, Koppers Totzek, Shell, Texa-

co, and Dow, [5153, 61]. So far only the Kop-pers Totzek, Texaco, and Lurgi processes havebeen used commercially for ammonia production

[51, 5457, 59, 341, 373375, 408]. The Shellprocess, demonstrated commercially in anotherapplication with a capacity equivalent to a world-scale ammonia plant, is also a potential candidatefor ammonia production processes.

In recent years little development work has

been done on complete ammonia plant conceptsbased on coal. The traditional leading ammoniacontractors have to rely on proprietary processeslicensed from different companies, which simi-larly tend not to have specific ammonia technol-ogy of their own. Again, compared to a steamreforming plant, the degree of integration isconsiderably lower; power generation facilitiesare usually separate. Thus it is difficult to identifyspecific ammonia processes for the individualcontractors and the following descriptions serveas examples, without striving for completeness.

The Koppers Totzek Process gasifiescoal dust with oxygen in the temperature range1500 1600 C at about atmospheric pressure.For a more detailed description of the gasification,refer to ! Ammonia, 2. Production Processes,Section 6.1.1 and [6066]. The cooled gas, free ofcoal dust and fly ash, contains about 60 % of CO.The next step is compression to about 30 bar,

followed by sulfur removal at38

C with chilledmethanol (Rectisol process). Steam is added forthe shift conversion, carried out stepwise withintermediate heat removal and with a standardHTS catalyst. A second Rectisol stage, operatingat 58 C and 50 bar, removes the CO2, and thefinal purification step is a liquid nitrogen wash.Any of the well known converter and synthesisloop concepts may be used, with no purge orminimal purge, due to the practically inert-freemake-up gas. Several plants are operating in

South Africa [329, 330] and India [331].The atmospheric-pressure gasification is a

considerable disadvantage of this process route,which substantially increases equipment dimen-sions and costs, as well as the power required forsynthesis gas compression. An energy input of51.5 GJ per tonne NH3 (LHV) has been reported.According to [80], the atmospheric ACGP gas-ifier could lower the consumption to 44 GJ pertonne NH3 (HHV).

Lurgi Process. Lurgi [52, 53, 75, 76, 78,79, 77, 332] offers a concept using its proprie-tary Lurgi dry bottom gasifier, described in



22/36

! Ammonia, 2. Production Processes, Section6.1.1. The moving-bed generator, which canhandle any sort of coal (ash content may exceed30 %), operates at 30 bar, and the product gascontains up to 15 % CH4, higher hydrocarbons,troublesome phenolic material, and tars. After

washing with process condensates to removeash and dust, the gas is cooled further withrecovery of waste-heat. Several process stepstreat the separated gas liquor to recover tar,phenols, and some ammonia. Shift conversion,Rectisol process, and liquid nitrogen wash arethe further operations in the production of make-up gas. The liquid nitrogen wash produces amethane-rich fraction, which is separately pro-cessed in a steam reformer, and the reformed gasrejoins the main stream at the Rectisol unit forpurification. The gasification has a power con-sumption of 32 34 GJ per tonne NH3, andsteam generation consumes 18 22 GJ pertonne NH3, resulting in a total energy consump-tion of 50 56 GJ per tonne NH3.

The Texaco Coal Gasification Process [53,6774, 333336, 359, 360] (see ! Ammonia, 2.Production Processes, Section 6.1.1) originatesfrom Texacos partial oxidation process for

heavy oil fractions and processes a coal waterslurry containing 60 70 % coal. A lock hoppersystem removes ash and glassy slag as a suspen-sion from the quench compartment of the gener-ator. The process can handle bituminous and sub-bituminous coal but not lignite. The further gaspurification steps used to arrive at pure make-upgas correspond to those described for an ammo-nia plant using the Texaco partial oxidation ofheavy oil fractions.

Ube Industries commissioned a 1000 t/d am-

monia plant in 1984 using Texacos coal gasifi-cation process [361, 362]. An energy consump-tion of 44.3 GJ per tonne NH3 is stated, which islower than the 48.5 GJ per tonne NH3 quoted foranother Texaco coal gasification-based ammoniaplant [52].

1.3. Waste-Heat Boilers for

High-Pressure Steam Generation

Because of its great influence on reliability andefficiency of all ammonia plants a special sectionfor the generation of high-pressure steam seems

to be appropriate. The operating conditions forthe boilers are more severe than those normallyencountered in power plants; on account of thehigh pressure on both sides the heat transfer ratesand thus the thermal stresses induced are muchhigher.

In steam reforming plants, for example, thetemperature of the gas from the secondary re-former has to be reduced from 1000 C to 350 Cbefore entering the HT shift vessel. In earlierplant generations two boilers were usually in-stalled in series, with a bypass around the secondto control the inlet temperature for the HT shift.Common practice for a long time was to use aboiler with water tube design. A famous exampleis the Kellogg bayonet-tube boiler, applied inmore than 100 plants. Besause of size limitationstwo parallel units were installed. For sufficientnatural water circulation these boilers needed asteam drum at a rather high elevation and aconsiderable number of downcomers (feed wa-ter) and risers (steam/water mixture). An alter-native tube-bundle design which can directlysubstitute the bayonet-tube internals was recent-ly developed. This concept uses twisted tubes[399].

In contrast fire-tube boilers are much better

suited for natural circulation and the steam drumcan sit in a piggyback fashion right on top of theboiler. This makes it possible to provide eachboiler with its own separate steam drum, whichallows a greater flexibility in the plot plan. In afire-tube boiler, the inlet tubesheet and the tube-sheet welds are exposed to the extreme tempera-ture of the reformed gas, which creates ratherlarge temperature gradients and therefore highexpansive stress. A positive feature of the design,however, is that debris in feed water (mainly

magnetite particles spalling from the water sideof the tubes) can collect at the bottom of thehorizontally mounted vessel without creatingdiffficulties and are removed easily by blow-down. Water-tube boilers, especially bayonet-types, are very sensitive in this respect, becausethe deposits may collect in the lowest and mostintensively heated part of the tube. In an extremecase of scaling, this may restrict the water flow tothe point where boiling occurs irregularly (filmboiling). The risk is overheating and tube failure.

The key factor which allowed the use of fire-tube boilers after the secondary reformer was thedevelopment of thin-tubesheet designs. Thick



23/36

tubesheets in this kind of service are too rigid andhave too high a temperature gradient, and theresultant stress on the tube-to-tubesheet weldscan lead to cracks. The inherent flexibility of thintubesheets assists in dispersing stresses and re-duces the risk of fatigue failure of the tube-to-

tubesheet welds and tubesheet-to-shell welds. Inall the various designs of this concept, the tube-sheet is only 20 30 mm thick. The hot inletchannel and the tubesheet are shielded by arefractory layer, and the tube inlets are protectedby ferrules. In one concept (Uhde [114, 115],Steinmuller [376]) the flexible tubesheet is an-chored to and supported by the tubes to withstandthe differential pressure, which poses some re-striction on the tube length. The tubesheet ofBabcock-Borsig (today Borsig) [112, 377, 378]is reinforced by stiffening plates on the back side(Fig. 14). Both solutions have full penetrationtube-to-tubesheet welds. Steinmullers waste-heat boiler product line is now owned by Borsig.A critical analysis of the two concepts is given in[379].

In the synthesis loop boilers at the exit of theconverter up to 50 % of the total steam is gener-ated. As much as 1.5 t of steam per tonneammonia, equivalent to 90 % of the reaction heat

can be generated. For this service also fire-tubeversions have been used, including Borsigssthin-tubesheet design. But compared to the sec-ondary reformer service, where the gas pressureis lower than the steam pressure, the conditionsand stress pattern are different. In the synthesisloop boiler the opposite is the case with the resultthat the tubes are subjected to longitudinal com-pression instead of beeing under tension. Severalfailures in this application have been been re-ported [116, 380].

The thick-tubesheet concepts in various con-figurations are more generally accepted now.Proven U-type designs (Fig. 15) are availablefrom Uhde [381, 392], Balcke Durr [108, 109,

383], KBR [113], and Borsig [386, 384]. Ahorizontal synthesis waste-heat boiler was devel-

oped by Balcke Durr with straight tubes andthick tubesheets on both ends [110, 385].

A special design is Borsigs hot/cold tubesheet.The hot and cold end of the tube are arrangedalternately, so that a hot shank is always to a coldshank and vice versa. The advantage is that thetubesheet can kept below 380C [111, 369].

Waste-heat boilers in partial oxidation plants,which cool the exit gas of the generator from1400 C to around 350 C, face additional diffi-culties. The gas contains soot and probably some

fly ash particles. Very high gas velocities andappropriate design are necessary to prevent anydeposition on the heat-exchange surfaces and toreduce the danger of attrition as well. A specialdesign for Texaco gasifications (orginally bySchmidtsche Heidampf GmbH and Steinmul-ler GmbH) is offered now in an improved version(Fig. 16) by Borsig [387]. Forced water circula-tion around the entrance nozzles helps to standthe high heat flux (700 kJ m

2s1) at this loca-tion. The Shell process uses propretary designs

(Fasel Lentjes) [388].Additional information on waste-heat boilers

is found in [100, 101, 3, 4, 372]. Many papers on

Figure 14. Reinforced tubesheet of the Borsig boilera) Tubesheeet; b) Tube; c) TIG welded root pass; d) Shell;e) Supporting ring; f) Stiffener plate; g) Anchor

Figure 15. Uhde U-tube synthesis loop boiler



24/36

this subject were presented at the AIChE An-nual Symposia for Safety In Ammonia Plants andRelated Facilities [389].

1.4. Single-Train Capacity

Limitations Mega-Ammonia Plants

The principle of the economy of scale is wellknown in production processes. For ammoniaplants a scale exponent 0.7 is widely used as arule of thumb:

cost2 cost1 capacity2/capacity10:7

The advantage of larger plants is not only theinvestment but also the reduction of fixed costs,such as labor, maintenance, and overhead ex-penses. But the simple formula should be han-dled with care because it is only valid as long as agiven process configuration is scaled up and thedimensions of the equipment for very largecapacities do not lead to disproportionate priceincreases. For example there may be fewer ven-dors or special fabrication procedures mightbecome necessary. So the scale factor couldincrease for very large single-train capacities.

Nevertheless the specific investment will de-crease with the increase of capacity. On the otherhand certain process developments like autother-mal reforming and exchanger reforming couldreduce the investment for the higher capacities.

For export-orientated plants at locations withcheap natural gas there is growing interest in verylarge capacities. Several studies [368, 394,402407] have been made by various contractorsto investigate the feasibility and the size limita-tions for steam reforming plants. According to

most of these studies a capacity of 5000 t/d seemsto be possible. With a conventional processconcept for the front end there is no restrictionfor producing syngas for 5000 t/d of ammonia.Steam reforming furnaces of the required sizehave already been built for other applications(methanol) and neccessary vessel diametersseem also to be fabricable. A limitation couldbe the synthesis compressor which can probablybe designed for 4500 t/d only. The KBR KAAPprocess has there an advantage because of its low

synthesis pressure. This also holds true for theDual Pressure Process of Uhde, where justthe loop which produces only two thirds of the

Figure 16. Borsig syngas waste-heat boiler for Texacogasification



25/36

ammonia is at high pressure whereas the once-through converter is at lower pressure. Anothercapacity limiting factor is pipe diameter in thesynthesis section. The largest plant so far wasbuilt by Uhde in Saudi-Arabia. It has a capacityof 3300 t/d and came on-stream in 2006.

2. Modernization of Older Plants

(Revamping)

With rising feedstock prices and hard competi-tion in the market, many producers have lookedfor possibilities to revamp or modernize theirolder, less efficient plants so that they can staycompetitive. Most revamp projects have beencombined with a moderate capacity increasebecause some of the original equipment wasoversized and only specific bottlenecks had tobe eliminated, not entailing excessive cost. Asthe market possibilities for a company do notincrease in steps of 1000 or 1500 t/d but slowlyand continuously, such a moderate capacity ad-dition will involve less risk and will be moreeconomical than building a new plant.

For a revamp project first an updated base-lineflow sheet of the existing plant should be pre-

pared from which the proposed improvement canbe measured [207209]. Depending on the ob-jective (energy saving and/or capacity increase)the following guidelines should be kept in mind:maximum use of capacity reserves in existingequipment; shifting duties from overtaxed unitsto oversized ones; if possible, simple modifica-tions of existing equipment are preferable toreplacement; the amount of additional equipmentshould be kept to a minimum [208].

To give an exhaustive list or description of the

individual modification options is beyond thescope of this article, but reviews on this subjectand useful information are given in [210222,303306, 342, 410, 414]. Section 1.1.2 describesmodifications that lower the energy consumptionin newer plant generations compared to the firstgeneration of single-train ammonia plants, andthis also represents an overview of the revampoptions for existing steam reforming plants.

Just a few of the frequently used revamppossibilities should be mentioned here. In steam

reforming plants it is often possible to lower thesteam/carbon ratio by using improved reformingcatalysts and copper-promoted HT shift cata-

lysts. More active LT shift catalysts lower theresidual CO content, which will reduce H2 loss(methanation) and inert content in the make-upgas. Drying of the make-up gas, addition ofhydrogen recovery from purge gas, and installinga more effective CO2 removal are other options.

With the aMDEA system, which can be flexiblytailored to fit into existing process configurations,it is, for example, possible to simply replace theMEA solvent with the aMDEA solution, adjust-ing the activator concentration accordingly toachieve zero or only minor equipment modifica-tion. Also hot potash processes have been con-verted in this way to aMDEA.

Other measures, involving more additionalhardware and engineering work, are introductionof combustion air preheating and reducing theprimary reformer load. This latter option is usedwhen the revamp objective is capacity increaseand the primary reformer is identified as a bot-tleneck. One possibility is to increase the duty ofthe secondary reformer and use air in excessof the stoichiometric demand. Elegant variantsof this principle are the Jacobs BYAS process[239, 289, 357] and the Foster Wheeler AM2

process [238, 292, 293]. Another method is toperform a part of the primary reforming in a pre-

reformer [1017] that uses low-level heat. Alter-native methods to enlarge the reforming capacitymake use of the process heat of the secondaryreformer in an exchanger reformer such as ICIsGHR [236] or Kelloggs KRES [27, 413]. Ifoxygen is available, installation of a parallelautothermal reformer or a parallel Uhde CARunit[3034], (see also ! Ammonia, 2. Produc-tion Processes, Section 6.1.1) could be consid-ered. Description of executed revamp projectsare given in [358, 224229].

Similarly, numerous modernization possibili-ties exist for partial oxidation plants, and theymay even outnumber those for steam reformingplants. Common to both plant types is the poten-tial for improvement of the synthesis loop andconverter. Application of indirect cooling andsmaller catalyst particles are frequently chosen toreduce energy consumption through lower pres-sure drop, reduced synthesis pressure, higherconversion, or a combination thereof. Apart fromreplacing existing ammonia converters, in situ

modification of the internals of installed conver-ters is a very economic approach. Topse [216,231] mostly uses its Series 200 configuration; and



26/36

Ammonia Casale [101103, 232, 233, 104107]its proprietary ACAR technology, with whichmore than 90 plants have been revamped so far.With a split-flow configuration Kellogg pro-posed an in situ revamp option to change its four-bed quench converter into a two-stage inter-

cooled converter (with two parallel beds for thesecond stage) using smaller catalyst particles of1.5 3.5 mm and 3 6 mm [366, 367]. A prom-ising loop modernization option for the futurewill be the Kellogg KAAP process [28, 234, 413].

The fact that more than 45 % of world ammo-nia plants are older than 30 35 years suggeststhat there is a major potential for revamp projects,even in plants which have already made mod-ifications. Revamp histories are a constant topicat the AIChE Annual Symposium AmmoniaPlant and Related Facilities [389] and a wealthof information and practical experience can befound there.

A special revamping option is the integrationof other processes into an ammonia plant [369].For example a CO production from a side streamupstream of the HT shift is possible [395]. Hy-drogen can also be produced from a side streamby using PSA. An example of coproduction ofmethanol is shown in Figure 17.

In the less integrated partial oxidation plantscoproduction schemes are easier to be incorpo-rated and there are several large installationswhich were directly designed to produce ammo-nia, methanol and hydrogen [396, 397].

3. Material Considerations for

Equipment Fabrication

Hydrogen Attack. In several steps of theammonia production process, especially in thesynthesis section, the pressure shells of reactionvessels as well as the connecting pipes are incontact with hydrogen at elevated pressure andtemperature with a potential risk of materialdeterioration [121123].

Chemical Hydrogen Attack. Under certainconditions chemical hydrogen attack [124,125, 127, 418] can occur. Hydrogen diffuses intothe steel and reacts with the carbon that is re-sponsible for the strength of the material to form

methane, which on account of its higher molec-ular volume cannot escape. The resulting pres-sure causes cavity growth along the grain bound-aries, transforming the steel from a ductile to abrittle state. This may finally reach a point wherethe affected vessel or pipe ruptures, in most caseswithout any significant prior deformation. Thisphenomenon was already recognized and princi-pally understood by BOSCH et al. [1] when theydeveloped the first ammonia process. The resis-tance of steel against this sort of attack can be

enhanced by alloy components which react withthe carbon to form stable carbides (e.g., molyb-denum, chromium, tungsten, and others). Therate of deterioration of the material depends onthe pressure of the trapped methane, the creeprate of the material, and its grain structure. Areashighly susceptible to attack are those which havethe greatest probability of containing unstablecarbides, such as welding seams [126]. The typeof carbides and their activity are strongly influ-enced by the quality of post-weld heat treatment

(PWHT). The risk of attack may exist at quitemoderate temperatures (ca. 200 C) and a hy-drogen partial pressure as low as 7 bar.

Numerous studies, experiments and carefulinvestigations of failures have made it possible tolargely prevent hydrogen attack in modern am-monia plants by proper selection of hydrogen-tolerant alloys with the appropriate content ofmetals that form stable alloys. Of great impor-tance in this field was the work of NELSON [120,127, 128], who summarized available experi-mental and operational experience in graphicalform. These Nelson diagrams give the stability

Figure 17. Methanol coproduction (side-stream loop)

a) Secondary reformer; b) Waste-heat boiler; c) HT shift;d) LT shift; e) Purification; f) Steam drum; g) Methanol con-verter; h) Catchpot; I) Syngas compressor; j) Let-down vessel



27/36

limits for various steels as a function of temper-ature and hydrogen partial pressure. In [129,130], CLASS gives an extensive survey, still validtoday, on this subject. Newer experience gainedin industrial applications required several revi-sions of the original Nelson diagram. For exam-

ple, 0.25 and 0.5 Mo steels are now regarded asordinary nonalloyed steels with respect to theirhydrogen resistance [124].

Physical Hydrogen Attack. A related phe-nomenon is physical hydrogen attack, which mayhappen simultaneously with chemical attack. Itoccurs when adsorbed molecular hydrogen dis-sociates at higher temperatures into atomic hy-drogen, which can diffuse through the materialstructure. Wherever hydrogen atoms recombineto molecules in the material structure (at second-phase particles or material defects such as dis-locations) internal stress becomes establishedwithin the material. Th

126777789 ammonia production plants

Documents