Higman Chris Gasification (Газификация угля)

Подождите немного. Документ загружается.

300

Gasification

pressure, opportunity for waste heat integration, particularly for the solvent regen-

eration in chemical washes, and availability of the solvent and toxicity.

•

Economic boundary conditions

. In particular the depreciation rate or pay-out time

specified for a project may influence the process selection. Typically, chemical

washes will tend to require less capital investment than a physical wash, but at the

expense of a higher utility demand for solvent regeneration.

Examination of the above criteria in any particular case will probably narrow the

field down to three or four serious contenders, sometimes even less. The chart in

Figure 8-3 provides assistance in this. Selection from this short list is then generally

a matter of pure economics.

8.2.2 Absorption Systems

Absorption processes are characterized by washing the synthesis gas with a liquid

solvent, which selectively removes the acid components (mainly H

S and CO

)

from the gas. The laden solvent is regenerated, releasing the acid components and

recirculated to the absorber. The washing or absorption process takes place in a

column, which is usually fitted with (dumped or structured) packing or trays.

The absorption characteristics of a solvent depend either on simple physical

absorption or on a chemical bond with the solvent itself. This provides the basis for

the classification of AGR systems into physical or chemical washes, which have

distinctly different loading characteristics.

The loading capacity of a physical wash depends primarily on Henry’s law and is

therefore practically proportional to the partial pressure of the component to be

Partial pressure of H

S/COS and

in gas feed, [bar]

ZnO

Rectisol

content, [ml/m

]

Sulfur in product gas, [ml/m

]

0.1

10 100 10000.1

0.01

Chemical solvents

Benfield

DGA

Amisol

Sulfinol

MDEA

DEA

Physical

and

chemical

solvents

Rectisol

Purisol

Selexol

Sulfinol

Hi-pure

ZnO

Molecular

sieves

Figure 8-3. Initial Selection of AGR processes (With permission: Lurgi)

Auxiliary Technologies

301

removed (Figure 8-4). This leads to the fact that the solution rate for any particular

operating pressure is approximately proportional to the volume of raw gas to be

processed.

In contrast, the loading capacity of a chemical wash is limited by the quantity of

the active component of the solution. Once a saturation level is reached only a minor

additional loading can be achieved by physical absorption in the solution. The solu-

tion rate is approximately proportional to the volume of acid gas removed.

Some mixed solvents have been developed using both effects. These are known

as physical-chemical washes.

Generally, solvent regeneration is achieved by one of or a combination of flashing,

stripping, and reboiling. Both flashing and stripping reduce the partial pressure of

the acid component. In physical washes, reboiling raises the temperature and reduces

the acid gas solubility. In chemical washes the increased temperature serves to break

the chemical bond. In such systems the acid components are released in the same

chemical form in which they were absorbed (Figure 8-5).

An additional class of washing systems, oxidative washes, regenerate the chemic-

ally absorbed sulfur by oxidizing the active component in the solvent and recovering

the sulfur in elemental form.

Chemical Washes

Amines. Solutions of amines in water have been used for acid gas removal for over

50 years. The principle amines used for synthesis gas treatment are mono- and dieth-

anolamine (MEA and DEA), methyldiethanolamine (MDEA), and di-isopropa-

nolamine (DIPA), the latter particularly as a component of the Sulfinol solvent.

Others amines used in natural gas applications, such as diglycolamine (DGA) or

triethanolamine (TEA), have not been able to make any significant impact in syngas

applications.

Loading capacity, [kmol/m

solvent]

Partial pressure, [bar]

Physical

solvent

Chemical

solvent

Figure 8-4. Equilibrium of Physical and Chemical Absorption

302

Gasification

MDEA is the most widely used amine today. It is more selective than primary

(e.g. MEA) or secondary (e.g. DEA) amines, due to the fact that CO

is absorbed

more slowly than H

A number of proprietary formulations have been developed to address specific

issues. For example, Ucarsol was developed to reduce corrosion with high CO

loading. BASF’s aMDEA includes an activator to accelerate CO

absorption, where

selectivity is not a requirement. Variation in the degree of promotion can influence

the energy requirement for regeneration. Exxon developed the Flexsorb family of

hindered amines specifically for high selectivity.

Typical performance data of different amine washes may be seen in Table 8-1.

The flowsheet of a typical MDEA wash is shown in Figure 8-6.

RAW GAS

CLEAN GAS

STEAM OR

WASTE HEAT

OFF GAS

RAW GAS

CLEAN GAS

MP OFF

GAS

STRIPPING

GAS

LP OFF

GAS

Figure 8-5. Regeneration by Reboiling (left) and Flashing or Stripping (right)

Table 8-1

Properties of Amine Solvents

Standard

MEA

Inhibited

MEA DEA MDEA

Molecular weight 61 105 119

partial pressure, bar <100 <100

Gas purity CO

ppmv 20–50 20–50

Solution strength, wt% 10–20 30 25–35 30–50

Solution loading, mol/mol 0.25–0.45 0.4–0.8 0.8

Energy demand, MJ/kmol CO

210 140

Notes: selective

Auxiliary Technologies

303

Physical Washes

The important characteristics for any successful physical solvent are:

• Good solubility for CO

, H

S, and COS in the operating temperature range, prefer-

ably with significantly better absorption for H

S and COS compared with CO

selectivity is an important issue for the application in hand.

• Low viscosity at the lower end of the operating temperature range. Although low-

ering the operating temperature increases the solubility, the viscosity governs in

effect the practical limit to lowering the operating temperature.

• A high boiling point reduces vapor losses when operating at ambient or near ambient

temperatures.

Rectisol. The Rectisol process, which uses cold methanol as solvent, was originally

developed to provide a treatment for gas from the Lurgi moving-bed gasifier, which

in addition to H

S and CO

contains hydrocarbons, ammonia, hydrogen cyanide,

and other impurities.

In the typical operating range of −30 to −60°C, the Henry’s law absorption

coefficients of methanol are extremely high, and the process can achieve gas purities

unmatched by other processes. This has made it a standard solution in chemical

applications such as ammonia, methanol, or methanation, where the synthesis cata-

lysts require sulfur removal to less than 0.1 ppmv. This performance has a price,

however, in that the refrigeration duty required for operation at these temperatures

involves considerable capital and operating expense.

LEAN

SOLUTION

ACID GAS

SEPARATOR

ABSORBER REGENERATOR

RAW

SYNGAS

CLEAN GAS

CONDENSER

RICH

SOLUTION

REBOILER

FLASH GAS

FLASH

VESSEL

Figure 8-6. Typical MDEA Flowchart with Single Flash Stage

304

Gasification

Methanol as a solvent exhibits considerable selectivity, as can be seen in Table 8-2.

This allows substantial flexibility in the flowcharting of the Rectisol process and

both standard (nonselective) and selective variants of the process are regularly

applied according to circumstances.

As a physical wash, which uses at least in part flash regeneration, part of the CO

can be recovered under an intermediate pressure. Typically, with a raw gas pressure

of 50 bar, about 60–75% of the CO

would be recoverable at 4–5 bar. Where CO

recovery is desired, whether for urea production in an ammonia application or for

sequestration, this can provide significant compression savings.

Figure 8-7 shows the selective Rectisol variant as applied to methanol production.

The incoming raw gas is cooled down to about −30°C, the operating temperature of

the H

S absorber. Both H

S and COS are washed out with the cold methanol to a

residual total sulfur content of less than 100 ppbv. The desulfurized gas is then shifted

outside the Rectisol unit, the degree of shift being dependent on the final product.

Carbon dioxide is then removed from the shifted gas in the CO

absorber to produce

a raw hydrogen product. This column is divided into two sections: a bulk CO

removal section using flash regenerated methanol, and a fine CO

removal section in

Table 8-2

Properties of Physical Solvents

Methanol NMP DMPEG

Chemical Formula CH

OH CH

N᎐(H

C苷O

O(C

Mol. Weight kg/kmol 32 99 178 to 442

Boiling point

at 760 Torr °C 64 202 213 to 467

Melting point °C −94 −24.4 −20 to −29

Viscosity cP 0.85 at −15°C 1.65 at 30°C 4.7 at 30°C

1.4 at −30°C 1.75 at 25°C 5.8 at 25°C

2.4 at −50°C 2.0 at 15°C 8.3 at 15°C

Specific mass kg/m

790 1.027 1.031

Heat of

evaporation kJ/m

1090 533

Specific heat

at 25°C kJ/kg.K 0.6 0.52 0.49

Selectivity

at working

temperature (H

S:CO

) 1:9.5 1:13 1:9

Auxiliary Technologies

305

which hot-regenerated methanol is used. The CO

removal section operates at lower

temperatures, typically about −60°C. The permissible CO

slip is dependent on the

application. For methanol synthesis gas 1mol% residual CO

in the raw hydrogen

is quite adequate. For hydrogen production based on methanation, typically 100 ppmv

would be appropriate. For ammonia where the gas is subsequently treated in a cryo-

genic nitrogen wash, 10 ppmv would be typical.

Following the solvent circuit, we see first an intermediate H

S flash from which

co-absorbed hydrogen and carbon monoxide are recovered and recompressed back

into the raw gas. The flashed methanol is then reheated before entering the hot

regenerator. Here the acid gas is driven out of the methanol by reboiling, and a

Claus gas with an H

S content of 25–30% (depending on the sulfur content of the

feedstock) is recovered. Minor adaptations are possible to increase the H

S content

if desired.

The hot-regenerated methanol, which is the purest methanol in the circuit, is used

for the fine CO

removal. The methanol from the CO

removal is subjected to flash

regeneration in a multistage flash tower. The configuration shown is typical for the

methanol applications with only atmospheric flash regeneration. For hydrogen or

ammonia applications where better absorption is required, the final flash stage may

be under vacuum, or it may use stripping nitrogen from the air separation plant.

Finally, the loop is closed with the flash regenerated methanol returning to the H

absorber.

Water entering the Rectisol unit with the syngas must be removed, and an addi-

tional small water-methanol distillation column is included in the process to cope

with this.

REFRIGERANT

MP FUEL

GAS

CO SHIFT

METHANOL

SYNTHESIS

RAW GAS

ACID

GAS

STEAM

FLASH

TOWER

ABSORBER

HOT

REGENERATOR

S FLASH

COLUMN

S ABSORBERRECOMPRESSOR

Figure 8-7. Flowsheet of Selective Rectisol Process (Source: Weiss 1997)

306

Gasification

Typically, the refrigerant is supplied at between −30 and −40°C. Depending on

application, different refrigerants can be used. In an ammonia plant, naturally,

ammonia is used, and the refrigeration system is integrated with that of the synthe-

sis. In a refinery environment, propane or propylene may be the refrigerant of

choice.

The Rectisol technology is capable of removing not only conventional acid gas

components but also, for example, HCN and hydrocarbons. Supp (1990, p. 83)

describes a typical hydrocarbon prewash system. Mercury capture using Rectisol as

a cold trap to condense out metallic mercury is also documented (Koss, Meyer, and

Schlichting 2002).

Selexol. The Selexol process was originally developed by Allied Chemical Corpo-

ration and is now owned by UOP. It uses dimethyl ethers of polyethylene glycol

(DMPEG). The typical operating temperature range is 0–40°C. The ability to oper-

ate in this temperature range offers substantially reduced costs by eliminating or

minimizing refrigeration duty. On the other hand, for a chemical application such as

ammonia, the residual sulfur in the treated gas may be 1 ppmv H

S and COS each

(Kubek et al. 2002) which is still more than the synthesis catalysts can tolerate. This

is not an issue, however, in power applications where the sulfur slip is less critical.

Selexol has a number of references for such plants including the original Cool Water

demonstration unit and most recently the 550 MW Sarlux IGCC facility in Italy.

The ratio of absorption coefficients for H

S, COS, and CO

is about 1:4:9 in

descending order of solubility (Kubek, Polla, and Wilcher 1997). A plant designed

for, say, 1 ppm COS in the clean gas would require about four times the circulation

rate of a plant for 1 ppm H

S, together with all the associated capital and operating

costs. In a gasification environment it is therefore preferable to convert as much

COS to H

S upstream of a Selexol wash. In a plant using raw gas shift for hydrogen

or ammonia, this will take place simultaneously on the catalyst with the carbon

monoxide shift. Where no CO shift is desired, then COS hydrolysis upstream of the

Selexol unit provides a cost-effective solution to the COS issue.

Other characteristics favorable for gasification applications include high solubilities

for HCN and NH

as well as for nickel and iron carbonyls.

The Selexol flowsheet in Figure 8-8 exhibits the typical characteristics of most

physical absorption systems. The intermediate flash allows co-absorbed syngas

components (H

and CO) to be recovered and recompressed back into the main

stream. For other applications, including H

S concentration in the acid gas or separate

recovery, staged flashing techniques not shown here may be applied.

Purisol. NMP or n-methyl-pyrrolidone is the solvent used in Lurgi’s Purisol

process. The operating range is 15°C to 40°C. The selectivity for H

S/CO

extremely high and largely independent of the operating temperature (Grünewald

1989). Solvent properties are included in Table 8-2. The characteristics are in many

ways comparable with Selexol.

Auxiliary Technologies

307

Physical-Chemical Washes

Some gas-washing systems exploit the principles of both physical and chemical

absorption and are known as physical-chemical washes. They generally use an

amine together with organic physical solvent. They can usually accept a higher loading

than an aqueous amine solution, thus reducing solvent rates. Furthermore, the

organic solvents applied in such systems accelerate the hydrolysis of COS to H

S in

the lower sections of the column, thus permitting an improved total sulfur removal

performance than a pure amine system. Other aspects, which still need review when

considering a physical-chemical system, are the potential for amine degradation,

which is generally unchanged compared with the equivalent aqueous amine system.

Their effectiveness at absorbing metal carbonyls is not documented and so must be

considered as unproven.

Sulfinol. Shell’s Sulfinol solvent in its original form was a mixture of DIPA and

Sulfolane (tetrahydrothiophene dioxide). The former provides a chemical solvent

and the latter a physical solvent. Meanwhile a modified solvent, known as m-Sulfinol

has been developed that uses MDEA as the chemical component. The original Sulfi-

nol formulation has been used successfully downstream of a large number of small

RAW

SYNGAS

TREATED GAS

LEAN SOLUTION

PUMP

STEAM

REBOILER

STRIPPER

MAKE UP

WATER

REFLUX PUMP

ABSORBER

RECYCLE

COMPRESSOR

RICH/LEAN

EXCHANGER

FLASH DRUM

ACID GAS

TO SRU

ACID GAS

CONDENSER

MECHANICAL

FILTER

HYDRAULIC

TURBINE

LEAN SOLUTION

COOLER

Figure 8-8. Selexol Flowchart for Selective H

S Removal (Source: Kubek, Polla,

and Wilcher 1997)

308

Gasification

oil gasifiers for the production of oxo-synthesis gas. The AGR at the Buggenum

IGCC is an example of a larger m-Sulfinol unit.

Amisol. The Amisol process was developed by Lurgi using a mixture of MEA or

DEA with methanol. It has been applied downstream of a number of oil gasification

units, but it has not established a wide market. Details can be found in Supp (1990)

and Kriebel (1989).

Oxidative Washes

Oxidative washes or liquid redox systems differ from other types of absorption

system in that the H

S in the acid gas is oxidized directly to elemental sulfur in the

absorption stage. The active agent in the solution is regenerated in a separate

oxidizing vessel, which also serves to separate the solid elemental sulfur from the

solution. The solvents of oxidative washes absorb essentially only H

S, but not

nor COS. This makes them suitable for applications where H

S must be

removed from a stream containing large quantities of CO

, even if the H

S partial

pressure is low.

There is no known existing application in a gasification environment, but such

washes exhibit potential as a substitute for a Claus plant, where the gasifier feed has

very low sulfur content and the sour gas is unsuitable for treatment in a Claus plant.

Earlier plants, notably the Stretford and Takahax processes, used vanadium-based

agents, which undergo a valence change from the pentavalent to the tetravalent state

during the absorption stage. Modern processes, of which Lo-Cat and Sulferox are

the best known, use chelated iron formulations.

The Lo-Cat process can be arranged in a number of different application-dependant

configurations of which that shown in Figure 8-9 is typical. Acid gas enters the

absorber, where the H

S is absorbed into the aqueous chelated iron solution. The

ABSORBER

S ACID GAS

CLEAN GAS

OXIDIZER

AIR BLOWER

VACUUM BELT

FILTER

WASH WATER

EXHAUST TO

ATMOSPHERE

SULFUR CAKE

Figure 8-9. Lo-Cat Flowsheet (Source: Adapted from Nagl 2001)

Auxiliary Technologies

309

ferric iron oxidizes the HS

−

ion to elemental sulfur according to reaction 8-1. The

iron is reduced to the ferrous state.

−

+2Fe

→ S

+2Fe

(8-1)

In the oxidizer, the sulfur settles out and is transferred to a vacuum filter where it is

separated from the solution as a cake. Air is blown into the oxidizer, where oxygen

is absorbed into the solution and oxidizes the ferrous iron back to the ferric state

(reaction 8-2) for recirculation back to the absorber.

2Fe

+½O

(l) +H

O. → 2Fe

+2 OH

−

(8-2)

The raw sulfur from the vacuum filter is typically 65% to 85% sulfur, the remainder

being water and dissolved salts including iron. This product requires further treat-

ment to meet generally accepted market quality, melting (to remove the water) and

filtering being important process steps. Nonetheless, the usual “bright yellow” color

specification for commodity sulfur is not met, and specialized applications need to

be located in the marketplace.

Generally, liquid redox systems are applied for small plants, in particular where

S concentrations are lower than can be handled by the Claus process (see Section

8.4). The Stretford process was regularly applied for Claus tail gas processing as

part of the Beavon tail gas treating process. A similar application using an iron

chelate process was put into service in 2001 (Nagl 2001).

8.2.3 Adsorption Systems

A second important group of gas treatment processes are based on the adsorption

of impurities onto a solid carrier bed. Some of these processes, such as molecular

sieve driers or pressure swing, allow in situ regeneration of the bed. Others, such as

S chemisorption onto zinc oxide, cannot be regenerated economically in situ, and

the beds require regular exchange.

The quantity of a gaseous component, which can be carried by any particular

adsorbent, depends not only on the characteristics of component and sorbent but

also on the temperature and pressure under which it takes place. This increase in

loading capacity with higher pressures and lower temperatures is illustrated in

Figure 8-10 and is utilized for the in situ regeneration of such sorbents as activated

carbon, activated alumina, silica gel, and molecular sieves.

The classic adsorption-desorption cycle uses both the temperature and pressure effect

“swinging” between high pressure and low temperature for adsorption (point 1 in

Figure 8-10) and low pressure and high temperature (point 2) for desorption. The

differential loading (L

−L

) is extremely high. The pressure swing cycle operates at

constant temperature T

between points 3 and 4. A temperature swing process oper-

ating at constant pressure between points 1 and 5 is possible but unusual in practice.