U.S. DEPARTMENT OF COMMERCE
National Technical Iflfwafcrtion Service
PB-271 128
Cbsap-Eers- 5-8
Final Report for Low
Pressure Tests of the
CPU-400 Pilot Plant
Combustion Power Co, Inc, Menlo Park, Calif
Prepared fir
Industrial Environmental Research Lab, Cincinnati, Ohio
Sep 77
-------
SECTION V
INTEGRATED SYSTEM TESTS
As defined in ecu.tract tasks LP-1, L?-8 and LP-11, three integrated low
pressure test se> ies were conducted to demonstrate the CPU-400 pilot
pi mt process controllability and combustion efficiency using manual
control, semiautomatic control, and automatic computerized control meth-
ods. Task LP-1 defined a two-hour test using the horizontal comLubtor,
inertial separator, and solid waste feed system IJPV< jped under a pre-
vious contract. This test was conducted for the primary purposes of
evaluating the combustion process controllability and characteristic
of the fluidized bed and solid waste feed material. Task LP-8 defined
a 24-hour test using the newly developed vertica' combustor, aluminum
removal vessel, and the intertial separator and the modified solid waste
feed system. This test was conducted for the purposes of evaluating
the combustion process controllability and efficiency, he :haracteris-
tics of the fluidized bod material, and for determination of various
combustor off-gas constituents. Task LP-11 defined a 48-hour test using
the vertical combustor, aluminum removal vessel, two stages of inertial
separators and the modified solid waste handling system. The testing
performed during th^s test series essentially evaluated the same areas
that were tested during the LP-8 testing, with the additional require-
5-1
-------
ment that the entire process be conducted under automatic (computerized)
control. All hardware and disciplines necessary for high pressure tur-
bine testing were evaluated during the performance of these tests.
TASK LP-1 TESTING (16-SEMES TESTS)
Test Summary
Fluidized operation started at elapsed time ET=458.5 minutes. In the
subsequent 340.5 minute period, numerous system adjustments and modifi-
cations were accomplished in preparation for the performance test inter-
val. Included were two periods of blower shutdown tot-sling 87.5 minutes
during which modifications were made to the twin pipe sjlid waste feed
system. These alterations eliminated plugging problen., at the airlock
feeder valve (splitter at inlet section, and pocket discharge section).
Also included was a period of bed makeup sand feeding (3500 pounds).
The continuous 2-hour pe.-iod for performance analyses was initiated at
ET=799 minutes and terminated at ET=919 minutes. The test was termina-
ted at ET=941 minutes. Total consumption of solid waste was 11,972
pounds, a new record for a one day operation.
Test Objectives
The primary objective of the test series was to demonstrate the contol-
lability of the hot gas system exhaust temperature produced by solid
waste combustion. Specific test objectives are cited below:
1. Test duration: 2 hours continuous with no auxiliary fuel
2. Exhaust temperature: 1450° F ± 20° F
3. Combustion efficiency: 90 percent (minimum)
5-2
-------
4. Superficial velocity: 7 feet per second
5. Measurement of weight and rate of solid waste feed
6. Measurement of calorific content of solid waste used
7. Measurement of moisture content of the solid waste
8. Measurement of correction factor for heat loss r"rom the fluid bed
Test Setup and Hardware Description
The LP-1 test setup, shown schematically in Figure 5-1, consisted of the
horizontal comb'jstor, the multi-tube (6-inch diameter) inertial separa-
tor, and the solid waste handling system.
Hot Gas System-
The combustor system included a pneumatic (air) solid waste injection
system, an auxiliary fuel (diesel oil) injection system, and a bark
flow bed heatup system. The gas from the fluid bed exhausted into the
inertial particle separator. Downstream from the temperature sensing
point, the gas was passed through a water spray section and then ex-
hausted to the atmosphere. Details of the horizontal combustor and the
other components which comprise the hot gas system have been previously
described in Section III.
Control Systetn-
The manual operation of the system was augmented with the analog control
system shown in Figure 5-2. During this test, the test conductor re-
tained an on-time option for direct control of the "olid waste feed
rates and bypass valve positions. The temperature of the cleaned gas
leaving the Inertial particle separator was the primary temperature to
5-3
-------
tn
i
NO.
1 BURNER
OIL LINE
SAND
HEATING
EXHAUST
PIPE
ce
EXHAUST
/
SOLID WASTE
FEED CONVEYOR
16 IN. AIRLOCK
FEEDER
VALVE
t
j
D
AIR SYSTEM
FOR OIL GUNS
NTRIC VESSEL ^\
1
FLUID BED COMBUSTOR
9 ft O.D. x 11 ft
MAXIMUM
WORKING PRESSURE
65 psi
V
1
1
-— 1
feic-4 NO.l AIR BLOWER
j_J L._, . . 1 7C UD
"~~ 4000-7000 scfm
COOLING
WATER
1st STAGE
INERTIAL
SEPARATOR
ASSEMBLY
FEEDER AIR
CONTROL VALVE
•SAND HEATING
EXHAUST PIPE
COOLING
WATER
BURNER AIR'
CONTROL VALVE
NO. 1 BURNER.
AIR LINE
NO. 1 BURNER
EXCESS AIR
COOLING AIR CONTROL
VALVE FOR SAND HEATING
PREHEAT
AIR BLOWER
10 HP
1000-1700 scfm
0 1.1 psi
WATER
AIR
NO. 2 BLOWER
SHUT-OFF VALVE
STACK
—TO
BAGHOUSE
-221-39
Figure 5-1 LP-1 Test Setup
-------
ATLAS STORAGE
AND FEED
SYSTEM
I
v-f)
BED TEMPERATURE -/
ATLAS CONTROL
SYSTEM
CONTROLLER
2
©~
SIGNAL 1
CONDITIONER
& TRANSMITTER
Q;5 Q^
COMBLISTOR f SEPARATOR / 1
1— OUTLET DilTI FT — ' L
TEMPERATURE TEMPERATURE
SECONDARY
AIP CONTROL
VALVING
4 \
/
J \
CONT .OLLER CONTROLLER
3 8
t t
v^y^ "~^v^x
Slu'.-- 2 SIGNAL 3
*• CONHITIONLR CONDITIONER
t :FJ;-.S»ITT:R & TRANSMITTER
r/c = THERMOCOUPLE
R = RECORDER T.PUT
221-230
Figure 5-2 Control System, LP-1 Test
-------
be controlled. The basic control technique utilized for this test was
the maintenance of tht ,j bed at a fixed temperature by modulating
the fuel flow and controlling the hot gas system outlet temperature by
controlling bypass air flow. Incorporated into the bypass air modula-
tion was a control loop which anticipated changes to the hot gas system
output temperature due to combustor outlet temperature variations.
Solid Waste Handling System-
The solid waste handling system (Figure 5-3) was developed under a pre-
vious contract and is described in detail in Appendix A. It consisted
of two major subsystems: the shredder/air classifier/pneumatic transfer
subsystem (Figure 5-4); and the Atlas storage/combustor feed subsystem
(Figure 5-5). The shredder/air classifier equipment consisted of the 75
horsepower Eidal Mini-mi 11 shredder, used to render a random lot of resi-
dential solid waste suitable for air transport into the fluid bed, and
the CPC-designed close coupled air classifier, which was used to remove
a large portion of the metal and glass from the solid waste.
From the air classifier, the shredded solid waste was transported to the
Atlas storage and reclaim unit. Figure 5-C shows the Installation of
•
the Atlas unit and the conveyors that meter the solid waste to the air-
lock feeder valve as demanded by the control system.
The task of the air-lock feeder valve. Figure 5-6, is to transfer the
soild waste frotf the ambient pressure zone of the conveyor to the pres-
surized solid waste feed pipe. A single 16-inch feeder valve configura-
tion was used to Insure an adequate supply and distribution of solid
5-6
-------
PACKER TRUCK
/' f
STOCK PILE
DUST KOP
CYCLONE.
EIDAL SHREDDER WITH
ANTI-BALLISTIC HOOD
-
WEIGHING CONVEYOR ,
BLOWER- -', •
AIR CLASSIFY ~
hrnM*d SolM Ki«t» Stori^r
'""' !
TRANSFER
CONVEYOR :j
mm IT r f nil
.'.
-y
SOLID WASTE SYSTEfl
CONTROL PANEL
ATI A'", f'!'!i/i VOP
NOTE: FOR LP-8 TESTING -
(a) FEED CHUTE REPLACED BY ; OAD SPLITTER
(b) SINGLE FEEDLR VALVE AND 6-INCH FFED LINES
REPLACED BY TWO VALVF' AND 5-INCH FFFD NNt'i
di [
I'-ANEl
EEED CHUTE/LOAD
SPLITTER
EITHER V*H i\ '^ I
-——.-FFFD ! !'ir
IA-000018
\l'"-l a'rl I P-8)
-------
ANTI-BALLISTIC HOOD
•:«•- - "'.&»
••
CONVEYOR HOPPER
Figure 5-4 Shredder/Air Classification/Pneumatic Transfer Subsystem
Figure 5-5 LP-1 Atlas Storage/Combustor Feed Subsystem
5-8
-------
16-INCH
FEEDER VALVE
Figure 5-6 16-Jnch Air-lock Feeder Valve and Drive Assembly (LP-1)
5-9
-------
waste to the bed. This valve has a volune of 2.6 cubic feet and rotates
at 11 rpm. During low pressure operation, the va^e operated with 2 psi
differencial pressure. Two 5-inch diameter feed Tines were used to car-
r> the processed solid waste feed to the combustor.
Test Results
The bed was fluidized after 7i2 hours of sand preheat time and :he control
conditions were initiated at a total elapsed time of 799 minutes. During
thr period between initial fluidization and start of the 2-hour perfor-
mance test, modifications were made to the feeder valve to circumvent
plugging problems at its internal splitter and lower discharge cavity.
Test Temperature Control -
Analog records of the key system temperatures, (Figure 5-7). show that
the required test temperature of 1450° F ± 20° F was met for the sepa-
rator exhaust during the 2-hour period between 799 and 919 minutes
elapsed time. In addition to the analog temperature recordings, numerous
temperatures were recorded with the digital data acquisition system. Air
flow, solid waste feed rates, and pressure data were recorded and solid
waste samples were taken at 15 minute -intervals. Analysis of this data
showed that the mean of the exhaust temperatu»e differed from the set-
point by 2.7° F and the maximum standard deviation of the exhaust tem-
perature was 6.9° F.
Superficial Velocity-
The superficial velocity through the fluid bed, based on air flow only,
-------
1430 F-jASK LP-1 PFRFORMANCE INTERVALS
<*0
ELAPSED TIME (MIN)
221-231
Figure 5-7 LP-1 Analog Record of Key System Temperatures
-------
was
6 feet per second nominal with variations (depending on jypass air)
f] 1 t1 r\ — C\ £ £ S\f\ +• r> rt ••* r- ,-T. f s*. »-\ i-J
of ->0 1 to -0.6 feet per second
Combustion Efficiency and Heat Loss-
Data and sample analysis results necessary for fuel heating value and
combustion efficiency calculations are summarized in Table 5-1. ne
heat loss rate was based on temperature recordings of the combustor
pressure shell, separator pressure shell and the interconnecting pipe
between the vessels. The lower heating value (column J) was the Parr
bomb calorimeter results corrected for the condensation of water vapor
in the combustion products. The correction was based on a dry, inert-
free combustible fraction based on C-jgH^oO-ig composition. The correction
was made to permit direct comparison (column K) of the Parr bomb Caliber
results to the heating value employed in the combustor heat balance cal-
culations in which the vapor form of all the combustion products was used
Post Test Inspection
The combustor was not disassembled immediately after the LP-1 test. Two
problem evaluation tests were performed and a third was initiated when it
was observed that poor bed fluidization was occurring.
When the combustor was opened for inspection it Became evident that a
clinker had formed on the fluidized bed surface. The amount of bed
material in the bed area itself was reduced to about 50 percent, while
significant amounts of sard and sand-glass agglomeration were found
behind the active area and in the settling chambers. The big clinkers
being formed on the bed surface consisted primarily of sand-glass
5-12
-------
Table 5-1. LP-1 TEST SUMMARY OF FUEL HEATING VALUE RESULTS
Ln
i
Perf.
inter-
i
2
3
4
7
g
A
A v
<, 0 1 : J
i.'dito
* 1 C .. -
i its
lib/
29.9i
51.45
JO. S3
jc . i e
Jl.53
^
27. 2S
29.15
B
i—
o.?n
:.:;;
0.321
C . "i , o
C . 2S6
0.235
= .:io
c
Ii.er;
' r j c 1 1 on
0.068 .
C.C&4
0.124
0.057
C.054
0..157
0.0.-.4
0.061
0
Av .
air
f low-
r ate
(IS/aiin)
299.8
2,8.4
299.3
302.4
302.8
303.4
504.5
306.2
F
Bed
temperature
range
1425 - !t;.5
1435 - iae,:
1453 - 1443
1443 - 14-' 7
1447 - '.443
1443 - >'ifi
Ij 4
4269.2
4250.7
_ .
422^.1
r
4232.7
h
he- j 1 1 ng
(Btu/ C
Cry It)
6619.7
C413.2
6513.1
6574.1
6072.0
1.
ifc/1.6
6694.1
6937. >
'
' J
v;1 uu
f ro^ Colori -
neter, Qr.
7,54.0
63J9 0
6573.1
7U19.0
C,,3.4
6....0
"*
7190.1
6896.0
I ,-
-
'.
6
-
6
-
'•
r~ "*'
. — _-
6090.6 I IOC.9
£3i&.3
6-61.9
63S'vG
Standard deviation
342. £
£.:;•
-------
agglomerations which were brittle enough to be broken by hand; however,
they proved to be firm enough not to be crushed by the bed action at
1300° F and 7 feet per second.
Investigation of the backheat history indicated that the final backheat
was performed at a slightly higher temperature than the other tests
(1780° F average and 1830° F peak) with used sand. The initial tests
were at comparable levels, but these tests had new or low time sand in
the bed. The phenomena that is believed to have occurrul is that since
the sand was well used, it had an increased amount of glass which has a
lower softening point than new sand (about 1500° F to 1550° F versus
2200° F); the level of backheat was enough to cause bed agglomeration
which was sufficient to build up a crust on the bed surface that could
not be broken down by bed action.
The baghouse was inspected and filter bags were found to have failed.
The puffback hand valve was jammed with sand and stalled the puffback
drive. This allowed the bags to load up and fail since they were not
being periodically cleaned.
An inspection of the inertial separator tubes after a total of 735
minutes of fluidized operation revealed that 15 tubes were plugged and
6 tubes were partially plugged.
The ash hopper under the inertial tubes was found xo be almost completely
filled. Apparently the ash transport system was overloaded. This was
not surprising as the horizontal combustor experiences high elutriation
5-14
-------
an
d, therefore, the inertial separator was experiencing a high ash/sand
Conclusions
The ^ollowing conclusions can be drawn from the l>-i test series:
1. The process of combusting solid waste in a fluid bed ;o:nbustor can
De maintained within a controlled temperature band sue!: as 1450 F
to 1470° F. This shoun. allc'* tne _,se of this type of combustor in
a gas turbine system.
9
he performance of the inertial separator system was ^satisfactory
in this test series because of tube plugging. The rerr.on for this
is believed to be insufficient velocities in the tube:.
3. The backheat temperatures for aged sand beds should L^ at more moder-
ate levels (below 1500° F) than were used in tn's series. This
should prevent s;ntering of the ash in the bed during bed neatup.
4. The combustion efficiency inJicdt-u by the coTibustib1,>::s analyzer was
greater than 98 percent. This i-s in agreement with t".e combustion
efficiency from heat balance which also indie ted high efficiency.
5. Testing conducted with the horizontal combustor demons"rated tnot a
combustor with high freeboard volume is necessary to rpduce bed elu-
triation. In addition, the combustor must reduce the tendency of
the feeding air flow to direct sand particles toward the exhaust
port where they are entrained into the exhaust gas stream and car-
ried out of the combustor.
6. An alumina/sand separator should be added to the hot gas subsystem
between the fluid bed combustor and the first stage inertial separator
5-15
-------
to reduce the high solids input to the first ~tage inertial separa-
tor.
7. The solid waste handling system performed moderately well during this
test series, however some feed system jamming occurred during periods
of high volurc demands.
TASK LP-8 TESTING (18-SERIES TESTS)
Test Summary
Fluidized operation started at elapsed time ET=140 minutes, after a sand
heating period of 121 minutes. Oil combustion was utilized to bring the
bed to the desired temperature for solid waste combustion. One stop
occurred during this time because of the loss of the facility air supply.
At ET=223 minutes solid waste combustion was started and continued until
ET=2428 minutes with th.ee interruptions. These interruptions occurred
at ET=310, 629, and 934 minutes and were caused by plugging in the feed
line on two occasions and in the feeder valve outlet on the other occa-
sion. The 24-hour control period commenced at ET-988 minutes and termi-
nated at ET=2428 minutes. At ET=2063 r.inutes the feed line under feeder
valve no. 2 plugged; the last portion of the 24-hour contrc period was
completed on one feed linp.
A period of 24 hours was completed with exhaust temperature controlled
to within 1450° F ± 30° F and 9.75 hours with the temperature controlled
to within 1450° F ± 20° F. Only solid waste was used as a fuel during
the 24-hour control period and automated solid waste feed was used.
During the 34.83 hour solid waste burn time 69,134 pounds were consumed.
5-16
-------
Bed depth increased 1 inch during the test; no bed material was added or
removed for bed maintenance.
Test Objectives
The priniary objective or this test series was to demonstrate the process
controllability and efficiency using semiautomatic control techniques.
Specific test-Objectives are cited below:
1. Test duration: 24 hours continuous
2. [xhaust temperature: 6 hour span. 1450° F 20 F
24 hour span, 1450° I- • 30° F
3. Combustion efficiency: 95 percent (two methods)
4. Residue removal: drain to maintain 35 t 3.5 1W bed AP
5. Fluid bed: maintain weight within * 10 percent (by ^P) <4 percent
elutriation per hour (-30 mesh)
6. Mass accounting of sand and inert material
7. Exhaust gas: determine percentage of the following chemical com-
pounds: CO, C02, CHX, HC1, and NOX
8. Partlculate: determine participate loading of exhaust gas
9. Superficial velocity: 4 to 8 feet per second
10. Post-test bed inspection for clinkers (none >2 inches)
Test Setup and Hardware Description
The LP-8 test setup, shown schematically 1n Figure 5-8, consisted of the
newly developed vertical combustor and alumina separator, the modified in-
ertial separator used during previous tests and the modified solid waste
handling system.
5-17
-------
U1
00
a
L
r^U
n-
t~ - >
*r. ft.
-b
.
-" --.;»-*-
Figure 5-8 LP-8 Test Setup
-------
Hot Gas System-
The vertical combustor svsteni includej a pneumatic solid waste injec-
tion system, an auxiliary fuel irWrinn system, and a hark flow bed
heatup system. The gas from Lne fluid bed was exhausted into the alumina
separator (first stage separator) which de-entrained the larger particles
from the gas stream, and was then routed to the inertial separator (se-
cond stage separator) for final cleanup of the exhaust gas. Downstrean
from the temperature sensing f1nt, at the second stage separator outlet,
the gas was passed through a water spray section and then exhausted to
the atmosphere. Residue from the second stage separator was pneumati-
cally transported to a baghouse filter. The combustor and particle se-
parators are shown 1n Figure 5-9. Details of this equipment and the re-
maining equipment elements that comprise the hot gas system have been
previously described in Section III. The modifications made to the in-
ertial separator since the completion of the LP-1 testing series are
cited below:
1. The original reverse flow cyclone tubes configuration was changed to
the "reduced cone angle" tube configuration shown in Figure 3-17(c).
2. Only 24 of the 43 cyclone tubes were used. See Figure 3-22 for the
tube arrangement..
3. The outlets of the non-active tubes were removed to provide better
flow manifolding to the active tubes and the resultant holes were
plugged with special seal adapters.
4. A baffle plate was welded to the top of the torroidal manifold at
the Inlet area to avoid possible high flow and high particle loading
in tubes adjacent to the inlet connection.
5-19
-------
5-° LP <' Tombustor ..md Particle Separator Configuration
-------
The temperature of fhe cle^mv qas le;iv'r-;i -he .f-onrj >:ta<;e 'pparator was
the .i--;-ar-y tei. i "• at jre to be conti ">1 led. Th-'» ta^ic control technique
utilu-J :~or »^i-. test was the rcnfrpHinq of t>-e HuH bed u-rriDftrdtiire
!\, ioii.j ' ,it.i nq Me ^ic 1 tiow; and conf - o-l in-" '"he- r^-i* qas -.iStem outlet
t-J'iper.'.*.urf? bv cfMncp. nq Oyoass ,iir fl M viro ,•?'1 ^ ^ *"i'-.r ir ; -)f-cond
st.ig-"1 ep.»r K.or:-. Inccrpordt"<-! into '.•••.- hy: .is'. . i r -" --v. i -i' "~n wdc- a
s.;D-1oic rt-nir.n ,inti ci oa f ed tn.-inq-?s •• ":" h> •• -jd-, ->/r,t--!. ..Mujt tempera-
ture duo to co'nbustor outlet tenoer-i' ;• r 4a<~'> -it inns ^ijr"-.q r'ie test,
the t-^st conductor retainer an on-ti"'t o^.-ti"1'1 *'-f '!'-' -''f •: nr-cl of the
solid wa^te feed rates an'1 bvoai:, ,-a1. /P '/':.s'*i Mr;. A -,'ir;ii • ie.j bl.>c»>
diau'rj.-n <^f the control svsf'"ir' 13 Pr"r' -'nt-'il in ^ijure ''i-iO. fl'is di^^rar:
incl jaes onlj tne jrinary active tl<~ tlar-
it». -efer to Section IV for a s:m:rvv nf * ie L?-3 control '.-'.teiii. The
temi".-rature, pressure and flow Censors that werp used to control, monitor
and anjiyze tne pilot plant operation nurinq the 24-nour test period are
shown ;n ri-jijre 5-8 and are describea in Tables 5-2 and 5-3. Fisher
c.,r,*r }]cr settinqs .ire shown in Tjbli- 5-^>.
So lid Waste Handling Systecn-
The LP-8 solid waste handlinq svstem (Fiqure S-3) was modified to im
prove the control of the solid waste feed rate and to reduce jamming ten-
dencies. These chanqes were made to the Atlas storaqe/combustor feed
subsystem (Figure 5-11) and did not affect the shreeder/ air classifier/
-------
r
T
4
~T
J
,._ J
c.-iO I. f'-.'.' Siirpl i t ; (-•: Lui.ti.il LuO:. l'-i
-------
Table 5-2. LP-b TUN.PERATURE SENSORS
T/Ca ]
I.D. Measurement
000
031
oo :
003
004
0.15
005
007
008
009
010
011K
014b
015
Fluid bed 4"
Fluid bed •}"
Huid bed /'
Fluid b~< d 9"
Fluid bed 9"
Fluid bed 9"
Corbustor plenum 9"
Corrbustor plenum 9"
Distributor plate surface
Distributor , iote surface
Fluid bed 0"
Alu.T.ir.ura vessel outlet.
Freeboard 10'
Freeboard 10"
016 CoOuc. for wall just Lelow feed-
line surface
017 Combur>tar doi.« - surface
016 Diesel oil supply - V
019 Aluminum vcs^fl wall - surface
020
021
024
025b
026
037
1st stage ir.^rtijl separator
top dome - surface
1st. stage inertial separator
wall - side surface
1st stage inertial separator
material removal line between
water injection point and
valve 009 r.ur.'oce
1st stage separator cutlet gas-1"
1st stage separator outlet pipt
surface
L. P. compressor outlet (jas - 1"
Use
Sol id waste feed
Control & bed operation
Backheat control
Mix air control
Bed set point adjust-
ment
Heat loss
Over temp
Heat loss
Ash flow
Hix.jd air control
Flow calc.
aOi;nc!ic,ions assnr i,-.tr:l with c.Kh T/C indicate lenqth thai, the T/C
projects into t'ic -libOciciLec Jifa. When "surface" is 'lOU'd, the
T/C is tact: welded ;o the noteu surface.
Control tliennococ-ple';
5-23
-------
Aflll F!,U FL f)W r,f NS
C ". ^
J c I b 0 1
1J1
107
120
" -.,-,,,„„„ " "
1
Solid tv;S!i; co'.ib pie: . \ ~>r'~ S
Compressor cutlet
Bed diff. pressure
,J|L
'.Mori f ic i j 1 vo ; •)< it, c Ic .
1 ! „ i c i .... 1 1 on arid C:'d ri'ii , i
/ ^rri Lor i ,>j
121
Si:. 1; ;•, , t..- . '.
i tor i inj
>: ^» L • i J - b:_: [.'. j u L i u t
K'5
141
147
160
I Anviind separator pross. drop
I
rr . dl i i.c- 1 air T"! ow sens ).•
I Jitf. prc;ss.
Food line 2 air flow Ltnsir
oi ff. pfess
i.,1 id v/:j..t:.- cc. ib . air i 1:,,
Barr neat, turner cAr f
Dicsei f... ' flo.v raL
'. clS t-j CO.: !;.
it-' tUS T.' fl i CO' 1 fi'i
: 1 i nu I .; i i... 'i.
Fl',;,!: z i f ; i >• t";. '-
. : : no t KJ
;-,. c kr,c.j t -:ontro"( 4 i il
-M j -,L-1 fu!_l control .•.
-------
Table b-4. LP-8 FISHER CUN1ROLLLR SETTING
Controlled
parameter
PB £
Reset
Range
Control
action
Rate
Control led
element
Control led
variable
Fi sher <>'U!.rol ler
No. 1
100
10
xl
Direct
-
Vdlve
005
Dump air
No. 2
100
0.1
xl
Reverse
0
Valve
007
Backheat
flow
No. 3
50
0.05
xlO
Reverse
0
Volve
002
Oil flow
No. 4
100
0.5
xl
Reverse
0.7'j
, , i 1 j ' '
control
system
'jol id
waste
feed
No. 5
50
0.1
xl
Di rect
-
.'a i ve
010
f; 1 f eed-
1 i ne air
flow
o. 6
oO
0.1
xl
irect
0
yaivi:
001
2 feea-
i ne air
flow
No. 7a
70
0.1
xlO
Reverse
-
Fisher #3
or »4d
Fi iher
set point
No. 3
15
0.5
A!
Reverse
-
Valve
055 by-
pass
valve
Exnaus t
mi xi ng
ai r
Cascaded with Controller #3 for fuel oil feed control or with
Controller *4 for solid waste feed control.
-------
I
I\J
Figure -^-}] LP-8 Atlas Storage and Combos tor Feed Subsystem
-------
pneumatic transfer subsystem (Figure 5-4). Refer to Appendix A for com-
plete details of the solid waste handling system. Specific changes im-
plemented for the LP-8 test series are itemized below:
1. The Atlas circumcykme inlet, which produced asymmetrical distribu-
tion and localized compacting of fuel, was replaced by a flat, splat-
ter plate which evenly distributed fuel over tne Atlas floor.
2. The efficiency of the Atlas sweeps were improved by adding an extra
bucket to each chain and incorporating digger spikes.
3. The diversion of dust collector reject back into the Atlas was im-
plemented by the addition of a rotary feeder.
4. A "clod buster" was added to the outfeed conveyor to break up com-
posted sol id waste.
5. Load levelers were incorporated on the transfer and weigh conveyors
to provide a constant volume of fuel per foot of conveyor.
6. The speeds cf the transfer and weighing conveyors were synchronized
with the Atlas conveyor sweep speed by a common variable frequency
controller which was controlled by the commanded volume signal sent
to the Atlas control system.
7. The single feeder valve installation was replaced by two CPC-designed
16-inch feeder valves. Each valve had a volumetric capacity of 2.6
cubic fee. per second and fed a separate 5-inch diameter combustor
inlet feed pipe.
8. The steel pant-leg chute which fed the feeder valve during previous
testing was replaced by a load splitter.
5-27
-------
Test Results and Analysis
Exhaust Temperature-
Figure 5-12 shows inertia! separator outlet temperature during the 24-
hour control period. The 24-hour and 6-hour temperature bands of 1450 *
30° F and 1450 ± 20° F respectively, have been superimposed. Control
within the required terrperature limits is apparent from Figure 5-12. A
statistical analysis was performed on the 287 separator outlet tempera-
ture readings from the data acquisition system. This analysis yielded
the following results:
Mean 1456q F
Variance 58.2° F
Standard Deviation 7.6° F
The results snow the mean differing from the nornnal point by 6 F and
the standard deviation indicating a ' 3.; val;e f •*• 22.8° " asr,um-;nq a
a normally distributed population of data.
Examination of the continuous recording of inertial separator outlet
temperature confirms that temperature was continuously controlled within
the specified 1imits.
Combustion Efficiency-
Heat Balance Technique - Combustion efficiency calculated by the heat
balance technique is presented in Table 5-5. These calculations utilize
the laboratory results for solid waste moisture fraction and inert frac-
tion (see columns B and C, Table 5-5). These results indicated an
average computed heating value of 7934.5 Btu/lb of combustibles
5-28
-------
-------
Table 5-5. LP-8 COMBUSTION EFFICIENCY RESULTS, HEAT BALANCE TECHNIQUE
Ferf.
inter-
val
1
2
t
4
5
5
7
8
9
10
11
12
A
Av
solid
waste
flo.:-
rate
(Uv
•ii n)
34.76
:-.f.63
36.81
36.23
34.29
33. E4
32.70
32.68
26. '15
27.92
26.72
IJ
Moisture
fraction
.3213
.3923
3 OS
.3027
.3058
.3£ .'.3
.2SP6
.2937
. 3087 .
.3033
.3351
.2857
C
I.iert
fraction
(Ir.cl.
ash)
.1160
.123J
.0934
.1169
.1404
.ice
.1891
.18.: 1
.133L
.1302
.1263
.1214
0
A;.
air
flow-
r.ite
( Ib'tnin)
312.7
312.7
312.7
312.7
312.7
31?. 7
3!".,
312.7
312.7
257.1
25M
257.1
E
Bed
terperature
ranee
(bF)
1350 - 1357
1357 - 1352
1352 - U51
1350 - 137r,
1376 - 1352
\r^ - 1353
1353 - 1360
1360 - 1330
1330 - 1275
1275 - 1276
1276 - 1275
1275 - 1300
F
Av.A'
ratoi
outUt
1502.6
100 M
150''. 3
15'<2.4
14T. .1
149.;. 5
149'.. 1
1490.1
ISM. 8
15K-.6
lOld.6
G
Heat
loss
rate
(Btu/min)
8068
75(.3
6985
65r)l
5782
590
5756
5842
6061
5211
5485
5.V.8
H
heating
va'ue, Q,
(Btu/ L
Ib coiio.)
7707.2
86P5.5
6379.0
7049.5
737C,.7
"137.?
8J&9.8
8546.3
7893.3
8-87.7
6175.9
71C4.7
7934. 5a
694. 15U
I
Heating
value
fron calori-
neter, Qr
(BtJ/dry^lh)
7220.6
7003.8
7315.4
7152.2
7264.9
'f43.1
tc^e.a
6434.9
6953.9
6606.5
7143.6
7010.8
J
Lower
heating
value, Q
(Btu^b cbfflb)
8069.0
8106.0
7907.0
7952.3
b467.3
8255.0
8^.8
6153.2
7980.2
7731.0
8178.2
7806.7
6086.6"
22H.2S;
K
Comt'USticn
efficiency
100 oc/ot
95.51
107.1
88.26
88.64
87.11
110.7
99.49
104.8
98.91
104.6
?9.97
92.03
98.09-
7.883-
I
CO
o
Stonciurd deviation
-------
while the bomb calorimeter indicates 8086.6 Btu/lb. This yields an av-
erage combustion efficiency of 98.09 percent.
Because of the disparity in the inerts mass accounting, it is believed
that the solid waste sample used in laboratory tests were not represen-
tative of the solid waste consumed. Table 5-6 presents com-
bustion efficiency calculated using an average inerts fraction of 0.183
obtained iron the inerts mass accounting. Moisture fraction was also ad-
justed assuming the missed portion of inerts contained insignificant
moisture. Results of this adjustment in inerts content raise the averag-
calculated heating value to 8298.4 Btu/lb of combustibles. Although
inerts disparity affects the heating value per dry pound, it does not
affect the heating value per combustible pound which makes the bumb
calorimeter test results for the lower heating value (column J, Table
5-6) valid even for erroneous inerts content.
Carbon Combustion Technique- The results from the carbon combustion ef-
ficiency calculations are summarized in Table 5-7. IP this method it is
assumed that unburned carbon in the flue gas will exist in the forms of
CO, HC (hydrocarbons), and C (elemental carbon) only. Therefore, com-
bustion efficiency can be calculated by the following equation:
12 ,.u \ . 12 k / .M \ i 24
IT a (An-n ) + TSTT o \ AH-.J + -^r C
a
5-31
-------
Table 5-6. LP-8 COMBUSTION EFFICIENCY RESULTS, MASi BALANCE TECHNIQUE
Perf.
Inter-
val
1
2
3
4
5
6
7
8
9
10
11
12
A
.'w .
solid
wast?
flcw-
rate
win)
34.76
35.87
36.68
36.61
36.29
34.29
33.84
32.70
32.68
26.35
27.92
28.72
B
hoiiture
fraction
.2369
.3561
.2B10
. 2600
.2906
.3546
.3008
.2987
.2905
.2840
.314S
.2*40
C
Inert
fraction
(incl.
*sh)
.1830
-
-
H
-
M
•
It
•
"
•
0
Av.
flow-
rite
(Ib/mtn)
312.7
312.7
312.7
312.'
312.7
312.7
312.7
312.7
712.7
257.1
257.1
257.1
E
Bed
temperature
range
135G - 1357
1357 - 1352
13^2 - 1350
ncn - '.376
1376 - 1352
1352 - 1353
1^3 - 1360
1360 - 1330
1330 - 1275
1275 - 1276
1276 - 1275
1275 - 1300
F
Av./U.
sep i-
outiet
tpn.i.
150!. b
151".. 6
i'.Ol.l
IM1.3
150'. f,
141). 1
1403.5
149. .1
149'.. 1
1513.8
151.!. 6
1511.6
r,
Heat
loss
rate
(BtJ/m1n)
. 8051
7547
6959
6435
5774
5935
5757
5844
6048
5203
5444
5337
Vjlues reproduced froin Taole 5-5.
cSUnCjrd deviation
H
rir.n-jtC'd
value ,Q_
(Btu/1-
Ib ccwb.)
8249.8
9221.3
75M.7
753P.1
77C-..0
9350.9
H314.4
8466.7
8294.9
3524.8
8680.5
7625.5
3298. 4 b
6C8.681-
1
Heating A
value
from Ciilorl-
meter,
(Btu/drV Ib)
7220.6
7003.8
7315.4
7152.2
7U4.9
C6--.8.1
6626.6
6434.9
6953.9
6806.5
7143.6
7010.8
J
Lo.'er
healing
value, Q.
(Btu/lt> cbni)
3069.0
81 "6.0
7907.0
7C52.3
RJ07. ':
6215.0
&4 J2 . fi
31i3.2
7980.2
7731.0
8178.2
78^6.7
8086. 6b
228. 25L
K
Con>uotion
efficiency
100 rjc/Qi
102.2
11?. 8
SS.97
5-. 79
91...
113.3
93.33
10 3. 3
133.9
110.3
106. 1
97.68
102. 6:
7.340"
OJ
ro
-------
Table 5-7 LP-8 COMBUSTION EFFICIENCY RESULTS,
CARBON COMBUSTION TECHNIQUE
Perfor-
mance
Inter-
val
1
2
3
4
5
6
1
8
9
10
11
12
Wt.
fraction
of CO,
I v\ <-
\ *>/
10.32
9.862
10.32
10.32
10.32
10.32
10.32
10.32
10.32
10.32
10.32
10.92
Wt.
fraction
of CO
(%)
2.607xlO"3
2.607xlO"3
2.607xlO"3
3.910xlO"3
3.910xlO"3
5.223xlO"3
3.910xlO"3
3.910xlO"3
2.607xlO"3
3.910xlO"3
5.223xlO"3
7.830xlO"3
Wt.
fraction
of CH
u) x
.7034xlO"3
1.045xlO~3
1.396xlO~3
1.748xlO"3
1.541xlO"3
:.396xlO"3
1.748xlO"3
1.541xlO"3
1.748xlO~3
1.748xlO"3
1.541xlO"3
1.748xlO"3
•
Wt.
ratio
of C
42.90xlO"6
42.40xlO"6
45.91xlO"6
42.90xlO"6
42.90xlO'6
41.40xlO"6
43.40xlO"6
43.40xlO"6
40.90xl-"8
55.29xlO"6
52.34xlO"6
55.88xlO"6
Combustion
efficiency
(.)
99.80
99 . 79
99.77
99. 76
99.77
99.76
99.76
99.76
99.78
99.72
99.72
99.69
a
99.76
b
0.0317
Mean
Standard deviation
5-33
-------
where: a weight concentration of CO? in flue gas
b weight concentration of CO in flue gas
c weight concentration of HC in flue gas
d weight ratio of C to flue gas
The C02» CO, and CHX measurements (columns A, B, and C in Table 5-7)
were obtained from the online gas analysis equipment. The free carbon
(column D ir> Table 5-7) was based on laboratory analysis of samples from
the first stage separator and baghouse. One sample was obtained from
the aluminum separator after the test while twelve samples were obtained
of the baghouse flyash during the test. As indicated in Table 5-7, this
method indicates high combustion efficiency (column E in Table 5-7) with
an average value of 99.76 percent.
Both this method and the heat balance technique indicate high combustion
efficiency. It can therefore be stated that the system has demonstrated
combustion efficiency in excess of the 95 percent required in the test
plan.
Material Accounting-
A mass balance was conducted on the inert portion of the solid waste fuel.
The material caught, in the vessels and exhausted through the stacks is
tabulated in Table 5-8. The bed weight increase was calculated by using
an initial height of 24 inches and a final height of 25 inches and a
density of 90 pounds per cubic foot. The material gathered in the se-
parator and baghouse was removed and weighed. The material exhausted out
the main stack was calculated from the measured particle loadings. The
S-34
-------
Table 5-8. LP-o INtRT/ASH MATERIAL ACCOUNTING
Parameter
(Ibs)
Bed growth
Aluminum tank sand/ash
Aluminum tank hard clinker
Flyash in inertia! separator
Hard clinker in inertial separator
Flyash out main exhaust
Flyash in baghouse hooper
Flyash out baghouse exhaust
Total
292
8760
132
783
23
321
2224
107
12,642
5-35
-------
material exhausted from the baghouse was estimated from visual compari-
son of the baghouse and main exhaust.
The total amount of inert ash material input to the system was estimated
from tne average inert fraction of the 12 solid waste samples taken dur-
in the 24-hour control period. The average inert fracion thus deter-
mined was 0.137; this figure times the 69,134 pounds of solid waste con-
sumed yields an estimated inert material input of 9470 pounds.
The reason for the large variation in the calculated and measured inerts
input is probably due to nonrepresentative solid waste samples. The
solid waste samples were taken from the transfer conveyo, and it is pro-
bable that a large portion of the inerts were on the belt ^nd did not get
sampled. The average inert fraction calculated from the accounting bal-
ance was 0.183.
Bed Maintenance-
The bed weight was mainta.ined within the required ± 10 percent. The ini-
tial bed weight was estimated at 7408 pounds and the final weight meas-
ured at 7700 pounds, which is an increase of 3.94 percent. This degree
of bed maintenance is verified by the bed pressure drop measurement pre-
sented in Figure 5-13. This figure shows bed differential pressure re-
maining almost constant until ET = 2063 minutes when a gradual increase
started due to one feed pipe operation. This is due to the reduced com-
bustor air flow when one feed pipe was shut down. Mo sand was added to
the bed nor was any bed material removed by the bed maintenance system.
5-36
-------
U1
LL!
CC.
Q
Q
LU
CO
\
iva isca ITSO
ELAPSED TIME (MIN)
221-233
Figure 5-13 LP-c iluU beJ Pressure Oroo
-------
Exhaust Gas Samplinq-
The gas analysis equipment described in Section IV and Appendix C was
used to monitor the exhaust gas constituents. The results are summarized
in Table 5-9. The CO? readings averaged 6.8 percent and are very consis-
tent. The CO readings averaged 41.7 ppm with a tendency to increase at
the end of the 24-hour period, probably due to operation on one feed
line. The CHX readings averaged 14.4 ppm and are quite consistent. The
NOX readings averaged 162.2 ppm and are consistent. The SOo readings
were lower than originally expected, and fell into the instrument ••eso'Ju-
tion range of 20 ppm. The 0? readings averaged 11.7 percent and are
very consistent. The HC1 readings averaged 90.3 ppm and show consider-
able fluctuation which is probably due to their being based on discrete
samples rather than averages as in the case for the other constituents.
A summary of the particle samples taken during the 24-hour period is pre-
sented in Table 5-10. These samples were taken at the inertial separa-
tor outlet by isokinetic sampling and the particle co-jnts were then made
with a Coulter counter. Two samples were taken at each 3-hour interval.
A total loading was performed on the first samples and counts for great-
er than 2 and 5 microns respectively, performed on the second sample.
It is apparent from this data that plugging of the inertial separator
probably occurred between the first and third sampling interval. This
can be seen in the sharp increase in total loading that occurred between
the third and ninth hour samples. Inertial plugging was verified in the
post-test Inspections which revealed that ash in the hopper had reached
5-38
-------
Table 5-9. LP-8 GAS SAMPLING VOLUMETRIC BASIS
Period3
1
2
3
4
5
6
7
8
9
10
11
12
Av.
co2
(*-)
6.8
6.5
6.8
6.8
6.8
6.8
6.8
6.8
6.8
6.8
6.8
7.2
6.8
CO
(ppm)
27.0
27.0
27.0
40.5
40.5
54.1
40.5
40.5
27.0
40.5
54.1
81.1
41.7
CHx
(ppm)
68
10.1
13.5
16.9
14.9
13.5
16.9
14.9
16.9
16.9
14.9
16.9
14.4 .
N0x
(ppm)
162.2
168.9
162.2
168.9
168.9
162.2
155.4
155.4
162.2
162.2
155.4
162.2
162.2
S0x
(ppm)
<20
<20
<20
<20
<20
<20
< 20
< 20
< 20
<20
<20
< 20
<20
°2
(*)
11.8
11.6
11.5
11.5
11.8
11.8
11.6
11.7
11.9
11.8
11.6
11.5
11.7
HC1
(ppm)
40.5
92.9
43.6
no.b
111.0
97.7
102.8
122.4
63.0
84.4
119.7
95.1
90.3
Periods denote 120 min intervals.
5-3?
-------
Table 5-10. LP-8 ''AR71CLL SAMPLING
Control
time
(hr)
3
6
9
12
15
18
21
24
Sample 1
Total loading
(grain/scf)
0.0566
0.1438
0.3472
0.3305
0.3975
0.3033
0.1784
0.6145
Sample 2
(grain/scf)
0.02541
J. 04996
n. 083 31
0.1/819
0.06970
0.11 -.56
0.14696
0.20323
(grain/scf)
0.06727
0.12985
0.16708
0.27140
1 13624
0.21918
0.22591
0.35940
5-40
-------
the tube bottoms. Vibrators were installed to keep the hopper
during future tests.
Superficial Velocity-
Superfidal velocity is presented in Figure 5-14. This data shows su-
perficial velocity was in the 5.8 to 6.0 fps range during two feedpipe
operations, and dropped to the 4.9 to 5.0 fps range on one feedpir'1
operation. The drop in velocity is due to the reduced comb us tor air flow"
caused by shutting down one feed line and venting the same amount to
maintain fluldizing air flow. Superficial velocity is well within the
4 to B fps range required by the test plan.
Post Test Inspection
After cooldown, the vessels were inspected to establish t*~,e system con-
dition. The combustor was in excellent condition. The bed depth meas-
ured 25 inches which verified the bed pressure drop measurements made
during the test. Approximately 10 pounds of alumina clinker material
formed on the exhaust pipe extending into the combustor. Five pounds of
deposit formed on the outside surface of the combustor exhaust duct se-
parated from the duct and were found on top of the bed below the exhaust
duct.
The and and ash collected in the first stage separator was within ap-
proximately 1.5 feet of the exhaust pipe. The weight of material col-
lected was 8769 pounds. This material had fused into one large clinker
and had to be broken apart for removal. Only a thin area on the sides,
top, and bottom remained loose. A large clinker (23.5 pounds) formed
-------
en
i
17KD
ELAPSED TIME (MIN)
221-234
Figure 5-14 LP-8 Superficial Velocity
-------
on the impingement baffle. Other large clinkers (108.5 pounds) formed
on the wall and top surface area where the exhaust was deflected by the
baffle.
In the inertial separator hopper, ash had built UD until it reached the
tube ash-outlets. Approximately 25 percent of the ash generated re-
mained in the hopper.
The baghouse hopper and filter bags were clean except for a slight ash
(
deposit.
During the test series, one case of splitter chute plugging and one case
of feed line plugging were experienced. A total quantity of 171,200
pounds of residential solid waste was received and processed. Shred
rate averaged between 1.4 and 2.0 tons per hour; however, the efficiency
of the air classifier was erratic due to its sensitivity to variations
in solid waste moisture content and compostion. Average of light frac-
tion separated ran between 75 and 85 percent of the received solid waste.
Conclusions
The following conclusions were drawn from the results of the LP-8 test-
Ing series:
1. In general the solid waste handling system performed adaquately
during the testing series. The shredding and air classification
operations on municipal solid waste produced a fuel form having very
satisfactory physical properties for energy recovery through com-
bustion 1n a fluldlzed bed reactor. However, the marginal operation
of the air classifier indicated that design efforts should be
5-43
-------
initiated to produce an air classifier less sensitive to variations
in solid waste moisture content and composition.
2. Test results indicated that the vertical fluidized combustor was a
highly efficient, easily fed, readily controlled reactor capable of
utilizing low quality fuels.
3. Results of low pressure operation of the solid waste fired, fluid
bed combustor showed an average combustion efficiency of 99.7 per-
cent.
4. No problem with fluidized bed residue buildup was experienced during
the testing series. Relatively large metal and inert particles en-
tering the active bed experience gradual oxidation or atrrition to
typical bed size particles and are eventually elutriated and col-
lected by the particle .separators. No bed material agglomeration
was experienced in tht 1270° to 1450° F bed temperature range.
5. Gas phase combustion above the fluidized bed has been reduced to ac-
cept'ble values without resorting to undesirable remedies such as
extensive internal bed structures, numerous fuel feed points, deeper
bed, multi-stage combustors, etc.
6. Earlier combustor freeboard and exhaust system deposit problems due
to the aluminum content of solid waste appear to have been solved.
7. The use of a residue removal system for the solid waste fired verti-
cal fluid bed combustor during low pressure operation would not be
required.
8. Trie addition of the alumina separator between the fluid bed combus-
tor and first stage inertial separator greatly reduced the high
solids imput to the first stage inertial separator.
5-44
-------
9. The material accumulating in the bottom of the alumina separator
must be continually removed or it will fuze into a large clinker.
10. Performance of the inertial separator during the testing series in-
dicated that, although the modifications made to the separator did
achieve longer trouble-free operation, the collecting hopper unload-
ing a.nd individual cyclone tube ash outlet plugging problem still
existed.
11. No evidence of serious hot gas system material corrosion or erosion
problems has been found.
TASK LP-11 TESTING (20-SERIES TESTS)
Test Summary
Continuous operation of a fluid bed combustor burning solid waste in ex-
cess of the required 48 hours under automatically controlled conditions
was successfully accomplishes with the completion of this test. The
second stage inertial separator outlet temperature was controlled by a
process control system to within 10° F of the desired set-point tempera-
ture for over 40 hours, with an additional 9 hours within 30° F. An on-
line gas analyzer sampled exhaust cases continuously, while discrete
samples were taken every 2 hours for HC1, ad every 3 hours for particle
count determination. A pre-test failure in the S02 system resulted in
no recording of S02 data, but successful operation of the remainder of
the on-line system resulted in a continuous record of 02, CO, C02, NOx,
and CHX. Sand and ash were successfully transported from both inertial
separators and the sand separator. Combustion efficiency computed by
the two methods indicated an efficiency of 99.7 percent. Thr superficial
5-45
-------
velocity was maintained within 4.0 to 5.4 feet per second throughout the
test with a nominal value between 5.2 and 5.3 feet per second.
Test Objectives
The primary purpose of this test was to verify the adequacy of the com-
puterized control system to automatics ly control the solid waste fueled
bed combustor and its ancillary equipment. Specific test objectives are
cited below:
1. Test duration: 48 hours continuous operation
2. Maximum variance temoerature from setpoint: + 10° F for 36 hours
and + 30° F for 12 hours,or + 20° F for 48 hours.
3. Temperature step changes:
a. from 1400° F to 1430° F and back
b. from 1400° F to 1370° F and back
4. Fuel: solid waste only
5. Combustion efficiency: 95 percent (two methods)
6. Continuous residue removal
7. Energy and material accounting: within 5 percent
8. Fluid bed height: ± 2h inches
9. Elutriation: less than 4 percent per hour of sand not passing a
30 mesh screen
10. Particulate: determine particulate loading of exhaust gas
11. Gas: determine percentage of the following chemical compounds:
S02, NOX, HC1, CO, C02, and CHX
12. Superficial velocity: between 4 and 8 feet per second
5-46
-------
13. Clinkers: none larger than can be removed by the residue removal
system
Test Setup and Hardware Description
The LP-11 test setup is shown schematically in Figure 5-15. Included
in this test configuration were the vertical combustor and particle se-
parators used during the LP-8 test series, an additional inertial sepa-
rator, and the modified solid waste handling system.
Hot Gas System-
The vertical combustor system used during the LP-11 test series was es-
sentially identical to the combustor system used in the LP-8 test con-
figuratior. The gas from the fluid bed was passed through the three
stages of particle separators, and was then exhausted to the atmosphere
downstream of the temperature sensing point on the third stage separator
after having been passed through the water spray. Residue from the se-
cond and third stage separators was pneumatically transported to the
baghouse filter.
In addition to the newly added inertial separator, other changes and ad-
ditions that were made to the hot gas system since the completion of the
LP-8 test series were: (1) the addition of pneumatic vibrators on the
hopper cones of both Inertial separators, (2) the enlargement of the ash
outlet on tne second stage separator hopper to 6 inches, (3) the addition
of a spinner duct to the alumina separator, (4) the addition of a sand
removal system to the alumina separator, and (5) the addition of the
necessary piping and control hardware required to pneumatically
5-47
-------
"!!
•. i
">
::i >•:-•(
>' 'i
r'*-
<•.->
'
'•;.*• '"I
t _; '•"••_!!i'i
I.
.I'Jt'
6--
m
Figure 5-15 LP-11 Test Setup
-------
transport the residue from the newly added third stage separator to the
baghouse filter.
The combus tor and particle separator configuration is shown in Figure
5-16. Details of the hot gas system have been described previously in
Section III.
Control System-
The temperature of the cleaned gas leaving the second stage inertial se-
parator (third stage separator) was the primary parameter to be control-
ed. The basic control concept used for this test was the controlling of
the fluid bed'and hot gas system outlet temperatures by modulating the
fuel flow into the combustor. Automatic control of the pilot plant was
implemented by coupling a process control computer through eight, two and
three-mode process controllers. Software was written and utilized to
maintain setpoint control of various key parameters, to control or moni-
tor numerous on/off and status functions, and to log test data for post-
test processing. A simplified block diagram of the control system is pre-
sented in Figure 5-17. Only the primary active elements of the test are
included. Start up, secondary, and inactive sections of the system have
been omitted for clarity. Refer to Section IV for a summary of the LP-
11 control system and software structure. Temperature sensors, and pre-
ssure sensors and switches that were used to control, monitor and ana-
lyze pilot plant operation during the 48-hour test period are shown in
Figure 5-15, and are identified in Tables 5-11 and 5-12. Fisher con-
troller settings are identified in Table 5-13.
5-49
-------
Figure 5-16 LP-11 Combustor and Particle Separator Configuration
5-50
-------
un
I
en
ATLAS
STORAGE
AND
FEED
SYSTEM
ATLAS
CONTROL SYSTEM
CONTROLLER 4
SIGNAL
CONDITIONER
AND
TRANSMITTER
SECOND
STAGE
SEPARATOR
X—X
SIGNAL
CONDITIONER
AND
TRANSMITTER
Fioure 5-17 LP-11 Simplified Control LOOD Block Diagram
-------
TABLE 5-11 LP-11 TEMPERATURE SENSORS
Thermocouple I. D.
Measurement3
000
001
002
003
004
005
006
007
008
009
010
013
014
015
016
017
018
019
020
021
024
025
026
027
028
029
031a
032
035
036
072
Fluid Bed 9"
Fluid Bed 9"
Fluid Bed 9"
Fluid Bed 9"
Fluid Bed 9"
Fluid Bed 9"
Combustor Plenum 9"
Combustor Plenum 9"
Distributor Plate Surface
Distributor Plate Surface
Fluid Bed 0"
Combustor wall
Freeboard 10"
Freeboard 10"
Combustor wall just below feedline - surface
Combustor dome - surface
Diesel oil suooly - 1"
Aluminum vessel wall - surface top dome
1st Stage Inertial Separator top dome - surface
1st Stage Inertial Separator wall - side surface
1st Stage Inertial Separator material removal
line between water injection point and valve
009 - surface
1st Stage Inertial Separator outlet gas - 1"
1st Stage Inertial Separator outlet pipe-surface
2nd Stage Inertial Separator internal ras - 10"
2nd Stage Inertial Separator top dome - surface
2nd Stage Inertial Separator side wall - surface
2nd Stage Inertial Separator outlet - 1"
(Well upstream of water injection point)
2nd Stage Inertial Separator material removal
line between water injection point and valve 0/0
surface
Propane supply - 1"
Low Pressure compressor outlet gas - 1"
Aluminum vessel inlet gas - 1"
Dimensions associated with each T/C indicate length that the T/C pro-
jects into the associated area. When "surface" is noted, the T/C is
tack welded to the noted surface.
3
Control Thermocouple
5-52
-------
TABLE 5-12 LP-11 PRESSURE SENSORS AND SWITCHES
Item Identification
Pressure Transducers
101
107
120
121
123
124
125
140
141
146
160
Position Transducers
200
220
221
222
223
225
226
231
232
241
242
Miscellaneous
400
402
403
404
405
406
407
408
Measurement
Plenum Pressure
Compressor Outlet
Bed AP
Distributor Plate AP
1st Stage Inertia! Separator AP
2nd Stage Inertial Separator AP
Aluminum Vessel AP
Feedline No. 1 Flow pitot tube
Feedline No. 2 Flow pitot tube
Combustor inlet pitot tube
Diesel flow to solid waste combustor
Outfeed conveyor level
Feed air line 1 valve position
Feed air line 2 valve position
Low Pressure air 'low control valve
Large overboard dump valve position
Inlet No. 1 to baghouse valve position
Inlet No. 2 to baghouse valve position
Bypass valve position
Diesel fuel valve position
Sweep speed (command)
Atlas outfeed speed (command)
Accumulated weight over conveyor
Oxygen analyzer output
Carbon monoxide analyzer output
Carbon dioxide analyzer output
Hydrocarbon analyzer output
Hydrochloric analyzer output
Sulphur dioxide analyzer output
NO and N02 analyzer
5-53
-------
TABLE 5-12. (continued) PRESSURE SENSORS AND SWITCHES
Temperature
LD. Switches
500
502
Pressure and Flow
Switches
600
601
604
615
620
660
661
662
Position Switches
706
721
724
730
731
733
73*
760
Miscellaneous
900
901
903
904
905
906
907
908
909
910
961
Measurement
Classifier temperature
Atlas hydraulic supply
Atlas internal pressure
Backheat fuel pressure
Baghouse internal pressure
Shop air pressure
Low Pressure Compressor filter AP
Backheet burner cooline air flow
Oil gun mix flow (Air)
Oil gun cooling flow (Air)
Diesel oil supply level
Low Pressure valve #729 position
Low Pressure comp. outlet valve #007 position
Belt position - transfer conveyor
Belt position - weigh conveyor
Exhaust cap status
Valve 005 status
Shredder vibration switch
Shredder conveyor status
Classifier status
Atlas sweep status
Atlas outfeed conveyor status
Transfer conveyor status
Weigh conveyor status
Feeder valve No. 1 status
Feeder valve No. 2 status
Igniter transformer status
Baghouse dump valve status
Flame sensor switch
5-54
-------
Table 5-13. LP-11 TEST, FISHER CONTROLLER SETTING
Controlled
parameter
PB %
Reset
Range
Control action
Rate
Control mode
Controlled
element
Controlled
variable
Fisher Controller
No. 1
100
10
xl
Direct
-0
Manual
Valve 005
Combustor
air flow
No. 3
50
2.0
xl
Reverse
2.0
Remote
Valve 002
Oil flow
No. 4
30
2.0
xl
Reverse
1.0
Remote
Atlas Control
System
Solid
waste feed
No. 5
40
10
xl
Direct
-
Remote
Valve 000
UFeedllne
air flow
No. 6
40
10
xl
Direct
-
Remote
Valve 001
#2 Feed line
air flow
No. 7
100
0.1
xl
Reverse
-
Remote
Fisher *3 or
I4a
Fisher
setpoint
I/I
I
en
control
C0ntroller #3 for fuel oil feed control or with controller 14 for solid waste feed
-------
Solid Waste Handling System-
Resjlts of the LP-8 test series and longer duration tests scheduled for
the LP-11 series established requirements for improved through-put and
reliability in the solid waste handling system. (Figure 5-18.) Changes
were made to both the shre.der/air classifier/pneumatic transfer sub-
system and the Atlas store/combustor feed subsystem (Figure 5-19). De-
tails are discussed in Appendix A and are briefly summarized below:
1. The firect-coupled metal air classifier used in previous tests was
replaced by an experimental wooden air classifier.
2. A shredder outfeed conveyor equipped with two paddlp *ype load
levelers to maintain a constant 3-inch level of waste feed at the
inlet of the new air classifier was added between the shredder and
air classifier.
3. A reject fraction conveyor was added to the air classifier reject
material outlet.
4. The two 16-inch airlock feeder valves were replaced ^y two 30-inch
Esco Rotafeeders (Figure 5-20). Each valve had volumetric capacity
of 30 cubic feet per minute.
5. The two 5-inch diameter combustor feed pipes were replaced by two
6-inch diameter feed pipes.
-------
CO',TRrL PW.CL
PACKER
PAM.
Figure 5-18 LP-11 Solid Waot- rianJlin^ S/stc'i
-------
in
s
KM FEEDER
VALVE (TYPICAL)
Figure 5-19 LP-11 Illas Storage and Combustor Feed Subsystem
-------
01
Ul
SO
VALVE NO. 2
DRIVE ASSY
VALVE NO. 1
WIVE ASSY
Figure 5-20 30-Inch Feeder Valve Installation
-------
Test Results and Analysis
Exhaust Temperature Control -
The basic test conditions called for controlling temperature at the out-
let of the third stage separator to 1400° ± 20° F. Acceptable variations
to this schedule were designated by the Project Officer. The alternate
condition under which this test was run allowed the control temperature
to devioce by ± 30° F from the setpoint but imposed a ± 10° F maximum
deviation for a minimum of 36 consecutive hours. Included as part of
this schedule was the requirement to increase the nominal setpoint from
1400° F to 1430° F for 4 hours and decrease it to 1370° F for another
4 hours. Deviations from specified limits were allowed during transi-
tion. These conditions were satisfied during the course of this test.
Specific key temperatures were sampled and recorded on hard copy at the
rate of approximately once per minute. Larger quantities of data were
sampled at the rate of roughly every 6*s minutes and stored on the com-
puter disc file for post-test processing. Figure 5-21 presents a plot
of the control thermocouple at this slower sampling rate. Elapsed time
began when the low pressure compressor was turned on. By ET * 43
minutes, the control temperature was within i ?0° F of 1400° F and con-
tinuous operation for 48 hours began. During this 48-hour (2880 minute)
period the control temperature exceeded ± 20° F for a total of 8 minutes.
At ET * 84 minutes, the temperature error was below -20° F for 4 minutes
and again at ET » 575 minutes for 4 minutes. At the latter time the
minimum test temperature of 1370° F was reached due to a problem with
the solid waste feed system which virtually shut off the fuel flow for a
5-60
-------
in
i
COMPLETE 36 HOURS -
WITHIN ± 10°F
BEGIN 48 HOURS BEGIN 36 HOURS
WITHIN
UJ
BAND
.' ±10°F BAND -,—| -:. - - —1 ,SETPOINT CHANGE
'
T '"
ELAPSED TIME (MIN)
221-236
5-21 LP-11 Third Stage Separator Exhaust Temperature
-------
short time. The dynamic oscillations of the control temperature due to
this perturbation damped to within ± 10° F by ET « 605 minutes. Froir
this time until test shutdown at ET « 3209 minutes, a total of 40 hours
and 25 minutes, the control system maintained ± 10° F except for two al-
lowed deviations during transitions from one setpoint to another. The
first of these occurred within the required continuous 36 hours and
while stepping up to 1430° F at ET » 1836 minutes. The second was at
ET = 2886 minutes during the step back to 1400° F from the 1370° F set-
point. The required 48-hour period was completed at ET = 2923 minutes.
Solid waste consumption continued however, for an additional 107 minutes
(hot shown) before shutdown.
/
Statistics of the control temperature are shown in Table 5-14. Except
for the statistically non-significant six data point period at the end
of the 48 hours, the recorded control thermocouple had a mean value
within 0.2° F of the desired setpoint. The variance and standard devia-
tion of each setpoint group is also shown. For comparison, the 24-hour
LP-8 test utilized process controllers only for the control of the first
stage inertial outlet temperature to 1450° F. The mean of that test was
1456° F with a standard deviation of 7.6° F.
These results illustrate some of the advantages of the hybrid control
approach; an inner analog controller loop based on measured sand separa-
tor top gas temperature and computer-generated setpoint. This inner
loop setpoint was re-evaluated every 10 seconds by an outer control
loop mechanized in the digital computer. Since both the setpoint and
process variable measurement (i.e., the second stage inertial separator
5-62
-------
Table 5-14. CONTROL TEMPERATURE STATISTICS
Elapsed time
duration
(sec)
40 - 1820
1830 - 2070
2080 - 2610
2615 - 2875
2890 - 2925
Number
of
points
308
42
93
44
6
Setpoint
(°F)
1400
1430
1400
1370
1400
Mean
(°F)
1399.8
1430.2
1 399 . 8
1370.0
1398.3
Variance
(°F)2
12.5
10.4
2.1
4.0
14.3
Standard
deviation
(°F)
3.5
3.2
1.4
2.0
3.7
-------
outlet gas temperature) for the integrating type of outer loop control
are contained in computer memory, very precise staady state control is
obtained.
The corollary is that the comparatively high bandwidth analog control
loop "floats" in setpoint as required but avoids needless loading of the
computer with high frequency calculations.
Exhaust Gas Samplinq-
An on-line gas analyzer was used for the determination of all gas con-
stituents except HC1. Single samples taken at 2-hour intervals were ti-
trated by the Volhard method for HC1 level determination. The data ob-
tained from the on-line system, averaged over 24 2-hour intervals, are
presented in Table 5-15 along with the discrete HC1 data. The S02 bat-
tery cell leaked and shorted the system, making the S02 recording mean-
ingless. The on-line system was zeroed to an atmospheric air background
and the hydrocarbon analyzer indicated a slightly negative production
from the combustion process. All the data indicated consistent readings
except the HC1, which had large fluctuations. This may be due to the
composition of the fuel during each period, a result of single point
sampling or a result of errors inherent in the technique used for HC1
determination.
The exhaust gas was sampled isokinetically using a wet impingement tech-
nique. (See Appendix C.) Two successive samples were taken at 3-hour
Intervals. The first sample was dried and weighed to determine the to-
tal particle loading; the second was subjected to analysis using a
5-64
-------
Table 5-15. LP-11 EXHAUST GAS COMPOSITION (VOLUMETRIC BASIS)
Period3
1
2
3
4
5
6
7
8
9
in
n .
12
13
14
15
16
17
18
19
20
21
22
23
24
Av.
co2
CO
6.0
6.0
" 5.8
6.0
6.0
5.8
5.8
5.8
5.8
6.0
5.8
5.8
5.4
5.4
6.2
6.0
6.0
5.8
5.8
5.8
5.5
5.5
5.5
6.0
5.81
CO
(onm)
30
I*
it
n
n
M
tt
it
•i
n
11
n
i*
ti
II
II
II
H
II
II
»
II
ft
II
30
CHx
(onm)
0
II
It
(I
II
M
II
If
M
II
II
M
M
II
II
II
ri
if
»
M
II
II
It
II
0
F
NO,
(onm)
157
140
124
130
124
135
140
140
135
135
140
140
135
150
170
145
UO
135
135
135
135
135
135
135
138.5
SO/
(onm)
NA
n
n
n
n
ii
11
ti
it
n
n
n
ii
M
»
"
H
If
II
II
II
M
n
n
-
°2
(%)
13.5
13.2
13.0
12.8
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.8
12.8
13.0
13.0
13.2
13.2
13.5
13.5
14.0
13.8
13.2
13.4
HC1
(npm)
210.0
242.8
173.8
197.1
197.1
164.3
145.5
116.9
84. b
141.5
187.1
141.9
199.8
151.2
188.9
115.2
119.2
160.2
256.9
122.5
140.4
138.9
129.6
137.8
161.1
aPeriods are at 120 minute intervals
DNA * not available
-------
Coulter counter to obtain a size distribution. Particle sampling re-
sults are cited in Table 5-16.
Bed Maintenance and Material Balance-
Pre- and post-test measurements indicated a bed height growth of 2 inches
(from 24 to 26 inches) or, based on constant density, an 8.3 percent
weight increase. The measured pressure drop across the fluid bed is
shown 1n Figure 5-22; the rise in pressure drop is a result of the bed
weight increase.
A tabulation of the inerts actually weighed, computed, or estimated ap-
pears in Table 5-17. The clinker and the sand or ash which accumulated
in each vessel or hopper was weighed after the test. Bed growth was
computed based on the height change mentioned earlier and the ash exit-
ing the main exhaust was computed using average particle loading. The
weight of ash leaving the sand bin vent and the baghouse exhaust was es-
timated by comparing the appearance of each exhaust density with the
known main exhaust density. Compared to the total solid waste fed, the
total inerts from Table 5-17 (Including flyash) corresponds to a 14.79
percent inert fraction.
Small solid waste samples were taken from the outfeed conveyor at 2-hour
intervals during the test. Moisture, inert fractions, and heating value
were determined from these samples. The average of all 24 samples yield-
ed an Inert fraction of iO.53 percent. The 4.26 percent difference be-
tween this value and that based on actual Inerts 1s probably due to the
laboratory method of sampling and processing the solid waste. The heavy
5-66
-------
Table 5-16. LP-11 PARTICLE SAMPLING
Control
time
(hr)
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
Sample 1
Total loading
(grain/scf)
0.0814
0.1454
0.2313
0.3704
0.3903
0.6146
0.7100
0.5004
0.3892
0.5133
0.5147
0.5185
0.5580
0.3882
0.5716
0.4001
Sample 2
Loading >5
(grain/scf)
0.0039
0.0272
0.0731
0.0423
0.1224
0.1713
0.1018
0.0997
0.0857
0.1679
0.0830
0.0844
0.1080
0.0719
0.0918
0.0508
Loading >2
(qtain/scf)
0.0135
0.0680
0.1234
0.0899
0.2140
0.3132
0.2664
0.1937
0.1863
0.3234
0.2348
0.2040
0.2753
0.1826
0.2100
0.1505
5-67
-------
en
i
C7>
CD
g'.
Of.
5.
o: •
o.
1 1
BEGIN 48 HOURS
! I
1 '
1 ' ' , 1 *
t , — . .
I" f
I
' -
-,- --,
- !
r -
- — r
I
I
.
-I i-
.COMPLETE 48 HOURS
j--p-j ^ j-j - -
f '
-------
Table 5-17. LP-11 INERTS ACCOUNTING
Location
Bed growth
Clinker in sand separator
Sand/ash in bottom of 1st stage separator
Sand/ash in sand bin
Ash out of sand bin vent
Ash in 2nd stage inertial separator
Ash in 3rd stage inertial separator
Ash in baghouse hopper
Ash out of bajhouse exhaust
Ash out of main exhaust
Amount
(lb)
550
31
1,145
10,200
22
253
53
4,925
650
550
(••}
3.0
0.2
6.2
55.5
0.1
1.4
0.3
26.8
3.5
3.0
TOTAL
18,379
ln'0.0
-------
fraction is likely to settle to the bottom both on the conveyor prior
to sampling and in the sample storage bag. Thus, since the complete
sample is not processed, a few large particles may escape causing dis-
tortions in the final result.
Combustion Efficiency Technique- The last column of Table 5-18 con-
tains the efficiency results for this test, using the heat balance tech-
nique. Mean efficiency for the 48-hour period was 99.68 percent, but a
one-standard deviation for the data of 7.91 percent was computed. The
table also shows the principal time varying data which went into the
computations.
As shown in the table, a 4.26 percent difference existed between inert
fractions derived from the mass balance and the average of laboratory
determinations.
In order to correct the measured ash/inert level to that of the inerts
accounting, the measured iiierts/ash level was multiplied by the ratio
of 14.89/10.53 which results in a 24 sample average that is equal to the
inerts accounting value. The inert fraction, column C, reflects this
correction.
Carbon Combustion Technique-.. In the carbon combustion method
for computing the combustion efficiency it is assumed that unburned
carbon in the flue gas exists in the forms of CO (carbon monoxide),
HC (hydrocarbons), and C (elemental carbon) only. Therefore,
5-70
-------
Table 5-18. LP-11 COMBUSTION EFFICIENCY, HEAT BALANCE TECHNIQUE
". '• t ; r -
vat
.1
12
I <",
37.; 3
i 3
1
i'j
:"
/ oistjre
j fraction
C
Inert
fraction
(incl.
,
C.1543
-.; :.-•« ! c.14'14
?< -M6Si | «?.!'•«
-j 0.3.170
0. Iv'O
12 i C -Vji-, | 0. 14^2
7i ' -,\ 3-2: 0. IT'S
;3 O.J43J
24 0 3^83
9£ 1 1 ~")f'j
'
o.;r/5
: ; j o 3550
9' , 0.3336
9? 0.3SJ6
28 0.43=2
ci 03? 19
t
0.1246
0.1 3H7
0.1211
C. U?9
0. 1532
0.1139
0.1207
0.1084
0. 1255
D
Av.
air
flow-
rate
-n.l
~.2',. 3
330 2
330.4
330.5
310.9
331.0
330.4
.130.4
33C.4
3^0.2
329.8
329.1
328.6
329.7
E
rar,:;e
1262. '•- 1144.!:
1144.i.l75i.'
1291.4-1185.6
1185.6-1342.2
1342.2-1326.9
1326.9-1299.7
1299.7-1339.0
M39. 0-1338 2
1338.2-1267.7
1267.7-131b.fe
1316.S-12C4.1
1284.1-1317.4
1317.4-1256.6
1256.6- 947.3
947. 3- 1 304.. T
F
Av.
2nd
sta.je
out'ot
nru
1400. ;
139J.I.
140,)..
1393.T
140T.I
1399.1
1 399.il
1399.(
i -.01. :
u.j.i
139'J.'.
1400.1
1399.'
14 16. .
G
Heat
loss
ratp
(Stu/nin)
H./J5.0
1CJ04.0
1"667.0
12119.0
103E4.0
9909. 2
C;238.5
8636.5
7788.9
73:>1.9
80 'j.. .5
7929.2
7080.7
6374.2
7 /Tj2 . 4
H
Co'iiouted
heating
(Btu/ C
Ib ccTb)
C786.6
8359.6
75S8.2
8SS4.1
91S9.2
8700.1
7013.0
8130.0
8309.3
69 .:6. 5
7930.1
7453.7
7027.0
7862.8
7323.8
I
Heating
value
from calori-
(Dtu/dry^lb)
75OJ.O
6878.5
6907.6
7052.0
7209.2
6697.0
7317.5
7186.2
7276.7
727C.5
6999.8
7484.9
7199.9
7177.4
7082.4
J
Lower
heating
value. Q
(3tu/ 1L) coin)
• ~. • ; C
76L,.,
7964.4
K 36. 2
8156.0
7*14.1
7S.S3.8
7d?2.6
7888.7
77?). 3
7832.3
BOO 7.1
7775.0
7707.8
7680.5
K
Co-"bustion
ef fi ciency
ICO Qr/Q
f/3.93
1C3.2
94.90
iO',.4
112.3
ir.9.3
96.87
103.0
IC'j 0
r.8.!5
101.2
93.09
100.7
102.0
95.36
-------
Table 5-18 (continued). LP-11 COMBUSTION EFFICIENCY, HEAT BALANCE TECHNIQUE
Perf
Inter-
val
16
K
13
13
20
a
ti
23
24
A
fv.
solid
/a< '.f
few.
rate
w
43.93
tf- .78
43.52
l?.?1)
40. 5t
51.56
•n. :s
i'j.47
41.98
B
Moisture
friction
0.372!
r«079
3.3445
.0.33CO
0.3932
0.39!»
0.3620
0.4280
0.4214
C
Irwt
fraction
(Incl.
ash)
0.1912.
0.1:33
0.1755
0.2219
0.1310
0.1160
0.1776
0.1698
0.1318
D
Ay.
air
flow-
rate
(Ib/nfn)
328.5
3?8.3
328.3
328.6
328.3
328.1
3?3.5
328.6
328.7
E
Bed
temperature
rapge
(&F)
1304.8-1299.2
1299.2-1293.5
1293.5-1310.8
1310.8-1325.0
1325.0-1335.6
1335.6-1336.3
1336.3-1274.4
1274.4-1288.6
1288.5-1333.6
F
Av.
2nd
stage
outlet
temo.
(°n
1429. 1
1428. S
1399. J
140C. )
1399.8
1399.1
1381.)
1369.)
1382.)
G
Heat
loss
rate
(Btu/min)
7K6.7
' 6674.0
r?02.8
6353.6
6358.3
6520.4
5361.2
5589.9
6665.0
H
Computed
heating
value. Or
(Otu/
Ib ccn.bj
8138.4
7601.7
7212.5
7761.8
7764.8
7794.4
7799.7
9112.2
8148.6
7912.8*
599. 65b
1
Heating
value
from calori-
meter. On
(BtuMry Ib)
66S1.1
7159.4
72f.2.3
6480.7
7014.6
7300.3
6868.0
6771.5
7266.7
J
Lower
heating
value, QL
(Stu/ib comj
7946.8
7816.9
8398.3
79«8.2
7680.9
8108.1
7999.0
8048.1
8085.6
7943.6*
195. 09b
K
Combustion
Efficiency
100 OC/Q,
(-:)
102.4
97.24
85.38
97.65
101.0
96.13
97.51
113.2
100 8
99.68*
b
7.908
en
ro
Standard deviation
-------
combustion efficiency may be calculated by the following equation:
J|a(AH
'2
where: a * weight concentration of C0£ in flue gas
b » weight concentration of CO in flue gas
c a weight concentration of HC in flue gas
d * weight ratio of C to flue gas
The further simplification of the above equation is based on the assump-
tion that 1 pound of carbon releases l",087Btu when converted to C02,
3965 when converted to CO, and about 2000 Btu when forming ethane,
(CoHg). the most likely hydrocarbon in the flue gas.
Hence,
14087 , . 3956 h + (2)(2000) c
n « ~4T~ a "21T 30
,a b 2c d »
(34~ * Iff * 37 * TJ> 14087
or n . a + 0.442b + 0.416c
a + 1.571b + 2.933c * 3.667d
Errors associated with failings fo these assumptions are expected to be
small. Therefore, the 99.7 percent average combustion efficiency compu-
ted by this method as seen 1n Table 5-19 1s considered to be an accurate
representation of the cr+ual efficiency.
he C02, CO, and CHX weight fractions presented In columns A, B. and C
of Table 5-19 were obtained by averaging the results of the on-line
5-73
-------
Table 5-19. LP-11 COMBUSTION EFFICIENCY, CARBON COMBUSTION TECHNIQUE
Performance
Interval
1
2
3
4
5
6
7
3
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
A
Wt. Traction
of C02
a
Mt. fraction
of CO
9.1 x 10'2 ! 2.9 x ID'5
9.T x 10~2
C
Ht. fraction
of CHX
0.0
N
8.8 x UT2 " .
9.1 x 1CT2
9.1 x NT2 " ;
8.8 x 10'2 " !
8.8 x 1
-------
analyzer over 2-hour intervals. The free carbon concentrations in col-
umn D were obtained by outside laboratory analysis on 24 samples taken
from the baghouse hopper during the test.
Superficial Velocity-
For this test,"the superficial velocity was specified to be between 4
and 8 feet per second. As shown in Figure 5-23, this requirement was
met for the duration of the test. With a fixed mass airflow throuqh the
hot gas system, the changes of superficial velocity are associated with
bed temperature variation and combustor bypass flow. Combustor bypass
flow is a controllable function that allows the bypassing of the fluid
bed by some of the compressor air flow. The air flow is mixed with the
gas coming off the bed. The primary function of the bypass air loop is
to provide an internal control loop for bed temperature maintenance.
During short periods of problems with support hardware, (e.g., the solid
waste feeder valve adapter) caused the bed temperature to drop to levels
below the nominal level of about 1325° F- Each of these drops is re-
flected in Figure 5-23.
Post Test Inspection
All major system components were in excellent condition at post-test in-
spection. At this time the combustor and first stage separator had been
used for a total of 158 hours of solid waste operation, while the third
stage separator had been operated for 101 hours. The second stage se-
parator had been operated for a total of 177 hours, since it was used
during some of the horizontal combustor tests. The refractory in the
5-75
-------
T
1 •
BEGIN 48 HOURS
' COMPLETE 48 HOURS
OURS
• -ji —4-
'-J
en
U-
>- *
t-^
0
_i
UJ -
5»
_1
O .
»— •
U.
cc
UJ
Q.
rs .
•
1
•-— v^-*%-^-"v-v^-%*-- -_ -v- vvv -~v*\ ^^---v*"^-''vv^^ \ .
1 1
1 • * ' • j '
. 1 : :
. : , i
• i - t • ! *
~\A ^:V ' : i t \\~~\--_\ '\': :
' ' i ' t ' ' t t r • r "*" ~;
....:. I....:..:. ill ;H J • !- ;
: ; rr . T i .~r"i
: ; j — _,,._ _._j_ i ... j. _j
i : ' M i ; - ' ; t
. . , . . . . j .
iiiiii'iiittiiitttiiiiiitiiit
in
—
MC ••• *«B MM I4BB k«M l-Ofc 1M9A !•» »«» IBM *HM a^M) »
-------
combustor showed no signs of wear or erosion.
A one cubic foot sample of bed material was examined for metal and found
to contain 0.33 pounds. It was therefore expected that 28 pounds (0.33
x 85) of metal would be present 1n the bed. A total of 316,000 pounds
/
of solid waste have been Introduced Into the bed since the vertical com-
bustor was commissioned. Allowing for attrition, previously calculated
to be 0.004 pounds oxidized per pound present per hour, the required
quantity of metal Introduced Into the bed to provide this result would
be 0.0318 percent of the solid waste. The metal content from 5 solid
waste samples taken during test run 20-4B averaged 0.032 percent, which
1s In agreement with the calculated value. During the model test dis-
cussed 1n Section III (Task LP-5), this value was calculated to be 0.17
percent of solid waste. The primary -..rerence 1s apparently due to the
present air classifier which removes most of the metal.
During the test the pressure drop across the distributor plate was ob-
served to Increase from 9 to 23 1wg. Post-test Inspection revealed a
considerable amount of dust buildup on the dlffusers. Subsequent bench
flow tests of dirty and cleaned dlffusers confirmed that the excess pres-
sure drop resulted from the dust. Since previous tests have not experi-
enced this phenomena, the cause has been attributed to the malfunction
of the baghouse discussed later 1n this section. This malfunction al-
lowed large amounts of flyash to escape Into the plant. The low pres-
sure compressor Intake located 10 feet from the baghouse was continually
exposed to dust resulting 1n high Intake air particle loading.
5-77
-------
The addition of the spinner duct in the first stage separator resulted
in a marked reduction in clinker formation over previous tests. A total
of only 31 pounds of clinkers were found in the duct. No erosion of the
duct was apparent. The transport of the captured inert material to an
off-line storage bin (about 20 feet away) was successfully accomplished.
As expected, about 10 percent of the total sand/ash collected by this
vessel built up at the bottom until it formed a conical entrant ; to the
discharge point.
The second stage separator had two of the 24 tubes visibly plugged at
the ash outlet, but the hopper had only small amounts of ash on the
walls. The newly installed vibrators were effectively preventing hopper
ash buildup.
The third stage separator hopper was likewise free of major ash deposits.
All 80 cubes, however, were plugged or partially plugged at the ash out-
let.
The puffback blower of the baghouse assembly was inoperable during ap-
proximately 6*4 hours of test due to a bearing seizure. This caused an
estimated 500 pounds of ash to escape to the atmosphere. Post-test in-
spection Indicated that all other baghouse functions operated properly.
Support of the LP-11 test series required the processing of approximately
290.000 pounds of residential solid waste. Efficiency of the shredder/
«1r classifier/pneumatic transfer subsystem was consistently Improved
over LP-8 data, with through-put averaging 2 TPH and better and air
separation yielding a cleaner cut of li'jht and heavy fraction. The
5-78
-------
results of experimentation with the wooden air classifier established
parameters for design of the full scale 5 TPH air classifier rind pneu-
matic transport system required for tht future LP-13 testing. Design
criteria established were as follows:
Feed rate: 5 TPH
loading; 0.5(lbs of solid waste/lbs of air)
Air classifier velocity: 1300 fpm
Blower design point: 6800 scf-n, 9 in. w.q., 22hp
Conclusions
The following conclusions were made from the results of LP-il testing
series:
1. Shredding and air classification operations on municipal solid waste
produce a fuel form having very satisfactory physical properties for
energy recovery through combustion in a fluidized bed reactor. A
solid waste handling subsystem with reliable components has been de-
veloped and extensively operated.
2. The vertical fluidized bed combustor has again been shown to be a
highly efficient, easily fed, readily controlled reactor of simple
design that is capable of utilizing low quality fuels.
3. Results of low pressure operation of the waste-fired vertical fluid-
ized bed combustor d ring the testing series showed an average com-
bustion efficiency of 99.7 percent.
4. Test results showed no problem with fluidijed bed residue buildup.
No bed material agglomeration was experienced in thp 1270 - 1450° p
bed termperature range.
5-79
-------
5. bas phase combustion above the fluidized bed was again within accep-
table values.
6. Combustor freeboard and exhaust system deposit problems due to the
aluminum content of solid waste have been solved.
7. The addition of the spinner duct in the first stage separator great-
ly reduce the formation of clinkers at the initial gas impact point.
8. The capability of the separator material removal systems to contin-
uously remove and transport sand and ash residue material from the
bottom of the separators to an external collection point was suc-
cessfully demonstrated.
9. While the newly installed pneumatic vibrators were effectively pre-
venting ash buildup in the hoppers of the inertial separators, pro-
blems of long term inertial separator performance degradation have
been encountered due to the tube plugging, particularly in the
final stage with very fine hot fly ash. Remedial development work
will continue at the more promising high press'ire conditions asso-
ciated with integration into the gas turbine cycle.
10.. Exhaust gas sampling instruments again indicated that the pollution
control effort required for the gaseous components of the CPU-400
process can be expected to be minimal.
11. No evidence of serious hot gas subsystem material rorrosiin or
erosion problems has been found.
12. The capability of the computerized central system to adequately con-
trol low pressure pilot plant operation was successfully demonstra-
ted.
5-80
-------
SECTION VI
TURBO-ELECTRIC SYSTEM
This section discusses the design and procurement of a turbo-electric
system for the high pressure configuration of the CPU-400 pilot plant,
and summarizes the results of a survey of candidate turbines for the
CPU-400 prototype plant. The pilot plant turbo-electric system was de-
signed and procured under contract tasks HP-1 and HP-2, respectively.
The prototype turbine survey was performed under contract task PI-2.
DESIGN AND PROCUREMENT OF THE TURBO-ELECTRIC SYSTEM
The turbo-electric system compresses the air used in the combustion pro-
cess, uses the combustion product gases to develop turbine shaft power
to drive the compressor and the power generator, and then generates and
controls electric power output. The turbo-electric system consists of
a 1000 kH turbine-driven generator, the switchgear and electric load
banks needed tc absorb the generated power, and the inlet and outlet
piping necessary to mate the t'irblne to the hot gas interface.
The basic components of this system have been predominantly supplied by
Mid-Continent-Ruston as a packaged turbo-electric unit. The design and
6-1
-------
installation of the turbine and the generator and switchgear into a
skid-mounted assembly, were accomplished by Mid-Continent Supply
Company. When delivered, the turbine system was modified sufficiently
to install the turbine combustor in a remote location, and to connect
the solid waste combustor and separators to the turbine. The electric
load bank and light bank, supplied by another contractor, were also
added to the turbo-electric unit.
Turb>-Generator Selection
The criteria for the selection of the CPU-400 pilot plant gas turbine-
generator were predicated on the results of system engineering studies
performed under contracts PH 86-68-198 and fiP-03-0054. From these
studies it was concluded that, as far as possible, the pilot plant gas
turbine-generator should be sized to mate with the existing pilot plant
combustor, and should be a readily available, standard production unit
requiring on^y slight modi fiction for use with the pilot plant hot gas
systti.ii. Most important, the turbine had to be an open cycle, constant
pressure machine, designed with an external combustion chamber and an
external conpressed air ouct. The pilot plant combustor had been sized
for a 1500 hp gas turbine-generator that would produce approximately
1000 kW of electrical power, essentially one-tenth the power output of
the full scale CPU-400 at full flow conditions. Also, in order to.
avoid excess fluid t»ed maintenance problems, its inlet tempera-
ture had to be maintained under 1500° F; this limited the choice of
6-2
-------
turbines to those with turbine inlet temperatures in the 1400°F to
1500°F operating range.
A review of the available turbines indicated that the Mid-Continent-
Ruston TA 1500 gas turbine best met these criteria. It was found to
be a well constructed heavy duty unit, amenable to experimental work;
it was reasonably priced; it had an acceptable delivery schedule; and
it was designed for use with an external combustor. The Ruston turbine
inlet temperature and compressor pressure output ratings matched the
temperature limits and operating pressure requirements of the pilot
p.lant combustor, and a 1000 kW generator was supplied with the turbine
as standard equipment.
Turbine Description
A cutaway druwing of the Ruston TA 1500 turbine configuration as nor-
mally supplied by ^id-Continent Supply Company is shown ,n Figure 6-1.
The turbine combustor is shown in its standard position with a cross-
over duct removed to permit a view of the rotary components. The
cross-over duct is normally connected between the compressor outlet
flanae, shown in the background, and the rectangular combustor fla.iae,
shown in the foreground.
The turbine consists of a gas generator assembly and a power assembly.
Included in the gas generator assembly ar
-------
EWAUVT OUCTi
COWM.SSOR POMfR TURSINt
OUTLET
Figure 6-i Pus ton T A-1500 Gas Turbi.ne
-------
are independent of one another and eift,er one may be dismantled without
interfering with the other. The combustion chamber can also be removed
witnout disturbance of the other components. Details of the turbine,
as installed in ie pilot plant facility, are shown in Figure o-2.
During operation, air at atmospheric temperature and pressure is drawn
into the 13-stage compressor through an air filter. As tne air passes
through the compressor, the moving blades produce a rise
-------
en
Figure 6-2 Turbine Installation into the CPU-400 Pilot Plant
-------
cycle pressure and maximum gas temperature in accordance with the
power required at the output shaft.
With the temperature and pressure having dropped considerably, the gas
flows to the power turbine, where some of its remaining heat energy
and a large portion of its pressure energy is converted into useful
power at the output shaft. This is achieved by expanding the gas down
to atmospheric pressure before it leaves the turbine and is discharged
to the atmosphere. The output shaft rotates at 6000 rpm and delivers
approximately 1500 hp into a speed reducing gearbox, the output shaft
of which rotates at 1800 rpm.
Power Generator
An Electric Machinery Company Model BEMAC II electric power generator
is coupled to the turbine gearbox output shaft. The generator is a
3-phase, 60 HZ alternating current machine rated 1250 kW at 480 volts.
Although its maximum rating is 1250 kW, it generates a nominal 1000 kH
in the pilot plant.
Switch Gear
The switchgear used to monitor the generator output and to couple the
generator to the load and'lamp banks is standard equipment supplied by
Mid-Continent Controls. It is housed in a metal enclosure and consists
of an a.c. ammeter, a voltmeter, a wattmeter, and a frequency meter for
indication of power generating conditions. Also included is a manually
operated 3 pole air circuit breaker with a 1600 ampere, 600 volt rated
capacity and a 50 kilo ampere interrupting capacity. Equipment for
6-7
-------
ground detection and generator winding temperatures are also incorpo-
rated. The switchgear, as installed in the pilot plant facility, is
shown in Figure 6-3.
Load Banks
A resistance load bank and a lamp load bank are used in dissipating
the generator output power.
Resistance Load Bank-
The resistance load bank is a forced air-cooled unit with a blower
supplying air to a plenum chamber containing the resistors. The
assembly is installed indoors and is provided with a hood for exhaust-
ing air to the outdoors through the building roof. The load bank is
rated at 1150 kW for dissipating the 1000 kW working output of the
generator at unity power factor. The resistors are controlled indi-
vidually by means of remotely operated contactors which provide load
selection in fourteen 75 kW coarse steps and the balance in four 25 kW
fine steps.
Lamp Load Bank-
The lamp load bank consists of two wall mounted units, each containing
eight lamps. Lamp spacing is based on ambient cooling with a maximum
power rating of 1.5 kW for each lamp. However, lamp ratings can be
reduced to a minimum of 500 watts. Each light bank is controlled
remotely by a single contactor.
6-8
-------
Figure 6-3 Switch Gear and Load Bank
6-9
-------
Piping and Control Desjq_n Consjjte_raj:jons
The thermal energy source for driving the turbine is obtained frnm *
parallel gas flow system Mtat contains a diesel fuel burning rombustor
in one path and a solid waste burning conibustor followed by three
stages of particle separators in the other flow path. Two basic prob-
lems inherent with the unique design of this subsystem were: 1) the
design of the interconnecting piping tc handle hot gases at tempera-
tun ; up to 1500° F while allowing for linear thermal expansion without
develoDing excess loads on the turbine flanges; and 2) the selection
o*: control methods and operating techniques tn start ind operate the
turbo-electric system with A large volume, optional combustor energy
source between the compressor and turbine. The follcwint) paragraphs
discuss the design philosophy used in solving these problems.
Hot and Cold Piping Design-
The terms hot" and "cold" are used in reference to the qas tempera-
tures within the piping runs between the combu1 f.ors and the turbine
inlet and the piping runs between compressor outlet and combustor
inlets, r siectively. The hot and cold piping runs aie depicted in
Figures 6-4 and 6-5.
Hot Piping Design - Pressure piping is being protected from high tem-
perature gases by internally insulating with Kaowool and lining the hot
gas pipinn with heat resistant stainless steel. Due to this insulation
the hot gas piping metal skin temperatures are approximately 400° F.
6-10
-------
Figure 6-4 Compressor Line Piping with Valves to the Fluid Bed Combustor
-------
en
i
DETAIL A
ROTATED 90°
1 FLUID BED COMBUSTOR
2 1ST STAGE PARTICLE SEPARATOR
3 2ND STAGE PARTICLE SEPARATOR
4 TEE
5 fcXPANSION JOINT ASSEMBLY
6 3rd STAGE PARTICLE SEPARATOR
10)
7 TURBINE INLET
8 COMPRESSOR OUTLET
9 ELBOW A'lD E^A'ISIfl'l BELLOWS
10 AID INLET VALVE (003)
11 AIP INLET VALVE (012)
12 PIJSTON r
Figure 6-5 Hot and Cold Pressure Piping Runs
-------
Linear expansion of the piping is acconvnodated by the use of bellows
expansion joint assemblies where required to prevent excessive turbine
flange loading. All the expansion joints used are tied with stress
rods so they do not expand axially, but do allow a lateral motion of
from Jj to 1 inch.
The thermal expansion between the combustor and third stage separator
results in a lateral deflection of the expansion joint downstream of
the separators. The thermal expansion between the separators and the
turbine results in a lateral deflection of the expansion joint between
the second and third stage separators. To accommodate these deflec-
tions, the particle separators are mounted on casters. Thus, they are
fixed in the vertical plane, but are flexible in the horizontal plane.
Cold Piping Design - The cold piping skin temperature can never exceed
the 400 F compressor outlet temperature; therefore, insulating the
piping was not required, and ordinary carbon s«.eel pipes were used.
However, the cold piping design was complicated by three factors:
(1) the angular displacement of the horizontal piping runs needed to
accommodate the Ruston combustor; (2) the two butterfly valves re-
quired to control air flow to the combustors; and (3) the difference
in length between the vertical piping at the fluid bed combustor inlet
and the verticle piping at the Ruston combustor and turbine compressor.
The expansion design philosophy is identical to that of the hot gas
piping 1f one considers the cold piping joined at the tee over the
Ruston combustor as being made up of three l-shaped piping runs.
-------
Each piping run consists of one horizontal leg and one vertical leg
and the necessary expansion joint assemblies reqiired to prevent
excessive flange loading. As in the hot gas piping all expansion
joints are tied with stress rods so that they do not expand axially
but do allow lateral motion of from S to 1 inch. Thus, the thermal
growth in one direction is absorbed by lateral deflections within
piping at right angles to the growth direction. These three piping
runs are: (1) the piping run between the tee inlet flange and the
turbine compressor outlet flange; (2) the piping run from the tee
down to the third stage particle separator inlet via the Ruston corn-
bus tor; and (3) the piping run between the tee outlet flange and the
fluid bed combustor inlet flange.
1. The piping run from the tee to the compressor outlet (see Figure
6-4). The horizontal run is mated to an expansion bellows and
elbow which is mated to an expansion joint assembly connected to
the compressor outlet flange. The thermal expansion in the hori-
zontal run results in lateral deflection of the expansion joint
in the vertical run mated to the compressor outlet flange. The
corresponding expansion in the vertical leg results in the tee
being raised approximately ^-inch.
2. The piping run from the tee to the third stage particle separator
(see detail A, Figure 6-5). The vertical piping in this piping
run consists of the tee outlet flange, the flow control valve and
6-14
-------
expansion joint assembly, a transition piece, and the Ruston com-
bustor inlet flange. The vertical expansion in this leg is essen-
tially equal to the vertical expansion in the vertical leg mated
to the compressor outlet flange. Therefore, the horizontal cen-
terline of the Ruston combustor is at a constant elevation above
the floor.
3. The piping from the tee to the fluid bed combustor. (See Figure
6-5). The horizontal piping run from the tee outlet and the ver-
tical piping run from the fluid bed combustor inlet are mated to
a curved elbow by two expansion joint assemblies. The expansion
in the horizontal run results in lateral deflection of the expan-
sion joint in the vertical leg mated to the fluid bed combustor
inlet. The differential expansion caused by the difference in
length between this vertical leg and the vertical legs at the com-
pressor and Ruston combustor result in lateral deflection of the
expansion joint in the horizontal leg.
Control Considerations-
Control of the Ruston gas turbine is centered around utilizing the
existing controls for the turbine and modifying them as required for
compatability with the solid waste feed system, solid waste combustor,
separators, generator, and load bank. Proportioning of fuel flow into
the fluid bed combustor to vary the turbine inlet temperature is used
to control power output. Thus variation of the fuel feed rate deter-
mines the turbine power output. Computer control of the system
will be Introduced as the system is developed.
6-15
-------
Overspeed Protection - One speed stabilizing and overspeed protection
system has been incorporated into the turbo-electric system. Because
the pilot plant combustor system contains a large quantity of pres-
surized hot gas when operating, release of this gas through the tur-
bine can result in very high overspeed conditions for the power tur-
bine generator shaft in the event of an electric load shed. Unre-
strained, the resulting high rotary speeds could cause severe physical
damage, if not destruction of the generator. To prevent such over-
speeds, a disc flywheel inertia, with a dual redundant caliper braking
system to apply restraining torque to the flywheel has been added to
the generator coupling.
The brake application techniques use compressor duct bleed air as the
energy source, and existing Ruston power turbine overspeed control
sensing elements for activation. Thus, the gas pressure source that
provides the turbine input energy is used as the power source for the
brakes. This allows the brakes to be automat'cally released as the
compressor speed drops since the pressure is reduced with lower com-
pressor speeds. The brakes are therefore released at lower speeds
when their operation is not-required for maintaining speed limits.
The braking systems (Figure 6-6) are activated at turbine overspeed
conditions of three percent and ten percent.
Three percent overspeed circuit - The three percent overspeed circuit
uses the Ruston turbine governor as the speed sensor. If the turbine
6-16
-------
M
^ CALIPtR °,
-------
speed increases to a level greater than three percont over the gover-
nor setpoint, two redundant microswitches are actuated. These micro-
switches in turn energize three solenoid valves. Two parallel redun-
dant valves exhaust the compressor air that holds the dump valves
closed. The dumo valves then open and exhaust a portion of the com-
pressor output flow, thereby reducing the energy available to the tur-
bine. The third solenoid opens compressor air to one pair of disc
brakes on the edge of the turbine inertia disc, thereby absorbing
energy. The microswitches are released as soon as the speed drops to
a value less than three percent over the governor setpoint. This in
turn allows the dump valves to close and the disc brakes to be released.
Ten percent overspeed circuit - A ten percent overspeed circuit uses
an unbalanced rotating weight in the gear box as the speed sensor.
If the turbine speed exceeds ten percent of the design speed for the
system (independent of any setpoint), the rotating weight shifts off
center and trips a valve, causing the compressor air holding the dump
valves closed to be vented and also causing servo pressure to be ex-
hausted. The venting of the air to the dump valves allows them to
open, releasing a portion of compressor air as in the case of the
three percent overspeed operation (note that the three percent over-
speed switch should have already caused the dump valves to be opened,
thereby making the operation redundant). The loss of servo pressure
causes the two other energy absorbing systems to come into operation.
One is a set of disc brakes (besides the three percent overspeed disc
brake units) that are energized by compressor air but restrained from
6-18
-------
actuation by servo oil pressure. The second system brought into opera-
tion is the load bank. Loss of servo pressure causes pressure switches
to be actuated which in turn energize all 75 kW resistors in the load
bank, thereby loading down the generator.
The valve that is tripped by the ten percent overs peed is not auto-
matically reset, but requires manual resetting. The system will
therefore be completely shut down and require a complete restart to
come back on-line following a ten percent overspeed condition.
SURVEY OF CANDIDATE TURBINES FOR THE CPU-400 PROTOTYPE PLANT
/
Since the development of a new gas turbine and associated hardware would
compound the operating risks and inflate the prototype cost, the tur-
bine survey was limited to existing industrial gas turbines with proven
operating reliability and design characteristics. Nine candidate tur-
bines sized in power output from 2 to 20 megawatts were evaluated for
performance with a solid waste combustion system, and for costs of
purchasing, installation, and operation. The 2 and 20 megawatt size
extremes corresponded to solid waste throughput rates of 120 to 1200
tons per day, respectively, and were based on an 80 percent utilization
factor. Thus, allowing 6 pounds per day of solid waste per person,
the candidate turbines covered a nearly tenfold capacity range that
would handle solid waste from municipalities with corresponding
populations of approximately 40,000 to 400,000 persons. Both perform-
ance and cost data resulting from this evaluation are shown in Table
6-1.
6-19
-------
Table 6-1. TURBO-ELECTRIC SYSTEM CANDIDATE SUMMARY
i
iv>
O
Parameter
Sytteai rating (•»)
Solid Midi flow (Ib/hr)
• (ton/day)
Specific consumption (Ib/kU-hr)
Turbo-generator price
Toul Instilled cost
Capital cost per W ($/kW)
Annual operating costs ($/yr)
Periodic maintenance cost
(one* **ch 3 years)
Electric power annual revenue
Annual saving for co«pr. potter
ToUl annual return
Manufacturer / Hoitels
General Electric
G3142
7.69
43.714
420
S.69
(1.315.000
1.8SC.US
241.46
130.164
65. 750
376.810
655.700
Jl. 032.010
65251
14.51
86.150
827
5.94
Jl. 780 .000
2.503.29?
172.52
142.092
89.000
710.990
1.352.843
$2.061.833
G5341
17.74
97.890
940
5.53
12.160. 000
2,996.126
166.89
150.275
108.000
869.260
2.006.464
I2.87S.724
Solar
Centaur
2.05
12.950
124
6.32
t 325.000
524.433
255.82
109.610
16.250
100.450
216.911
$ 317.361
West In-house
W41G
3.74
24.736
237
6.62
t 825.000
1.189.4C2
318.04
119.3M
41.250
183.260
268. ie:
% 4S2.049
HlOt..
a. 98
SO. 814
4M
S.66
$1.000.000
1.471.114
164.04
12S.652
50.000
440.020
717.631
J1.157.6S1
HI 9 11,
15.39
99.970
960
4.SO
$1.200.000
1.791.328
116.40
112.871
60.000
754.110
1.453. Old
t2.207.l2l
Turbodyne
1A
3.91
31.530
322
8.57
$ 824,300
1.166.669
298.38
119.C39
41.215
191.590
330. 2M
$ 521.850
98
19.48
124.240
1.190
6.39
Sl.873.000
2.60J.20e
U3.48
US, 750
93,650
954.520
1.939. S90
S2.944.410
-------
The system ratings given in the table are as-installed with allow-
ances for installation effects. The turbo-generator prices shown were
provided by the manufacturers as estimated values. The total instal-
lation costs were obtained by estimating modification and installation
costs, then adding this to the purchase price. All other costs and
revenues are estimated values without firm quotations or detailed
designs to support them; however, they are felt to be reasonable
and conservative values. Operating costs and revenues that are
functions of system running time are based on 7,000 hours or 80 per-
cent utilization per year.
All of the systems covered by the Table 6-1 data involve one turbine
of the model indicated at the top of each column. This study ex-
plored the performance and costs associated with a broad rang" of
capacities from 2 to 20 megawatts in single turbine systems only.
Because of this limitation, two factors tend to cause the lower
capacity systems to be more expensive to buy and operate on a per
kilowatt or per solid waste ton basis. The first factor is the
economics of scale in manufacture of the larger gas turbines. As
the data shows, the capital cost per kilowatt capacity is lower for
the larger machines. It varies from a low of $116.40 per kilowatt
for the 15.39 megawatt W191G turbine to a high of $318.04 per kilo-
watt for the 3.74 megawatt W41G turbine. The second factor is that
the operating labor costs are treated as being the same for each
system, since it is likely that the same number of operating
personnel will be required for around-the-clock, 7 days a week
6-21
-------
operation of each independent system, regardless of its capacity.
This imposes a rather severe but realistic penalty on a single
turbine system under 5 megawatts in capacity. Thus, the net annual
dollar return for turbo-electric systems under about 300 tons per day
combustion capacity reduces to negative dollars (costs) or at best
zero return.
The data indicated one significant general conclusion: to obtain the
greatest dollar benefits from the turbo-electric system using a single
turbine unit, the solid waste combustion capacity should be sized
greater than about 500 tons per day. Sizing the plant greater than
about 700 tons per day combustion capacity provides decreasing re-
turns for the turbo-electric system. This conclusion applies only to
the turbo-electric subsystem and does not Incorporate cost and return
trends associated with the solid waste handling or solid waste com-
bustion subsystems involved in a total CPU-400 system. Therefore, the
optimum sizing of a total plant could shift away from the capacities ,
mentioned above due to the other subsystem effects. However, since
the scope of this study was limited to evaluating turbo-electric
systems only, the analysis of the other CPU-400 major subsystem
effects on total system costs and returns were not included. Such
considerations are appropriate to a more comprehensive evaluation
document or to those design studies and proposals dealing with a
unique installation where requirements are more firmly established.
6-22
-------
It should also be pointed out that CPU-400 systems need not te re-
stricted to single turbine units for the larger capacity installations.
Considerations of operating flexibility, reliability, maintenance
ease, labor cost reduction, and add-on plant expansion capability may
favor the use of several smaller turbine modules rather than one large
unit. For example, the use of three or four Centaur turbines *ith
reasonable capital costs per kilowatt capacity could provide a 400 to
600 ton combustion capacity plant competitive with a single large unit
installation; and provide greater operating flexibility and potentially
better reliability.
Further system analysis, performed after this study was completed, re-
sulted in a significant uprating of the solid waste throughput and
power generation capabilities of the CPU-400 systems by also burning
nondewatered sewage sludge. This uprating, of course, is associated
with the increased mass flow going into the turbine on which no com-
pression work was expended. For an existing turbine this uprating is
possible only 1f the area of the turbine nozzle car be increased to
compensate for the increased mass flow through the turbine without
exceeding the compression ratio design range.
It was established th*c the nozzle area of the Solar engine could be
easily modified for an increase of 12.5 percent. This would result in
a power increase to 3 megawatts, its rated level, an
-------
combustor, and its air flow rate matches the capacity of the fluid bed
combustor now being tested in the pilot plant. Therefore, it is
readily adaptable into the pilot plant system. Also, since the pilot
plant hot gas system combustor and particle separators are compatible
1n size to the units proposed for the prototype plant, the use of the
Solar engine would mlnirize the design risks in scaling from pilot
plant hardware to prototype hardware.
6-24
-------
SECTION VIi
REFERENCES
1. Staff members of Office of Solid Waste Management Programs,
Environmental Protection Agnecy. Resource Conservation,
Environmental, and Solid Waste Management Issues. In: Second
Report to Congress, Resource Recovery and Source Reduction. U.S.
Environmental Protection Agency, Washington D.C. Publication
Number SW-122. p. 1-17 March 1974.
2. Atkin, M.L. Operation of an Open Cycle Gas Turbine Burning
Pulverized Bituminous Coal. Department of Supply, Australian
Defense Supply Services, Aeronautical Research Laboratories
(Melbourne). ARL/ME 118: 30 p. July 1967.
3. Atkin, M. L., and G. A. Duke. The Operation of a Ruston and
Homsby Gas Turbine on Queensland Coal. Department of Supply,
Australian Defense Supply Services, Aeronautical Research
Laboratories (Melbourne). ARL/ME 131: 32 p. Aoril 1971.
4. Kaiser, E. R., C. 0. Zeit, and J. B. McCaffery. Municipal
Incinerator Refuse and Residue. In: Proceedings of 1968
National Incinerator Conference, American Society of Mechanical
Engineers (ed.). New York, American Society of Mechanical
Engineers, United Engineering Center. 145 p. May 1968.
5. Fleischer, L. R. Investigation of Corrosion-Deposition.
Phenomena on Gas Turbine Blades. (Final Report, EPA Contract
68-03-0049). Westinghouse Research Laboratories. Pittsburgh,
Pennsylvania. 28 p. June 1972.
7-1
-------
SECTION VIII
APPENDICES
Page
A. Solid Waste Handling System 8-2
B. Gas Composition and Particle Analyses of Model
Fluid Bed Combustor Exhaust Gas 8-31
C.. Gas Composition and Particle Analyses of Exhaust
Gas from the Full Scale Vertical Com!>ustor 8-47
8-1
-------
APPENDIX A
SOLID WASTE HANDLING SYSTEM
INTRODUCTION
The solid waste handling system configuration used in the CPU-400 pilot
plant during the contract period 30 June 1971 to 31 January 1973 was,
for the most part, designed during a previous contract. However, the
solid waste handling system did undergo some changes during this period
in order to provide support for the horizontal combustor testing per-
formed during the LP-1 test series and the vertical combustor testing
performed during the LP-8 and LP-11 test series. Most significant of
the changes were the incorporation of large 30-inch diameter feeder
valves, the improvement of air classifier efficiency, and the design
and incorporation of an automated feed control system, thus integrating
the entire solid waste feed system with the combustor and pilot plant
control system.
In general, all elements of the solid waste system performed adequately
to support test objectives. Changes brought about by combustor oper-
ating requirements and variations in interface equipment and process
variables were accomplished such that in final configuration, predicta-
ble performance capability had been achieved.
8-2
-------
The solid waste handling system as configured to support the pilot plant
low pressure test program is composed of two major subsystems - a
shredder/air classification/pneumatic transfer subsystem and an Atlas
storage/combustor feed subsystem. A diagram of the solid waste handling
system used during the LP-1 and LP-8 testing series is shown in Figure
8-1. Figure.8-2 depicts the solid waste handling system used during
the LP-11 testing series.
SHREDDER/AIR CLASSIFICATION/PNEUMATIC TRANSFER SUBSYSTEM
This subsystem consists of equipment and facilities designed to accept
and process "as received" solid waste from packer trucks, separate
combustibles from non-combustibles, and transfer the combustible fuel
fraction to a storage facility for subsequent burning.
Shredder Input Conveyor
Municipal and residential waste delivered by local packer trucks is
dumped on the facility floor adjacent to the conveyor in an area capable
of accommodating 2 to 3 loads (20,000 to 30,000 pounds). By means of
an M-60 Bobcat skip loader with a 15.6 cubic foot bucket, the material
is transferred to the hopper of the shredder conveyor to an even depth
of approximately 2 feet deep by 4 feet wide. The conveyor assembly con-
sists of a 10 gauge steel and angle fabricated hopper, and a Prab steel
belt conveyor with a straight incline of 20 degrees. The belt is 48
inches wide, 3/16-inch thick, with a 6-inch pitch and 4-inch high
traveling side wings. Pusher flights 4 inches high located on 48-inch
centers were removed when experience indicated no advantage was gained
and the flight Induced fallout of garbage onto the floor. The drive
8-3
-------
oo
PACKER TRUCK
STOCK PILE
LOADER
EIDAL SHREDDER WITH
ANTI-BALLISTIC HOOD \
|t
BLOWER-^ ;
AIR CLASSIFIER
DUST KOP
ill"! i 11 I I i
SOLID WASTE SYSTEM
CONTROL PANEL
NOTE: FOP LP-8 TESTI'jr, -
(a) FEED CHUTF PI PLACED R- MAD SPL1TTEP
(h) SIMPLE FEEOcP VAL Vf AY,' t'.-l'.CH f : i i' LlNfS
PLPLAflDB» TWi' . AL.'i1- A',; :-i'(1H I i ! [) I 1'I! ^
CYCLONE
UF.ir.nrritj CONVEYOR
' TRANSFER
; , v A:
ATLAS CONVEYOR
I'ANfl
IA-?2|-368
Figure 8-1 Solid Waste Handling System Schematic (LP-1 and LP-8 Testing)
-------
P.TEL.'^lIC
TRA.'.S?:3T
CONTROL PANEL
DUST RET'JR',-,.,
CFF^fs '?\ \
rc.L-.wr, ^^, \
BLOWER (IK
AIR CLASSIFIES <-
SPLITTER,
00
I
en
PACKER TRUCK
CO\'.:t3R CONVErTR I !
"*-\ INSTR'J-E'.TATION^ *:'-'_':'
EIDAL '
SHREDDER
\ (
OUTFEED XR£JEC* F=.::T:::J
ccr,;E:.:R (i) CO.-..EI:? (i:
'DEERIS
BOX
V-..-I;. 2:
221-r::
Figure 8-2 Solid Waste Handling System Schematic (LP-11 Testing)
-------
arrangement has a 3 hp, 230/460 volt, 3-phase, 60 cycle motor with vari-
able speeds from 3 to 18 fpm which can be manually set and then elec-
trically controlled for on/off condition automatically from the control
panel by demand current from the shredder. This conveyor dumps solid
waste directly into the top of the Eidal shredder, and can easily
deliver the waste material at a rate of 2 tons per hour.
Shredder
The Eidal Model 75 mini-mill shredder (Figure 8-3) is a patented belt-
driven vertical rotor shredder powered by a 75 hp motor. The rotor
shaft has double bearings on its lower end, thus it can receive feed
over the entire top opening of the shell. The shredder, without feed
chute, is approximately 4 feet 6 inches high, 7 feet long, and 4 feet
11 inches wide. The top opening of the shell is 4 feet 4 inches in
diameter. An anti-ballistic hood fabricated by CPC encloses the top of
the shredder.
The weight of the complete mini-mi 11 without feed chute or feed conveyor
is approximately 14,000 pounds with the weight of the rotor at 1800
pounds and the fly wheel weight at 2800 pounds. The unit is equipped
with pre-1 ubricated antifriction bearings, a choker bar to assure a
finer grind, a blower and enclosure with filters built around the motor
to keep it from picking up dirt and small debris, and a quench water
deluge and trickle water source for safety and grind control.
The 75 hp drive motor is of open drip-proof construction suitable for
vertical operation. Power requirement is 220/440 6J Hz, 3-phase.
8-6
-------
Figure 8-3 Shrtddtr and Air Classifier (LP-1 and LP-8)
8-7
-------
During operation, material dumped Into the top of the shredder 1s Ini-
tially smashed oy the top breaker bar and then gradually drops down
through a tapered barrel past gear style grinders, becoming smaller
and smaller until 1t passes the choker bar. During the LP-1 and LP-8
testing the shredded waste was then ejected into an exhaust chute,
which formed the Interface with the air classifier. During the LP-11
testing the shredded material was ejected onto a conveyor wh'ch then
carried 1t to the air classifier.
Shredder Outfeed Conveyor
The shredder outfeed conveyor added for the LP-11 testing series is a
cleated slider belt conveyor driven by a 3/4 horse power constant speed
motor. This conveyor has a straight i icline of 30 degrees and is
mounted on a 2-foot by 13-foot steel frame. It is equipped with two
paddle-type load levelers that maintain a constant level of approxi-
mately 3 inches on the conveyor belt. This conveyor is capable of
transporting the shredded material at a rate of 2 tons per .hour.
A1r Classifier
The air classifier seoarates light combustibles from the heavier metal
fractions of shredded waste. Material ejected from the shredder during
operation 1s drawn up a zig-zag path with increasing velocities. Light
material Including paper, cardboard, light plastics, textile fibers,
dirt, etc., are entrained In the main stream and delivered to the pneu-
matic transfer system. During the LP-1 and LP-8 testing series the
heavier materials were allowed to drop out the bottom of the air classi-
fier. The desired cut of material was controlled by adjusting blower
8-8
-------
ppm. Shop air jets were attached to the exhaust chute of the shredder
to allow rapid cleaning of the chute in case a slug of particularly wet
material was encountered, which could create temporary blockage of flow
Into the air classifier. During the LP-11 test series the heavier
materials were dropped out onto a reject fraction conveyor for transfer
to a debris box..
The air classifier used during the LP-1 and LP-8 testing series (Figure
S-'') was a CPC designed and fabricated sheet steel device. The unit
was rectangular in cross section with outer dimensions approximately
3 feet wide by 2 feet deep by 5 feet high. The Input section was close-
coupled to the shredder exhaust chute and the output was on top, coupled
to a material transfer pressure blower which induces an air flow of from
4000 to 5000 cfm.
While the operation of this air classifier was satisfactory during the
LP-1 test series, the efficiency of this unit during the LP-8 test
series was erratic. It was found to be sensitive to variations In solid
waste moisture content and composition. This marginal operation of the
air classifier led to the design, development and test of a small, con-
veyor-fed, zig-zag wooden unit followed by a second larger wooden unit
(Figure 8-4) which subsequently became operational fcr the LP-11 tests.
This unit Incorporated movable zig-zag sections which permitted adjust-
ment of the air flow throat width to provide control of separation effi-
ciency for various loading, moisture content and air flow combinations.
The overall dimensions of this unit were 16 Inches in depth, 30 Inches
In width, and 96 Inches In height. Shredded solid waste was introduced
8-9
-------
8-4
Air Cl«s1f1er
8-10
-------
Into the upper end via the 2-foot wide shredder outfeed conveyor which
1n turn was fed directly from the shredder exhaust. As previously
mentioned, the reject material from the bottom of the air classifier
was transferred to a debris box via the small reject fraction conveyor.
The light fuel fraction was exhausted to the Atlas via a TVhp Reese
materials handling blower.
Reject Fraction Conveyor
The reject fraction conveyor added to the air classifier for the LP-11
testing series is a cleated slider belt conveyor driven by a Vhp con-
stant speed motor. This conveyor has a straight Incline of 30 degrees
and is mounted on an 8-inch by 11-foot steel frame. It has the capa-
bility of transporting material at a rate of 600 pounds per hour.
Pneumatic Transfer Components
A materials handling blower provides the motive force for pneumatically
conveying shredded waste fuel from the air classifier Into the Atlas
storage unit. A dust control system provides a final air filter before
exhausting system pressure to atmosphere.
Blower. Pneumatic Transfer-
The blower is a Reese 11 HE steel plate exhauster, 1150 cfm at 16 inches
of water gauge complete with 7S hp, 460 volt, 60 Hz. 3-phase TEFC motor,
V belt variable speed drive.
Shredded material enters the air Intake duct and 1s driven by the fan
blades Into a 12-Inch diameter duct connecting to the cyclone located
on top of the Atlas storage facility.
8-11
-------
Cyclone (Atlas Storage Unit)-
The Atlas storage unit cyclone is located at the top center of the Atlas
storage unit, suspended from a bridge structure. The cyclone is fabri-
cated of 10 gauge steel; 44 inches in diameter and 60 inches high with
a 24-inch diameter inner chamber, 36 inches high. The i-iner chamber is
fitted at the top with a 24-inch to 10-inch transition adapter which
connects to a 10-inch round duct leading to the blower intake of the
dust kop cyclone. The storage room cyclone's function is to decelerate
and distribute or separate out the shredded waste into the Atlas storage
unit. Concurrently, dust and explosive vapors are drawn up through the
inner chamber and out to the dust cyclone.
Dust Kop Cyclone-
,". single cyclone separator rated at 1200 to 2800 cfm at 10.7 inches of
water is located on an elevated (18 feet) platform adjacent to the
Atlas storage unit. The unit is supplied by the AGET Manufacturing
Company and comprises a series 3000 blower with 10-inch inlet driven
by a 7H-hp, 230/460, 3-phase, 60 Hz motor turning at 3600 rpm. Its
function is to process dust-laden air from the Atlas storage unit by a
principle in which the centrifuging effect of the cyclone, created by
rapid rotation of the high velocity airstream, forcibly precipitates
dust out of the bottom and into a sealed drum via a 7-inch diameter
flexible hose. Cleaned exhaust air then exhausts through the center
duct of the cyclone to atmosphere. For LP-8 and LP-11, the dust col-
lected was returned to the Atlas floor via appropriate ducting and a
rotary air lock valve.
8-12
-------
.. A-J. ^.< • —^«fc^fcr««j
-------
ATLAS STORAGE/COMBUSTOR FEED SUBSYSTEM
This subsystem provides for the storage of shredded sol id waste and an
automatic retrieval or outfeed system designed to deliver the desired
volume and weight of fuel to satisfy fluid bed combustor demand.
Atlas Storage Unit
This unit is a cylindrical structure 16 feet in diameter and 20 feet
tall with a capacity for storing 3400 cubic feet of shredded waste.
Recovery rate capability is variable from 100 to 1000 cubic feet/hour.
Recovery is accomplished by four chains of sweep buckets located on the
floor of the cylindrical bin. Each sweep chain is fixed at one end to
a powered rotating "pull ring" encircling the storage area, the other
ends being free or trailing. As the pull ring rotates around the
periphery of the bin, the sweep chains trail toward the center, con-
tacting the stored material at the outside of the pile, filling the
buckets which sweep material onto a single outfeed drag conveyor re-
cessed in the floor (Figure 8-6). At the output end of the outfeed
conveyor, transfer conveyors then move the material to the combustor
feeder valves.
Sweep Conveyor DHve-
The sweep conveyor system is powered by a hydraulic drive package con-
sisting of the Houdallle fixed displacement hydraulic motor, and one
power unit consisting of a Vlckers variable displacement piston pump
coupled to a TVhp, 1800 rpm, TEFC, a.c. tectrlc motor. During the
LP-8 testing, the efficiency of sweeps was improved by adding an extra
8-15
-------
SHREDDED. WR CLASSIFIED SOLID WASTE FUEL
Figure 8-6 Photographic Views of the Atlas Storage Unit Interior
8-16
-------
bucket to each chain and incorporating digger spikes,, and the drcum-
cyclone Inlet, which produced unsymmetrlcal distribution and localized
compacting of fuel, wat replaced by a flat, splatter plate which evenly
distributed fuel over the Atlas floor. Control 1s manual at a remote
control panel, or automatic through central control room circuitry.
Outfeed Conveyor Drive-
This unit 1s powered by a hydraulic drive package consisting of one
Houdaille fixed displacement hydraulic motor, and one power unit con-
sisting of a Vlckers variable displacement piston pump coupled to a
5 hp, 1800 rpm, TEFC, a c electric motor. Control 1s manual at a
remote control panel or automatic through central control room circuitry.
Output speeds range from 13.5 rpm to zero. For the LP-8 and LP-11
testing series a rotating clod buster was located at the output end,of
the conveyor to break up lumps of packed or composted waste.
Control System (AVTEK)
Several modes of operation are available to provide the flexibility re-
quired to support large combustor operations or to satisfy lower volume
requirements of model combustor tests. Briefly, these modes include:
Mode I Non-production - Maintenance
Mode II Manual Speed Control - Direct manual control of sweeo
drive and outfeed conveyor
Mode III Manual Volume Demand with Automatic Volume Control -
Manually set to desired volume output and then auto-
matically maintained
8-17
-------
Mode IV Process Demand with Automatic Volume Control - The
"process demand" signal for volume output is supplied
from the fluid combustor operational control source
(i.e., fluid bed and exhaust temperature)
Constant volume output is controlled through a paddle sensor located on
the output end of the outfeed conveyor trench (Figure 8-7). When
material level is low in the trench, the paddle sensor signals the
volume control system which speeds up the sweeps. As this deficiency
is corrected, the sweeps slow down to provide the proper material height
in the trench. If, during operation, the stored waste free-flows to
fill the trench completely, this condition will stop the sweep system.
However, if free-flowing continues, and the trench remains full, the
volume outfeed rate will be higher than required. It then becomes
necessary to slow the conveyor down to maintain the required volume
output. (This is done automatically). As soon as free-flow stops, the
system will return to normal operation.
Instrumentation provided to monitor Atlas operation includes: a pres-
sure Indicator, located on side wall of Atlas bin to measure internal
pressure In Inches of water; a demand load Indicator, demand volume of
solid waste required expressed 1n percentage of feed rate to the com-
bustor; and an actual load Indicator, measured output delivered, ex-
pressed in percentage feed rate.
Transfer and Weigh Conveyors
The transfer and weight conveyors (Figure 8-2) carry the waste material
8-18
-------
Figure 8-7 Atlas Outfeed Volume Sensor
8-19
-------
stored in the Atlas to the feeder valve inlets. During the LP-1 testing
series these units were manually controlled from a remote control panel
(conveyor motor control panel) located adjacent to the AVTEK. At this
time the transfer and weigh conveyors were operated at a constant speed
and no synchronization y ;~he speeds of these conveyors with the Atlas
sweep speed, feed rate, or the feed volume signal from the pilot plant
control room was exercised. During the LP-8 and LP-11 testing, the
speed of the transfer and weigh conveyor were controlled by a common
variable frequency controller and the ratio of the speeds of these units
was then constant. By using the solid waste volume demand signal from
the control room as the control signal for the variable frequency unit,
the transfer and weigh conveyor speeds became a direct function of the
required volume. Therefore, since the commanded volume signal and the
level and paddle height signals control the speed of the Atlas outfeed
conveyor, a constant level of material is maintained on the transfer
and weigh conveyors as a function of the commanded volume signal and the
Atlas outfeed conveyor control loop.
Transfer Conveyor-
Material leaving the Atlas outfeed conveyor trench is picked up on a 29
degree inclined conveyor for transfer to a weighing conveyor. The trans-
fer conveyor is a 22-1nch wide by 22-feet long belt with 14-inch cleats
on 16-inch centers. It is driven by a Il2-hp gear motor and chain drive.
Rotating paddle-like levelers were added for LP-8 and LP-11 to insure
constant level of material on the conveyor for more accurate feed rate
control.
8-20
-------
Weigh Conveyor-
Thls portable unit is installed between the Atlas transfer conveyor and
the combustor feed valve(s) to provide automatic readout and totalized
weight of material being fed to the combustor. It is a self-contained,
3600 pounds per hour unit composed of the following standard equipment:
1. 20-inch wide by 12-foot long rubber belt feeder with 1-inch
high flange and full length skirt boards.
2. Head pulley drive utilizing a reducer and drive chain driven
by a 1 hps a.c. squirrel cage reversible, induction motor
operation on 230/460 volt, 3-phase, 60 cycles.
3. Merrick Model 440 weightometer including: multiple idler sus-
pension; temperature compensated load cell; solid state milli-
volt to volt amplifier including output indicator and load cell
excitation power supply.
4. Punning time meter with thumbwheel reset.
5. Silid state mv to v integrator including: rate integrator;
seven digit totalizer; zero and span adjustments; test button;
impulse transmitter.
The model 44C w^yhtometer is basically a mechanical suspension system
which transmits tne weight on the conveyor idlers to an electronic load
cell. The millivolt output from the load cell is proportional to the
material loading on the conveyor per lineal foot of belt. The millivolt
signal can be amplified directly or combined with a speed measurement
and amplified to produce the desired rate signal. If the conveyor speed
8-21
-------
Is constant, the weight measurement Is directly proportional to the
rate of material flow on the conveyor belt (pounds per hour).
Load Splitter/Feed Chute
For early, single feeder valve tests performed during the LP-1 test
series, a rectangular duct containing a motor-driven serrated wheel was
located between the weigh conveyor and the feeder valve to Induce sepa-
ration of Input material to each side of the split feed valve. For LP-8
and LP-11 test series, a stainless steel pant leg chute was provided
to split the load for two feeder valves. (See Figure 8-2).
Feeder Valve(s)
A single, CPC-des1gned rotary air-lock feeder valve provided the feed
mechanism to the horizontal fluid bed combustor tested during the LP-1
test series. The valve body was of welded steel construction and was
approximately 16 Inches by 30 Inches by 30 Inches overall dimensions.
It Incorporated an eight bladed, adjustable tip rotor capable of de-
livering 30 to 60 pounds per minute of shredded waste to a positive
pressure pneumatic conveyor (nominally 3 psl) at a nominal rotational
speed of 11 rpm. The original design supplied the full Input of
shredded waste to a single 6-1nch diameter feed line to the combustor
(see Figure 8-8). Later combustor design changes dictated a requirement
for two 5-1nch diameter feed lines. This was accommodated 1r a modifi-
cation of the valve with Installation of divider plates 1n each of the
rotor pockets, thus splitting the valve Into two separate sections,
each supplying an Individual 5-Inch feed line (see Figure 8-9). The
splitter unit mentioned above was accordingly added to aid 1n splitting
8-22
-------
Ftgurt 8-8 11-Inch Air-lock Pttdtr Yilvt (LM and IP-8)
8-9 Air-lock Fttdtr Vilvt, Hodlflod for Dull 5-lnch Fttd Systtm
8-23
-------
the Input before it reached the valve. The valve dividers were removed
when the system was redesigned to provide two individual low pressure
valves for feeding the vertical combustor during the LP-8 test series.
The single valve drive system utilized during the LP-1 test series in-
corporated a 7.5 hp, 1760 rpm, 230/460 volt, 60 cycle, 3-phase a.c.
motor driving a 38.4:1 gear reducer. Output of the reducer to the
valve is further reduced from 46 rpm to 11 by sprockets and an ASA No.
160, 2-inch pitch, single width chain assembly. This drive system was
replaced by two 20 hp drive/reducer units driving two feeder valves for
LP-8 and LP-11 (Figures 8-10 and 8-11).
For the LP-11 test series, two rotary air-lock feeder valves were em-
ployed in the system to transfer solid waste fuel from conventional
conveyors at ambient pressure to 6-inch pressurized pneumatic transfer
lines leading to the fluid bed combustor. The air flow through these
lines is controlled by valves 000 and 001. The ESCO Rotafeeders are
30-inch diameter, stainless steel cast valves, utilizing 13 feed pockets
with 13 axial sealing edges between the pockets. A typical 30-inch
feeder valve installation in the system is depicted in Figure 8-11.
The nominal volumetric handling capacity of each valve is 36 cubic
feet per minute when rotating at a constant 11 rpm. Each valve is
driven through a gear motor unit rated at 20 hp.
Sealing against the pressure is provided by adjustable hardened knife
edge blades; four blades on each side of the rotor are continuously
sealing against pressure at all times. High pressure air which fills
8-24
-------
Figure 8-10 IP-8 Air-lock Feeder Valve and Drive Assembly
Figure 8-11 Typical 30-Inch Air-lock Feeder
Valve and Drive Assembly. .
8-25
-------
the pockets as they pass through the pressure zone is expanded twice
from the high pressure to the low pressure slue of the valve before
final release to atmosphere. Feed air purge lines assist in blowing
solid waste from the pockets as they pass over the valve outlet. To
Insure concentricity of the valve body, it is water-jacketed and heated
to near feed air temperature. This permits closer running clearances
and reduces air loss. The rotor shaft bearings consist of graphite im-
pregnated sleeves which are cooled by a water system operating in the
0.2 gpm and 5 to 8 psig range.
Modifications were made to the feeder valves after completion of the low
pressure test series to eliminate certain anomalies encountered during
these tests. They included: rework of the shaft face seal installation
for higher preload and increased floating freedom; rework of drive mount
to eliminate chatter and associated resonances with the chain drives;
and rework of the transition section between the rotor outlet cavity
and the feed pipes.
Subsystem Control
The Atlas storage and combustor feed subsystem contains six motors and
one solenoid valve. Originally, during the LP-1 testing, only the motors
associated with the Atlas bucket sweeps and Atlas outfeed conveyor were
variable speed units. Thes? motors were controlled through the Atlas
control unit (AVTEK) and could be controlled either locally at the AVTEK,
or remotely by a commanded volume signal from Fisher controllers in the
pilot plant control room (Figure 8-12). When operated remotely, the
Atlas outfeed conveyor sp*>«d is basically a function of the commanded
8-26
-------
TR-72-4
CD
I
CO^USTOR
d£D
T/ u
EX-AJST
GAS
T/C '
r
A:A°.M
hIGM LO'.J
n n
i 1 1
MV XVITTER
"" SOJRCE " »2 •
I
TO RECCROEf
n n
MM
TV /'"ITTER
* SC.SCE " -3
1
TO PtCC-OEf
COTRCL ROOM RACK »1
)
i
CO'!T.~LLE:r<
12
1
TO RECORDER
.
CGT.TPCLLE.R
• 3
~1
P;
y
k
V 1 T I
DOLE ri
^ ,
^
EIGHT
P.L
PiC
hEI
K.
1
A
P
r .
DLE
Gr^T
( • LtU
^ tf^ ,_ ^ . ». r- r--> rr: -r-
T " f'f ' 1 r 3
/ "N S.-Lit
^X
;.-EC
Figure 8-12 LP-1 Solid Waste Metering Control
-------
volume and the Atlas sweep bucket speed 1s controlled by the command
along with the error between the desired and actual material height 1n
the outfeed conveyor.
As previously mentioned, for che LP-1 testing the transfer and weigh
conveyors were controlled manually from the conveyor control panel lo-
cated next to the AVTEK, ind were operated as constant speed units with
no connection to the main pilot plant control circuits. However, during
the LP-8 and LP-11 testing the speeds of the transfer and weigh conveyors
were controlled by a common variable frequency controller which was con-
trolled by the commanded volume signal routed to it from the input of
the AVTEK control loop. The transfer and weigh conveyor speeds were
then a direct function of the commanded volume signal, and the commanded
volume and Atlas outfeed control loop maintained a constant level of
material on these two conveyors. (See Figure 8-13).
In this control scheme the motors are automatically operated in a pre-
set sequence. Starting the feeder valve motor causes the weigh and
transfer conveyors to be started by powering the variable frequency
source. The variable frequency unit contactor then causes the outfeed
conveyor motor to be energized which in turn causes the sweep drive to
be energized. Loss of any link in the feed metering system causes the
motor upstream to be shut down. The solenoid valve operating the lubri-
cation system for the sweeps and outfeed conveyor 1s slaved to the Atlas
controls and requires no special attention.
8-28
-------
T3
I
00
Q.
•o
I
00 —.
ro
vo
CO
o
Ol
v>
re
re
ex
o
o
1
POWER INPUT
60 v A
\°52/
C
«MI
RcCTIFItR
AND
VOLTAGE
CONTROL
LOGIC
AND
CONTROL
MODULE
INVERTER
TRANSFER
CONVEYOR
MOTOR
WEIGH
OWEY0.1
MOTOR
COMMAND VOLUME
MEASURED LEVEL IN
JUTFEED CONVEYOR
n
POSITION FEEDBACK
CONVEYOR MOTOR CONTROL PANE
OUTFEED
DRIVE
SERVO
CONTROL
/051B\
^^
S«EE?
DHIVE
SERVO
CONTROL
POSIT:C:I
FEEDBACK
-------
The solid waste feed rate is controlled by equipment installed in
rack 1 in the control room. This equipment incorporates two controllers,
a digital temperature indicator, a millivolt source and associated cir-
cuitry. At rack 1 the solid waste feed can be controlled either manually
or automatically. In manual mode, the demand signal for increased or
decreased solid waste feed req"ires manual adjustment of the controllers
through use of thumbwheel settings. In the automatic mode, the desired
bed and exhaust temperatures are set into the controllers by setpoints.
The controller then compares the actual bed and exhaust temperatures to
the setpoints and then either increases or decreases the commanded vol-
ume signal based upon the difference between the measurement and the
setpoint. The setpoints can be set in locally at the controller, as was
done during the LP-8 testing, or it can be set in remotely from the com-
puter, as was done during the LP-11 testing.
8-30
-------
APPENDIX B
GAS COMPOSITION AND PARTICLE ANALYSES
OF MODEL RUID BED COMBUSTOR EXHAUST GAS
INTRODUCTION
On July 23, 1971, testing was conducted to establish the chemical
qualitative and quantitative composition of the off-gases and fumes
from a fluid bed combustor burning municipal solid waste. The results
of the test were used to establish the principal corrosive constituents
of gases and flyash and their concentrations to permit simulation of the
fluid bed exhaust during the turbine blade corrosion-deposition testing
subsequently conducted by Westinghouse Research Laboratories.
The Model Number 1 low pressure fluid bed combustor, burning shredded
and air classified municipal solid waste, was used to generate the
off-gases and flyash particulates which were analyzed by two
Independent outside laboratories.
DESCRIPTION OF THE ANALYTICAL TECHNIQUES USED FOR THE COLLECTION
AND DETERMINATION OF THE OFF-GAS CONSTITUENTS
The Trapelo West Laboratory, contracted by CPC for the outside
laboratory support, established a program for analyzing the composition
of the combustor off-gas by using mainly wet chemical techniques. In
-------
this manner, each gas constituent could be determined at least by one
specific method.
Nitrogen, Oxygen and Oxides of Carbon
The Orsat analyzer determined these gases by selective absorption and
volumetric measurement of the remaining gas volume.
Carbon Dioxide -
Carbon dioxide was measure by bringing the gas sample of known volume
Into contact with a solution of potassium hydroxide:
C02 + 2 KUri K2CO^ + H^
In the reaction, C02 disappeared from the gaseous phase and remained
as <2^3 ^n ^e liquid phase; hence, the loss in volume of the gas was
equ^l to the C02 content.
Carbon Monoxide -
Carbon monoxide was determined by bubbling the gas sample coming from
the C02 analyzer outlet through an aqueous solution of CuCl. Here the
carbon monoxide was oxidized to C02 which again was removed as <2C03
effecting another decrease of the original gas sample volume.
Oxygen -
Oxygen was determined by the reaction with pyrogallol (1. 2, 3 -
trlhydroxy benzene) under alkaline conditions. The end products of the
reaction, CO2 and organic adds, were held as potassium salts in the
absorbing solution. Since all C02 had been removed from the sample
before bringing it Into contact with the alkaline pyrogallol, the
8-32
-------
decrease in sample volume was due to the removal of oxygen.
Nitrogen -
Nitrogen was determined from the balance between original gas sample
volume and volume decrease.
Hydrogen
A gas sample collected in a 2.0 liter bottle was brought into the
thermal conductivity detector system of a Varian gas chromatograph.
Hethane
The gas mixture was spilled from the sampling bottle into the detector
system of a Varian gas chromatograph using hydrogen flame ionization
detection.
Oxides of Nitrogen
In this test the gas sample was bubbled through a jet impinger
(modified Greenberg-Smith impinger or Midget impinger) in which a
Saltzman reagent solution of known concentration reacted with N02 of
the sample changing its color to red. Then the gas was brought into a
second Impinger containing a.potassium permangenate solution to oxidize
NO to N0£ which in return could be detected in a third Impinger again
by a solution of the Saltzman reagent.
Sulfur THoxide
Sulfur trioxide was determined by the modified Shell Method. Sulfur
trioxide was removed from the sample gas stream by an Impinger
containing 80 percent isopropyl alcohol and 20 percent water. The
8-33
-------
sulfur dioxide was collected in the next two impingers in series by cm
aqueous 3 percent hydrogen peroxide solution. The impingers were keft
at 0°C by immersion in an ice water bath. After finishing the sample
collection, the solutions were combined and the sulfate concentration
of the solution in each impinger was then determined by titration with
.01 molar barium percholorate using thorin as an indicator.
Hydroger Chloride
Hydrogen chloride was collected in an impinger train containing a
standard solution of sodium hydroxide. Then the chloride was determined
by titration with silver nitrate using potassium chromate as an
indicator to determine the end point.
Hydrogen Fluoride
Hydrogen fluoride was sampled using dilute sodium hydroxide in the
impinger absorption train. For analysis perchloric acid was added to
the absorption solution and the hydrofluoric acid was separated out by
steam distillation, followed by a fluoride ion determination using the
S.P.A.D.N.S. dye method for a photometric procedure.
Ammonia
The gas stream from the sampler was sent through an impinger containing
an alkaline solution of potassium mercuric iodide I^Hglij. This
compound combined with NHj to form a yellowish-brown colloidal
dispersion whose intensity of color was directly proportional to the
amount of NH-j originally present:
2 K2HgI4 + NH3 + 3KOH —> NH2HgOHgI + KI
8-34
-------
Hydrogen Sulflde
During sampling the gas was bubbled through an" absorption mixture of an
alkaline suspension of cadmium hydroxide in the Impinger. The sulflde
1on was then reacted with a mixture of paraamlnodimethylaniline and
ferric chloride to produce methylene blue which 1s determined
colorlnetHcally.
Phosphorous Pentoxide
After Bubbling the gas through an acidic ammonium molybdatj solution,
a yellow colloidal solution of ammonium phosphomolybdate (NH4)3
P04* 12 Mo03 was produced for photometric determination of the
phosphate.
Water Vapor
The moisture content was determined by condensing the water vapor of the
exhaust gas in a set of three jet impingers immersed in an ice water
bath. The first and third Impinger were empty while the second
contained a scrubbing liquid (diluted sulfuric add, ^SCfy). The
determination was carried out as a volumetric measurement.
Partlculates
The high volume sampling technique was employed to measure the amount
(mass) of particulates In a known gas volume per time unit. As
filtering media a pre-welghed glass fiber filter with an average pore
size of 0.5 micron was used (standardized by the Bay Area Air Pollution
Control Dlstrick). After finishing the particulate collection, the
filter Mas dried and weighed again.
8-35
-------
FLUID BED MATERIAL DESCRIPTION
The fluid bed material was washed 16-mesh "Del Monte" sand. A total
weight of 483 pounds of sand was loaded Into the combus tor. The void
fraction of the sand was 0.39. Pertinent Information Including typical
particle distribution and typical chemical composition of this sand
are summarized below:
Particle Distribution:
02 larger than 6 mesh
1.6% " ' " 16 "
88.52 " " 30 "
8.5% " " 50 "
1.32 " " 100 "
0.12 smaller than 100 mesh
Chemical Composition:
S102
A12°3
CaO
MgO
K20
Na20
Fe203
T10,
86.o:%
8.17%
1.18%
0.07%
1.96%
1.67%
0.22%
0.12%
SOLID WASTE FUEL DESCRIPTION
The fuel for this test was residential solid waste collected as
normal street pickup. The truck was diverted from Its normal dump site
to the CPC facility where the delivered refuse was shredded and air
classified on July 21st. The choker bar of the Eldal shredder was set
In one-half to produce "mini-fuel" with a reduced particle size to aid
In feeding the model combustor. This finer shred grind setting
Increases the amount of Inert material ('1ne glass anc metal fragments)
which 1s carried over with the lighter materials from the air
8-36
-------
classifier. Typically, for a fine shred, the inerts will be on the
order of 15 to 20% by weight of the fuel. The moisture content after
grinding was 30%. In the two day period between the grinding and
combusting, the moisture content decreased by approximately 3%.
TEST CONDITIONS
The testing was initiated at approximately 8:40 A. M. on 23 July 1971.
As nrev>~jsly described in Section III, the testing was performed
using the CPC-designed Model Number 1 fluid bed combustor in the test
setup shown in Figures 3-60 and 3-61. The sample collection and
distribution system is shown schematically in Figure 3-62. The
combustor was operated at 1400° F with the off-gases at a nominal
1600° F in the exhaust duct. The static sand height was 2 feet, and
the superficial velocity was 5.2 feet per second. No techniques nor
additives were added to suppress any constituent of the exhaust gas.
The sampling line temperature was maintained at 360 ± 30° F throughout
the test. Pertinent sampling and combustor temperature information is
summarized in Table 8-1.
DISCUSSION OF TEST RESULTS
The laboratory test results for the off-gases and particulates have
been summarized in Tables 3-17 and 3-18 respectively. There are some
differences 1n the description of the sampling test procedure between
the Trapelo-West report and the CPC recordings. During the first run,
sampling for the SO analysis was carried out for 20 minutes while
the gas collection for the HC1, HF, and ?^S determination took 30
8-37
-------
TABLE 8-1. TEST SUMMARY SHEET
CO
I
Tim
8:40"- "
10:30 -
10:40 -
10:55 -
11:08 -
11:09 -
11:16 -
11:23 -
11:45 -
12:15 -
12:26 -
12:56 -
1:04 - 1
1:10 - 1
1:14 - 1
1:29-1
10:30"
10:40
10:55
11:08
11:09
11:16
I Sampling flows _T
Sampling [spec, scrub ( coll.
type (l/rcin. 1/nin. cfm •
T*
. .
1
M«ss spec, on dieseli 3.0 :
- -
Mass spec, on 3.0
Grab bottle Orsat
-
11:23! Grab bottles hydro-
carbcn/MO '
1
2.0
2.0 |
-
-
3.5
-
11:45 - 3.5
1
12:15 SO sampling
\sa.rDlinfj dura-
tior. was 20 min.) - 1.75
1^:26 Reheat bed
12:56 KC1 . HF. P-0, - 2.0
1:04
-
- 3.5
-
3.5
i
•1C Mass spec. 3.0 - , -
:13 - - 4.5
:29
:35
NH3»20 - 2.0
-
1
4.0 i
1
Sampling
derations
tot. min. .
.
10
15
13
-
7
"
2?
Tetnperalur
es
Fluid 1 Inertlal
bed "- Tn "
' .
1405 | 1455
-
1360 1450
-
1410 1470
_
20 1430 1500
11
30 i 1460 1570
8
6
3
15
6
1460 1630
-
1450 1570
cut
.
1210
-
irco
Corr*ntS
Stcr*. fli;id1/e. i
prehejt cycle.
Gas off/dicsel
gas
on
Oil off/solid wa
-------
Table 8-1 (continued). TEST SUMMARY SHEET
Tine
1:36 - 1:51
1:52 - 1:55
1:56 - 2:11
2:11 - 2:14
2:14 - 2:25
2:25 - 2:40
2:40 - 2:45
2:45 - 3:00
3:00 - 3:02
3:02 - 3:10
3:14 - 3:21
3:24 - 3:39
3:39 - 3:47
3:47 - 4:02
4:02 - 4:40
4:40 - 4:45
Sampling
type
M2S
Sampling flo.s
spec.
1/min.
-
-
W/NO.
.
-
Grab bottles (Orsati
hydrocarbon . NO ) | -
sox
i
, -
HF/HC1, PjOj ' -
-
Uq.
scrub
1/min
2.0
-
2.0
-
2.0
Part.
coll.
cfro
-
4.0
Sampling
durations
tot.min.
IS
3
Temperatures
Fluid
bed
1420
"
15 1450
j
5.0
*
-
- | 5.0
i
2.0 ; -
5.0
NH-/H.O • - 2.0
J £
3
-
15
5
15
5
15
1420
1410
-
1440
-
1460
l
H2S ' - i 2.0
Inertial
in
1600
-
1570
-
1520
1550
-
1550
-
1660
15 1430 ' 1S60
out
1300
-
1310
-
1250
1260
-
1270
-
1320
Comments
Interrupted sampling due
to feeder valve jam.
put oil on between
3:11 - 3:14
1280
1 ! ':
5.0 8 . - -
*0/N02 - 2.0 - 15 KOO 1500 1240 !
.
.
. . _ . .
_
.
: ' i
14 l/n1n
15 : 1460 1610 1310
i
Changed over to Trapelo
high velocity particle
sampler
00
IO
-------
minutes (see Table 8-1). As the Trapelo West calculations were based on
the assumption of a 15 minute collection time the results regarding
these four gases had to be recalculated. (See Table 3-17.)
GAS ANALYSES
The analytical results revealed six major constituents for the
combuscor off-gas, namely ^, 02, COg, Hydrocarbons, ^0 and CO. All
other constituents appeared in such small quantities that their
concentration can better be expressed in parts per million (ppm) units
rather than In percentages.
The amounts of ^ 02» C02» and ^ are typical for combustion exhaust
gas while the concentrations of unburned hydrocarbons and CO are
unusually high and low, respectively.
The results of the Orsat analyzer had to be corrected due to the fact
that the determination of ^ (as customary) was done by totaling the
C02, CO and 02 percentages and subtracting the same from 100 neglecting
the presence of the other major constituents. The correction
coefficient was determined for both sets of Orsat analyzer results as
91.5 95.6
Cj - "TOO" * .915 and C2 - ~TO"0~ 3 .956, respectively.
The corrected gas composition is shown in the right side column of
Table 3-17.
The differences in gas composition between the first and second
sampling sequences can be explained not only by the fact of a varying
garbage fe.ed composition due to the heterogenous character of municipal
8-40
-------
refuse but also by deviations in the test procedure during the ?1rst
sampling sequence.
As can be seen In the test summary sheet (Table 8-1) some operational
difficulties were encountered during the first two hours of the test.
In order to maintain the bed temperature at 1400° F the solid waste
feeding was Interrupted three times by reheating cycles with diesel
oil. Past testing has periodically shown that the model fluid bed
combustor retains partially decomposed diesel oil 1n the sand next to
the distributor plate. The samples collected after each reheat cycle
are suspected of being contaminated by the oil products that persisted
1n the combustor during the start of each solid waste feed period. This
1s especially true for gases like hydrocarbons whose concentration
dropped from Initially 5.9 percent to 2.4 percent 1n the second run
where no reheat cycles were used. The same explanation can be given for
the concentration variations of sulfur oxides and phosphorus pentoxlde
which are closely related to the presence of hydrocarbons.
Therefore, we can assume that the analyzed gas samples collected after
1:10 P. M. represt.it more accurate results. Accordingly, the last five
test samples of the first run taken after 1:10 P. M. (NH3, H20, H2S,
NO, N02) show a much better correlation with the corresponding results
of the second test run.
The discrepancy between the NOX analyses and the results of a separate
NO and NO* determination are probably due to a different sampling
technique. NO and N02 were determined separately on-line while the gas
8-41
-------
sample for the NOX analysis was collected in an evacuated glass flask
and analyzed later in the Trapelo laboratories. It is likely that some
of the NOX gases reacted with the water vapor in the flask to form
nitrous or nitric acid and thus became undetectable by a later
application of the Saltzman method.
The NOX concentration appears to be rather low compared with data from
other combustion processes. ~o confirm these results the Trapelo West
Laboratory rechecked both the analytical procedure and the calculations.
The measured concentrations of SOX and HC1 appear to correlate well with
studies about the composition of municipal refuse. According to some
data from Battelle and Bureau of Solid Waste Management typical
municipal garbage contains HC1 in a concentration range of 80 to 200 ppm
and SOo always in a significantly small amount. The concentrations of
HF and ^S were expected to be rather low since municipal refuse contains
only traces of these compounds. In addition, HgS is thermodynamically
unstable at higher temperatures, therefore, Its presence in the hot
exhaust gas is very unlikely. As the analytical method to determine
HF is known to be rather specific and sensitive, the results can be
considered as a reliable Indication that only traces of HF are present.
The combustible gases hydrogen and methane could not be detected in the
gas samples by the employed gas chromatograph whose sensitivity is
restricted to 1 ppm. This result 1s somewhat surprising, at least for
CH4 since the exhaust gas consisted of 2.4 - 5.9 percent hydrocarbons
which normally are accompanied by small amounts of methane.
8-42
-------
the difference in the results of the Pfls determination has already been
discussed previously in connection with the heating problems during the
first test run. It means that a P^ concentration in the garbage feed
flue gas of 0.1 ppm is more likely than higher concentrations caused by
remaining diesel oil fractions in the combustor bed.
FLYASH ANALYSES"
The flyash collected at the exit of the inertial separator residue
collection tank (the secondary particulate collection point) was
analyzed by the Trapelo West Laboratory as well t-s by another outside
laboratory, the Metallurgical Laboratories Incorporated in San
Francisco.
Trapelo West determined the anions Cl", F", ^2®$* and S04~2 and also
ran emission spectroscopic analyses for the primary and secondary
particulate collection. Metallurgical Laboratories Incorporated
determined the concentration of the major constituents of the flyash
from the secondary particulate collection point by wet chemical methods
while the minor constituents were derived from an emission spectrum.
The method used by the Metallurgical Laboratories for determining the
anions employed a fusion of the flyash with ^2003. The solution of the
fused salts in water could be used for determining the anions as
follows:
f*2^5 ty tne ammonium phosphomolybdate method
$04 by standard titration with
8-43
-------
Cl by standard tltration with AgN03
F by steam distillation with HC104, photometrlcal
determination of the fluorine - Zr - alizarin lake
The results of the flyash analyses are summarized in Table 3-18. It
must be noted that the percentages listed are converted values for metal
oxides.
The results obtained from the different flyash samples, especially the
two secondary particle collection samples analyzed by two different
laboratories, show some discrepancies for the major components. Since
both laboratories were sent parts from the same sample these differences
must primarily be associated with the low emission spectrograph accuracy
for the major components. The wet chemistry approach used by the
Metallurgical Laboratories Incorporated for the h.1gh concentration
components are expected to be the more accurate levels.
No attempt was made to sample isokinetically for measuring participate
concentrations because the model combustor test set-up did not use
staged particle separation equipment which was under deveopment for the
full scale combustor (see Appendix C) system. In accordance with the
test plan, the partlculate chemical concentrations are prorated for the
full scale gas cleaning system to the ^articulate loading goal of
0.001 grains per cubic foot. This particle loading can be converted
Into parts per million by the following relationship;
0.001 grain/eft » 1.9 ppm by weight
Expected concentrations of flyash constituents 1n the cleaned gas
8-44
-------
entering the turbine can be determined by multiplying 1.9 ppm by the
weight fraction of the respective constituents.
RECOHCNQATION FOR THE TURBINE BLADE WERIAL TEST
From the results of the analyses of the combustor exhaust gas and
flyash the constituents cited below were recommended for use 1n the
turbine blade material testing. The constituents listed are based on
the test results summarized in Tables 3-17 and 3-18. All constituents
with concentrations below 0.05 ppm were considered to be negligible.
Gases
Particulates
°2
ra2
CO
HC
SO,
HC1
S102
A12°3
Na20
K^
CaO
MgO
Fe203
0.8 ppm
0.5 ppm
0.1 ppm
0.06 ppm
0.2 ppm
0.06 ppm
0 . 1 ppm
75. M
14.8S
5.25
0.2%
2.31
2.4%
10 ppm
90 ppm
0.1 ppm
Testing was conducted by the Westlnghouse Research Laboratories,
Combustion Systems Research Group under EPA contract 68-03-0049.
As previously mentioned in Section III, the test results showed no
attack other than oxidation. Oxide scales were left Intact even though
the deposit spa!led, and no localized attack was found other than
normal penetration of oxide at grain boundaries in spite of the
8-45
-------
presence of chlorine, sulfur, and alkali metals 1n the test environment.
Conservative extrapolation of the data Indicated that a machine life In
excess of 10,000 hours should be obtainable.
8-46
-------
APPENDIX C
GAS COMPOSITION AND PARTICLE ANALYSES OF
EXHAUST GAS FROM THE FULL SCALE VERTICAL COMBUSTOR
INTRODUCTION
This appendix describes the equipment us-.d for the full scale vertical
combustbr exhaust gas composition analyses, and discusses the techniques
capJoyed in the determination of the exhaust gas stream particle loading
and size distribution. Carbon monoxide, carbon dioxide, hydrogen chlor-
ide, sulfur dioxide, hydrocarbons, oxygen and oxides of nitrogen were
the exhaust gas compounds selected for gas composition and particle
analyses. Results of these analyses are presented in Section V.
GAS COMPOSITION ANALYSIS DESCRIPTION
The gas analysis systems (Figures 4-1 and 4-2} consist of a sampling and
distribution system, analytic instrumentation and analog data recorcers
Integrated Into a mobile rack mounted complex. The sampling system pro-
vides particle removal, gas cooling, drying, temperature and flow control.
Also Incorporated into this system, where interference gases are con-
cerned, were methods for reducing the component of interest. In addi-
tion, a method of zero setting has been included within the sampling dis-
tribution manifold.
8-47
-------
From the sampling point in the exhaust gas line, the (jas sample is
routed through a particle filter located at the sampling probe outlet.
The particle filter is rated at 5 microns and can function at tempera-
tures above 1500° F. After passing through the filter, the gas sample
is cooled by radiation losses of the sample line to the- heat exchanger.
The heat exchanger further cools the sample to about 50° F and removes
water to this temperature saturation point. A probe removes cooled
sample gas from the heat exchanger after which the gas is heated and
controlled to a temperature of 100° F. It is then pumped and regulated
to the distribution manifold from which the sample flow rates to the
analysis instruments are keot within 10 percent or better with indivi-
dual needle valve controllers.
The rack mounted analysis instruments include: two infrared analyzers
(Beckman Model 315A), a flame ionization detector (Beckman Model 400),
two electrochemical reaction cells (Theta Sensor Models LP-800AS and
LS-800AN), and one polarographic analyzer (Beckman Model 715).
Carbon Monoxide and Carbon Dioxide
These two gases are measured with Beckman Infrared Analyzers, Model
315A(S) (Short-Pathlength). Each analyzer automatically and continuously
determines the concentration of the particular component of interest in
the gas mixture. The analysis is based on a differential measurement of
the absorption of infrared energy.
To measure the differential absorption of infrared energy, the Instrument
uses a double-beam optical system contained in the analyzer section. Two
8-48
-------
infrared sources are used, one for the sample energy-beam, the other for
reference energy-beam. The beams are blocked simultaneously 9.25 times
per second by a chopper - a two-segmented bidde rotating at 4-5/8 revo-
lutions per second. Li the unblocked condition, each beam passes through
the associated cell and into the detector.
The sample cell is a flowthrough tube that receives a continuous stream
of sample. The reference cell is a sealed tube filled with a reference
gas, which is selected for minimal absorption of infrared energy of
tho:e wave lengths absorbed by the sample component of interest.
The detector consists of two sealed compartments separated by a flexible
metal diaphragm. Each compartment has an infrared-transmitting window
to permit entry of the corresponding energy beam. Both chambers are
filled, to the same sub-atmospheric pressure, with the vapor of the com-
ponent of interest (CO or 007). Use of thi,s substance as the gas charge
in the detector causes the instrument to respond only to that portion of
net difference in energy due to the presence of the measured component.
During operation, the presence of the infrared-absorbing component of
interest in the sample stream causes a difference in energy levels be-
tween the sample and reference sides of the system. This differential
energy increment undergoes the following sequence of transformations:
1. Radiant energy - In the sample cell, part of the original energy of
the sample beam is absorbed by the component of interest. In the
reference cell, however, absorption of energy from the reference
beam is negligible.
8-49
-------
2. Temperature - Inside the detector, each beam heats the gas 1n the
corresponding chamber. Gas in the reference chamber is heated more,
however, since energy of the reference beam is greater,
3. Pressure - Higher temperature of gas in the reference chamber rais.es
pressure o-f this compartment above that of the sample chamber.
4. Mechanical Energy - Gas pressure in reference chamber distends dia-
phragm toward sample chamber. The energy increment is thus expended
in flexing diaphragm.
5. Capacitance - The diaphragm and an adjacent stationary metal button
constitute a two-plate variable capacitor. Oistention of the dia-
phragm away from the button decreases the capacitance.
When the chopper blocks the beams, pressures in the two chambers equalize
and the diaphragm returns to its undistended condition. As the chopper
alternately blocks and unblocks the beams, therefore, the diaphragm
pulses, thus changing detector capacitance cyclically. The detector is
part of an amplitude modulation circuit that impresses the 9.25 Hz infor-
mation signal on a 10 MHz carrier wave provided by a crystal-control led
radio frequency oscillator. Additional electronic circuitry in the
oscillator unit demodulates and filters the resultant signal, yielding
a 9.25 Hz signal. The 9.25 Hz signal is routed to one amplifiet,control
section for amplification and phase inversion, then back into the analy-
zer section for synchronous rectification. The resuK'ng fullwave-recti-
fied signal is returned to the amplifier/control "ection for filtering
and additional conditioning, as required, to drive the meter and recorder.
8-50
-------
The meter nadlng is a function of the concentration of the component of
interest in tie sample stream. Whe the instrument is put into opera-
tion, it is adjusted so that a reading of zero corresponds to a concen-
tration of 0 percent of the component of interest, while a fullscale
reading corresponds to the highest concentration in the operating range
covered.
Hydrocarbons
The hydrocarbons in the exhaust gas are detected with a Beckman Model
400 hydrocarbon analyzer. This analyzer utilizes the flame method of
detection to determine the concentration of hydrocarbons present in a
gas stream or ambient air. The flame formed when hydrogen burns in air
contains a negligible number of ions. Introduction of mere traces of
hydrocarbons into the flame results in a complex of ionization, producing
a large number of ions. A polarizing voltage applied between the burner
jet and the collector produces an electrostatic field in the vicinity
of the flame. The electrostatic field results in an ion migration
whereby the positive ions are attracted to the collector and the negative
Ions to the burner jet. Thus, a small ionization current is established
between the two electrodes. This small current is directly proportional
to the hydrocarbon concentration in the flar;ie and is measured by an elec-
trometer amplifier circuit.
The amolifier circuit provides the driving voltage for an indicating
meter and recorder. The magnitude of the electrometer amplifier signal
1s indicative of the number of carbon atoms passing through the flame.
8-51
-------
The "carbon counting" characteristic of the detector is proportional to
both the sample flow rate and the nature of the sample. For example,
a flow of 10 cc/minute of a given sample provides a signal approximately
twice as great as that produced by a flow of 5 cc/minute of the same
sample. Also, a given volume concentration of a typical Cc hydrocarbon
produces a signal twice as great as an equal volume concentration of a
C-j hydrocarbon, or six times as great as an equal volume concentration
of a single carbon (CH^) hydrocarbon. Instrument characteristics, such
as sensitivity, are thus typically reported in terms of CH4 (methane)
equivalent.
Oxygen
The oxygen in the exhaust gas is measured with a Beckman Model 715 pro-
cess oxygen monitor. The monitor consists of two basic modules: an
amperometric oxygen sensor and an amplifier unit. The sensor detects
the partial pressure of oxygen present; the amplifier measures the magni-
tude of the sensor signal, which may then be read directly on the meter
or used to drive a 10-millivolt recorder.
The amperometric oxygen sensor contains a gold cathode and silver anode.
The two electrodes are seoarately mounted within a PVC body, and are
electrically connected by a potassium chloride electrolyte. A constant
potential is impressed across the two electrodes. A gas-permeable teflon
membrane separates the electrodes from the process sample, and fits
firmly against the gold cathode. Oxygen from the sample diffuses through
the membrane and is reduced at the gold cathode. The resultant electri-
cal current flow between the anode and cathode is proportional to the
8-52
-------
partial pressure of oxygen 1n the sample.
The amplifier contains a meter, controls, and a solid-state, drift-free
feedback amplifier which measures the current level generated by the
sensor. The meter has two scales that correspond to the two ranges pro-
vided by the range switch, and a third scale for dissolved oxygen, per-
cent saturation, or partial pressure measurements. Air calibration is
made on the 0 to 25 percent scale by setting the meter pointer to 20.9
percent via the calibrate control. The calibrate control is equipped
with a locking device to prevent inadvertent changes of the calibration
setting.
Sulfur Dioxide and Oxides of Nitrogen
These two gases are monitored with Theta Sensor Model LS-800 analyzers.
One analyzer (Model LS-800AS) measures both NOX and S02. Its response
is identical to NOX and S02» therefore allowing for direct subtraction
of the S02 measurement in order to determine NOX. Each unit is a single-
channel Instrument designed to provide quantitative measurement of the
S02 or S02 and NOX as found in the stack emissions.
The principle of operation .utilizes a combination of semi-permeable mem-
brane and selective redox reactions within the completely sealed trans-
ducer. The electrical signal from the transducer is directly propor-
tional to the concentration of the pollutant in the gas environment being
\
monitored. The response Is linear over the 0 to 5000 ppm range covered.
The transducer is a totally enclosed, sealed plug-in unit. The Instru-
ment is equipped with a set of thtc* range selector switches (L, M, H)
8-53
-------
allowing the selection of a desired sensitivity for full scale meter
readings of 500, 1500, or 5000 ppir respectively.
Hydrogen Chloride
Hydrogen chloride in the exhaust gas is monitored by use of the Volhard
titration method. Discrete samples are taken at 2 to 5 hour time inter-
vals to determine the average HC1 concentrations present in the exhaust.
A volume of the exhaust gas was first bubbled through a solution of so-
oium hydroxide. The liquid was then analyzed in the laboratory for the
concentration of sodium chloride. As previously mentioned, an attempt
to procure an on-line instrument for HC1 measurement using a gas chroma-
tograph was abandoned when the instrument subcontractors determined that
it was beyond the state-of-the-art.
OFF-LINE DETERMINATION OF GAS STREAM PARTICLE LOADING AND
SIZE DETERMINATION
The remaining pages in this appendix describe a technique for off-line
determination of the loading and size distribution of particles in a
gas stream and provide a general proce&jre with typical results.
Particle analysis has historically resided somewhere between a pure
science and a black art. There are three reasons for the problems in-
volved in particle analysis. First, it involves probability. It deals
with the representative sample as opposed to an exact measurement. The
particle counts determined for one sample are not necessarily identical
to the particle counts determined for a different sample taken from the
same system unde»- seemingly identical conditions. Similarly, if a sample
8-54
-------
is divided into two equal parts, the analysis of each part will probably
provide somewhat different results.
The second problem encountered in particle analysis is that it involves
sizes and counting of parts that generally cannot be seen with the un-
aided eye. This together with the need for relatively large sample
counts to increase the confidence level in the results makes optical
count methods tedious and prone to error.
The third problem involved with particle distribution analysis is the
susceptibility to the introduction of errors in the results wh~" great
care is not used in sampling or analyzing the data. The results can be
drastically altered by particles from the atmosphere, oarticles intro-
duced by the act of connecting and disconnecting linos to the sampling
system, particles lost or extra particles introduced by sampling at
non-isokinetic conditions, particles introduced by the laboratory equip-
ment used in the analysis, particle loading variations resulting from
leaks in the sampling system, etc. While such problems are subtle, they
ran have a devastating influen ;> the results.
It is for these reasons that it is imperative to have a proven procedure
to determine particle size distribution and to have the personnel in-
volved in the sampling and analysis sufficiently familiar with the equip-
ment, technique, and philosophy to recognize when problems are occurring.
Techniques for Determination of Particle Size Distribution
The determination of particle size distribution in a gas stream involves
three operations: system sampling, sample analysis, and data reduction.
8-55
-------
The general approach to be followed for each of these operations is
discussed below.
Isokinetic Sampling-
The basic technique for obtaining a sample for later analysis is to pass
a representative sample of the particle laden gas through a liquid,
collect the particles in the liquid and allow the gas to escape to tKe
atmosphere. A schematic of the sampling system is illustrated in
Figure 8-14.
The main consideration is to insure that the sample is a representative
sample. To accomplish this, the sample must be collected isokinetically.
In other words, the velocity of the gas entering the sampling system
must be equal to the velocity of the gas that bypasses the sample tube.
If the velocity of the gas entering the sampling system is greater than
the velocity of the surrounding gas, the inertia of some large particles
would orevent them from following flow streams and entering the sampling
tube (Figure 8-15a). The sample will therefore tend to be light in the
number of large particles counted. Similarly, if the sample velocity is
less than the velocity of the surrounding gas, the inertia of some heavi-
er particles will cause them to continue into the sample tube rather than
follow the flow stream around it, thus tending to increase the number of
large particles in the sample (Figure 8-15b). Assuming that the gas
follows the perfect gas law, the cated volume flow rate for the sys-
tem shown should be:
"••" '»
8-56
-------
SAMPLE PROBE
EXHAUST STACK
COLLECTION BOTTLE
BYPASS
VALVE
VACUUM
PUMP
COLLECTION FLUID
MOTE: Impingers are designed for a certain range of flow rates
and should be used only within that range.
Figure 8-14 Sampling System Schematic
8-57
-------
FLO;;
STREAM
FLOW
STREAM
Ficjure 2A Hyper-Isokinetic Samplinn
PATH OF
HEAVY PARTICLE
PATH OF
LHHT PARTICLE
-*• LIGHT PARTICLE
FLOW
S REAM
HEAVY
PARTICLE
Figure 8-15 Non-Isokinetic Sampling
8-58
-------
where: As » .ample tube entrance area
Aa •*• stack area
TS = sample absolute temperature in the flowmeter
Ta * stack absolute temperature
PS = sample absolute pressure in the flowmeter
Pa = stack absolute pressure
Qs *• volumetric flow rate of sample
Qa - stack volumetric flow rate
Assuming standard temperatures and pressures
530 (P ' + 14.7) „
(Td' + 460) 14.7 d
where: Ta' - stack temperature in degrees Fahrenheit
Pa' = stack gas pressure in psig
By maintaining the established flow rates through the impinger for a
specific period of time, the sample will represent a known quantity of
gas. For example, a flow rate of one cubic foot per minute measured at
the flow meter and maintained for 5 minutes will result in a sample which
contains all the particles from 5 cubic feet of gas at the temperature
and pressure of the gas in the flow meter (normally close to standard
conditions) . However, this assumes no errors in sampling. If a leak
exists 1n the sampling train or the sample tube, and the sample tube
pressure is below atmospheric pressure, the measured flow rate will rep-
resent the sample flv,w plus a leakage flow. The collected sample will
therefore represent something less than the 5 cubic feet previously
computed. If the sample system pressure is above atmospheric pressure.
8-59
-------
the leak will be outwards, the particles will probably follow the main
stream rather than the leakage flow path and the collected sample will
be representative of an unknown volume somewhat in excess of the 5 cubic
feet calculated. Either case will result in an error in loading and
probably in the distribution.
Another area of potential error is the introduction of particles other
than those entering the sampling tubes. These can be introduced by the
act of filling the sampling bottle with collection fluid, connecting or
disconnecting the sampling bottles or handling of the particle laden
sample after collection, or before or during analysis.
On the other hand, particles entering the sample tube can leave the
sample air stream and be caught in voids in the sampling train thereby
resulting in a lower than actual loading. This can occur if condensa-
tion takes place within the sampling train before the sample enters the
impinger or if large temperature changes take place between the sample
point and the collection bottle. Such temperature changes cause thermal
gradients across the flow stream which in turn cause cross stream flow.
This cross stream flow leads to the deposition of particles on the
sampling train wall.
Particle Analysis-
An automatic particle counter is used to determine the particle quantity
and distribution in the sample. Th- sampling configuration is illus-
trated in Figure 8-16. A measured quantity of particle laden fluid is
drawn through an orifice of known size. The fluid is an electrolyte.
8-60
-------
x—-^ VACUUM
/ V^PUMP
"STOP
cnu-iT"
LL:C
TRODL
START
COUflT"
LLCCTRODI:
COLLECTION
BOTTLE FOR
SAMPLED FLUID
ORIFICE
MERCURY
COLUMN
ELECTRODl
Figure 8-16 Particle Analyzer System
8-61
-------
thus allowing for a current to be passed through the orifice which is
the major resistance to current. Particles passing through the orifice
act to reduce the cross sectional area available to current flow, there-
by increasing the current path resistance. By measuring the change in
resistance at the orifice, it is possible to determine the size of par-
ticles going through. Within the working band of the orifice, the
change in measured resistance is proportional to the current, to the
signal amplification and to the volume (cr diameter3) of the particle.
The unit is usually set up to measure the number of times the change in
resistance exceeds a set value during a single draw of sample through
the orifice. The unit used to measure the change in resistance is a
pulse height analyzer. A standard pulse counter counts the output of
the analyzer. Since the resistance limit threshold for the pulse height
analyzer corresponds to a particle iize, the unit will essentially
count the number of particles that exceed a selected size. By starting
and stopping the counter using a mercury column switch network which
draws in a specified quantity of particle laden fluid, it is possible
to count the number of particles in a known quantity of fluid that
exceeded a pre-selected size. The number of particles that exceeded
the pre-selected size in a cubic foot of gas can then be determined by:
1
-------
Using different resistance thresholds, correspond to different particle
sizes, for a number of measurements, it is possible to obtain a distri-
bution for the particles collected from the sample gas stream.
As with all particle counting systems, errors can be encountered with
this sizing technique. Any change that can cause an apparent change in
the measured resistance can cause errors in counting. Electrical fields
such as those emitted by soark plugs can generate currents in the sample,
thus causing high particle counts. Mechanical vibrations in the equip-
ment can also cause apparent resistance changes between the probes. In
some cases, it has been observed that mechanical vibrations can mask
small particles going through the orifices, thereby causing a reduction
in the small oarticle counts. Another problem area is in the orifice.
Particles on occasion tend lo partially block the orifice. Any parti-
cles going through the orifice will then appear larger (greater change
in resistance) than they actually are.
The only good technique to avoid erroneous readings due to these condi-
tions is to be familiar with the characteristic signal from thr sensor
(displayed on an oscillosope). When the signal is apparently bad, the
operator should check the orifice for plugging and check the area for
excessive electrical or mechanical noise.
Another area that can cause erroneous readings is to attempt to use the
instrument outside of its designed range. Counting too many particles
(a heavily contaminated sample) will cause '•oincidence problems where a
number of small particles go through the orifice concurrently and are
8-63
-------
therefore ineasured as a single large particle. Attempting to count too
few particles is equally bad because background counts can exaggerate
the data. Also, attempting to count exceedingly small particles with
respect to the orifice size can lead to a high count due to noise.
Data Reduction-
Reducing the data into a form suitable for evaluation of a system in-
volves a number of steps, especially if the data is to be used for
evaluation of particle separators or other devices for which is it de-
sired to compare two readings. The reduction of particle analysis
data can be greatly improved by familiarity with characteristic trends
of oarticulate distribution Such knowledge makes it possible for the
data evaluator to detect and compensate for sampling and analytical
errors, and equipment deficiencies.
Particle distribution for a number of particle generators including a
fluidized bed generated exhaust tend to follow a straight line when
plotted on log-log paper (a special graph paper which is not available
commercially but which can be generated via a computer program and peri-
pheral plotter). The analytical data is initially normalized to a unit
gas volume and then plotted .on the loq-log paper. The plot is then
examined for deviations from a smooth curve and corrections made as
required. If loading and distribution for single samples only are
required, the resulting curve is used directly for distribution deter-
mination. This involves dividing the curve into a number of segments,
each representing a range of particle sizes, computation of the number
of particles in each increment by simple subtraction, computation of
8-64
-------
the total volume represented by the particles in each increment using
a mean diameter and finally, computation of the equivalent weight of
particles in each increment using a standard density. From these last
figures, a plot can be generated which displays the total weight of
particles exceeding the micron size indicated by the graph ordfnate.
The effort required for this complete data reduction process can, of
course, be greatly reduced by having available and using appropriate
computer programs.
If two distributions are to be compared, such as for determination of
particle collection efficiency in a separator, it is necessary to fur-
ther condition the data from the analysis prior to loading evaluation.
First, the ratio of the two loading curves is calculated and plotted
(plotting is usually done on a log-log graph). This clot presents the
separator efficiency versus particle size, efficiency being the ability
of the separator to remove all particles of the noted sizo. As with
the initial distribution curves, the efficiency curve should be smooth
and the efficiency should continuously increase as the particle siza
increases. Following curve smoothing and conditions by the data evalua-
tor, the efficiency curve and the separator input loding curve are used
to generate a new output loading curve. This is done since the output
sample is usually prone to background noise and introduction of errors
through contamination of the sample. (In some cases, background noise
has caused an apparent decreasing and even a negative efficiency in a
scpai ator output. This can occur when the ratio of particle counts over
the size range of interest is in excess of 1000:1). Following this data
8-65
-------
conditioning, the two curves are reduced for loading as for the single
curve case.
General Procedures and Typical Results
The general procedures to be followed for sampling, analysis and data
reduction are presented below. Sample worksheets are frequently used
to provide typical results.
Sampling Procedure-
The volumet"1c flow rate of gas in the exhaust stack or sampling point
must initially be determined. 'This is normally achieved using a cali-
brated orifice with known coefficients and correcting for the tempera-
ture and ,1 -fissure within the stack. If the velocity is high (the usual
case in fiuidized combustion systems) the particles will probably be
evenly distributed across the stack and the velocity profile will be
relatively flat. If not, corrections should be made for the velocity
profile and samples should be taken at a number of points in the stack
to insure chat the sample is representative of the gas stream.
A sampling tube should then be installed in the stack. The tube should
be sized to allow for a flow compatible with the sampling system (note
that impingers are designed for specific ranges of flows). The collection
bottle should be filled with an appropriate amount of collection fluid
and the impinger installed. The impinger should then be connected to
thfr sampling train and checked for leaks by turning on the vacuum pump
with the imoinger Inlet blocked. Any flow in the flow meter indicates
a leak.
8-66
-------
The sampling train should be connected as close as possible to the
sample point at the start time and the calculated sample flow rate set.
Temperature and pressure within the flow meter should be recorded for
later flow rate verification. Also, any abnormal condition encountered
during sampling should be noted.
Following sample collection, the iiroinger should be withdrawn from the
collection bottle and any particles deposited on the Inlet washed Into
the bottle using the proper solution. Final sample volume should be
recorded in order to allow conversion of the data to a unit gas volume.
(See sample data sheets).
On occasion, a background sample should be taken. This involves placing
a filter on the end of the sampling probe and taking a sample as though
it were a standard exhaust emission sample. It is important to follow
the same procedure used for standard sampling when taking the background
sample. Sometimes, an investigator will use greater care with the back-
ground sampling and analysis than Is used for exhaust gas sampling and
analysis. This makes the background data "look good" but of little use
when using background results for stack analysis correction.
Data Analysls-
T(... particle counter should first be calibrated using the orifice selec-
ted for sample analysis. A typical data sheet is s-hown in Figure 8-17.
The main objective 1s to determine instrument settings that correspond
to a known particle size. Usually a pollen is used for calibration and
the breakpoint (point at which half 'he particles are larger than the
8-67
-------
Initial Calibration
Orifice 70 H _
Calibration particles 12 - 13
Date
Resistance
20.000
Settings
Trla 6.7S
I/A 16
I/I
1/4
NOTE: Establish settings such
that roaJino* at a t)
setting of 10 are twice
the readings that occur
with a t) setting of 50.
(tu at nax.)
Background at tj • 10
1195
Readings at t] • 10
1. _1285
2.
3.
4.
5.
SIM
1253
1/2 (average) ' 10( 6313
631 • half count
1310
1270
6313
Readings at t, • 54
1. 622
2. 644
3. 638
4. 630
5. 628
SUM 3162
(set to obtain above half count)
(trial and er.-or)
Check
Som
632
Calculate K* (Based on O3 rather than V)
K» • oi 0 1n
I x A x tj
. 83)
10.0
x 16 x 54
Figure 8-17 Initial Calibration Data Sheet
8-68
-------
threshold and half the panicles smaller than the threshold) is deter-
mined by trial and error. Once the settings for the known particles
are established, settings for other particle sizes can be calculated.
It has been found from past experience with the counter that settings
should be established such that the operator has to adjust only the
amplifier gain. The adjustment for this setting can be varied from
1/8 to 64 In 10 steps, resulting In a possible threshold variation
corresponding to an 8 to 1 particle diameter range. (It should be
noted that the instrument settings are for the Inverse of the amplifier
gain and the Inverse of the orifice channel.). This range is usually
sufficient for particle loading and distribution. Equivalent particle
diameter for each setting of the arrpHfler ia1n control can be calcu-
lated from:
) ('/'I) (I/A) tj (4)
where: K' » the calibration factor previously calculated
I/I « the current control setting
I/A » the amplifier control setting
tj « the lower threshold setting
D * the equivalent particle size In microns
The upper threshold (tu) Is adjusted to the maximum setting for all
standard counter operations and the trim Is set at the value used during
calibration.
-------
By selecting appropriate single fixed values for II and tj, all diame-
ters of normal interest can be covered by the available amplifier con-
trol settings.
In recalibrating the unit, it has been found best to set up the original
calibration settings and then adjust the trim knob to obtain the appro-
priate bre.ak point of the calibration sample with these settings. The
data analysis sheets can then be updated to the latest calibration point
by using the latest trim setting and UM'IPI th>> oinnnallv selected values
for the other settings. (See Figure r-18).
After calibration, the system is ready for parti le analysis. Some of
the specific procedures to bp followed when obtaining distribution data
are:
1. The counter should be wanned up for one hour prior to analyzing
any samples.
2. Samples should be analyzed the same day that they are collected.
3. It it is necessary to filter the sample in order to avoid
plugging, use a mesh in the same size range as the sampling
orifice. Filter the sample prior to any dilution and then use
the dilution fluid to Wash the particles through the mesh.
4. All samples should be within 50° F of each other when analyzed.
'5. When splitting (or reducing) the sample size, pour sample be-
tween two beakers a number of time, reducing the amount trans-
ferred each time until the desired.volume is obtained in one of
8-70
-------
_UJ72
Orifice 70 M.
iallbratlon particlci 1? - IS
Orlgi.ial -iSttlngj ri'ji-: f'.tjtlisi. ic'.tingi such
thit rLiidinc;'. ut a ti
i^aJin-js t.'iat O':cur
I/A ]/ rtltn « t| s^ttiitn of 50.
(tu at max.)
at half count
d at t^ • 10 2
~"5
Peidings *t t^ • 10
1. 1.1;?
"1/2 (average) • 10 J~5694
3. 1148
569 • half count
4. 1179
5. 1167
Sum 5694
Readings at ti =_54 (use orlginil tj setting ind adjust trim)
1. ^41 l.ew trlr s»'tin(j » 9__
2. 515
3. j40
Check * Sun • 565
*• 589 7
5. 612
Sum ?828
Figure 8-18 Recalibratior. Data Sheet
8-71
-------
the two beakers. If only a small portion of a sample Is to be
used, the smaller sample should be obtained by a number of
splitting steps.
For example, if it 1s desired to use only 1/8 of a sample,
split the original sample into two parts, split one of the new
volumes into two additional parts and finally split one of the
smaller volumes Into two Darts again.
6. Measure the orifice resistance with the counter unplugged.
7. Wash sample tube with distilled water prior to introducing
a new sample.
8. H1pe orifice at least every fifth reading. Following each run,
increase the armature current to the maximum, operate the sys-
tem through two counting cycles and then make a final check on
the 5u range. This final check count is to be noted at the top
of the data sheet.
9. The following number of repeats shall be made for each count:
Count Repeats
10 - 100 8
100 - 1000 4
1000 - 6000 3
Counts should not exceed 10,000
Typical sampling data sheets are presented on the following pages. They
are self-explanatory and are set up to provide the instrument operator
who Is familiar with the equipment a step by step analytical procedure
for sample analysis.
8-72
-------
Test I. Q.__/3-V2
Probe Position
Date
On
Test
Floor
Sample Flow Rate
Time
Sarplc 54% volume (AXB)
Final 'sotone voluir.* In impinger (0)
L following purt1clc collection and
impinger «ash
(A) 06 •
(B)
S ,
(0 # 66 7 (
(0)
200 ,
SCFM
Minute*
Cubic f«a
•1
Portion of sample used for
dilution test
(E)
Isoton added for initial dilution (F)
| u-st. (liotc: Add sufficient
i I'.'^tinc to brint; in-, tial i;rtp count
j coi.r ICgS.COO MX. Fluid Mill
j ,->r&ojo!y be clouay but not opaque).
Initial I
Dilution—, Total sarple size for initial
and ! dilution test (t»F) (G)
Test I
I Initial counts
I i
A -
Trim •
M "
2S
MX
Resistance
.1
/75
2oo
Figure 8-19 Dilution and Volume Factor Summary Data Sheet (Probe C)
8-73
-------
—
Dilution required ( L )
|OOW
Counts for 30 V 1 (H) Z/ 1 z£
(D~~r-L£¥z
(J
Sum ft * I * J) If.
i./ ?. *t
£> i 7/2
Avor.qe (K) (L) 2O9Of^
3
(M) 2
Note: round o^r to nearest unit (WWMM)
Selected sample size
Final _' Isotone added
Dilution
Total sample (N » 0)
NOTE: - -
1
2
3
4
1 f M • 5 then
6
7
8
i» - i
(N) /^^?
0)~^5'o
IP) 2J'0
^
ISO*
100
50
set M *1 25 i. nl and O •
2b
C^
25
,_25j
0
100
100
75 - •!
100
125
150
;"J
The volucw (gas represented by sampling volume (SO
'V, • C
Is:
Volume
Factor
Calculation
x E x N
ff F
.05
x too
feet
Figure 8-20 Dilution and Volume Factor Calculations Data Sheet (Probe C)
8-74
-------
Single Thre^old Test I.D. > 3 ' v i
70 p Orifice Prote Location /£
.'sotont Liectrolytt Date, S' 3
'••JC^ <^T Oac'KOro^n<2 at /?TA> sotting ;; flS
Trla C.2g Final count on f.:.yy sotting JO't'V^ /OtO^I /
-------
Ttst I. D. /3-12
Probe Position _/ &
Dite 5-3-72
On
Test
Floor
I
Sample Flow Rate
Time
Sir-pic <;as volwv.e (AXB;
Final Jsotonc volu
fo'i'powinr, ;.,rtic~pe
(A) Ot scfti
(B)
7
30O
Minutes
CuMc feet
ml
iCii.". in and
Inititl
Portion of sample used for (E)
in.tui miuticn ttit
Isoton adcod for initiil rtilution (F)
test. (:..••-•.•: A«JJ su. • ,cpp.'..t
Isotoro to i'f\r,:; in1, t ,,;"i ;.•.•:,• cc-nt
idd
Test
fUtchir,; SCs'.
(i)
/75
ml
200
Figure 8-22 Dilution and Volume Factor Summary Data Sheet (Probe B)
8-76
-------
Fin*l _
Dilution
Volume
F«ctor
Counts for 50 V 1 (H) A* 9V/
* (D—73-?ys-
j nr/'n
SU.TI (H + I * J) K)~7/TJ8~
Averse (K) L) /59 7'/
1 3
Dilution required ( L ) (;•;) 2
)oo3o
Note: round «FT 10 r.carest unit (I/AMMM)
Selected sample size (N) /OO
Isotone «(3iieJ (0) /oa
Total sam/.e (N + 0) (?) yoo
NOTE: r _
1 _ . ,
2 fl50 0
3 100 IOC
4 bO 100
If M • 5 then setN -i 2S . ml and 0 • 75 - ml
6 25 100
7 25 125
8. 25 150
]_25J J75_
M
The volutne (gas rcpresente-1 by sampling volume (50 ^1) Is:
V, • C x.OS x £ x N
o~ ff ?•
-f<6£7x .01 x 25 , /go
cub'c feet
Figure 8-23 Dilution and Volume Factor Calculations Data Sheet (Probe B)
8-77
-------
Threshold
70 /I/ Orifice
Isotone Electrolyte
err 1^50 +210
Sample Temp. 75. 5 _
Aperture resistance ^0 ^
Hitching resistance setting 2O ><
Test I.D.
Probe Location
Date 5-3-72
Screening prior to test T_X_ N
Screen size
Background atMf setting 2fS
Tr1i>
Final count onf.~A> settingBJ>2,
(Should nutch inittjl countj"
tu set at max.. Separate/Locked switch to separate
Xft I
Count
till
3W6
Sum
Average
/07/t
J-tt
«•>
3fc
_JL5_
23 22
ilf
22
/o.o
M
i.6
Figure 8-24 Raw Data Particle Count Data Sheet (Probe B)
8-78
-------
t. o.
Prob* Position
D*U
//?
On
Tint
Moor
L
(A)
(B)
(0
Jsotone volume 1n Impinger (0)
following p-irticlu collection and
w.isn
Flow Rite
Time
Sample ",t\ volume (AXO)
_M1nutei
Cubic f««l
Initiii I
Or,uHon_J Total
T«t
Portion of sample uted for {£}
tnitul dilution test
idJcil for Initial dilution (F)
,"i^to: Adu Sufficient
lo t.rini; mi ttal /.•/:<• count
i lie clou.iy but not o^«qu«).
\\te for InitUl
tea (l»F) (C)
InitUl counts
'/8
t. •
rUtehlng
(MX
ml
225
Figure 8-25 Dilution and Volume Factor Surrmary Data Sheet (Probe A)
8-79
-------
Tlnal .
Dilution
Dilution required ( L)
|653a
Note: round OFF to nearest unit
Selected sample size
hotone adUori
Total san^le (N » 0)
MJTE:
IfM
2
3
4
5
6
1
8
then set
lounts for 50 \ 1 (
Su«
(H * I » J)
Average (X)
1
t^j/f
150*
100
50
25
25
25
25
-25_
H) 75V7
I ) ? S £• 7
J y?rr^
X 1307
L
H
?* }i
0
M) 225
0
ml and O •
O
225
o" "
100
100
75 • ml
100
125
150
JJ^-.
Volune
Factor
Calculation
The voluae (gas represented by stapling volucw (50 ^t) U:
V« • C xU)5 x I x N
IT f f
•61 x Jo x 22S •
-------
Single Threshold
70 // Orific*
Isotone Electrolyte
Test I.D.
/3 ~
Probe Location
Sample Tenvp. 7V. 7
Amperture resistance 20K*
-
Matching resistance setting 2,0 KL
Date S'3-7j
Screening prior to test Y X ft
Screen size 3O •* 2 f
--
Background at /.?TX/ setting
Trim
Final count on ^-•yv setting ?255 J5
-------
Data Reduction-
The raw laboratory data is initially normalized to a standard gas volume
(left hand side of sample Particle Counting Data Reduction Sheets,
Figures 8-28 and 8-29). This is accomplished by multiplying the raw
count by the ratio, selected unit volume to sample volume factor
(noted on the laboratory data sheets). A preliminary distribution plot
can then be generated from the normalized data (Figure 8-30).
The data reduction from this point depends on the type of data required,
single point loading, or particle removal system efficiency and loading.
The following paragraphs describe the approach used for efficiency ana
loading evaluation. For single point loading analysis, the efficiency
calculations are omitted and loading is directly calculated. However,
the particle count curves should be reviewed to omit effects of coinci-
dence and background.
Preliminary efficiency calculations - The preliminary efficiency calcu-
lations are used to correct the data for machine and procedural deficien-
cies. Efficiency is defined as the percent of particles removed by the
separator or unit under test. Since this efficiency is dependent on
particle size, it is not single values and must be expressed as a func-
tion of particle size.
The number of particles of a certain size entering the test unit is
mathematically equivalent to:
\
AU -» o (5)
R-82
-------
CO
I
oo
CO
re1.. s'.ro;~ Paw
A.cri;« Co.nt
I 72Z-8-
*7l^-~.
.^ssi
17
C.3
__?_7
Li
/o.o
3.7,5
D;'..nsir-?a-; CDunt/L'nit V
(F « C/B)
tl
_2'?*J
—l~l(Ljt I
JO 3^
::r?^;i5
/.-?.jii
£1^24
590
zn
Test I. 0. /J -VZ
o«tt j-3-y<
Upstrean Probe /£
Downs trej:i Probt /A
(H)
A
A*/
2.o ! M«
.S.o
6.0
6~T'
0.0
3 IV
J./V
(J)
M
7.
3-/V
3-/V
J-0
6.
7.YI
J.o
6.3
fO.O
fO.O
(D
75.21
13.7$
93.3?
11.66
7V.
13.11
Figure 8-28 Particle Counting Data Reduction Sheet (1)
-------
CO
I
Test I. 0. is -
cut* _ s -
Upstretn Probe /C
Dot«r.strean Probe /A
A) t'putrej- $J-pli Vo'--o Fjctir ,.,,'?„•} cubic fefft.
8) D:..-;irc4~. S:-sle Vali.-? r*ctor"r.,,,.j j
C) Selected BJse (Lr.it) Vol,.-; _,o'{
0) £>/ Selected for Ar.jl/sis J _ "
feet,
cubic fiet.
microns.
(E)
IJ.
(•*•
H3
r*
^ ^i
31*
>v^
P Q
J»'°
£ -7
7 f ^
/O.o
CCuli '.
(F)
'Jp'.troi- Sjw
A.cr:70 Cj'.Rt
Cc..-r.r?i- P'.v
A-.e.".^; Caunt
/0S"i ?*.*./
i4(z
2. J /.?
J P ? 54 $
7 3 -> i 8
/3V//^>
/5?*4
3?^^g
<£*£
fit r {
I4
k O
* a.
C
-Q-
-------
8 9 10 11 12 13 14 15
1x10
1x10
1x10'
1 2
Figure 8-30 Preliminary Distribution Curves
8-85
-------
where N is the number of particles greater thai size p. The number
leaving the separator 1s equivalent to:
. /dN out
du
v •• v,
AP
(6)
The efficiency at pa (percent of particles of micron size pa removed by
the separator) is therefore:
» -100
(P « wa)
liliil ' dN
dp dp
dN in
dp
(7)
N. y
By selecting a sufficiently small micron size increment, the above ex-
pression can be approximated by:
/
ty-100
(ll « Ma)
AN in AN out
Ag AM
\
100
AN in
Au
AN In - AN out
AN~Fn
u = w.
W " V";
where AN - j N(
-{N(p = p + Ap)f = -dN
and N » number of particles greater than size p
(9)
(10)
8-86
-------
Assuming that the distribution tends to follow straight lines or straight
2
line segrents when plotted on log-log paper, then:
In N - In Ni = (In rf - (In u,)2
1 V (ID
In Nj - In N2 (In Uj)2 - (In vj
where: NX = number of particles larger than
and: Nj, N-. uj, and u- are known from test results
This equation converts to:
V/here:
(in N! - In N?) (In.)2 - (In vl)2
(In Wl)2 - (In M2)
2
Using formulas 9, 10, and 12 it is possible to determine the separator
efficiency for any particle size using known input and output counts at
two different particle sizes. Typical results are shown on the right
side of the "Particle Counting Data Reduction Sheet". Note that in all
but the extreme cases the efficiency at each selected micron size is
determined using two different pair of data points. The dual answers
are then averaged to compensate for non-linearity in the data. Also
note that the calculation accuracy is dependent on the AW used in the
calculation (equation 10). The 6w selected should be small enough to
provide accurage results and yet be sufficiently large to be compatible
with the equipment used for the calculations. If it is too large, the
approximation that dN/du = -AN/AM does not hold. If it i* too large,
the computational equipment will be unable to accurately determine the
small differential (AN) between two large numbers.
8-87
-------
Data comoensation - The resulting efficiencies are next plotted on log-
log paper (Figure 8-31). At this point, it becomes apparent where back-
ground influences have adversely affected the data. Assuming that the
particles are essentially similar in shape and densities, inertial sepa-
rator efficiency should continuously increase with particle size. How-
ever, when the count becomes small enough to be adversely influenced
by background contamination, the apparent efficiency begins to decrease.
The effer1 jf background count is illustrated 11 Figure 8-32 where the
unprocessed laboratory reading: are compared to a typical background
distribution. At the two micron point, the background count is rela-
tively low and introduces an error of approximately 3 percent. At the
ten micron level, the background appears to be responsible for almost
all of the counts for probes A and B. This large particle count results
in a low apparent efficiency for the separator and must therefore be
compensated for when reducing the data. Compensation is normally accom-
plished for inertial separators and similar devices by assuming a linear
efficiency when plotted on log-log paper as illustrated in Figure 8-31.
Compensated distribution - Once the compensated efficiency curves are
available, new particle distribution plots can be generated. The method
used is to assume that the data for the highest loading (probe C) is
most nearly correct (note that the background count has the least in-
fluence on this curve) and then calculate the other distributions based
•jpon the corrected efficiencies.
-------
99.8
EFFICIENCY FOR
QA-B
©B-C
UNCOHPENSATED
DATA (EFFECT OF
\BACKGROUND)
Figure 8-31 Efficiency Curves
8-89
-------
PROBE
El 1C
A IB
0 1A
UNPROCESSED LABORATORY READINGS
6 7 8 9 10 11
Figure 8-32 Comparison with Background
8-90
-------
From equation 10,
dN
JL
100
or
\JUt
,
dw
(13)
(14)
By assuming the plot of efficiency to be a straight line between points
1 and 2, f
(1 •
\
where:
In-
ioo
n s the efficiency for particles of size u
A A
(15)
By substituting the above into equation 14. selecting an appropriately
*
small micron increment, and using the 2 micron count as the baseline,
equation 14 can be aporoximated by:
Wa + Ay - 2
Nout = Nout
- i5oj
]
in'1
f / \\
(» \)
ln \wj
In (^\
In.
7 . 41"
V Tool
(} M
^ \ 100/J
AN,
in
where:
» 2 + 2M Ay - Ay
Mn
\
J (16)
(17)
(18)
8-91
-------
A computer solution for the compensated curves using the example data
is presented in Figure 8-33.
Loading data - The particle loading 1s next determined from the distri-
bution plot. Considering equation 18, N^n represents the number of
particles within a ± AU span about ua- Assuming that the typical parti-
cle Is spherical and that the average volume of each particle is:
——
particle 6
The volume represented by the Nin particles is therefore:
(19)
vp = "a - AM = T" ANin
The total volume of particles whose diameter is greater than y is
a
therefore:
(20)
total
• L
M » 1
AN.
in
where:
U * 11 + 2MAu -
a
(21)
(22)
-3,
By using an average density, d(gra1n/micron ), the total loading of
particles greater than diameter u 1s:
d
M > 1
(¥) 1N'°
grains
unit gas volume
(23)
8- i
--)._
-------
1x10
CORRECTED PLOTS OF A AND B
BASED ON LINEAR EFFICIENCY CURVES GO 1C
1x10'
12 3 4 5 6 7 8 9 10 11 12 13 14 15
Figure 8-33 Comoensated Distribution Curves
8-93
-------
This expression has a limit as M -»•<->. Calculations will therefore indi-
cate a panicle point to cut off the summation. This is illustrated on
the following calculation sheets for the example case.
Loading curves can be generated from the results as indicated in Figure
8-34. The total loading (determined by weighing all the particles ;n a
collected sample) is plotted on the one micron line in order to show
the limit for the loading curve.
Density calculation - The. average density of the particles in the system
can be determined using a dry collection of the particles, i.e., (sample
of flyash from the test separator). This is accomplished by weighing
the dry sample, mixing the weighed sample in a measured quantity of the
analyzer solution and then proceeding through the complete analysis as
though the sample were collected in the normal fashion. The weight of
the oarticles counted by the analyzer is equal to the total dry weight
multiplied by the ratio (volume drawn through analyzer to volume of
sample solution). Calculation of tre total volume of particles repre-
sented bv the test sample is calculated using equation 21. Knowing the
vol'ime and weight, the average density can be calculated.
8-94
-------
IxlO1
3 4 5 6 7 8 9 10
IxlO
Figure 8-34 Loading Curves
8-95
-------
f. D
5-3-72
a/
/233
&7 I .02ft/
7<<(< .*»J>_L
fioo
f~t'*/t Sitr
( ^1
X >2
X >3
* >5
x > /
x > /o
x > /5
PX >2S
X >'^0
X >^0
.,' */*0
X > 3fti)
X >5»o
X >£0£>
^ >f2.0t>
64*4 !»4
(y^-ysefj
.12 66 7
.03997
. 0+41*)
.0/J58
,f>0 372
.OOO H '
.OO002
«v^//i
if f>»f/t'en/
Figure 8-3j Loading Analysis Sheet (Probe C)
8-96
-------
Location
To /a/ /Ovcliriy Q/ Me
Rrtstle Si»,
_f^2
<* < 3
L^-xJ"**
^"••"/'cf_2_
.02795
_3 <«.5 .0/626
_i< x u /__L004L2A.
_7^^._^jo_\.0oo ?jt_
/O<_x_< /£ .QQO /3
/C < x < ?5 >. ooooi
< /.loo
fi>*tit/t S
X > 2
X >3
>' > 5
x > 7
x > /o
x > /5
x > *5
X >'^0
x >Jo
)< ? /(Q
x > 3*>«>
_^^».
>f ^ £ 0 0
X" >/Z00
^r;
.OS /SO
oztss
.00529
. oo/o<>
.000 //
,ooot/
«*'*
Aur/tC
Figure 8-36 Loading Analysis Sheet (Probe B)
8-97
-------
I.D.
a/
i 3
W,-j
-*/UF ;
.009/b
/?/,
2)fl/e 5-3-72
fivfujf Si if
( ^>
X >2
x >3
A' > 5
-
x > 7
X > /0
x > /S
x >a5
X >40
X >3o
X ?/6o
.• > lo*
X i.5»»
X >&00
X >t2.o9
££*•/>'» 1
(*r*"/SC£t
.o'on
.002/7
. OQO 2Z
. 66OO 2
t*:^ ft
/l
------- |