SUITABILITY OF FEEDSTOCKS FOR THE SASOL-LURGI FIXED BED
DRY BOTTOM GASIFICATION PROCESS
“Gasification Technologies 2001, San Francisco, California, October 7-10, 2001”
JC van Dyk∗, MJ Keyser, JW van Zyl
Sasol Technology, R&D Division, Department of Carbon to Syngas Research,
P. O. Box 1, Sasolburg, 1947, South Africa
Abstract
Coal is a crucial feedstock for South Africa’s unique synfuels and petrochemicals industry and it is used by Sasol as
a feedstock to produce synthesis gas (CO and H2) via the Sasol-Lurgi fixed bed dry bottom gasification process.
A detailed knowledge of coal characteristics is essential to predict gasification behavior when a specific coal source
is to be gasified. The principle aim of this paper is to highlight and discuss those specific coal characteristics,
which affect gasification behavior and stability. In order to determine the suitability of a coal source for gasification
purposes, the coal is characterized and the results compared with historical data. Benchmark data, obtained when the
gasifiers are operating without problems and with relatively high stability, is used as a reference. Coal from sources
with extreme properties (e.g. ash content <10% to as high as 35% or “brown coal” with moisture content of
approximately 30%) can be gasified in a Sasol-Lurgi fixed bed dry bottom gasifier provided that certain operational
changes are implemented. Other properties, like high caking propensity for example, require blending to acceptable
levels and /or mechanical modifications. The coal characteristics discussed in this paper are not the only properties
affecting gasifier stability, but are those properties which are easily measurable on laboratory scale and can be
related to gasifier performance. Interpretation of coal characterization data gives an indication of expected gasifier
performance, and the suitability of a specific coal source for Sasol/Lurgi Fixed Bed Gasification. Data on a number
of coal sources will be discussed.
Keywords: Sasol-Lurgi fixed bed dry bottom gasification, coal characteristics, South African coal sources, other
feedstocks.
1.
INTRODUCTION
Coal is used as main feedstock for South Africa’s unique synfuels and petrochemicals industry
and it is used by Sasol as a feedstock to produce synthesis gas (CO + H2) via the Sasol-Lurgi
fixed bed gasification process. South Africa, as well as many other countries in the world, will
for many years to come rely on its abundant coal resources for energy and petrochemical
products.
The Sasol plants located in Secunda and Sasolburg (South Africa) gasify approximately 30
million tons of bituminous coal to synthesis gas, which is converted to fuels and chemicals via
the Fischer-Tropsch process. A total of 97 fixed bed dry bottom gasifiers, 17 at Sasolburg and
80 at Secunda, have a combined production capacity of approximately 5.1 x 106 m3n/h dry crude
gas, which is equivalent to approximately 3.6 x 106 m3n/h pure synthesis gas. These production
rates are well in excess of the design capacity and were achieved by continuous debottlenecking
and optimization.
∗
Corresponding author. Tel.: +27 16 960 4375; fax.: +27 11 522 4806; e-mail: johan.vandyk@sasol.com
The coal from the sources used by Sasol vary substantially in terms of chemical and physical
properties and this directly relates to gasifier behaviour. The ability of fixed bed gasifiers to
handle a variety of different feedstocks is seen as a significant advantage over other gasification
technologies. At the Schwarze Pumpe site in the former East Germany, 7 fixed bed dry bottom
Lurgi gasifiers are utilized for treatment of solid wastes, such as plastics, sewage sludge, rubber,
contaminated wood, paint residues and household wastes [10]. Other distinct characteristics of
fixed bed gasifiers are the following:
• It uses lump coal and limited grinding is required. Coal used for fixed bed gasification is
mined, crushed down to <70mm and screened at a bottom size of 5-8mm.
• Coal with a high ash content can be gasified without severe losses in thermal efficiency,
since the ash is not extracted in the molten state.
• High “cold gas” thermal efficiency is achieved through counter-current operation, which
allows the gas and solid product streams to exit at relatively low temperatures.
• Low oxidant requirements due to the high thermal efficiency.
• Valuable co-products like tars, pitches, oils and chemicals are produced.
• A H2/CO ratio of 1.7 to 2.0 is produced directly which is suitable for Fischer-Tropsch
synthesis without the need for additional water-gas shift conversion to adjust the H2/CO ratio.
The Sasol-Lurgi gasifiers has some limitations, e.g.:
• Limited ability to handle excessive fine coal or coal with a high caking propensity.
• Broad particle size distributions can lead to excessive coal segregation, which in turn may
cause channel burning and unstable gasifier operation.
• Pressure drop can limit gas throughput in certain instances.
• Relatively high steam consumption.
The principle aim of this paper is to highlight and discuss those specific coal or feedstock
characteristics, which affect gasification behavior and stability. Due to the large variation in coal
properties from various sources, detail coal and feedstock characteristics are essential to predict
gasification performance when a specific coal source is to be gasified.
2.
COAL MINING AT SASOL
The coal used by Sasol for Sasol-Lurgi fixed bed dry bottom gasification in South African has a
low rank, is inertinite rich, and has properties which may vary significantly from one mine to the
next.
Sasol Mining (Pty) Ltd. is responsible for coal mining in the Sasolburg and Secunda regions and
supplies coal to Sasol’s synthetic fuels and chemical plants. The division operates regional
operations comprising the Sigma Colliery and Wonderwater strip mining operations at Sasolburg
and the Secunda Collieries, which consist of six underground operations and a strip mine near
Trichardt.
The combined run-of-mine output from the Sasolburg and Secunda operations increased from a
total production of ±20 million tons in the ten year period 1954 – 1964, to almost 51 million tons
2
per annum in 2000. Supplies to the Sasol Chemical Industries (Sasol 1) plant in Sasolburg are
supplemented with external purchases of 0.9 million tons in 2000.
During the 2000 financial year, the company supplied 46.7 million tons of saleable coal from the
50.9 million tons of coal extracted to the operations of Sasol Synthetic Fuels (SSF) at Secunda
and the Sasol Chemical Industries (SCI) at Sasolburg. The company commenced its coal export
operations at Secunda during the 1997 financial year and produced 3.2 million tons of export
quality coal, which was exported mainly to Europe, during the 2000 financial year [8].
TABLE 1
SASOL MINING (Pty) Ltd. PRODUCTION HIGHLIGHTS [8]
1999
Production (millions of tons)
2000
50,9
49,0
Total production
Sigma Colliery including Wonderwater
5,1
5,5
Secunda Collieries:
6,5
Bosjesspruit Colliery
7,4
Brandspruit Colliery
8,7
8,6
Middelbult Colliery
9,0
8,8
Twistdraai Colliery
5,6
5,9
Twistdraai Export Colliery
6,0
5,1
Syferfontein (underground and strip)
9,1
8,6
Colliery
Saleable production from all mines
49,4
47,0
External coal purchases from other mines
0,9
0,7
Total sales excluding exports
49,9
49,0
Sales to SCI (Sasol 1), Sasolburg
6,2
6,5
Sales to SSF, Secunda
40,5
39,4
International sales
3,2
3,1
3.
COAL CHARACTERISTICS
PERFORMANCE
AND
THE
EFFECT
ON
GASIFIER
New coal sources and areas under exploration for utilization in Sasol-Lurgi fixed bed dry bottom
gasification are characterized in detail and the results compared with historical data in order to
determine the suitability of a coal source for gasification purposes. Benchmark data, obtained
when it was possible to operate the gasifier without problems and with relatively high stability, is
used as a reference.
The following tests are conducted on coal sources to determine suitability for gasification
purposes:
• Proximate analysis
• Ultimate analysis
• CO2 gasification reactivity
• Particle size distribution
• Ash melting properties and ash composition
• Caking properties under 26 bar pressure
• Thermal fragmentation (atmospheric pressure)
• Mechanical fragmentation
• Fischer Assay
3
•
•
•
Total sulphur
Heating value
Maceral analysis and rank
Data obtained on a number of coal will be discussed.
3.1
Proximate analysis
Ash content gives an indication of the amount of inorganic material in the coal from a source and
includes mineral matter inherent in the coal structure, as well as out-of-seam inorganic
contamination. The measurement of ash content is used by Sasol Mining as an operating tool to
prepare blends with an ash content within the agreed limits and with a relatively small variation
in ash content over time. The budgeted ash content of the coal blend requested by gasification at
SSF in Secunda is ±28 % (air dry basis). Gasification at the Sasolburg plant operates on a coal
blend with an ash content exceeding 32 % during certain periods.
FIGURE 1
VARIATION IN ASH CONTENT (air dried basis)
50
% Ash (average - air dried basis)
45
Secunda
Biological
sludge
40
35
30
25
20
15
10
5
0
A
3.2
Non SouthAfrican
Sasolburg
B
C
D
E
F
I
II
a
b
c
d
e
CO2 reactivity
CO2 reactivity is determined in order to get an indication of the expected rate of the gasification
reactions. Reactivity is expressed as a mass loss per time at 50% burn-off under a CO2
atmosphere.
The reactivity of coal from the various sources used by Sasol vary between 2-5 hr-1, although
coal sources with lower and higher reactivities have been gasified in the past. Some of the non
South-African coal sources tested for gasification purposes have showed reactivities of as low as
0.5 hr-1 and also as high as 9 hr-1. However, it is uncertain what the lower limit for reactivity is
below which the gasification reactions will become too slow for complete conversion.
4
3.3
Particle size distribution
Particle size distributions of coal blends are determined in order to estimate or predict which size
distributions are more likely to cause unstable operation due to pressure drop effects. Pressure
drop problems manifest themselves in a variety of ways, and include grate traction loss (due to
bed fluidization), channel burning (leading to unacceptable gas outlet temperatures) and solids
elutriation (carry-over). Probably the best known estimation method for pressure drop is the
Ergun equation, which gives pressure drop as a function of bed voidage ε , viscosity µ, fluid
density ρ, superficial velocity Us and particle diameter dp [2]:
(1 − ε ) µ U s + 1 . 75 (1 − ε ) pU
∆P
= 150
L
ε 3 d p2
ε 3d p
2
2
s
…(1)
When dealing with particle size distributions instead of uniformly sized particles, the particle
size dp has to be replaced by φ d p , where φ is the particle sphericity and d p the average particle
size reflecting the mean surface area (also referred to as the Sauter mean diameter). The Sauter
diameter of a coal sample with a specific particle size distribution is calculated as follows:
1
dp =
…(2)
 xi 
∑i  d 
 p ,i 
where i = screen number
xi = fraction (mass %) on screen i
dp,i = diameter (mm) of screen i
Experience has shown that d p is a useful parameter for predicting which PSD’s are more likely to
result in gasifier instability. Extensive research on the effect of coal types and PSD on pressure
drop and how the data fit the Ergun equation have been conducted, but will not be reported here.
The value of d p is extremely sensitive to the smaller particle sizes, or the so-called “tail” of the
PSD. As illustrated in Table 2, a 10% change to the coarser side resulted in a change of only
3% in the Sauter diameter, while a 10% change in particle size to the finer fraction resulted in a
7% change.
Extensive operating experience has shown that size distributions with a Sauter diameter below a
specific value can result in unstable gasifier operation. Inefficient screening due to screen
overload causes misplacement of fine coal, which can easily reduce the Sauter diameter to
unacceptably low values resulting in highly unstable gasifier operation.
5
TABLE 2
EFFECT OF CHANGE IN PARTICLE SIZE ON SAUTER DIAMETER
Fraction (mm)
Standard
Coarser
Finer
composition
fraction
fraction
-5.8
42.1
52.1
42.1
-3.7
33.3
23.3
23.3
-2.8
24.6
24.6
34.6
% Change in
3
7
Sauter
diameter
3.4
Thermal fragmentation
It is known that when lump coal from certain origins (e.g. South-African low rank inertinite rich
coals) is exposed to high temperatures (700oC), it will tend to undergo fragmentation (primary
and secondary fragmentation)[1]. Primary fragmentation occurs during devolatilization, while
secondary fragmentation occurs during combustion of the char by burnout of carbon bridges
connecting parts of the particle. In the case of fixed bed gasification, fine material formed in the
gasifier may lead to the same kind of hydrodynamic problems as was described previously.
Thermal fragmentation of coal is measured by placing a sample with a specific predetermined
size distribution into a pre-heated muffle oven at 100oC under atmospheric pressure [9]. The
coal is then heated to 700oC (final temperature) at a rate of ±12oC/min. The experiments are
conducted under nitrogen with a reaction time of 60 minutes at the final temperature. After the
sample is cooled under nitrogen and screened again, the change in size distribution is calculated.
The percentage thermal fragmentation of coal is given as a percentage decrease in Sauter
diameter. The smaller the percentage decrease, the better the thermal stability.
Thermal fragmentation is defined as:
% Thermal fragmentation =
d
p
before test − d after test
p
d
p
before test
x 100
…(3)
As illustrated in Figure 2, weathering / oxidation and moisture content affect the thermal
fragmentation of coal sources. An extensive study revealed that the effect of moisture
contributes to ±75% of the thermal fragmentation of coal [1]. This is not only surface moisture,
but a combination of surface moisture and inherent moisture captured within the pores and the
coal structure. Although moisture contributes significantly towards fragmentation, it is also
affected by a complex interaction with other factors.
6
FIGURE 2
THERMAL FRAGMENTATION OF COAL SOURCES (effect of moisture and weathering)
70
% Thermal fragmentation
Secunda
Sasolburg
60
50
% Thermal
fragmentation
(dry)
% Thermal
fragmentation
(w et)
40
30
20
10
0
A
3.5
B
C
D
E
F1
F2
F3
I & II
Caking
Caking of coal particles can be described as the softening or plasticity property of coal, which
cause particles to melt or sinter together to form larger particles when heated. Caking of coal
within the gasifier can cause pressure drop fluctuations and channel burning, resulting in
unstable gasifier operation. In severe cases oxygen break-through can occur, which can be a
safety hazard due to the possibility of downstream explosions.
The caking propensity of coal is determined by pyrolizing a coal sample with a specific
predetermined size distribution in an argon atmosphere at the typical gasifier pressure, i.e. ±26
bar. The sample is screened afterwards and the increase (if any) in particle size determined.
This test is unique and was developed in-house by Sasol for characterizing coal under conditions
similar to those prevail within the gasifier.
Pressure significantly influences the caking propensity of coal. Coal with a medium to low caking
propensity shows no caking at atmospheric conditions, and a highly caking coal will have a much lower
caking propensity at atmospheric conditions than at 26 bar. Atmosphere (i.e. nitrogen, CO2, etc.) does
not have a significant effect on the caking propensity of coal.
According to previous experience, coal from sources with a relatively high caking propensity resulted in
unstable gasifier operation and are therefore mixed with other coal sources having a low caking
propensity to obtain an acceptable blend. Normal blends used for gasification at Secunda have a caking
property of ±20 % and coal blends used in the Sasolburg plant have no or a very little caking. The
variation in caking properties between the different coal sources used for gasification in Secunda and
Sasolburg, as well as non-South African coal sources tested, are given in Figure 4.
7
FIGURE 3
CAKING PROPERTIES OF SASOL’S COAL COURCES
Secunda
Sasolburg
100
Non SouthAfrican
90
High risk
80
% Caking
70
60
50
40
Uncertain area
30
20
Safe operating
region
10
0
A
3.6
B
C
D
E
F1
F2
F3
I & II
a
b
d
Ash fusion temperatures and ash composition
The ash fusion temperature (AFT) of a coal source gives an indication to what extent ash agglomeration
and ash clinkering is likely to occur within the gasifier. Ash clinkering inside the gasifier can cause
channel burning, pressure drop problems and unstable gasifier operation.
The results of an AFT analysis consist of four temperatures, namely the initial deformation temperature,
softening temperature, hemispherical temperature and flow temperature. Ideal gasifier operation is to
operate at a temperature above the initial deformation temperature in order to obtain enough
agglomeration to improve bed permeability, but to operate below the ash melting temperature to prevent
excessive clinkering. Ideal coal sources will thus have a big difference between the initial deformation
temperature and the melting temperature. Secunda and Sasolburg coal sources currently used for
gasification have an ash melting temperature > 1350oC and an initial deformation temperature of
>1300oC. Non South-African coal sources tested (China, India and Ethiopia) have ash fusion
temperatures similar to the Sasol coal sources. This does not imply that coal sources with lower ash
fusion temperatures are not suitable for the Sasol-Lurgi gasification process, provided that gasifier
operation is adopted accordingly.
The ash composition, specifically the Ca and Fe content in the coal, gives a fair indication of the
expected ash fusion behaviour. A Ca and/or Fe rich coal source normally has a low ash fusion
temperature due to the fluxing properties of the Ca and Fe minerals.
Although the standard AFT analysis is currently used as the only prediction tool for ash fusion
temperatures of coal, literature studies have showed that this may not represent the actual flow
temperature of certain minerals and mineral phases [6]. Fe, for example, in a specific phase can slag at
temperatures as low as 700oC and then solidify again. This is not reflected by a standard AFT analysis.
Sasol Technology, R&D Division, is currently investigating this issue. Visual investigation of actual
ash produced from a fixed bed gasifier showed that the coarse and fine ash is sometimes extracted from
the gasifier with the important middle fraction being absent. This can possibly be explained by the fact
that some minerals already slag and clinker at low temperatures.
8
4. MAXIMUM THEORETICAL PURE GAS YIELD
In order to compare the maximum theoretical pure gas yield of different coal sources, a detailed
experimental evaluation, together with thermodynamic modeling, is conducted on the coal to simulate
the gasification process. The theoretical pure gas yield predictions are subject to certain assumptions:
•
The thermodynamic model calculates the theoretical pure gas yield of the remaining fixed C
after pyrolysis.
• Experimental data for pyrolysis gas production and composition is added afterwards to obtain
the maximum theoretical pure gas yield.
Inputs in model:
• Coal composition from proximate analysis.
• Pyrolysis product yields.
• Composition of pyrolysis gas.
Assumptions made in model:
• Percentage unconverted C reporting in the ash is 3% of total amount of C in the coal feed,
but can be adjusted according to experimental data.
• CH4 and water-gas shift reactions are at chemical equilibrium. CH4 formation and water-gas
shift approach to equilibrium can be adjusted to match actual CH4 production figures and
H2/CO ratios.
• No kinetic, hydrodynamic or particle segregation effects are taken into account. This means
that the model is not affected by load, particle size, coal reactivity, diffusion effects, heat and
mass transfers. All of these factors may have an effect on the actual carbon conversion and
yield predictions.
Examples of theoretical PG yield predictions for a few South African coal sources and one non
South-African source are shown in Figure 4. Exact agreement between predictions and plant
data are normally not obtained, but experience showed that the predicted trends are valid (e.g. if
blend composition changes and an increase in yield is predicted, then an increase is measured on
the plant).
9
FIGURE 4
COMPARISON OF MAXIMUM THEORETICAL PURE GAS YIELD BETWEEN COAL SOURCES
Pure Gas Yield (Nm 3/t DAF)
1700
1650
1600
1550
1500
1450
1400
1350
I & II
X
I & II plus
15% X
Y
Non South
African
5. CONCLUSIONS
The coal characteristics discussed in this paper are not the only properties affecting gasifier
performance and stability, but are those properties that are measurable on laboratory scale and
are easily related to gasifier performance. Interpretation of these results gives an indication of
expected gasifier performance, and also the suitability of a specific coal source for Sasol-Lurgi
fixed bed gasification. Interpretation of standard coal analyses and uniquely developed
laboratory tests, together with previous experience gained by Sasol over the past 50 years,
has put Sasol in a position to identify suitable coal sources for a Sasol-Lurgi fixed bed dry
bottom gasification process.
Sasol is furthermore in the unique position not only to have the ability to characterize and test
coal from various sources on laboratory scale, but also to test new and current coal sources on an
isolated commercial scale Sasol-Lurgi MK IV test gasifier at the Secunda site. These full scale
tests require approximately 4000 tons of test coal, which allows a 6 day test run on a MK IV
Sasol-Lurgi fixed bed dry bottom gasifier at a coal feed rate of ± 50 tons/h. Full scale test results
are supportive to laboratory scale coal characterization data, since past experience showed that
gasifier performance is usually not determined by one or two coal characteristics, but is
dependent on the combined effect of all properties due to the large degree of interaction
between them. This facility is used and will assist in future to further optimize product yields
and to increase throughput by means of a thorough understanding of those coal characteristics
that influence gasifier performance [7].
It is strongly recommended that full scale test work be conducted to confirm the design basis for
new plants with unfamiliar coal sources. Sasol has experience and a proven track record on
commercial scale test runs in Sasol Two and Sasol Three with South-African coal sources as
well as with American coal sources, (e.g. DGC, EL Paso, CarteOil, Tristate, Phillips-Petroleum).
10
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Van Dyk, J.C., Thermal friability of coal sources used by Sasol Chemical Industries (SCI) for
gasification – Quantification and statistical evaluation, M.Sc. Thesis, University of the
Witwatersrand, Johannesburg, 1999.
Kunii, D. and Levenspiel, O., Fluidization Engineering, Robert E. Krieger Publishing Company,
Huntington, New York, 1977.
Keyser, M.J., Van Dyk, J.C. “Full Scale Sasol/Lurgi Fixed Bed Test Gasifier Project:
Experimental Design and Test Results”, Paper presented at the 17th Annual International
Pittsburgh Coal Conference, Pittsburgh, USA, 11-15 September 2000.
Adding value to life, Sasol Annual report 2000.
Stubington, J.F., Linjewile, T.M., The effects of fragmentation on devolatilization of large coal
particles, Fuel, Vol. 68, 1989, pp. 155-160.
Alpern, B., Nahuys, J., Martinez, L., Mineral matter in ashy and non-washable coals – Its
influence on chemical properties, Comun. Serv. Geol. Portugal, 1984, t. 70, fasc. 2, pp. 299-317.
Keyser, M.J., Findings on the statistical evaluation of tests 14 to 23 on the test gasifier with a
standard grate and standard grate/uniflo grate comparison, 28 January 2000.
Sasol Annual Report 2000.
Van Dyk, J.C., Development of an alternative laboratory method to determine thermal
fragmentation of coal sources during pyrolysis in the gasification process, Fuel 80 (2001), pp.
245-249.
Hirschfelder, H., Buttker, B., Steiner, G., Concept and realisation of the Schwarze Pumpe, FRG
‘Waste to Energy and chemicals centre’, IChemE Conference ‘Gasification in Practice’,
Assolombarda, Milan, Italy, 26-27 February 1997.
ACKNOWLEDGEMENTS
The author gratefully acknowledges support and inputs received from the co-authors MJ Keyser,
JW van Zyl as well as P van Nierop and also wishes to thank B Ashton and S du Plessis for their
work contribution to the successful completion of this paper.
11