AGILITIES EVALUATION
 OF  HIGH  EFFICIENCY  BOILER
 DESTRUCTION OF  PCB  WASTE
July 1980
               By:
               J.E. Cotter and R J. Johnson
               Contract Number 68-02-3174, Task Number 11
               EPA Project Officer: D.C.Sanchaz
 INDUSTRIAL ENVJRONMENTAL RESEARCH LABORATORY
 US. ENVIRONMENTAL PROTECTION AGENCY
         !ANGLE PARK, N.C. 27711	

-------
 AGILITIES  EVALUATION
OF HIGH EFFICIENCY  BOILER
DESTRUCTION OF PCB WASTE
 uly 1980
               By:

               J.E. Cotter and R J. Johnson
               Contract Number 68-02-3174, Task Number 11
               EPA Project Officer: D.C. Sanchez
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, N.C. 27711	

-------
     FACILITIES EVALUATION OF HIGH EFFICIENCY
           BOILER  DESTRUCTION PCB WASTE
                        by
                   J.  E. Cotter
                   R.  J. Johnson

                     TRW, Inc.
                  One Space Park
         Redondo Beach,  California  90278
              Contract No. 68-02-3174
            Program Element No. C1YL1B
      EPA Project Officer:  David C. Sanchez

   Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
         Research Triangle Park, NC  27711
                   Prepared for
       U.S. Environmental Protection Agency
        Office of Research and Development
               Washington, DC  20460

-------
                                 ABSTRACT

     A rendering plaint byproduct, yellow grease, was found to be contamin-
ated by PCB's from a transformer leak.  The PCB content (under 500 ppm)
determines the method of disposal under 40 CFR Part 761.  For this evalua-
tion, destruction in a high efficiency boiler was evaluated as an alternative
to landfill disposal.  The process steam boiler belonging to the waste
owner, Seattle Rendering Works, was evaluated as a candidate site for waste
destruction.  The logistics and fuel handling requirements were found to be
feasible to set up 1n a short time, and the boiler size and residence time
were determined to be likely to allow high destruction efficiency.  With
99.9% destruction of PCB's, the downwind concentration was estimated by
diffusion modeling to be less than OSHA limits for industrial exposure*
Fuel characteristics of the yellow grease were used to support the recom-
mendation for 100% grease fired as fuel.
     A second high-efficiency boiler candidate was also evaluated.  The
Shuffleton power plant, operated by Puget Sound Power & Light Company,
operates three boilers from a common oil fuel supply system.  The size and
facilities at this site were determined to satisfy all the prerequisites
for high efficiency boilers (40 CFR Part 761), and to best be operated by
blending the waste with the normal fuel oil supply.  A 30% waste blend was
evaluated and found to be completely miscible and feasible with respect to
logistical support.
     A verification test bum was recommended and outlined for either can-
didate site. Current EPA protocol and policy developments for PCB destruction
were found to be appropriate for the preparation of a candidate facility
test poan and an example public notice.
     This report 1s submitted in fulfillment of Contract Nol 68-02-3174, Work
Assignment No. 11, by TRW Environmental Engineering Division, under sponsorship
of the U.S. Environmental Protection Agency.  This report covers the period
10 January 1980 to 1 April 1980, and work was completed as of 1 July 1980.

-------
                                CONTENTS
Abstract
  1. Background information ............................................ 1
          Source of contaminated waste ................... ,...,.... ..... 1
          D1 sposal optl ons .................. . ....... . .................. 1
          Prerequisites for high efficiency boiler destruction ......... 2
          Candidate sites for waste destruction ........................ 2
          Scope of work reported ............. * ..... . .................. .3
  2. Seattle rendering works analysis .................................. 4
          Facility background ................................... . ...... 4
          Plant operations ............................................. 4
          Reconmendations of the feasibility of using the Seattle
            rendering boi 1 er ...................... . , . . . ..... . .......... 6
          Recommendations for facility modifications requirements ...... 9
          Conduct of the waste Incineration ........................... 10
          Envi ronmental impact ........ . ............................... 13
  3. Shuffleton power plant analysis ................................. '16
          Fad 11 ty background ......................................... 16
          Plant operati ons. „ .................... . ...... * ........ . ..... 16
          Recommendations on the feasibility of using the Shuffle ton
            boilers for PCB waste incineration ........................ 18
          Recomnendations for facility modifications .................. 21
          Conduct of the waste incineration ........................... 21
          \
          Environmental impact ............ . ........................... 23
References ...................... .... .............. . ....... . • • . ........ 25
Appendix A.  Local  meteorological  data ........................ - ...... 26
Appendix B.  Test Plan for evaluation of PCB destruction efficiency
            at the candiate fadHtes ................................. 2S
Appendix C.  Example public notice .................................... 36
                                    11

-------
                         1.  BACKGROUND INFORMATION

SOURCE OF CONTAMINATED WASTE
     A quantity of tallow (yellow grease), produced  from  the  rendering of
chicken packing by-products, was contaminated by a transformer  leak at a
packing plant 1n Billings, MT.  The Incident occurred 1n  1979,  and by mid-
1979 the contaminated yellow grease had already been sold to  distributors who
had combined the contaminated material  with other stocks.   Routine FDA Inspec-
tions Identified the PCB contamination, and subsequent analyses of blended
stocks were carried out.  The FDA notified the EPA Region X headquarters in
September 1979 of the results of those  analyses where the PCB content was
greater than 50 ppm, and therefore subject to EPA jurisdiction  (1).
     All grease stocks containing more  than 50 ppm of PCB were  traced to two
tanks at 2900  llth Avenue SW, Seattle, WA; the grease is owned by Seattle
Rendering Works, Inc., Seattle, WA.  The tanks contain (as of the date of this
report) approximately 500,000 Ibs and 600,000 Ibs, or about 150,000 gallons
total (2).  FDA inspection of the grease revealed PCB concentrations  ranging
from 116 to 391 ppm, (3) based on four samples taken from the tanks,  and
duplicate assays.
DISPOSAL OPTIONS
     The contaminated grease, having a  PCB content of under 500 ppm,  does not
require destruction in an EPA-approved Incinerator (1).  The  final  ruling
governing PCB disposal, 40 CFR Part 761 (1), Identifies destruction  In high
efficiency boilers or disposal in chemical landfills as acceptable  alternatives.
     "High efficiency" boilers are defined to include power generation boilers
and Industrial boilers that operate at a high combustion efficiency  (99.9%)  as
defined by the percentage ratio of C02  to CO+C02 concentrations 1n  the combus-
tion gases.  These boilers are assumed to be capable of achieving 99.9%  or
greater PCB destruction efficiency.
     There are two approved locations in EPA Region  X where PCB contaminated
waste may be disposed.  These chemical  landfills are Chem-Nuclear Systems
(Arlington, OR) and Wes-Con (Grand View, ID).  It was estimated by  Region X
that landfill ing would cost a waste owner around 8
-------
 PREREQUISITES FOR HIGH EFFICIENCY BOILER DESTRUCTION  (1)

      The  Regional  Administrator may  grant approval for PCB Thermal destruction

 In a boiler 1f a  nuntier of prerequisites  are met:

      1.   The boiler 1s rated  at a minimum of 50 million BTU/hour,

      2.   The PCB  contaminated waste  comprises no more than ten percent (102)
          of the total  of volume of fuel,

      3.   The waste will  not be  added to the combustion chamber during boiler
          start-up  or shut-down  operations.

      4.   The combustion  emissions  will contain at least three percent (32)
          excess oxygen and the  carbon monoxide concentration will be less
          than 50 ppm for oil  or gas-fired boilers or 100 ppm for coal-fired
          boilers,

      5.   The combustion  process will be monitored continuously or at least
          once each  hour  that  the  PCB contaminated wastes are being burned
          to  determine  the  percentage of excess oxygen and the carbon monoxide
          level  1n  the  combustion  emissions,

      6.   The primary fuel  and waste  feed  rates are monitored at least every
          15  minutes whenever  burning the waste,

      7.   The carbon monoxide  and excess oxygen levels are monitored at least
          once  an hour  and  1f  they  fall below the levels specified, the flow
          of  wastes  to  the boiler  1s stopped Immediately, and

      8.   Records are maintained that Include the monitoring data 1n (5)  and
          (6)  above  and the quantities of PCB contaminated waste burned each
         month.  When burning PCB wastes, the boiler must operate at a level
         of  output no  less than the output at which the reported carbon
         monoxide and excess oxygen measurements were taken.

CANDIDATE SITES FOR WASTE DESTRUCTION (2)

     A survey of potential boiler or Incinerator sites by Region X produced

the following Initial possibilities:

     1.  The steam boiler at the Seattle Rendering Works  plant, located at
         5795 S. 130th  Place,  Seattle, WA;

     2.  The utility boilers at Puget Sound  Power & Light's Shuffleton Plant,
         located at 1101 Lake  Washington Blvd., Renton, WA;

     3.  A kiln at Ideal  Cement Co.,  1n Seattle, WA;

     4.  A utility boiler at the Pacific Power & Light Centralla (WA) plant.

-------
     Sites (3) and (4) subsequently dropped out of consideration,  leaving  the
Seattle Rendering and Puget Power boilers as candidates.
SCOPE OF WORK REPORTED
     EPA Region X requested support in evaluating candidate boiler facilities
from IERL-RTP.  The required support was provided by TRW  under Contract  No.
68-02-3174.
     The scope of work Included an evaluation report for  each facility,  includ-
ing general background information, description of boiler and operating
procedures, and recommendations for modifications required to meet regulatory
requirements and good engineering practice,  The scope also Included  9 Test
Plan for the evaluation of PCB destruction efficiency, and the preparation of
an example public notice of intent to conduct verification testing.

-------
                     2.   SEATTLE RENDERING WORKS  ANALYSIS

 FACILITY BACKGROUND
      The Seattle Rendering Works, Inc.,  facility 1s  located at 5795 S. 130th
 Place, Seattle,  WA, adjacent  to the Duwamish  River and the Foster Golf Course,
 The plant location  1s  Indicated on the topographic map, Figure 1.  The
 vicinity of the  plant  Is  mixed  Industrial-commercial-residential.  The nearest
 residential  areas are  Immediately west of the plant  (about 100 meters), and
 about 400 meters north, on a  bluff approximately 50  meters high.
      The plant produces by-products from edible  rendering operations, Includ-
 ing tallow (or yellow grease).   It operates three shifts.  Seattle Rendering
 1s  also 1n the business of distribution  of bulk  sales of grease, through
 purchases of smaller batches  from other  rendering plants.  The manager 1n
 charge 1s Mr. Wes Benefiel, President.
 PLANT OPERATIONS
      The boiler  at  Seattle Rendering Is  a  dual-fuel  (gas and/or #6 oil) water-
 tube  Cleaver-Brooks  model  (Delta  D-60) with a  design capacity of 35,000 Ibs
 steam/hr.  The oil  burner  1s  a  low-pressure, air atomizing type.  The rated
 gas-fuel  efficiency Is 78$, so  the  design  heat Input 1s 46,000,000 Btu/Hour.
 The furnace  volume  Is 755  cu.ft.  in the  Immediate combustion zone, excluding
 the upper radiant section.  Typical stack  temperatures at maximum firing are
 550-560'F.  The  combustion zone temperatures are  1n excess of 2200°F at the
 back  wall.  A steam-heated preheater 1s  capable of raising Incoming oil temp-
 eratures  up to 100-240eF.  The oil feed  pressure  1s maintained by a gear pump,
which  Is  protected by a duplex 30-mesh strainer on the suction side (4).
     The boiler has the standard set of safety controls, Including a programmed
burner shut-down In the event of flame failure;  low water conditions; low fuel
oil temperature; low fuel  oil  pressure;  low atomizing air pressure; or excess
steam pressure.   Under normal  operation, the oil  metering valve and the
secondary air flow butterfly are regulated by a modulating motor driving both
linkages (5).
     The boiler operates  at full modulation from a low fire to high fire
position, regulated by steam pressure, which 1s directly affected by steam
demand.  Since a PCB destruction program would be required to operate at a

-------
r"T;*^-,
                                   '*c."iTiTTi"v77- ':/
                                     ::£*:•  /  ,'/•
                                   S-t-r . 	•V*- \'J_.   ' / '•
'•V'i."!?"•'  }\

;"•••'••"•"l •»•  .
:..RJYerton.Hei
                  Figure  1.   Seattle  Rendering  Works location

                                          5

-------
 steady rate of waste fuel  consumption, there would be a change  from normal
 boiler operating practice.   If the fuel waste were consumed at  maximum  fuel
 feed rates, then some of the generated steam would have to  be dumped when
 process demands dropped off.  In the case of waste fuel  consumption at  some
 Intermediate rate, auxiliary heating must be made  up  by firing  natural  gas
 simultaneously during high  demand periods, and excess steam would have  to be
 dumped at low demand conditions.
      Facilities for storage of yellow grease at Seattle Rendering Include two
 270,000 Ib capacity heated  storage tanks.  The grease 1s pumpable at about
 100°F, and congeals at lower temperatures.  Other  storage and transport equip-
 ment  Includes two 50,000 Ib Insulated trailers, and  one 20,000 Ib  Insulated
 trailer.   Very large heated storage tanks are rented  at terminal facilities,
 such as at Harbor Island.(6).
      The  processing operations are not expected to be hindered by the proposed
 PCB  waste Incineration.  Presently there 1s  a spare storage  tank at the site,
 and  one of the large Insulated trailers (and the smaller one also)  can be made
 available.
 RECOMMENDATIONS ON THE FEASIBILITY OF USING  THE SEATTLE  RENDERING BOILER
 Suitability of Yellow Grease as  Boiler Fuel  (7)
      The  Inspection of a grease  sample from  one of the contaminated shipments
 (Fujltsukl  Maru II, port tank 09-15-79)  produced the  results, shown in  Table
 2-1; two grades of fuel  oil are shown by comparison.   These Inspection  re-
 sults provide some confidence that the grease could be burned undiluted,
 since:
      • The viscosity 1s between that of 15  and #6 fuel oil, so atonrization
        should  be  good with  a burner that 1s  usually  run with 16 oil.
      • The fuel value of the grease 1s  932  of a 16 oil value.
      • The ash  content  1s  low.
      Furthermore,  a  telephone report obtained from Mr.  Keith Markegard at
 Pierce Packing*confirmed that they had burned 600,000  Ibs of grease in their
 steam boiler, with  good combustion characteristics.
*Bmings, MT.

-------
TABLE 2.1   LABORATORY INSPECTIONS

Viscosity
SUS S 122*F
SUS g 200'F
Density, Ibs/gal
@ 122°F
@ 180-F
Btu/lb
Moisture %
Water & Sediment X
Carbon %
Hydrogen %
Oxygen X
Sulfur %
Ash %
Grease

143.7
59.8

7.459
7.266
16,779
0.41
2.8
76.6
11.9
10.5
0.01
0.12
#5 Fuel Oil

40
•

@ 60', 8.0
•
18.250
0.05
1
87.5
10.2
-
1.1
—
16 Fuel Oil

300

& 60*. 8.33
18,000
0.05
2
88.3
9.5
-
1.2
0.1

-------
 Estimated Destruction Efficiency
      The residence time of the PCB's  in  the  furnace  combustion zone has to be
 determined as  part of the evaluation  of  destruction  efficiency.;. At a waste
 feedrate of 6 GPM, the actual  volumetric flowrate of combustion  gases  through
 the 755 cu.ft.  furnace volume  is 43,097  ACFM, yielding a  residence tine of
 1.1 seconds.   The  residence time will  be twice as  long at a feedrate of 3 GPM,
 since the gas  volumetric rate  is half as much.
      The estimated destruction efficiencies  under  these conditions depends on
 a number of factors,  including:
      t  Comparison of residence time  and temperature to those specified in
         40 CFR,  Part  761,  Subpart E:  1200-C  (2191°F) 9 2  seconds, or 1600'C
         (2912°F) 0 1.5 seconds dwell:
         {These conditions  are  for waste  Incinerators, burning PCB wastes of
         any concentration).
      •  Estimated  efficiency of the Cleaver-Brooks burner, and the ease of
         atomization of the  grease fuel,  as compared  to 16 fuel oil.
      •  The likelihood of achieving high  destruction efficiencies with PCB
         concentrations  of a few  hundred  ppm, vs. percentage range concentra-
         tions found in  some other wastes  (it's always easier to remove high
         percentages of  high concentrations).  99.935  destruction efficiency 1s
         often obtained  with high  efficiency  incinerators and high PCB concen-
         trations.
      The  likelihood of  achieving 99.9 + % PCB destruction is estimated to be
 very  good at a 3 GPM  feedrate, since the time and temperature relationships
 may approximate those of good  incinerator practice.  It is less certain that
 a 6 GPM feed rate would yield  equivalent destruction efficiency, but the
 operation at 6 GPM would certainly be preferred for  reducing the program
 length.  Therefore, the verification test should determine the destruction
 efficiency and combustion temperature at both feedrates.
     The boiler operation under automatic control will allow full modulation
 from  low fire to high fire, regulated by steam pressure.  High fire fuel
 demand would be about 6 GPM, since the Btu content of grease and #6 oil are
similar.  Low fire demand is about 25* of high fire, or about 1.5 GPM  (4).

                                       8

-------
Recommended Exceptions to Standard Prerequisites  for Boiler Specifications
and Operations	
     Although the Cleaver-Brooks boiler at Seattle Rendering  does not meet the
standard prerequisite of 50 million Btu/hr heat Input (at  full output),  1t
comes close at 46 million Btu/hr.
     It 1s recommended that the standard prerequisite of a maximum blend of
10% waste (with the remainder fuel oil) Is less appropriate for  the Seattle
Rendering program than a 100% waste fuel choice.   The 10%  blend  1s based on
the possibility of using any waste, Including those with very low fuel value.
As already noted, the Btu content of yellow grease 1s very close to #5 or #6
fuel oils, and the combustion characteristics are judged to be good.  Further-
more, there are no present facilities available for blending  of  grease and
fuel oil.
RECOMMENDATIONS FOR FACILITY MODIFICATIONS REQUIREMENTS
Logistics
Verification Testing;  Use 50,000 Ib (-6700 gal) trailer,  go for 1.5 days,
then switch to small (3000 gal) trailer for 0.75 days while  large  trailer is
returned to port for reloading.  This assumes that flexible  lines  can be
directly coupled to the oil feed pump for the trailer.
Post Test;  Transfer from 50,000 Ib trailer to heated 270,000 Ib storage tank
(9 days storage).  The trailer will have to make 6 round trips to  get the
tank filled.  Since turn-around time (gate-to-gate) for a  trailer  1s  about  1
hour for loading, (or unloading) and the port-to-plant transit time  1s  about
0.5 hr, 2 loads/day will be'easy.  Three  days of waste  fuel  can be brought
into the plant in one day.
A copy of each loading ticket should be recorded.  If the  transport  of  the
grease falls Into a hazardous waste category,  compliance  with the necessary
state and DOT manifest regulations will be required.
Trailer-to-Tank Transfer
     A temporary dike must be set up around the transfer area, sized to retain
up to 10% of the storage tank capacity  (10% X 4839 cu.ft.  -  484 cu.ft.). This
can be made up with sandbags and 6 mil  polyethylene.  Vent control will  not be
required because of the low volatility  of grease and PCB's.

-------
     The  trailer and transfer hoses should be dedicated during the program.
The existing  300 GPM transfer pump can be used, along with existing transfer
hoses and quick-disconnect couplings.
Waste Fuel Handling System
     This 1s  Illustrated schematically 1n Figure 2.  Modification or additional
equipment requirements are as follows;
     a.   Provide booster pump (20 GPM centrifugal, 0.5 HP) prior to existing
          tank to oil feed pump (Note: existing 300 GPM pump cannot be used as
          both a loading and feeding pump, since these two requirements will
          conflict).  Disconnect existing oil feed and return lines.
     b.   Tee  from bottom discharge line to booster pump suction: break con-
          nection at branch to second storage tank, to Insure that no contam-
          inated grease could be run down to second tank.  Provide valves on
          suction and discharge to allow removal for repair if necessary.
     c.   Install 2 inch line from booster pump discharge to existing oil pump
          suction—run through back wall of boiler room to minimize run length.
          Provide 1 Inch Insulation, and return line to storage tank to maintain
          high flow velocity to prevent setting up,
     d.   Install pressure gages upstream of basket strainers at oil pump stand.
     Additional measurement points that should be installed for adequate boil-
er monitoring during PCB Incineration include:
     •  Flowrate indicator and recorder downstream of oil regulating valve.
     •  A1r windbox pressure indicator.
CONDUCT OF THE WASTE INCINERATION
Verification  Testing
     Background emission levels will be determined by running the PCB emissions
testing procedure while the boiler is burning #6 fuel oil at a steady feed rate
A PCB analysis will  be required for fuel oil samples.
     Normally, three separate tests are conducted at the same conditions for
verification of the  efficiency of PCB waste Incineration.  In this program the
performance at multiple grease feedrates is needed to decide on recommended
operating conditions.   Therefore, the verification test plan will be:
                                     10

-------
                                   CONNECTION TO
                                   EXISTING RETURN LINE  (c)
CONNECTION TO
EXISTING FUEL OIL
PUMP SET & FILTERS
(c), INSTALL PRES-
SURE GAGES (d)
                                    DISCONNECT OIL FEED &
                                    RETURN LINES     (a)
                                   BOILER ROOM
TEMPORARY 2" LINE (INSULATED)  (c)
    V          I  BL1
BLIND EXISTING
    \        y  CONNECTION TO
      V	S\ 2ND
                v
     270,000 Ib
       STORAGE
                           300 GPM PUMP

                         EXISTING TRANSFER LINE
                                               TEMPORARY  2" RETURN  LINE(c)
                                                BOOSTER PUMP  (a)
                 Figure  2.  Waste Fuel  Handling Schematic.
                                      11

-------
     •  Duplicate tests at 3 GPM feedrate.
     t  Duplicate tests at 6 GPM feedrate.
     Boiler fuel control must be set on manual during these tests.
     If automatic boiler fuel control 1s used during subsequent Incineration,
the feedrate should be kept from going below the recommended feedrate by dump-
Ing steam 1f process demands are low.  If the verification test demonstrates
that at least 99.92 PCB destruction 1s achieved, then the remainder of the
waste can be burned with confidence.
     The emissions tests during the verification burn will be carried out
according to the GCA draft test manual (8).  The test procedure will use an Inte-
grated sampling train with a solid adsorbent (florlsll) section that is effective
in capturing PCB's 1n low concentration.  The sampling will be conducted
Isokinetically, with the sample probe positioned at a number of pre-specified
points during the traverse of the boiler stack, at two ports 90  apart.
The sampling effort Is expected to take 3 hours to get sufficient PCB capture
for analysis detection.  The adsorbent material and impinger rinses will be
sealed and taken to a laboratory that Is experienced in PCB analysis.  Both
gas chromatography (GC) and gas chromatograph mass spectroroetry (GC/MS)
analysis methods will be used for analysis of all PCB isomers.  A quick
prescreening test for total chlorinated hydrocarbons will be used to evaluate
verification tests.  A detailed description of the sampling and analysis
methodology for PCB destruction verification 1s given in Appendix B.
Operations and Surveillance
     The boiler will be started up on natural gas fuel to reach steady opera-
tions, as determined by stack measurements of temperature, CO, and 02.  The
grease circulation loop will then be started, Including the feed preheater.
When the feed temperature reaches 140-150*F, the grease fuel can be directed
to the burner and the gas fuel gradually reduced, such that stack temperatures
and steam pressure remains constant.  The grease fuel should not be heated
above 180*F, since it may smoke at any point where 1t comes Into air contact
(probably the storage tank).
     Surveillance of the boiler and fuel handling areas must be maintained  on
a 24-hour basis, with the following data to be recorded:  CO, 02*  fuel  rate.

                                      12

-------
     t  Boiler shut-down should be done 1n a similar manner to start-up, with
        gas fuel used to finish off the firing before shut-down.  The grease
        should be drained from the fuel line before any shut-down; otherwise
        it should be kept hot and circulating.
     t  If a fuel leak 1s detected in the system that cannot be corrected by
        a fitting adjustment, then the grease circulating loop must be shut-
        down and drained prior to repair.  The grease already spilled should
        be allowed to congeal and then be scraped up and returned to the fuel
        storage tank.  Scraping congealed grease Into the hatch of an unload-
        ing trailer may be possible.  Final spill cleanup can be done with a
        hexane solvent.
     •  The contaminated waste will be transported 1n a dedicated trailer,
        and transfer hoses will also be dedicated to the disposal program.
        This procedure will avoid the mixing of contaminated grease with non-
        contaminated grease.
Cjean-Uj)
     The following equipment clean-up will be required:
        -  Trailer
        -  Transfer hoses
        -  Transfer pump
        -  Booster feed pump
        -  Feed and recycle lines, including fuel oil pump.
     The entire transport-and-fuel feed system should be flushed with uncon-
tamlnated grease and combustion continued until the grease PCB concentration
drops below 50 PPM.   The temporary connections  can then be removed and the #6
fuel oil feed loop restored.
ENVIRONMENTAL IMPACT
     The anticipated Impact of PCB waste burning at the Seattle Rendering
facility was modeled with a point-source dispersion model, actual stack dimen-
sions,  and a range of local  meteorological conditions, as reported from the
PSAPCA* station  at Tuckwlla/South  Center,  about 2 miles  south  of the plant,
(Appendix A.).
 *Puget  Sound Air Pollution Control  Agency
                                     13

-------
      The flat-terrain  dispersion  model  1s only  useful for Impact analysis up
 to the point of a significant  terrain change.   The prevailing winds at the
 source location (35Z of the  time)  are south to  southwest, with neutral (Class
 "0")  atmospheric  stability.
      There is a sharp  terrain  rise of 50 m. only 400 m. 1n a prevailing (SSW)
 direction from  the  rendering plant source.  The calculated effective stack
 height  is 40-65 meters.  Model predictions at 400 m. of PCS concentrations at
 the plume center-line (which  is worst-case conditions) are as follows:
Feed Rate
6 GPM
6 GPM
3 GPM
3 GPM
Wind Speed
10 m/sec
4 m/sec
10 m/sec
4 m/sec
No PCB Destruction
6.4 yg/rn
16.0 yg/m
1.6 yg/m
4.0 yg/m
99. 9X PCB Destruction
6.4 ng/m
16.0 ng/m
1.6 ng/m
4.0 ng/m
     The modeling estimates are based on calculated emission quantities as
follows:
     •  6 GPM feedrate using worst case of 400 ppm PCB's:  9110 SCFM stack
        discharge, 0.12 gm/sec uncontrolled, 0.12 mg/sec controlled emission
        rates.
     •  3 GPM feedrate, using moderate case of 175 ppm PCB's 4555 SCFM stack
        discharge, 0.03 gm/sec uncontrolled, 0.03 mg/sec controlled emission
        rates.
     The ground-level concentration of the plume under flat-terrain assump-
tions may be more realistic at 1.5 to 2 km downwind, after the plume has
passed over the bluff.  The atmospheric PCB concentrations for a 99.9J effi-
cient destruction operation are in the range of 0.1 to 0.4 ng/m .
     Techniques for ambient monitoring of these extremely low PCB concentra-
tions are still being developed, but it 1s recommended that an attempt be
made to make such a measurement at the most likely downwind position, in
support of the verification test.  The ambient monitoring methodology 1s
described In Appendix B,
                                     14

-------
     An analysis of downwash concentrations of PCB's was also done, using the
          * an area source.  In a downwash model, dispersal of pollutants
occurs In the Immediate, ground level vicinity of a source, rather than
through an elevated plume.  Calculated atmospheric concentrations of PCB's
were estimated to be around 0.2 ng/m3 with 99.9% destruction,  The OSHA
limit for worker exposure to PCB's on an 8-hour day, 40-hour week basis 1s
1 mg/m3 with 42 percent PCB chlorine content, so that the downwash concentra-
tion certainly does not threaten worker safety.
                                    15

-------
                      3.  SHUFFLETON POWER PLANT ANALYSIS
 FACILITY BACKGROUND
     The Shuffleton Power Plant, owned by the Puget Sound Power and Light Co.(PSPL)i
 1s  located at 1101 Lake Washington Blvd., Renton, WA.  The plant 1s on the
 southern edge of Lake Washington, as Indicated on the topographic map,
 Figure 3.  The vicinity 1s generally Industrial and recreational.  The near-
 est  residential areas are about 1 km east of the plant.
     Shuffleton 1s operated at a steady level of 330,000 Ib/hr of steam 1n
 each of three oil-fired boilers.  The plant was constructed in 1929, and 1s
 expensive to operate.  Consequently, Shuffleton 1s only run when the antici-
 pated load demand in the PSPL service area cannot be met by other generating
 sources (9),  The plant superintendent is Mr. Bill Clark.
 PLANT OPERATIONS
     The Shuffleton boilers are balanced draft type (FD, ID fans); there are
 16 Peabody mechanical burners in each boiler, burning #6 fuel oil heated to
 200"F.
     The available measurements Include air/fuel  ratio, oil pressure, wlndbox
 pressure, furnace draft, fuel oil temperature, air preheater temperature and
 pressure, and oxygen in the stack gases.  Fuel oil 1s not presently metered,
 although an orifice exists in the fuel oil line.
     The fuel oil is supplied from large storage tanks to two smaller service
 tanks, each having a working capacity of 800 bbls.  The tanks are run alter-
 nately, on about a 5-hour cycle, while 3600 to 3900 bbl per day of fuel oil
 are consumed.  The tanks have open hatches at the top which could be used
 for pumping in other fuels to get a mixed fuel.  The fuel is kept at 150'F
 in the service tanks.  The fuel oil 1s pumped out of the service tanks by
three 80 GPM pumps, and distributed through heat exchangers to raise the fuel
temperature to 200*F.  The fuel is then Injected Into each of the burners 1n
the three boilers.   The viscosity of #6 oil is sufficiently lowered by the
200°F preheat to get good burner and combustion performance (9).
     The operations of the Shuffleton power plant are not expected to be
hindered by the proposed PCB waste Incineration,  as indicated by the follow-

                                     16

-------
               ^wLdWTsn rW^r
               v / -  py ft • - •• ^m
               fi1'  fr^ **^!  v:\w
Figure 3. Shuffleton (Puget Power) Plant location
           17

-------
 1ng discussion on feasibility.   Three-shift  operations  coverage of the boiler
 and fuel  handling system will  provide  the necessary surveillance of combus-
 tion performance that 1s needed for PCS  Incineration.   There Is reason to
 believe that close surveillance 1s  mandatory for maintaining good combustion
 at Shuffleton.   The plant has  received frequent citations for opacity vio-
 lations (10), despite a test of the No.  1 boiler which  showed particulate
 emissions within compliance (11).   A combustion'consultant's report submitted
 to Puget  Power Identified possible  causes of poor combustion to Include
 Inadequate fuel  atomlzatlon (due to low  fuel  oil temperature) and dirty
 burners (12).

 RECOMMENDATIONS  ON THE FEASIBILITY  OF  USING  THE SHUFFLETON BOILERS FOR PCS
 WASTE INCINERATION
 Suitability of Yellow Grease as a Boiler Fuel
      The  first  consideration for suitability 1s the determination of nrisd-
 b111ty, or degree of mixing, of grease and 16 fuel oil, since using a blended
 fuel  1s the technically simplest method  for  waste Incineration 1n a large
 boiler.  Yellow grease and 16  fuel  oil mixing tests were conducted by North-
 west  Laboratories at two blend ratios  (10* and 30% grease, by volume) and two
 temperatures (180*F and 200*F).  The results  of these four tests were consist-
 ent;  the  blends  were misclble,  with no stratification,  at all conditions (13).
      The  Inspection of a grease sample from  one of the  contaminated batches
 by Northwest Laboratories produced  the results reported In Table 2-1; two grades of
 fuel  oil  are shown by comparison.
     The  previous  Inspection results provide  further confidence that the
grease will  combust well  with 16 fuel oil, since:
     •  The  viscosity  1s  between that of  #5  and 16 fuel oil, so burner
        atomlzatlon with  the blended fuel should be good.
     t  The  fuel value of the grease 1s 93* of a #6 oil value.
     •  The  ash content 1s low.
Estimated Destruction Efficiency
     The residence time of the PCB's 1n the boiler combustion zone 1s a con-
sideration 1n the evaluation of destruction efficiency.  At a typical fuel

                                     18

-------
feed rate of 38 6PM, and 4-5% excess cwygen, the calculated rate of combus-
tion gas flow 1s 1111 std. cu ft/sec (14). The estimated residence time 1n the
lower half of the boiler 1s 2.5 seconds, with another 2.5 seconds dwell  in
the upper zone (15).  The temperature 1n the combustion zone 1n the vicinity
of the burners, has been measured at 2800»F (16).
     The estimated destruction efficiencies under these conditions depends on
a number of factors, including:
     t  Comparison of residence time and temperature to those specified  in
        40 CFR, Part 761, Subpart E: 1200°C (2192°F) @ 2 seconds, or 1600°C
        (2912°F) § 1.5 seconds dwell:
        (These conditions are for waste incinerators, burning PCB wastes  of
        any concentration).
     •  Efficiency of the burners and the ease of atomization of the grease
        and 16 fuel oil blend.
     •  The likelihood of achieving high destruction efficiencies with
        blended PCB concentrations of less than a hundred ppm, vs. percentage
        range concentrations found in some other wastes (it's always easier
        to remove high percentages of high concentrations).  99.99% destruc-
        tion efficiency 1s often obtained with high efficiency Incinerators
        and high PCB concentrations.
     The likelihood of achieving 99.9% PCB destruction is estimated to be
good, given the similarity of the boiler conditions to good incinerator  prac-
tice, and the compatibility of the grease and #6 fuel oil.
Becoroiended Exceptions to Standard Prerequisites for Boiler Operations
     It is recommended that the PCB incineration be conducted with a grease-
to-fuel oil ratio of 30/70 instead of the standard prerequisite ratio of
lo/90.   There are good arguments for using a 3055 blend, since the length  of
     needed to complete the effort will  be shortened by a proportionate
      , and the grease 1s not expected to cause any combustion problems  or
      ble heat reductions.  The environmental  consequences of burning a  SOX
blend, and the logistical limitations of supporting a higher grease percent-
     are discussed below.
                                      19

-------
 RECOMMENDATIONS FOR FACILITY MODIFICATIONS
 Logistics
     The two alternating service tanks (800 bbl working capacity) will  be
 used for temporary storage of the grease (as a blend), and for making up a
 blend in the off-line tank; this should be done while the tank Is filling
 with #6 oil, to get the benefit of agitation during the fill.
     If a 10% blend 1s made up, 1t would require 3200-3400 gallons of grease
 to be added to the service tank, and one trailer (6500 gal. capacity) for
 support.  The single trailer could probably split a load between two alternate
 tank fillings, and then return to the port.  The plant-to-port round trip,
 with filling, 1s estimated to take 3 to 4 hours.  This turnaround time will
 not be limiting, given a 5-hour cycle time with the tanks.
     If a 302 blend 1s made up (as recommended), it would require 9600-10,200
 gallons of grease addition per tank, and two trailers for support.  They
 would both be required to return for reloading following each tank filling.
 Blends much greater than 302 would cause a logistical limitation.
Traller-to-Tank Transfer
     The rundown line between two large storage tanks and thessrvlce tanks has
 a connection that could be used for transfer, but top-filling of the service
 tanks is preferred for the agitation benefit.  The transfer area must have
 a temporary dike set up, sized to retain a trailer load (under 1000 cu.ft.)
 using sandbags and polyethylene sheet.
     The transfer pump and temporary power connections must be provided, along
 with transfer hoses.  The transfer pump should have about a 200 GPM capacity,
 to prevent any unnecessary hold-up time for trailers.
CONDUCT OF THE WASTE INCINERATION
Verification Testing
     Background emission levels will  be determined by running the PCS emis-
sions testing procedures while the boilers are burning uncontamlnated #6 fuel
oil.  A PCB analysis will  also be done on fuel oil samples.
                                      20

-------
     The schedule will Include emissions testing on three sequential  days  of
operation, to determine what day-to-day differences may occur.   This  series
of tests will be done with one batch of blended grease and fuel  oil each  day,
so that PCB's will not be burned unless emission testing 1s In  progress.
Continuous oxygen and carbon monoxide monitoring of stack gases  will  also  be
done during the verification test sequence.  If the verification test demon-
strates that at least 99.9% PCB destruction is achieved, then the remainder
of the waste can be burned with confidence.
     The emissions tests during the verification burn will be carried out
according to the 6CA draft test manual (8).  The test procedure  will  use  an
Integrated sampling train with a solid adsorbent (florisil) section that  1s
effective in capturing PCB's 1n low concentrations.  The sampling will be
conducted 1sok1netically, with the sample probe positioned at a number of
prespecified points during the traverse of the boiler stack.  Since equal
amounts of fuel are supplied to each of the three identical boilers,  then 1t is
sufficient to sample one stack rather than all three.  The sampling effort
1s expected to take 3 hours to get sufficient PCB capture for analysis
detection.  The adsorbent material and Impinger rinsed will be sealed and taken
to a laboratory that is experienced 1n PCB analysis.  Both gas Chromatography
(GC) and gas chromatograph-mass spectrometry  (6C/MS) analysis methods will be
used for analysis of all PCB Isomer,  A quick prescreenlng test for total
chlorinated hydrocarbons will be used to evaluate verification tests.  A
detailed description of the sampling and analysis mehtodology for PCB destruction
verification is given in Appendix B.
Operations and Surveillance.
     The grease should not be pumped Into  the fuel  oil service tank until it
nas been established that the boiler Is operating satisfactorily, as evidenced
by the combustion parameters holding at the criteria specified by regulation
40 CFR 761, and visual Inspection of the burner flame.   After the blended
fuel firing  1s initiated, a relnspectlon of the combustion parameters will
be required.
     Surveillance of the boiler and fuel handling  areas  must be maintained  on
a 24-hour basis, with the following data to be  recorded:  CO, 02, fuel rate.
Periodic measurements of COz will  also be  done.

                                      21

-------
Clean-Up
     The following equipment clean-up will  be required:
     -  One to two trailers
     -  Transfer hoses
     -  Transfer pump
     If any spill or leak occurs during trailer unloading,  the equipment  used
to scrape up the congealed grease will also require cleaning.   A hexane sol-
vent can be used for cleaning, and added to the fuel oil  1n small quantities.
The blended grease and fuel  oil will probably not be considered a contaminant
in the fuel oil handling system, since the PCB content will be well  below
50 ppnu  The system will be purged by the routine use of uncontamlnated fuel
oil.
ENVIRONMENTAL IMPACT
     The anticipated Impact of PCB waste burning at Shuffleton power plant
was modeled with the EPA "Valley" model, which takes Into account a certain
amount of terrain variation.  The computer program was run by the PSPL
quality staff, using terrain and stack dimension parameters already set  up for
the Shuffleton plant and vicinity.
     The model was run with a low wind speed of 2.5 m/sec and a prevailing
southwest wind, over a range of atmospheric stability conditions.  Local  wind
speed and direction data are given in  Appendix A.   The most stable condi-
tions ("F" stability) usually yield the highest ground concentrations of a
pollutant, since dispersion by air mixing is minimal.  The following normal-
ized data factors were computed for stability Classes A thru F (A being the
most unstable).  The data factors are multiplied by an estimated emission rate
to get a maximum 1-hour concentration at the indicated downwind distance.


Data Factor +
Location of maximum
1 Hr PCB concentra-
tion @ ground! evel
Stability Class
F
3.95

2 km NE

E
2.11

2 km NE

D
0.027

12 km NE

C
0.129

4 km NE

B
0.245

2 km NE

A
0.20

near plant

                                      22

-------
     The maximum 1-hr PCS concentration can then be estimated, using "F"
stability conditions, from the emission rates determined as  follows:
     a.  lot blend, 3200 gal. of 400 ppm (worst case)  grease burned over  5-hour
         period yield 0.24 g/sec (no control), or 0.24 mg/sec (99.935 control-
         led) emission rate;
         Worst case model prediction for downwind PCB  concentration «
         3.95 X 0.24
         " 0.9 vg/m3 (no control), or 0.9 ng/m  (99.9% control)  at 2 km NE  of
         plant.
     b.  30% blend yields three times the emission rate above—0.71 g/sec
         (no control), or 0.71 mg/sec (99.9% control).
         Estimated downwind PCS concentration at 2 km  NE then will be:
         2.7 yg/m3 (no control), or 2.7 ng/m  (99.9* control).
     These estimates can be compared to the OSHA exposure limit  for PCB's.
The OSHA allowable concentration for worker exposure on an 8-hour day,  40-hour
week basis 1s 1 mg/m3 with 42 percent chlorine content 1n the PCB's.
     Techniques for ambient monitoring of these extremely low PCB concentra-
tions are still being developed, but It 1s recommended that  an attempt  be made
to make such a measurement at the most likely downwind position, 1n support
of the verification test.  The recommended methodology 1s described 1n
Appendix B,
                                     23

-------
                                REFERENCES


  1.  40 CFR Part 761, Polychlorlnated Biphenyls (PCB's)  Manufacturing,  Pro-
     cessing, Distribution 1n Commerce, and Use Prohibitions,  Federal
     Register 44, 31514-31568, May 31, 1979.

  2.  K. Feigner, status review meeting, Region X, 01-14-80.

  3.  Letter (09-28-79) from R. V. Mlecko (FDA) to 0.  Donaldson,  EPA
     Region X.

  4.  Letter (01-03-80) from H. I. Bosler, Bosler Energy  Systems, to  K.
     Feigner, Region X, also, personal communication  with  M. I.  Bosler.

  5.  Operating Instructions, Cleaver-Brooks Burner, Model  B.

  6.  VI. Beneflel, personal communication.

  7.  Laboratory Report to Seattle Rendering Works (01-15-80),
     Laboratory Report to Seattle Rendering Works (01-18-80),  fuel inspec-
     tions of yellow grease performed by Northwest Laboratories, Seattle,
     Washington.

  8.  Test Plan for Evaluation of PCS Destruction Efficiency  1n Industrial
     Boilers, GCA June 1979.

  9.  Personal communication, W. A. Wood, Assistant Superintendent, Puget
     Sound Power & Light Company.

10.  Puget Sound A1r Pollution Control Agency, Inspection  reports.

11.  Atmospheric Emission Evaluation. Alsid, Snowden  & Associates report to
     P5P&L Company, February 1979.

12.  Boiler Combustion Efficiency, Shuffleton Power Plant, Puget Sound
     Power & Light Company, Wei land, Lindgren and Associates,  March  1979.

13.  Laboratory Report to Seattle Rendering Works (01-28-80),  mlsdblHty
     tests of yellow grease and #6 oil.

14.  Air Pollution Engineering Manual, AP-40, U.S.  Environmental  Protection
     Agency.

15.  PSP&L Company Drawing RB430, Boiler cross section,  Shuffleton Power
     Plant.

16.  Personal  comnunlcatlon, T. Van Decar, PSP&L Company.
                                    24

-------
                                APPENDIX A
                        .LOCAL  METEOROLOGICAL  DATA
1.  Local wind speed  (1/80)

2.  Local wind direction  (1/80)

3.  Location of met station  (#17)

-------
         - STA1IONI 1UK*11A. SOUIH CENTER                •••4MBIE.NT A|H QUALITY UATA tt£f>OH1*.*     DATA REf0HT • AO90R46I-I
       '    AGCNCVI PUCCT SOUND AIR POtL.CONT AUTH STATEl WASHINGTON        SITCI  IT0Z-KSO          VEARI »VT»
       I {  POLLUTANT! NINO SP RESULTANT           UMlTSl HPH               DATA-FORMAT I JUN.A      HOCL C00t« 61103
         —iwr AVC SHOWN UMDEN ENDING HR          PMOJECTI 01              COMV FACTONI NONE USED
PACE I  20
MONTH I JANUARY
NETHOOt 20
RUN-DATE fS/07/7*
                                                                          HOURS PST
           tw
                  «t   •?   13   •*   is   i*   or   oo   o*   i*    n    ie    u    in    is   |6    IT    10    i»   20   zi   77   n   7* MM  HN
at
41 • 1
02 T 4
-03-*— 1
•* r i
06-S-4
•0«-f — «
Us
u s
-»S-H-
ir N
-!•-? -
i» r
2ft S
•ZI-S--
Z2 H
21 I
-24- M-
25 r
26 r
2T-S-
2« H
30 T
31 W
.* 7.^ ?.» 0.6 3.b t*t 4.1 I
.2 ».? 1.6 l,« 2.0 3.0 X.4 4
T 1.* 3.6 S.J 2.1 7.1 6.1 1
*— 0. 0 —•.••- 0.1 -O.I •-».*- 1.4- 4
0.« 0.2- 0.2--O.J -1.5 0.4 0
0.0 0.1 1.7 |.S 0.6 0.2 0
2.* 1.3 7.* 3*1 7.4 3.0 y

I J
4 1
T 10
T— 5
1 1
0 0
• 4

S 4
T 3
1 10
6 I
3 1
6 1
7 -1,
6
1

» B.4 3.1 Z.I 3.1 T.r 3
1 4.6 3.« 5.4 2.2 7.) 4
7 14.7 11.3 l«.4 13.* l*.t 4

* S
« 2
1 ?

5 7
ft 3
1 6

• r
S 1
7 S

• 4.
7 7
7 7

.3 1.
1 7
5 4.
3 7
1 7

.• 7
>1 2
.3 7
0 1
5 1
« 1
6 1
5 1

» 3
» 3
4 2
1 3
3
0
«
4
a
*
4
^

1 7
I 1
• 7
1 1
•
3
1
3
S
0
7
7

7 7.0 1
3 S.4 I
S 0.5 4
5|4.4 •
010.7 7
S 3.7 I
1 *.« 2
7 6.6 3
S 7.S 7
•14.7 6
6 0.3 7
6 6.* 2
1 *.4 4
6 7.4 7

          -NO	31	31	31-— 31 —31-  31-  31 - -31— »l  - 7»-   3»   31   30   3»   M   30   31 -  31   31   31   31   31   31   31 -
           HN    7.T  2.T  7*4  7.1  7.7  7.4  7.9  2.7  7.0  3.5  3.«  4.6  4.4  5.4  S.4  4.5  4.0J  4.1  3.0  3.7  7*9  7.4  2.6  2.3     3.3
           N>    *.* 14.4 II.•  T.I  a.7  7.5  8.2  *.4 |0.| |7.7 |4.2 14.S 11.3 13.* 17.4 17.3 II.•  ».* 10.I  *.0  9.1  7.7  T.O  7.||*.5
           	-  - 	TOTAL-NUHBtM OF HOUMLV AVCNAGESI   737         -     • IMOICATCS 4 II HOURS AVA1LAW.E f» OAIL* AVERAGE

-------
        i\jMtt\.fc» SOOTH cunt*                »**AMtitMY AIM OUALIIV OAT A Mti»om*»»     DATA HUMNT  • AO?OIU«I-I   PAKI  21

MCMCVI PMtf SOUND AW POLL.CONt  AUlM S1A1CI WASHINGTON        SITEl |782-*S6          VCAHI 1979                MONTH I JANUAItr

POLLUTANT! HIND OH RESULTANT           UM115I DtGNfES           0ATA-FOMMATt KKKX.      POLt COOCI 61104          NLTMOOt 2*

H* AVC SHOWN UNDER CNOIN6 MR          PROJECT! 81              CONV f ACTOWI NONE USED  NOTE! ••• DEMOTES CALM    RUN-OATI 05/«7/79
OA 8)   01   02   03   04   OS
                          057  Mil
                     •S3  034  32*
                317- 3**  357  349
                343  3S6  343  039
                •95  IM  !•!  3**
                34*- 62*  357  34*
                          2M
                2M

                IM  IM  216  I8S
                          351  357
                161 - |*3  174  192

                MS  857  Ml  161
61 M
•2 T
64 T
6SF
-MS
67 S
M N
-*9 T-
16 II
11 T
-12 f
13 S
14 S
-IS ••»-
16 T
17 II
18 T
19 T
26 S
-n s-
22 N
23 T
-24 «
25 T
2*r
27 S
26 S
2* M
-3* T
31 M


342
694
-355
•31
1S2
— .
143

-IM
IM
•58
. - .


16*
221
196
-359
172

.- .._.

1*8
685
169
- oMtl
•rw
892
351
326
2M
-
113

- 162
191
36*
_ — .
IM

_ ... .
247
1*8
— 8M-


_. 	

185
• 174—187
356



354

_.. ....
101
HOURS PST
67 68
•75 Ml
358 338
35* 025
069 077
•42 34*
352 357
12*

-1*5
138 085

18* 17*

116 12*
_ 	

326
1*6-
2*3 »9S
IM 19*
1*7 -2»5

1*3


^A6K f. 6^8E
CW9 «V9



340-346
Ml 354
0*
•*7
326
••2
0*0
3^6
3*5
201

•206

•26
1*2

13*
-


2M

1*2
..

17*


IV*



_
••1
10
IM
•M
353
•76
32*
351

143

185
•36


147
-


17*

168
206
1*6
181



._ . . .



•28
II
1*2
• 17
343
•M
346
346

691
337

638
206





172
228
IM

200
194
•13

IS2
2*4-


•85
625
12
121
62*
612
691
33*
352
681
•93



185
34V
133
337
205
IS*

215


183
172
6*3

221
248


•M
• 12
13
118

• 11
•*3
349
341
110
•62


M7

3SO

-

117
209

215
3M
16*
199
•23
014
163
243
642

614

1*
IM
337
344
•97
348
343
124
040
331


19*
355
12*
003

IM
191
217
216
•36
157
187
624
•38
155
263
•22

352

IS
•71
343
352
•90
34*
354
105
•73
359
349

229
337
12*
33*

132
175
237
216
•66
174
176
OM
• 18

345
•23
049
•18

16
049
346
003
094
346
0*2

•58

333

214
346
006
337

116
187
198
211

204
103
•24
•34
142
352
3**
••5
•OS

17
•76
352
3*0
OM
353
006

•52
1*2
32*
214
210
3SZ
06*
335-

103
192

2*7
201 -

212
•21
65ft

«M-
3*0
•91

351
18
6M
3*0
3*0
•95
349
•92


179
331
213
202
354
344
249

•96
171

101
2M

232
•24

177
•25

•M

348
If
161

353
•M
349
235

1*3
173
339
194
162
3*6
611


659
179

173


6M
626

IM
624


694
• 11
28
IM
•01
351
•*•
347
0*5

126
137
321
169
IM
341
632


•to
195

IM

IM
•90
•03

174
• 17



•37
21

35«
344
•47
342

219

145


176
33*
342


• 10
186

312

211
152


176





22
116
624
327

346
351
11*
1*5
IM


199
323



335
204

329
175




181
613

127


23
690
350
320
105
344
296
097
142
181

199
195
330




205
2M
337
164



IM
162
•49


•7*
150
24 NAM
184 121
3*«
327 340
112 355
35* 3*0
357
320
206
147 359
349
195 357
192 229
029 355
3*0
339
20S
100 335
22* 22*
205 247
355 355
3*0
2M 211
176 232
026
196
176 221
34* 352
3*0
127
355
354
                1*8

               	197- M*
                1*4   174  177  1*7
                1*4   186  1*9  1*1
               -•I9-+44— 175-196
                              26)
                     2M  1*2  1*8
                ITS— IT*—186--


                	   -  355  334
                     33*  342  662

NO	1?—-I* -  IS	IS-   I* • 18-  17—26-16  - IT  - 20   25   23   2*   2«

MX    359  3*6  34*  355  357  3*0  356  357  3*5  353  346  352  3*0  355  359
                                                                                 25   26   24   21    22   IS   17   21   16

                                                                                3*0  3*0  3*0  353   351 358 351  358  356 3*0
                 TOTAL NUNKR OF MOUNLV AVCMMCSl   462

-------
            IS78
flTnOSPHCRlC SflnPLING NETWORK
SOUND am POLLUTION CONTROL REGION
             urn ,.«.o
-------
               APPENDIX B.  EXAMPLE TEST PLAN FOR EVALUATION
                     OF PCB DESTRUCTION EFFICIENCY

STACK SAMPLING
Sampling Methods
    The sampling train to be used for PCBs will  be a modified Method 5  train  as
described 1n "A Preliminary Procedure for Measuring the PCB Emissions from
Stationary Sources?: W. J. Mitchell, U.S. EPA, August 26, 1976.   A solid
adsorbent tube containing 7.5 grams of pre-extracted 30/60 mesh  Florisil is
Placed between the third and fourth Impingers.  Alternative adsorbants  are
XAD-2 and Tenax GC.
    The sampling and velocity traverse will be along two diameters of the
stack for a nominal 40 inch diameter stack, a total of 44 points will be
sampled to provide a representative sample of the flue gas composition. Total
sampling time at a flow rate of approximately 1.0 ft3/min. should be long
enough to yield approximately 10 wg of PCBs, based on a destruction efficiency
of 99.9 percent.
    Sampling will be isokinetlc (±10 percent) with readings of flue gas
Parameters recorded at every sampling point during the traverse.  In the
event that isokinetic sampling cannot be maintained, the train will be  shut
down and the problem remedied.  In the event that either steady operation is
not maintained, or monitored gas parameters (CO, C02, 02) are out of the
specified limits, the testing will be stopped until conditions are stabilized.
Steady operation of the boiler will be the responsibility of operating
Personnel and the verification burn coordinator.
lest Matrix
    A series of PCB emission tests at steady boiler operating conditions
WH1 be conducted, such that three separate days of testing are carried out
for a single waste fuel feed rate.  If multiple waste feed rates  are eval-
uated, two successive days of testing will be scheduled  for each  of  tiie
Proposed feed rates.  The test schedule  is Indicated 1n  the test  matrix,
Table A-l.
                                     29

-------
       TABLE  II-1   TEST MATRIX
Sample
Morisil resin
Condensed water
Impinged participates
Train wash
Florisil blank
Train wash blank
Waste fuel
No. 6  fuel
Polyurethane plug
Source
PCU train
D.I.G. hexane
D.I.G. acetone
Trickle value (inte-
grated)
Trickle value (inte-
grated)
High-vol sampler
Container
Amber glass
PCB train impingers     Amber glass
Amber glass
Amber glass
Amber glass
Amber glass
Amber glass
Amber glass
Species
Analyzed
PCDs
PCDs
PCDs
PCD background
PCD background
PCD
PCD
PCB's
  Number
(Maximum)
     5
     1
     1
     4
     1
     8

-------
    In order to provide background data on emissions from the boiler, a pre-
liminary test will be conducted with No. 6 fuel oil nnlj  nnt dnj prior to the
initiation of the PCB test burn.  This test will be for the same duration
and at the same conditions as those to be used for the PCB sampling.   The
test will use a PCB train and will provide emissions data for the boiler as
operated at normal load and firing No. 6 fuel oil.  A measurement of the
precision of stack collection of PCB will be provided by the simultaneous
operation of duplicate PCB trains on the second day of the burn.
Boiler and Stack Gas Parameter Measurements
    During each of the tests, boiler operational parameters will be as
Indicated in Table A-2.  Continuous monitoring of flue gas for CO, COo,
and 02 will be provided.  The monitors will be equipped with a gas condition-
ing system and recorders.  Monitors will be calibrated prior to use and as
required.  Parameters will be continuously recorded during each test.
    Other flue gas parameters will also be measured Including the flue gas
velocity, static pressure and temperature.  These measurements will be made
prior to each run for the determination of proper nozzle size and sampling
rate.  These are also shown in Table A-2.

     Table A-2.  SUPPORTING PARAMETER MEASUREMENTS DURING TEST

     Parameter measured                            Method

C02in flue gas                       Continuous gas analyzer
CO in flue gas                       Continuous gas analyzer
Og In flue gas                       Continuous gas analyzer
Flue gas velocity                    Calibrated pi tot tube/manometer
Flue gas temperature                 Thermocouple
Flue gas pressure                    P1tot tube/manometer
percent moisture In flue gas         Moisture  gain 1n Method 5 tests
Fuel flow rate                       Positive  displacement or differential
                                     pressure  flowmeter
                                    31

-------
Sampling Requirements
    The stack sampling program will require approximately four days of tes-
ting with four team members on site.  A total of three to four tests and one
background test will be conducted.  Other requirements to be provided Include
the following:
    •   three 20 amp circuits accessible to stack;
    •   three 20 amp circuits for monitors;
    •   area for sample recovery.
Quality Control
    In order to prevent any contamination of samples with materials which
may interfere with analysis, the highest level of quality control will be
observed 1n all field testing.  All solvents and resins will be checked for
purity before the sampling program; filter paper and glass wool will be
solvent-cleaned.  Appropriate blanks will be included to provide background
information.  The sampling trains will be all glass and precleaned to remove
any residues.  At no time will plastics or grease be used on the organic
sample portions of the sampling train.  Sample packing lists will be included
for each test with information regarding source lot numbers and preservation/
extraction techniques.
    Each PCS train will be spiked in the field with a solution containing a
known amount of deuterated tetrachloroblphenyl Internal standard.  Laboratory
analysis of the internal standard  will provide information on sampling and
analysis efficiency.
                                    32

-------
AMBIENT MONITORING
Sampling Method and Schedule
     Sampling will be with high-volume samplers, modified as  described  1n
"A Method for Sampling and Analysis of Polychlorlnated Blphenyls  (PCBs)
in Ambient A1r," EPA-600/4-78-048 (Figure A-l).  The airborne PCB 1s
collected on a series of two precleaned polyurethane foam plugs housed  1n
the sampler throat, a 16-cm long threaded, aluminum tube.  Motor  exhaust
1s diverted from the sampler Intake by a duct,  now rate through the
sampler will be maintained at between 20 and 35 ft3/m1n for a sampling
period  time coordinated with the verification test.
     Wind directions expected to prevail during the test period will  be
used to select two nearby downwind sites.  Positioning of the two on-s1te
downwind monitors will be done each morning prior to the start of the
sampling period.  This will be done on the basis of forecast weather, on-
$1te wind conditions, and expected plume dispersal.
ANALYTICAL METHODOLOGY
Stack Sampling
     The following procedures apply after sample recovery 1n the  field.
A prescreenlng analysis will be used, requiring perchlorl nation  of PCBs
with antimony pentachloride.  This procedure converts all PCB Isomers 1n
a sample to a single Isomerlc species, decachloroblphenyl (DCB).   This
approach can only be used for confirmatory purposes and to increase
sensitivity.  Because different amounts of DCB are produced by each Isomer
(mono-, b1-, trl-, etc.), adjustment factors are necessary to convery DCB
values to specific Aroclor mixtures.  Blphenyl, a known emission  from com-
bustion sources and present 1n Aroclor mixtures, 1s converted In  the
perch!or1nation process to DCB.  Thus, 1f blphenyl 1s suspected  to be
present 1n significant amounts, 1t must be determined separately to
correct DCB values.  The exposed Flor!s1l sorbent trap 1s extracted with
hexane 1n a Soxhlet, cooled, and then concentrated to about 5 ml.  Contents
of the first two (water filled) 1mp1ngers are extracted with hexane, dried
by passing through columns filled with anhydrous NagSO^, and added to the
         extract.  The probe and Implngers are rinsed with acetone then
                                     33

-------
                                     Faceplate
                                      Stainless Steel Throat Extension
                                              Polyurethane Foaa
                                              Plug Location
                                     Throat Extension
                                           Wire
                                           Retainer
                                            Motor Unit
                                          Adapter
                            -  Exhaust  Duct
                              (3 • minimum  length)
Figure A-l.
Assembled sampler and shelter vrlth exploded
view of the filter holder.
                          34

-------
hexane.  The rinses are dried with  anhydrous ^SCty and added to the com-
bined extracts.  The combined extracts  are then cleaned by extracting with
concentrated H2S04 (1f the cleaned  extract 1s still colored, liquid chroma-
tography on Fieri$11 can be used).   The extract 1s then made to 25 ml
volume and split Into four portions: three 5-ml volumes for perchlorinatlon
and one 10-ml aliquot for confirmation  studies by GC/MS.  After perchlorina-
tlon, the solution 1s extracted four times with hexane and made to 5 ml.
Analysis for decachloroblphenyl 1s  performed by 6C-ECD.  Chromatographlc
parameters are:
     t  Chromatograph - Any suitable Instrument.
                                     4     fi^
     t  Detector - Electron capture, H or   N.
     •  Column - 1.8 m x 2 mm ID, 3% OV-210 on Supelcoport, 100/120 mesh.
     t  Temperatures - Column 280'C, others not specified.
     •  Carrier gas - Not specified (N2 or H2)» 30 ml/m1n.
     t  Detection limit - Not specified, but standards of 25 to
        50 pg/yl are suggested.  The overall detection limit 1s
        specified as 10 ng DCB 1n a 5 ml perchlorinatlon aliquot.
        This Indicates a minimum detection limit of 2 pg/yl Injected.
     Results are reported 1n terms  of ng DCB per cubic meter of combustion
effluent sampled. GC/MS 1s used to  verify the  presence of PCBs by  pattern
matching with Aroclor mixtures.  The precision of  the DCB analysis Is
estimated to be 10-15X, and recovery of PCBs through the entire sampling
and analysis procedure 1s estimated to  be 85-95%.
Ambient Samples
     Polyurethane foam plugs will be collected from  ambient monitors and
returned to  the laboratory 1n precleaned amber glass jars with Teflon-Hned
caps.  The samples will be Soxhlet extracted for  3 hours with  hexane.   After
cooling, the extracts will be transferred to a Kuderna-Danlsh  (K-D) eva-
porative concentrator and the volume reduced to  5  ml.   The  extract will  then
be divided Into a reserve plus two  portions for perchlorinatlon  according
to the method  outlined above.
                                     35

-------
                     APPENDIX C.  EXAMPLE PUBLIC NOTICE

                  NOTICE OF INTENT TO CONDUCT VERIFICATION
                           TEST OF PCS DESTRUCTION
1.0 BACKGROUND
     A quantity of PCB-contaminated yellow grease, produced as  a  rendering
byproduct, 1s currently stored at 2900 llth Avenue SE,  Seattle, WA.  The
contaminated material, amounting to 1,100,000 Ibs  (150,000  gallons), 1s
owned by Seattle Rendering Works, Seattle, WA.   The PCB1  content  of the
grease (116 to 391 ppm range) qualifies 1t for destruction  in a high effi-
ciency boiler or, alternatively, disposal  1n a chemical landfill  under 40
CFR Part 761.  The Regional Administrator of U. S. Environmental  Protection
Agency  Region   1U will review the proposed destruction of the waste 1n a
high-efficiency boiler to  assure that  the prerequisites of 40 CFR Part 761
are satisfied.
      A  high  efficiency boiler has been selected by the owner for use in
disposing  of PCB - contaminated yellow grease.  It is the U.S.  EPA Regional
Office  intent to require a verification bum to demonstrate that PCB con-
taminated  waste can be destroyed effectively.
     (1)  The boiler capacity is rated at 46 million Btu/hr heat Input,
          as compared to the prerequisite SO million Btu/hr.
     (2)  The recommended waste feed rate will be 100% waste as fuel,
          rather than the prerequisite of 10% waste with 90% normal
          fuel.  Both full and half feed rates will be demonstrated.
 The Condensed Chemical Dictionary, Van Nostrand Reinhold Company, 1977, page
696.  "polychlorinated blpnenyl (PCB).  One of several aromatic compounds
containing two benzene nuclei with two or more substltuent chlorine atoms.
They are highly toxic.  Their chief use is in heat exchange and Insulating
fluids in closed systems.  FDA has prohibited their use 1n plants manufacturing
foods, animal feeds, and food-packaging materials, and has established strict
tolerances on their presence 1n many food products (2.5 ppm).
NOTE:  Because of their persistence and ecological damage from water pollution
their manufacture has been discontinued In the U. S. (1976)."

                                    35

-------
1,1   OBJECTIVES OF THE VERIFICATION BURN
     The purpose of this verification 1s to accurately measure the destruc-
tion efficiency of the  designated high efficiency boiler for PCB, using the
recommended operating conditions.  The test results are expected to reveal
minimum destruction-efficiencies of greater than 99.9 percent based on other
test results In related equipment.  If the destruction efficiency is confir-
med at greater than 99,9%, confidence will then exist that the unit can
destroy the PCB waste,  consistent with the intent of 40 CFR Part 761.  This
will then serve to support the operational disposal of all the contaminated
stocks by the owner,
2.0  DESCRIPTION OF PROPOSED TEST FACILITY
2.1       EXISTING FACILITY
     It is proposed that the waste grease containing PCB will be incinerated
in the process steam boiler at Seattle Rendering Works, located at 5795 S.
130th Place, Seattle, WA.  The boiler at Seattle Rendering Is a dual-fuel
(gas and/or 16 oil) water-tube Cleaver-Brooks model (Delta D-60) with a design
capacity of 35,000 Ibs  steam/hr.  The oil burner is a low-pressure, air
atomizing type.  The rated gas-fuel efficiency 1s 782, so the design heat
input is 46,000,000 Btu/hr.  The furnace volume is 755 cu. ft. in the imme-
diate combustion zone,  excluding the upper radiant section.  Typical stack
temperatures at maximum firing are 550-560 F.  The combustion zone tempera-
tures are in excess of  2200 F at the back wall.
     The boiler has the standard set of safety controls, including a programmed
burner shut-down in the event of flame failure; low water conditions; low fuel
oil temperature; low fuel oil pressure; low atomizing air pressure; or excess
steam pressure.
     Facilities for storage of yellow grease at Seattle Rendering include two
270,000 Ib capacity heated storage tanks.  The grease is pumpable at about
100 F, and congeals at  lower temperatures.  Other storage and transport equip-
ment includes two  50,000 Ib Insulated trailers, and one 20,000 Ib Insulated
trailer.
                                    37

-------
2.2  REQUIRED  MODIFICATIONS TO THE EXISTING FflHTI TTTF<;
     The normal oil-feed supply to the boiler will  be blocked  off,  and  a
temporary waste fuel line Installed.   The temporary Une will  run from  a
grease storage trailer to the fuel feed pump  of the boiler with  a circulating
return line,  the trailer and transfer lines  will  be dedicated during the
verification burn.
     Additional measurement points that will  be Installed for  adequate  boiler
monitoring during PCB Incineration Include:
     (1)  Flowrate Indicator and recorder downstream of oil regulating
          valve.
     (2)  A1r wlndbox pressure Indicator.
     (3)  Continuous carbon monoxide  and excess oxygen monitoring 1n the
          flue gas.
2.3  OPERATIONAL REQUIREMENTS
     The Seattle Rendering boiler normally operates at full modulation  from
a low fire to high fire position, regulated by steam pressure.  Since the
PCB destruction verification bum will be required to operate  at a  steady
rate of waste fuel consumption, process steam generation will  be held
constant during the test period.
     The  following  table compares the proposed verification test conditions
with boiler and operating condlsltons presented In 40 CFR  Part 761.
         EPA Requirement                              Proposed
(Federal Register 40 CFR Part 761)                Verification Test
1.   Boiler must be rated at a minimum      1.  46 MM BTU/hour rating
     of 50 MM BTU per hour.
2.   011  'flow containing PCB's  must be less 2.  100 % contaminated fuel.
     than 10* of the total fuel flow.
3.   011 containing PCB must not be added   3.  Same (gas-firing will be used)
     during startup and shutdown.
4.   Before burning PCB contaminated        4.  Same
     oil, the owner/operator must conduct
     tests and determine that the
     combustion emissions contain at least
     3% excess oxygen and no more than
     50 parts per million of carbon
     monoxide.
                                    38

-------
5.   The company must notify the
     appropriate EPA Regional Adminis-
     trator at least 30 days prior to the
     test and has supplied the combustion
     emissions data required 1n No. 4.

6.   The combustion process 1s monitored
     continuously or, for boilers burning
     less than 30,000 gallons of mineral
     oil annually, at least once each
     hour that PCB contaminated oil 1s
     being burned.

7.   The primary fuel and mineral oil
     feed rates are monitored at least
     every 15 minutes whenever FCB
     contaminated mineral oil Is being
     burned.

8.   The carbon monoxide and excess
     oxygen levels are checked at least
     every 15 minutes whenever burning
     PCB contaminated mineral oil.
5.  Emissions will  be determined
    for proper combustion prior
    to verification burn.
6.  The combustion process will
    be monitored continuously
    for all boilers.
7.  The fuel feed rates will be
    monitored continuously for
    all boilers.
8.  Proposed bum will require
    CO, C02-  Excess 02 and
    opacity levels to be
    monitored continuously.
                                    39

-------
3.0  TECHNICAL SUITABILITY OF CONDUCTING THE VERIFICATION BURN
3.1  FUEL CHARACTERISTICS
    The laboratory Inspection of grease samples Indicated below  Indicate
that the contaminated waste can be burned as a boiler fuel substitute  for
No. 6 fuel oil, with good combustion characteristics.
    1)  The viscosity Is between that of 15 and #6 fuel  oil,  so  atom-
        Ization should be good with a burner that 1s usually  run with
        *6 oil.
    2)  The fuel value of the grease 1s 93% of a 16 oil  value.
    3)  The ash content 1s low.
3.2  SUITABILITY OF THE BOILER
    The waste fuel will be fired at full rate (6 gal Ions/minute) and half/rate
(3 gallons/minute).  The residence time of PCB and temperature  in  the  furnace
combustion zone are as follows:
    Fuel Rate(gallons/minute)  Temperature (*F)  Residence Time  (sec)
                3                2200  (with             2.2
                               supplementary gas
                                   firing)
                6                    2200                1.1

    The residence time of PCB's and temperature for 3 galIons/minute firing
rate are similar to those specified for conmercial waste Incinerators  as
called out in 40 CFR, Part 761, Subpart E (2191°F at 2 seconds).
3.3  SUITABILITY OF THE TEST PLAN
    Normally, three separate tests are conducted at the same  conditions for
verification of the efficiency of PCB waste Incineration.  The  verification
bum performance at multiple grease feedrates 1s needed to decide  on recom-
mended operating conditions.  Therefore, the verification test  plan will  be:
    (1)  Duplicate tests at 3 GPM feedrate.
    (2)  Duplicate tests at 6 GPM feedrate.
    The boiler v/111 be started up on natural gas fuel to reach  steady opera-
tions, as determined by stack measurements of temperature, CO,  and Og.  The
grease circulation loop will then be started.  When the feed temperature
reaches 140-150 F, the grease fuel will be directed to the burner and the

                                    40

-------
 gas fuel gradually reduced, such that stack temperatures  and  steam pressure
 remains constant.
     The emissions tests during the verification burn will  be  carried out
 according to (an EPA-approved) test manual.  The test procedure will use an
 integrated sampling train with a solid adsorbent that 1s  effective 1n  cap-
 turing PCB's 1n low concentrations.  The sampling will be conducted 1so-
 Ic1net1cally, with the sample probe positioned at a number of  prespedfied
 points during the traverse of the boiler stack.  The sampling effort 1s
 expected to take 3 hours to get sufficient PCB capture for analysis detection.
 The adsorbent material and 1mp1nger rinses will be sealed and taken to a
 laboratory that 1s experienced in PCB analysis.  Both gas chromatography  (GC)
 and gas chromatograph mass spectrometry (GC/MS) analysis  methods  will  be
 used.
 4.0  ENVIRONMENTAL SUITABILITY OF  CONDUCTING THE VERIFICATION BURN
      The A1r Quality  analysis results were obtained for worst-case
 conditions as well as expected performance.  The worst case analysis is
 based on  an  absolute  failure  mode  (0% destruction of PCB contaminants, with
 waste concentrations  higher than reported  analyses  and maximum waste fuel
 feed rate).  The expected  performance mode Is  based on 99.9% PCB destruction.
 and average  PCB concentrations 1n  the waste with mid-range feed rate of waste
 fuel.
     Dispersion model  estimates, using local terrain  and meteorological data,
predict a  maximum downwind  concentration of PCB of 16,000 nanograms per cubic
ineter (ng/m ) for worst case conditions, and 4 nanaograms per  cubic meter for
expected  performance.  This can be  compared to the OSHA standard limit value
for PCB's.  The OSHA allowable for  worker exposure on an 8-hour day, 40-hour
                            3
^eek basis 1s 1 million ng/m , so that worst-case conditions would produce
% concentration about  60 times lower  than Industrial  health standards.
/^solute  failure would only be caused by combustion  failure, which would be
detected  and  correc*ed if 1t happened during a verification burn.
     The  vicinity of the Seattle Rendering  plant  Is  mixed  Industrial-commer-
cial-residential.  The nearest residential  areas  are Immediately west  of the
                                     41

-------
plant (about 100 meters), and about 400 meters north, on a bluff approxi-
mately 50 meters high.  The maximum downwind concentrations are predicted to
occur 1n the prevailing wind direction, at the bluff 400 meters north from
the plant.  Ambient monitoring of PCS concentrations at these nearby resi-
dential areas will be carried out during the verification burn.
      No significant water or  land  contamination 1s  expected  from the
verification burn.
FUEL SAMPLING
    Fuel samples will also be obtained and analyzed for PCBs, and the following
parameters: N, S, Cl, C, H, ash, water, sediment, and caloric value.  The
fuel samples will consist of:
    •   A trickle sample obtained during each run
    t   A sample of No. 6 oil normally used 1n the boiler
                                    42

-------