REDUCTION OF ATMOSPHERIC POLLUTION
FINAL REPORT ON RESEARCH ON
REDUCING EMISSION OF SULPHUR OXIDES,
NITROGEN OXIDES.AND PARTICULATES
BY USING FLUIDISED COMBUSTION OF COAL
Main Report (Vol.1 of 3)
Environmental Protection Agency,
Office of Air Programs,
Research Triangle Park,
North Carolina 27711
-------
NATIONAL COAL BOARD
FINAL REPORT
. JUNE 1970 - JUNE 1971
REDUCTION OF ATMOSPHERIC POLLUTION
MAIN REPORT
PREPARED FOR
'ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF .bIR PROGRAMS
411 WEST CHAPEL HILL STREET
DURHAM, NORTH CAROLINA 27701
REFERENCE NO. DHB 060971
FLUIDISED COMBUSTION
CONTROL GROUP
SEPTEMBER 1971
NATIONAL COAL BOARD
LONDON, ENGLAND
-------
FOREWORD
This document describes studies carried out by the National Coal
Board of the U.K. and jointly financed by the NCB and the National Air
Pollution Control Administration. (now the. Office of Air Programs) of
the U.s. Environmental Protection Agency. The work was performed
over the period June 1970 to June 1971 at the NCB research establishments
at Che1tenham (CRE) and Leatherhead (BCURA). The Contract Manager for
the Office of Air Programs was Mr. P.P. Turner and the OAP representative
in the U.K. was Mr. E.L. Carls.
~ .
Hi
..
-------
Foreword
Table of Contents
Page No.
1.
Summary and Conclusions
Introduction
2.
1
3.
Material Used
The Research Programme
4
3.1
3.2
Coals
Limestones and dolomites
6
6
7
4..
4.1
Pilot~Plant. Experimental Work
4.2.3..1
4.2.3.2
4.2.3.3
8
8
8
9
9
13
14
14
17
17
21
25
26
26
31
Experimental techniques
4.1.1 Measurement of 802 concentrations
4.1.2 Measurement of NO. concentrations
x
4..L.3 Measurement of corrosion of metals
4.2 The 36 in combustor (Task I)
4.2.1 Description of the plant
4..2.2 Experimental programme
4.2.3 Results
4.3
The 48
4.3.1
4.3.2
4.3.3
4.4
Pittsburgh coal and Limestone 18
Pittsburgh coal and Dolomite 1337
Welbeck coal and U.K. limestone
x 24 in pressurised combustor (Task II)
Description of the plant
Experimental programme
Results
32
36
36
40
40
The 27 in combustor (Task III)
4.4.1
4.4.2
4.4.3
4.5
Description of the plant
Experimental programme
Results
The 12 in corrosion combustor (Task IV)
4.5.1
4.5.2
4.503
Description of the plant
Experimental programme
Results
46
46
50
50
v
-------
4.6
40503.1
4.5.3.2
4.5.3030
Test Series 1: 100 h test without limestone
Test Series 2: 500 h tes~ without limestone
Test Series 3:.500 htest with limestone
addition
The 6 in combustor (Task V)
4.6.1
40602.
4.603
4.7
Description of the plant
Experimental programme
Results
406.3.1
406.3.2
4.603.3
U.Ko limestone with Welbeck, Park
Illinois and Pittsburgh coals
Limestone 1359 and Illinois coal
Limestone 18 and Pittsburgh coal
Hill,
Results of corrosion Investigations
4.7.1
4.7.2
4.7.3
Appearance of Steels after Tests
Weight loss Results
Meta1lographic Examination
5.
The Mathematical model (Task VI)
Laboratory Experimental Work
6.1 Coal Studies (Task VII)
6.2 Limestone and dolomite studies (Task VIII)
6.
6.2.1
6.202 .
6.2.3
6.2.4
602.5
6.2.6
6.2.7
6.2.8
6.2.9
7.
Discussion
Introduction
Pore. structure on calcination and sulphation
'Impermeable' layer formation on su1phation
Porosity at elevated temperature
Effect of thermal cycling on S02 uptake
Deposition on UoK. dolomite
Surface area
A simple test for stones
Reaction rate measurements
7~1 Emission of SO 2
7.1.1 Introduction
7.1.2 Eff ect of Plant Design
7.1.3 Effect of CalS Ratio
7.1.4 Effect of Combustion. Temperatur~
7.1.5 Effect of Fluidising Velocity
701.6 Effect of Bed Height
7.1. 7 Effect of Operating Pressure
701.8 Effect of Recycle
Page No.
50
50
52
52
52
55
56
56
56
56
63
63
67
67
69
71
71
73
73
73
79
79
81
82
83
83
83
86
86
86
86
91
93
99
101
105
108
-------
7.1.9
Page No.
Effect of Additive Particle Size
109
114
119
121
128
129
137
138
139
141
142
142
142
142
144
144
144
146
146
7.1.10 Effect of type and source of additive
7.1.11 Effect of coal type
7.1.12 The approach to. steady state
7.1.13 Accuracy of the results
7.1.14 The mathematical model
Emission of nitrogen oxides
7.2
7.3
7..4
7.5
7.6
Alkali and chloride emission
Combustion performance
Emission of Par~icu1ates
Corrosion and Deposition
7..6. 1
7.6.2
8.
Acknowledgements
9.
References
10.
Glossary
Appendix 1.
Appendix 2..
Appendix 3.
Appendix 4.
Appendix 5..
Appendix 6.
. Appendix 7.
Appendix 8..
Appendix 9.
Deposits
Corrosion
7.6.2.1
70602.2
70602.3
7.6.2.4
7.602.5
7.6.2.6
Eff ect of time
Effect of different coals
Effect of f1uidising velocity
Effect of operating pressure
Effect 'of additive
Comparison of 'above bed' and 'in bed' results
147
149
150
Experiments with the
Experiments with the
Experiments with the
Experiments with the
36 in combustor (Task I)
48 in x 24 in pressurised combustor (Task II)
27 in combustor (Task III)
12 in corrosion combustor (Task IV)
Experiments with the 6 in combustor (Task V),
Mathematical model of . sulphur retention (Task VI)
Coal studies (Task VII)
Limestone and dolomite studies (Task VIII)
Methods for determination of nitrogen oxides
The appendices are bound in two separate volumes: Appendices
1, 2, and 3 in one volume, Appendices 4, 5, 6, 7, 8. and 9 in another.
-------
SUMMARY AND CONCLUSIONS
Scope of the Programme
The U.K. National Coal Board (NCB) and the U.S. National Air
Pollution Control Administration (NAPCA - now the Office of Air
Programs, OAP) entered into an Agreement starting on June 1st 1970,
jointly to sponsor a programme of research to determine the usefulness
of the f1uidised combustion system in reducing the emission of sulphur
and nitrogen oxides, and particulates from the combustion of coal.
The programme was scheduled to-cover a period of twelve months
and involved the following work at the NCB's Leatherhead and Che1tenham
laboratories:
(i) Experiments on a number of pilot-scale combustors to
measure the effect on emission of sulphur and nitrogen oxides and
particulates of a selected range of process conditions, e.g. (a) coal
type, (b) the quantity, size and type of additive (limestone/dolomite)
fed to the combustor to retain sulphur, (c) combustion conditions as
regards temperature, pressure and fluidising velocity, (d) plant scale,
and (e) design-features such as bed depth and the recycling of
incompletely reacted fuel and additive.
(ii) Experiments on selected pilot-scale combustors to assess
the extent to which the addition of limestone or dolomite to coal in a
fluidised bed in a large test rig influences the corrosion, erosion and
deposit formation on specimens representative of typical evaporator,
superheater and reheater tube metals.
( Hi)
Laboratory scale experiments to characterise the coals and
additives used.
(iv) Development of a mathematical model to assist in correlating
the factors influencing the pollution control characteristics of a
fluidised combustion system.
ix
-------
The features of the combustors used, the range of operating
conditions explored, and the aggregate number of test hours accomplished
are shown in Tables Sol and So2oIt will be noted that experiments
were carried out at combustion pressures up to 5 atmospheres absolute,
fluidising velocities up to-ll ft/s, bed temperatures up to 1680oF, using
four different coals, three limestones and two dolomites, and that the
test running tUne totalled 5,300 hours.
Table
Sl
Main Features of the Pilot-Scale Combustors
Location BCURA, Leatherhead CRE, Che1tenham !
,
,
Designation 27 in 48 inx 24 in 36 in 12 in 6 in
(pressurised) (corrosion)
Bed Cross 27 in dia 48 in x 24 in 36 in x 18 in 12 in x 12 in 6-in dia
Section \
Bed Depth f t 1.5 - 2 3.5 - 4 2 - 7 2 2 - 3
Operating 1 up to 5 1 1 1 I
Pressure
AtmoAbs.
F1uidising 6 - 11 . 2 2 - 8 3 2 - 3
Velocity
ft/s
Coal Rate 200 - 300 300 - 500 75 - 300 20 - 25 4 - 6
1b/h
Total 2150 430 1000 1100 600
Running ,
Hours
-------
r-- - -u.- - ~:~~:~-~:
v a 1:.\ abl~s
CoaJ
~
Asb Content
\101.atile Matter
Sulphur
Chlod.ne
H20 as fed
Ash Fusion
Size
Bed Depth
'I'empE.~at.u~e
Fluidising Velocity
. -
Excess Air
Recyde ~f Cylone
Fines
Add i.tive
CaiS ID".)}
X'Citl.O
%
of
ft/s.
Table
S2
The Vat'.l.ab1es Explored
.
- .
Pilot-Scale Combustor(s)
Non-Pressurised
Pressurised
UoSo Pittsburgh
UoSo Illinois
UoKo Welbeck
U., S 0 Pittsburg~
UoK, Welbeck
U oKo Park Hill
%
%
12 - 18
37 -. 47
10 3 - 404
001 - 006
1 - 10
13 - 18
30 - 41
103 - 301
001 - 006
1 - 6
2100 - 2600
-1/16 in
%
%
1800 -:. 2600
-1/8 ~n and -1/16 in
ft
;)
F
105 - 7
1420 - 1680
305 - 4
1470
2 - 11
2
%
-12 to +29
11 to 33
Zero, Partial, Full
Partial
UoS, Limestone 18
U,S, Limestone 1359
U,K.. Limestone
UoSo Limestone 18
UoSo Dolomite 1337
UoKo Dolomite
UoSo Dolomite 133'7
0-6
0-3
xi
-------
Main C.:;.nc lusl.ons
Tbe ma i 0 c.::;;ocl u.s tons reacbed from tbe work are:-
(i) That witb fluidised combustion and the addition of limestone
(cr d:'lomi te.) the emission of sulphur oxides from coal burning power
plant can readily be controlled to meet the very rigorous restrictions
(100 ppm v/v S02) planned for certain densely. populated areas in the U.S.
'F.:r a p..:.wer plant. burning a 3% sulphur coal this would involve feeding
sufficient additive to retain 95% of the sulphur, Under the best
combinatiDns of operating conditions about 1.8 times the stoichiometric
quantity of additive would be required,) which for a 100 MW plant would
invc1ve supplying 160 ton/day of limestone or 280ton/day of dolomite.
The less st.'ringe" t restrictions that have been proposed for buil t-up .
areas (300ppm), and for power stations generally in the U.S. (700 ppm),
would require sulphur retentions of 85% and 67% respectively for 3%
sulphur
-------
The orde.r in which the operating variables .are c0nnnent.ed upon
takes 10t0 account. both their relative importanc.e.and s.::me.of the inter-
actIons, e.g. thr)ugh their effect on gas and solids residence times.
Since there was 110 s'ignificant difference in the. S02 emiss Lon from
the different combustors when operating under comparable conditions,
it has been concluded that differences in scale and design ~eatures could
in general be ig-ncred when considering the effects of changing the
va.:riab 1 es,
GalS m,J 1 ratio: The S02 is reduced asymptotically to zero as the feed
rate ot additive to the' fluidised bed is increasedn The percentage S02
redur.tion obtained at a given operating condition is a function of the
molar rat.io of added calcium to sulphur in coal, and is almost independent
of the sulphur content of the coal, since the react.ion is approximately
first order with respect tG 502 conc.entration. Clearly, in order to obtain
a specified concentration of S02 in the offgas.whel1 burning a coal of' high
sulphur cootent, it if:' ne('essary to achieve a higher percent.age S02
reduct:i..:m by using a higher GalS mol ratio.
For a given coal the lowest values of t.he GalS mol ratio were
required at. (i) low fluidising velocities (i.e. 2 to 3 ft/s), (ii) a bed
temperature of around 15000F, and (iii) when most of the fines larger t.han
about 10 ,.lm were recycle.d.
Bed Temperature: The optimum bed temperat.ure was between 14000F and
1600G'F. The level. of S02 emission and the change. In.-emission with change
(..f temperature on each side of the optimum appeared to depend on the type
of additive used and to some extent on the GalS ratio employed. The
.i 01: rease io e.mlssion co ei ther side of the minimum tended to be greater
at l~w thaD.at high GalS ratios, i.e. under conditions where the fraction
d calcium suJphated was higher. The optimum temperature was found to
'be 1500-1550oF for limestone additive and l4000F-1500op for the dolomite.
The data suggest that t.;) maintain the same level.of sulphur emission
o
(e..g. 85% sulphur retention) at, for example, 100 P above the optimum,
would involve increasing the GalS ratio by a factor of about two.
The
effect of changing bed temperature was not investigated on the pressure
combust.:>T.
xiii
-------
The rapid increase in sulphur emission at bed temperatures
above about 15S0oF is unexpected from laboratory measurements of the
reaction rate between CaO and S020 Since the effect of temperature
appears to be reversible, Leo the S02 reduction reverts to a high
value almost immediately the bed temperature is reduced,. it cannot be
accounted for-by irreversible factors such as sintering or sl~g
formation at the particle surface. One tentative explanation postulates
that an oxygen-containing species (e.g. hydroxide. ions derived from
traces of water, which is k.nown to be difficult to remove) are
involved in the conversion of CaS03 to c.a.S04' It is possible that above
the optimum temperature the hydroxide ions become more.mobile and hence
less able to participate in the reaction.
o
At low temperatures (e.g. below 1350 F) .sulphur.retention with
dolomite was higher than with limestone, the reason being the lower
calcination temperature. of the MgC03 in dolomite which leads to the
development of pore structure below the temperature.of calcination of
the CaC03"
Fluidising Velocity: Increase in fluidising_velocity resulted in an
increase in sulphur emission. An empirical correlation was derived
and is reported later in the summary, Increase in velocity without
any compensating action results in reduction in both gas and solids
residence times. .To maintain the same sulphur retention, say 85%,
at 8 H/s as at 2 ftls fluidising velocity. the Ca/S mol ratio (at a bed
o
temperature. of. 1500 F and without recycle) would-have. to be increased
from about 2 to about 4.
Bed Height:
Increase in bed height usually resulted-in a reduction
io S02 emission. An empirical correlation is reported later in the
summary. In principle it should be possible to counteract the adverse
effect of increasing velocity by a proportionate increase in bed height.
At atmospheric pressure the attendant increase in pressure loss for other
than a small increase in bed height could be prohibitive. In addition
because. the tube bank required would occupy only a part of the bed height
the effectiveness of increasing bed height maybe-reduced by the
formation of large gas bubbleso The effect of bed. height in super-
charged boilers is potentially of greater significance.
Here the deep
xiv
-------
banks d close packed tubes may assist in breaking lip large gas bobbles
and hence may improve the contact between gas and s.Jlids. Further, tbE
increase io pressure loss due to increasing bed height is less imp~ttan[
under pressure.
Fines Tecycle: A high proportion of the additive is elutriated tr:m
the bed before being fully uti1ised. By efficient recycle of fines
larger than 10 ~m, to the bed, S02 reduction was ~~: :~;£ed significanrly,
e.g. frem 73% to 99% at a fluidising velocity 2 ft/s aod a Ca/S m~l
ratl.O of 1.6.
Operating pressure: The effect of operating pressure on S02 reductbo
was negligible when dolomite was used as an additive. This is to be
expected with a reaction which is first order with.respect to the partial
pressure of S02" With limestone as an additive, tbe reduction obtatoed
at 5 atm. was appreciably lower than with dolomite, and lower than with
limestone. at. atmospheric pressure. This was also to be expected, since
at l4700F, calcination of limestone to give.a.porous structure wauld n0t
occur at operating pressures above 2 atm. Penetratbn af the particle
by S02 would therefore be difficult and only a surface layer ~f
sulphate would form. It was found, however that the performance of
the limestone was better than this reasoning would. imply, and this
suggests that the exposure of fresh. surface by attrition plays a
significant.role. Nevertheless, from the point.of.view of both the
CalS mol ratio and the total quantity of additive required t~ attain a
target level of sulphur retention, dolomite was superior to limestone.
To retain 85% of the sulphur, for example, the estimated CalS m)l
rat10s for dolomite.and limestone were 1.1 and.3.25.respectively, and
the est.imated quantities of additive were, respectively, 7.6 Ib and
12 lb per Ib af sulphur removed.
. Paortic1e Size:
I t was found., for coarsely crushed limes tonet that
the percentage S02 reduction increased when the. particle size of
limest'one was reduced, probably due to the consequent increase io
available reaction surface"
On the other hand,.with d~lomite there
was n~ effect of particle size, suggesting that access to internal
surface is not a limiting factor for'dolomite.
xv
-------
With additive ground to -lZ5 ~m or -150 11m, it was found that
the fluidising velocity had a profound effect on.SOZ reduction. Whereas
at low velocity (3 ft/s) the fine additive improved,. SOZ. reduction, the
reverse was true at high velocity (8 ft/s). The data of Pope, Evans
and Robbins suggest that, with a CalS mol ratio oLZ.6.in beds 10 in
deep fluidised at 12 ftls, Limestone 1359 gave.80% S02 reduction when
ground to -44.~m, and 60% reduction when. ground to -74 \lm. Evidently
the 502 reduction is very sensitive to the size of finely ground
particles, so that the Pope, Evans and Robbins data are at least
qualitatively consistent with those for -150 ~m limestone from the
present study. Limestone ground to -150 \lm may have too short a
residence time at high velocity to achieve a high.degree.of sulphation,
whilst superfine material will become highly sulphated, since its
residence time will not be markedly less than that of -150 ~m material.
However, superfine material may cause serious.gas~c1eaning problems.
Type of additive: The type and source of additive affects the reduction
in 502 that can be achieved. At. atmospheric pressure Limestone 18 was
the most effective additive on both mo1ar.andweight.bases; the least
effective on a mol basis was Limestone 1359, and on a weight basis
Dolomite 1337. To achieve the same level of retention with the
poorer Limestone 1359 as with the Limestone. 18 would require an increase
. of. up. to .100% in.. the. CalS mol. ratio. As. mentioned. previously, for
operation under pressure, both the dolomites were superior. to Limestone 18
on weight. and molar bases.
An important.finding from the point of view of simplifying
prediction of suitability of stones was that measurements made at
.room temperature. and at combustor temperatures.showed the same
accessibility of the structure to gases of similar molecular size to 5020
Temperature cycling (as may occur in some plant designs when elutriated
particles are recirculated) does not affect. the. pore. structure significantly
from the point of view of 502 uptake. An empirical reactivity test was
considered to be the most economic method for classifying stones. For
limestones the results of laboratory experiments. give pessimistic
predic.tions of plant performance. This. is thought to be because
these tests do not take into account the beneficial effect of attrition
xvi
-------
in the combustor which results in removal of. the_sulphated surface
layer. The effect of attrition was particularly important for
limestones in the pressurised combustor andfor.Limestone 1359 at
atmospheric pressure. For dolomit~ access to the.internal surface
of the particles does not appear to be a limiting factor.
Whereas thermal. losses are incurred from the sensible. heat
requirement and the heat of calcination of an additive, the heat of
sulphation represents a thermal gain. Up to a Ca/S mol ratio of
about. 2 using limestone (sufficient to retain 85% of. the sulphur of
a 3% sulphur coal under. good operating conditions) it is estimated
that the heat of sulphation will counterbalance the sensible heat
requirements and heat of calcination.
With dolomite. however the nett
thermal loss would be about l~% of the coal heat input. Under
pressure, calcination of CaC03 is inhibited. and. the thermal loss
incurred by. using dolomite would be negligible forCa/S mol ratios up
to 2.
. Type of coal:. - The most. impor.tant coal property - in this. .context is the
sulphur content, which determines not only the quantity of additive
required for a given Ca/S mol ratio, but.also the percentage S02
reduction (and hence.ca/s mol ratio needed)to-me~t set. limits of S02
emission. Since the S02 absorption reaction-. is first order with respect
to S02 concentration, it could be expected that.the.same relationship
between percentage reduction and Ca/S mol.ratio would hold.for all coals
irrespective of the sulphur content. However, the experimental results
showed that, for a given Ca/S mol ratio, similar S02 reductions were
obtained for three of the coals, but the reductions.were up to 15%
higher with Welbeck coal. Differences in the rate of sulphur release
have been found between coals and might partly account.for differences
in performance. A more likely explanation of the higher S02 reduction
'with Welbeck coal is its low sulphur content, which had the consequence
that additive was fed at a lower rate and hence had a longer residence
time. This could.have resulted in the higher. degree of sulphation,
particularly if. particle attrition was an important effect.
xvii
-------
Plant design: As mentioned earlier,it was concluded that despite
the difference ,Jf scale and design over therange.of . eombus tors used
(Table 81) t.bere was nu significant difference between the S02
r.edoe:tluns0btained in different combustors with the same operating
CoOTJdi t i c'n!':.
Nevertheless, direct application of the present results
to c(}mbustors.ofc~mmercial s-1ze requires some. caution.
. Some .factors
rNhi "h might a1 ter the S02 emission, such as the depth of the tube bank,
have already been mentlaned. Further, vbservation of a radial
djB~ribu[ion of S02 concentration in the freeboard of one vf the larger
pilot pl~nls suggest.8 that coal feed spacing, if greater than that used
in the pi bt-plants, mflY assume significance in cOIImlerc.ial boilers.
Mathematical rn-:.:,del and correlati'0n of data:
.The mathematical model
has been developed to give fairly satisfactory. prediction of the
consequences of changing some operating conditions" Additional
devekpmeot is needed (a) to take further account. of attrition of
a.dditive dnd (b) t':. extrapolate the resul ts to combustors that differ
significantly fruID the present pilot plants. The model in its
present form has not been useful in correlating. the experimental data.
However anumbe.r of . empirical correlations have been derived as follows:
There is.an approximately exponential relationship between the
S02 reductbn and, the GalS mol ratio of the form
R = 100 (1 - exp(-MG) )
where R is the percentage S02 reduction
G is the Ca/~. mol ratio
M.is.anempirical constantdepending.on the coal,
limestone and operating conditions.
The effect of fluidising velocity on S02 reducti::m may be
approximately correlated by
A ~ XliV
where A is the absorption ratio, defined as R/(lOO-R)
v is the fluidising velocity
Xl is an empirical constant depending on the
GalS mol ratio and other operating conditions.
xviii
-------
The effect of bed height on S02 reduction may be approximately
r.crrelated by
A = X2 H
where H is the bed height
Xi is an empirical constant depending on the
Ca/S mol ratio and other . operating conditions.
Emissi0n of nitr~gen oxides
EmissiGnof NOx from the pressure combustor (50-200 ppm) was
significantly lower than from the non-pressurised.combustors (300-600 ppm).
All the. combustors produced less NOx pollution, both with and without the
use of additives, than is common with conventional plant. The reason
for the superior performance is uncertain. .NOx emission could not be
correlated with S02 emission, although on some occasions a decrease in
the S02 emission (due to feeding limestone or dolomite) was accompanied
by an increase in the NOx emission. It was concluded that more
information was needed on the mechanism of NOx formation before a
contribution could be made towards reducing emission"
Emission of. alkalis and chlorine
As expected,. the low combustion temperatures in.f.luid.bed combustors
resu1 ted in low alka.li emissions"
The combustion:gases from the pressure
combustor contained about 2 ppm of Na, Le. about one tenth of the lowest
cc,t!centration re.ported for the gases from conventional plant. The
concentration of Kwas less than 0.5 ppm.. Higher.emissions were measured
when limestone was added to' the pressurised. combustor instead of dolomite
(5 ppm of Na and 1.5 ppm :.;.f K) and from one of . tbe. non-pressurised
combustors that was being operated at the higher bed temperature of 15600F
(6 ppm of Na and 3 ppm of K). As expected most of the chlorine of the
coal was released into the combustion gases.
Emissi0n of particulates
Particulate matter elutriated from fluidised bed combustors comprises
5 to 15% cf tbe carbon and 80 - 100% of the ash and additive. By using
primary and secondary cycl0nes having collection efficiencies of 90% at
about 10 um it was possible to collect 95 - 98% of this material to give
dust emissions of 0.2 - 0.6 grlscf.
Within this range the emission was
xix
-------
approximately proportional to the feed rate of ash plus additive.
Increasing the fluidising velocity increased.elutriation from the bed,
but because of more efficient cyclone operation with higher gas flow
rates there was little effect on emission.. Fines.recycle in the 36 in
combustor increased the dust emission to 1.4 gr/scf. The pressurised
combustor.had an internal recirculation cyclone in addition to" primary
and secondary cyclones, and dust emissions in the range 0.05-0.1 grlscf
were obtained.
Based on these results it is unlikely that there would be any
problem in meeting projected statutory limitations on particulate
emission.
Corrosion.and deposition
The addition of limestone or dolomite had no significant effect
on corrosion or deposition of tubes in the bed or in the gas space
under the range of operating conditions likely. to be experienced in a
commercial plant.
The amount of material settling on the turbine blade cascade
~t the outlet of the pressure combustor was slig~t and was judged to
be unlikely to.affect turbine performance.. . There were no signs of
sintered deposit~ or of erosion.
-------
1.
INTRODUCTION
The need to reduce air pollution by sulphur dioxide, nitrogen
oxides, carbon monoxide, hydrocarbons and particulates resulting from
combustion of fossil fuels in power plants, is widely recognised.
Operating techniques are available at the present time to
limit carbon monoxide and hydrocarbon emission to acceptable levels.
Particulate emission can in most circumstances be limited to
currently acceptable levels using conventional, albeit expensive
equipment.. The situation as regards the emission of sulphur and
nitrogen oxides is much less satisfactory. Because of the low
concentrations of the pollutants and the large volumes of gas which
must be treated, it is at present inherently expensive to remove
sulphur oxides. Insofar as nitrogen oxides are concerned there are
as yet no commercial processes. An alternative approach is the
development of new fuel burnipg processes in which the emission of
pollutants during the combustion process is reduced. One such
process which appears to be particularly promising is the
combustion of the fuel within .a f1uidised bed to which limestone
or dolomite is added.
of reductions in the
This process also offers the possibility
capital and op~rating costs of a f1uidised
with conventional plant, due to the following
bed boiler, compared
factors:
1.
The placing of steam tubes directly in the fluidised bed
results in a reduction in tubing requirements because of
the higher heat fluxes attainable in the bed.
2.
The low bed temperature at which combustion takes place,
1400 - 17000F, may reduce tube corrosion and fouling.
3.
The higher volumetric heat release rates possible in a
f1uidised bed combustor may permit smaller unit sizes.
4.
The low vo1ati1isation of alkalis at f1uidised bed operating
temperatures may mean that the exhaust gases can be passed
through a gas turbine without deleterious effect. This
would enable the combustor to be operated at elevated pressures
to further reduce the unit size and also to improve the thermo-
dynamic efficiency.
- 1 -
-------
5.
Low quality fuels may be burnt in a fluidised bed.
Experimental work on fluidised bed combustion has been in
progress since 1965,1 both in the U.K., at CRE and BCURA, and in
the U.S., by Pope, Evans & Robbins and the U.S. Bureau of Mines.
The early work was chiefly concerned with combustion efficiency
and heat transfer, ~ut since 1969 the anti-pollution aspects have
become of major importance. In the U.S., research on this aspect
has been co-ordinated and financed by the Air Pollution Control
Office (now the Office of Air Programs) of the U.s. Environmental
Protection Agency. In addition to Pope, Evans & Robbins and the
U.S. Bureau of Mines, other organisations currently carrying out
work on fluidised bed combustion for APCO include the Argonne
National Laboratory, Esso U.S.A., Consolidation Coal and
Westinghouse. The research work up to June 1970 had demonstrated
the following:
1.
The combustion efficiency of a single stage fluidised bed
is in the range 90 to 98%, depending on the operating
conditions.
2.
The combustibles losses are almost entirely as elutriated
carbon particles. On combustion efficiency considerations
the emissions of CO and hydrocarbons are negligible.
3.
The overall combustion efficiency can be improved to more
than 99% by burning the elutriated fines, either by recycle
to the main bed or in a separate burn-up cell.
4.
The heat fluxes to the steam tubes in the bed are high, as
expected.
5.
The particulates leaving the boiler are not so fine as when
using pulverised fuel, and are more readily removed from the
flue gases.
6.
The S02 emissions are reduced plightly by combustion in an
inert fluidised bed~
-------
7.
Under some operating conditions, the S02 emission can be
reduced by more than 95% when limestone is added to the bed.
8.
The system pressure can be increased to at least 6 atm
without significant reduction in combustion efficiency.
At this stage the data on combustion and heat transfer were
reasonably comple~, but there were many gaps in the data on
reduction of pollution, in particular
1.
The effects of the limestone addition rate, bed temperature,
fluidising velocity and bed height were not clearly defined.
It was not known whether it would be possible to attain the
desired S02 retention under the operating conditions indicated
as being the most economic, i.e~ high temperatures and high
fluidising velocities.
2.
It was not known why there was a variation in the level of
percentage sulphur emission from coal to coal at similar
operating conditions, nor was this effect predictable.
3.
Ther~ were few data on the comparative,performance of
alternative lUnestones and dolomites.
4.
The effect of coal and additive particle size was not known.
5.
The effect of fines recycle (containing part-utilised additive)
was not clear, since cdnflicting results had been obtained.
6.
The effect of system pressure when feeding limestone or dolomite
was not known.
7.
There were apparent differences in performance between the 6 in.
rigs at ANL and CRE when operated under similar conditions.
8.
There were few data from large pilot plants and 'it was not
known whether there would be an increase in the emission of
pollutants on scale-up.
9.
It was not known whether the add~tion of limestone to a bed
would affect the rate of corrosion of the steam tubes.
- 3 -
-------
2.
As part of the overall research effort towards an understanding
of these problems, a programme of research has been carried out at the
NCB research establishments at Cheltenham (CRE) and Leatherhead (BCURA)
jointly financed by the U.S. Environmental Protection Agency and the
NCB. This programme made particular use of the large pilot scale
facilities available and also of the NCB's expertise in the. fields
of corrosion, materials characterisation and mathematical modelling.
This document describes the work which was carried out and presents
the main results together with an assessment of the information
obtained. Details of the individual research programmes carried out
are presented in a series of Appendices.
THE RESEARCH PROGRAMME
The main objectives of the research programme were
1.
To assess the effectiveness of the fluidised bed combustion
process, with and without the addition of limestone or
dolomite, towards the reduction of S02 emission showing, by
experiments over a range of operating conditions, those
operating parameters that significantly affect the attain-
ment of the immediate target emission (300 ppm v/v) and to
indicate how these data may affect plant design.
2.
To gather, over the same range of operating conditions, data
on the levels of NO emission that occur during the combustion
x
process when the 502 is partially absorbed by added limestone
or dolomite, relating these data to the operating conditions
where possible.
3.
To measure the particulates elutriated from the fluidised
bed combustor in order to provide data for the design of a
particulates removal system which will reduce the emissions
to atmosphere to an acceptable level.
4.
To contribute towards an understanding of the way in which
the porous properties of limestone or dolomite affect S02
retention under the conditions prevailing in a fluid bed
combustor,and to develop a simple method of classifying
limestones and dolomites according to their utility for
502 retention in the fluidised bed combustion process.
- 4 -
-------
L----
5.
To develop a mathematical model of the retention of S02 in a
fluid bed combustion system to allow the performance of new
plant, with respect to S02 emission, to be predicted at the
design stage. from the design and. other basic data.
6.
To study corrosion of typical steels used in boiler
construction when immersed in a fluid bed burning coal, both
with and without the addition of limestone/dolomite.
The research programme to meet these objectives was organised
as eight Tasks, as follows:
Task
II
III
IV
V
VI
VII
VIII
Main Objective.
I
To compare the performance of the 36 in rig
with that of the 6 in rigs at CRE and
Argonne, and to extend the range of operat-
ing conditions for which experimental data
is available.
To obtain, for operation under pressure, data
on (i) the emission of sulphur and nitrogen
oxides and (ii) corrosion/deposition of
boiler metal and turbine blade specimens.
To .carry out long-term t~~ts to asse~s the
effect of limestone addition on corrosion
of evaporator, superheater and reheater'
metals immersed in the fluid bed.
To obtain data on corrosion of evaporator,
superheater and reheater materials for'
lower f1uidising velocities.
To obtain data on sulphur retention for a
range of coals and limestones, in particular
to allow comparison to be made with the 6 in
rig a t Argonne.
To complete the development of a mathemetica1
model of sulphur retention and to compare its
predictions with the results of laboratory
and rig experiments, and to up-date it as
appropriate.
To investigate the distribution of sulphur in
a range of coals and in the residue from the
rigs.
To investigate the pore structure and
related factors that affect sulphur retention
by lime.
- 5 -
Plant and/or Location
36 in combustor
CRE.
48 in x 24 in
combustor
(pressurised)
BCURA .
27 in combu.stor
BCURA.
12 in'corrosion
combustor
CRE.
6 in combustor
CRE.
Mathematical work
BCURA.
Laboratory work
CRE/BCURA.
Laboratory work
BCURA.
-------
3.
MATERIALS USED
The experimental work in the programme was carried out using four
coals, three limestones and two dolomites.
3.1
Coals
Typical analyses of the four coals are given in Table 3.1.
Table 3.1
Typical Analyses of Coals Used
I
Coal
I Proximate analysis
I
[Total moisture
Ash
I Volatile matter
i
I .
I U1t~mate analysis
i
I Carbon
i Hydrogen
i
r Nitrogen
: Sulphur
i Oxygen + errors
\
i Ch10r ine
% a.r.
% a.r.
% a.r.
% d.b.
% d.b.
% d.b.
% d.b.
% d.b.
% d.b.
(Btu/1b)
) Calorific value (d.a.f.)
i Swelling Number
) Gray King coke type
: Ash Analysis
CaO
MgO
Na20
K20
siO
. '2
%
%
%
%
%
Size
-
; (As Received)
Illinois
9.8
11.8
46.6
67.8
4.5
1..3
4.4
8.5
0.2
14300
4i
D
10 .1
1.0
1.7
1.8
40.8
I
i
-! in
Pittsburg Park Hill We1beck
1.6
13.5
41.1
71.7
4.5
1.4
2.8
4.4
0.1
15100
8
'G9
8.0
1.3
0.7
1.6
45.8
-! in
The sources of the coals were as follows:
2.1
16.5
39.2
68.2
4.4
1.3
2.5
5.3
0.1
14750
1
D
2.2
1.7
0.8
3.6
46.0
-li in
4.2
18.2
38.3
67.5
4.3
1.5
1.3
5.1
0.6
14400
1
C
1.8
1.4
1.8
3.2
57.5
-li in
Illinois coal: Illinois seam No.6, Peabody Coal Company, Mine 10,
Christian County, Illinois: supplied by the Argonne National Laboratory.
This coal has been referred to as 'Peabody No.lO' or 'Argonne' in interim
reports.
- 6 -
-------
Pittsburg coal: Pittsburg seam, high volatile 'A'
Humphrey P1I"epara.tion Plant, Osage,. West Virginia.
referred to as 'Humphrey No.7' in interUn reports.
bituminous coal,
This coal has been
Park Hill.:
Yorkshire Coalfield, Park Hill Colliery, Nr. Wakefield.
.,' We1beck: Midland Coalfield, N .Notts Area, Welbeck Colliery,
Nr. Mansfield. Top Hards seam.
3.2
LUnestones and. dolomites
Typical analyses of the limestones and dolomites are given in
Table 3 . 2.
Table.3":.2
Typical analyses of lUnestones and dolomites
Composition %
Component
Dolomite U.K.dolomite Limestone . Limestone U .K.limestone
1337 18 1359
CaO 28.9 29.3 45.7 55.7 55.4
MgO 22.9 21.5 1.4 0.3 0.3
Ji20 + C02 47.4 46.3 36.6 43.6 43.5
Si02 0.5 - 13.6. 0.5 0.7
Fe203 0.2 - 0.3 0.1 0.1
S03 - 0.1 - - -
, Total 99.9 97.2 97.6 100 .2 99.8
The sources of those materials were:
Limestone 18: Supplied by Fuller Industries Inc., Fort Meyers,
Florida. This has been referred to as 'U.S. Limestone No.18' or
'T.18' in interim reports.
LUnestone 1359: Supplied by M.J. Grove LUne Co., Stephens City,
Virginia and prepared at the Argonne National Laboratory. This
has been referred to as 'Argonne' limestone or 'Limestone 1359'
in interim reports.
- 7 -
-------
U.K. limestone: Supplied by J.Gregory & Sons, Kidsgrove, Stoke-on-Trent.
U.S. dolomite 1337: Supplied by Charles Pfizer & Co., Gibonsburg, Ohio.
U.K. dolomite: Supplied by Steet1ey (Manufacturing) Ltd., Worksop, Notts.
This has been referred to in interim reports as 'U.K. dolomite' or
'Do1of1ux' .
4.
PILOT PLANT EXPERIMENTAL WOR~
The experimental programmes for the pilot plants were sub-divided
into test series, eachcompi-isiIig from 1 to 15 tests. The operating
conditions in each test series are summarised in Table 4.1. This
experimental programme is' somewhat different from that given in the
research proposa1s2, mainly because of late delivery of American coal
and limestone. Also additional tests were introduced as the programme
proceeded; these changes were agreed with the OAP representative in
the U.K.
4.1
Experimental techniques
4.1.1 Measurement of S02- concentrations
Three methods have been used for the measurement of the S02
concentration in the off-gas from the pilot scale combustors:-
The iodine method
This was the standard method used to determine the S02
concentration in all the pilot plant work. Gas was withdrawn from
a sample probe by means of a vacuum pump and bubbled through a
solution of iodine. The volume of gas required to deco1ourise a
solution of known concentration was measured.
Infra-red analyser
Hartmann-Braun infra-red analyserswere used to monitor the
S02 concentration in Tasks I, II and III. The agreement with the
iodine method was good in Tasks II and III, but in Task I the
ana1yser indicated concentrations about 12% lower than those
measured by other methods.
The hydrogen peroxide method
This method was used in Tasks I,'IV and V as a check on the other
methods. Gas was bubbled through a solution of hydrogen peroxide and
the sulphate produced was determined gravimetrically by precipitation
as barium sulphate.
-------
4.1.,2
Meaaur~eut of NOy concentrations
It was intended. that infra-red absorption.NOx analysers would
be used for this work, but these were not delivered. in time to be
used. The two methods which were used to determihe the.concentration
of NOx are described in detail in Appendix 9 and summarised below:
The flue gas sample was withdrawnthrougha.silica probe and
bubbled through water to remove 802' HCI and other soluble gases.
(A steel probe.with water injection was used for the pressurised
combustor). About 5% of the NOx was also dissolved in the water,
the exact amount being determined using Saltzman's reagent. Two
alternative methods were used to determine.the remaining NOx content
of the gas.
The modified Saltzman's method
A sample of the S02 free gas was taken.in an evacuated 400
ml sample bottle containing 40 ml of 8altzman's.reagent. After
30 min the solution was withdrawn and fresh reagent was added.
This was repeated.atintervals' of 30 min until the colour developed
by the solution was negligible. All of the solution was bulked
and the intensity of the colour was measured using a spectrophoto-
. meter.
The BCURA NOx box
In the BCURANOx.box the S02 free gas is first passed
through an oxidiser in which any NO present i~'converted to N02'
The gas is then passed through a cell containing.a platinum gauze
electrode moistened by a wick dipping into an electrolyte solution
in which an active.carbon electrode is immersed. .The electrical
current through the external circuit between the electrodes is
dependent on the N02 concentration in the gas,.and.is measured
. .a microammeter. . Before' use the instrument is calibrated using
using
standard gas mixtures.
4.1.3
Measurement of corrosion of metals
The corrosion of metal specimens at various temperatures was
investigated.under Tasks II, III and-IV. Composite tubes were
formed by clamping together i inch wide rings of the metals to be
- 9 -
-------
Table 4.1. Summary of pilot plant research programme
(X, 0 and Z indicate the operating conditions in each test series)
Task I II
Test Series 1.1 1.2 1.3 2 3 4 5 6 1 2. 3 4 5
Illinois
Coal Pittsburgh X X X 0 X 0 X 0 X X X X X X 0
Park Hill
We1beck X X
Coal -1587 or -1680 X X X 0 X 0 X X X X X 0
size, ).lm -3175 X 0 X X X
U.K. limestone X
U.K. dolomite X
Additive Limestone 18 X X 0 X 0 X 0
Limestone 1359
Dolomite 1337 X 0 X X X X X
Additive -125 or -150 0
size,).lm -1587 or -1680 X X 0 X 0 X X X X 'X 0
-3175 X X X X
0 X X X X X X X X X
0.5
CalS 1 X X 0 X X
1.5 0 X X X X
mol 2 X X 0 X X X X X X X X 0
ratio 3 X 0 X X X
4 X
6 X
1290
Bed 1380 X X
1470 X X 0 X 0 X X X X X 0
temperature, 1560 X X X 0 X X X X
of 1650
1700
2 X 0 X X X X X 0
Fluidising 3 0 0
4 X X X 0 X X X
velocity, 6
ft/s 8 X X X X
11
1.5
2 X X X 0 X 0 X X X
Bed 3 X
height, 3.5 X
ft. 3.75 X X X X X 0
4 X X 0 X
6.5 X
System 1 X X. X 0 X 0 X X 0 X X
pressure, 3.5 X X
atm. .abs. 5 X X X X 0
Fines No X X X 0 X 0 X X X X
recycle Yes 0 X X X X X 0
- 10 -
-------
Table 4.1 (continued). Summary ~f pilot plant research programme
(X, 0 and Z indicate the operating conditions in each test series)
Task III IV V
Test Series 1 2 3 4 1 2 3 1 2 3 4 5
Illinois X X 0
Coal Pittsburgh X X 0 X 0 Z X 0 X X X X 0
Park Hill 0
Welbeck X X
Coal -1587 or -1680 X X X X X 0 X 0 X.O X
size. ].Jm -3175 X X 0 X 0 Z X 0
U.K. limestone X X 0 X X
U.K. dolomite
Additive Limestone 18 X 0 X 0 0 X 0 X
Limestone 1359 X X 0
Dolomite 1337 Z X
Additive -125 or -150 0 0 0
-1587 or -1680 Z X X X X 0 X X 0 X
size, ].Jm -3175 X X 0
0 X X X X X X 0 X
0.5 X X 0
Ca/S 1 X 0 X X 0 XO X 0
1.5 X 0
mol 2 X 0 Z X X X X X 0 X
ratio 3 X 0 XO X 0 X X 0 0 X 0
4 X X
6
1290 X
Bed 1380 X
1470 X X 0 X X 0 X 0 X 0 X
temperature, 1560 X 0 X 0 Z X X X X
of 1650 X
1700 X
2 X
F1uidising 3 X X X X X 0 X 0 X 0 X
4
velocity, 6 X 0
ft/s 8 X X 0 X Z X 0
11 X
1.5 X
2 X X 0 X.O Z X 0 X X X X X 0 X 0 X 0 X
Bed 3 X
height, 3.5
ft. 3.75
4
6.5
System 1 X X 0 X 0 Z X 0 X X X X X 0 X 0 X 0 X
pressure, 3.5
atm. abs. 5
Fines No'" X 0 X X X X 0 X 0 X 0 X
recycle Y~s X X 0 X 0 Z X 0 X X X X
... In Task III, no recycle refers only to the external system.
- 11 -
-------
Steam/air
inlet
~m/air
~
outlet
~..,
C>
holes
91ands
I II
35 approx.
I ,.
4 3 approx.
- Fig. 4.1
General
arrangement of corrosion tube
"
'.:J
'\~
, 'it--
Corros i on spec i mens
/EXisting
---
------
--
rig tube
I.
assembly
20jlO
'V
J
-------
examined. The 'rings were 2 in diameter f::>r TaskS' : j t :lnd IV. as Sh')1..TJ}
in Fig. 4.1, and I in. dia for Task II. The surface-of the rings was
ground to a smooth finish of 10 ~in or better, and each ring was
cleaned and weighed before assembly into the tubes. Several such tube
assemblies were installed in the fluidised bed and each tube was
maintained at a different constant temperature, using air as coolant
(or steam in some tests in Task III).
Tube wall temperatures were
measured using thermocouples inserted into holes drilled in tWG vr
three rings on each assembly. In Task IV, cooled tube assemblies
were a1so installed in the freeboard, and uncooled metal plates
(coupons) were used, both in the bed and freebcardo
After a test the tubes and coupons were removed, photcgraphed
and examined visually, Deposits and scales were sampled where they
existed and where they could be removed withaut damaging the metal
specimens. Rings and coupons were. then selected.for . more detailed
microscopic.examination of' the fireside surfaces and thickness
measuremenets of scale and deposit with a micrometer.
The majority of the test specimens were then descaled by
immersing them for five minutes in a bath of molten sodium hydroxide
containing-2%.oLsodium.hydride maintained at 680.~ 71SoF. They
were then water quenched and washed to remove.all traces of scale.
The specimens were then re-weighed and a selection.of.both desca1ed
and undescaled rings and coupons were. sectioned for meta1lographic
examination.
A number of 'blank' determinations were carried out to
assess the accuracy of the weight 10ss determinationso Specimens
were weighed and descaled without being exposed-in.. the fluid bed
combustors. .The-results suggested that over a.period of 100 hours,
weight loss changes of less than 2 IJg/cm2h..were..not.significant.
The.36 in ce.>mbustor (Task I)
4.2
The wo-rk carried out on the 36 in combustor.is described in
detail. in Appendix I.
- 13 -
-------
4.2.1
Description of the plant
A flow diagram of the 36 in combustor plant is shown in Fig. 4.2
and the details of the combustor vessel are given in Fig. 4.3. The combustor
was a mild st.eel column with refractory lining, having an internal CT'OSS-
section of 36 in b> 18 in. The air distributor comprised a plate with
1/2 in diameter stand pipes on a 3 in square pitch, each stand pipe having
. four. 0.147 in diameter holes drilled horizontally around the top. The
height from the distributor to the gas offtake was.l5 ft. Coal and
additive were fed pneumatically from separate hoppers through adjacent
lines to the centre of the bed cross-section, 5 in above the air jets.
The off-gases passed through primary and secondary cyclones and then to
the stack. These cyclones were 90% efficient at 12 ~m and 9 ~m
respectively.
The primary cyclone fines could be recycled to the
bed via an external line, if required. The heat generated in the
bed was removed by 14 water-cooled tubes (2 in inside diameter)
arranged on a4!.in triangular pitch within the 2 ft deep zone above
the distributor. The water flow-rate was normally controlled
automatically to maintain the bed temperature at a set level. The
excess heat generated in the free-board was removed by injecting
,steam, in order to limit the off-gas temperature to, a maximum-of l7000F.
The normal-.operating procedure was to ,set:. up an operating
condition, allow r.hf.' system to stabilize and then to carry out a four
hour mass balance. The main gas sampling point was after the cyclones,
but there.were, also sampling points in the freeboard. Samples were
taken for determination of 02' CO, C02 and CH4 by chromatograph. 502
was determined by bubbling through iodine or hydrogen peroxide
solutions and was also continuously monitored using a Hartmann-Braun
infra-red analyt:er. NOx was measured using Sal tzman' s method. NH3
and Cl were determined by analysis of the peroxide solution. All the
outlet solids streams were sampled and solids samples could also be
extracted directly from the bed. The plant was operated without fines
recycle except for a single experiment, Test 4.6.
4.2.2
Experimental programme
Tests were carried out using Pittsburgh coal with Limestone 18
and Dolomite 1337, and Welbeck coal with U.K. limest.one. Six series
~
-------
I
-
CJI
I
Coal
Air
To Cooling
Tower
Acceptor
Condenser
- - - -)
c::::)
- (: ::)
L-
CJL-
a.CJ
~-
",0
CJCJ
eX
L-
CJ
-
~
E
o
CJ
-
-------
~1 ~
- 0
N ..--
"U
~
- 0
- 0
o .Q
- Q)
an Q)
~
LL
I: 't:
tD 0
- a.
N a.
:J
en
: ~
o ~
an '0
II)
x 0\
; 0 -
tD -e N
c." . -
~
Explosion Relief
To Cyclones
Bed Ash
Filling Point
36"x 18" Cross - Sect ion
Throughout
11. Temperature Control
Tubes, 2,'/8 in 0.0.) on
4'/2 in Triangular - Pitch.
Coal
Acceptor
Fines Re,cycle
Air Distributor
36"
Fig. 4. 3 Task I.
The
36 in. combustor
-------
---. - _._-~ -~._._~-~-~
-,
of 'tests were carried o~t,as shown in Table 4010 ~h~ plant was
operated continuously for periods of up to 5 days soparated by
one or mere weeks du~ing which the coal and additivo were prepared
and the plant was checked. Test Series 1 required three periods of
operation, designated Series 1.1, 1. 2 and 1. 3, but ~he remaining
test series were each completed in a single operating period.
Typical size distributions of the coals and additives as prepared
are given in Table 4.2.
Tab.le 4.2", Typical size distributions of prepared ceals and additives
.... -0-
I . Percentage in size range Coizes in ~m)
Material Median
+3175 +1680 +1003 +500 +250 +~2S +66 +45 -45 ~m
I' -3175 -1680 -1003 -500 -~50 -125 -66
Pittsburgh coal, -i in; 0.3 17.7 17.1 22~6 15.0 ~.5 7.7 3.5 6.6 648
I Pittsburgh coal, -10 B.S. 0 0.2 11.9 34.4 21.2 U.9 8.8 3.1 7.5 456
We1beck coal, - i in 0 25.9 20.2 18.2 9.7 ().3 5.9 2.7 11.1 890,
I Limestone 18, -i in 0 11.1 14.7 20.9 20.0 U.6 7.1 2.6 8.0 453
Limestone 18, -10 B.S. 0 0.1 5.9 15.1 22.0 2~.5 12.2' 5.0 17.2 207
I Dolomite 1337, as rec'd. 0 14.9 28.1 30.3 8.9 ~.2 5.7 2.3 5.6 874
Dolomite i337, -10 B.S. 0 0.9 7.1 12.6 9.3 U.O 27.6 10.3' 13.2 122
I U.K. limestone, -A in 0 11.1 13.9 16.3 22.7 U.2 8.2 2.9 10.7 391
.
4.2.3
Result~
Pittsburgh coal and Limestone 18
Test Series 1, 2 and 5 were carried out using.Pittsburgh coal and
Limestone 18 without fines recycle. The results obtained are summarised
4.2.3.1
in Tables 4.3 and 4.4
There was little difference in S02 reduction ~etween the temperatures
o
of 1470 and 1560 F and all data at these temperatur~s,are shown in Fig. 4.4.
'Increasing fluidising velocity over the range 3 to G ft/s affected a
substantial decrease in S02 reduction, but changing the limestone particle
size had little if any effect. The 802 reduction f~ll by 30% when the
.0,
temperature was decreased to 1380 F in Test 1.4.
- 17 -
-------
Table 4.3
Summary of results obtained with Pittsburgh coal and Limestone 18,
prepared to -1680 ~m, without fines recycle (Task I)
F
00
Test No. 1.1 1.2 L3 1.4 1.5 Datum1 1.6 1.7 1.8 1.9 Datum' 1,10 1.11 i
I
I
Fluidising velocity, ft/s 4,0 4.0 4.0 400 4,0 4.0 4.0 4.0 4.0 400 3.0 3,0 3.1 'I
of I
Bed temperature, 1470 1470 1470 1380 1560 1560 1560 1560 1560 1560 1470 1470 1470 i
Bed height, ft 2.3 2.2 2.2 2.2 2,.2 2.2 2,2 2,1 2.2 2.1 2.3 203 2.1
I
Oxygen in. off-gas, % 4.1 303 4.3 3.8 304 4.0 4.2 301 3.3 I 2,6 206 I 3.0 209
Unburnt carb0n, % 7.8 9.2 I 7,2 7,8 6.2 I 7.6 7.9 7,9 8.5 8.2 6.7
- ! - J
I 1652* I
Maximum freeboard temp, of 1593 I 1562 1591 1567 1638* - 1634 1648* 1657* - i 1582 1582
I i
I I
i 1 I
Ca/S mol ratio 0 0 2,2 2,2 I 2,2 0 1.3 202 3.3 I 1.2 0 I 2.2 3.3
I I
S02' ppm 1750 2050 400 1020 360 2100 880 510 180 840 2020 330 42
S ret.ention, % 9 18 83 56 85 - 61 78 93 65 - 86 97
S02 reductiont % 0 0 81 50 83 0 58 76 91 60 0 84 98
C1-, ppm - 72 ... - - - 61 86 24 21 - 49 32
NH3' ppm - 3 - - . - - 0 6 2 0 - I 2 2
i
I
* Limited by injection of steam
T Based on a datum level of 2050 ppm
1 ILl<>,'~v~ ~~/-.-- ~~- Jt.~ ~;
-------
Table 4.4
Summary of results obtained with Pittsburgh coal prepared to -3175 ~m and
Limestone 18 prepared to -3175 ~m and -125 ~m, without fines recycle (Task I)
I-'
1.0
Test No. Datum 2.1 2.2 2.3 Datum 2.4. 2.5 Datum 5.1 5.2 5.3 5.4 5.5
F1uidising velocity, ft/s 8.1 8.1 8.1 8.0 4.0 4.0 3.0 4.0 4.0 8.0 7.8 8.0 8.0
Bed temperature, of 1560 1560 1560 1560 1560 1560 1560 1560 1560 1560 1560 1560 1560
Bed height, ft 2.1 2.1 2.1 2.1 2.1 2.1 2.0 2.1 2.1 2.2 3.5 3.6 3.6 .
Oxygen in off-gas, % 2.2 2.9 2.9 2.5 3.0 3.2 2.6 - 3.2 3.2 2.4 2.8 3.0
Unburnt carbon, % - 13.5 12.4 12.2 - 8.4 - - 8.4 11.4 14.8 18.6 -
..
Maximum freeboard 0 1634* 1634* 1634* 1688* 1627* 1603* 1600* 1596*
temp, F - - - - -
Maximum limestone size, ~m - -3175 -3175 -3175 - -125 -125 - -3175 -3175 -3175 -3175 -3175
Ca/S mol ratio 0 1.1 2.3 2.9 0 1.0 1 0 1. 7. 1.6 1.9 5.7 6
S02 ppm 1830t 1200 900 650. 2100+ 1050 1050 2400'r 850 1180 1040 80 11
S retention, % - 44 62 73 - 54 - - 67 55 62 97 -
S02 reductiont % 0 34 56 68 0 49 49 0 65 51 57 97 100'
I
, (550
NOX, ppm 424 470 515 - - - - 305 I - 445 - 560
(325
C1-, ppm - 23 30 40 - 20 - - I 60 65 65 30 -
NH3' ppm - 4 4 0 .- 0 - - I 2 4' 2 3 -
I
* Limited by injection of steam.
t The sulphur content of the coal varied during these tests and the datum
estimated for' each test from the coal's sulphur content, as follows:
S02 concentration was
Test 2.1
Tests 2.2 to 2.5
Tests 5.1 to 5.5
1830 ppm
2050 ppm
2400 ppm
.1- ,~.!~'-l(.i#t'~ .:&-C[?":'"'/--~fi.(, ~
'r-
-1_- .~-/,:,~'r;,;.
-'-"'~":":'-. J~:" -~ :-~,-
-------
100
90
'/
.
80
c 70
0
.-
en
en
E60 .
Q) .
N /
o
V)
50 /~
c
c
o
.-
t; 40
:J
1J
Q)
L
o! 30
U1
Pittsburgh coal
Limestone 18
o .
Bed temperatures: 1470 & 1560
2
Symbol Velocity Size Bed depth
.ft/s f;4m ft
rm 8 -3175 n
o 4 -3175 2
. 4 -1680 2
, 3 -1680 2
~ 3&4 -125 2
o
1
2 3 4
CG/S mor ratio
5
6
Fig.4.4. Task 1. S02 reduction in the 36 in. combustor for
Limestone 18 and Pittsburgh coal. 1470-1560oF
-20.-
-------
In these tests the concentrations of CO andCR4 were below ~he
minimum level of detection of 200 ppm. At 4 ftls the bed temperature
+ 0
was uniform to - 10 F, but there was a tendency' for the maximum
temperature to occur above the coal feed point. .' At 8 ft/s with a
2 ft deep bed there was a hot spot above the coalfeed,aqout 600F
, f
above the mean bed temperature. This did not occu~ in the deeper beds
used in Test Series 5. A high proportion of the limestone was
e1utriated from the bed; about 80% at 4 ft/s and, 90% at 8 ft/s.
. \ .
Gas samples. taken 2 ft above the.bed in.Te.st '1.1 and in Test
., \
Series 5 showed a marked radial distribution of S02 concentrations,
as shown in Table 4.5.
Table 4.5
Variation in S02 concentration 2.ft above the bed
S02 concentration, ppm
Test No.
Above feed 3 in radius 6 in r'adius 12 in radius Stack
,
1.1 2800 - 1750 1400 1750
Series 5 datum 2350 1950 i250 r. - 2400
" I
5.1 1300 - 400 .. ~ 850
5.2 2150 - 1409 " - 1180
4.2.3.2
Pittsburgh coal and Dolomite 1337
Test Series 4 and 6 were carried out using\P~ttsburgh coal and
, \
Dolomite 1337. The tests were carried out without, 'fines recycle,
"
except for a single experiment, Test 4.6, which,,;was carried out in
,\ .
order to allow comparison with the pressurised.~bmbustor, Task II.
The results obtained are summarised in Table 4.6. .,
As with Limestone 18 there was little diffkrence between S02 reductions
obtained at 1470 and 1560oF. However, in contrast.to limestone, there was
" 0
no decrease in reduction when the temperature.wasreduced to 1380 F in
: ,
Test 4.2 (cf Tests 4.3 and 4.4). All the data obtained at the three
temperatures are shown in Fig. 4.5.' As with lim~stone, increasing the
f1uidising velocity caused a substantia1'decrea~e in S02 reduction. At
8 ft/s, and a Ca/S mol ratio of 5, increasing t~ebed height from 2 to
7 ft appeared to give a small improvement in th~'alre~dy, high S02 reduction.
- 21 -
"
~.. '
t.
-------
Table 406
Summary of results obtained with Pittsburgh coal and Dolomite 1337 (Task I)
!
Test Noo Datum 4.1 4.2 4.3 4.4 4.5 406++ Datum 6.1 602 6.3 604 6.5
i
Fluidising ve10city, £tIs ( 400 4.0 4.1 3,8 4.0 201 2.2 8.0 8.0 8.0 8.1 800 I 8.0
! ! I !
! i i
I Bed temperature, :)F I 1470 I 1470 1380 1560 I 1470 1470 I 1470 ! 1560 1560 1560 1560 i 1560 I 1560
I I I '
I Bed height, ft I 2.1 i 20] 2.1 201 2,1 3.7 I 3.8 207 207 207 400 I 7 5.5
! Maximum c,Jal i ! I
size, 'um ! 1680 I 1680 1680. 1680 1680 1680 1680 3175 3175 3175 3175 I 3175 I 3175
i ' I ' \
I I ! ' ,
I i I I I I
I I i I I i
I Oxygen in oft-gas, %. t 2,5 I 3.,0 3.7 209 3.8 2 <.8 3.1 2.9 300 3,3 304 207 3.0 '
I I I I !
!
Unburnt. <"arbon, % - j 9,1 8.0 11005 '9.0 8.0 0,7 12.9 12,9 15,0 11~(5 1.5.6
1M, « b d ! I I
i I ! f
:J -, i 1591 1530 , 1560* 1594 1510 1522 1648*/ 1638* 1629 1636
I ax:unum tr.ee car temp, F - i l620 i
! ! I \ , i
I I
1680 '11680' I' 1680
3,1 12.6 2.7
; I
620 I
I
75 j
, I
72: .
I
44 I
I
6 I
.
I
!
! I Maximum d.,) l':;mi t.e
~; I CalS m,:,l ratie
. !
, I S02, ppm
I ,
I S retenti.,;)o., %
I 802 reduction, %
size, ~m
I
i
,
I
!
;
I
I
~ 2240
f
\
i
!
380
85
83
o
o
NOX, ppm
I 234
I 46
\ 34 I
~ i
! Cl-,
!
I NH3,
l
ppm
ppm
244
54
6
I
i
!
!
I
i
, 380
!
I 86
:" 83t
I
I
I
42 I
6 I
I
I
! 1680
, .
I
1680 I
2.2 j
I
~
! 2.7
,
600 I 580
76 I '75
73' 74
I.
360 I
29 I
7 I
j
,
*
With fines recycle at about five times the coal feed rate.
Limited by injection of steamo
+'t'
.j.
Based on datum level of 2240 ppm.
Based on datum level of 2190 ppm.
x,
I
I
I
16801
1.61 o.
20 II' 2190
99 -
99" 0
I
392 I
38 I -
8 i
I
i
I '
! 3175 I
II 2,5 !
840 I
I 67 I
I 6,ixl
I I
I 3175 !
503 !
I
260 I
,
90 I
!
88xI
I
!
I
3175 3175 i
3175 i
!
5.4 5.2 5.0 '
~
280 155 I 280 \
.
!
I
88 94 I 89,!
87x 93K I 87 'i
I j
I
i
390 390 360 400 425 !
!
62 59 59 I 5S 60 !
I
I
I i
4 9 15 I 7 12 i
j
, !
L
-------
100
90
eo
c 70
o
.-
'"
'"
; ~ 60
N
o
4/)
. c ~O
.-
c
-0
-
+'
U 40
:::J
'0
t)
L.
-.! 30
2
10
o
Fig. 4.5 Task I
,
/
, III/-
. m~'
,/
Symbol Ve loclty Size . Bed depth
. ft/s J-im ft
(Q] 8 -3175 n
. 4 -1680 2
, 2'15 -1680 4
R 2.25 -1680 4
withrec)'C1e
1
234
Co/S mol ratio
6
.:~
S02 reduction in the 36 in. combustor for.
Dolomite 1337 and Pittsburgh coal. 1380-1560 of
-23-
-------
Table 4.7
Summary of results obtained with We1beck coal and U.K. limestone
prepared to -3175 ~m, without. fines recycle (Task I)
3.1 3.2 3.3 3.4 I
Test No. 3.5 3.6
F1uidising velocity, ft/s 7.9 8.0 8.0 7.9 8.0 4.1
Bed temperature, of 1560 1560 1560 1560 1560 1560
Bed height, ft 2.1 2.1 2.1 2.2 3.8 2.3
Oxygen in off-gas, % 2.9 3.0 3.4 3.4 3.3 3.1
Unburnt carbon, % 10.9 9.7 9.2 11.9 12.3 6.2
Maximum freeboard, of 1650* 1636* 1648* 1647* 1611* 1656*
CatS mol ratio 0 1.8 2.1 2.8 2.8 3.0
S02 by iodine, ppm 1480 910 740 540 540 420
S retention, % 14 45 53 65 68 73
802 reduc tiont % 0 38 50 64 64 72
I
1 C1-, ppm 430 565 455 380 455 410 I
. I 8.5
I NH3' ppm 11 8 11.5 13 10 I
!
i ~
* Limited by injection of steam.
t Based on datum level of 1480 ppm.
- 24 -
-------
Fines recycle at 2 ft/s resulted in a dramatic reduction in S02 emission
to only 20 ppm at a Ca/S mol ratio of 1.60
The effect of bed temperature on S02 redu~tion was investigated by
varying the bed temperature over the range 1420 to l6200F ~hile maintaining
all other operating conditions as in Test 6.3. The S02 concentration
indicated.by..theinfra-red analyser responded almost immediately to
changing temperatures and the variation ofS02 reduction with
temperature is shown i~ Fig. 7.7,
The fines recycle rate in Test 4.6 wasab9ut..five.times the coal
\ .
feed rate. . The CO and CH4 concentrations in the,.off-gas were, as .before,
below the minimum level of detection of 200 ppm.. .Tbe.bedtemperature
'f 'h' b + 1 0
was un~ orm to w~t ~n a Qut - ° F at a fluidisingvelocity of 4 ft/s
+ 0
and within - 20 F at 8 ft/s. Over 90% of the dolomite feed was
elutriated from the bed at all operating conditions other than with
fines recycle. Even with recycle at a fluidising ~elocity of 2 ft/s,
about 80% of the. dolomite eventually passed through the primary cyclone.
It. .was not .possible. to maintain a bed of .more. than'.. 6 .ftdeep, even when
feeding coarse.dolomite at a CatS ratio of five.'
'.
402.3.3
Welbeck.coal and U.K. limestone
Test.Series 3 was carried out using Welbeck coal' and U.K. limestone
prepared to minus l.in, (-3175 ~m),. The effect.of the Ca/S.ratio and bed
height were investigated at 8 ft/s and a.single'test.wascarried out at
4 ft/s. The results of these tests, which were all without fines recycle,
,
are summarisedin Table 4.7.
The effect of bed temperature was investigated by varying it over
the range 1470 to1630oF during a period.of two;hoursand monitoring
the S02 emission using the infra-red analyser.. ~Asiwith,dolomite, the.
response to changing temperature was almost instan~aneous, (Fig. 7023)0
The.S02.reduction obtained is plotted against the ~ed temperature in
Fig. 7.7.,
I
'I
.r
- 25 -
-------
The bed temperature was uniform to within:!: 200F and there was
no 'hot spot' above the coal feed point.
About 80% of the limestone
was elutriat.ed from the bed at 4 ftis, and about 90% at 8 hIs.
NOx
measurements were not made during these tests. The CO and CH4
concentrations were again below the minimum level of detection of 200 ppm.
The 48 x 24 in Pressurised Combustor (Task II)
4.3
Tbe work carried out on the pressurised combustor is des,cribed in
detail in Appendix II.
4.3.1
Description of the plant
A flow.diagram of the pressurised combustor plant is shown in Fig. 4.6
and details of the combustor are given in Fig. 4.7. The.combustion vessel was
a refractory-lined shaft of approximately 48 x 24 in cross-section, 10 ft
high, housed in a cylindrical steel shell, 22 ft long x 6 ft lod. having domed
ends and capable of withstanding an internal pressure of up to 100 psig. The
walls of the combustor were designed to withstand.a.pressure.differential of
+ .
about - 5 psi.
The air . distributor plate was fittedwith.4,65 bubble. caps on a H in
square pitch, each containing four horizontal jets, 1 in above the
distributor plate. .Co~bustion air, preheated.to.400~5000F, was. supplied,
via a plenum chamber, through the distributor plate. Combustion was
initiated by preheated combustion air derived from three gas burners
located in the plenum chamber. After coal combustion was established,
the gas burners were isolated.
The coal and.additive were premixed and.fed.viaa pressurised
hopper.system to four pneumatic feed lines, each.of.which terminated in a
nozzle that protruded 4 in into the bed.
from each end of both of the 48 in walls.
The nozzles were located 12 in
Heat transfer tubes were
innnersed in the bed on a
row, with horizontal and
respectively.
Twelve of
staggered pattern of 18 rows of 7 or 8 tubes per
vertical pitches of approximately 6 in and I! in
the tubes were air-cooled, six preheating the
combustion air and six forming corrosion specimen assemblieso The
co'rros~on specimen assemblies were similar to those described in '::,ection
4.1. The majority of the remaining 53 tubes could be used as water-
cooling circuits, the remainder being d:'::;:;Ll.=:::;,. The bed temperature was
controlled by varying the number of water-cooling circuits in use.
- 26--
-------
Cooling
water
Cool
I
l\:)
-J
I
Air
J,
J ..
. .....
I _... . -.-
-
Nitrogen
'. - ~~.. '.~-'.- - .
mbustor
. .
g: Slort-unontv II
Steam ---.--f-J I
, I
: I
rc!>wn's _Sta!'!:~~ O!'~J
QOS .
.
Cooling
water
disposal
Disposal
stack
I Pressu re
let-down
Dirty
water
disposal
w. .~,F~.9 4.6 Task n. Flow diagram of 48in x 24 in pr-essurised
combustor plant I
Cascade
section
,Water
sprays
ft -blast
injection
Water
sprays
Wet
sample
SI udge
collection
, Collection
of fines
Vent
-------
10,
110
120
130
14"
150
16.
170
18,
19.
20.
NOTES ON THE DIAGRAM OF THE 48 in x 24 in PRESSURISED COMBUSTOR
L
Water supply to water-cooled circuits
2,
First-stage dust-collecting cyclone
3,
Internal recirculation cyclone
4,
Balance air supply
5 c
Gas burners for start-up
6,
Bed removal pipe
7,
Pressure shell
8.
Cooled liner
90
Combustor casing
Second-stage dust~collecting cyclone
Air supplies to air preheating tubes and corrosion assemblies
Cascade housing
Sampling point for alkali content of gases
Mixing baffle
Sampling po~nt for NO gases
x
Spray nozzles
Deposition probe (Not used in the APCO programme).
Dust sampling and 02' C02' CO, probe
Exhaust to pressure 1et~down valve
Dip-leg from recirculation cyclone
- 28 -
-------
16
17
12
OIT"" IMOW... ("-SCAM 01' TUftllle
k"OII ,~ 'T1.O't''''' 1801'ON,
CoalllOllOH ...NO DI'OIITION
2 10
3
-9
-8
7
-~
Coal
4
20
"
"
.,
:g 0°0 0°0 0°0 0°0 .~
o 0 I) 0 ...;
00000'
~o 0°0°0° 0 °0°0 °0°0 °0;'4.
00000
o 00 00 00 0° o'
°0 °0 °0 °0 °0"
o 00 00 00 0° 0
°0 ""0 °0 °0 °0,'
Ash
KCYtON 0triI ~ A' IH(JWIIt8 ""'-"NT Of
",.1 .. fllUDIIID MD
Fi 9 4.7
Task n.
The
48in x 24 in
pressuri se d
combustor
-29-
I
I
J
-------
.-"
. .~! _.-
.~.-
++-- .
~.-
-~j -'-
. L-/
'---- - I -
--I
.
I
(
)
.
I
----1
Target rods
\
,
\
I
Combustion gases
L.J-.J
Fig. 4.8 Task n. Arrangement of blade cascade an d
target rods
-30-
-------
A bank of four rows of uncooled tubes, on a nominal.3 in triangular
pitch, was located immediately above the bed to' act as a hcrizontal baffle.
Water-cooled circuits were provided in the, freeboard zone to limit
the exit gas temperature to l5000F.
The waste gases passed through a 10 in.diameter internal recirculation
cyclone with a dip-leg passing back into the bed'. The amount. of material
being recycled back t.o the bed was unknown" but 'a~ailable evidence suggests
that. it was small and the system probably behaved as a non-recycle system.
Subsequent.ly the partially-cleaned gases passed th~ough two 10 in diameter
cyclones arranged in series from which collected, fines were removed con-
tinuously by positive blowdown systems. The cleaned gas'es were then
accelerated t.hrough a gas turbine blade cascade, (Fig. 4.8) quenched and
cleaned in a venturi scrubber and hydroclone befpre being exhausted to
the stack via a pressure let-down valve.
The main gas sampling point was immediate~y after the cascade. CO,
C02 and CH4 were continuously monitored on infra-red analysers and 02 on
a paramagnetic analyser. S02 was determined by:bubbling through iodine and
in the later tests was also continuously monitored using a Hartmann Braun
infra-red analyser. NOX was measured using both the Saltzman method and
the BCURA NOX box. All the, outlet solid streams ~ere sampled and t.he bed
level was maintained at the required level by periodically extracting
material directly from the bed.
4.3.2
Experimental programme
, .
Tests were carried out using Pittsburgh.cda1:.with Dolomite 1337 and
Limestone 18, and using Welbeck coal with. U.K. limestone. Five series of
tests were carried out, as shown in Table 4.1. :',The plant was operated con-
tinuously for periods of up to 4 days separated!:by ::me <;>r more weeks during
which the coal and additive were prepared and the plant was checked. Size
distribut.ions of t.he coals and additives are given in Table 4.8
I'
- 31 -
-------
Table 4.8
Typical Size Distributions of Prepared Coals and Additives (Task II)
I Percentage in size range (s ize~. in !.lm) ,
I Median '
I
Material -1588 -794 -500 size !
I -211 -152 -75
+1588 + 794 +500 +211 +152 +75 urn
Welbeck coal 0 14.4 21. 7 41.7 4.8 5.9 11.5 400 ~
Pittsburgh coal 0 21.2 17.4 22.1 12.5 12.4 14.4 360 I
~
Do lomi te 1337 (Tes ts 2 and 3) 0 75.5 17.8 2.4 0.9 1.2 2.2 1100 1
! Dolomite 1337 (Tests 4 and 5) 0 27.8 17.5 13.3 8.5 12.6 20.3 400 t
I Limes tone 18 0 16 15 2.1 18 10 20 270 ~
i ,
Fluidising velocity, bed depth and bed temperature were
maintained substantially constant throughout the programme at nominal
values of 2 ft/s, 3.75 ft and l4700F respectively. Operating pressure
was either 3.5 or 5 atm absolute.
The test conditions were largely predetermined by the coal/additive
mixture prepared before a test. Measurements were made continuously
throughout a test series and test periods were selected in retrospect to
. .
coincide with periods when the S02 emission had reached equilibrium. The
duration of these. test periods was from 4~ to 31 hours.
4.3.3
Results
The main results are tabulated in Table 4.9 and the data on the
reduction of S02 emission are plotted in Fig. 4.9. The main conclusions,
as far as Task II is concerned are:
(1)
Reduction of S02 emission was not affected by pressure
in the range 3~ to 5 atm.
Altering the median size of the additive (in this case
Dolomite 1337) from 1100 to 400 micron did not affect
(2)
the reduction in S02 emission. The change in size was
produced by widening the size distribution, i.e., by
increasing the proportion of fines.
-. 32 -
-------
Table 4.9.
Summary of results obtained on 48 in x 24 in pressurised combustor (Task II)
F1uidising velocity:
Bed temperature:
Bed depth:
Top size of coal & additive:
2.0 ft/s
14700F
3.75 ft
1588 ~m (1/16 in)
I
(..:)
(..:)
I
Test 1.1 1.2 1.3 1.4 1.5 2.1 2.2 2.3 2.4 3.1 3.2 4.1 4.2 5.1 5.2 5.3
Coal We1beck Pittsburgh Pittsburgh Pittsburgh P ittsburg~ol.
Additive U.K. Dolomite Dolomite 1337 Dolomite 1337 Dolomite 1337 Limestone 18 1337
Duration of test period
h 41 12 271 141 26 131 21 29 41 11 8 251 ,31 23 16 8
Pressure 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 5 5 '5 5 5 5 5 5
02 in exhaust % 3.8 4.2 4.6 5.1 4.8 3.0, 4.5 4.9 6.3 5.3 5.1 5.8 5.3 3.6 3.8 5.3
Carbon loss % 3.1 2.9 2.2 2.2 1.4 2.9 3.0 3.9 2.7 '1.8 1.1 0.9 1.3 1.2 0.5 0.8
Median size of additive 750 1100 1100 400 270 400
~!Il
Ca/S mol ratio 0 0.77 0.73 0.91 0.92 1.41 1.88 1.93 1.88 1.65 1.80 1.54 2.02 1.87 2.38 1.08
S02 in exhaust ppm 1020 620 430 100 150 210 90 70. 110 100 100 130 50 690 470 310
Sulphur retention % 9 52 62 90 86 92 96 97 95 95 95 94 98 67 79 85
502 reduction % 0 39 57 89 85 89 96 96 95 95 95 94 98 66 77 85
'so 3 in exhaust ppm 9 14 14 8 5 1 0 1 - - - - - - - -
NO in exhaust ppm 40 - 80 120 130 - -' 80 130 90 100 160 - 140 190 90 90 140
x
C 1- in exhaust ppm 490 590 640 460 330 80 70 50 - - - ~O 40' 30 50 50
Na in exhaust (gas)ppm 1.1 0.9 0.7 0.4 0.4 0.7 0.4 0.2 - - - 1.1 1.1 1.6 2.7 0.7
K in exhaust (gas) ppm 0.3 0.2 0.2 0.2 0.2 0.2 0.2 0.1 - - - 0.3 0.4 0.7 0.9 0.2
Note: Equilibrium S02 emission may not have been reached in Test 1.2.
-------
100
,
I
~ I
~. 9
'C ~l
o :;;;,0
en CJ,O X
en ~/
~ 80 Cbl
::,
L. .&1
::} .f:::
~ (hI
a. ~I
-
::} Cb
en 70 ~I
c ~I
c :;j/
0 tl
+'
~ 60 ~I
1) I
& / °
I
, ~y
. .
./
,
50
,0
1'0
CalS mol ratio
2'0
3'0
Fluidising velocity: c 2ft/s
Bed depth : 3 -75 ft
Bed temperature: c 1470°F
Acceptor Pressure: atm. abs.
Co al Median size: 31/2
Type micron 5
Welbeck U.K. Dolomite 750 0
Pittsburg. Dolomite1337 1100 ~ .
Pitt sburg Dolomite1J37 ~OO X
Pitt sburg limestone 18 270 .
Fig. 4.9 Task n.
Reduction of sulphur emission in the
pressurised com bustor
-34-
-------
(3)
The percentage S02 reduction with Welbeck/U.K. dolomite
appeared to be slightly higher than with Pittsburgh!
Dolomite 1337, although the range of Ca/S ratios
investigated did not overlap sufficiently for a firm
conclusion to be drawn.
(4 )
Reductions obtained with Limestone 18, although con-
siderably lower than with dolomite (on a Ca/S basis),
were surprisingly high in view of the fqct that
limestone does not calcine at these conditiorts.
The time taken to reach an equilibrium S02 emission was
substantial when dolomite was the additive - about 20 to 50 hours
when the initial bed contained no dolomite.
The approach to steady
state is discussed further in Section 7.1.12.
The major loss of combustion efficiency in all tests was in
the carbon of size less than 50 micron elutriated from the bed.
Combustion in the freeboard, based on heat balance between the
top of the bed and the top of the freeboard, amounted to 2 to 3% of
the coal heat input.
The temperature of the bed showed a small, but consistent, rise
of 30 to 400F between a level near the bottom of ,th~ bed and one near
the top. At any particular level the variation in 'temperature was
+ 0 .
about - 20 F, although no measurements near a coal feed point were
made.
The dust burden of the gases passing over ~he cascade was about
150 lb dust/l06 Ib gas (about 0.1 gr/scf) when burning Pittsburgh coal.
Tbis was about twice the dust loading obtained when burning Welbeck
coal. The inert input was considerably higher with Pittsburgh coal,
partly because of its higher S content, and partly because of the
higher Ca/S ratios generally used with the Pittsb~rgh coal.
The dust was sized by Coulter count and found generally to have a
median size of 6 to 8 micron with about 90% less than 30 micron. As
with all determinations of size distributions of this order, however,
the value of the data, particularly from the point of view of turbine
- 35 -
-------
erosion, is open to question.
Earlier tests. with Welbeck coal had
shown that most of the larger dust particles were in the shape of
'platelets' which would not be regarded as erosive. The accuracy of
the Coulter counter technique for this type of particle is open to
doubt.
The cascade blades were not cleaned between tests and the
appearance did not vary
a light accumulation of
adhering to the concave
greatly from test to test. Usually there was
dust on the leading edge with flecks of dust
faces and negligible deposition on the convex
faces. There were no signs of sintered deposits or of erosion.
Photographs of the cascade blades are given in Appendix 2.
The target rods usually had a small knife-edge of deposit on
their leading face, and had developed.a considerable 'tail'. Cross-
sectioning and examination under a microscope revealed no evidence
of corrosion or erosion.
The data obtained on corros~on of tubes in the bed is given in
section 4.7, together with data from other pilot plants.
4:4
The 27 in CombustGr (Task III)
The work carried out on the 27 in combustor is described in detail
in Appendix III.
It was originally intended that only corrosion investigations would
be made on this plant, but subsequently the objectives of the work were
extended to include investigations into sulphur emission and when possible,
nitrogen oxides emission.
4.4.1
Description of the plant
A flow diagram of the 27 in combustor plant is shown in Fig. 4.10
and a section through the combustor in Fig. 4.11. The cylindrical
section containing the fluid-bed was 27 in diameter and the shaft above
this section diverged to 46 in diameter for the lower 2 ft of the
freeboard before returning to a cylindrical section 24 in diameter.
The height of the bed zone was 2 ft and the total height of the freeboard
was 11 ft. The walls of the bed zone and freeboard consisted of twenty
- 36 -
-------
rnal
Heat cyclon
exchanger
4----
--+
To stack
.--.......-- _. --.' .
Internal
cyclone
I.D. fa n
Freeboa rd
I
1:.).
-1
I
Lock
hopper
Cool ing
water
circuits
Fluid-bed
Corrosion
Probe cootin
Air distributor
-.
Air
inlet
I Bed
, material
to wa ste
+-
F. D. fan
-.J'. . ,'< -- . 4..,-
Fig. 4.10
Task m. Flow diagram
of 27in. combustor plan t
Coal
storage
hopper
Weighing
conveyor
Limestone
Tran sfer
hopper
+-- Conv.eying
aIr
I Cyclone fines
t to waste
~
Conveying air
-------
. .,
hopper actuator
Dip leg Bed load ing
lock hopper chute
)(
0
L.
a.
c Sight glass
t
0
...
-------
water-cooled jackets. Combustion produc.ts.leaving the freeboard
passed. into a vertical water-c.ooled heat exchange-rbe.f:)re enter'ing the
external dust cleaning cyclone.
The bed was supported on a bubblecapdist['ib1)~;,i plate, air for
combustion being supplied via a plenum chamberbcatE:~d directly bel::1w
the distributor plate. Crushed coal and additive were metered into a
hopper, transported pneumatically to the combustn and injected into
t.he fluid-bed via a 'y' 'feeder. The two 1. in diamete'r . arms of the 'y'
formed injection points which were radially. equidistant.. from the axis
of the combustor at a distance apart of 0.3 ft approximately.
A "mushroom" gasburner fit:ted in the cent.re of the distributor
plate supplied towns gas 'for start up. During.normal operation it was
used to distribute fines recycled from the external cyclone.
A low efficiency cyclone' positioned in the freeboard recycled ~50
micron particles to the bed via a dip-leg and lock-hopper system, whilst
a high efficiency cyclone external to the combustor removed solids
greater than 10 micron from the combustion productsn A controlled
portion of ,the fines collected by the external cyclone could be recycled
pneumatically to the bed via the "mushroom" burner and the remainder was
discharged to collection drums.
Access p~rts for bed cooling tubes, corrosion tube assemblies and
other probes were provided in the walls of the bed section.
The 2 in
diameter tubes were arranged at 5 in centres horizontally and 6i in
centres vertically: the dip leg from the int.ernal cyclone prevented the
use of intermediate tubes to form a triangular pattern: the tubes were
.cooled with either steam or air.
Four corrosion.tub~s were used for
t.he first. two test series.
Higher bed temperatures were required for
the third series and only two tubes were used, to reduce the cooling
load on the bed.
The iodine method was the main ,technique used for determining
the S02 concentration in the gases, but an infra-red analyser was
also used from l~st SerlB5 2 ~nward~ NOX w2£ d€ter~ifled oSJng
Saltzman's method and the BCURA NOX box.
- 39 -
-------
4.4.2
Experimental programme
Four series of tests were carried out at atmospheric pressure
using -3175 micron Pittsburgh coal. The main operating conditions of
these tests are detailed in Table 4.1. Two different American
acceptors were used, Limestone 18 and Dolomite 1337, the former being
examined at two levels of particle size.'
The plant was run continuously for 100,500 or 1000 hours, each
run period being separated by several weeks for refurbishing and
checking the plant plus preparation of coal and additive. Test Series 1
was run at constant conditions throughout, whereasqtheremaining three
.....~
test series were sub-divided into periods having different operating
conditions. . .The actual test periods in Test Series 2 and 3 were
normally of 16 hours duration, preceded by a settling down period of
minimum duration of 18 hours.
In Test Series 4, which ran for 100
hours only, the settling down period for the tests was 6 hours and
the actual test period was 4 hours.
.4.4.3
Results
Test Series 1 was a 500 h corrosion test.carried out under constant
operating conditions without limestone addition. The corrosion data
obtained in this and subsequent test series are given in section 4.7,
together with corrosion data from other pi1ot.plants.Other results
obtained are given in Table 4.10.
Test Series 2 was primarily a 1000 h.corrosion.test..with addition
of Limestone 18 and was carried out under approximately constant
combustion conditions.
However two sizes of limestone, minus 3175 and
minus 150 microns, were used over a wide range of Ca/S mol ratios, both
with and without recycle of fines from the external cyclone. The
results obtained. in this test series are given in Table 4.10 and a
comparison of the S02 reductions obtained using the two different
sizes of additive is shown in Fig. 4.12. LowerS02 reductions were
obtained when using the finer limestone.
- 40 -
-------
100
0/
(0/
o/~
01
I
~ ~o
o 0
-3175pm Limestone
c
o
.- 80
cI)
cI)
I I I
1 2. 3
CalS mol ratio
-150pm Limestone
o
o
1 2 3
CalS mol ratio
4
o With externa I recycle
~ No externa I recycle
Fig. 4 .12
Task m.
502 reduction .in the 27in combustor for
Limestone 18 and Pittsburgh coal. e ftls.
1500-1575°F Test Series 2
-41-
-------
Table 4.10.
Summary of results obtained in Test Series 1 and 2 on 27 in combu8tor. (T..k III)
,.
,
Test No. 1 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9.1 2.9.2 2.10 2.11 2.12 2.1) 2.14 2.U ~ . l' . 1 2.16.2
Operating conditions
Acceptor Nil L L L L L L L L L L Nil L L L L L L L
Acceptor size - C .C C F F F F F F F - C C C C C C C
CatS mol ratio 0 1.58 0.44 3.09 1.86 0.88 2.47 0.74 2.47 1.63 1.45 0 1.09 2.37 3.89 0.62 2.45 1.93 2.47
F1uidising velocity, ft/s 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8
External fines recycle Yes Yes Yes Yes Yes Yes Yes Yes Yes No No Yes Yes Yes Yes Yes Yes No No
Bed temperature, of 1510 1500 1510 1530 1560 1570 1550 1550 1570 1564 1550 1570 1560 1550 1560 1550 1560 1500 1520
Bed weight, 1b 280 250 270 220 210 210 220 230 240 310 250 270 260 220 210 270 210 280 210
Gas composition
C02 % v/v 14.6 14.3 14.1 14.3 14.6 14.8 14.1 14.7 13.9 14.3 11.9 14.9 16.2 15.2 16.9 14.8 15.4 13.7 10.7
CO % v/v 0.72 0.66 0.68 0.69 0.66 0.65 0.72 0.69 0.71 0.54 0.57 0.70 0.68 0.71 0.67 0.75 0.71 0~64 0.48
02 % v/v 3.2 3.7 3.5 2.8 3.9 3.5 3.8 3.9 3.8 4.2 7.6 3.5 3.4 3.6 3.6 3.6 3.8 4.0 7.6
S02 ppm v/v 2050 930 1760 290 1370 1520 830 1610 790 1110 940 16/K> 1220 500 310 1550 530 910 680
NOx ppm v/v 530 430 - - - 380 490 500 390 - - - - - - - - - -
Unburnt carbon % 16 14 16 16 14 9 11 18 7* 20* 7+ 6* 13 12 11 9'" 11 16* 16*
Sulphur retention % 21 69 42 91 52 46 71 44 72 62 57 40 59 83 90 48 83 72 72
502 reduction % 0 55 14 86 32 26 59 21 61 46 54 18 40 76 85 - 24 74 56 67
I
~
~
I
Acceptor:
L denotes Limestone 18.
Acceptor size: C denotes -3175 micron, F denotes -150 micron.
'" No estimate of unburnt carbon in exhaust dust is available.
+ No estimate of unburnt carbon in cyclone fines available.
I -
;<{.~.: ,,""
. .. -... :.J' .
:.~~ ;J~~.~' '~;'..~.
~ 1.~~ '-.
-------
Table 4.1L
Summary of results obtained in Test Series 3 ~n
the 27 in combustor. (Task III)
Test No. 3.1 3.2 3.3 3.4 3,,5 3.6 .3.7 1
,
i
Ope.rating conditions I
i
!
Acceptor L L L L L L D \
I
Acceptor size C C C C C F M I
i
Ca/S mol ratio 2.68 2.69 2.80 2.63 2067 2.,66 L74 I
I
I
Fluidising velocity ft/s 6 11 7 11 8 6 8
External fines recycle Yes Yes No No Yes Yes Yes I
I
,
Bed temperature, of 1530 1570 1550 1550 1660 1560 1550
:
Bed weight, 1b 220 210 220 260 200 260 250
Gas composition i
;
,
C02 % v/v 13.2 14.9 14.9 13.5 .14.0 14.5 14.3
CO % v/v 0.69 0.65 0.63 0.57 0,64 0.72 0..63 :
02 % v/v 3.7 5.2 5.1 6.1 4.4 3.6 4.4
S02 ppm v/v 300 480 380 480 860 ;
i 340 640
NO ppm v/v - 330 510 410 500 470 530 \
x
I
Unburnt carbon % 10 8 12 15 6 14 15
Sulphur retention % 90 80 85 81 I 66 88 76
,
S02 reduction % 85 77 82 77 58 83 69
\
Acceptor:
L denotes Limestone 18,
D denotes Dolomite 1337.
Acceptor size:
C denotes -3175 micron,
F denotes -150 micron.
M denotes -1587 micron,
- 43 -
-------
Table 4" 12,
Summarl. of results obtained in T-sst
the 27 in combust or 0 (Task Ill)
S.e:des 4 on
. 4~4.2 ' -- 46l4> . ~~8 I
'
, Test No, 4,,3 4.,4 405 ' 4,.9
,
. ! .
\
i Operating conditi,.mS ! ' : (
i Acceptor I I
L D D D D D .D t D D i
I I ;
! Acceptor size C M M M M M M ~ M M I
I I i
1 Ca/S mol ratiQ 2,,94 1.89 1088 0,62 0061 L 79 0066 I 4,78 ' 5,56
I i
! I I ;
I I
i F1uidising velocity, ft/s 8 8 8 8 8 8 8 ! 8 I 8 !
!
i I i I
j External fine:s re,',:;yc1e Yes -Y,as Yes Yes Yes Yes y,~s ~ N',,;, I No
I of ~ i
Bed 1430 1420 1540 1560 1680 I 1680 ' 1
temperature, 1430 1 1650 1550 ,
1 I I '
! Bed \Jeight, 1b 350 310 210 230 200 180 2'50 200 210 '
i I I i
- . - I ! i
! Gas I
composition , I :
i i !
I Ct). % v/v 12.,3 I 1405 13.,7 1203 1209 1401 13 0 1 17 05 14,,8 j
I 2 I !
I co I
% 'V Iv 0,64 i 0060 0072 0,74 10075 I 0,69 0063 ; O. is I 0 .6(;
. I I
i ii I
J 02 % v/v 4,,0 i 4.5 403 3.,8 4,.1 3,,8 4,,4 1 3.,7 56
I .
I ' I ~
ppm v/v 550 ' 530 .450 1160 1730 1060 650 750 I 370 J
I S02 ! '
, 1
I i
J I I I ! \
; NO ppm v/v 410 I 480 I 480 480
- ! - - - -. I
1 x
\ !
I ' i
carbon % 2"" I 8~ 5* 00.4* 5* 5* 5* 7'~ 29*
J UnbuI'nt ~ ~I
i I
~ r I ,'- "
, I ~
i Sulphur retention % 78 Ii 78 82 I, 54 31 58 72 69 I 82 !
\: ! j I :1
\ 80- n~du.c.tion % 73 i 7/.j. 78 i 44 16 48 69 64 I 82
I ~ i t
j L r I' :1
., I
Accept_0r ~
L de':lQte~ -:. imeston<:: 18,
D denotes Dolomite 13370
Acceptor si7.e:
C denotes -,3175 micron,
M denbtes -1587 microno
* No estimate of unburnt carbon in exhaust dust is available,
- 44 -
-------
Fig. 4. 13
100
/
/1
c
o
~ 40
u
:J
'U
t)
L 3
~
o
20
10
o
1
2 3 ~
Co IS mol ratio
5
6
Bed temperature: oF' 1420 1550 1650 1680
With Internal recycle - . . -
.With internal and ~ 0
externa I recycle - C
. .-
in the. 27 in combustor for
and Pitt.sburgh coal 8 ft/s.
Task m. S02reduction
Dolomite 1337
(Test Se rles 4)
-45-
-------
The main objective of Test Series 3 was to determine the effect
of £luidising velocity using Limestone 18 by operating at 66 and 11
ft/s, both with and without external fines recycle. . Three additional
tests were carried !Jut with (a) a high bed temperature, (b) fine
limestone and (c) :olomite l337.The results obtained. in Test Series 3
are given in Table 4.11. With -3175 micron Limestone 18, the S02
emission increased slightly with increasing fluidising velocity,
(Tes ts 3.1 to 3,4). Al though re.cyc1e of fines from the external cy clone
increased S02 reduction when operating at 6! ft/s"it had no effect at
the higher velocity of 11 ft/s. The S02 reduction obtained with fine
stone at 66 ft/s, (Test 3.6) was only slightly (2%) below that
.. .
\
obtained with coarse stonen This is in contrast to the large difference
at 8 ft/s in Test Series 2. An attempt to operate_at a fluidising
velocity 'of 4 ft-is was abandoned aft.er partial de-fluidisation occurred.
. .
The main ,objective of Test Series 4 was . to determine the,.Ie.ffect of
temperature on S02 reduction using 1.')olomite 1337.with and without.
external fines recycle. A single test was also carried out with Limestone 18'
at a low bed temperature.. The results are given. in Table 4.12 and the S02
reductions obtained with Dolomite 133'] are.plotted.in Fig.4.U. The S02
reduction obtained in Test 4.7 is high, corresponding to about 100%
utilisation of th~ stene. but there is no reason to suspect the validity
of this test. The 502 reductions obtained at 1650~1680oF were considerably
.1ower than those at &oth 1420 and 15500F, A comparison.between the results
at 1420 and.1550oF.indicates that the effect,of.temperature is dependent
o
on the Ca/S mol ratio, with optimum temperatures of' about 1550 F at a
ratio of 1.8 and of 14200F at a ratio of 0.65. Fig. 4.13 shows that
l~''':':i ,.,..,,:!,:;,,>,:~~.~e.rnal fines recycle impr.:)ved the S02 reduction at afluidising velocity
\.'l:..,j}\"!:l":'i'r!'~<\::Jt8 ft.l So
4.5
The 12 in corrosion combustor (Task IV)
. .
'. ':
The work carried out on the corrosion tfbmbustor is described in
detail in Appendix IV.
',.' ,;
", .
',. .
. .
. .
..;.
f-
...
:" ~ .
": "';;;'''-.
. --,. :." .
-
..
. .'
~'~iT;~;;" :..>
".:~t;.'
'..
.." :',".
''';;'~Y:;' .;
'.
-,46 -
..
.-:
-------
Description of the plant
4,5..1
A flow diagram of the 12 in corrGsi:m ccmbusto.t plant is shown
I
in Fig. 4.14 and the details of the combustocvesse \ . are given in
Fig. 4, L5. The combustor was a 8 ft high mild steel c.. I urnn, refractory
lined, with an internal cross section of 12 in square,
The air
distributor comprised a plate with 9 short.vertical.~ubes arrange~
on a 4~ in square pitch, each tube having six! in diameter holes.
drilled horizontally, around the top, Coal was inje~ted ~t the
centre of the combustor just ab:)ve the air.distributor nozzles.
When limestone addition was required, the limestone ~a~ pre-mixed
with the coal.
A constant bed. height was maintained.by awei:r:, the excess bed
ash pas£ing into a sealed hopper. The off-gases passed through an
internal pr~mary cyclone and external high efficiency secondary cyclone
and then to. the stack. Fines from the internal cyc1cne were recycled
to the bed by means of a dip-leg.
Four rows of 2 in diameter air-cooled tubes wete immersed in
the bed .,:m a 6 in triangular pitch; these remc.ved the excess heat
generated by combusti:m. Four of these tubescompri.,f,ed corrosion
specimen assemblies and these were cooled to the re~~ired. test
temperatures.
Two specimen tubes were also. situated. in. the freeboard
2 f t above the top of the ash bed.
The construction of these
specimen tube as~emblies is described in sectic-n 4.1,),
The plant was. designed to operate continuously for extended
periods with a minimum of staff and was fully automated with
comprehensive safety systems.
Mass balances were nc.rmally carried
out :)ver a period of 100 h. Gas samples .were r.akeo from the
freeboard, about 3 ft above the bed surface, for analysis of C02'
02' CO andCH4by chromatograph. Other gas samples were taken
after the secondary cyclone. These were analysed.for S02 by the
iodine and hydrogen peroxide methods, for S03 by the 'Shell
Condensation' method and a1so for Na, K and Cl-. An the outlet
solids streams. were sampled and solids samples could a1so be
extrac.ted directly from the bed.
- 47 -
-------
Coal and
transport air
Vibratory ----
sieve
~\
Coal
lock,
hopper
Coal
feed
hopper
Air
Coal feed
cyclones
Fines
collection
. cyclone
Internal
fines
return
system
Specimen tube
+- Cooling air
Speci men tubes
+--
-+- Cooling air
Control tube
~
. .
FJuidlsed bed
Air
Gas pre- heater
Ash collection
hopper
Task N. Flow' diagra m of 12.in. corrosion
combustor pia nt
-48-
Fig. 4. 14
-------
~1 a ~ Q or -
",juo"z'n ... .,.a
I
.;..
CO
,
,
....
,
I
,
I
I
I
,
I
I
r'~
~2t.
:-.
:
,
I'
I
~
51""'" ...... ......, .. ~OI."-U . ~..... 'Q
NATIONAL COAL BOARD
k , D. UU'LlSI8OUT
STOKE ORCHARD, CHE1.TENtWoO.
SO<.aII_OfO,..eI'
,
,
,
:.-.
,
I
~t.t ':,
t.t.t- .1
, , :'-.-i
' '~
FACE '0'
8 ~'~T_- II.OPIO f.
C ~:::'-;::'T'=~:'";~~~ ,
0......1., -10
Fig. 4. 15
Task
IV;
"
: I
, ..'~I !
I !~I ! I
ii==r
I I
I
I
I
,
I
,
I
I
I
~
NOTE :-
FOR 01 TAilS OF TU8ES
SU £_N 2SI!I0l!
The
12 in.
corrosion
~
FACE ',.,'
'0. ..0< 2 IIAIICT (...... _rlC8o"'~ ro.
SHTlO") UI OIIG - ...tala/JllI
TO TL[
It FLUID' 5£0 REACTOR
CORROSION RIG. WK. 1
combustor
1\
1\
"
,..(1'",' II
"
"
"
:'
r,
l
"
"
"
:'
"
,...n',.,,' II
, '''c.l'e.
"
"
O---~---'
;:+ ~$-+-:
,I j !,
,.. ,
:_"'_.:t-t~
VIEw ON UNOERSIOE
,---
,
~
1,;---
''''I '.~'
VIEW ON TOP OF Rf."'C"O~
Ol~..,~!t 0' ,...sUL""'O.... "'NO
IU'''"CTORY IN'l880/1/8
\\ .......s.."",,- ',0.'"
10 ....J1. h."w..,'
, &O\.T- 'fZ"-"... 1~'
. c...s..(,
T STuD ~ w.." .. t'\.G
, 80110'" P\.II'I
~ '.O.H ~U8[ CO'llf."
.. 80TTO'" StC"ON
) TOP fuel CO'llIR
1 TO. SIC no...
I
C ", 0....... \.- 10(""
~
".'/.. QAT(
--~
,..cc'O'
OF RE AC TOR
'.CI '0'
l~4 au
IS. on
..& on
IN 1:..a , 1
IN 1:880/1/.
IN 1880/5 t
IN le80 5 S
IN 1..0 3
I'" Z.eO/5 8
DA~IL (.2..0",1
-------
4.5.2
Experimental programme
Three corrosion tests were carried out and these wer~
designated Test Series 1, 2 and 3, as shown in Table 4.1. Each
test series consisted of an extended period of operatlJn under
nominal constant standard conditions. These were a fluidising
velocity of 3 ft/s, a bed temperature of l560oF, a bed depth of
2ft, total recycle of primary cyclone fines and an oxygen
concentration in the flue gas of about 2%. Test Series land 2
were carried out to obtain corrosion and deposition data using
Pittsburgh coal without additive and were of 100 hand 500 h
duration, respectively.
Test Series 3 was carried out to determine
~" ,
the effect on corrosion, if any, of addition of Limestone 18 to
the bed and was in all other respects a duplicate of Test Series 2.
The coal and limestone were both prepared to a top size of 1680 j.im.
4.5.3
Results
The data on corrosion are given in Section 4.7 together with
corros~on data from other pilot plants.
4.5.3.1
Test Series.1 : 100 h test without limestone
The combustor was' operated continuously except for one shut-down
of 6 h durationcaused.by a coal feed stoppgge due to the coal feed
rotary valve jamming. A short preliminary test had shown that with
Pittsburgh coal, coking would occur in an uncooled coal feed pipe.
A new water-cooled feed pipe was used in Test Series I and this
overcame the problem of coking. However it was apparent that the
new position of. the pipe led to bed ash being occasionally transported
up the cyclone dip-leg. Thus it was necessary to add -1680 ~m shale
daily in order to maintain the required bed heighto The results,
other than corrosion data, are summarised in Table 4.13.
4.5.3.2
Test Series 2 : 500 h test without limestone
The combustor operated continuously except for one short
interruption of a few seconds, which was the result of a power cut.
The position of the water cooled probe ~ad been changed, with the
result that there was an overflow of bed ash during the test.
However the combustion efficiency was only slightly better than in
Test Series 1, indicating that the fines recycle system may not
- 50 -
-------
Table 4,13,
Summary of results obtained on 12 in corros~on combustor (Task IV)
Coal
Additive
Pittsburgh
Limestone 18
. F1uidising velocity
Bed temperature
Bed depth
Top size of coal and additive
3 ftl s
15600F
2 ft
1680...m
5
Test Series 1 2 3 I
Duration of h 100 500 465 !
test, i
Mass balance period, h 0-100 0-115 283-383, 0-100 334-434
Oxygen in off-gas, % L9 2,3 201 201 2.4 I
Unburnt carbon 10ss, % 8.2 6,5 700 603 I 1,2 '
Off-gas temp, of. 1425 1475 1475 1410 I 1410 I
I CalS mol ratio I
! 0 0 0 2.4 109
in off-gas, ppm v/v 2270 2090 2050 I 52 ~
~ 502 200
! I.
i 502 reduction, % 0 0 0 90,5 I 9705 !
!
I I
i co in off-gas, ppm v/v 330 660 .900 600 600
i ~
! -
C1 in off-gas, ppm v/v 57 64 74 55 - (
,
, I
I Na in off-.gas, ppm wt, 604 207 4.7 4,5 3,7
,
i K in off-gas, ppm wt. 300 L2 1.6 007 0,9 !
t I \
I
~ Ca in off-gas, ppm wto 69 , 16 ! 21 98 135 ~
I . f ppm v/v 5 I 5 5 1 7 . r
' 503 ~n of -gas,
~ I
- 51 -
-------
have been functioning correctly. Two mass balances were carried
out during. the test and the results are summarisedin Table 4.13.
The performance of the plant remained constant during the 500 h test.
4.5.303
Test Series 3 : 500 h test with limestone addition
~
The combustor operated continuously,. except for four min,or
interruptions of several minutes duration during which the bed
. . 0
temperature dld not fall below 1400 F, until 35 h before the
planned end of the test. A failure of the coal feed electronic
control unit then caused ~he plant to be shut down~ .For three hours
before shut-down the combustion conditions varied between reducing and
t'\
oxidising.as attempts were made to rectify the fault. Two mass
balances of 100 h were carried out during the test and the results
are summarised in Table 4.13. These results suggest that, sometime
between the two balance periods, the fines recycle system began to
function correctly. Thus the unburnt carbon loss fell to 1.2% and
the S02 reduction increased to 97.5%. despite a lower Ca/S mol ratio.
4.6
The 6 in combustor (Task V)
The work carried out on the 6 in combustor is described in
detail in Appendix V.
4.6.1
Description of the plant
A flow diagram of the 6 in combustor plant is shown in Fig. 4.16
and details of the bed section of the combustor are shown in Fig. 4.17.
The combustor was of circular cross-section and was constructed throughout
of stainless steel.. The lower section was 6 in. in diameter and 6 ft
high; this expanded over approximately 1 ft into a1 ft 5 in diameter
section,4 ft in hei.ght. The whole combustor could be heated electrically
by external wall heaters. These were used for start-up and then to
maintaina.uniform temperature throughout the freeboard. Air supplied
from a plenum chamber, passed through a distributor plate made from a
drilled flat plate covered with three layers of i in diameter alumina
balls. The pre-mixed coal/additive feed was pneumatically conveyed to
the bed, which it entered tangentially, approximately i in above the
alumina balls. Excess heat was removed'by a water-cooled metal coil
immersed in the fluidised bed. The bed height was maintained constant
by emptying surplus ash through a tube in the centre of the distributor.
- 52 -
-------
Fines
Cdtchpot "
Fines Return Line
Combustor
Condenser
Fluidised Bed
Cool ing
Coil '/
Vibrator
t
Ash
Cooling
Water
Reservoir
Fig. 4. 16" Task ~. Flow d i a~ira m
-53-
of
6in." combustor
pi an t
-------
........
....-
....-
....-
....-
.........
,. -
u-,
(2:-4--
---'.-- ---:.. ~
'.J- - - _.J.J
ck~E-=- -=--
1..1- ..=- -=- .::. ~
- --
-------
The gases leaving the combustor could be directed through two
alternative cyclone systems, both comprising primary and secondary
cyclones, for operation with or without fines recycle. In the
'with recycle' system the primary cyclone was, vertically above the
bed and the fines were recycled via a dip-leg. Gas ~amples were
,
withdrawn at a point about 2 ft after each second~rycyc1one, as
appropriate, and bubbled through iodine or H202 solution for
determination of 802' Samples were also taken for analysis of
I
02' CO, C02 and CH4 by chromatograph.
Each test was carried out as a one day (l6hour) run comprising
plant start-up, approach to equilibrium, a six hoJr mass balance and
shut down.
4.6.2
Experiment.al programme
Tests were carried out using all of the four c9als, Illinois,
,
Pittsburgh, Park Hill and We1beck, and the three limestones. A total
of forty one-day tests was carried out and.these were grouped into five
test series as shown in Table 4.1. Except for two tests with fine
Limes tone 1359, all the coals and limestones, were', prepared to minus
10 BS mesh (1680 ~m).
Typical size distributions.of the.materia1s as
prepared are given in Table 4.14.
Table 4.14
ica1 .size distributions .6:f prepared coals and limestones (Task V)
I Percentage in size range (sizes ,in ~m)
Material Median
! +1680 +500 +250 +125 +63 ~m
I -1680 -500 -250 -125 -63
! !
~ .. 0.3 49.2 25.1 13.9 6.1 5.4 494
I Ill~nol.S coal
! Pittsburgh coal 0 36.6 23.5 17.0 12.8 10.1 340
! Welbeck coal 0 29.4 37.1 17.3 6.4 9.8 364
4.9 I 716 I
, Park Hill coal 0 63.5 16.7 8.9 6.0 !
~ Limestone 18 0.1 21.0 22.0 22.5 12,.2' 22.2 207 J
~
I Limes tone 1359 I 0.1 36.7 24.8 16.6 8.7 13.1 353 j
I
j U.K. Limestone j 0.5 52.1 19.1 10.1 7.2 11.0 537
j I I
"
- 55 -
-------
4,6,,3
Results
U,K, limestone with Welbeck, Park Bill, Illlo:is ond
~-_._-,-
P i.t t sbutgh coals
4,6,3. J
The 8ffect of the CalS mJ'] ratio on S02 redu.~: t,:n was determined
far all four c;-)als with UcK, limest':me at a fluid:Lslng 'iebcity of
, , 'J
3 £t/s and at 1470 F. The resul ts obtained are summarised 1:n Table l~ .15
and are plotted in Fig. 4.18. Possible reasons f0t the higher S02
redu::tlGn with ~.Jelbeck ccal than with the other three coals are
discussed in section 7.1.11.
Additional tests with Welbeck cGal,
ind uded in Table 4.15, showed an improvement in S02 reduc.tlon when
the f luidi.sing velocity was lowered to 2 ft/s. but n;:: effect .Jf
internal fines recycle or of. subsL ,lchiometdc combus t1.:::m condi tio'Os.
4.6.3.2
Limestone 1359 with Illinois coal
The effect c,f the CalS mel ratio c.o S02 reduction was determined
for Limestone 1359 with lIlln,elis coal in order to al1,)w comparison (i)
between Limest:;ne 18 and Limestone 1.359 and (d) between the CRE and
ANt 6 in combust,ns. The ,]perating (.:mditlC'ns and te,$IJlts obtained
are summarised in Table 4,16, and the 802 reductbns at 14700F are
plotted in Fl.g, 4019. The 802 reducti,]n was imp-rcyed by deepening
the bed from 2 tc 3 ft and by the use of finely crushed (-125 ~m)
')
limestone. The. S02 reducti~ns obtained at 1290~F were very low
(15% and 18% at CalS mol ratl'JS f::' 1.1 and 202 respectively) and
clearly showed that operation at this temperature is uosatLsfactory
wilh Limest)ne 1359,
Lime5t~ne 1359 was less effective than U.K.
limesrcne (Figo 7.16).
406.3.3
LimestQne 18 with PittsbuI&h and Welbeck coals
Limest:m(;' 18 was 1J8ed in five tests with Pittsburgh coal and in a
single test with Welbeck coal.
The operating conditions and results are
summari sed in Tab 1 e 4.17.
With Pittsburgh coal the Lime~tone 18 was more
effective tban U.Ko limestone (Fig. 7.15).
The single result with Welbeck
suggests .that this might not be the case with Welbeck coal, but a
meaningful e,::;mparison is not possible,
Compa.ratlve te.sts at a CalS mct ratio of 0.8 starting with (a) Ume-
free ash, (b) hme-free shale, (~) lime rich ash fr::;m a test at Ca/S of
206, confirmed thate.quilibrium c.onditions were being reached in the
short time al L:ywed for stabll isatico.,
- Sf) -
-------
100
10
Cool Symbol
Illinois 0
Welbeck ~
Pork.Hi II 'V
Pittsburgh 0
6
90
80
70
c:
0
.-
en
en
.- 60
E
Q)
N
0
V> 50
c:
c: tJ
o
.-
t 40
:J
U
Q)
c..
0 a
0- 30
Park Hill, Illinois
& Pittsburgh.
o
20
Fluidising velocity: 3ft/s
o
Bed temperature: 1470 F'
Bed depth: 2ft
o
1
2
Co IS mol ratio
3
Fig. 4. 1 e To sk Jr.
SO reduction in the Sin. combustor
2 .
for U.K. limestone with different coals
-57-
-------
100
m
90
80
/
/
o
70
A
60
[!J
o
~
o
.~ 50
~.
'1:1
ell
a::
N '0
o
(/)
A
/0
/
30
Bed Additive Symbol
depth. ft. size. J.lm
2 -1680 0
3 -1680 A
2 - 1 25 E1
20
10
o
o
1
2
Co/S Mol Rat io
3
FIg. 4.19 Task Y S 02 reduction in the Sin. combustor for
limestone 1359 and IllinoIs coal 3 ft/s 1470 of'
-58-
-------
Tab le 4.15.
Summary of results obtained on the 6 in. combustor with U.K. limestone and
We1beck, Park Hill, Illinois and Pittsburgh coals at a bed temperature
of 14700F and a bed height of 2 ft; (Task V)
V1
\0
Test No. 1.5 1.611.7 J 1.8 2.1* 2.2* 2.3 2.4\2.5 2.6 2.7t 2.8 2.9 2.10 2. 1112 . l~t 2. 13 2.14 2.15* .114.2/4.3 4.4 5.1
Coal type .I We1-
Illinois ~ We1beck Park Hi!' Pittsburg' beck
F1uidising velocity, ftls 3 3 3 3 2 2 2 3 3 3 3 2 3 3 3 3 3 3 3 3 3 3 3 3
Oxygen in off gas, % 2.6 2.2 2.3 2.6 3.1 3.1 3.1 3.1 2.3 2.2 0.3 2.6 2.0 2.5 2.0 2.3 2.1 2.5 2.0 ~.3 2.4 2.2 2.4 2.8
Unburnt carbon loss, % 4.6 3.4 3.7 4.5 2.6 2.6 6.9 7.9 7.1 5.3 13.4 5.1 0.9 9.4 0.4 0.1 0.3 9.2 8.7 ".3 9.5 0.7 8.7 14.0
Ca/S mol ratio 1.1 2.2 3.2 0 0 0.8 0.8 0.8 1.8 2.9 0.8 1.8 0.5 1.3 2.7 0 1.1 2.6 0.8 0 1.0 2.1 3.1 1.6
S02' ppm 1785 1000 229 3896 1087 488 513 599 207 33 612 158 1690 1045 237 2091 1210 322 547 980 1372 938 241 342
S retention, % 55 66 94 1 7 58 56 54 82 98 61 86 15 30 79 12 49 86 54 5 41 57 90 79
S02 reduction, % 55 67 94 0 0 55 52 45 81 97 48 85 - 25 78 0 42 84 50 0 31 52 88 71
Cl-, ppm 179 182 152 160 438 408 456 267 411 289 325 410 136 89 68 175 149 132 337 121 75 89 95 303
NH3' ppm. . - 34 25 17 26 11 10 6 ..., 51 - - 16 15 25 60 31 36 30 7 - - - - 12
.. ..
* Tests with primary fines recycle
t Test at substoichiometric conditions
-------
Table 4.16 Summary of results obtained on the 6 in combustor with Limestone 1359
Illinois coal (Task V)
Test Noo 1.1 1.2 1.3 1.4 3.1 3.2 3,3 3.4 3.5* 3.6*
Fluidising velocity 3 3 3 3 3 3 3 3 3 3
Bed temperature, of 1470 1470 1470 1470 .1290 1290 1470 1470 1470 1470
Bed height, ft 2 2 2 2 2 2 3 3 2 2
Oxygen in off gas, % 2.9 2.8 2,6 2.9 2.7 2.4 2.7 2.6 2.4 2.5
Unburnt carbon loss, % 4.2 3.8 3.3 3.3 7.3 6.5 5.0 4.9 4.1 3.6
Ca/S mol ratio 0 1.5 2.2 3.3 1.1 2.2 1.1 2.1 1.1 3.6
802' ppm 4023 2118 1450 680 3376 3245 1930 1136 1523 278
8 retention, % 0 47 64 78 24 26 49 70 61 92
802 reduction, % 0 47 63 78 15 18 51 72 61 93
Cl-, ppm 134 126 150 145 141 161 126 125 139 97
NH3' ppm 20 90 .9 23 34 23 12 11 21 9
* Tests with-125 ~m limestone.
- 60 -
-------
Table 4,17 Summary of results obtained on the 6 in combust~r with Limestone 18
and Pittsburgh and Welbeck coals at a temperature of 14700F
and a bed depth of 2 ft (Task V)
Test No. 4.11 4.5 . 4.6 J 4 8714 . ~k 4.9! 5.2
, Coal Type .( Pittsburgh-~ t Welbeck
Fluidising velocity, ft/s 3 3 3 3 j 3 3
Oxygen in off gas, % 2.3 2.5 202 2,3 2,1 20.5 206
Unburnt carbon loss, % 6.3 806 9.2 706 9,4 1101 1109
CalS mol ratio ° 0,9 107 206 0,9 0.9 109
502' ppm 1980 1137 581 185 1238 1115 236
S retention, % 5 50 75 92 .;5 60 8.5
802 reduction, % ° 43 71 91 38 ~3 80
C1-, ppm 121 83 69 65 69 50 363
NH3' ppm - ° - 21 22 24 3
* Test with lime rich bed.
t Test with shale bed.
- 61 -
-------
----.--
~
Table 4-18-
Typical Analyses of Alloys Tested
I
"
"
i
I I Normal* I
1 Material Nominal Composition % I
I ! I Oper ating !
AlSI ! C ' p S 8i ~empo
Type BSS Cr Nj Mn I Mo Ti Al Fe Other
NOe I .1 Max, ! Max, Max Max, F
-- J
- r - -
~ 0, 25 I
Medium Carbon Steel 3059/5 0,05 0,'05! Bal. 900 ;
! f '
,
, i
1% Cr, ~1. Mo Steel 3059/ LOO ~ 0 0 15 007 0-05 0,051 0,6 0,50 BaL 1050
i
J I
620 ' t
~ max I
~ I
\
2!% Cr, 1% Mo Steel 3059/ 2025 i 0-15 0.7 0,04 0,041 OoS LOO Bal. noo i
I. i
622 '~ maxo ! I
~ !
, I
i. I
12% Cr Steel EN S6A 410 12 0, is LO 0,04 0,03! 1,0 Bal. -
max, ! I
!
1 Austenitic 316 EN 58J 316 17 12 0.09 200 0,05 0,05 ; 0,2 2,S Bal. - I
I I
max, I
! I
Austeni.tic 347 EN 58G 347 18 11 0008 2.0 0,05 0,031 LO BaL Nb 10 1220
j I
max, I XC min I
! I
I 0,031
Esshete 1250 15 10 0,15 6,0 0,04 LO 1.0 BaL Nb 1% - I
I I
I I
Nimonic PE 16 17 43 1 0010 00021 0.3 300 L2 102 BaL - i
! i
! I i
-
*Ref, CEGB9 Modern Power Station Practice, Vo1,2
-------
4.7
Results of Corrosion Investigations
The corrosion of metal specimens at various temperatures was
investigated under Task II (48 in x 24 in pressurised combustor), Task III
(27 in combustor) and Task IV (12 in combustor), the latter combustor
being designed specifically for corrosion work.
analyses of the metals investigated.
Table 4.i8 gives
4.7.1
Appearance of Steels after Tests
When the tubes and coupons were removed from the combustors they
were examined visually. . In general the condition of the fire-side surfaces
appeared similar for all the tests.
Direct metallic erosion of specimen tubes by the ash particles had
not occurred. This observation was expected because the particle
velocities, associated' with fluidising velocities,of 2 ft/s to 8 ft/s,
would be quite low. It has been shown that in conventional coal-fired
boilers, flue gas velocities of around 100 ft/s a~e necessary before
steel is eroded by particle impaction3.
Visual examination of the test specimens .also suggested that the
scouring action by the ash particles was not sufficiently erosive to
remove the protective oxide cover and expose clean metal surface to the
environment~ This was confirmed by the specimen weight loss results.
These showed that the rate of corrosion. decreased with time, thus
indicating the formation of a stable protective oxide layer.
On the otner hand, the presence of the fluidised bed did have some
influence on the formation of deposits on the tubes.. The tubes
situated in the bed were mostly covered with a thin, even, hard and
adherent. deposit «0.1 rom thick).
A thicker (up to 1.5 rom) but more
loosely adherent ash-type deposit was sometimes formed on the underside
,
of the tubes cooled to 750-9000F which weresituated.close to a coal
feed nozzle.
1-:
The underside of the tubes. situated above the bed in the 12 in
combustor (Task IV) were covered with a layer of fine ash, and coarse
1
ash particles formed a layer on the top of these tubes. Both types of
layer were loosely adherent and were not found on the tubes situated in
the bed.
- 63 -
-------
Table 4.19.
Task II - Average weight loss of specimens
t
Wt 0 loss in jJ.g/cm2h I
Test I
Temp. 1% Cr 2!% Cr 12% Cr Type Type Esshete PE 16 ;
of 347 316 1250
660 24 2 4 j ,
35 1 - I - ;
Series 1 1020 - - 2 4 4 4 4
1
(We1beck + U ,Ko Dolomite) 1160 357 307 5 4 5 7 3
i
100 h duration 1350 - - 5 8 8 30 5
!
1460 - - - 17 16 50 8 !
!
Series 2 630 99 151 18 7 8 - 8
(Pittsburgh + Dolomite 1337) 1140 - - 16 28 23 48 14
81 h duration 1330 - - 18 48 54 210 36
1460 - - - 109 124 215 43
750 24 31 4 6 3 - 3
Series 3 and 4 1030 8 6 15 7 6
- -
(Pittsburgh + Dolomite 1337) 1220 11 12 40 60 11
- -
145 h duration 1450 18 14 90 40
- - -
Series 5 700 50 31 - - - - -
(Pittsburgh + Limestone 18 1040 - - 4 20 8 14 6
+ Dolomite 1337) 1200 - - 4 13 4 76 14
100 h duration 1450 45 20 60 45
- - -
Note:
Towards the end of Test Series 2 the bed partially def1uidised and some
sintering of the bed occurred. This probably caused the higher weight losses
for this test.
-------
Table 4.20.
Task III.
Average weight loss of specimens
[
Wt. loss in ;:.g/cm2h
I Test
Temp. M.CoS. 21% Cr 12% Cr Type Type Esshete
I of 347 316 1250
I 750 60 75 4 5 4
Series 1B 3
I (Pittsburgh) 930 400 - 7 6 4 8
1110 - - 3 - 3 7
500 h duration
I 1290 - - 4 9 10 16
750 8 4 1 <1 <1 1
I Series 2 930 36 38 3 1 1 1
(Pittsburgh + Limestone 18) 1110 - 270 2 ' 2 3 4
I 1000 h duration 1290 - 3 .5 9 12
-
Series 3 750 10 14 2 <.:1 1 1
I (Pittsburgh + Limestone 18) 930 - 30 2 <1 1 1
500 h duration
~ote: Test Series 1A was terminated by a failure of the cooling med~um after 147 hours.
resulted in the specimen tubes being grossly overheated for a short period and the weight
Nere not representative. They have: been omitted from the above Table.
This
losses
- 65 -
-------
Table 4,21, Task IV.
Average weight losses of specimens
Wtc loss in wg/cm2h I
Test Position Temp Type I Type Esshete PE 161
of M.C.S. 2i% Cr 12% Cr 347 I 316 1250
-_....
690 91 92 - - 1 2 -
940 163 179 - - 2 3 -
Series 1 In bed 1100 - 579 - 7 8 9 -
(Pittsburgh) 1320 - - - 11 11 21 -
100 h duration 1560 - - - 12 9 20 -
780 62 74 - - 1 2 -
In
1340 - - - 5 4 10 ,-
freeboard
1500 - - - 3 2 4 -
770 48 32 2 - -1 1 -
920 88 78 6 - 2 1 -
In bed 1090 - 539 5 3 3 4 -
Series 2 1310 - 3 3 4 6 3
-
(Pittsburgh) 1560 - - 308 7 5 7 3
500 h duration
760 25 25 1. - <1 <1 -
In .1310 1 1 1 5 1
- -
freeboard 1520 - 3 - 1 3 2
-
730 41 , 42 2 - 1 1 -
930 84 98 6 - 2 2 -
Series 3
In bed 1100 - 222 3 2 2 3 -
I (Pittsburgh
I
I + Limestone 18) 1320 - - 7 6 6 13 5
Ii 465 h duration 1560 475 3 2 3 2
- -
I
! 760 18 18 2 - <1 <1 -
i In 1310 - - 2 2 2 6 2
J
Ii freeboard 1490 - - 2 2 3 12 2
~
-------
The layers of material on the tubes were no thicker after 500
or 1000 hour operation. than they were after 100 hours, They were
insignificant compared with those occurring in conventional powdered
coal-fired boilers, and were predominantly ash.
There appeared to have been very little surfa~e attack of the alloys
tested at their usual operating temperatures. This includes the low
alloys, namely, medium carbon and 21% Cr steels, WhlCh were only slightly
blistered. These two alloys had somewhat thicker scales when tested above
4 ~ ~
their normal maximum operating temperatures of 900 F and 1100 F respectively.
Weight Loss Results
4.7.2
The weight loss results for all three Tasks are summarised in
Tables 4.19, 4.20 and 4.21.
The maximum.penetration rate on the fireside of conventional
boiler tubing t.hat can be tolerated is generally assumed to be about
1.5 x lO-6.in/h. This is equivalent to a specimen weight 10ss rate of
30 ~g/cm2 if there is no preferential loss around. the circumference of
the tube nor any intergranular penetration of the metal surface. This
figure has been used as a guide in assessing data.from the present
experiments.
4.7.3
MetallographicExamination
Specimen rings and coupons were sectioned, mount.ed and polished.
They were examined under a microscope and a few were also examined by
electron probe microanalysis. For normal operating conditions, materials
maintained at or below their accepted maximum. working temperature in con-
ventional plant did not suffer significant intergranular..penetration by
sulphur.
The surfaces of. the austenitic steels. from .the 100h tests were
mostly smooth* with the exception of the coupons in Task IV held at
8500C in the bed. The low chromium steels exhibited somewhat rougher
surfaces as might be expected from the higher specimen weight losses.
The surfaces of the specimens from the 500 hand 1000 h tests were on
the whole slightly rougher.
There appeared to be little difference
in the number or magnitude of surface irregularities between the tests.
*
defined as surface irregularities ~
4 ~m.
- 67 -
-------
The oxide layer for all steels was up to 10 Ilm thick with no
significant effect of test duration or metal temperature,
~
J";'.~
;~.
,,~
The austenitic steels showed little evidence of sulphide
penetration at temperatures below 11100F (maximum observed was 1OI-Im
in isolated instances) and for test durations of up to 1000 h. In
o 0
the range 1290 F to 1560 F sulphide penetration of 10 to 20 ~m was
general in 100 h tests and, exceptionally, up to 70 I..im in the
1000 h test.
100 h
For the low chrome steels, penetration was generally absent in
o
teets at temperatures below 930 Fo It was about 50 Ilm in 500 h
o
at temperatures of 1110 F.
tests
Specimens of 2!% Cr held at 11100F for 500 h in Test 2 of
Task IV were examined by electron probe microanalysis, confirming
sulphide penetra.tion of about 50 I..im.
- 68 -
-------
THE MATHEMATICAL. MODEL (Task VI)
5.
The mathematical model, and comparisons. with the results of a
number of test series and with results from work. carried out by Pope,
Evans and RobbinsS are described in detail in Appendix 6.
The construction. of. a mathematical.model to.represent the sulphur
retention
process was embarked upon with. the,. twin objectives of gaining
an unde.rstanding of the behaviour of the plants in the experimental
programme, and of developing a technique for use in, extrapolating the
results to other plants, fuels or limestones, for p~rposes of economic
assessments.
It is assumed - that the rate controlling. process is transport.
within the additive particles rather than transport of S02 to the
pa:tt.ic1e. surface. The rate of mass of 802 absorbed per unit time per
unit mass of additive has been related to the S02.concentration by
means of a rate.constant. Values of the rate. constant have been
determined in.laboratory tests (Section 6.2.9) .and related to particle
size and degree.of.sulphation (Fig. 6.2). In the mathematical model
these relationships. have been assumed to apply.to particles within the
fluidised bed combustor.
1.n general form, the model falls into two broad sections.
The
first section. calculates from the input data the.terminal.velocities of
particles in up to ten size groups, and thence.. the. bed composition in
terms of weights of ash and limestone of different particle sizes. The
mean residence time of each group of particles is also calculated. The
second section evaluates the absorption of sulphur-dioxide in and above
t.he bed by an iterative. procedure. - An initia1..value.is assumed for
the mean. reactivity of the bed and from it are.calculated the sulphur
retention and the. mean S02 concentration in the bed, Knowing the
residence time of each close size fraction and the relationship between
reactivity and degree of sulphation, the value of the degree of
sulphation is calculated for each size fraction,such that the calculated
total sulphur retention would be obtained with the ~alculated mean 502
concentrat.ion. An improved estimate of the mean bed reactivity is
calculated from these values of degree of sulphation, and hence improved
I.
-. 69 -
-------
values .of..sulphur. retention and mean S02.concentration are obtained.
The procedure is repeated until a satisfactory.degree.of convergence
is obtained, . the whole process taking. less than. a minute of computer
time.
.,.
~...'
The mathematicaL model predicts. that. the.. extent. Qf sulp~ation
of particles small enough to beelutriated from. the.. bed should be
very small compared with those partic1es::which stay in the bed. However,
the utilisation of the fine particles should be-.increased significantly
by efficient recycling to the bed.
~.
-tJ
The utilisation of particles only just small enough to be
elutriated from. the bed is. very sensitive. to.. the.. value. of the elutriation
rate constant,..whieh,.determines the residence time.of these particles in
the bed.
Whereas. the..c.omparisons.with experimentaL.te.ats.using dolomite
are on the whole good, it was obvious even before running the mathematical
model. that. there would be difficulty in getting agreement in the case of
the limestone tests.
Considering, for example, the tests in Task V using
Limestone 1359, .for which the laboratory reaction rate measurements showed
that,the.reactiv:ity.dreps.to.zero at utilisations ranging from about 10%
for.. the. coarse. particles to. about. 50% for. the. very. fine particles. A
mean value would be about 20%. In a typical. test on the six inch diameter
combustor,addition~of this limestone at a Cats mol. ratio of 1 gave a
retention.pJ..about.4S%: this was. in spiteofthe..fact.that some 50% of
the stone was of small enough size to be elutriated from. the bed. Even
if no limestone were elutriated, the retention could. not be explained on
the. basis ,I!)f . the kinetic data.
It was obvious therefore that an attempt would have to be made to
. allow.feDr the. exposure of fresh, unsulphated.stone.by abrasion of the
sulphated surface layer whose impermeability results in the limit to the
utilisation capacity of the stone. A crude computational procedure for
doing this has been introduced into the model by assuming that the
sulphated particles consist of complet~ly unsulphated cores surrounded
by concentric shells of completely sulphated material. The rate at
which particles are "regenerated" by attrition is calculated as the
.. 70
-------
weight of particles in the bed divided by the time needed to reduce
the particle diameter by an amount equal to double the shell thickness,
and then multiplied by the chances that a particle will stay in t.he bed
for a sufficiently long period of time for this to occur.
rhe mean utilisation.used in the absorption calculation for
deriving the reaction rate is then reduced by.an amount corresponding to
this rate.of "regeneration". The effect of this is that the actual
particle utilisation can exceed the limiting value found in the
laboratory experiments. It is not, of course"allE>wed to exceed 1.
Corresponding adjustments are made in the calculations of bed
.compositionto.allow for size reduction.of. the particles by abrasion.
Ideally, experimentally determined values of abrasion indices
for the various materials and rigs should be inserted into the
mathematical model, .but since there has not been time to derive them
independently, values giving satisfactory agreement have been
determined by trial and error.
6.
LABORAT.ORY . EXP'ERIMENTAL WORK
6.1
Coal studies (Task VII)
The work carried out to study the release of sulphur from coals
is described in detail in Appendix VII.
The experimental work carried out comprised the determination
of the various forms of sulphur in coals, cyclone fines from the 36 in
combus.tor - and~ carbonised.. coal (char) prepared. under contro lled condi tions
in the laboratory. The laboratory char was. pre~ared.by passing fine coal
particles. 53 to 75 ~m dia, suspended in argon carrier gas through a
furnace at 1470oF. The particles were fed and collec.ted. by water-cooled
probes and the. residence time in the furnace.was.lOO.ms... The results
obtained are given in Tables 6.1 and 6.20
The. ratio of.organic to pyritic sulphur is. considerably higher
in the combustor fines and laboratory char than in the coal. This
indicates that the rate of release of pyritic. sulphur is generally more
rapid than that of. organic. A comparison of the sulphur retention for
the laboratory.. char shows that the rates of release of both organic and
- 71 -
-------
pyritic sulphur for We1beck coal are higher than the corresponding
rates for Pittsburgh. .. It. is .not clear whether this is . also the case
for the combustor fines because of the large amount of inert material
and calcium sulphate in the fines.
For Pitt.s.burgh. coal the ratio of organic/pyritic sulphur in the
laboratory. char. is close to that of the cyclone:. fines, indicating that
the pattern of. sulphur release is similar. This is not the case for
Welbeck. coaL.. Although there were oxidising conditions in the combustor
and reducing conditions in the laboratory furnace, it is surprising that
there should be agreement for one coal and not for another.
Table"6.l
Forms of sulphur in coals
~
Sulphur % as received Ratio
Coar
Total Organic Pyritic I Sulphate Org./Pyr.
Illinois 3.95 2.45 1.35 0.15 1.8
Pittsburgh 2.75 1.44 1.25 0.06 1.1
Welbeck 1.25 0.73 0.50 0.02 1.4
Park Hill 2.40 1.01 1.35 0.04 0.7
Table 6.2
Forms of sulphur in 36 in combustor fines and laboratory char
I
i Coal Pittsburgh Welbeck
!
.
Designation 36 in fines Lab. char 36 in fines Lab. char
% sulphur as received
Total 2.4 2.4 1.3 1.1
Organic 0.3 2.24 0.15 1.06
Pyritic 0.03 0.17 0.06 0.06
Sulphate 2.1 - 1.1 - !
I
retention I
% sulphur
Total - 37 - 30
! Organic - 65 - 54
I
Pyritic - 5.4 - 3.6
Ratio: organic/pyritic 10 13 2.5 17.5
- 7? -
-------
6..2
Limestone and dolomite studies
6.2.1
Introduction
The reasons for the differences in the sulphur retention of
different stones are generally accepted to be ascribable to differences
" . 6.7
in physical. structure rather than in chemical compos~tlon .' Certainly
porosity formation on calcination has been established as. an important
factor (for example, potter6 has found for calcined stones a correlation
between pore volume, measured by mercury porosimetry, and their sulphur
oxide capacity), although it is not known definitely why. various types
of stone calcined under the same conditions should give products having
markedly different pore structures.
It. is clear therefore that basic studies aimed at obtaining
further understanding of the retention of.sulphur.oxides by stones
should include measurements of the pore structures of those stones. Such
an investigation was carried out in Task VIII and since in the present
work stones were fed into the fluidised combustors in the uncalcined state,
particular emphasis was placed on following the development of pore
structure at various..stages of calcination and determining how the pore
structure .LS modified during sulphation,
Quanti tative measurements of rates. of rea.ction of S02 wi th the
various s.tones.. ha.ve. also been made in order to provide data for use in
mathematical modelling. of the S02 absorption process (Section 5 and
Appendix 6).
6.2,,2 . Por.e struc..ture...on calcination..and sulphation
The main body.of the work concerned the use of the mercury
injection porosimetry and displacement density methods to follow the
development.of pore structure on rapidly heating.stones and then heat
soaking them for. various times in chosen atmospheres.
Mercury injection porosimetry is used for determining the
distribution of pore volume with respect to pore entrance diameter
(p.e.d.) 0 The..method depends on the principle that a certain minimum
pressure is required to force mercury through an entrance that is not
wetted by that liquid; the smaller the p.e,d. the greater is the
pressure required.. The cumulative volumes of mercury that enter the
- 73 -
-------
the sample as the pressure is increased in. steps are determined and
the pressures are converted into p.e.d. values by use of a computer
~.
program; the curve of cumulative pore volume ~ p.e.d. shows the
pore volume . lying behind entrances with p.e.d. .va1ues within any
chosen limits in the range of p.e.d. covered.. This curve may also be
expressed in the first differential form,. or as.. a histogram. In the
present work, wi th.an assumption that the pore entrances were 'circu1ar,
the range of..p..e.d.. covered was 59 ~m to 0.014 pm. All the primary
cumulative. pore.., volume .~. p..e.d. data are colle.cted. at the end of
~-..~
.
Appendix 8.
~
The displacement density of a solid, is the mass, of the solid
divided by the. volume of fluid it displaces on immersion. The further
a particular.fluid penetrates into the solid the lower .wil1 be the
displacement. volume and. the higher the,correspanding.density.
Mercury
under 1 atm. pressure and helium were used as fluids.
The mercury was
assumed to penetrate no pores having p.e.d. less than 14.5 ~m, and
the 'mercury density' was regarded as an approximation to the super-
fida1 particle,,:,density of a stone sample. Helium will enter pores
with p. e. d. down to 0.26 nm (0. 00026 ~m); the,' helium dens i ty'
. approximates to. the. density .of the solidmatter..af,.stone samples.
The redprocals..of the mercury and helium densities give the volume
(Vx) of pores with p.e.d. in the range 14.5 ~m to 0.26 nm. Porosity
values were calculated. on a unit weight basis and as percentages of the
'particle' volume (viz, reciprocal of the mercury density).
.Exploratory..experiments were.made.in.which.U...K.,dolomite and
Limastone 1359 particles were dropped into a futnace at l4700F to achieve
rapid heating and left in the furnace for specified times before removing
them for weighing and pore structure examination; the furnace atmosphere
was such that. either it initially contained.air",no ,at,tempt at flushing
being made, or it was flushed with nitrogen. A further series of
a
experiments on similar. lines was carried out with the furnace,at 1470 F
or 16500F, being flushed with a specified gas mixture - composition by
volume, 15% C02' 3% 02' 82% N2,with and without 2000 ppm of 802 being
present: the residence times of samples ,in the furnace were i, 1 and 4 h;
- 74 -
-------
st.ones examined were Dolomite 1337, U.K. dolomite, and Limestones
18 and 1359; particle size was -1000 +300
iJnl.
The experiments with
502 were intended to simulate the behaviour of stone particles in the
combustor. The results have enabled conclusions to be drawn as follows:
(i)
The development of porosit.y":' on calcination is a time and
temperature dependent: process; some examplesoLamounts of porosity
formed aresbown in Table 6.3.
Table 6.3
Examples.,oLaIIWup.ts of porosity (Vx, % v/v) formed in stones on
ca1cination*, and the extents of calcination** (in brackets)
Temperature and time of calcination
Stone Raw l4700F l6500F
Stone
. 1/4h 4h 1/4h 4h
,
Dolomite 1337 8 24 (34) 33 (52) 52 (85) 54 ( 90)
U.K. dolomite 5 33 (57) 41 (69) 54 (96) 56 (100)
Limestone 18 7 14 (15) 17 (30) 45 (85) 48 ( 96)
Limestone 1359 2 5 (4) 14 (19) 44 (86) 46 ( 91)
Using the gas mixture 15% C02' 3% 02, 82% N2' zero S02'
** Weight loss expressed as a percentage of weight loss for
*
complete calcination.
The amount of porosity (% v/v) formed is for each stone a
linear function of C02 loss.
(ii) On calcination, porosity is generated initially much more rapidly
in dolomite than in limestone (see Table 6.3). This difference is
most marked. in a C02 bearing atmosphere at 14700F. The effect is
evidently associated with the decomposition of the MgC03 in dolomite.
(iii) There is little, if any, porosity with p.e.d. in the range
0.014 ~mto 0.26nm, with the calcined-only materials or, with the
possible exception of Limestone 1359, with the calcined-sulphated
ones.
t
The term porosity. herein means Vx' unless ~therwise stated.
- 75 -
-------
~~
45
I
. ~
40
35
~
o
~
~25
Co
::J
V>
20
o
Fig. 6. 1
VI DOlomite 1337
A U.K.Dolomite
B Li mestone 18
o Limestone 1359
.
VI .1650 F
. 1470.F
A 1470.F
.
o 1650 F
o 1470.F
1/4
1
4
Time: h
Variation of level of sulphation with time and
temperature
-76-
-------
-------------- --
--------
(iv) Sulphation* varies with time, as shown in Fig. 6.1. It is
emphasised that the stones were not precalcined before being exposed
to 502
(v) Sulphation caused a reduction in porosity.
It was found that
sulphur is accommodated in the smallest pores (as detected by the
mercury porosimetry) in the stone, e.g. with Dolomite 1337 at 16500F
most of the porosity filled, blocked off, otherwise eliminated on
sulphation lay in pores with p.e.d. between 0.1 ~m and 0.03 wm; with
Limestone 18., part of the sulphation took place in pores with p.e.d.
between 0.3 ~m to 0.1 ~m.
(vi) As the results in Fig. 6.1 clearly show, Limestone 1359 is poor
compared with the other stones in its ability to absorb 502' The
studies of calcination showed that, of the stones, Limestone 1359
is the poorest.in quantity of relatively large.pores (say, with p.e.d.
in the range 14.5 ~mto 0.3 ~m) whose presence would facilitate the
more rapid ingress of.SOZ to the interior of the particle and lead to
more rapid. reaction. Contrawise, Limestone 18 shows the most
pronounced development.of pores in this range ofp.e.d. On sulphation
the volume of.pores with p.e.d. > 0.3 ~m shows little change with the
two dolomites,.viz.,. the 'transport pores' retain their identity, but
with Limestone 18 progressive reduction in. the volume of these pores
occurs as the level. of sulphation increases.
As comparison of the
resu1.ts.in.Table..6.3 with those in Fig. 6.1 shows, the level of sulphation
reached under the present conditions does. not depend solely on the amount
of.porositydeveloped.in the stone on calcination' alone. A fairly high
porosity is developed in Limestone 1359 after 4.h.at.1650oF but the level
of sulphation remains relatively low (evidently owing to the lack of
transport pores, as mentioned above) and indeed the porosity in the stone
a
after 4 h at 1650 F in the SOZ bearing atmosphere is as much as 42%.
. Comparison of the latter with the mercury porosimetry data suggests that
the sulphation may have caused narrowing of some pore entrances to <0.014
urn diameter (cf., 6.Z.3 below).
*
Sulphation is defined here as the ratio of the quantity of sulphur
taken up by the sample to that taken up if all the CaO were
converted to CaS04'
- n -
-------
(vi i)
It appears that at least two criteria of pore structure have
to be met if a stone is to be a good S02 absorber:
(a) extensive volume as large, viz, transport, pores for
ease of access to the interior of the stone;
(b)
extensive volume, for accommodation of the sulphur, a~
relatively small pores - which have a higher surface/
volume ratio than larger pores.
In the case where one factor is missing, as in Limestone 1359 with
little porosity in large pores, or, in Potter's work~ with Iceland Spar
containing low porosity, then the sulphur uptake is liable to be poor.
(viii) The mechanism of sulphur fixation is not clear but some points
arise from the present work:
(a) Since sulphur is found in the smallest pores the 802 probably
does not react to sulphate at the point of first contact with the
stone, in which case the S02 must migrate in some form until a
certain favourable site or condition is obtained.
. 0 0
(b) Increasing the temperature from 1470 to 1650 F does not
cause the increase in rate of su1phation with the dolomites that
might be expected from chemical kinetics considerations.
This
effect may be related to the observation in the pilot plant work
that the efficiency of removal of S02 passes through a maximum
at about 1560oF.
(c) With dolomites, at least, a highly speculative suggestion,
based on the observation made during the course of the work (see
Section 6.2,4 below) that the last traces of water were difficult
to remove, is that an oxygen-containing species, perhaps hydroxide,
localised on the internal surface of the stone is necessary in the
formation of. the. sulphate; increase of temperature above a certain
level could cause such a surface species to become mobile resulting
in reduced formation of sulphate.
.The chemistry of the sulphur retention process, and the
possible influence of water on this and on the formation of the
pore structure itself, would seem to be worthy of further study,
particularly as such work may lead to an understanding of the
mechanism, and hence perhaps elimination of, the fall-off in
sulphur retention above about l5600F referred to above.
- 7R -
-------
6.2.3 'Impermeable 1ayer'formation on su1phation
It has been possible to demonstrate the presence of an 'impermeable'
layer or 'shell' within the particles of a sample of Limestone 1359,
sulphated in the laboratory using the method briefly described in 6.2.9 .'
This was not found within Limestone 1359, sulphated Ln a combustor, nor
within sulphated samples of Dolomite 1337 and the U.K. dolomite.
(i)
(ii)
The method used was as follows:
Mercury porosim~try curves of the sulphated stones were measured.
Further samples of the sulphated stones were taken and the
particles broken individually using a razor blade; the product
was then sieved to remove particles smaller than 300 ~m and the
porosimetry curves of the +300 ~m particles were measured.
The porosimetry curves for the Limestone 1359 sulphated in the laboratory
show that there is considerably porosity within the whole .sulphated
particles not reached by mercury in the porosimeter but that some or all
of this porosity is made accessible on simple cleavage of the particles.
No such increase of accessible porosity was seen with a Limestone 1359
sample that was calcined but not sulphated, nor with the plant-sulphated
stone or the sulphated dolomites. Evidently. in ,the 1aboratory-sulphation
of Limestone 1359 some pore entrances have been narrowed to <0.014 ~m
diameter; the presence of such narrowed entrances, which may be regarded
as a form of 'shell', will aggravate the poor su1phation behaviour of the
stone. Such 'shell' . formation is presumably associated with the less
open pore structure in this stone in comparison with the dolomites which
do. not show the effect. '. It is possible that with .theHplant-sulphated
Limestone 1359, a 'shell' was destroyed, or was prevented from forming, by
attrition in the f1uidised bed.
6.2.4
Porosity at elevated temperature
It is known that certain materials such as glasses and synthetic
graphites contain pores with entrances through which a particular gas
molecule will pass at one temperature but not at another, viz,
accessibility of pores to gases may change with temperature.
with increase of temperature, accessibility would increase
Usually,
- 79 -
-------
(' closed-pore penetration'), the extent of the effect being related
to the molecular size of. the gas, but a decrease has been observed
('open-pore closure').
As .pore.struc.ture measurements, in general, and on stones in
particular,havein the past been made at room temperature or lower,
it was considered necessary in the present work to investigate the
extent of any change in the accessibility to relevant gases of pores
in stones when the temperature is increased to combustor temperature.
The methods used. to detect either type of accessibility change
depend on the principle that gas that penetrates a pore at a particular
temperature becomes trapped in the pore when the temperature is changed
to that at which the pore is no longer accessible to the gas. After
removal by evacuation of the free gas, the trapped gas may be released
for measurement by bringing the temperature of the solid back to the
original temperature of penetration.
The U.K. dolomite and Limestone 1359 were examined in the
temp~rature range, room temperature to 1470oF. The principle
limitation of the work is that all the samples examined, which included
some from the exploratory experiments mentioned in 6.2.2, became
perforce fully calcined in vacuum before the experiments were started.
Four inert probe-gases were used, of increasing molecular s~ze:
helium,.neon, krypton and xenon, the last being similar to S02 in
molecular. diameter. Gas release was measured using.a mass spectrometer.
With both stones there was no open-pore closure effect
wi th any
of the gases.norany.c1osed pore penetration by helium or neon. Large
apparent closed pore penetration effects by xenon were observed with
some samples of both stones, but with the dolomite and probably the
limestone, the effects were largely due to residual adsorption at room
temperature of xenon in open fine pores, this gas not being pumped
away with the free gas and subsequently appearing on increasing the
temperature along.with any released from previously inaccessible pores.
But the total times taken for the xenon to be released at l4700F in
vacuum were between 10h and 100h and it is expected that any
- 80 -
-------
contribution of.xenon due to closed-pore penetration would have
been released more slowly than that merely residually adsorbed:
this is taken.to mean that the penetration of any closed pores is
a slow process compared with the rate of sulphur uptake by the
stone and is likely to be of only secondary importance, if at all,
in the combustor. Apparent closed-pore penetration effects were
also observed. with krypton, but these were much smaller than, and
analogous to, those with xenon.
There were.some qualitative differences in quantity of
residually adsorbed xenon released between the -1000 +300 ~m dolomite
and limestone samples, presumably reflecting corresponding differences
in open fine-pore structure. The results are difficult to interpret,
however, and are influenced by calcination history which varied among
the samples;attention.has been called to the differences because of
the other differenc~s.that have been observed.between the stones, for
example, in 802 sorption capability.
Finally, the. following incidental
during the course of the experiments.
observations were made
(a) Traces of 'water' were retained by the U.K. dolomite samples,
but not by the limestone ones, for extensive times at 14700F under
reduced pressure. This water is presumed to have. been combined as
Mg(OH)2' the last vestiges of which were difficult to decompose.
(b) 802 was. released from the U.K. dolomite in vacuum on
increasing the temperature from 1470° to 16500F;the quantity of
802 originally present. in the material was estimated from this to be
<;15 ppm (w/w). Although this 1S a very small. quantity, the observation
may have some_significance, possibly in the problem of the reduction in
802 uptake on increasing the temperature in the combustor to above
15600F.
6.2.5
Effect of thermal cycling on 80, uptake
Elutriated stone fines are subjected, if recycled, to rapid
temperature changes and it has been suggested that such 'thermal
cycling' might have an effect on the capability of the recycled fines
for absorbing 802' Experiments were designed to test this possibility.
- 81 -
-------
The sulphur retentions of thermal cycled and non-cycled samples
were compared" The two samples were sulphated simultaneously with
both samples i.n the same furnace:0ne sample was taken through a
thermal cycle experiment whilst the other was maintained in the
hot region at constant temperature,
In this way it was intended
that only the effect of the thermal cycling should be apparent and
possible differences that might have arisen between separate
experiment.s, for example, due to variability in gn flow, avoided"
-1000 +300 I-im and - 53 ~m samples of the U.L dolomite and
Limestone 1359 were used,
1500 to 10000 ppm; C02' 0
o
temperature was 1470 F.
with four gas-flow compositions: S02'
to 75%; 02' 2 to 10%, all v/v" The furnace
The cyclic temperature changes were not seen
to have any systematic effect on the sulphur retention of the limestone
or the dolomite.
Incidental observations made in the course of these experiments
were:
(a) Change of particle size of the stone from -1000 +300 urn to
-53 ,Jm resulted in an increase in sulphur uptake for all of the
atmospheres used;
the effect was larger with the limestone"
(b) Increase of the C02 concentration to 75% caused. a substantial
decrease in the S02 uptake with both stones and both particle sizes,
With -1000 +300 ~m material the effect was much more marked with the
limestone"
(c) Overall, under the conditions of the experiments, in the
-1000 +300 ~m size grade, the dolomite is the better S02-absorber.
As expected, reducing the particle size to -53 urn appears to largely
,
eliminate the difference in S02 uptake between the two stones,
particularly at the C02 concent.ration of 75%.
6.2.6
Deposition on U.K. dolomite
A brief optical microscopical examination was carried out on
U,K. dolomite particles from the bed of the pressurised combustor, after
completion of Test Series I. The major portion of ash was first removed
with a magnet and the -300 urn dolomite particles were examined.
Generally,
they were covered with a shell of fused brown deposit;
were chipped or broken and in others the
some of them
- 82 -
-------
shell apparently had become detached from the particle. The shell
is presumably a slag formed with ash in the combustor: no such shell
was observed ..on. the samples prepared in. the laboratory.. No comment
can be made on the role of the shell in the sulphation process.
6.2.7
Surface area
A few measurements of surface area, using the BET method, were
made in the course of the work.
6.2.8
A simple test for stones
The most simple and satisfactory laboratory-method of placing
unknown stones in order of acceptor ability would seem to be a
sulphation experiment wherein stone particles are dropped into a
tube at reactor temperature through which an appropriate 502 bearing
gas stream. flows.(see.6.2.2). Subsequent analysis for Ca, C02 and
S02 content would enable curves to be constructed showing the
variation of the.level of sulphation with time (cf. Fig. 6.1).
6.2.9
Reaction rate measurements
Measurements were made on Dolomite. 1337, the U.K. dolomite and
Limestones 18 and 1359 to obtain data on reaction rates specifically
for use in mathematical modelling of the 802 absorption process.
The reaction. rate constant was derived as a function of sulphation,
particle size and temperature.
The .materials.. used were taken from the uncrushed material
supplied for' the plant experiments. These were crushed in the
laboratory and.size.graded for the reactivity-experiments.
Mixtures of OZ,.NZ' C02 and SOZ in various. proportions (constant
for each experiment) were passed through a thin bed of the stone, and
. the amount of 802 absorbed by the stone was determined after the lapse
of successive. intervals. of time. The amount of S02 absorbed was
determined' from the absorption of the residual SOZ' in the gas stream
after passage. through the bed, in hydrogen peroxide solution and
. titration with standard alkali. The experiments were repeated at
different. tempera.tures and using different particle sizes.
- 83 -
-------
0.12
0'10
tf)
en
.::£
M 008
E
+J
C
o
+J
tf)
C
00'06
u
: d>
1+.1
C
L
C
o
:;:; 0,04
I u
I c
(l)
a:::
0'02
o
1
2
3b
1 DOlomite 1337, 177~m
2 U.K. Dolomite, 177}lm
3 limestone 18, 177~m
3a" ",1414~m
3b" " , " 75%. CO 2
4 Li mestone 1359
1
30
2
10
20 30
Sulphation: 010
40
Fig. 6, 2 Comparison of reaction rate consta nts of different ston~~
c 81.-
-------
The. choiceof.a static bed for the experiments was governed
by the need to ensure that the 80Z concentration change through the
bed was as small as. possible, in order to..simplify interpretation of
the results.. This necessitates use of a thin bed, and hence
rules out the possibility of using a f1uidised bed.
Argonne National Laboratory have found in previous work8 that
over the range of gas velocities likely to be encountered in fluidised
combustion beds,. the reaction rate between stone and 802 is substantially
independent of gas velocity. In the present work, velocities were kept
as close as possible to those used in the plants, thus reducing as far
as possible any errors due to dependency of reaction rate upon mass
transfer of. 802'. to the particle surface.
Reaction rate constants are expressed in units of m3/kgs. This
is a rate of mass of 80Z absorbed per unit mass.of.stoneper unit of 80Z
concentration expressed-in mass per unit volume, i.e., kg 80Z absorbed
per kg stone per second per (kg 80Z ~m3 gas mixture).
The results showed that the reaction rate constants.of the four
stones'at zero su1phation did not Giffer widely either with stone type,
with particle size,cwith COZ concentration, or with. temperature, generally
lying in the range 0..05 to 0.1 m3/kgs, the limestones lying near the
bottom and. the dolomites. near the top of this...range.... At. su1phations
greater than 10%, however, there is considerable variation in rate
between different stones, between different particle sizes, and with
temperature. and C02 concentration.
Fig. 6.Z.i1lustrates a few of the results...Curves of rate
constant ~sulpha~ion are shown for 177 ~m particles of the four
stones, in OZ/NZ/SOZ mixtures at l470oF, together with curves for
larger particles (1414 ~m) of Limestone 18 in air/SOz mixtures and
02/NZ/C02/802 mixtures at partial pressures of C02 in the neighbourhood
of those obtaining in combustion at 5 atm pressure.
The results show the reactivity of the various stones to be
consistent with their behaviour in the 'drop-tube' experiments referred
to earlier (6.2.2 and 6.2.8) and show that the latter method is suitable
for a simple assessment of the efficiency 'of a stone in retaining sulphur
in the bed.
- 85 -
-------
7.
DISCUSSION
7.1
Emission of 802
7.1.1
Introduction
The main finding of all the experimental work described in this report is
that it is technically feasible to reduce the 802 emission from coal-.fired
fluid-bed combustors to the levels required by existing ordinances or likely future
legislation by injecting into the bed sufficient lime in the form of limestone or
dolomite.
The effect on S02 emission of changing operating conditions is discussed
under the following headings:
(a) Calcium to sulphur ratio
(b) Comb us tion temperature
(c) F1uidising velocity
(d) Bed height
(e) Operating pressure
(f) Fines recycle
(g) Particle size
(h) Additive type
(i) Coal type
The process data were obtained from five combustors.
ConsequentlYt before
commenting on the effects of the variables, the extent to which the data may be
considered to be homogeneous is examined. The time taken to reach steady operating
conditions and the mathematical model are discussed at the end of the section.
The accuracy of the data obtained is also discussed later (section 7.1.13).
The reproducibility of the S02 emission in replicate tests was found to be ~ 29%
of the value. For a coal with 3% sulphur (Pittsburgh) this leads to
reproducibilities of percentage S02 reduction of~ for example t 15% at 50% S02
reduction and t 5% at 85% reduction. Differences between single results within
these limits may not be significant, but smaller differences between sets of results,
e.g. curves through a number of points, are meaningful.
7.1.2
Effect of plant design
The main differences between the five pilot-plants can be summarised as
fo llows :
(i)
Bed cross-section:
2 2
This ranged from about 0.2 ft to 8 ft .
Two were
circular, two rectangular and one was square.
- 86 -
-------
(iO Cooling surfaces in the bed: These ranged from closely packed banks
of 1 in. cutside diameter tubes extending over almost the whole height of the bed
(48 in x 24 in pressurised combustor), to widely spaced 2.37 in, outside diameter
tubes extending over only a part of the bed height in the 36 in combustor.
In the pressure combustor there were several rows of tubes at and innnediately
above the bed sUT.'faces to minimise splashing and carryove'r'.
(iii) Cooling of the freeboard: Excessively high temperature in the freeboard
was prevented by direct steam-injection in the 36 in combustor, by cooling coils
in the pressure combustor and by water cooled walls in the 27 in combustor.
In the 6 in combustor the freeboard was heated to avoid the temperature dropping
below the bed temperature.
(iv) Feeding of additive: Additive was fed through a single nozzle
separately from the coal feed in the 36 in combustor but in the others it was
premixed upstream of the coal feed point.
(v) Feeding of coal:' The area of bed served by each coal feed nozzle in
the large combustors ranged from 2 ft2 for the pressure combustor to 4 - 4~ ft2
for the 27 in and 36 in combustors.
(vi)
Operating pressure and velocity:
The 48 in x 24 in pressurised
combustor~ the 12 in combustor, and the u in combustor, were only equipped to
operate at 10w velocities' (3 itjs or lower).
Observations on the effect of
higher veloc.i ties d~pend upon the resul ts from the 27 in and 36 in comb us tors.
Operation under pressure was of course restricted to the 48 in x 24 in
c.ambustoro
(vii)
Recycle of carryover material:
The 36 in combustor was equipped with
an external fines recycle system. The 27 in combustor had both internal and
external systems: the internal system was in use at all times and reference
in the remainder of the report to operation without recycle refers only to the
external system.
recycle systems.
The other combustors had only internal and hence less positive
The performances of the combustors under closely similar operating
condi dons have been compared in Figs. 7.1, 7.2 and 7.3 by plotting the
percentage reduction in 802 against the Ca/8 mol ratio. The main comparison
was between the 36 in and the 6 in combus~ors? and Fig. 7.1 shows that their
performances are identical. The performance of the 27 in combustor (Fig. 7.2)
is slightly better than that of the 36 in combustor, probably because of the
internal fines recycle in the 27 in rig. Comparison between the performances
- 87 -
-------
100
80
c
o
,-
II)
II)
E
(1) 60
N
o
l/)
c
5 40
+J
U
~
~
~ 20
o
0'
/X-
/
o
;:,
1 2
CalS mol ratio
3
4
Pittsburgh coo I .
Limestone 18 (-1680~m)
Bed depth 2ft
Fluidising velocity 3ft/s
o
Bed temperature 1470 F
No recycle
o 6 in combustor
x 36;n combustor
Fig,71. Comparison of. 502 reduction in
6in and 36 in Combu stors
-88-
-------
80
c
o
~60
E
x/
/0
x/
/ 0
Pittsburgh coal
Limestone 18(-3175~m)
Bed depth 2 ft
Fluidising velocity 8ft Is
o
Bed temperature 1560 F
o 36in combustor(no recycle)
x 27in combu stor (with
internal recycle only)
c
o
+'
lJ
:J
1J 20
Q)
L..
o
If)
c 40
o
o
0--
o
1 2
Co IS mol rat io
3
Fig.'" 2. Comparison of 502 reductio n in 27in
an d 36in Combu stars
c100
o
.-
(/)
(/)
.-
o x
yxx
/xx
x
/
Pittsburgh coo I
Dolomite 1337 (-1587 \lm)
Bed depth 3'8 ft
Fluidi sing veloc ity 2ft/s
o
Bed tempera tu re 1470 F
With recycle
o 36 in combustor
x 48in x 24in combustor
E
~80
o
If)
c
c
o
~ 60
u
:J
1J
~ 0
0--
1 2
ColS mol rat io
, 3
Fig.?,3. Comparison of 502 reduction in 36in
and 48in X 24in Combustors
-89-
-------
100
80
..!
o
c
o
:;; 60
u
::::J
'0
V
L..
N 40
o
II)
20
o
o
123
CalS mol ratio
Welbeck coal
U.K. Limestone
Bed depth 2ft.
Flui dising velocity 3 ft.ls
Bed temperature 14700 F
No recycle
ANL data A
CRE data 0
4
Fig 7.4
Comparison of performance of the 6 in. combustors
at CRE and ANl
100
80 ' 0 Illinois coal
/~A Limestone 1359
..! Bed depth 2ft.
0 Fluidising velocity 3 ft/s
c:
0 60 / Bed temperature 14700 F
....
c: No recycle
G.I
- ANL data A
G.I
L.. 40 CRE data 0
i... I!i.
::I /
.s::
Q.
::I
II) 20
o
o
2
3
1
Co IS mol ratio
4
Comparison of performance of the 6in. combustors
at C RE and ANl
-90-
Fig 7.5
-------
of the 36 in and 48 in x 24 in combustors with :recy'.:~e. Figo 7.3~ shows the S02
'reductic,n in the former to be slightly mere effident~ but the difference could
be aC0uotE-d fo'(' by the more effective r'ecyc.1ing system of the 36 in combustor.
At the cutset of the pr'ogramme it was thought that the differences
betwe€n the sulphur retention performances of the 6 in cO'.rnbustors at the CRE and
the Argonne National Laborato'des might be due to differences in design CT'ab le 7.1).
When burning Welbeck coal (1.6% sulphur do a. L) the sulphur retention at the ANL
was app'.recl:;ibly lower than that obtained at the CRE (Fig. 7.4) 0 It was found,
however, that sulphur retentions obtained in the two combustors when burning high
sulphur Illinois coal we're in dose agreement (Fig. 7.5), and it was concluded
that the design differences were unlikely to cause differences in sulphur retention.
Although it: .:.an be concluded that design differences between the pilot
scale combustors did not affect S02 reduction" direct application of the present
results to full scale requires some caution. Departure from certain features of
the rigs in scaling~>up could lead to significant differences in performance. For
example~ incr-easing the aI'ea served by a coal feed point might result in inferior
performance, as suggested by the twofold differences in 802 concentration found
ov'er the C['OSS seedon of the 36 in combustor. The effects of other design
factO'J:'s are discussed latero
Table 7,,1
Design di fferences between CRE and ANL 6 in Comb us tors
CRE
ANL
Dia. of upper part c£ freeboard
Cocling
Air di.stributor
Insulat.ion of freeboard
Coal and additi.ve feeds
Air preheat temperatu're
Met.hod of 502 Determination
17 in
Internal water coil
Drilled plate
Yes
Premixed
Not preheated
Wet methods (H202
and iodine)
6 in
External
Bubb le cap
No
Separate
8500F
Infra red
analyseT
The effect of Ca/S ratio
7.1.3
As would be expected~ and as predicted by the mathematical model, the
percentage reduc.tion in S02 emission increases with increase in the Ca/S mol ratio.
Sections 4. 1 ~ 4.2." 4.3 and 4.5 ccn tain several plots showing the variations of
percentage reduction with Ca/S mol !.atio~ and many of the curves show that the
502 emission can be reduced to
significant operating variable
would have, been useful to have
very low levels. The Ca/S ratio is the most
determining the reduction in S02 emission, and it
a correlation :)f the percentage reduction with
- 91 -
-------
Ca/S ratio. The mathematical model gives little guidance on the form of a
possible correlation, since it does not compute the percentage reduction
analytically but by an iterative method.
Examination of the curves of percentage reduction vs Ca/S ratio suggests
that there is an approximately exponential relationship of the form
R = 100 (1 - exp(- MC» where R is percentage S02 reduction, C is ,the CaJS mol
ratio and M is a constant depending on coal, additive and operating conditions.
Detailed analysis of the results suggests that some of the curves do not fit the
exponential relationship well, possibly because of anomalies introduced by effects
of particle residence time, attrition, etc., or, as in the mathematical' model, by
the S02 release profile in the bed. The exponential relationship has therefore
been used only for small extrapolations of data to permit comparisons to be made
at a common Ca/S ratio.
An indication of the quantities of additive required to meet the target
levels of S02 emission expected for (a) built-up areas (300 ppm) and (b) some
densely populated areas in the U.S. (100 ppm) is given in Table 7.2. The
quantities are for a combustor operated under the optimum conditions of
temperature, with a low fluidising velocity and with fines recycle. The
stoichiometric requirements for calcium carbonate are 3.15 lb/lb of sulphur, and
for dolomite 5.75 lb/lb of sulphur.
- 92 -
-------
Tab 1 e 7. 2
Additive requirements for a 100 MW power plant boiler
(900 short tons/day of 3% Sulphur Coal)
Parget level of emission ppm v/v
Quantity required 300 100
1. Ca/S ratio 1.1 1.8
2. Weight of additive/lb S
removed
(i) Limestone lb 4.1 5.9
(ii) Do 10m te lb 7.6 10.8
3. Quantity of additive required
(i) Limestone ton/day 110 160
(ii) Do"1omi te ton / day 180 280
4. Quantity of sulphated additive
to be disposed of
(i) Limestone ton/day 110 150
(ii) Dolomite ton/day 150 210
The Effect,of Ccrmbustion Temperature
It was found in laboratory studies of additive reactivity that the rate
of reaction of additive with 802 increased with temperature up to about 16500F,
but t.he vari ation was not in accordance wi th the manner expected from the
7.1.4
Arrhenius relationship.
The results of pilot-plant studies of the effect of
combustion temperature on the reduction in 802 emission are shown in Figs. 7.6
and 7.7. Data interpolated from these curves together with those from other
comparisons, are summarised in Table 7.3. They show that for both limestone and
dolomite there appears to be an optimum temperature for removal of S02 in the
range 1400 to l600oF. The shape and temperature of the peak is not consistent.
This may be the consequence of experimental errors associated with the few data
at widely spaced temperatures or may be a real effect resulting from the
influence of other operating variables. For example, it is known from laboratory
studies of Limestone 18 that the reaction rate constant varies with temperature in
a manner that depends on the fraction of lime utilised (Fig. 7.28).
With fresh
lime (less than 20% converted) the reaction rate constant continues to rise with
increase in temperature up to at leas t 16500F, but as the fraction of lime
reacted increases, a peak develops in the curve, and becomes progressively more
- 93 -
-------
100
c:
o
.-
en
en
.-
~8
tit
~
c:
.-
560
.-
+'
U
:J
'U
4)
'-
~ 40
°
100
c: 80
I 0
I '-
. en
en
E
Cb
N
o
V)
c:
c
o
'-
+'
U
I. -6
Q)
L.
f!. 20
)(~
x
Pittsburgh coof ,.' .,..x-- '
, .'
L stone 18 (-1680fJm) ,
Bed depth: 2ft
FJuidising vel; 4ft/s
No recyc Ie
36in combustor
Co/Smol ratio: 2.2
1400 1500 1600
Bed temperatu re: of
x
/
/x"'x
x
/
. '
. . ;~ ,:.~ x.;....... ,
Pittsburgh coal
Dolo~ite 1337(-1680J.Lm) .
Bed depth: 2ft
F/uidising'vel: 4ft/s
. No recycle. 36;n combustor
Co/S mol ratio : 2'7
1400 1500 1600
, °
Bed temperature: F
1700
Pittsburgh coal
Limestone 18 (-3175 f-lm)
Bed depth: 2ft
Fluidising vel: 8ft/s
With recycle
27in combustor
CalS mol ratio; 2'8
o
1400 1500 1600
Bed temperature: of
~
Pittsburgh coal ~
DOlomite 1337 (-1587f!m) "~
Bed depth: 2ft
-------
100
tI)
:~ 80
E
Q)
070
If)
c
.- 60
c
o
~
~ 50
u
Q)
L.
~ 40
°
Fi g. 7. 7.
1400
0-00
'-00
O~
o 0
\0
1500 1600 1400
Bed temperature: of
Welbeck cool
u.K. Limestone (-3175 J-Lm)
Bed depth: 2ft
F"lu idising velocity: 8ftts
Cots mol ratio: 2.8
No recycle
36in combustor
o
° -O--O~8
00""
0"0
1500 1600
Bed temperature: of
Pittsburgh coal
Do 10 mite 1337 (-3175 ~m )
Bed depth: 4ft
F"luidising velocity: 8ftts
Cots mol ratio: 5'3
No recycle
36 in combustor
Variation of 502 reduction in temperature surveys
-95-
-------
Table 7.3
Effect of Combustion Temperature on
Reduction in S02 Endssion
\0
C\
Additive and Nominal Nominal CajS % Reduction in S02 Emission at:
Coal and Task Bed Ht. Velocity Recycle mol
Top size pm Top size pm ft. ft/s Ratio 12900F 13800F 14700F 5600F 6500F
Pittsburgh Limestone 18 I 2 4 No 2.2 - 50 81 79 -
1680 1680
Pittsburgh Limestone 18 III 2 8 Yes 2.8 - - 81* 80* 62*
3175 3175
Illinois Limestone 135.5 V 2 3 No 2.2 19 - 64 - "-
1680 1680
We1beck UK Limestone I 2 8 No 2.8 - - 72** 64** 36**
3175 3175
Pi ttsburgh Dolomite 1337 I 2 4 No 2.7 - 72 74 73 -
1680 1680
Pi ttsburgh Dolomite 1337 III 2 8 Yes 0.5 - - 61* 45* 22*.
3175 1587 ._--
Yes 1.9 - - 76* 74* 57*
No 4.8 - - - 77 64
Pittsburgh Dolomite 1337 I 4 8 No 5.3 - 95** 94** 89** 72**
3175 3175
*
**
Interpolated from plotted data (Fig. 7.6)
Interpolated from temperature surveys (Fig. 7.7)
-------
pronounced whilst the peak temperature faUso If this beha','i:":.u::- c'curs t:nder
combustion conditions in the plant» it could be expe-:-:ted that the shape and
tempe,rature of the peaks in the curves shown in Fi gs, 7,6 and ].] would depend
on factors (e.ge Ca/S ratiop fluidising jfelocity~, bed height, :recyding par::ide,
size,
additive and coal characte:::istic.:s) that innuenc~ the f'ra'. den d 1 ime
utilised.
It is known that calcium sulphate decomposes at temperatures above
19000F, but the exis tence of optimum temperatures below l6000F was une:xpect,ed
both from the laboratory work and from the good retention performance obtained
by Zielke, Lebowitz, Struck &'Gorin" working at l8000F with dolomiteo It is
confirmed however by the behaviour of limes tone., found by Esso Research Engineering
C 10
o.
The explanation may be in the fac.to!'s that affect the conversivn of the
initial reaction productp CaS03 (unstable)~ into CaS040 If9 for example?
hydroxide ions are involved it is possible that there will be a temperature above
which the stability of hydroxide ions on the internal surface of the particles
will decrease, and hence the nett retention of S02 will be lvwer. The
possibility of hydroxide ions being present was demonstrated in the pore structure
studies (Appendix 8, Section 2.40208)u
Reduction of tempe'ratu're below the optimum ul timately resul ts in a
major increase in sulphur emission when limestone is the additivep whereas the
effect with dolomite is small. When limestone is used~ reaction is largely
governed by the temperature-dependent rate of the calcination reaction which
develops pores for, penetration by S02 and internal surfaces for reactiono
is probably this factor that gives limestone a more clearly defined optimum
temperature range than dolomite; The MgC03 in dolomite cale:ines at a lower
temperature than CaCO) and hence gives access to internal surfaces that would
otherwise be masked by an i.mpermeable layer of CaS040 The higher S02 emission
with limestone at low temperature could limit reduction of load by reduction of
It
temperature °
The critical and possibly anomalous effects of temperature on
sulphur emission warrant careful consideration in any futur'e worko
- 97 -
-------
Veloc i ty : ft/s .
10 5 3 2
-
a::
"" ~
.!"
6 al V c
o
85 .-
1 /~ en
/ en
E
- /
Q: P ~
~4 /- BON
o 81 0
I \I)
o / .C
o
yo- ""
~ / +
Q: /
~ 2 +",, ""
0 .
", ~-
--x-- --A-
0 0.1 0.2 0.3 0.4 0,5
1/Velocity: ft-'s
Symbol Coal Add itive Size Bed depth Bed temp. ColS
IJm ft 8F morratio
x Pittsburgt 18 3175 2 1560 1.7
o " u 1680 2 1470 2.2
6 Welbeck UK.l's1cn8 " 2 " 0.8
V " " " 2 " 1.8
+ " " 3175 2 1560 2.8
C .Pittsbur91 18 " 2 " 2.7
3ein. 27in and 6in combustors
Fig. 7. 8. Empirical relation between SOa redu ct ion
and fluidising velocity
-98-
-------
7.1.5
Effect of Fluidising Velocity
Increase in fluidising velocity is accompanied by the following effer ts:
(a)
(b)
Reduc. don of gas res idence tj me in the bed.
Higher voidage of the bed, with reduction in the
weight of solids present in a bed of given height~
(c)
and hence reduction of solids residence time.
Bypassing of a greater proportion of the gas in
bubbles.
(d)
Increased throughput of coal, and hence of
additive for a given Ca/S ratio and given
pr~portion of excess a~r; this results in a
further reduction in solids residence time.
These effects can be expected to result in poorer sulphur retention when
the fluidising velocity is increased.
showing effects of fluidising velocity.
Table 7.4 summarises the results of tests
Most of the results show that, as
expected, the percentage reduction in 502 emission falls with increase in
fluidising veloc.i ty. The results can be approximately correlated by:
A = XI/V
where A is the Absorption Ratio, defined as the ratio of 502 absorbed to S02
e.mitted~ Le. (% Reduction)J(lOO - % Reduction) and V is the velocity; the value
of the constant Xl depends on the operating conditions. The validity of this
simple relationship can be judged from Fig. 7.8.
It is estimated that to reduce sulphur emission to 300 ppm for a 3%
sulphur coal at a bed temperature of 1500oF, it would be necessary to supply about
twice the quantity of additive at 8 ft/s as at 2 ft/s.
- 99 -
-------
TABLE 7. 4
Effect of Fluidising Velocity on
% Reduction in 502 Emission
......
o
o
I
'
coal Type Addi ti ve Type Si zes jJm Task Nominal Temp Recycle CalS % Reduction in S02 Emission at Nominal
Bed Ht. of Mol Veloci ty (ft/s) of:
Coal Addi ti ve it. Ra t.i 0 2 3 4 6.5 ~ 11
Pi ttsburgh Limestone 18 -3175 -3175 I 2 1560 L7 54 38(1) -
No - - -
-3175 -3175 III 2 1560 Yes 2.7 - - - 85 81(2) 77
-3175 -3175 III 2 1560 No 2.7 - - - 80 ,. 7'
-1680 -1680 I 2 1470 No 2.2 84 81 i
- - "1 -
=1680 - 125 I 2 1560 No 1.0 - 50 50 - -; -
.
We1bec:k U,K. Limes tone -1680 -1680 V 2 1470 No 0.8 53 45 - - - -
-1680 -1680 V 2 1470 No 1.8 86 81 - - "7 -
-3175 -3175 I 2 1560 No 2.8 - - 69 (3) - 64 -
NOTES:
(1)
(2)
(3)
Extrapolated
Interpolated
Extrapolated
from Ca/S = 106
f'l'om smoothed curve
from Ca/S = 3.0
-'
i .
I
-------
70L6
Effect of Bed Height
Increase in bed height increases the residence t.ime of both gas and s':did8
and would be expected, therefore, tv improve the percentage reduction in. 502
emissicn " Tab Ie 7 - 5 summarises the resul ts of comparison tes ts showing the elfe,: t
of bed height. The first of the comparisons, for Welbec:k -::cal with U.K. limestonE:,
showed no change in 502 removal when the bed height was nearly doubled, but the
remaining comparisons do show improvement. With Pittsburgh coal and Dolomite 1337
at a Ca/S mol ratio of 5.0, the percentage reduction was high even at the lowest
bed height, and increasing the bed height would not be expected to cause mo'!'e than
a small improvemenL The tests with Illinois coal and Limestone 1359 shew
substantial improvements in S02 removal on changing bed height from 2 ft to 3 ft,
suggesting that, as might be expected, a bigger improvement would be obtained at
low rather than at high sulphur retention.
With the exception of the first comparison in Table 7.5, the results can
be approximately correlated (Fig. 7.9) by the relationship:
A
=
HX2
where H is bed height and A is the Absorption Ratio.
X2 depends on the other operating condi~ions.
Bed depths of about 2 ft are needed in atmospheric pressure plant to
The value of the constant
ensure good mixing of'coal, and even at the highest fluidising velocities likely
to be used, this depth is adequate to accommodate the heat transfer surfaces
immersed in the bed.. The fourfold increase in bed depth theoretically needed to
compensQte for an increase in fluidising velocity from 2 ftls to 8 ftls would
cause an intolerably high pressure loss in a non-pressurised combustor. On the
other hand, deep beds are needed in pressurised combustors to accommodate the
required amount of heat transfer surface; fortunately high pressure drops are not
a limiting [actor at high pressure. The bed depth needed is broadly proportional
to the 'pressure; it is also proportional to the fluidising velocity, and might
largely compensat.e for the adverse effect on 802 reduction of increase in
velocity..
In the present pilot-scale tests, and in any commercial application of
fluid-bed combustion, upwards of 80% of the gas passes through the bed as bubbles-
According to well-established theory there is a continual exchange between the gas
in the bubbles and that in the rest of the bed which results in the high gas/solid
contacting efficiency of fluidised beds.
As the bubbles progress upwards through
- 101 -
-------
12 -
a::
~
~
10 90 c
o
,-
en
en
"E
o G>
85 t\I
~
c
80 '-
/ c
./'V ~ 0
70 :;;
2 ~ ......... ----- X 60 ~
~--A-- 50 u
Q)
f-
o 1 2 3 4 5 6 7 ~
o
Bed depth: ft
Symbol Cool "Additive Size Velocity Bed temp, Co IS
~m ft/s " of mol ratio
X Wel beck U.K,L'stone 3175 8 1560" 2.8
o Pittsburgh 1337 " " " 5.0
A Illino is 1359 1680 3 1470 1 . 1
'V " " " " " 2'2
36in and 6 in combustors
Fig.l9. Empirical relation between 502 reduction
and bed depth
-102 -
-------
TABLE 7.5
Effect of Bed Height on %
Reduction in S02 Emission
.....
o
UJ
Coal Type Addi ti ve Task Nominal T~ Recycle GaJS % Reduction in S02 Emission with Nominal Bed
and size, pm Type Vel of Mol Ht (ft) of: .
ftls Ratio 2.0 2.7 3.0 3.8 4.0 .5 6.9
Welbeck U.K. Limestone I 8 1560 -No 2.8 64 - - 64 - - -
-3175 -3175
Pi ttsburgh Dolomite 1337 I 8 1560 No 5.0 - 85 (1) - - 87(2) 87 92 (3)
-3175 -3175
Illinois Limestone 1359 V 3 1470 No 1.1 37(4) - 52 - - - -
-1680 -1680 73(5)
2.2 64 - - - - -
Superscripted data extrapolated from the following GalS Ratios.
(1)
(2)
(3)
(4)
(5)
GalS = 5.4
GalS = 5.3
GalS = 5.2
GalS 1. 5
GalS = 2.1
-------
the bed they coalegce and form fewer, but larger, bubbles moving with ever-
increaging velocity. In the absence of any correcting influence this results
in decreasing contact efficiency with increaging bed depth. The presence of
tubes in the bed ig known to help in breaking up these large bubbles and
restore the high gagJsolid contacting efficiency. The effect of increasing
bed height can therefore be expected to depend, amongst other factors, upon
the amount of heat transfer surface and its distribution in the bed. Because
of this, increase in bed height may be expected to be more effective in
pressurised combustorg in which the bed is fully packed with tubes.
- 104 -
-------
7.1.7
The Effect of Operating Pressure
Increase in operating pressure whilst maintaining f1uidising velocity
and excess air at the same levels reduces the solids residence time in the bed
in proportion to the increase in pressure (unless of course the depth of the bed
is increased in proportion). From this point of view lower sulphur retention
might be expected under pressure than at atmospheric pressure.
Increase in pressure can also be expected to inhibit calcination and
hence pore formation.
It will be seen from Fig. 7.10 that at atmospheric
pressure the partial pressure of C02 in a boiler is below the equilibrium value
o u
at temperatures above about 1380 F, and calcination can therefore occur in this
region. At operating pressure greater than 2 atm, the partial pressure of C02
in a boiler is above equilibrium at 1470oF, and since calcination is therefore
impossible, poor performance would be expected in a'pressurised'combustor with
limestone. The equilibrium partial pressure for calcination of MgC03 at all
combustor temperatures is very high, so that the MgC03 in dolomite can be
expected to calcine in pressurised combustors giving access to internal surfaces
and satisfactory sulphur retention.
The results of the tests in the pressurised combustor are detailed in
Table 4.9 and shown in Fig. 4.9 of Section 4. Dolomite was markedly more
effective than limestone on the basis of the CalS mol ratio. The quantity of
additive required to achieve a required level of emission is also less with
dolomite than with limestone on a weight basis. For example, to achieve 300 ppm
when burning a 3% sulphur coal would require about 7.6 lb/lb sulphur for dolomite
as compared with an estimated 10 1b/lb sulphur for limestone. This result was to
be expected from theoretical considerations and from the results of the porosity
measurements from calcination experiments in the laboratory (Appendix 8).
The
I .
performance with the limestone additive was, however, not as poor as had been
expected. This may be due partly to chemical and physical characteristics of
the particular additive (Limestone 18) used in the tests, and partly to attrition
in the combustor of the CaS04 layer on the surface of the particles.
Comparison of the performance of the pressurised and'non-pressurised
combustors (Table 7.6) whefi using dolomite as the additive shows that sulphur
retention efficiency for the pressurised combust6r is well above that for the
non-pressurised without recycle of carryover, but is below the performance
attained with positive recycle.
- 105 -
-------
1100
Temperature: of
1200 1300 1400 1500 1600
1800 2000
VI
Q)
L
Q)
~
a.
VI 0'4
0
E 0'3
+'"
0
0'2
(\I
0
U
.....
0 0'1
Q)
L
::::s
VI
VI
Q)
L
a.
0
.-
+'"
L
C
a.. ..':)~ CaD + CO2
0'01 t.,' Fiel d
.'<)
.~
~
~
0'002
6'5
Approximate
range of conditions
in pressurised fluid-
bed c0'l0r
Ca COJ
Field
~~
G
rJ
II.
c:;:,'"
<.J
Approx i ma te range
of conditions in
atmospheric pressure
fluid-bed combustors
4'0
Fig. Z 10. Equil ibrium diagram for the CaC03/C02
system in relation to conditions in
fluidised bed combustors
,106,
6'0. 5'5 5'0 4'5
',Reciprocal temperature: 104jOR
-------
, I
I
It is considered that the slightly poorer performance of the pressurisE'C
','0.mbustcI' compared with the non-pressurised combustor with recycle re,flect~ the
:nterlor efficiency of the recycle system on the pressure combustor and r.hc:t ",n}'
e([el:t of pressure is small.
Table 7.6
Comparisons of i. Reduction in S02 Emission
at Atmospheric and at High Pressure
Coal:
Pittsburgh
14700F
Temperature:
Pre,ssure Additive Task Bed Ht. Velocity CaJS Mol Recycle %'Reduction
Atrn.Abso Type Size ft. ft/s Ratio in S02
)lm Emiss~on
1. Dolomite -1680 ,1 3.7 2.1 2.0 No 80 (1)
1337 Yes (2) 97 (3)
3,5 -1587 II, 3.8 1.9 2.0
5 -1587 II 3.8 2.0 2.0 Yes (2) 98
1 Do 10m te -1680 I 3.8 2.2 1.6 Yes (4) 99
1337 Yes (2) 94 (5)
5 -1587 II 3.8 2.0 1.6
1 Limestone -1680 I, V 2.0 3.0 1.9 No 77 (6)
18 Yes (2)
5 -1587 II 3.8 2.0 1.9 66
1. Limestone -1680 I, V 2.0 3.0 2.4 No 87 (6)
18 Yes (2)
5 -1587 II 3.8 2.0 2.4 77
,
NOTES: (1) Extrapolated from CaJS = 2.2
--
(2) Recycle system thought to be ineffective
(3) Extrapolated from CaJS = 1.9
(4) Positive recycle of fines
'(5) Interpolated from Fig. 4.11
(6) Interpolated from Fig. 7.1
- 107 -
-------
7.108
The Effect of Recycle
Fines elutriatedfrom fluidised combustors contain unconsumed caTben.
and the loss of carbon can be reduced by recycling the fines. Since the
elutriated fines contain incompletely utilised additive it can be expected t.h""t
fines recycling will also improve the percentage reduction in S02 emission.
Table 7.7 summarises the data on the effect of recycling. It should be
. emphasised that the only positive recycle systems were those employed in the
36 in and 27 in combustors;
the 24 in x 48 in pressure combustor. and 12 in and
6 in non-pressurised combustors had cyclones with internal dip-leg. and were
probably less effective than the efficiences for which they were designed.
Table 707
Effect of Fines Recycle on .
% Reduction in S02 Emission
Coal Type Additive Task Nom NOIll Temp Ca/S Mol % Reduction
and size, Type Bed Vel of Ratio Without With
pm. and size, .Jt. ft.' ft/s Recycle Recycle
pm
Pittsburgh Limestone 18 III 2 8 1550 to9 56* 62(1)
-3175 -3175 (2) *
III 2 6.5 1540 2.7 80 8:
III 2 ~0.9 1560 2.7 78 (3) * 77
Pittsburgh Limestone 18 III 2 8 1550 1.9 50 (1 ) * 33
-3175 - 150
Pittsburgh Dolomite 1337 I 3.8 2 1470 1.6 73(4) 99
-1680 -1680
Welbeck UK Limestone V 2 2 1470 0.8 53 55
-1680 -1680 I
V 2 3 1470 0.8 45 75
* With internal recycle only
NOTES: (1) Extrapolated from Ca/S = . 1.6
(2) Extrapolated from CaJS = 2.8
(3) Extrapolated from Ca/S .:: 2.6
(4) Extrapolated from CaJS .:: 2.2
- 108 -
-------
The c.omparison for Pi ttsburgh coal with Do10mi te 1337 obtained horn the
36 in .~ombus to,;:, shows that a substantial improvement was obtained.
One. of tht:: . we
compa'.dsons for We1beck coal with U.K. LiJnestone from the 6 in combustor shows a
marked imp',o:ovement, but the other does not.
The remaining comparisons were obtained from tests in the 27 in .dg which
had a medium efficiency internal cyclone, in operation at all times. It is
possible, however, that the hopper system did not operate sufficiently qui.ckly t':;
cope with the greater dust loading when the external cyclone was in operation"
This may account for the anomalous effect of recycle in the third and fou~th
comparisons in Table 7.7.
It is concluded that recycle of fines (particularly where the cyclone
has a low cut size) is beneficial and that in those circums tances where this WdS
not clearly' shown, the effectiveness of the recycle system is open to doubt.
It may no.t ,however, be feasible to exploit fines recycle fully in
commercial plant. Recycling fines involves engineering problems and, except in
a few special circumstances where highly efficient internal cyclones can be
envisage.d, the recycled material causes a thermal load on the fluid-bed. These
problems are clearly greater at high fluidising velocities and it is likely that
less effi.e-ient re-circu1ati(;>n c.yclones would then be used to limi t the amount of
material in circulation. It should also be noted that, as discussed in
Section 7.5 fines recycle increases particulate emission.
7,1.9
Effect of Additive Particle 5ize
The 'rates of -many- gas-soli::ds reactions are proporti.onal to the solid
surface area, and are tnerefore an'inverse function of particle size. In
fluidised combustion systems, therefore, reduction in additive particle size should
improve, 502 removal, particularly for additives where
obscured by an impervious shell of calcium sulphate.
size may result in rapid e1utriation of the additive
the internal surface becomes
However, reduction in paxticle
and hence reduce the residence
time in the bed from as much as several hours to as little as a few seconds, and the
net effect may be a reduction in 502 removal efficiency. Further reduction in
particle size below that at which rapid e1utriation occurs will not greatly affect
the residence time of particles in the bed, 'and hence would be expected to improve
502 retention. Therefore there may oe an optimum particle size for coarsely
crushed additive, below which retention decreases to a minimum and then increases
for very fine particles. With both coarse and fine additive, the 502 removal
efficiency w~uld, in general, be expected to fall ~ith increase in ve1ctity,
- 109 -
-------
100
80
c:
o
.-
en
en
E
4> 60
N
o
V)
c:
c:
o >
.- 40
+'
U
::J
~
o
0'" 20
o
x
~
X/>
X/
X .
x
o
1
2 .~
ColS mol ratio
4
Pittsburgh coo I
Limestone 18
Bed depth: 2ft
Fluidising- velocity: 8ft/s
.
Bed temperature: 1550 F
With recycle
27in combustor
Particle size JX -3175 J-Lm
. \0 -150 I-Lm
Fig. Z 11. Effect of add itive particle size on
502 reduct ion (high fluidising velocity)
-110-
5
-------
100
o
80
c
o
.-
en
en
./
E
4>60
C\I
o
tf)
c
o
/Y'
x
X/
/
~40
+'
U
::J
U
4>
'-
...! 20
o
o
1
2 3
Ca/S mol ratio
4
III i no i s coa 1
Limestone 1359
Bed depth: 2 ft
Fluidising veloc j ty: 3ft/s
Bed temperature: 1470°F
No recycle
6in combustor
P~rticle size {X -1680l-Lm
o -125f-.Lm
Fig.712. Effect of additive particle size on
502 reduction (low fluidising velocity)
-111-
-------
100
c
o
'j;) 80
en
E
~
0-
-
",,'"'t)
"
/ ,,"
X /
~///
/;//J / 36 in combustor
/ / Bed depth: 2ft
;6 &/ particlesize{X-1680pm
/ 0 . 0-3175~m
/
C\I
o 60
(/)
c
c
.2 40
+-'
U
:J
U
~
L 20
o
0'
FJuidising velocity: 8ftts
Bed temperature: 1560°F
No recycle
o
1
2
3 4
CalS mol ratio
5
6
Fig. Z 13. Effect of additive particle size on 502
reduction ( Limestone 18)
100
c
o
~ 80
E
~
~
~
C\I
~ 60
c
.-
OX
/
27 in & 36 in combus tors
Bed depth: 2 ft
Particle Size{X -1680pm
o -3175 J-tm
Fluidising velocity: 8ftts
°
Bed temperature: 1560 F
No recycle
c
.2 40
+-'
U
:J
U
~ 20
~
o
o
1
2
3 4
CalS mol rat io
5
6
Fig. 7.14. Effect of additive particle size on 502
reduction ( Dolomite 1337 )
-112 -
-------
The resu1t~ of tests which had the main objective of determining the
effect of particle size are shown in Table 7.8 and in Figs. 7.11 and 7.12.
Additional data obtained over a range of fluidising velocities and bed height
have been extrapolated to common operating conditions using the correlations in
Sections 7.1.5 and 7.1.6 and are given in Figs. 7.13 and 7.14.
Table 7.8
Effect of Additive Particle Size
on % Reduction in S02 Emission
Coal Type Additive Task Nom Nom Temp Recycle CalS % Reduction for
Type Bed Vel of Mol Additive Size(~m) of
Ht. [tis Ratio -3175 -1680 150 ~ - 125
-
ft.
Pittsburgh Limestonl I 2 4 1560 No 1.7 54 68(1) - 69(2)
18
III' 2 6.4 1550 Yes 2.8 86 (3) - 83 -
III 2 8 1550 Yes 1.0 38(4) - 28(4) -
III 2 8 1550 Yes 2.0 65 (4) - 50 ( 4) -
Illinois Limestone V 2 3 1470 No 1.1 - 39 (5) - 62
18 81 (5)
V 2 3 1470 No 3.6 - - 93
NOTE S: (1) Extrapolated from Cal'ii == 1.3
(2) Extrapolated from Cal'ii == 1.0
(3) Temperature -- 1525°'F
(4) Interpolated from 'F ig. 7.11
(5) Interpolated from 'Fig. 7.12
Fig. 7.13 indicates that a reduction in the limestone particle size from
-3175 )lm to -1680 ~m improved 802 reduc.tion; this reduction in size would not be
expected to affect e1utriation significantly. Fig. 7.14 indicates that a similar
size reduction for dolomite has no effect~ suggesting that access to internal
surface is not a limiting factor fOl dolomite.
- 113 -
-------
Reduction of additive particle s-ize to lesS- than about 150 1iin results in
poorer sulphur retention at high 'Velocities CFig. 7.11) But improyes retention at low
velocities (Fig. 7.12).
The former is in contrast to the results puBlished By rope~ Eyans and
Robbins5; they show that for Lbnestone 1359 at a Ca]S''lIlo1 ratio of 206, with
velocities of 12 ft}s in a Bed only 10 inch deep,8C>~ retention was oBtained when
the limestone was ground to -44 pm, and 60% retention w1th.~74 pm limestone. This
is in line with the initial comments, namely, that decrease in size down to a certain
level will result in a reduction in S02 retention, But that further decrease in size
is likely to result in an improvement in S02 retention.
7.1.10
Effect of type and source of additive
It was shown in Section 6.2 that the structure (porosity) of the calcined!
sulphated limestone particles varies with the source of the stone, and that dolomites
behave in a different manner from limestone.
In the pilot-scale programme three limestones and two dolomites were used.
Data comparing Limestone 18 with U.K. Limestone are shown in Fig. 7.15 and Limestone
1359 with U.K. Limestone in Fig. 7.16. From these figures it can be deduced that
Limestone 18 is more effective than U.K. Limestone which in turn is more effective
than Limestone 1359.
Further data for operation at atmospheric pressure are taBulated in
Table 7.9. These show that Limestone 18 is more effective than Dolomite 1337 except
at low temperatures (1380oF). Comparison of the data at l4700F in Table 709 with
Fig. 7.15 suggest that Dolomite 1337 is similar to U.K. Limestone.
Thus on a molar basis the above four additives can be placed in descending
order of effectiveness as follows:
Limestone 18
U.K. Limestone and Dolomite 1337
Limestone 1359
This order of effectiveness may be compared with that ind{cated by the
laboratory sulphation test, Figo 6.1. In this test, the Dolomite 1337 was slightly
superior to Limestone 18 and both were markedly better than Limestone 1359. The
difference between the behaviour in the f1uidised bed and in the laboratory is thought
to be due to attrition of the particle surface in the fluidised bed. In the
laboratory test (Section 6.202) it was found that access of S02 to the interior of
limestone particles was reduced with increasing degree of su1phation.
- 114 -
-------
t-'
t-'
\J1
TABLE
7.9
Effect of Type of Additive on
% Reduction in $02 Emission
Coal Type Additive Task Nominal Tempo Recycle F1uidising Cals % Reduction
and Size~ Type and bed Ht. of velocity mol in,S02
].lm Size, ].lm ft. ft/s ratio em1SS10n
Pittsburgh Limestone 18 I 2 1380 No 4 2.2 50
-1680 -1680
Do1arilite:':1337 I 2 1380 No 4 2.2 (1) . 66
-1680
Limestone 18 I 2 1470 No 4 2.2 81
-1680
Dolomite 1337 I 2 1470 No 4 2.2 (2) 67
-1680
Limestone 18 I 2 1560 No 4 2.2 82
-1680
Dolomite 1337 I 2 1560 No 4 2.2 (2) 53
-1680
Pittsburgh Limestone 18 I 3.8 1560 No 8 5.7 96
-3175 -3175
Dolomite 1337 I 3.8 1560 No 8 5.7(3) 90
-3175
We1beck Limestone 18 V 2 1470 No 3 1.9 78
-3175 -1680
U.K. Limestone V 2 1470 No 3 1.9(4) 83
-1680
(1)
(2)
Extrapolated from CalS
Extrapolated from CalS,
Extrapolated from CalS =
2.6
2.7
(3)
(4)
5.3
1.8
=
Extrapolated from Ca/S
=
=
-------
100
/
o
/ x
o
01
c
o
en
en
80
E 60
Q)
N
o
tI)
c
.-
c 40
o.
......
U
:J
"0
Q)
L.. 20
...!
o
x
x
I
o
1 2
CalS mol rat i 0
3
Pittsburgh coal (-1680 ~m)
Bed depth: 2 ft
Fluid ising velocity: 3ft/s
o
Bed temperature~ 1470 F"
No recyc Ie
3Sin and Sin combustors
o Li mestone 18
X U.K. Li mestone
Fig.Z 15.Compari'son 01 502 reduction with
.Limestone 18 and U.K.Limestone
-116-
4
-------
100
c 80
0
"-
en
en
E
Q)
C\I 60
0 x
t/)
c
.-
c
.~ 40
+-'
U
:J
'C
Q)
L.
~ 20
o
1
2
CalS mol ratio
3
Illinois coa I (-1680 J-lm)
Bed depth: 2ft
Fluidising velocity: 3ft/s
o
Bed temperature: 1470 F
No recycle
Sin combustor
o Limestone 1359
x U.K. Limestone
Fig.Z 16.Comparison of 502 reduction with
Limestone 1359 and U.K. Limestone
-117-
4
-------
In Limestone 18 this was due to progressive reduction of the large 'transport pores';
in Limestone 1359 there was an aBsence of ttransport pore~' and an i~permeable surface
layer of sulphate was rapidly formed. Attrition of the particle surface would thus
account for the small improvement in sulphur retention for Limestone 18 and the large
improvement for Limestone 1359, both relative to the performance of Dolomite for
which access to the internal surface is not a limiting factor. The absence of a
sulphated surface layer on Limestone 1359 particles from one of the fluidised bed
combustors (see Section 6.203.) provides direct evidence that attrition of the
surface layer was occurring in the bed.
It should be remembered. that the calcium 'content of dolomites is considerably
lower than that of limestones (see Table 302, Section 3), and that Limestone 18 has a
somewhat lower calcium content than other limestones.
above order of effectiveness is changed to:
Thus on a weight basis, the
Limestone 18
U.K. Limestone
Limestone 1359
Dolomite 1337
With Limestone 18 only slightly superior to U.K. Limestone. An anomalous result,
indicating a slight superiority of U.K. Limestone over 'Limestone 18, was obtained
with We1beck coal. This anomaly might be due to the low sulphur content of the
coal.
It was shown in Section 701.7 that under pressure Dolomite 1337 is
markedly superior to Limestone 18 on both molar and weight 'base:!:\ but the retention
by the uncalcined Limestone 18 was better than expected, again probably because of
attrition of the surface layer of sulphateo U.K. Dolomite used, with Welbeck coal
gave better percentage S02 reduction than Dolomite 1337 with Pittsburgh coa1~ but
this might again be, an effect due to the coal rather than the dolomite.
Efficient utilisation of additive is important from the point of view of
minimising thermal losses. In addition to sensible heat losses, calcination is
endothermic:
CaC03 = CaO + C02 43 kca1Jgmo1 (298oK)
MgC03 = MgO + C02 24 kca1/gmo1 (298oK)
However, su1phation is strongly exothermic:
CaO + S02 + ~02 = CaS04 + 120 kca1/gmo1 (298oK)
- 118 -
-------
With a high utilisation of stone, e.g. 85% reduction of 502 emission from a 3%
sulphur coal with a CaJS mol ratio up to aBout 2, tnere is approxiuately a thermal
balance for limestone at atmospheric pressure, and a heat loss witn dolomite of up
to about l!% of the thermal input to the Boiler. At a CalS mol ratio of 4~ the
heat losses incurred to achieve 85% 502 reduction are increased to Z% for limestone
and 4% for dolomite. Under pressure, calcination,of CaC03 is inhibit'ed, and the
heat losses are consequently reduced. For example, tneneat loss for 85% 502
reduction with dolomite under pressure would Be negligiBle at CalSuol ratios up to
2, and only 2% of the coal heat input at a Ca/~mol ratio of 4.
7.1.11
Effect of coal type
The most important coal property isobvl0usly the sulphur content, since
the actual feed rate of additive required for a given percentage SOZ reduction is
approximately pr~portional to this. Also, the percentage S02 reduction which must
,be achieved in order to obtain a particular S02 emission is lower with a low-sulphur
coal. Secondary considerations are the amount of calcium present in the ash~ which
affects the emission of 80Z without addi~ive, and the total amount of ash, which
affects the residence time of solids in the bedo In order to allow direct
comparison of the effect of the operating variables with the various coals? the
effect of sulphur and calcium content has been eliminated by plotting the results
as the percentage S02 reduction'against the molar ratio of added calcium to total
sulphur in the coal, Fig.,1.17. This approach is based on the assumptions that
the process of 802 reduction is directly proportional to the S02 concentration in
the gas, and the close agreement between the results from the four coals plotted on
this basis confirm that this is a good appro:rl:mation. RoweV'er, at any given CaJS
mol ratio the percentage S02 reduction fOT WelBeck coal was higher than for the other
three coals; there may a1so be some gma11er differences between the other coals.
The better reduction with Welbeck coal could not be explained in terms of
the coal rank, as measured by the volatile matter or caking capacity of the coal.
It was found that the rate of release of pyritic sulphur from coal is about an order
of magnitude more rapid than that of organic sulphur, with the actual r~tes differing
between coals. The total rate of sulphur release' from Wel£eck coal was more rapid
than for Pittsburgh (the other coals were not tested), and this may have led to a
lower proportion of the sulphur release in the ~ombustor being from elutriated fines
in the freeboard, where reaction with stone is inefficient.
- 119 -
-------
IVV
20
Cool Symbol
Illi nois 0
Welbeck !:l
ParkHill 'V
Pittsbu"9h 0
/)
90
80
70
c
o
.-
en
en
'E.60
Q)
(\I
o
\I) 50
c
Park Hill, Illinois
& Pittsburgh
[]
.-
c
'0
.-
t 40
~
'C
G)
L.
'I- 30
.~
o
Fluidising velocity: 3ft/s
. 0
Bed tempera ture : 1470 F'
Bed depth: 2ft
o
1
2
CalS mol ratio
3
Fig.~1~ Comparison of 502 reduction for different
coals with U.K.Limestone
-120-
-------
An alternative explanatLon is that Because of the low feed rate of
additivE' with the loW'''',sulphur WelBed~: coal" the residenc.e tUne of cca'rse additive
partie les in the. bed is longe'l'. This -may' Be expected to allow a higber
utilisation of the stoneq particularly-'if to.e p-r:0cess of attrition is involved~
and hence~ since the CaJ~mol ratto ts the same, a higo.er reduction of SO
2
emission is obtained.
7.1.12
The approach to steady state
In the context of this discussion, the approach to steady state is
the attainment of a new leve'l of S02 emission after making a change in ope:rating
conditions. The time to reach steady state is not always the same as tbat
requi.red for the concentration and sulphation of additive particles in the bed
to reach their final values.
Changes in bed composition often occur slowly. For example~
starting with an additive~free bed in the pressure combustor~ a steady S02
emission was established after about 60 bours when dolomite was fed at a CaJS
mol ratio of about 0.8 (Appendix Ir~ Fig. A2.3.l). Comparison of the analyses
of, the bed material at this time with those during the subsequent 40 hours
operation showed that some bed composition changes were still occurring. On
the other hand~ the fina~ 802 emission level was reached after only 25 hours in
a test started with a partly sulphated bed (Fig. 7.18). Although the app:roach
to steady state was slow in both instances~ the response of 802 readings to
fluctuations in coal feed rate (and possibly the CalS ratio) was rapid.
The behaviour with limestone in the pressurised combustor was
different. The 502 emission reached its minimum value within about six hours
of starting to feed limestone. The retention efficiency was lower and the
fluctuations in 502 emission greater in amplitude and frequency than with
dolomite. It might be expected that with 1imestone~ which is significantly
more reactive when the extent of sulphation is low, steady state conditions would
be achieved before the total calcium invento'ry had stabilised.
This would
explain a much more rapid app:t'oach to steady state S02 emission with limestone.
Experiments at atmospheric pressure in the 6 in combustor also showed
the relative unimportance of achieving a stable limestone concentration.
After six hours' operation, the percentage S02 reductions were:
(a)
(b)
43% after starting with a lime-f'ree bed~ and
38% after starting with a lime-rich bedo
- 121 -
-------
-----.....,....,.-,~,
.a
- -
u 600
Q)
.a
c
~400
0
+'
C
Q)
>
c 200
--
0
0
U
0
-
I >
I -- \0
- ;>
t\:) .......
t\:) ~30
I
00
a.
0
20 30 40 50
Ti me from start of test: h
Fig.7.18. Variation of 502 emission and of total and unconverted
CoO inventories in 48in X 24in .Combustor
+'
C
~ 100
c
o
U
N
o
\I)
o
Inven tory of
uncon ver fed Co 0
o
o
o
00
n rO-()-, 0 0
o 0 0 Ou 0
o oep 0
o
o
cP
10
60
70
-------
-1-1800
E
a.
a.
;n1600
+'
u
.::J
u
~1400
a.
c
o
~ 1200
::J
.D
E
81000
'6
+'
~ 800
+'
C
8
Q) 600
u .
.-
x
o
.-
u 400
L
::J
.s:;
a.
S 200
c.J)
Tes t 2.5
---
Test 2.6
Pittsburgh coal (-3175 J-im>
Limestone 18 (-150 f-Lm >
. . {Test 2.5: 0.88
CalS mol ratio Test 2.6 :2.47
Fluidising velocity: 8.1 ft/s
Bed temperature: 1530°F
27in combustor
Fig:7.19.Variation of 502 emission after increasing rate
of limestone addition in 27in Combustor
t
Change of rate of
limestone addition
o
310 320 360 370
Time: hours from start of Test
Ser i es
-123-
-------
2000
cI)
+'
g 1400
1:)
~
a.
1200
c
o
.-
+'
cI)
, ~ 1000
E
o
u
+'
C
Q) ,
"E: 600
8
Q)
1:)
'- 400
X
.Q
1:)
5 200
.c.
a.
;:)
V)
-------
"" 0100houf.
1~~
..~
~ C",,, 1-
,
i~
~ .S_"'J 1- 5 0
c:
.:i -
[11""
'1
4" -~ 0
.cJ
:>
'",,"
I,. -
.~>
'-1:1"
.~~ 0 030hours
..
- ..
\
- .. -2 0 4)
r.. ~ E
~~ ......
~~
-,
..;:!
It
- -~- 1 0
..."-""'
......t'
,~
- ....
.
I" .. ""'
- ""1-- - - 2 'OOhaur.
~
.2!... iI - -
I
..r
I~t
I t
D 5
O2 : 0'0
J
10
Fi g. 7. 21
4~ _II
I
-"'r'\ \i"]
....
- ""I -
~
~ - -~ -
rl
~ \, -
'-,
""' -
)
~ \.. -
..
~ ~
~ '-
\
~ :-
(
~ ,-
I -
~ -"""
- -
...:;:. 1-
....... ~ .......
....
~ 1-
~~-
I
o
I
100
S~ : ppm
01 00 ho Uf'
50
'0
0030hours
20
4)
E
......
10
21.ODhours
J
200
Variation of 502 emission with 02 concentrati on
in 48in. x 24in. combustor
-------
2400
Limestone feed
! started. CaIS= 1.7
Velocity 4 ftls
2000
E
a.
a. 1 500
Limes tone.feedl
rate increased
Ca/S = 1'6
c
o
+'
o
~
+'
C
Q)
U
c
8
(\11000 .
o
~
Veloc ity
increased I
to 8ft/s +
500
Coal feed rate.
increased I
CalS = 0-9
o
5
10 15 20 25 30
Time: hours from start of Test Series
35
Fig.. 'Z 22. Variations of 502 em i ssion after changes in
operating conditions in 36in Combustor
-126-
-------
1700
Q)
L. 1500
~
+'
o
L.
~ 1400
E
Q)
+'
u.. 1600
o
~ 1300
CO
1000
E
a.
a.
o
10 20 30 40 50 60 70 eo 90 100 110120
Elapsed time: m in
Fig."7. 23. Variation of temperature and 502 emission
during a bed temperature survey in the
36in Combustor
-------
The approach to steady state S02 emission during a long term test on
the 27 in combustor is shown in Fig. 7.19 for operation with -150 ~ 1imestone~
and in Fig, 7,20 for -3175 pm limestone. It will be seen that the S02
concentration had achieved the new levels after approx. 7 hours' operation.
The sensitivity of 802 emission to fluctuations in coal feed rate
at constant rate of air input (depicted by the 02 concentration in the flue
gases) is shown for the 48. in x 24 in combustor in Fig. 7.21. It will be seen
from Fig. 7.22 that~ in the 36 in combustor~ after doubling the throughput
(increasing f1uidising velocity from 4 ftis to 8 ftls and doubling the coal feed
rate), and taking into account the attendant fluctuations in the change~ steady
conditions were achieved within six hours. Comparison of the behaviour shown
in Fig. 7.21 with that in Fig. 7.22 suggests that~ unless there is adequa~e
time for the bed composition to change, fluctuations in the sulphur input rate
are not matched by corresponding changes in the rate of retention, and result
in large fluctuations in S02 emission.
The very rapid response of 802 emission to changes in temperature
(Fig. 7.23) shows that the adverse effect of high temperature is reversib1e~
and is therefore not due to sintering or slag formation on the particles.
7.1.13
Accuracy of the results
The reproducibility of the data depends on errors arising from
determination of the 802
reproducibility of those
the absorption of 802.
The reproducibility of determination of the 802 concentration, assessed from
36 tests in Tasks I~ III, IV and V in which no additive was used, was found to
be approx. * 10%, which is within the generally accepted limits (Boiler
Availability Committee Bulletin No. MC/3l6). In considering the
reproducibility of operating conditions,it has been shown that variations of
Cai8 mol ratio, bed temperature, fluidising velocity, bed height and additive
particle size have a smaller effect when the percentage reduction in 802
emission is high (i.e. when the emission is low), so that it can be assumed that
the reproducibility of S02 absorption will be better at high percentage reductions.
Tbe reproducibility of 802 emission during 'tests in which additive was used~
has been assessed from 21 tests (10 replicate sets) in Tasks I~ II, III and V~
and found to be ! 29%. Taking into account the! 10% analytical error
~)
~)
concentration in tbe gas; and
operating conditions that affect
- 128 -
-------
associated with the emission without additive~ this results in a reproducibility
of percentage reduction of ! 15% at 50% S02 reduction or ~ 5% at 85% reduction
when the datum S02 emission is 2000 ppm. Differences between single results
within these limits may not be significant~ but smaller differences between
sets of results~ e.g. between curves through sets of points, are meaningfuL
7.1.14
The
mathematical model
Application and testing of the mathematical model against plant data
has shown that good predictions are obtained for tests with coarse dolomite
(pressurised combustor, Test Series 2 and 3). This is because with coarse
dolomite the processes of elutriation and attrition, for which rate constants
are uncertain, are not rate controlling.
The agreement obtained for this
case demonstrates the validity of the treatment of sulphur retention in the
mathematical model. For additives of wide size distribution the predicted
sulphur retention is sensitive to the values used for the elutriation rate
constants of the close size fractions. For limestones, and for Limestone 1359
in particular, the predicted
attrition rate constant.
retention is also dependent on the value of the
Published elutriation rate data show considerable scatter, but it can
be seen in Fig. 7024 that in order to achieve good agreement with experimental
results for Limestone 18 without allowing for attrition, it was necessary to
use values at the.extreme minimum of the range of published elutriation rates,
(the rate constants decrease with increasing value of the empirical constant n)o
Fig. 7025 shows that by allowing for attrition of particles in the bed the
results for Limestone 1359 can be predicted when using reasonable e1utriation
rates (n ; 1). Attrition is assumed to be described by the equation
dx
dt
;
- K x
d
where
x is the particle size
-1
Kd is the attrition rate constant, h
t is the residence time in the bed, h
- 129 -
-------
100
~
o
c4
o
.-
+'
C
~ 30
+'
Q)
Q:
90
"-
E6
a.
:]
en
....
o
10
Model predictions
o Plant results ( Test Nos.)
o
1.0
2.0 3'0
CalS mol ratio
4,0
Fig. 7.24
Comparison of predictions for Limestone 18 with.
experimental results of Task I, showing effect
of decreasing elutriation rates (Increasing value
of, n) with no attrition
-130-
-------
100
L.
~
"5. 60
~
en
't5 50
o Mea sured retention (Test Nos.>
01'4
~
o
Kd= 0
o
1-0
2.0 3,0
CalS mol ratio
4'0
Fig. 7.25
Comparison of predictions for Limestone 1359 with
experimental results of Task "Sl, showing effect of
increasing attrition rate constant, Kd (n = 1)
-------
In most of the computations it was assumed that all of the S02 was
released at the bottom of the bed, Lower 802 reductions were predicted
when it was assumed that 802 was released throughout the bed, particularly
at high CajS mol ratios, Figo 7.260 802 release throughout the bed is
thought to be the reason for the reduction being less than 100% in tests
with CajS mol ratios of 5 at a fluidising velocity of 8 ft/s (36 in
combustor, Test Series 6, Table 4.6).
The effects of varying bed depth and throughput (velocity) are shown
in Fig, ],270 The model predicts an improvement in retention on deepening
the bed, mainly due to the increase in solids residence time when the bed
is deepero The residence times for non-e1utriated solids corresponding to
curves C, D and E in Figo 7,27 are: 11 h, 21 hand 30 h respectively,
. .
allowing utilizations of 059, 071 and .78 respectively at a Ca/8 mol ratio
of L10
The effect of increasing throughput is the reverse of that of
increasing bed depth, mainly because of the reduction in particle residence
time, There is an additional effect (relatively small in the case tested,
but which could possibly be large) due to increased elutriation from the
bed at the higher gas velocities involved.
. Reaction rate measurements for a range of temperatures were only
determined for one stone, Limestone 18, Figo 7.280 These showed that
whereas unsu1phated stone increased slightly in reactivity as the
temperature was raised from l4720Fto l652oF, stone of 20% utilization or
more decreased in reactivity in this range. This rate data has been used
in conjunction with other input data corresponding to tests on the 36 inch
combustor,with the results shown in Fig. 7.290 A maximum of retention is
predicted at about l4720F in good agreement with experimental findings
(Section 701,4).
- 132 -
-------
100
90
I
L
::3
.r:.
0.50
::3
CI\
-
o
~40
o
c
o
~30
~
+'
Q)
a:: 20
10
Fig. 7.26
o
....-
,,"
"
/
'"
;'
/
;'
/
"
"
/
/
/
/
/
/
/
/
/
/
/
/
/
/
I
I
I
/
/
/
/
/
/
/
I
/
I
I
/
I
'/
'/
'I
'I
OTest 2.3(3'5atm)
. Test 2.4( 5atm)
502all formed at
. bottom of bed
- - - - 502 formed throu
out the- bC!d
Plant results:
Model:
o
0'5
2.0
1'0 1 '5
Co IS mol ratio
Comparison of predictions for Dolomite 1337 with
experimental results of Task II showing the
effect of 502 release through the bed
"--' . ,.~:;
-------
100
90
c:
'- 60
(...
:J
.r:.
E-50
:J
en
....
040
o
0--
c:
.~ 30
.....
c:
Q)
.....
Q)20
0::
10
o
Fig. 7.27,
1
A 2ft deep bed: 4ft/s
B 4 " ... It:" II
C .2" " ".:8"
D 4 " " ": II "
E 6" " "=",,
2 3
CalS mol ratio
Pred i ct ed effect of opera tin 9 cond itions
for Li meston e 18 (Kd = .025, n = 3)
-134-
-------
0.08
~
.; 0 '05
c:
a
+'
~ 0'04 ~o/o
o ~
u
~ 0'03
c
L.
c:
o 0.02
o
1300
1350
1400
1450 1500
Temperature: of
1550
1600
1650
Fi g. 7. 28. Variation 01 reaction rate constant with
temperature and °/outilisation of limestone:
Limestone 18 (-2060 +1000p.m)
,~i CC."
-------
100
80
1652eF
c
o
u
c
.- 60
L
:J
~
a.
-
:J
en
1292 eF
1It-
040
~
o
c.
o
+'
C
(J)
Q) 20
0::
o
1
2
ColS mol ratio
3
4
Fig. 7. 29. Effect of temperature on sui phur
retention predicted by mathematical
model: Limestone 18 in '36in Combustor
-13 R-
-------
,_2
Emission of nitrogen oxides
Since the i.nfra-'red absorption NO:x analysers o'.cae'red for thl S w.Tck
were not: delhreT€d in time to be used~ it was not possible to foHow the
'\.'"adatL-:-n in NO emhsion with changing opexating con~ii t.i.:ms v
.X
determinations of the NO concentration in the off-.gas we.re made using t.he two
x
methods described in Section 4.1.2 and Appendix 9, namely the modified
Salt.zman chemical analysis and the BCURA NO Box. In the data p'resented in
:x
Appendices l~ 2 and 3 the method of determination used is spedfieds, but the
ag:reeme.nt between the two methods was so close that the mean .va1ues only are
H::'we.ve:r.
given in the tables of Section 4.
About 80% 0f the NO
x
combusto:cs were within the range 300 to 600 ppm vivo
determinations made in the 27 in and 36 in
Attempts were made to
correlate the variations within this range with the plant. operating conditions.,
but the:re was no correlation with fluidising velocity 7 bed temperature:) .CaIS
mol ratio or 502 emission. However~ in the 36 in combustor there was a
tendancy for the NOx concentration to increase as the S02 emission was
reduc.ed in some of the test series (Tables 4.4 and 4.6). Also some result.s
(36 in c.ombustor, Test Series 4" Table 4.6) indicate that the NO emission
x
may be reduced by about 100 ppm when the coal size' is reduced from -3175 to
-1680 11m.
The NO .emissions from the 48 in x 24 in pressurised combu.stor were
x
significant.ly lower than those from the atmospheric pressure combustors
(40 t() 190 ppm~ Table 4.9) but no reason can be confidently assigned for thiso
The hi.ghest. NOx concentrations were obtained at the lowest S02 emission levels.,
but this may be due to chance.
- 137 -
-------
,I
7.3
Alkali and Chloride Emission
The presence of volatilised alkali in flue gases is believed to be
important in the formation of alkali-rich deposits on superheater tubes in
o
conventional boilers, although the mechanism of deposition is not clearly understood.
Measurements of the concentrations of volatilised sodium and potassium
in the flue gases were made in tests in the 48 in x 24 in cambustor and the 12 in
combustor using electro-static precipitatore as described in Appendix IV and by
Ounstedllo A small cyclone, normally used in this method to remove very fine
dust particles before the sampled gases are passed through the electro-static
precipitators, was omitted in ,the 12 in combustor tests when it was found that
no dust was being collected. It is assumed that the volatilised alkalis are
present as sub-micron aerosols.
Chloride in the off-gases was determined in
tests in the 36 in, 48 in x 24in and 12 in combustors.
In, the pressure combustor, the volatilised Na content of the gases
was 1 to 2 ppm by weight, representing 1 to 2% of the Na input when burning'
Pittsburgh coal, and less than 1% of the input when burning Welbeck coal which
has a higher Na content. The volatilised K content of the gases was less than
0.5 ppm by weight representing 0.2 to 0.3% of the K input. When limestone was
used as the additive, the Na and K contents of the gases were 5 ppm and 1.5 ppm
by weight, representing 3% and l~% of the inputs respectively.
Alkali emissions from the 12 in combustor, both with and without
limestone addition, were generally higher - 4 to 6 ppm by weight of Na and 1 to
3 ppm by weight of K, representing about 5% and 1%, respectively, of the Na and
K input. These slightly higher emissions could be due to the slightly higher
temperatures (1560oF compared with about 14700F in the pressure combustor) or it
could be that dolomite (used for most of the tests in the pressure combustor,
but not in the 12 in combustor) inhibits the emission of alkalis.
Emissions from conventional plant have not been systematically
investigated, but Na emissions of 10 to 40% of the input have been reported
'11
(Ounsted ).
The considerably lower emissions obtained with f1uidised
combustion are due to the lower operating temperatures.
In the 12 in combustor the Na. and K contents of the dust collected by
the gas-cleaning cyclone and of the dust escaping the cyclone were similar to
those in the coal ash (i.e. no enrichment dccurred). The same was true of the
dust collected by the two-stage gas cleaning system in the pressurised
combustor but the dust escaping the second-stage cyclone showed higher contents
of both Na and K (typically contents increased from 0.5% to 1% and fram 0071.
- 138 -
-------
to 106% respectively) compared with the coal ash. Earlier work (BetheU12J
showed that the Na can be concentrated slightly in the finer: fractions of the
asho
The chloride content of the flue gases was typically 60 ppm by weight
when burning Pittsburgh coal, representing (as expected) about 70% of the input
chlorine, and 500 ppm when burning We1beck coal (which has a higher chlorine
content), representing more than 90% of the input chlorine. Measured values
of ~he chloride content of the gases showed a wide variation C! 50%) but this
reflects the difficulties inherent in the determination, and no correlation
with operating variables eQuId be found.
7.4
Combustion Performance
The combustion performance of a fluidised bed combustor is known to
be affected by such operating parameters as excess air level, fluidising
velocity and fines recycle. In order to siinulate conditions in a boiler it
was decided that tests would be carried out as far as possible with oxygen
concentration in the off-gas in the range 2.5 to 3.5%. If combustion of the
feed coal were complete, this would correspond to an excess air level of 12 to
18%. With operating conditions at which combustion was incomplete, e.g. high
v~locities without fines recycle, it was necessary to. increase the coal feed
rate in order to maintain ~ombuStibnrate~ This higher coal feed rate caused
a reduction in the" excess air level and in some cases substoichiometric feed
rates were reached.
The excess air levels have been calculated from the mass
balances given in the Appendices, and typical results are given in Table 7.10.
For Pittsburgh coal in the 36 in and 6 in combustors (Tasks I and V)
without fines recycle, the carbon loss at fluidising velocities in the range
2 - 4 ft/s was about 8%. The carbon loss was not affected by bed temperature,
but more combustion occurred in the freeboard at lower bed temperature.
Increasing the fluidising velocity to 8 ftls caused a significant fall in
combustion efficiency, resulting in carbon losses in the range 13 to 19%.
At 2 ft/s in the 36 in combustor, the carbon loss was reduced to 1% by total
fines recycle from a cyclone with a cut size of 10 IJm.
Carbon losses from the pressurised combustor (Task II) were in the
range 1 - 3%, the lower value. being most'frequent1y obtained with the high rank
coal. Very little combustion occurred in the freeboard of this combustor~
most probably as a consequence of the configuration of the tubes at the top of
the bed which limited splashing.
- 139 -
-------
Table 7.10
Eff-ect of. operatiug condition. on the
loss of unburnt carbon
Coal PittsbuTgh
Task .r II
Test 4.5 4.6 1.4 1.1 5.1 2.1 5.4 2.2 4.1
F1uidising velocity, ftls 2 2 4 4 4 8 8 2 2
o 1410
Bed temperature, F 1410 1380 1470 1560 1560 1560 1460 1460
Fines recycle No Yes No No No No No Yes Yes
Pressure, atmospheres 1 1 1 1 1 1 1 3.5 5
Excess air, % 7 16 14 15 9 3 -4 22 18
% Unburnt carbon 8 1 8 8 8 13 19 3 1
Coal Pittsburgh
Task III IV V
Test 2.3 2.12 2.4 3.1 3.2 3.4 2 3.2 4.2
F1uidising velocity, ftls 8 8 8' 6 11 11 3 3 3
. Q 1560 1560 1560 1525
Bed temperature, F. , 1570 1550 1560 1560 1470
Internal fine. recycle Te. Yel Yes Yes Yea. Yes Yes Yes No
External fin.. recycle Yes Yes Yes Yes Yes No - - -
Excess air, % -4 1 4 -5 9 1 5 12 4
% Unburnt carbon 16 12 9 11 9 15 .7 1 9
Coal Welb~ck Illinois Park
Hill
Task t II V V V
.
Test 3.6 3.3 1.3 2.1 2.8 2.4 5.2 3.1 3.3 2.10
Fluidising velocity, ft/s 4 8 2 2 2 :3 :3 :3 3 3
o 1560 1560 1460 1470 1470 1470 1470 1290 1470 1470
Bed temperature, F
Fines recycle No No Yes Yes No. No No No No No
Pressure, atm 1 1 3.5 1 1 1 1 1 1 1
Excess air, % 10 10 22 15 10 10 2 1 10 4
% Unburnt carbon 6 9 2 3 5 8 12 1 5 9
- 140 -
-------
',-
Tn Test 2 of the 12 in combustor (Task IV) the recycle system was nomInally in
operation, but it is thought that the good agreement with those tests witboUL
r-e-:.y,= Ie Ln the 36 .in and 6 in combustors indicates that the recycle sysceIDwas
maltunctloning. The carbon loss of 1% with fines recycle in Test 3,2 in the
12 in combustor is in good agreement with tests in the 36 in and the "8 in x
24 in combustorso
:Even with internal and external fines recycle in the 27 in c.ombust'Jr.
the carbon loss was of the same order ..as in the other combustors without
'['ecyde, The poor combustion efficiency is thought to be due to quenching of
combustion in the freeboard by the water cooled walls"
From combustion efficiency considerations the losses of combustible
gases La all but the 27 in combustor were negligible, comprising less than
200 ppm of CO in the off-gas. However in the 27 in combustor there was.about
006% of COin the off gas, equivalent in heat loss to about 3% unburnt carbon..
Typical car:bon losses for the three other coals are given in Table 7.100
These results indicate that there is little difference between the combustion
performance of the four coals used,
705
EmissIon of Particulates
Particulate matter emitted from f1uidised bed combustors consists of
unburnt carbon, ash and a~ditive particles. For a combustor operated without
fin~s recycle the carbon elutriated from the bed represents 5 to 15% of the
teed coal, and 80 to 100% of the ash and additive is also elutriatedo Almost
all of these fines. can be collected by efficient cycloneso
In the 36 in and
12 in combustors the primary and secondary cyclones were 90% efficient at
about 10 um, and particulate emissions were in the range 001 to 0.6 gr/scf,
representing I to 5% of the non combustible material in the feed. In the
pressurlsed combustor, which had an extra cyclone, the dust loading at the
turbine cascade was only 000.5 to 0.1 gr/scfo Particulate emission from the
27 in combustor was an order of magnitude higher because the cyclone efficiency
wa~ lower,
I'he emission increased approximately in proportion to the feed rate of
ash plus additive. There was little effect of fluidising velocity; although
th~re was increased e1utriation at higher fluidising velocitie~ the cyclones
operated more efficiently and there was thus no increase in emission:
- lAl -
-------
In order to achieve a satisfactory combustion efficiency it is
necessary to burn ~he elutriated carbon fines, either by recycle to the main bed
or by separate firing to a burn-up cell. Both of these methods~ but particularly
fines recycle~ will result in an increase in particulate emissions since the
ash and additive fines have to be re-collected. In the 36 in combustor the
particulate emission increased to 1.4 griscf when the primary cyclone fines
were recycled to the bed.
Based on these results it is unlikely that there would be any problem
in meeting projected statutory limitations on particulate emission.
7.6
Corrosion and deposition
7.6.1.
Deposits
Under normal operating conditions, the amount of material adhering
to the tube surfaces was negligible. The slight accumulations which occurred
on tubes near coal nozzles or on tubes above the bed generally had the same
composition as 'the ash, although in some isolated instances there was a
slight enrichment in alkalis or sulphur.
This fredom from deposition as compared with conventionally-fired
boilers is thought to be due to:-
(a) the lower vapour pressures of the alkali salts at the lower
temperatures in the fluid bed and consequently the reduced
tendency for volatilisation of the alkalis. This is in
contrast with conventional pulverised 90al-fired boilers
o
where flame temperatures are about 2900 F and where the
fireside deposits that form are frequently enriched with
sodium, potassium and sulphur;
(b) the 'scouring' action of the fluidised bed.
The former is probably the more important factor.
7.6.2.
Cgrrosion
7.6.2.1.
Effect of time
Figure 7.30 shows the average specimen weight loss of a typical low
alloy steel (21% Cr) and a typical austenic steel (AISI 316) as a function of
time for tests with Pittsburgh coal. Results for 127. Cr steel were similar to
2
On each graph a line indicates the 30 pgicm h rate
those for the austenitics.
of metal was'tage which is assumed to be tolerable.
- 142 -
-------
. 200
C'I
E
u
-
0')8
E
U)
U)
o
+J 6
.c
0')
'-
~
30jJgI cm:! h
/
300
A
100
A
11.9.0] - .-"'-
.",""
.",
-
.",""
,--
,--
"
"..
"..
"
'"
/
/
""
/'
~ OOO~
/ - -_9.>()---- ---- ---0
/ 0- - - -=--=-- - ~ 0<- - -
--. --- x
2'~ X Chrome steel
A
o
- 7. o--F X
Austenitic Type 316 steel
o
4
/
"
10' f", /
\'l." '/
./
./
./
/'
./
/'
/'
/'
./
/'
""
/'
//' 1100'F- - d
,/ C --
--
/ 8 ----
/' --
/ ----~
c/ --.- -- -- & - ~Qg:.f - ~ 0
/d.", _-0- - ---
,/,--'- ---- 0
-:::: 0- -
o
2
o
200
400
600
Ti me : h
800
1000
C Metal temp. circa 1270°F
A " " ,, 1100°F
o " " ,, 900°F
x " " ,, 750°F
~.lJf c
15
10
5
c
I')
I
o
or-
en
en
O'
U)
U)
0'4 Q)
c
~
u
.-
.c
+J
0'3 ~
Q)
-
o
>
:J
0'2 Jf
0'1
-------
It will.be seen that the rate of metal loss is non-linea'!' with
respect to time.
This is somewhat academic for the 127. Cr and austeniti~
steels where metal loss rates are extremely low~ but is important for the
low alloy steels.
. There can be little doubt that these steels would have
an acceptable rate loss over the life time of a boiler (e.g. 100~000 hours)
at temperatures up to at least 9000]'. It is also probable that .the rate
loss would be acceptable for 2i% Cr steel at near its normal maximum
. a
operating temperature of 1090 F~ although tests of longer duration (i.e.
longer than 1000 h) would be needed to prove this point.
7.6.2.2.
Effect of different coals
rn addition to Pittsburgh.coal~ Welbeck (0.5% Cl content) was burned
in Test 1 of the 48 in x 24 in pressurised combustor. Comparisons with Tests 3~
4 and 5 show little difference in weight loss rates for the two coals.
i,n line with ea.rlier evidence obtained on the 12 in combustor suggesting
nature of the coal does not have a major effect. on corrosion in fluidised
combustion.
This is
that the
7.6.2.3.
Effect of fluidising velocity
It was mentioned in the introduction that the effect' of erosion by the
fluid bed ash particles was not known. Erosion might prevent deposit formation
and so reduce corrosion or it might accelerate attack by preventing the formation
of a protective oxide film. In the work reported herej wastage could in no case
be attributed directly to erosion.
It is possible to compare the weight losses after 500 h tests in 12 in
and 27 in combustors in which the fluidizing velocities were 3 and 8 ftls
respectively. The results show some scatter but do not indicate any marked
effect of fluidising velocity.
7.6.2.4.
Effect of operating pressu're
A comparison of the results from 100 h tests in the 48 in x 24 in combustor
(at 3.5 to 5 atm pressure) and 12 in combustor (at atmospheric pressure) taken
together with earlier evidence shows that weight loss measurements are in
reasonable agreement~ as are the results of metallographic examinations. It
is concluded that pressure does not affect the corrosion of tubes immersed in
the bed.
- 144 -
-------
TABLE 7 0 11
Com arison of Wei ht los-gData.fr0r4..500..hTests-
With and without L~e~toneAddit~on 12~.C&.mbUstOT)
--.-..,--~~ . - 2 .,.-..--
Nominal Wt ..1o~s-.iate, j)glcm h
Alloy Tube Wall Position
Temp.- W~tftou t With
.JF . t:bties-tone. . . LUnestone
Medium carbon 750 In Bed 48 41
steel ]50 ABove Bed 25 18
930 In oed .88 84
. .
21J4% 750 In bed 32 42
C-r ~ 1% Mo 750 ABove Bed 25 18
930 In Bed 78. 98
12% Cr 1320 In bed 3 ']
1320 Aoove. Bed 1. ..'" 2
. - .
AISI 316 1320 In Bed 4 6
1320 Above ced 1 2
1320 In oed 3 6
AlSI 347 1320 ABove Bed 1 2
1560 In Bed .7.. 3
. . .
1320 In Bed 6 13
Esshete 1250 1320 ABove Bed 5 6
1560 In Bed 7 3
1320 In Bed :3 5
PE 16 1320 ABove Bed 1 2
1560 In Bed 3 2
- 145 -
-------
7.6.2.5.
Effect of additive
The effect of limestone addition on metal wastage in tbe 12 in
combustor is shown in Table 7.1ltwhere the average spe~imen weight losses in
the tw6 500 h tests" with and without limestone addition are given. The
overall picture confirms that there was little difference in metal wastage
between the two tests.
Microscopic examination of specimens from the two 500 b tests
indicates that there was more attack of specimens which had been tested with
limestone addition. Particularly with the high chromium steels at 11000F,
l3200F and l5600F somewhat more sulphide penetration (up to 40~m) had
occurred. It is possible that limestone feed to the bed resulted in more
sulphide penetration of the alloys, because of the higher sulphur potential in
the bed. However, sulphide attack, of the same amount as reported here, was
also found in earlier tests with Newstead coal and no limestone addition.
Limestone or dolomite addition to the pressurised combustor produced
similar results, the weight losses showing little change and the meta110graphic
examination showing a slight increas"e in sulphur penetration.
7.6.2.6.
Comparison of 'above bed' and 'in bed' results
Tubes above the bed are in less intimate contact with ash and carbon
particles. and as might be expected, weight losses for specimens placed above
the bed were substantially lower than those situated in the bed.
- 146 -
-------
8.
ACKNOWLEDGEMENTS
The NCB Contract Manager wasD.H. Broadbent, assisted by S.J. Wright.
The research programme was directed by A.D. Daiuton (CRE) and H.R. Hoy (BCURA).
- - . -
The proj ect co-ordinator was..D.J'. ..Loveridge.. The pilot. plant experimental
.. . .. . .
work at.CRE.wasadministered by J. McLaren. The following personnel were
involved in the Tasks into which the programme was divided. .
Task I
Task II
Task III
Task IV
Task V
Task VI
Task VII
Task VIII
The project leaders were D.C. Davidson and D.F. Williams:
." .'
A.A. Randell was responsible for operation of the plant:
D.G. Cox and ~. Highley carried out the data processing
and assessment of results.
The project leader was A.G. Roberts: D.M. Wilkins was
responsible for operation of the plant and J.E. Stantan
carried out the data processing and assessment of results.
The project leader was.D.J. Loveridge: M.H. Barker was
responsible for operation of the plant and D.M. Wilkins
carried out the final data processing and assessment of
results.
The project leader was M.J. Cooke:
B.J. Bowles was
responsibl~ for operation of the plant; .E.A. Rogers carried
out the corrosion studies.
The project leader was D.C. Davidson: A.W. Smale was
responsible for operation of the plant; D.G. Cox, J. Highley
and J. Holder carried out the data processing and assessment
of results.
The project leader was D.W. Gill: . he was assisted by
F.V...Bethell and B.B. Morgan.
The project leader was D.C. Davidson:
was carried out by R.F., Littlejohn.
the experimental work
The. project leader was D.H.T. Spencer: he was assisted by
A.A. Herod and B.A. Napier. Additional work to obtain data
for the mathematical model was carried out by F.V. Bethell
and G. McDonald.
Work on emission of NO was carried out by J.T. Shaw.
x
- 1/,7 -
-------
The monthly and quarterly progress reports were prepared by
D.C. Davidson, and D.J. Loveridge. Assessment of the experimental
results for,.the Main Report.wascarriedo.u.t.by J.E. Stant an, J. Highley
. .
and A~G. Robe~ts. . The Appendices to the Main Report were edited by
J. Highley and W.K. Joy.
The OAP representative in, the U.K. wasE.L. Carls, and his
. .
valuable contribution both in theexperimen.tal work and in the prepara-
tion of the progress reports and the final report. is. acknowledged.
"
- 148 -
-------
9.
. 8.
REFERENCES
L
Skinner, D.. G., 'The fluidised. cembustion of coal.
A review of the
findings of research up to the beginning of 1969', National Coal
Board, London, 1970.
2.
U.K. National Coal Board proposal to U.S. National Air Pollution
Control, Administration for 'Research on reducing emission of
sulphur and nitrogen oxides using £'luidised combustion of coal',
May 1970.
3.
Raask, E., Wear, Q, 301-315, 1969.
4.
. Modern .Power. Station Practise, Vol. 2, U.K. ..Central Electricity
Generating Board, London, 1963-64.
5.
Pope, Evans and Robbins, reports to NAPCA. on development work on
fluidised bed combustion, Nos... 42, 43, 47-54.
6.
Potter, A.E., Amer. Ceramic Soc. Bull., 48 (9), 855.-858, 1969.
7.
Falkenberry, H.L. and Slack, A.V., Chem. Eng. Prog., 65 (12),
61-66, 1969.
Jonke, A.A., 'Reduction of.. Atmospheric Pollution by the Application
of Fluidised Bed Combustion~' Argonne National. Laboratory Monthly
Progress Report to NAPCA, No.3, October 1969~
9.
Zielke, C.W., Lebowitz, H.E., Struck, R.T. and Gorin, E., Jnl. Air
Pollution Control Association, 20 (3), 164-169, 1970.
Esso Research Engineering Co., British. Fat.ent..l, 183, 937, 1970.
10.
11.
Ounsted, D., J. lnst. Fuel, 31, 474-499, 1950.
Bethell, F.V., BCURA report to NCB, Doc.No. FCP 4, 1969.
12.
- 149 -
-------
10.
Glossary
Cats mol ratio:
Sulphur retention:
. Mola of calcium in additive.
Mols of sulphur in coal
. Sulphur in coal. -. Sulphur in gas
.Sulphur in coal
x 100%
S02 reduction:
S02 emission without additive - S02 emission with additive
S02 emission without additive
Absorption ratio:
Utilisation of additive:
S02 reduction
100 - S02 reduction
Mols of aulphurretained br additive
Mols of calcium in add1tive
x 100%
Volum.~floW.Tateof &s atcombuation tem erature and
Fluidising velocity:
CroBB Bect10n of-cambuBtcirneg1ect1ng tubes
Excess air:
Unburnt carbon loss:
Air,iuput~. Stoichiometric air for coal input
Stoichiometric: air: for coal input
Unburnt solid carbon
Carbon in coal input
x 100%
- 150 -
x 100%
ressure
x 100%
------- |