November 1985
   PRACTICAL GUIDE - TRIAL BURNS FOR HAZARDOUS
                WASTE INCINERATORS
P. Gorman, R. Hathaway, 0. Wallace, and A. Trenholm
            Midwest Research Institute
               425 Volker Boulevard
           Kansas.City, Missouri  64110
           EPA Contract No. 68-03-3149
                 Project Officer:
               Donald A. Oberaclcer
            Thermal  Destruction Branch
        Alternative Technologies Division
 Hazardous Waste Engineering Research Laboratory
              Cincinnati, Ohio 45268
 Hazardous Waste Engineering Research Laboratory
        Office of Research and Development
       U.S. Environmental  Protection Agency
              Cincinnati,  Ohio 45268

-------
                                  NOTICE
     This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication.   Mention of trade
names or commercial products does not constitute endorsement or recommendation
for use.
                                       ii

-------
                                    FOREWORD


     Today's rapidly developing and changing technologies and Industrial products
and practices frequently carry with them the Increased generation of solid and
hazardous wastes.  These materials, if improperly dealt with, can threaten both
public health and the environment.  Abandoned waste sites and accidental releases
of toxic and hazardous substances to the environment, also have important environ-
mental and public health implications.  The Hazardous Waste Engineering Research
Laboratory assists in providing an authoritative and defensible engineering basis
for assessing and solving these problems.  Its products support the policies,
programs, and regulations of the Environmental Protection Agency, the permitting
and other responsibilities of State and local governments and the need of both
large and small businesses in handling their wastes responsibly and economically.

     The manual concentrates on those aspects of a trial  burn that are the
most important and those that are potentially troublesome.  The manual contains
practical explanations based on experience of Midwest Research Institute (MRI)
and others 1n conducting trial burns and related tests for EPA.  It includes the
comments of several industrial plant owners and operators.  It is directed mainly
to incinerator operators, those who may conduct the actual sampling and analysis,
and those who must Interpret trial burn results.  It will also be useful for
regulatory personnel and others that need to understand trial burns.
                                             David G. Stephan, Director
                                 Hazardous Waste Engineering Research Laboratory
                                      iii

-------
                                    ABSTRACT


     The manual concentrates on those aspects of a trial burn that are the
most Important and those that are potentially troublesome.  The manual contains
practical explanations based on experience of Midwest Research Institute (MRI)
and others 1n conducting trial burns and related tests for EPA.  It Includes
the comments of several Industrial plant owners and operators.  It 1s directed
mainly to Incinerator operators, those Mho may conduct the actual sampling and
analysis, and those Mho must Interpret trial burn results.  It Mill also be
useful for regulatory personnel and others that need to understand trial burns.
Potential trouble spots that have been encountered are:  (1) trial burns
frequently take more time and effort than an operator anticipates;  and
(2) failure to meet the trial burn requirements.
                                        1v

-------
                                 CONTENTS
Foreword	
Abstract	    iv
List of Figures  . . .	viii
List of T*blts	    is

Sections
  I.      Introduction 	     1
  II.     Overview of * Trial Burn .	     3
               A.  What Docs a Trial Burn Involve	     3
                      1.' Regulatory Limits	     3
                      2.  Permit Conditions	     4
                      3.  Sampling and Analysis Activities 	     4
                      4.  Trial Burn Time Requirement! 	     5
                      5.  Assessiai Potential Performance Problems  .     5
               B.  What Types of Sampling and Analysis are
                     Typically Required.	     S
                      1.  Selecting the S&A Matrix	     7
                      2.  Identifying S&A Methods	     7
                      3.  Adverse Stack Sampling Conditions. ....    10
                      4.  Sample Train Sealing Problems	    10
                      5.  Need for Specialized Methods	    13
               C.  What Skills, Equipment, and Facilities are
                     Needed to Conduct a Trial Burn. .	    13
                      1.  Facilities and Equipment	    13
                      2.  Staff ing Needs	    13
                      3.  Selecting a Contractor	    13
               D.  What are the Major Cost Factors Associated
                     with a Trial Bun	    15
                      1.  Planning and Preparation 	    15
                      2.  Sampling and Analysis.	    15
                      3. 'Quality Assurance	    15
                      4.  Estimating the Costs	    16
  III.    Planning for a Trial Burn	    17
               A.  What Equipment/Instrumentation is the Incinera-
                     tor Required to Have	    17
               B.  How Should the Operating Conditions be Selected  .    19
                      1.  Operating Parameters that Affect Permit
                            Conditions	    19
                      2.  Use of Pretest or Hinibums. .......    20
               C.  How Should Trial Burn POHCs be Selected?. . . .  ;    20

-------
                         CONTENTS (continued)
             D.  What Types and Quantities of Waste are Needed
                   and How Can They be Prepared	    22
                    1.  Quantities of Waste	    22
                    2.  Types of Waste	    22
                    3.  Waste Preparation	    22
                    4.  Mixing . . .	    23
                    5.  Time Requirements	    23
             E.  How Many Runs are Necessary and How Long is
                   Each Run	:	    23
             F.  How Many People will be Needed, With What
                   Experience	    24
             G.  How are POHC Stack Sampling Methods Selected. .  .    24
             H.  What Detection Limits are Required for the
                   Sampling and Analysis Methods	    24
                    1.  Waste Feed Detection Limits	    24
                    2.  Stack Gas Detection Limits	..    24
                    3.  High Concentration of Volatile POECs  ...    28
             I.  What QA/QC Needs to be Done	•   28
             J.  How is it Best to Plan for the Possibility that
                   Trial Burn Results are Outside RCRA Require-
                   ments	    28
IV.     Conducting the Trial Burn. . .	    31
             A.  What -is Involved in Preparing for the Tests  ...    31
              •      1.  Schedule	•  . .  .    31
                    2.  Sampling Crew	    32
                    3.  Equipment. . . .	    32
                    4.  Facility Readiness 	    36
                    5.  Process Data	    36
                    6.  Data Sheets and Labels	    37
                    7.  Safety Precautions 	    39
                    8.  Observers	    41
             B.  What is Involved in Conducting the Actual Sam-
                   pling, and What are the Problems that May
                   Occur	    41
                    1.  Equipment Setup.	    42
                    2.  Preliminary Testing	    43
                 '   3.  Actual Testing	    43
             C.  What is Involved in Analysis of Samples and
                   What are the Problems that May Occur	    45
                    1.  Sample Check-in	    46
                    2.  Analysis Directive 	    46
                    3.  Sample Inhomogeneity	    47
                    4.  Analytical Interferences 	    47
                    5.  Saturation of GC/MS Data System	    47
                    6.  High Blanks	    47
                    7.  Poor Precision	    47
                    8.  Recovery Efficiency	i ......    48
                    9.  Actual Versus Expected Results 	    48

-------
                         CONTENTS (continued)
             D.  How are the Sampling Data and Analysis Data
                   Converted to Final Results	    48
                    1.  Blank Correction	    43
                    2.  Significant Figures and DRE	    51
                    3.  Rounding Off DRE Results	    51
                    4.  Reporting DRE with a "<" or ">" Sign ...    51
             E.  How are the Data and Results Usually Reported .  .    52
V.      References	    63

-------
                                  FIGURES






No.                                                                '   Page



 1   Modified Method 5 sampling train (MM5), 	    11



 2   Volatile organic sampling train (VOST)	    12.



 3   Example of computer labels	    40
                                    Vlll

-------
                                  TABLES

No.                                                                   Page

 1   Tine Factors Involved in a Trial Burn	      6

 2   Sampling Methods and Analysis Parameters 	 .  .      8

 3   Example Analytical Procedures	      9

 4   Capabilities Necessary for Trial Burn Sampling and Analysis.  .     14

 5   Incinerator Equipment/Instrument Requirements for Trial
       Burn	     18

 6   A Typical Example of Sampling Personnel Required . .  	     25

 7   Procedure for Identifying Necessary Stack Sampling Methods .  .     26

 8   Example Calculation for Estimating POHC Concentration in
       Stack Gas, at DRE of 99.99%	'	     27

 9   Example QA for a Trial Burn	     29

10   Example List of Sampling Equipment and Supplies Typically
       Used	     33

11   Example List of Data Forms	     38

12   Potential Problems that May Occur During Tests 	     44

13   Data Necessary for Calculating DRE	     49

14   Incinerator Operating Conditions	  .     S3

15   Concentrations of POHCs in Waste Feeds	  .     54

16   Calculated Input Rates for POHCs in Waste Feeds	     55

17   Concentrations of Volatile POHCs by VOST in Stack Effluent .  .    .56

18   VOST Blank Correction Values 	     57

19   VOST Sample Volumes	     57

-------
                            TABLES (continued)



No.                                                                   Page



20   Blank Correction Values for Semivolatile FOHCs 	     57



21   Destruction and Removal Efficiencies 	 .......     58



22   Modified Method 5 Test Data	     59



23   Continuous Monitoring Data	     60



24   General Analysis of Aqueous Waste	     61



25   General Analysis of Organic Waste	     61



26   Example DRE Calculations	     62

-------
                                 SECTION I

                               INTRODUCTION
     On Hay 19, 1980,  tbe  U.S.  Environmental Protection Agency (EPA) pub-
lished regulations under the authority of the Resource Conservation and Re-
covery Act  (RGRA)  for  hazardous waste  incinerators.  These  regulations  re-
quire that new and existing incinerators adequately destroy hazardous organic
compounds and maintain acceptable  levels of  particulate  and chloride  emis-
sions.  Owners and operators  of incinerators are  required  to  demonstrate
the performance of the facility by means of a trial burn.  As a consequence,
industry and control agency personnel have become involved in planning for,
conducting, and interpreting the results from trial burns.   This  manual  is
written to assist those individuals in their efforts.

     The manual concentrates on those  aspects of  a trial burn  that are  the
most  important  and those  that  are potentially troublesome.  The  manual
contains practical explanations based on experience  of  Midwest Research
Institute (MRI) and others in conducting trial burns and related  tests  for
EPA.  It includes the comments of several industrial plant owners and oper-
ators'.  It  is directed mainly to incinerator operators,  those  who may con-
duct  the actual  sampling and analysis, and those who must interpret trial
burn  results.  It  will also be useful for regulatory personnel and others
that need to understand trial burns.

     One of the major  objectives was  to make  this Guide readily usable.
For that reason, the discussion  is brief and avoids dwelling on detail.   A
question and answer format is used to  relate the  material to operator con-
cerns.  Each subsection begins as a question that could well be posed by  an
incinerator operator  who  needs to  conduct  a trial burn.   The  narrative
following each question provides answers to the question or provides infor-
mation pertinent to  the question.   For each question, the  most important
considerations are discussed,  and  potential trouble spots are  identified.

     This Guide addresses  multiple components of the trial burn process  in-
cluding planning and preparation, sampling and analysis  for  the trial burn,
process monitoring during  the trial burn, and data reduction and  reporting.
The Guide does not directly address the preparation of the Trial Burn Plan,
but it does address some planning aspects that affect Trial  Burn Plan prep-
aration and subsequent interpretation  of the trial burn  results.

     The remainder of  the  Guide is divided into three sections.   Section  II
presents an overview of the trial  burn process  and requirements.   Section
III discusses planning-for the trial burn.   Section IV discusses  conducting
the trial burn and reducing and reporting data from the  trial burn.

-------
                                SECTION II

                         OVERVIEW OF A TRIAL BURN
     This section  summarizes  different aspects of the trial bum.  It de-
scribes the  trial  burn process and  requirements for the trial burn.   Basic
information is provided to help answer four questions:

     A.   What does a trial burn involve?
     B.   What types  of sampling  and analyses are  typically involved?
     C.   What skills, equipment  and facilities are needed to conduct the
          trial burn?
     B.   What are  the major  cost factors associated with  a  trial burn?

A.  WHAT DOES A TRIAL BURN INVOLVE?

     When an incinerator operator is faced with the need to perform a trial
burn, -the first questions that come to mind are:  "What do I do for a trial
burn?".and "What does the trial burn do  to me?."  From the  operators' per-
spective, the  key  trial burn  considerations  are the regulatory  limits that  .
must be achieved,  the permit  conditions  that result from  the burn,  and  the
extent of  sampling and analysis  activities  required.   Potential trouble
spots that have been encountered are:  (1) trial burns frequently take more
time and effort  than an operator anticipates; and (2) failure to meet the
trial burn requirements.  Each of these  considerations is discussed below.

1.  Regulatory Limits

     The trial burn provides  regulatory  agencies  with data  that will allow
them to issue an operating permit.  Consequently, the trial burn is directed
to testing the plant to show that  it achieves the RCRA limits,  under the
desired plant operating conditions.  Those RCRA limits are:

          Destruction and removal efficiency  (DRE) > 99.99% for all subject
          principal organic hazardous  constituents (POHCs).
          Farticulate emission <  180 mg/dson  (corrected to 7% 02).
          Hydrogen chloride (HC1) emissions <  4 Ib/hr, or > 99% removal ef-
          ficiency.

     In addition to the  above standards, state permit  officials may add
their own  individual trial  burn  and permit  conditions to  the  federal stan-
dards .                         .                                         .

-------
2.  Permit Conditions

     From the operator's standpoint, operating conditions imposed by a per*
mit need to allow the plant to incinerate the types and quantities of waste
they expect to handle, at the necessary feedrates, and within an acceptable
range of operating conditions.  That is, the permit conditions need to pro-
vide the plant with the desired flexibility, within limits that are reason-
ably achievable.  Based on the trial burn results, the operating permit may
specify certain criteria such as:

          No wastes  nay be incinerated which  contain any Appendix VIII
          compound having a  higher heating value  (HHV) below  that  of  the
      ,    most difficult to incinerate POHC used in the trial burn.
          Maximum cdncentration of certain POHCs in waste feed.
     •    Maximum waste  feedrate, and/or maximum  total heat input rate.
          ^aviipnm air feedrate, or maximum flue gas velocity.
          Minimum combustion temperature.
          Maximum carbon (CO) monoxide content of stack gas.
          Maximum chloride (Cl) and ash content of waste feed.

Additional criteria are discussed in Reference 1.

     The trial  burn  involves testing at conditions  that meet  the plant's
operating needs while  meeting .the three RCRA limits.  It may be necessary
to test at more than one operating condition in  order to satisfy all those
needs.- .For example,  it might be difficult  to  achieve a high heat input
rate (i.e., design heat input rate) with a waste feed that contains desired
high levels of  Cl and ash.  These  factors  are discussed more fully as a
part of planning activities in Sections III-B and III-C.

3.  Sampling and Analysis Activities

     Each test  run  in the trial burn includes sampling of the waste feeds
and the stack effluent.  These samples are then split into a series of sub-
samples to be  analyzed for POHCs, Cl", HHV, ash, etc.  The sub samples are
then analyzed  for the  subject POHCs  by  rather  complex methods  that  include
analyses by gas chromatography/mass spectrometry (GC/MS).  Analysis results,
along with waste  feedrates  and stack gas  flow rates measured  during  each
run, are used  to  calculate the DREs.  Usually,  samples of  ash  and  scrubber
waters are also taken and analyzed for the subject POHCs.  Although not re-
quired by RCRA, regulatory agencies may impose  other additional  sampling
and/or analysis requirements.  More detail  on sampling and analysis  pro-
cedures is included in Section II-B.

     For any trial burn,  at  any  one set of operating conditions  (and  waste
feed characteristics),  EPA documents recommend three replicate runs.   How-
ever, it may be acceptable to make three or  more runs with each run done  at
different conditions or with different waste feed  characteristics.  In this
a  Federal Register. Wednesday, May 20,  1981, Vol. 46, No.  97,  40  CFR 261,
   Appendix VIII, p. 27477.

-------
regard, there appear to be differences in what  is  acceptable  from  case-to-
case, so plans oust be approved by the responsible regulatory agencies before
the trial burn.  The Trial Burn Plan  should  specify  the  number  of  runs  and
the test conditions for each run.

     An important  thing  to  remember in planning for three or more runs is
that the quantities of waste required are substantial.  Each run may require
4 to 8 hr of  plant operating time.   It  is probably  best to also burn the
same, or very similar,  wastes during nontest periods  (i.e.,  at night)  in
order to'maintain  reasonably steady conditions over the test period.  The
total trial burn period can require a rather large quantity of the specified
waste(s).   Those quantities  are  also specified as part  of the  Trial Burn
Plan.

4.  Trial Burn Time Requirements

     A major  factor  in  performing a  trial burn is time.  Many steps are
involved in the  trial  burn sequence of events listed in Table 1.  Some of
the  steps  have time limits  dictated under  RCRA.  For others,  adequate
time must be  allowed.   For example,  the many samples  obtained  in  a  trial
burn,  and  the complexity  of POHC analysis, make  it desirable  to allow
1-1/2 months  to  complete  the analyses and another half month to prepare a
detailed report  of all  results.   General guidelines for time requirements
are included  in Table 1.  In specific instances greater amounts of time may
be required.   If  an operator is unfamiliar with trial burns, consultation
with other operators, consultants  or  agency  personnel  early in  the process
can provide more  exact  estimates  of time requirements for specific situa-
tions.

     In addition to the time required to adequately prepare and conduct the
trial burn, time is also required for preparation of the Part B Permit Appli-
cation.  Frequently the applicant will be working on trial burn preparations
and  responding  to  letters and comments on the RCRA permit simultaneously.

5.  Assessing Potential Performance Problems

     Probably the  most  important  question  faced by the operator is "Will  I
pass?"  (i.e., meet the  RCRA  requirements).   The trial burn can  be  designed
to include several different operating  conditions  including  some where  po-
tential incinerator  performance problems are minimized.  Another  alterna-
tive  selected by  some  plants is  to  conduct an unofficial preliminary
"minibura" (i.e.,  one  run) prior to  the actual trial burn.  This  miniburn
provides some indication of the results that can be expected, but it oust
be done at least 2 months before  the  scheduled  trial burn in order to  com-   .
plete  all analyses,  evaluate the results, and  make  whatever changes are
required.

B.  WHAT TYPES OF  SAMPLING AND ANALYSIS ARE  TYPICALLY REQUIRED?

     The primary objectives  of the  sampling  and analysis  (S&A) program  are:
(a)  to quantify POHC input  and output  rates to determine whether DRE  re-
.quirements are met;  (b) to measure input and output  rates of chloride;  and

-------
    TABLE 1.  TIME FACTORS INVOLVED IN A TRAIL BURN
Notification to submit Part B application.

Evaluate all conditions at which plant desires to be
permitted (1 month).

Prepare trial burn plan and submit to EPA (required 6 months
after notification). .

Prepare responses to EPA on any questions or deficiencies
in the trial burn plan (1 month).

Hake any additions or modifications to plant that may be neces-
sary (1 to 3 months).

Prepare for trial burn.

     *    Prepare for all S&A, or select S&A contractor (2 to
          3 months).

     *    Select date for trial burn, in concert with S&A staff
          or contractor (completed 1 month prior to test).

     *    Notify all appropriate regulatory agencies (1 month).

     *    Obtain required quantities of waste having specified .
          characteristics.

     *    Calibrate all critical incinerator instrumentation
          (2 weeks).

Conduct trial burn sampling (1 week).

Sample analysis (1 to 1-1/2 months).

Calculate trial burn results (1/2 month).

Prepare results and requested permit operating conditions for
submittal to EPA (1/2 to 1 month).

Obtain operating permit.

-------
(c) to determine  stack effluent particulate concentrations.  The two most
important considerations are selecting the S&A "matrix" (i.e., selecting the
streams to be  sampled and analytes to be measured) and identifying appro-
priate S&A methods.  Specific problems which can be encountered are adverse
stack conditions,  sample  train  sealing problems,  and  the  need for special-
ized S&A methods.  Each of  these  factors  is  discussed briefly below and in
more detail in Sections III-F and G and IV-A and B.

1.  Selecting the S&A Matrix

     The main  focus  of the sampling activities is collection of the waste
feed and the  stack effluent samples, the  latter  being the most complex.
Usually, the ash  and scrubber waters are  also  sampled and analyzed.   The
main focus of  the  analysis  activities is  on  the POHCs.  The stack S&A also
includes determination  of HCL and particulate emissions,  but these  methods
are relatively simple compared to those for POHCs.  A discussion of sampling
and analysis needs can also be found in References 1, 2, and 3.

     Overall, the  S&A typically required consists of  the following,  as a
m-in-jptmn     .                  •

          Obtain representative samples of each waste feed  stream to the
          incinerator.  Analyze those samples for the selected POHCs. and
          for HHV. Cl, and ash.   (Remember'that  the  input  rate  of each
          waste feed  must also  be determined in order to  compute the POHC
          input rate which is used in the calculation of DRE.)

          To achieve a "representative" waste feed sample,  liquid waste feeds
          are often  sampled once  every 15 min and composited  in each run.
          Solid waste  feeds must  also be sampled using the best practical
          method of  obtaining representative samples  of each type of solid
          waste used in the trial burn.

          Sample stack emissions  to determine stack gas flowrate, HC1,par-
          ticulate concentration, and to determine concentration of POHCs.

2.  Identifying S&A Methods  •

     An example of S&A methods  that  could be specified for a trial burn is
shown in Tables 2  and 3.   These  tables  identify  the  main references that
are available on recommended S&A  methods, particularly Refs.  2 and  3.   These
documents  contain valuable information but  do  take  considerable time  to
understand.  They are best'utilized by personnel  experienced  in S&A methods.
These references  are the  best  sources to identify the methods that can be
used in  a  Trial Burn Plan.  However, experience  helps a  great deal in  se-
lecting the most appropriate of the available recommended  methods.

     Determination of stack  gas flow rate and particulate  emissions  is  done
according to the conventional stack sampling method commonly referred  to as
Method 5 (MS).  This method encompasses EPA Methods  1-5  and is defined in
detail in 40 CFR Part 60, Appendix A.  HC1 emissions  are  sampled by modify-
ing the Method 5 train  to include a caustic  impinger.

-------
                 TABLE 2.  SAMPLING METHODS AMD ANALYSIS PARAMETERS
                               Sampling
                              frequency           Sampling                       .
       Sample                for each run          method      Analysis parameter


1.  Liquid waste feed  Grab sample every 15 min     S004      V&SV-POHCs, Cl*, ash,
                                                              ult. anal., viscosity,
                                                              HHV

2.  Solid waste feed   Grab -sample of each       S006, S007   V&SV-POHCs, Cl", ash,
                       drum                                   HHV

3.  Chamber ash        Grab 1 sample after all      S006      V&SV-POHCs,
                       3 runs are completed                   EP toxicity

4.  Stack gas          Composite              '   MM5 (3 hr)   SV-POHCs, particulate,
                                                              H20, HC1

                       Three pair of traps,      VOST (2 hr)  V-POHCs
                       40 min each pair

                       Composite in Tedlar          SOU      V-POHCsC
                       gas bag

                       Composite in mylar        M3 (1-2 hr)  C02 and 02 by Orsat
                       gas bag

                       Continuous (3 hr)         Continuous   CO (by plant's
                                                 monitor      monitor)
   VOST denotes volatile organic sampling train
   MM5 denotes EPA Modified Method 5
   M3 denotes EPA Method 3
   SXXX denotes sampling methods found in "Sampling and Analysis Methods for
     Hazardous Waste Combustion."3

   V-POHCs denotes volatile principal organic hazardous constituents (POHCs).
   SV-POHCs denotes semivolatile POHCs.
   HHV denotes higher heating value.

C  Gas bag samples nay be analyzed for V-POHCs, only if VOST samples are saturated
   and not quantifiable.

-------
                  TABLE 3.  EXAMPLE ANALYTICAL PROCEDURES


1.





2.

,


3.


4.











Sample
Liquid waste feed





Solid waste feed




Ash


Stack gas
a. MM5 train
Filter and
probe rinse
Condensate

XAD resin
Caustic impinger
b. VOST
c. Tedlar gas bag
d. Gas bag
e. Cont. monitor
Analysis
parameter
V-POHCs
SV-POHCs
Cl"
Ash
HHV
Viscosity
V-POHCs
SV-POHCs
Cl"
Ash
HHV
V-POHCs
SV-POHCs
Toxicity


Participate
SV-POHCs
Cl"
SV-POHCs
SV-POHCs
Cl"
V-POHCs
V-POHCs
C02, 02
CO
Sample
preparation
method
8240
8270
-
-
-
•
8240
8270
-
-
-
8240
P024b, P031
-


M5
P024b, P031
-
P021a
P02U
-
•
-
-
—
Sample
analysis
method
8240
8270
E442-74
D482
D240
A005 •
8240
8270
D-2361-66 (1978)
D-3174-73 (1979)
D-2015-77 (1978)
A101
A121
C004


M5
A121
325 . 2
A121
A121
325.2
A101
A101a
M3 (Orsat)
Continuous monitor

Note:
    Four-digit numbers denote methods found in "Test Methods for Evalu-
    ating Solid Waste," SW-846.2
    Numbers with prefixes of A, C, and P denote methods found in "Sam-
    pling and Analysis Methods for Hazardous Waste Combustion."3
    Method No. 325.2 (for Cl') is from "Methods for Chemical Analysis
    of Water and Wastes," EPA-600/4-79-020, March 1979.4
    Numbers with prefixes D and E denote methods established by the
    American Society for Testing and Materials Standards (ASTM).
    M3, M5 refer to EPA testing methods found in the Federal Register,
    Vol. 42, No. 160, Thursday, August 18., 1977.s

Tedlar gas bag samples will be analyzed for V-POHCs, only if VOST sam-
ples are saturated and not quantifiable.

-------
     Sampling of  stack  effluent for POHCs, in order to determine DRE, may
require  from  one  to three separate methods  (or more), depending on  the
number of  POHCs to  be quantified  and  their characteristics  (e.g., volatile
or semivolatile), and on  the  detection  limits  that  are  required to  prove a
DRE of 99.99%.  These methods are:

     1.   Modified Method 5 (MM5) - for semivolatile POHCs.

     2.   Volatile  Organic  Sampling Train  (VOST)  • for volatile POHCs.

     3.   Gas bags * for volatile POHCs.

     4.   Special methods - for certain POHCs  which cannot  be  sampled  with
any of the above methods.

     Semivolatile POHCs commonly  require use  of MM5.   This one sampling
train, shown  in Figure  1, provides for determination of particulate, HC1,
and the  SV-POHCs.  However,  the probe  rinse  must  be evaporated and  the
filter desiccated to  determine particulate.   Thereafter, these components
can be  extracted, and  combined with  extracts of the XAD  resin and the
condensate, for analysis  by  GC/MS to determine SV-POHCs.  A small ^aliquot
of the condensate must  be removed before  extraction to  quantify Cl   in the
condensate, as well as in the caustic.

     A diagram of the VOST train  commonly used for  sampling volatile POHCs
is shown in Figure  2.   This train,  unlike M5  or MM5,  does not  involve  tra-
versing the stack.  However,  the VOST preparation and analysis procedures
are quite complex.  Those interested in the detailed procedures should refer
to Ref. 6.

3.  Adverse Stack Sampling Conditions

     Adverse  stack  sampling  conditions  are frequently encountered  at haz-
ardous waste  incinerators.   Problems that have been encountered include
cyclonic flow, very high  temperature stacks  (1600°  to  1800°F), and high
moisture content  (saturated  with H20 at 150°F with  droplet  carryover).
These potential problems  should be considered during planning, and appro-
priate actions should be  taken.  More complete discussions  of  cyclonic flow
and moisture are  included in Section IV-B-2.

4.  Sample Train  Sealing Problems

     Both  the VOST  method, and available  guidance  on the MM5  method,  state
that no  grease be used  on any of  the  connections  in the train (i.e.,  ball-
joints).   Teflon  or Viton 0-rings have been  used  in VOST,  and in MM5, to
provide  adequate  seals  without use of grease.   Added care must be taken to
ensure leak-check integrity  of the sampling  trains,  with some added risk
that a  test  may have to  be repeated  if any sampling  train  fails the post-
test leak  check.
                                     10

-------
                                                 Fillet
              Quail
Ilteimocouple
     Reveite-f ype
     Pilot lube
                                                                                          llteimonteter
                                                                    0   ®__(3)__0   (5
                                                                             ICE IAIU
          I Modified Greeribuig-Smllli. jjeyeijed. Imply
          i GieenbiNB-Smllh, 50ml of Double DlUllled In Glati H2O
          I Gieenbuia-Smllli. IOOn.1 of 0.1 N KOII
          I Modified Gieeil>iuo-Sinllli Eniply
          I Modified GieenbufO-SnUlli. S|O2
W) Condemei
® XAD Ketln Caihldge

 * Ice Walef Jacket
                           Figure 1.   Modified Method 5  sampling  train  (MM5)

-------
Glau Wool
Parllculala
Filter
    t
 Slack
 (or Tetl
 System)
            Probe
                  Condensale
                  Trap Impinger
                                                          Vacuum
                                                          Indicator
     Tenax

     Charcoal Backup
                                                               Rolometer
Empty    SI lea Gel
                                  Pump
tr
 Dry Gat
 Meier
                                                       Exhaust
                                                       I l/mln
                            Figure 2.  Volatile organic sampling  train (VOST).

-------
5.  Need for Specialized Methods

     Although the majority  of POHCs are sampled with  either  VOST or MM5,
specialized  sampling  methods must be used  for some POHCs.   Those  POHCs
which require specialized methods  are identified in Appendix B of Ref. 3.
This reference should be consulted during the planning stage  to assure that
proper methods are used.

C.  WHAT SKILLS,  EQUIPMENT, AND FACILITIES ARE NEEDED TO  CONDUCT A TRIAL
      BURN?

     The incinerator  facility operator  is  responsible for  conducting the
trial burn.  The  facility  must provide the types and  quantities  of waste
needed, and operate the plant during the trial burn at the conditions under
which they  desire to  be permitted.  However, the specialized sampling and
analyses required in a trial burn are beyond the capability of most  facili-
ties.  A facility that has  most of  the  necessary  capabilities still may
decide to use  an  experienced contractor because of the specialized  nature
of the methods  and the fact  that  the trial burn may  be only a one time
need.  The  important  considerations  in  the decision  are  facilities and
equipment requirements - and  staffing needs.   Each of these factors,  plus a
brief consideration of contractor selection, are addressed below.

1.  Facilities and Equipment

     Whether the  operator  uses a  contractor  or their own staff  for the
trial burn  S&A,  certain capabilities are  required.   The  facilities and
equipment that are usually  necessary are shown in Table 4.

2.  Staffing Needs

     Decisions regarding use  of in-house or contractor expertise  to  conduct
a trial burn also depend on  staffing needs.  At a minimum, the trial burn
staff should be knowledgeable in stack sampling methods, have experience in
analysis of low  concentrations of organics in complex matrices,  and be
familiar with  calculating  and reporting trial burn results.  Knowledge of
process monitoring  is also  helpful.  A  detailed discussion of the number
and  capability  of  S&A personnel  required  is  included in  Section III-E.

3.  Selecting a Contractor

     Trial  burn procedures  are relatively  new, and  are much more  sophisti-
cated than  a normal EPA Method 5 test for particulate  emissions.  There are
about  10 to 20  organizations in the United States who have trial burn S&A
experience,  and probably several more who  are capable of doing so.  If a
facility would  like to know  who to contact for S&A service, they  should
make  inquiries  at state and  federal  regulatory agencies or contact other
incinerator facilities who  may have  already conducted  trial burns.
                                     13

-------
  TABLE 4.  CAPABILITIES NECESSARY FOR TRIAL BURN SAMPLING AND ANALYSIS



1.   Sampling equipment for solid waste feeds (especially drummed wastes).

2.   Stack sampling equipment, usually including the following:

          EPA Method 5 equipment, and all associated test equipment (e.g.,
          EPA Methods 1, 2, and 3).

          Method 5 equipment adaptable to Modified Method 5 (greaseless)
          and associated XAD resin preparation,  extraction, and analysis
          facilities.

          Volatile Organic Sampling Train (VOST) Equipment with at least
          18 pairs of VOST traps.  Also, all facilities needed for prepar-
          ing, checking, and analyzing the traps.

          Gas bags and associated sampling equipment (see Figure 7, Ref-
          erence 3).

          Field laboratory equipment for sample  recovery.

3.   Facilities for analyzing all samples, includingr

          Laboratories containing relevant safety equipment such as hoods
          and equipped with sample preparation equipment including Soxhlet
          extractors, separately funnels, continuous extractors, blenders,.
          Sanifiers or other equipment to homogenize waste feed samples,
          sodium-sulfate drying tubes, Kuderna-Danish glassware, etc.

          Equipment for preparing VOST traps to  allow simultaneous heating
          and purging of the traps.  Ideally the traps should be prepared
          and stored in an organic-free laboratory.

          All required compounds to prepare calibration standards and sur-
          rogate recovery spiking solutions.

          Computerized GC/MS instrumentation.

          Established QA procedures for assessing precision and accuracy
          of analytical methods.

4.   Knowledge and preferably experience in all of the sampling and analysis
     methods and calculation/reporting of results.

5.   Process monitoring experience, especially quantification of waste feed-
     rates and documentation of plant operating conditions.
                                     14

-------
D.  WHAT ARE THE MAJOR COST FACTORS ASSOCIATED WITH A TRIAL BURN?

     The three  major  cost components of the  trial  burn are planning and
preparation, sampling and analysis and quality assurance (QA).  Each of the
components must be included in trial burn budgets.

1.  Planning and Preparation

     One of  the first cost factors for a trial burn is preparation of the
Trial Burn  Plan,  including preparing  responses  to questions  and additional
information requested by the regulatory agencies.

     The second cost  factor  is plant  additions  and  modifications needed  to
comply with  RCRA regulations (see Section  III-A).   These  may include CO
monitors, waste feed flow monitors,  etc.,  and stack sampling  ports  and
scaffolding needed for the trial burn. .

     A third cost factor is associated with acquiring/storing the types and
quantities  of  wastes necessary for  the  trial burn  (see Section III-C).

2.  Sampling and Analysis

     The major  cost  factor is the sampling and analysis required by the
trial burn.  In general, this cost usually  ranges from  $30,000  to $150,000,
depending on the number  of runs,  the  number of samples  to  be taken in each
.run, and the analysis required on each sample.

     The number of runs to be conducted in  the  trial burn  is one  of the ma-
jor cost  factors.   EPA recommends three  replicate  runs at each operating
condition to be tested.   Thus, a minimum of three  runs are usually done.
If two  operating conditions  are involved,  then six  runs may be  necessary.
In some  cases,  an array of operating conditions  are tested, with only one
run at each  condition.

     In each run, all influent and effluent streams  are usually sampled  and
analyzed as  described earlier (see Table 2).   The  number  of these  samples
and the complexity of their analysis  obviously  affects  the cost.  Ordinarily,
the field sampling activity, including all  preparation  for sampling,  accounts
for one-fourth,  to one-third  of the S&A cost.   Analysis of samples usually
accounts for one-third to one-half the cost, with  the  remaining costs for
data reduction, calculations, and reporting of  results.

3.  Quality  Assurance

     Analysis costs,  especially for  the POHCs,  are  a rather large cost fac-
tor, partly  because analytical QA activities typically  include:

          Replicate analysis of some  samples.

          Analysis of samples  spiked  with POHCs or  surrogates.

          Analysis of blanks.


                                     15

-------
          Analysis of calibration standards.

          Analysis of blind audit samples.

     All of  the  above can easily involve analysis  of  twice  the  number  of
samples actually  taken.   All these samples then must be multiplied by the
number of analyses to be performed on each sample.  Some QA is an essential
part  of  a trial  burn,  but excessive QA can  rapidly  increase the cost.

4.  Estimating the Costs

     Determining  the  cost for a trial burn is highly  site dependent.   In
general, the trial burn costs will depend on the following factors, as dis-
cussed in the preceding sections:

          Number of runs
          Number and type of waste feed samples
          Number of effluent samples
          Number of different analyses performed on each sample
          Complexity of the QA/QC plan
          Modifications required to prepare facility for trial burn

The normal range  for  a trial burn sampling and analysis program conducted
by an outside contractor is $30,000-$150,000+.  This range does not include
plant Codifications,  preparations  on-site for the test, or preparation of
the permit application and trial burn plan.

     The breakdown of costs for a trial burn is roughly:,  one-third for the
field sampling program;  one-third  for sample analysis  and project QA/QC;
and one-third preparation,  engineering calculations and reporting.  These
are rough  estimates and  frequently  the  analysis  portion of  the  program
can involve as much as half of the total cost.

     In summary,  the  sampling and  analysis  part of  a trial burn  is costly,
and each time another run,  another sample,  or another  analysis is added to
the test plan the cost will rise.  Each such addition needs to be carefully
considered in order to hold costs at the lower end of the range.
                                    16

-------
                                SECTION  III

                         PLANNING FOR A  TRIAL  BURN
     The  probability for success of a trial burn is enhanced by go.od plan-
.ning.   The major objectives of  the  planning process are:   (a)  to  select
trial burn conditions  that provide the plant adequate operating  flexibility;
(b) to  assure that the trial burn will be conducted in a manner acceptable
to regulatory agencies;  and  (c)  to make  the  trial burn  cost  effective.   Key
questions  addressed  during planning  are:

     A.    What equipment/instrumentation is  the incinerator required  to
           have?

     B.    How should operating conditions be selected?

     C.    How should trial burn  POHCs be selected?

     D.    What types and quantities  of waste are needed and  how  can they be
           prepared?

     E.    How many runs  are  necessary and how long  is each run?

     F.    How many people are needed and with what  experience?

     G.    How are POHC sampling  methods  selected?

     H.    What detection limits  are required for the sampling and analysis
           methods?

     I.    What QA/QC needs to be done?

     J.    How is it  best to  plan for the possibility that trial  burn results
           may be outside RCRA requirements?

A.   WHAT  EQUIPMENT/INSTRUMENTATION  IS  THE  INCINERATOR REQUIRED TO HAVE?

     The  incinerator is  required by  RCRA to  have the equipment/instrumenta-
tion shown in Table  5.  The  regulatory  agencies  also may require monitoring
of  other important operating parameters (e.g.,  scrubber water  flow rates,
venturi scrubber AP,  etc.). Minimum or  maximum levels for each parameter
may be  specified in  the  operating permit.  Analysis of  waste feeds  may also
be  required if the  operating permit stipulates  limitations  on HEV, Cl, or
ash content of waste feed.
                                     17

-------
TABLE 5.  INCINERATOR EQUIPMENT/INSTRUMENT REQUIREMENTS FOR TRIAL BURN
        Equipment to maintain particulate emissions below 0.08 grains/
        dscf.

        Equipment to maintain 99% HC1 removal or HC1 emissions below 4
        Ib/hr.

        Equipment that provides 99.99% DRE on POHCs.

        Stack test ports and scaffolding.

        Valves, taps, etc., for sampling all waste feeds, liquid effluents,
        ash, etc.

        Equipment to maintain noncyclonic flow in stack when testing.

        Continuous CO monitor.

        Continuous waste feed flow monitor.

        Continuous monitor for combustion gas velocity or air input
        rate.

        Continuous combustion temperature monitor.

        Automatic interlock system to shut off waste feed under the fol-
        lowing situations.

        a.   Low combustion temperature.

        b.   High CO concentration.

        c.   High combustion airflow to incinerator or high combustion gas
             velocity.'
 Established based on trial burn results or state statutory limitations.
                                   18

-------
B.  HOW SHOU1D THE OPERATING CONDITIONS BE SELECTED?

     Operating conditions  for  the trial burn are  selected to provide the
plant with operating flexibility.  Important considerations are the key  op-
erating parameters that affect permit conditions and the use  of pretests  or
niniburns to help establish those conditions while meeting RCRA requirements
(e.g., DRE).

1.  Operating Parameters that Affect Permit Conditions

     The operating  conditions  selected for the trial  burn must represent
the worst case conditions under which the incinerator may expect to operate,
and therefore needs  to be permitted to  operate.   The  conditions  selected
may include any or all of the following:

     •  .  Waste containing hardest- to-burn POHC (lowest HHV) .

          Highest concentrations of all POHCs selected.

          Maximum waste feedrates.

          Maximum combustion airflow rate (minimum residence  time) .

          Maximum CO level in stack gas.

          Mim'"11"11 'combustion temperature.

          Minimum HHV of waste.

          Maximum thermal input  (Btu/hr) .

          Minimum 02 level in stack gas .

          Maximum Cl content of waste feed.

          Maximum ash content of waste feed.

                   or. maximums on other operating  conditions  (e.g., venturi
          scrubber AP, scrubber water flow rate and pH).

     Obviously, it is very difficult to achieve all of the  above  at  any one
set of  operating  conditions.  In fact, some  of the conditions are almost
direct  opposites  (e.g.,  maximum airflow rate  but minimum 02 in stack gas).

     The  first  six items in the above  list are probably the most important
and may be  achievable in .one set of operating conditions that also include
some of the other conditions.   If so,  one trial burn (three runs) at those
conditions  may  suffice.   If not, additional  runs that include the  other
conditions  may  be necessary.  Of course,  operating conditions which result
in permit conditions most favorable to each  individual  facility  will  have
to be determined  on  a case-by-case  basis.
                                     19

-------
     The major problem with the worst-case conditions is that they maximize
the chance of  failure  (not meeting RCRA  requirements).   Since  the plant
wants to pass,  the  exact conditions must be carefully selected, balancing
operating needs against  increasing  chance of failure.   Plant  operating  ex-
perience is very important in these decisions.

2.  Use of Pretest or Miniburns

     Preliminary testing and miniburns can be extremely valuable in helping
to select operating  conditions  for the.actual trial burn.  The  following
types of miniburns may be useful.

     The hardest  to  burn POHC,  at high concentration,  can be used in  a
miniburn that  is  conducted at the lowest temperature  and the highest CO
level.  If the  results show a DRE exceeding 99.99%, then  it is likely that
99.99% DEE will be  achieved regardless of any other operating conditions.

     At high Cl input rates, a well designed scrubber will not usually fail
99% removal  even  at  minimum conditions.   A pretest could verify that pre-
sumption.

     Achieving  the particulate  limit  causes  problems more frequently than
does achieving  DRE.  A pretest with EPA Method  5' will help identify any
problems and help in selecting conditions for the  trial  burn.   The pre-
test can also  uncover  specific  sampling and analysis problems that may  act
be readily apparent.

     Mist carried over from a recirculating scrubber solution or alkaline
scrubbers can  have a drastic impact on particulate emission measurements ,•
especially if  the scrubbers  are not equipped with efficient mist elimina-
tors. . It may  be  advisable to conduct  a preliminary particulate  test, well
in advance of the actual trial burn, to identify possible problems.

     For existing plants,  any of the above pretesting could be done prior
to submitting the Trial Burn Plan for approval.  For new plants, pretesting
will haveto be part of the approved Trial Burn Plan.

C.  HOW SHOULD TRIAL BURN POHCs BE SELECTED?

     POHCs for  the trial burn should be selected during development of  the
trial burn plan.  The  selection is in conformance with the regulatory ap-
proach laid out in the Guidance Manual for Hazardous Waste Incinerator Per-
mits  (Reference 1).  In  addition to the regulatory criteria,  the following  •
two considerations should be taken into account:   (1) maximum flexibility of
operating conditions under the permit; and (2) ease of sampling and analysis
during the trial burn.

     Currently  the regulation requires that.a DRE of 99.99% be demonstrated
for the selected POHCs.

     In addition, the  following limits will result  from  the  selection  of
POHCs for the trial burn:
                                    20

-------
           Appendix VIII compounds  in any subsequently burned waste feed must
           be  present in concentrations lower than the POHC in highest con-
           centration during the trial burn.

           Appendix VIII compounds  in any subsequently burned waste must have
           a heat of combustion ranking higher than the POHC with the lowest
           heat of combustion during  the  trial burn.   (Heats of combustion
           for all Appendix VIII compounds  have been  determined and can be
           found in Reference 1).

 Because of these  limits,  the  POHCs  chosen  for trial  burn testing must in-
 clude the Appendix VIII compounds in the waste feed,  usually the compounds
 in the highest concentration and with the lowest heat of combustion.  "Ap-
 pendix VIII"  refers to  the Appendix  of the  hazardous  waste regulations which
 lists compounds which  are  considered hazardous (see  40 CFR  261 Appendix
 VIII).

      It is important that  the  Appendix VIII compound present  in highest
 concentration in any proposed waste feed be present in the feed during the
 trial burn at  the  maximum concentration expected, in order to obtain the
 necessary permit conditions.  Likewise, it is important that the compounds
 with the lowest heat of combustion be present in the  waste feed used during
 the trial burn at  sufficient  levels to determine 99.99% DRE (see Section
 II.H)..

      In selecting POHCs for a  trial  burn, sampling and analysis implications
 also must be considered.  From this point of view, Appendix VIII compounds
 fall, into three categories:

           Volatiles - compounds which can be sampled  using the VOST (in cer-
           tain cases other methods may be more appropriate, as discussed in
           Section III.H)

           Semivolatiles - compounds  which can be sampled using the Modified
           Method 5 train

           Other - compounds which must be sampled using -different techniques;
           special trains,  colorimetric methods, etc.   These include compounds
           which degrade easily in water or  which have special interferences
           or  are  otherwise difficult  to quantify using GC/MS analysis.

      Ideally, all trial burn POHCs could be selected  from either the volatile
 or semivolatile group.   This minimizes the  number of  sampling trains used in
 the field and  simplifies  the  analysis.  If possible  the "other" category
 should be avoided, because more specialized equipment may be needed, which
 will have to be  cleared by permit  reviewers  in advance of the test, and
 may result in higher sampling and analysis  costs.  An additional considera-
 tion is to avoid  POHCs which also might show up as products of incomplete
 combustion from  the burning of  the waste  (e.g.,  chlorinated benzenes,
 ethanes, and  methanes).

      All of these considerations must be taken  into account when selecting
 POHCs for a  trial  burn.  One solution which has been, used at incinerators
.which hope to  burn a wide variety  of wastes,  is  spiking of a low heat of
 combustion compound  (e.g.,  carbon tetrachloride or perchloroethylene) in
 significant concentrations (5-10%) into the waste feed.

                                     21

-------
D.   WHAT  TYPES AND QUANTITIES  OF  WASTE ARE NEEDED AND  HOW CAN THEY BE
      PREPARED?

     The  response addresses  calculation of waste  quantities,  assuring
adequate  supplies  of  waste by type, and preparation  of  wastes.   Specific
problems  addressed  include nixing  of synthetic or  spiked  wastes  and time
requirements.

1.  Quantities of Waste

     The  quantity  of  waste required is dependent on the waste feedrate to
be used during each run,  the  number of  runs, and  the  duration of  each run.
Waste feedrate and number of runs are selected by the incinerator Operator,
and  are specified  in  the Trial Burn Plan.  The. sampling time required  in
each run  is  usually 3 to 4 hr plus 1 hr to line out the unit before start
of testing, and  1  to  2 hr for contingencies (plant operating problems  or
sampling  problems).   Considering these,  a quantity  of waste sufficient  for
8 hr of operation  should  be available for each  run.   If  the trial burn  in-
volved only three runs, at one set of operating conditions, then waste suf-
ficient for 24 hr of operation should be available.

2.  Types of Waste

     Sufficient quantities of waste must be available for each type of waste
feed that is used.  Each type of waste must have all the specific character-
istics' that are required to meet selected operating conditions.   For example,
the waste to  be  burned during a trial  burn might include  both  continuous
feeding of an organic liquid and intermittent  feeding of  drummed solids.
Each of these wastes must meet certain specifications selected for the trial
burn, including POHC concentrations, heating value, Cl and  ash content, etc.

3.  Waste Preparation

     Three methods can be used to prepare the required quantities of wastes
possessing the correct characteristics.  These three methods pertain mainly
to POHC characteristics  but may be  used to achieve any of the necessary
characteristics.   The three methods are:

     a.   Use actual wastes,

     b.   Use synthetically prepared wastes, or

     c.   Continuously  spike  POHCs  into the waste during the trial burn.

     Method (a) usually  is desirable,  if it is possible to acquire actual
wastes that have  the  necessary  characteristics  or to  achieve those  charac-
teristics by  blending of  actual wastes.  Method (b) usually involves using
actual  wastes mixed  with purchased chemical  compounds  (i.e.,  POHCs).
Method (c) is  similar to method (b) except that it applies mainly to con-
tinuous liquid feeds,  with the purchased chemical(s)  continuously  pumped
into this feedline.
                                    22

-------
4.  Mixing

     All of  the  above three methods require  that the waste feeds be well
nixed, but mixing is especially important for methods  (b) and  (c).   Lack  of
good mixing for any waste feed can, and has, caused problems in trial burns.
For method (c) the trial burn may involve continuous spiking of POHCs (pure
components or  mixtures  thereof)  into a liquid waste feed line.  When this
is to be  done, a connection (1/4 or 1/2-in. valve) must be provided, with
another sample tap located further downstream.  It is also highly advisable
for the plant  to install an inline mixer between these two connections to
help ensure  that the  "spiked"  components are  well mixed with the  waste  and
that the samples collected are representative of this mixture.
                  •
                                               «.
5.  Time Requirements

     Another important factor in waste preparation is time.  The quantities
of waste  involved can be rather large, and it may take several weeks  to
acquire sufficient quantities of wastes to prepare a homogeneous batch  with
the proper characteristics.  Storage space for these "special" wastes,  over
some time period, can impact normal plant operations.  Finally, some addi-
tional time  may  be needed  to sample and analyze the wastes  to  be  sure they
have the necessary characteristics.

     Adequate  time also  must  be allowed for numbering, weighing, and sam-
pling of  drummed solids  before the trial burn.  Since the number of drums
may exceed 300,  the  problem of weighing and sampling initially may not be
realized.   Also,  samples of drummed solids  must be representative of those
drums used in each run.  Representative samples may be obtained by sampling
each drum during each run, or by "staging" the drums to be used in each run
and sampling them prior to the trial burn.

E.  HOW MANY RUNS ARE NECESSARY AND HOW LONG IS EACH RUN?

   .  This question was discussed in Sections II-A-3 and II-D-2.  Additional
points offered as guidance are:

          Each run will  require at least 2 to 4  hr.   It is best to plan
          only one run  per day, except in  special cases when  sampling  is
          less complex than usual.   Quite often,  when an  incinerator op-
          erator hears that the sampling time required for each run  is  3  to
          4 hr,  it  is assumed that the sampling  crew can do two  runs each
          day.  However, the  sampling  crew has about 2 hr of work in pre-
          paring  for  each  run and at least 2 hr  of work after each  run is
          completed to recover,  label,  and package each sample.   In many
          instances a variety of problems do occur, both in plant operations
          and  in sampling, so that one 3 to 4 hr  run may involve  a 12 to  16
          hr day  for  the sampling  crew.  The  most reasonable  assumption is
          that one run can be completed each day.

          EPA  recommends three runs at each set of operating  conditions to
          be tested.
                                    23

-------
          If several  sets  of operating conditions are to be tested, regu-
          latory agencies may allow fewer than three runs at each condition.

          Conducting more than six runs may not be cost effective.

F.  HOW MANY PEOPLE WILL BE NEEDED, WITH WHAT EXPERIENCE?

     Personnel required  for  sampling during the trial burn usually number
between 5 and  10,  depending on the complexity of the sampling.  A typical
example, reflecting the  sampling  plan shown earlier  in  Table  2,  is pre-
sented in Table 6.

     The personnel list  in Table  6 is only an example.  In some cases one
person can do multiple jobs depending on sampling frequency and complexity,
and physical layout of sampling locations.  Also, quite often the crew chief
performs one of  the  sampling activities, again depending on complexity of
the sampling activity.  Plant personnel may perform the process monitoring.
However, the data  should be separate from any normal plant operating log,
and usable in the Trial Burn Test Report.

G.  HOW ARE POHC STACK SAMPLING METHODS SELECTED?

     A general procedure to  identify the appropriate  POHC  stack  sampling
methods is outlined in Table 7.   When both  volatile  and  semi volatile POHCs
are present, both  MM5 and VOST are  needed.   Analyses performed on these
samples must  provide  the  necessary detection limits  for  the  POHCs ,  as
mentioned in Tables 7 and 8.

H.  WHAT DETECTION LIMITS ARE REQUIRED FOR THE SAMPLING AND ANALYSIS METHODS?

1.  Waste Feed Detection Limits

     Analyses of POHCs in waste feeds must  be  capable of detecting  the  ex-
pected concentrations, which usually are above 10,000 ppm (1%).  But, it  is
desirable that the detection limit be 100 ppm (commonly  achieved  by recom-
mended analytical  techniques).  A POHC at this concentration or above may
be considered (under RCRA) to be  "significant."

2.  Stack Gas Detection Limits

     Detection limits  required  for POHCs in  stack  gases are  discussed  in
Table 8.  The rule-of -thumb  that  can be used in most  cases  is:

     100 ppm in waste feed = 1 Mg/m3  in stack gas at  99.99% DRE
This  equation can be used  to  estimate stack gas concentrations  for any
waste feed concentration  (i.e., 2,000  ppm in waste = 20 pg/m3  in  stack gas
at 99.99% DRE).
                                     24

-------
        TABLE 6.  A TYPICAL EXAMPLE OF SAMPLING PERSONNEL REQUIRED
      Job
Number of
personnel
Experience required
1.  Sample liquid
    feed (once every
    15 min)

2.  Drum solid
    sampling and
    recording (once
    every 5-10 min)

3.  Sampling ash and
    scrubber waters
    ev^ry 1/2-1 hr

4.  Stack sampling
    MM5

    VOST

5.  Process monitor to
    record operating data
    every 1/4-1/2 hr and
    determine waste feed-
    rates

6.  Field laboratory
7.  Crew chief
                 Technician with sampling ex-
                 perience and safety training.
                 Technician with sampling ex-
                 perience and safety training.
                 Technician with safety training.
                 Experienced console operator and
                 technician for probe pushing.

                 Experience with VOST operation.

                 Engineer or other person expe-
                 rienced in plant operations and
                 trial burn requirements.
                 Experienced chemist for check-in
                 and recovery of all samples, and
                 preparation of sampling equipment
                 for each run.

                 Person experienced in all aspects of
                 trial burn sampling to direct, all
                 activity and solve problems that may
                 occur.
              Total
                                    25

-------
   TABLE 7.   PROCEDURE FOR IDENTIFYING NECESSARY  STACK SAMPLING METHODS
Step 1.   Determine Whether Each PQHC is a Volatile or Semivolatile
          Compound

          Volatile compounds are  generally those that have boiling points
          below 130°C.  Most can  be sampled with VOST  or gas bags.  The
          best way of determining the sampling method needed is to refer to
          Appendix B of Ref.  2.  If the POHC shows sampling by "particulate/
          sorbent" then MM5 is  required.  If it shows "sorbent" or "gas bulb"
          then VOST or gas bags will be the sampling method.  Regardless of
          whether or not  a POHC is a volatile or semivolatile, some POHCs
          require special  sampling  methods as indicated  in Appendix 8  of
          Ref. 2 (e.g., formaldehyde).

Step 2.   Estimate Concentration of Each POHC in the Stack Gas, Assuming a
          DRE of 99.99%                           .

        .  Estimation of the concentration' of each POHC requires some knowledge
          or approximation of POHC concentration in waste feeds, waste feed-
          rates, and stack gas  flowrates.  Using that information, concentra-
          tions of each POHC in the stack gas can be estimated, for an assumed
      •_   DRE of 99.99% (see example calculation in Table 8):

          For semivolatile POHCs,  MM5 is suitable for any stack gas concen-
          tration above 1 |Jg/m3.

          For volatile compounds,  VOST should be used when stack gas concen-
          trations fall within the range of 1 to 100 ng/L.  However, if the
          estimated stack gas concentration exceeds 100 ng/L,  then gas bags
          should also be  used  and analyzed in the event  that  the VOST sam-
          ple concentrations saturate the GC/MS data system.
                                    26

-------
     TABLE 8.  EXAMPLE CALCULATION FOR ESTIMATING  POHC CONCENTRATION
                 IN STACK GAS,  AT DRE OF  99.99%
Basis;

Waste feed flowrate approximately 4 gpo
Density of waste approximately 9  Ib/gal
POHC concentration in waste feed  is near  minimum significant level
  of 200 ppm (200 ppm = 0.000200  g POHC/g feed)

Stack gas flowrate unknown but total beat input to incinerator is
  approximately 30 x 106 Btu/hr with 100% excess air
Calculation;

Rule of thumb (applies in most,  but  not  all  cases):

Each 100 Btu of heat input produces  about  1  dscf of flue gas at 0%
  excess air, or 2 dscf of dry flue  gas  at 100% excess air

1.  Flue Gas Flowrate:

           »-«»• ET   a                • 5>°°° —/-
2.  Waste Feedrate:
                   fa i*»\  / ACA  »\
                                  = 16,300  g/min


3.  POHC Input Rate:
                  (.,- .   >
           16,300 &  feed I  I 0.000200 fi  V""! Is 3.26 g POHC/min
                  nui  I  \          g  *^^/« i       =


4.  POHC Stack Output Rate,  at  99.99%  DRE:

          (3.26 g/min) (1.0-0.9999) -  0.000326 g/min

5.  POHC Concentration in  Stack Gas (at  99.99% DRE):
                                   27

-------
     Some POHCs may require special sampling/analysis methods or may show
low recovery  efficiencies  for MM5 samples.  Therefore,  each case  must  be
considered separately  to  ensure that the  detection  limit for the  methods
used are  low  enough to quantify those specific  POHCs at  the  concentrations
expected in the stack at 99.99% DRE.  Consultation with analytical chemists,
or with  the  authors and EPA project officer  given in Ref. 3 can be most
valuable in this regard.

3.  High Concentration of Volatile POHCs

     Since the GC/MS analytical techniques for MM5  and  VOST samples can
easily achieve a  detection limit of 1 M8/°3» there is usually no problem.
However, high stack concentrations  of some volatile  POHCs may exceed the
range for VOST samples  (i.e.,  that  saturate the GC/MS).   Those samples  re-
quire use of  gas  bags in order  to  determine if 99.99% DRE was or  was not
achieved.  The  gas  bags are analyzed by  transferring a small volume of
sample onto a VOST  trap prior to GC/MS analysis.   For example, 5 L may be
taken for analysis.   This  quantity is 4 times less than the quantity sam-
pled by VOST under normal sampling conditions.

I.  WHAT QA/QC MEEDS TO BE DONE?

     It is important in planning the trial burn to stipulate exactly what
QA will be done  and to know why it is needed.  Some QA  activities may  be
desirable but are not essential in specific cases.  Blanket statements that
"full QA" will be employed in the trial burn are not definitive, and exces-
sive QA can drastically increase costs.  An example  list of  basic  QA for a
trial burn is given in Table 9.  Preliminary discussion of  QA procedures
with the responsible regulatory agency is  recommended prior to submittal  of
the Trial Burn Plan.

     One example  of a QA activity  that may be  specified without adequate
thought is "chain-of-custody."  The number of samples collected in a trial
burn normally numbers 100 to 300.  Adherence to chain-of-custody procedures
for all of these  samples requires  considerable time and effort, with its
associated cost impacts.  Unless there is  reason to believe that sample  re-
sults will be a part of some  judicial proceedings, chain-of-custody  proce-
dures on  the  samples may be an  unnecessary  added cost when traceability
procedures would suffice.

J.  HOW  IS IT BEST  TO PLAN FOR THE  POSSIBILITY  THAT  TRIAL BURN RESULTS  ARE
      OUTSIDE RCRA REQUIREMENTS?

     There is always the  haunting possiblity that the Trial Burn  results
may show failure to meet one or more of the RCRA requirements.  This result
is more  likely  when the trial burn is conducted under "worst case"  condi-
tions at which the plant wants to be permitted to operate.

     A miniburn  and other preliminary testing  (e.g., Method 5)  can help
identify problems before  the trial burn,  and,  after modifications,  avoid
failure during  the  trial burn.  Another alternative  is to conduct runs  at
two sets  of  operating conditions.  One set would be worst case, while  the  •


                                     28

-------
         TABLE 9.  EXAMPLE QA FOR A TRIAL BURN
All equipment used  in S&A activities should  have  written cali-
bration procedures.   Procedures and  documentation of the most
recent calibration should be available.

Traceability procedures (not necessarily chain-of-custody) should
be established to ensure sample integrity.

A  GC/MS  performance  check  sample  should be analyzed each day
prior to sample analysis.  If results are outside acceptable lim-
its, samples should not be run.

All samples  from  at least one run should be analyzed in tripli-
cate to assess precision.

A minimum  frequency of check standards  (5% is suggested) should
be  used  with each  sample batch.  Analysis  of actual samples
should be suspended if check standards are outside of the desired
range.

Blank samples should be analyzed to assess possible contamination
and corrective measures should be taken as necessary.  Blank sam-
ples include:

     Field blanks - These blank samples  are  exposed  to  field  and
     sampling conditions and analyzed to assess possible contami-
     nation  from  the  field  '(a minipnim of  one for  each type of
     sample  preparation  or  the  number specified by the  appropri-
     ate method).

     Method blanks -  These  blank samples  are prepared in the
     laboratory and  are  analyzed to assess possible  laboratory
     contamination  (one for each lot of samples analyzed).

     Reagent and solvent blanks - These  blanks are prepared  in
     the laboratory and  analyzed to determine the background of
     each  of the  reagents or solvents used  in an  analysis  (one
     for each new lot number of solvent or reagent used).

Field audits and laboratory performance and systems audits may be
included in  some  cases.   Cylinders 'of audit  gases for  volatile.
POHCs are available from EPA.

A minimal  level  of calculation checks (e.g.,  10%) should be es-
tablished.
                           29

-------
other set could be conditions that increase-the chance of passing and that
the plant could tolerate for continued operation.   The latter would be less
desirable and would  not be cost efficient if the plant passed under worst
case conditions.  Therefore, it  may  be possible to test only under worst
case conditions.  If  the plant  fails the RCRA requirements, then perhaps
the contingency plan  could be  to request a variance and retest under the
other set of  operating  conditions,  as soon as possible after the results
from the first test are available.
                                    30

-------
                                 SECTION IV

                          CONDUCTING THE TRIAL BURN


      Many questions arise when  preparing  for and conducting a trial burn.
 Some questions  which may be  asked are:

      A.    What  is  involved in preparing for the tests?

      B.    What  is  involved in conducting the actual sampling?

      C.    What  is  involved in analysis  of samples?

      D.    How are sampling and  analysis data converted-to final results?

      E.    How are  the data and results  usually reported?

      Each of the above  questions are broad and cover many specific items.
 Subsequent sections of this manual will attempt to address these questions
 by discussing specific  areas  that are  most important and  areas  that may
.cause problems.  In  formulating answers,  it has been assumed that the in-
 cinerator operator has  obtained all necessary approvals  of the Trial Burn
 Plan and is preparing to implement that plan.   At that  point, the realities
 of the test come to the forefront and answers are needed to many questions
 like those discussed below.

 A.  WHAT IS INVOLVED IN PREPARING FOR THE TESTS?

      Preparations  for the test  are numerous; several of the most important
 items are scheduling, sampling  crew activities, equipment preparation and
 calibration, facility readiness, process data,  data sheets and labels, and
 safety precautions.  One potential problem tha-t should be addressed  during
 preparation is  how to coordinate with  observers  during  the trial burn.

 1.  Schedule

      Many scheduling problems can occur if the lead time required is not
 anticipated. These include:
                                     31

-------
          Item
Time Often Required
Hake additions or revisions to Trial

  Burn Plan

Acquire all wastes needed

Select test contractor

Pretest site visit by contractor

Notify all regulatory agencies

Install necessary sampling access
Test contractor begins preparation

Numbering, weighing all drum solid feed
Conduct test

Sampling equipment teardown

Analysis of samples

Report of results from contractor
Submit results to EPA
        Varies
        Varies
3 months before test

1 month before test
1 month before test
complete 1 week before test
2-3 weeks before test
2-3 days before test
3+ days
1 day after last test
1-1 Jj months after test
2 months after test
3 months after test
     If a minibura  or other preliminary testing is involved, it should be
done at least 2 months before the Trial Burn, which would  increase  some  of
the required time intervals shown above.

2.  Sampling Crew

     If a sampling/analysis  contractor is used, that contractor will have
to make crew selections  and  assignments at  least 2 to 3 weeks prior to the
test.  Many of those crew members and the analytical personnel will perform
the complex activities for equipment preparation and calibration, prepara-
tion of all absorbent traps (MM5 and  VOST)  and special cleaning of  all
sampling containers.  Other logistical arrangements for transporting equip-
ment and personnel  to the site must also  be made.   For these reasons,  a
firm date for the test should be established, in concert with the S&A con-
tractor, at least 1 month prior to the test.

3.  Equipment

     The large  amount of sampling equipment needed  for a Trial Burn is
usually surprising  to the operator.   A list of some of this equipment is
shown  in Table  10.   What that list does not show is the detailed prepara-
tion procedures for much of the equipment.   For example:
                                    32

-------
        TABLE 10.  EXAMPLE LIST OF SAMPLING EQUIPMENT AND SUPPLIES TYPICALLY USED
Method 5 - particulate train

Console with pump heaters
Sample box with pump heaters
Umbilical-sampler hookup (gooseneck)
Umbilical cords and adapter (elec.)
Probe tips (4/set)
Probes (type-SS, glass)
Extra quartz inserts
Extra glass inserts
"S" type pitot (ft-5)
Rails - dexangle
Port-rail clamps (collar size 4")
Probe-support
Sample box guide attachment
Potentiometer
Spare thermocouples
Intercom (with cable)
Manometer, inclined
Manometer "U"_
5-impinger foam inserts (for sample box)
Digital pyrometer
Umbilical thermocouple adapter
Submersible pumps
Latex tubing for condensers
Glass tape (high temp)
Console supply briefcase
Method 5 - glassware

Aluminum cases
Impingers
"U" connector
90° connector
Cyclone
Filter holder with, frit
Cyclone bypass, 90°
Teflon sleeves (45/50)
2-liter impinger bottle
Condenser
XAD's
U-tubes
2 L bottle foam inserts
Socket flask - 500 ml
Plastic caps for probe ends
Miscellaneous clamps, gaskets,
  stoppers
Method 4 - moisture train

Probe
Midget impinger
Midget connectors
Glass wool
Silicone grease
Micrometer valve
Vacuum pump
Vacuum gauge
Impinger box/ice bath
Dry gas meter w/thermometer
Spring clips
Vacuum tubing
Stopwatch
Integrated gas train

Grab saaple
  Probe
  Squeeze bulb
•  Gas bags
Integrated gas bag train
  Probe
  Midget impingers
  Micrometer valve
  Pump
  Rotameter
  Box w/bag insert
  Bag
  Pitot tube (S-type)
  Inclined manometer
  "U" manometer (H20)
  Vacuum gauge
  Purge fitting
  Miscellaneous tubing
  Glass wool
  Sealant
Analysis
  Orsat analyzer
  Spare Orsat parts
VOST equipment

Teflon line
Rotameter
                                            33

-------
                                 TABLE 10.  (continued)
VOST equipment (concluded)

Spare 602 polymer
Probes
Dry gas meter
Manifold
VOST - glassware and fittings
  Condenser
  Fittings
  Lopinger
  Teflon tubing
  Glass stopcock
  Glass tapered joint
Spare glassware
Teflon sleeves for glassware
Tenax traps
Clean coolers
Recorder
Nitrogen cylinder
Air cylinder (THC)
H2/N2 cylinder (THC)
Calibration gases
Gas regulators
Teflon tubing
Ascarite trap
Silica gel trap
Silica gel
Ascarite
Heated lines
Controller
Variacs
Probe filter
Parts box
Fyrite 0? and CO?

 •Fyrite (21% scale)
C02 Fyrite (20% scale)
Sample pumping line consisting of:
  - hose from probe to filter
  - filter
  - aspirator bulb w/check valve
  - rubber connector tip
Spare parts
  Filtering yarn
  Diaphragm
  Gaskets
Spare chemicals
Orsat - stand
02 buret
C02 buret
Manifold
Graduated buret
Leveling bottle w/tubing
Orsat sampler
Mylar bags
Continuous monitors

   ditioning manifold
    analyzer
CO analyzer
0, analyzer
Safety equipment item

Hard hats
Hard hat liners
Safety shoes
Safety glasses - goggles
Slip on shields for glasses
Ear plugs
Neutralizer
Ear protectors
Face masks (full face respirators)
Dust respirators
Climbing belt and lanyard
Gloves high temperature
Rain gear
First aid kit
Water jug
Saf-T-Lok
Fire extinguisher
Face shield
Viton gloves
Jumpsuits
Restricted area sign
Black and yellow ribbon
Tarps
No smoking sign
Respirators
Cartridges - organic vapors and acids
Eyewash bottle
Fire blanket
                                            34

-------
                                 TABLE 10.  (continued)
Computer and associated items

Hardware -
Pocket computer
Printer cassette interface
Cassette recorder
Cassette to interface connecting cable
Ribbon cartridge
Printer paper
Level II basic reference manual
Extra blank cassette
Software
Program printout copies
Laboratory-general equipment

Lab tool box
Oven
Kimwipes
Chix wipes
Wash bottles
Glass wool
Barometer
Plastic for lab floor
Graduated cylinders
Pipettes
Funnel glass
Beakers
Thermometer (0-125)°C
Bulbs for disposable pipettes
Brushes and soap
pH paper
Ultrasonic cleaner
Triple beam balance w/weights
Ramrods with brushes
Wash tub
Filter holder clamps
  Distilled-deionized H20 double-distilled
  B and J Acetone (1 pt/part. Tests) gal.
  Silica gel (2 run/lb) (large can)
  Spare Fyrite chemicals
Particulate (Method 5)
  Filter paper (glass fiber)
  Sample bottles (glass)
  B and J methanol
  60 ml poly bottles w/caps
  0.1 n KOH
General items

Spare hardware equipment
Electrical tool box
FM 2-way radio
Heat gun
Sample labels
Ice bags
Lab notebooks
Air tank for blow-back
Portable welding unit
Label tape and dispenser
Data sheets
SOP's
Traceability sheets
Timers for consoles
Rubber bands
Clipboards
Paper clips
Scotch tape
Masking tape
Laboratory-chemicals

General -
  Orsat chemicals:
  3 oz Oxorbent (02)
  3 oz Cosorbent (CO)
  3 oz disorbent (C02)
  3 oz burette solution
                                           35

-------
          Calibrate all MM5 consoles.
          Condition and check all VOST traps.
          Clean and pack all MM5 resin traps (XAD).
          Preweigh all MM5 filters.
          Clean all glassware and sample bottles.
          Purchase special reagents and solvents.
          Modify equipment for special sampling situations.
          Reserve time for use of analytical instrumentation.
          Collect and pack all necessary field laboratory equipment.

     In many cases, the complexity of the sampling plan or lack of available
plant facilities requires provision of a large field trailer for the samples
and sampling equipment.  This trailer also  serves as the  field  laboratory.

A.  Facility Readiness

     Facility readiness is  critical  to  conducting the  test  as  scheduled.
Checking operational  readiness  of  the incinerator and its components, in-
cluding  critical  instrumentation (especially flow meters),  is  vital and
should be done  early enough to correct any problems identified.  When the
plant is operated under worst case conditions, unanticipated problems often
occur.  The plant should be operated under test planned conditions prior to
the tests to minimize costly delays during the test.

     Other facility  readiness needs  are identified during the pretest  site
survey.  This survey will  identify most of the sampling needs, especially
those related to the stack  sampling  ports and sampling  platform, which can
require some installation work by the plant.  Frequently the contractor will
need to rent a trailer to be used on site for sample workup and storage.  A
suitable location  for the  trailer should be identified during the survey.
The survey also will identify other needs such as electrical supply require-
ments.  (These requirements are usually much larger than the plant expects.)
The survey should  be conducted  at  least 1 month  prior to  the test  to  allow
time for modifications to the facility.

     Facility readiness also  includes  preparation of all the wastes to be
used in  the  tests.  Waste preparation is especially important for drummed
solid waste.  It  is  highly desirable to have all drummed waste on site at
least a  week prior to the test  and  have  the drums arranged in a staging
area in  the  order  that they will be  used.   These preparations will facili-
tate numbering, weighing, and sampling of each drum.  Drum preparation can
require  considerable  time  and  effort on the part  of the plant operating
staff and the sampling crew.

5.  Process Data

     Process data  recorded  during  a  Trial Burn is  of equal importance with
the sampling activity, for two reasons:

          Process  data are necessary for computation of DRE.
          Some  process readings recorded during the Trial Burn may,  and
          probably will, become the  limits  specified in the  operating per-
          mit (e.g., minimum combustion chamber temperature).

                                     36

-------
     The average  feedrate  of each waste and fuel input stream must be de-
terminable for each run by flovmeter, tank level change, drum weights, etc.
Those feed rates,  and analysis results for waste feed samples are used to
compute thermal input rate  (BTU/hr)  and POHC input  rate.  These parameters
are used to calculate DRE.

     Three steps for process monitoring are recommended:

     1.   Before the  test,  determine what process  data must be taken and
reported to EPA.

     2.   Record all  possible  data,  to help identify any problems during
the test, or in the test results.

     3.   Before the test, establish the acceptable range for each critical
operating parameter.

     Item 3 above  is  not as  simple as  it may appear.  For example, combus-
tion chamber temperatures  for  the Trial Burn might  be 2000°  ± 50°I.   Ques-
tions 'then arise  as to what if the temperature range is not maintained at
all times during the test:

          Is sampling to be interrupted if the temperature goes outside the
          established range?
          How long  can the temperature be outside  the  range,  or how far
          outside the  range, before  ordering an interruption in sampling?  .
          If sampling is interrupted, how long must the temperature be back
          within range before sampling can be restarted?

     This one example demonstrates  the complexity  of questions that  fre-
quently arise.  These questions  should be anticipated and guidance devel-
oped before the Trial Burn to assure trial burn operating conditions that
meet plant  needs.  There  is often  precious little time to make those
decisions when the questions are faced during a test.

6.  Data Sheets and Labels

     Preparation of all data sheets and sample labels that will be needed
for the test is important.

     Many different data sheets are needed for a Trial Burn as shown Table 11
The units of measure must be shown for every item on every data sheet.  Too
often, data are taken  (i.e., numbers recorded) without showing the units  of
measure  (e.g.,  °T,  gal/min,  etc.).   Instrument factors (e.g., Rdg x  100  =
°F) should be  noted during trial burn preparation  to assure that data are
accurately compiled.   Data sheets  may be a better  record than copies of
strip charts, since the latter do not  show units  of measure  or multiplica-
tion factors and are often difficult to interpret for other reasons.

     All data sheets  should  be prepared before  the  test to ensure that all
necessary data are recorded.  Specific assignments  should be made as  to who
                                    37

-------
                  TABLE 11.  EXAMPLE LIST OF DATA FORMS
Traverse Point Locations
Preliminary Velocity Traverse
Method 5 Data Sheets
Isokinetic Performance Work Sheet
MS Sample Recovery Data
Integrated Gas Sampling Data (Bag)
Orsat Data Sheet
VOST Sampling Data

Drum Weighing Record
Drum Sampling Record
Liquid Waste Feed Sampling Record
Fuel Oil Sampling Record
Drum Feed Record

Process Data (Control Room)
Miscellaneous Process Data (In-Plant)
Tank Level Readings

Log of Activities

Ash Sampling Record
Scrubber Waters Sampling Record
Sample Traceability Sheets
GC/MS Data Calculation Sheets
Note:  Units of measure must be shown for each item on each data sheet.
                                    38

-------
is responsible  for  each sheet.  The data to be recorded can be identified
during the pretest survey, and the data forms prepared thereafter.

     Sample labels  should also be prepared prior to the test.  Labels can
be prepared most efficiently using computer printed labels like the example
shown in Figure 3.   Replicate labels which each contain all essential in-
formation including  a  unique sample number for each  sample  are printed.
Replicate labels are needed  in order to place  one on  the sample container,
one on the  outside  of the final packaged sample,  one in the field labora-
tory log book and leave the  fourth  on  the sheet of  labels.  The last  label
provides a  quick way  of  checking which samples were  taken,  since some
changes may be  made in the  field.   Blank labels are also always provided
for such changes or additional samples that may be taken.

     Preparing the computerized labels requires careful thought to identify
each and every  specific 'sample that will be taken during each  run,  includ-
ing all replicates  and blanks.  Label preparation also helps identify all
the sizes and types of sample  containers that will be needed, how they must
be prepared, and the number of each that is required  (including spares).
This activity often shows that 50  to  100 individual  samples  (and  labels)
will be involved in each  run;   Given this magnitude of samples, preprinted
labels with specific sample  names and  a consistent  numbering system should
be prepared before  the Trial Burn to help avoid confusion and errors that
can occur if labels are prepared later in the field.

7.  Safety Precautions

     Most plants and sampling  crews utilize common safety equipment such as
safety glasses,  steel-toed shoes  and hard hats.  However,  an outside con-
tractor's sampling crew needs  to be made aware of all plant safety require-
ments and any special  hazards that  may exist,  especially  with regard to
particularly toxic  components  in  the feed streams  or  the  stack effluent
(e.g., high CO levels).

     Sampling personnel, need to be  instructed  on any  special safety equip-
ment and procedures for liquid waste  sampling or  sampling of any  other
hazardous waste  feeds.  The  need  for protective equipment such as  specific
types of  gloves, goggles,  respirators,  chemical  resistant suits,  etc.,
should be established  early enough  in the planning stage so that the sam-
pling crew can prepare for their use.  Once at the test site, plant person-
nel must ensure that the  sampling personnel are  informed  of the plant's
safety procedures,  especially if they could  impact on the test program
(e.g., evacuation of the  sampling area caused by  a process  upset  in an
adjoining portion of the plant).

     One special note  about safety  for sampling of drummed wastes  is en-
countering "bulging" drums.   A bulging drum can, of course,  indicate  pres-
sure buildup in the drum.   If a bulging drum is encountered in the course
of drum sampling,  only experienced plant personnel should' attempt to open
it.  Furthermore, a drum  can be under  pressure, even  if  it is  not  bulging.
It is recommended  that plant  personnel be assigned, to assist the sampling
crew and be responsible for  opening  all drums to be sampled.
                                    39

-------
RUN * 1         101
Liq. Organic Waste Feed

Proj, *     DATE1
Plant Nane
RUN i 1         102
Aqueous Waste Feed

Proj, *     DATE I
Plant Nane

RUM t 1         103
Kiln Ash Sluice Water

Proj. t     DATE!
Plant Name
RUN * 1         101
Liquid Scrubber Effluent

pioj. *     DATE:
Plant Nane

RUN t 1         105
MOST Trap
Pair il» Tenas:
Proj. *     DATE!
Plant Nane

RUN i 1         106
MM5 Caustic Solution

Proj. i     DATE!
Plant Nam?
RUN * 1      .101
l.iq. Organic Waste Feed
               i
Proj. *    DATE!
Plant Nane
RUN t 1        102
Aqueous Waste Feed

Proj. *    DATE!
Plant Nane

RUN * 1        103
Kiln Ash Sluice Water

Proj. *    DATE!
Plant Nane
RUN * 1        101
Liquid Scrubber Effluent

Proj. *    DATE!
Plant Nane
RUN * 1        105
VOST Trap
Pair *1, Tenant
Proj. *    DATE!
Plant Nane

RUN t 1        106
MMS Caustic Solution

Proj. t    DATE!
Plant Name
RUN * 1           101
Liq. Organic Haste  Feed

Proj. *    DATEt
Plant Nane
RUN * 1           102
Aqueous Waste Feed

Proj. *    DATE!
Plant Nane
RUN t 1           103
Kiln Ash Sluice  Mater

Proj. * .   DATE!
Plant Nane
RUN * 1           101
Liquid Scrubber  Effluent

Proj. *    DATE!
Plant Nane
RUN # 1
MOST Trap
Pair tl» Tena::
Proj. *    DATE!
Plant Nane
105
RUN * 1           106
MM5 Caustic Solution

Proj. *    DATE!
Plant Nane
                       Figure 3.  Example of computer labels.

-------
8.  Observers

     The. operator  frequently does not realize  until a test starts  that
trial burns bring  out everyone with a vested interest and  even  those who
are just interested.   Observers  may include regulatory authorities, extra
operating and maintenance personnel, responsible plant management, and many
others .who otherwise  are seldom present.  They  all  usually congregate  in
the control room.  These "observers" usually have reason to be present,  but
their numbers can create problems.

     Observers want to  ask questions and to have lengthy discussions with
the plant operators.  Some of  this may be necessary, but it can  divert  the
operators'  attention  from  their primary function  and responsibility.
Similarly,  the observer may  want to ask questions of the sampling crew at
times when they  must  give their full attention  to their sampling  respon-
sibilities.   Also,  suggestions made by "observers" to operators or samplers
are sometimes  interpreted as  a  directive to change how they  are doing
something.

     To help avoid  the above problems, the following should be done prior
to the trial burn:

          Instruct  all  operating personnel  and  samplers not to  make any
          changes unless  directed  to do  so by their supervisor  or other
          designated individuals.

          Require that  each  observer minimize discussion or interference
          with operators or samplers during busy periods, especially during
          test periods.

          Assign one plant person as the primary contact for all observers,
          and request  that  the observers direct questions  and comments to
          that person first.

          Since most observers are interested in what is being done and how
          it is  being done,  have descriptive material  available and,  if
          needed, make  arrangements  for  them to discuss the test plan and
          sampling/analysis  methods  with appropriate personnel  at  appro-
          priate times before or after the actual test periods.

B.  WHAT IS  INVOLVED  IN CONDUCTING  THE  ACTUAL  SAMPLING, AND WHAT  ARE THE
      PROBLEMS THAT MAY OCCUR?

     The main factors involved, in the actual sampling for a  trial burn are:

          Equipment setup
               Sampling train setup
               Setup waste feed  sampling
                                    41

-------
          Preliminary testing
               Velocity traverse
               Cyclonic flow check
               Moisture measurements

          Actual testing
               Waste feed sampling
               Process monitoring  and determination  of  waste feed rates
               Sampling of ash and scrubber waters, etc.
               Stack sampling
               Sample recovery
               Labeling and sample packaging/storage
               Preparation of equipment for next run

          Equipment dismantling and packing

     Brief discussions of the above  items  and  procedures  to  avoid  problems
that may occur, are presented below.

1.  Equipment Setup

Sampling Train Setup—
     The first job  of the sampling crew after arriving at the facility is
unloading and  setting up of equipment  (usually into a field  laboratory
trailer) including  setup  for  the stack sampling.   Setup  on  the  stack for
MM5 is- usually the  most difficult  step.  First,  relatively heavy equipment
(40 to  80 Ib)  must  be moved up to the sampling platform.  Second, support
rails or a monorail must be installed to  allow the MM5  train to traverse
the stack.  These rails often must extend  outward  from  the port  a  distance
equal to the  stack  diameter plus  about 2  ft.   The rails must be rigidly
secured to support  the sampling train over its entire length  (~ 8 ft) and
allow free movement with  no interfering objects (e.g., guard  rails).  The
problem usually  encountered is  the lack of means  to support the rail,
especially at the outer end which may extend 4 to 8 ft further out than the
platform.   (Some platform  designers assume that a  6-ft diameter stack can
be sampled from a 2-ft wide platform).  Also, the sampling ports are  almost
always  at about  the same level as the platform guard rail,  so part of the
guard rail must be  removed  to provide clearance for the  sample box,  if not
done earlier.

     Stack samplers have necessarily developed various  means of  supporting
the rails, but each test site always  requires something slightly different.
A pretest survey may identify some modifications the  plant  can make  to fa-
cilitate the  stack  setup,  consistent with  the  design of the  rail system to
be used.  But ideally, the width of the platform would be at least equal  to
the diameter of the stack plus 2 ft.

     Another common problem encountered during stack setup  is inadequate
electrical outlets.  As  a minimum, .at least four 110-v, 20-amp  electrical
outlets should be available on the stack sampling platform.   If  at all pos-
sible,  these circuits  should be dedicated  to the test, without interference
from other plant equipment.
                                    42

-------
Setup for drummed- waste sampling—
     The equipment setup period may include numbering, weighing, and sampling
of drummed wastes.  Plant personnel will be needed to assist in this activity.
Adequate time must be set aside for this work, and the plant needs to preplan
the work  so that the wastes are  properly staged and equipment provided
(scales, forklift, etc.).   Preparation of drummed waste  is  done most ef-
ficiently with 2 or 3 person teams.  Each team can process a drum in 2 to 5
min.

2.  Preliminary Testing

Velocity Traverse-
     One of  the  first preliminary tests  is a  preliminary  velocity  traverse
the stack and determination of stack moisture content.  Values obtained are
used to determine sampling train conditions, so the preliminary measurements
need to be  made  at test conditions.  .This  preliminary  test will require
additional time on the first test day.

Cyclonic Flow—'
     Another preliminary test is a check for cyclonic flow in the stack, as
required by EPA Method 1.  Sampling under cyclonic flow conditions requires
special equipment  and methods.   Alternatively, a flow straightener can be
installed in the stack,  as far upstream of the test ports as possible.  This
installation could involve considerable delay to design,  fabricate, and in-
stall' the flow straightener.   Installation may require a plant shutdown.
If there is any reason to suspect that stack flow may be  cyclonic, then the
plant should either  have the flow checked or install a flow straightener
well in advance  of the  tests to  avoid the possibility of having to  delay
testing at the last minute.

Moisture Measurements—
     Another possible problem  is  high moisture content of the stack gases
from wet scrubbers.  For scrubber stack test, crews should determine moisture
content at saturated conditions prior to the particulate  run to ensure that
the runs are conducted under isokinetic conditions.

3.  Actual Testing

     Sampling usually  consists of  taking representative samples  of  all
influent and effluent  streams,  especially waste feeds and stack effluent,
during  each of  three runs.  Process  monitoring,  including  collection of
data needed  to determine all waste feed  rates,  is also  done during each
run. When recording process data, common practice is  to read and record the
instrument  readings  once every 15 min, even if they  are  also continuously
recorded.   These manual readings provide  a  good written record for  in-
clusion in  test  reports.  The results reported should include notation of
momentary excursions.   Otherwise, the operating permit might  not  contain.
any allowance for those types of  occurrences  that are  a part of  normal
plant operations.

     During  the  test one person  who  knows the conditions under which sam-
pling should be  interrupted, and who  is  in radio contact with the stack


                                    43

-------
test  crew  at  all  times  should be  responsible  for  process  monitoring.   That
person must notify the  crew to interrupt  sampling whenever  deemed neces-
sary, but  especially when a serious process  upset  or a  shutdown  occurs
during a test.   Such transient conditions could  have  a drastic  impact on
the samples,  so  all sampling must be stopped immediately.  In that event,
sampling can  be  resumed after the desired plant operating conditions have
been reestablished.

     Immediately after each run, the sampling crew must recover all samples,
properly label each,  and package them for storage and shipment.   Samples
are usually double-bagged  with  protective wrapping and stored in coolers,
many of which must.be iced each day.  After that work has been completed,
all sampling  equipment  must be prepared for  the  next  run.   All this may
sound relatively  simple, but there are numerous  problems that can'occur:
Some of these problems  are listed in Table 12  and  are briefly explained
below.
                    TABLE 12.  POTENTIAL PROBLEMS THAT
                                 MAT OCCUR DURING TESTS
                    Plant operating problems
                    Determination of waste feed rates
                    Weather
                    Sampling equipment problems
                    High filter vacuum due to filter
                      loading
     Plant operating problems during a run are not uncommon, and are due in
part to the fact that the plant is being operated under "worst case" condi-
tions.  This  scenario may cause  operating  conditions  to go outside the
specified range for the test, or upsets may occur to cause a plant shutdown.
Such situations require  interruption of all sampling until desired condi-
tions are reestablished.

     Any interruptions in sampling, for whatever reason, can have an impact
on proper  determination of  waste feed rates, especially liquid wastes.
This may not  be much of a problem if  the plant  is equipped with reliable
waste flow meters.   But,  if tank  level  changes  are  the basis for deter-
mining feed rate, then  levels have to  be read whenever  the  sampling  is  in-
terrupted and again when it is restarted.

     Another problem  that  can often occur during testing is bad weather.
This can alter  plant  operating  conditions, but usually  it impacts  the  sam-
pling activity.  Heavy rain, lightning, or high winds are common conditions
that will  require interruption in  sampling,  since  it  is unsafe for the
sampling crew to remain up on the stack unless it is be enclosed.

-------
     Most of the other types of problems that occur during a test relate to
the stack sampling equipment.  Equipment used to conduct stack sampling per
Method 5 include pumps, heaters, thermocouples, etc., all of which are sub-
ject to mechanical  failure.  Another  type  of problem  is  high pressure drop
across the  sampling train,  usually caused by material on  the particulate
filter used  in  the  train.  When this occurs, sampling must be interrupted
for 30 to 60 min to change filters.

     The most serious problems related to the stack sampling are failure to
achieve prescribed  sampling  rate  (i.e.,  isokinetic sampling) or  failure  to
pass final  leak checks.   Either situation can invalidate  a  run, probably
requiring that  it be  repeated.  The plant  must be aware  that this can hap-
pen, requiring additional quantities of waste and additional time.
          •
C.  WHAT  IS  INVOLVED  IN ANALYSIS  OF SAMPLES AND WHAT  ARE THE PROBLEMS THAT
      MAY OCCUR?

     Even in a  relatively simple trial burn, 100 to  200 samples may be
acquired for analysis.  Each sample will  need to be  analyzed for several
different parameters  (HHV, Cl, ash) and several analytes (POHCs).  Numerous
problems can occur  in the analysis phase  of a trial  burn  due to the  com-
plexity of  sample analysis,  the variety  of sample types, and the detection
limits required for some samples.  Most frequently,  these problems occur ia
the analysis of samples  for POHCs.  Some of the most common problem  areas
are:

          Sample check-in

          Analysis directive

          Sample inhomogeneity

          Analytical  interferences

          Saturation  of GC/MS detection instrumentation

     Usually the analytical  results are  reported to the  project  leader who
is responsible  for  using  them to  calculate final results (DRE) and to pre-
pare a test  report.  The project leader first examines the analytical re-
sults.  Potential problem areas which may appear in the data are:

     •    High blanks

          Poor precision for analysis of replicate samples

          Poor  accuracy  for  recovery of  surrogate  recovery components
          spiked into the sample prior to  analysis

          Nonconformance with "expected" results

Each potential problem area is discussed below.
                                    45

-------
1.  Sample Check-in

     Samples taken during a trial burn are usually brought  to  the  analyti-
cal laboratory  for transfer  to an analytical task leader.  At that point,
when the  samples  are checked in and transferred, the project leader needs
to cross-check  that  each sample taken in the field has arrived and is in-
tact.  Any missing samples  can in this way be  immediately  identified  and
hopefully located.   Also, any extra samples that have  been taken can be
identified.   The  sample  information should be recorded in a bound labora-
tory record book- (LRB) and sample traceability  sheets should be used,  even
if strict chain of custody sheets are not used.

2.  Analysis Directive

     After completing sample  check-in, the project leader should prepare  a
written directive for sample analysis specifying the following for each and
every sample:

          Sample number and type of sample (e.g., 141 - waste feed).

          Notation of any safety or hazard considerations in handling/ana-
          lyzing samples.

          Analysis parameters and analytes.

          Analysis methods for each parameter or analyte.

        .  Analysis sequence (where necessary).

          Indication of  whether sample  is to be analyzed in duplicate (or
          triplicate).

          Samples to be  spiked with analytes and surrogate recovery com-
          pounds for determination of percent recovery efficiency.

The directive should also  indicate expected concentrations, analysis pri-
orities, and the schedule for reporting of analysis results.

     Preparation of  the  above directive is complex but  very important.   It
documents the analysis needs  and will be very  helpful  to  those who .must
perform the analyses.  Host importantly, it helps assure that all  the neces-
sary analyses will be done as needed to satisfy the trial burn requirements.

     The analysis sequence can be especially important for MM5 samples.   If
the same  sampling train  is  used to  collect particulate and POHC  samples,
the particulate  filter must  be desiccated and weighed prior to extraction
for POHC  analysis.   Another  example is analysis of a filter for POHCs and
metals.   Since  one precludes  the other, it would probably  be necessary to
cut the filter  in half.   Such  examples are presented here  to  demonstrate
the importance of written specifications for the sequence in which analyses
must be done.
                                    46

-------
3.  Sample Inhomogeneity

     When analyses of samples are initiated, a problem often encountered is
nonhomogeneity  of  samples, especially waste  feed  samples.   Liquid waste
feeds  often  separate during  shipment/storage.   Steps must  be taken to
homogenize samples before  portions  or aliquots are  removed  for  analysis.
Solid waste  samples  often  present even more difficult problems that  should
be discussed with the analysis task leader.

4.  Analytical Interferences

     Waste feed  samples  are organically complex and usually contain many
constituents other than POHCs.  These other constituents may interfere with
the POHC  analysis,  especially if the POHC concentrations are  low  in com-
parison to  the other constituents.   Some  type  of  sample cleanup may be
required.                                          .

5.  Saturation of GC/MS Data System

     Saturation  of the GC/MS  data system may  occur  for any type  of sample,
but especially VOST traps, VOST samples are analyzed by thermal desorption,
and cannot be prescreened  or diluted.  If saturation occurs, gas bag samples
must be analyzed.  Saturation can be a complex problem if there are several
volatile POHCs with a wide range in their anticipated concentrations.  VOST
results may  be essential for determining DRE of some POHCs present at low
concentrations  in waste  feeds,  but those POHCs may be masked by high con-
centration of  other POHCs in the VOST  samples.   Such 'problems  are  best
avoided by careful  preplanning,  including selection of POHCs  and/or their
concentrations in the waste feeds.

6.  High Blanks

     High blanks may occur for any  parameter  or analyte, but probably more
frequently for POHCs.  Blank levels can be used to "correct" sample levels,
but this  is  not possible if blank levels are  near  or exceed  sample levels.
If both the blanks and  sample levels  are  high, causing the uncorrected
sample value to  yield a DRE below 99.99%, no useful  information  is obtained
for that  sample.  Every precaution must be taken in the laboratory and in
the field to prevent contamination of samples.

7.  Poor Precision

     The  QA  protocol usually  requires triplicate analysis  of critical  sam-
ples  from at least one  run.   Wide variability in these  triplicate  analysis
results may  occur  if samples were not homogeneous,  if the samples contain
some  interfering component,  or  if other problems occur  with the  analytical.
technique.   In any case,  the precision obtained provides an indication of
the possible variability in results reported for each sample  and  analyte.
For  POHCs,  knowledge of this variability  may be quite  important  if the
calculated DREs  are  close  to  99.99% (e.g., 99.989%).
                                     47

-------
 I.  Recovery Efficiency

     Another  normal  part of the QA/QC  is  spiking of samples with  known
 unounts of POHCs and/or surrogates to determine recovery efficiency for the
 malysis method.  The  desired recovery range is  normally  70  to  130%,  but
 ictual results may  in some cases be much  lower or much higher for  any of
 several reasons.  Sample  results usually are not corrected based on  re-
 covery efficiency results,  but  knowledge of  the recovery efficiency is im-
 portant for the same reason as knowledge of the precision.   Also, poor pre-
 cision or poor  recovery efficiency may indicate  a need to reanalyze  the
 samples.

 9.  Actual Versus Expected Results

     In certain cases the. project leader has some idea of what the analyti-
 cal results  should be.  For example, the amount of POHC added into  a waste
 feed tank might  be  known,  so there is  some  expected concentration  of  the
 POHC in those samples.  Usually the analytical results are in good agreement
with expected results,  but in some instances the analytical  results  may
 disagree with the expected value.  The  analyses and  calculations then  must
 be  rechecked,  and the  QA/QC data scrutinized more  closely for  a clue.
Mathematical errors  are the most common cause of such problems, but other
possible causes, such  as  a poorly mixed feed tank,  cannot be overlooked.

D.  HOW ARE TEE SAMPLING DATA AMD ANALYSIS DATA CONVERTED TO FINAL
     -RESULTS?

     This section addresses the calculation methods used to convert labora-
 tory data on organics, and field data on flow rates,  into DRE numbers.  The
 values necessary to  calculate DREs,  and how they are obtained,  are listed
 in  Table 13.  A  brief review of the method  used  to  calculate DRE is pre-
 sented at the end of this section.  First, however,  it  is necessary to give
 attention is  given  to the areas of blank correction, significant figures,
 rounding of  DREs,  and the need  to use  "<" and ">" signs in reporting  DRE
 data.

 1.  Blank Correction

     Because  achievement  of 99.99% DRE often results  in stack  concentra-
 tions that  are  at or below  ambient or  laboratory levels for POHCs, con-
 tamination  of samples  can be a  significant problem.  The purpose of blank
 correction  procedures  is  to account for any portion of the sample  results
 that represent contamination, or something other  than the value  intended  to
 be  measured  (e.g., stack emissions).

     The underlying  philosophy of the procedure  is  based  on  a  paper pre-
 pared by the American Chemical Society  Committee  on  Environmental Improve-
 ment (ACS)7 and on experience  in  conducting and  interpreting trial burn
 data.  The ACS paper assumes  that blank values are random  samples that vary
 because of  preparation, handling,  and  analysis  activities. Under this as-
 sumption, blank  values can be  treated  statistically.   The "best estimate"
 for the blank for any particular sample is the mean  of  the available blanks.


                                    48

-------
                             TABLE  13.  DATA NECESSARY FOR CALCULATING ORE
      Measured value
Example
 units
              How value is obtained
Mass flow rate of feed


Volumetric flow of feed

Density of feed

Concentration of POHC in  feed

Total quantity of POHC in
stack sample

Volume sampled of stack gas
Total quantity of POHC  in
blank samples

Volumetric stack flow rate
g/nin


L/min

8/«L

M8/8

Mg


Nm3
MB


Nma/nin
Measured during  test or calculated from flow and
density.

Measured during  test.

Density analysis from lab.

•Analysis of waste feed samples

Reported by lab  for each sample taken during test.
For VOST  this  is  found in the sample train data.
For M5  this  is  found with the MS train data.
For gas bags this is reported as the volume analyzed
by the  laboratory.

Analysis  of  "blanks"
Reported  as  result of pitot traverse with M5 train
Note:  1 |Jg = 10 6  g
       1 ng = 10"9  g
       Nm3 = normal  cubic meters =  dry  standard  cubic meters

-------
The  ACS  procedure also enables determination  of  whether  a  sample is  "dif-
ferent from"  the  blank. 'If the sample value is not  significantly different
from the blank value, a sample cannot  be blank  corrected.   Even so,  the
measured sample value does provide  an  upper  bound for the emission value
.and  may  still provide sufficient  information  for determining if  the  re-
quired ORE of 99.99%  was met.

      The blank correction procedure applies mainly  to  stack  emission  sam-
ples and consists of  the following:

      a.   Assemble data for each POHC from all  of  the field  and trip blanks.
An example of such data for VOST might be:

                               Run 1          Run 2         Run  3

      POHC A     Field  blank    0.008  Mg       < 0.002 Mg    0.004 Mg
                Trip blank     0.005  Mg         0.004 |jg    0.003 Mg
     b.  Determine  whether or not the field blanks are -statistically  dif-
ferent  from the trip blanks by using the paired t-test  (consult  a  statis-
tics text) .

     If  the field blanks are significantly different than the  trip blanks
use the  field blank  data only.  If the blanks are not significantly differ-
ent use  all of  the blank values.

     c.  Calculate  the average and  standard deviation of the blanks  (many
calculators have statistics functions which allow you to do this easily) .

     d.  Determine  whether or not each measured sample  value is "different
from" the blank value by using the following test for each sample:
     s - sample value
     b = (blank average) + 3  (std  deviation  of blanks)
     If s is  greater than b,  then  the  sample is  "different"  from
       the blanks.

     e.  If  the measured sample value is different from the  blank value
 then the blank correction procedure  is applied:

     Blank corrected emission value  (Mg)  = measured  sample value
       (Mg) - average blank value  (Mg)

   ..; f .  If  measured sample  value is  not different from blank value then
 the measured  sample value is  used as an upper bound emission value and the
 emission rate is  considered less than  or  equal to  the measured value.   This
 results  in  the reporting of  emission  concentration  and  mass emission  rate
 with a  "<"  sign.   As a consequence,  DRE would be reported with a ">" sign.
                                     50

-------
2.  Significant Figures and DRE

     DRE is usually  reported with one or  two  significant  figures  depending
on the  accuracy  of the measured values which go  into the calculation of
DRE.  It is  important to note that a reported DRE of 99.99 or 99.999% has
only one significant figure.  The reason  for  this  is  that what  is actually
being measured is  the penetration, which  is the amount of a  compound which
is not destroyed.  That is:

     DRE = 100% - Penetration

For a DRE  of 99.99%, the penetration is  0.01% (one significant  figure).
For a DRE of 99.9916%, the penetration is 0.0084%  (two significant figures).

     The DRE is reported with the same number of significant figures  as the
least accurately measured value  used in the calculations.  The  controlling
measurement that determines  the  number of significant figures  is usually
the stack concentration.  GC/MS methods can normally only report  concentra-
tions with one or  two significant figures.  This will result in a DRE with
the same number  of significant figures as reported concentrations, unless
another measured value (waste  feed concentration,  waste feed flow rate,  or
stack gas flow rate)  has fewer significant figures.

3.  Rounding Off DRE  Results

     The rules  on  this are stated in  the Guidance Manual for  Hazardous
Waste Incineration Permits:- ".  . .if  the  DRE was  99.9880 percent, it
could not  be rounded off to 99.99 percent."   In other words, your  cal-
culated value, after rounding  to the.proper number of significant figures,
must  equal  or exceed 99.99% to be acceptable.   (Note:   This same  rule
applies to rounding HC1 results to 99%.)

4.  Reporting DRE with a M<" or ">" Sign

     As mentioned in  the section on blank corrections, if the sample  is not
"different" from the  blank (greater than  the  average blank plus three stan-
dard deviations), then it cannot be blank corrected.  As  a consequence, the
DRE will be  reported with a ">"  sign.  This  reported ">" value will also
occur when the  POHC  in the sample is undetected (below detection limit of
the analysis  method).  But,  as  long as the DRE is >  99.99%, this is  not  a
problem.

     In cases where both the blanks and  samples  have high values,  a DRE
below 99.99% may be  preceded by  a ">" sign (i.e.,  > 99.96%).  Such a  number
is  useless in evaluating achievement of  99.99%.   Experience in using the
recommended  sampling methods and avoiding contamination  is  the only way to
minimize this possibility.

     Occasionally,  a sample  may saturate the GC/MS with  the POHC in ques-
tion.   This  will result in an emission rate  with a ">"  sign and a DRE with
a  "<"  sign.   If such a DRE  is below 99.99% the incinerator  clearly  fails.
If  it is  above  99.99% (i.e., < 99.9964%), the number is  useless.   To avoid
                                     51

-------
                            f '..•.-•••;
such problems,  alternate  sampling methods should be  used,  based on pre-
liminary estimates of the stack concentrations that may exist.

     The conclusion  of  this  section is:  always  design  the sampling and
analysis so that  passage/failure  of the 99.99% criterion is determinable.
This can best  be  done by preliminary  estimates of  POHC  concentrations  in
the stack (assuming 99.99% DRE) and with selection of sampling and analysis
methods having appropriate upper and lower limits of detection.  Experience
in use of these methods to avoid contamination is also a key factor.

E.  HOW ARE THE DATA AND RESULTS USUALLY REPORTED?

     The results  should be  reported in a format which includes all infor-
mation and data necessary to calculate final results, presented in as clear
and succinct format  as  possible.  This will  include:  a  description of  the
operating system;  the  operating conditions  during the test;  the measured
quantities of  POHCs, HC1, and participate in  all samples;  and the cal-
culated results.   Example formats for presentation of these data are pre-
sented in Tables 14 through 25.  Using part of the data in these tables, an
example calculation of DRE is shown in Table 26.
                                    52

-------
         TABLE 14.  INCINERATOR OPERATING CONDITIONS3

Parameter
Organic waste flow rate,
kg/min (Ib/hr)
Aqueous waste flow rate,
kg/min (Ib/hr)
Heat input rate,
GJ/hr (106 Btu/hr)
Combustion chamber
temp., «C (°F)
Calculated residence
time, sec
Stack height, m (ft)
Stack exit velocity, m/s
(fpn>)
Stack temperature, °C (°F)
Run 1,
11/3/82
3.76 (497)
•
6.13 (811)
8.37 (7.93)
1053 (1925)
2-5
11.6 (38)
10.7 (2,110)
810 (1490)
Run 2,
11/4/82
4.01 (542)
5.38 (712)
.9.01 (8.54)
1066 (1950)
2.4
11.6 (38)
10.3 (2,030)
749 (1-380)
Run 3,
11/4/82
4.50 (595)
4.90 (648)
10.50 (9.96)
1094 (2000)
2.2
11.6 (38)
11.3 (2,230)
766 (1410)

Data collected by reading plant monitoring instruments at regular
intervals.  Values shown are averages for each run.

Determined by measuring storage tank liquid levels at start and
finish of each run.

Calculated front chamber volumes and stack flow rates.
                              53

-------
                            TABLE 15.  CONCENTRATIONS OF POHCs  IN  WASTE FEEDS (pg/g)
tn


Volatile POHCs
Carbon tetrachloride
Trichloroethylene
Benzene
Toluene
Semivolatile POHCs
Phenol
Naphthalene
Aqueous waste
Run 1 Run 2 Run 3
< 2a < 2 < 2
< 1 < 1 < 1
< 3 < 3 < 3
94 110 100
42,000 34,000 b
< 100 < 100 b
Organic waste
Run 1
6,400
5,900
2.7
1,800
4,200
510
Run 2
6,000
5,500
260
2,400
1,000
350
Run 3
4,700
4,300
140
1,900
b
b
•
            a  Results reported as less-than values represent limits of detection.

               MM5 sample voided for this run due to equipment problems.  Therefore, the waste  feed
               samples were not analyzed for semivolatile POHCs.

-------
     TABLE 16.  CALCULATED INPUT RATES FOR-POHCs
                  IN WASTE FEEDS
                                Input rates  (g/min)

Volatile POHCs
Carbon tetrachloride
Trichloroethylene
Benzene
Toluene
Semivolatile POHCs
Phenol
Naphthalene
Run 1

24
22
0.010
7.3

270
1.9
Run 2

25
22
1.1
10

180
1.4
Run 3

21
19
0.63
9.0

b
b

Combined input rates for both waste feeds.

Samples not analyzed for semivolatiles since MM5 sample
was voided for this run.
                        55

-------
TABLE 17.  CONCENTRATIONS OF VOLATILE POHCs BY VOST IN
             STACK EFFLUENT (Not Blank Corrected), ng/L


Carbon tetrachloride
Trichlorocthylene
Benzene
Toluene

Carbon tetrachloride
Trichloroethylene
Benzene
Toluene
Carbon tetrachloride
Trichloroethylene
Benzene
Toluene

1st
Pair
2.3
20
2.2
6.2

1st
Pair
2.3
17
2.0
21
1st
Pair
3.1
4.8
6.0
15

2nd
Pair
0.47
1.8
2.3
0.99

2nd
Pair
1.7
1.8
7.4
7.5
2nd
Pair
0.58
0.95
7.1
9.7
Run 1
3rd
Pair
0.57
1.6
2.2
2.1
Run 2
3rd
Pair
1.7
1.0
2.6
4.3
Run 3
3rd
Pair
0.45
0.66
6.2
5.7

Average
1.1
7.8
2.2
3.1

Average
1.9
6.6
4.0
11
Average
1.4
2.1
6.4
10
                            56

-------
        TABLE 18,  VOST BLANK CORRECTION VALUES
                             Average         Standard
                           blank value       deviation
                                (ag)             (ng)
POHCs
Carbon tetrachloride          < 2                0
Tricnloroethylene             < 1                0
Benzene                       < 3                0
Toluene                         3.7              1.8
            TABLE 19.  VOST SAMPLE VOLUMES
                         (Dry Standard
                         Liters)
     Run no.         Pair no.         Volume  (L)
1
1
1
2
2
2
3
3
3
1
3
6
1
3
6
. 1
3
6
18.4 .
18.1
17.5
18.4
18.5
18.5
18.9
18.9
19.0

        TABLE 20.  BLANK CORRECTION VALUES FOR
                     SEMIVOLATILE POHCs
                                    Blank correction
            Compound                   value  (|jg)
                                           3.4

                                           6.0


                            57

-------
       TABLE 21.  DESTRUCTION AND REMOVAL
                    EFFICIENCIES (DREs)

Volatile compounds
Carbon tetrachloride
Trichloroethylene
Benzene
Toluene >
Semi volatile
compounds
Phenol
Naphthalene

Run 1
99.99966
99.9975
a
99.9973
Run 1
99.9985 >
99.96
VOST
Run 2
99.99942
99.9977
99.972
99.9926
MM5
Run 2
99.99996 >
99.98

Run 3
99.99946
99.99906
99.914
99.9916
Run 3
99.9996
99.986

Waste feed concentration < 100 M8/8 ia this run.
                        58

-------
        TABLE 22.  MODIFIED METHOD 5 TEST DATA



Volume of gas sampled (Nm3, dry)
Sampling time (min)
Percent isokinetic
Moisture content (%)
Percent 02 (dry)
Percent C02 (dry)
Stack flow rate (actual m3/min)
Stack temperature (°C)
Stack flow rate (Nm3/min, dry)
Run 1,
11/3/82
2.277
140
95.0
15.8
10.5
7.8
355
809
73
Run 2,
11/4/82
2.101
140
96.8
13.0
10.8
.7.7
341
749
76
Particulate concentration
  (mg/dscm)                         842        523
  (gr/dscf)                           0.367      0.228
  (mg/dscm corrected to-7* 02)    1,125        719
  (gr/dscf corrected to 7% 02)        0.491      0.313
  (Ib/or)                             8.1        5.2

Chloride emissions
  (g/min)   .                         31.1       37.1
  (Ib/hr)                             4.1        4.9
Note:  Run 3 was voided by a broken probe liner.
                          59

-------
                  f*-.-
       TABLE 23.  CONTINUOUS MONITORING DATA*
                                 Run  1,      Run  2,
                                11/3/82.     11/4/82
Oxygen (1)
  Range                         7.1-11.0     8.3-11.0
  Average                         9.4           10.5

Carbon dioxide (I)
  Range                         7.2-10.6     7.2-8.1
  Average                         8.5           7.6

Carbon monoxide (ppm )
  Range             v           < 1-5.8      <  1-5.3
  Average                         1.4           1.8

Total hydrocarbons (ppm )
  Range                V          < 1           <  1
  Average                         < 1           <  1
   Concentrations on dry gas basis.

   Total EC reported as propane.
                         60

-------
      TABLE 24.   GENERAL ANALYSIS OF AQUEOUS WASTE

Parameter
Heating value, kJ/kg
(Btu/lb)
% Chlorides
% Water
I Ash
Saybolt viscosity
(sec)
Run 1
1,800
780
0.39
88.54
0.70
28.9
Run 2
1,720
730
0.36
93.89 *
0.78
28.2
Run 3
1,550
660
0.26
89.67
0.74
29.5

      TABLE 25.   GENERAL ANALYSIS OF ORGANIC WASTE

Parameter
Heating value, kJ/kg'
(Btu/lb)
% Chlorides
% Water
% Ash
Saybolt viscosity
Run 1
34,140
14,680
1.03
2.15
1.53
32.1
Run 2
34,390
14,800
1.26
3.18
2.13
30.1
Run 3
37,110
15,960
0.72
5.73
2.36
30.3
(sec)
                           61

-------
                        TABLE 26.  EXAMPLE DRE CALCULATION
     The following is a sample calculation showing the method  used to  con-

vert the analytical results to DREs for trichloroethylene  in Run 2 using

the VOST sample.


           W.    W
              Win


DETERMINE INPUT RATE (W. )
                       in


     W   s (organic waste flow rate x TCE concentration) +

                   Table 14                Table  15


           (aqueous waste flow rate x TCE concentration)

                    Table 14               Table  15




     W   = (4,010 g/min) (5,500 (jg/g) *  (5,380 g/min)  (< 1 Hg/g) =

           22 .x 101 Mg/nin = 22 g/min (Table 16)
CALCULATE OUTPUT RATE (W  J
                        out


     Stack .flow rate = 76 Nm3/min (Table 22)




     VOST concentration . avg =      1.8 + 1.  _ 7 g ng/L  (aot ^1^ corrected)



                   (Concentration values taken from Table  17)




     Blank correction



          VOST »/ l n8/s**Ple (Table W * < O.Q5 qg/L

                 18.5 L/ sample (Table 19)



          Blank corrected value =7.8 ng/L - < 0.05 ng/L « < 7.8  ng/L = < 7.8



     VOST output rate



          Mass flow = (< 7.8 Mg/m3)  (76 Nm3/min) (1 x  10"8 g/pg)



                    = < 0.00059 g/min (corrected)



CALCULATED DRE



         - 22 g/min - < 0.00059 g/min    n
         —         __  . .            X 1UU
                   22 g/min



         = > 99.9973% (Table 21)
                                         62

-------
                                 SECTION V

                                REFERENCES
 1.   U.S. Environmental Protection Agency/Office of Solid Waste, Washington,
     D.C.   Guidance Manual for Hazardous Waste Incinerator Permits, March
     1983.

 2.   U.S. Environmental Protection Agency/Office of Solid Waste, Washington,
     D.C.  Test Methods for Evaluating Solid Waste - Physical/Chemical Meth-
     ods.  SW-846  (1980),  SW-846 Revision A (August 1980), and SW-846 Revi-
     sion B,  July  1981.

 3.   Harris,  J.,  D. Lars en, C. Rechsteiner, and K.  Thurn.  Sampling and
     Analysis Methods  for Hazardous  Waste Combustion,  First Edition.   Pre-
     pared  for U.S. Environmental Protection Agency,  Contract  No.  68-02-
     3211 (124) by Arthur  D. Little,  Inc., December 1983.

 4.   Methods  for  Chemical Analysis  of Water and Wastes.   EFA-600/4-79-020,
     U.S. Environmental Protection Agency, March 1979.

 5.   Federal  Register, Volume  42, No.  160, August 18,.  1977.

 6.   U.S. Environmental Protection Agency/Industrial Environmental  Research
     Laboratory.   Protocol for the Collection and Analysis of Volatile POHCs
     Using VOST.   EPA-600/8-84-007, March 1984.

.7.   Guidelines for Data Acquisition  and Data Quality  Evaluation in Environ-
     mental  Chemistry.   Analytical Chemistry,  52(14):2242-2249,  December
     1980.
                                     63

-------
                                   TECHNICAL REPORT DATA
                           (fleur rod Imunrtions on t/it went btfore camplttint/
                                                                     ACC6SSIOWNO.
i TITLE
PRACTICAL GUIDE - TRIAL BURNS  FOR  HAZARDOUS
WASTE  INCINERATORS
                                                          I. (WONT DATC
                                                            November  1985
                                                          •. PIRP ORMING ORGANIZATION CODE
7 AWTMORIS)
 P. Gorman, R.  Hathaway, D. Wallace,
 and A. "Trenholm
                                                          • . Pf RP ORMING ORGANIZATION ME'QRT NO
                                                           8034  -  L
 . PERFORMING ORGANIZATION NAMC ANO AOORISS
 Midwest Research  Institute
 425 VoUer Boulevard
 Kansas City,  Missouri  64110
                                                          10. PROGRAM iLfMENT NO.
                                                           ABRD1A
                                                            68-03-3149
12. SPONSORING AGENCY NAMI ANO AOORf SS
 U.S. Environmental  Protection Agency
 Office of  Research  and Development
 Hazardous  Waste  Engineering Laboratory
 Cincinnati.  Ohio  45268
                                                          13. TYPE OP RtPORT ANO PERIOD COv£«EO
                                                           Research.  Final  1984-1985
                                                          14. SPONSORING AGENCY COOC

                                                           EPA/600/12
18. SUPPLEMENTARY NOTIS
        The manual  concentrates on those  aspects of a trial burn that  are  the
   most important and those that are potentially troublesome.  The manual  contains
   practical  explanations based on experience of Midwest Research Institute (MRI)
   and others  1n conducting trial burns and  related tests for EPA.   It  includes
   the comments of  several Industrial  plant  owners and operators.  It  Is directed
   mainly to  Incinerator operators, those who may conduct the actual sampling and
   analysis,  and those who must Interpret trial  burn results.  It will  also be
   useful fo'r  regulatory personnel and others that need to understand  trial  burns.
   Potential  trouble spots that have been encountered are:  (1) trial  burns
   frequently  take  more time and effort than an  operator anticipates;   and
   (2) failure to meet the trial burn  requirements.
                               KIV WORM ANO DOCUMINT ANALYSIS
                 DESCRIPTORS
                                            b.lOENTIPIERS/OPf N SNOIOTINMS
c. COSATi Field Croup
I. DISTRIBUTION STATIMENT

    RELEASE TO  PUBLIC
                                            IS. SECURITY CLASS (Ttti* *ifO*l
                                               UNCLASSIFIED
71. NO. Of PAGES
  •  73
                                             M. S8CURITY CLASS (TKu ftft/
                                               UNCLASSIFIED
                                                                       32. PRICE

-------