Office of Water &
           Waste Management
           Washington D.C. 20460
SW-719
September 1978
An Evaluation of the
Resource Recovery
Demonstration Project,
Baltimore,  Maryland

Executive Summary

-------
AN EVALUATION OF THE RESOURCE RECOVERY DEMONSTRATION PROJECT,

                     BALTIMORE, MARYLAND


                      Executive Summary
              This report fSW-719) was prepared
        under contract for the Office of Solid Waste
          by A.  J. Helmstetter and R. A. Haverland
            U.S. ENVIRONMENTAL PROTECTION AGENCY

                            1978

-------
     This report has been reviewed by the U.S. Environmental Protection
Agency and approved for publication.  Its publication does not signify
that the contents necessarily reflect the views and policies of the
U.S. Environmental Protection Agency, nor does mention of commercial
products constitute endorsement or recommendation for use by the
U.S. Government.

     An environmental protection publication (SW-719) in the solid waste
management series.

-------
                                    PREFACE
     This report is a complete technical, economic, and environmental evalu-
ation of the Landgard® Resource Recovery Demonstration Plant at Baltimore,
Maryland.  It was prepared for EPA by A. J. Helmstetter and R. A. Haverland of
Systems Technology Corporation.  Because of its bulk the report is presented
in four volumes:  an Executive Summary, the operational evaluation, an analysis
of problems, and the appendices.  Intended particularly for resource recovery
planners and administrators, the Executive Summary briefly describes the
Baltimore application of the Landgard® concept for the processing of mixed
municipal solid waste.  In addition, it presents an introductory problem
analysis of most of the major innovations that proved ineffective, caused
serious shutdowns, and required redesign or abandonment.  The second and third
volumes are detailed in-depth accounts of the evaluation, prepared primarily
for the designer.  Of the four volumes, only the Executive Summary has been
prepared for wide distribution.  The second, third, and fourth volumes are
available through the NTIS, Springfield, Virginia  22161.
                                      Ill

-------

•". •


       BALTIMORE LANDGARD© FACILITY

-------
                                   CONTENTS
Preface	ill

Figures	    VI

Tables	   VI1

Unit Conversions	   Vlii

Acknowledgment  	    1x


          Introduction  	     1

          Background  	     3

          Plant Evaluation	     4
               Process Description  	  .....     4
               Plant Performance	    12
               Plant Modifications	    18
               Cost Evaluation	    19

          Specific Problem Areas  .	    24
               Scaling from Prototype to Large-
                 Scale Unit	    24
               Variation from the Design of
                 Proven System  	    25
               Designing for Heterogeneous
                 Municipal Solid Waste  	    25
               Program Management	    26

          Summary	    27

-------
                                    FIGURES
Number
   1      A Plan View of the Baltimore Landgard® Facility ........    5


   2      Process Flow Diagram of the Baltimore Landgard®
            Facility  	    6
          Schematic of a Hammermill Shredder
   4      Schematic of the Storage and Recovery Unit	    8


   5      Schematic of the Ram Feeders	    9


   6      Schematic of the Kiln	10


   7      Schematic of the Gas Purifier	11


   8      Plot of Cumulative Refuse Processed Over Time	14


   9      Process Heat and Material Balance (SI Units)   	   16


  10      Process Heat and Material Balance (English Units)  	   17

-------
                                    TABLES
Number                                                                    Page






  1       Chronology of Landgard® .	   2






  2       Demonstration Project Financing (to February 1977)  	  19






  3       Scenario Operating Parameters 	  21






  4       Projected Cost Summary  	  23

-------
                                       LIST OF UNIT CONVERSIONS
Description
SI
English Equivalents

Length
Area
Volume
Mass
Pressure
Temperature
Energy
Density
Energy /Mass
Mass Loading
Concentration
Unit
meter
centimeter
millimeter
micrometer
square meter
cubic meter
liter
kilogram
megagrams
kilopascal
Celsius
joule




Symbol
(m)
(cm)
(mm)
(Vim)
(m2)
(m3)
(1)
(kg)
(Mg)
(kTa)
(C)
(J)
(kg/m3)
(MJ/kg)
(g/DSCM)
(yi/D
Unit
3.28 feet
0.394 inches
0.039 inches
1.0 micron
1.76 square feet
35.31 cubic feet
0.264 gallons
2.20 pounds
1.10 tons
0.145 pounds per
square inch
5 Fahrenheit/9-17.8
9.48 x 10"1*
.0624
431
0.437
1.0
Symbol
(ft)
(in)
(in)
(y)
(ft2)
(ft3)
(gal)
(Ibs.)
(ton)
(Ibs. /in2)
(F)
(Btu)
(Ibs. /ft3)
(Btu/lb.)
(gr/DSCF)
(ppm)

-------
                                ACKNOWLEDGEMENT
     This evaluation program was performed under EPA Contract No. 68-01-4359,
"Technical and Economic Evaluation of the EPA Demonstration Resource Recovery
Project in Baltimore, Maryland."

     The EPA Project Officer was David B. Sussman of the Office of Solid
Waste, Washington, D.C.

     Testing was carried out at the demonstration facility in Baltimore,
Maryland, with the cooperation of the city plant staff and the Monsanto on-
site engineering staff.  The economic evaluation was performed in conjunction
with Arthur Young & Company.  The contribution of both of these groups has
been greatly appreciated.  The contribution of Dr. H. G. Rigo along with other
staff members is also acknowledged.

     Systems Technology Corporation would like to express its gratitude to the
above-named individuals and all others associated with this evaluation.
                                       ix

-------
                                   SECTION 1

                               EXECUTIVE SUMMARY
INTRODUCTION

     The demonstration of the Landgard® resource recovery concept in Baltimore,
Maryland was one of the first attempts in this country at large scale resource
recovery from mixed municipal solid waste.  Partially funded by grants from
the U.S. Environmental Protection Agency  (EPA) and the Maryland Environmental
Services, the plant was designed by Monsanto EnviroChem for the City of
Baltimore to demonstrate the feasibility of using pyrolysis as an integral
phase in the recovery of energy, glassy aggregate, and ferrous metals from a
mixed municipal solid waste stream.  The plant was designed to receive, shred,
and pyrolyze 907 megagrams (1000 tons) per day of unsorted residential waste,
recover the usable residues and energy from the processed waste, reduce the
volume of material requiring ultimate disposal, and to perform these operations
within all environmental standards.

     To evaluate the operational experience of the Baltimore plant, the EPA
commissioned SYSTECH to analyze the technical, environmental, and economic
aspects of the plant.  Consequently, a four-volume report was prepared by
SYSTECH, which is intended to provide information that will assist others
involved in implementing resource recovery plans.  The four-volume report is
available from the National Technical Information Service.

     The single most important fact that must be understood concerning the
Baltimore Landgard plant is that it was a demonstration of the application of
a technology in a novel configuration, and that technical modifications in
such a situation are not unusual in order to achieve effective operation.
However, in the case of the Landgard demonstration, the modifications were
extensive, costly, and satisfactory results were at times elusive.  Neverthe-
less, contrary to common perceptions, the project has in many ways been a
success.  Not only did it provide invaluable experience in the use of innova-
tive equipment combinations and techniques, but it also ultimately proved the
feasibility of using the Landgard concept (though not the original design) to
process mixed municipal solid waste.

     Of major importance is the fact that the major processing component, the
rotary processing kiln, has been demonstrated to be an excellent primary
reaction vessel.  Numerous operational deficiencies and abnormalities such as
refractory failures, process instability, and control difficulties were
encountered in the kiln operation during the early part of the demonstration.
However, these problems have been effectively resolved by converting the kiln
from a pyrolytic reactor to a substoichiometric combustion reactor (starved-

-------
                       TABLE 1.   CHRONOLOGY OF LANDGARD®
Fall 1968

Spring 1969

Spring 1974

Nov. 1974
Process Development

- Bench scale prototype, Dayton, Ohio
  .3 to .6 TPD capacity
- Small scale prototype,
  35 TPD capacity
- Small scale prototype, Kobe City,
  35 TPD capacity
- Full scale prototype, Baltimore, Maryland
  1,000 TPD capacity
                         St. Louis, Missouri
                                    Japan
Sept. 8, 1972
Jan. 10, 1973
Jan. 31, 1975
Nov. 1, 1975

Dec. 31, 1975
Jan. 1, 1976
Nov. 5, 1976
Nov. 6, 1976
Jan. 31, 1977
Feb. 18, 1977
Baltimore Demonstration

- EPA grant awarded
- Contract approved
- Construction completed, plant commissioned
- Plant could not meet guarantees—modifications
    required
- Supplemental agreement for modifications signed
- Modification work begun
- Modifications completed
- Operation begun
- Monsanto recommends, project termination
- Monsanto personnel leave site, city continues
    with project

-------
air incinerator) , replacing the installed kiln refractory with a different
material and installation method, and modifying the method of air introduction
and the amount of heat supplied by auxiliary fuel for the process.  After the
remedial provisions were implemented, there has been no downtime attributed to
rotary kiln failure, and the long-term reliability of this component will be
determined with continued operation.
     The secondary processing equipment, consisting of the gdl> puAx^-te/t  (after-
burner) and the pollution control device, is currently being redesigned  to
perform at acceptable levels of reliability.  The £}
-------
native plant configurations to exploit the potential of the Landgard concept,
as well as to profit by the experience gained thus far.

PLANT EVALUATION

     As prescribed by the EPA in the contractual award to SYSTECH for the
evaluation of the Baltimore plant, the following requirements have been met:

     1.  Description of the initial plant processing configuration;
     2.  Determination of the process balance and efficiency;
     3.  Assessment of the effects of the process on the environment;
     4.  Summary of the proposed and implemented modifications to improve the
         process performance and reliability;
     5.  Detailing of the proposed and implemented modifications to improve
         the process performance and reliability;
     6.  Assessment of the current cost to process waste with the Landgard
         process and the projected cost when all modifications will be com-
         pleted, with the assumption that the modifications will produce the
         desired Improvements.

Process Description

     The Landgard process is based and centered on the pyrolysis of shredded
municipal solid waste and the subsequent JJl A^itu. combustion of the pyrolysis
products.  Figures 1 and 2 show the basic layout of the Baltimore Landgard
Plant where the principal components are a rotary kiln, the primary reaction
vessel, and a Q(U> pafu.^eA., the secondary reaction vessel.  Figures 3 through
7 show details of the major equipment components.

     After the incoming waste is weighed, deposited, and shredded* (figure
3) it is fed at a controlled rate to hydraulic ram feeders (figure 5) which
in turn feed the shredded waste to the rotary kiln (figure 6).  Since the kiln
operates at a negative pressure (minus one inch of water), the ram feeders
extrude the waste through passage-restricted cylindrical tubes to maintain the
kiln air seal.  As the waste tumbles down the declined rotating kiln, it
undergoes various thermal processes and reactions.  While the partially com-
busted gases, called the kiln-off gases, exit at the feed end of the kiln to
flow to the goA pu/u.£t&t. (figure 7), the combustion residue, consisting of
ash, other inert solids, and some unburned carbon char, is discharged at the
other end of the kiln to fall into a water quench tank.

     The thermal processing begins when the shredded waste is extruded into
the hot rotary kiln.  As the waste continuously tumbles down the kiln, it is
dried and its volatile content is vaporized to form a gas mixture containing
carbon monoxide, hydrogen, methane, and other hydrocarbons, along with inert
dilutents.  This gas is partially combusted in the kiln, with the heat release
limited to the exact level needed to sustain the endothermic pyrolysis
   Originally it was then discharged into a storage and recovery silo
   (figure 4).

-------
                                               DRAINAGE CHANNEL
Figure 1.   A plan view  of the Baltimore Landgard® facility.

-------
                                                                         MAGNETICS
"D BURNERS

   COMBUSTION AIR
                                                GASES        SOLIDS
                                              -*•	KILN
                          SPILLAGE
                          &SLAG

                               BURNERS [T
                                                                                            1
                                                                                      DEHUMIDIFIER

                                                                                             \
                                                                                      EXHAUST TO
                                                                                      ATMOSHPERE
Figure  2.   Process flow diagram of  the Baltimore Landgard® facility.

-------
 VERTICAL FEED CHUTE AND
 EXPLOSION RELIEF DUCT
FENWALL
EXPLOSION
SUPPRESSION
BOTTLE
FENWALL
EXPLOSION
SENSOR
                                       ROTOR SIDE PLATE
                                                         REMOVABLE ACCESS DOOR
                                                CUTAWAY OF DISCHARGE GRATE
     SHREDDER DISCHARGE CONVEYOR
               Figure 3.  Schematic of a hammermill  shredder.

                                        7

-------
               STORED MATERIAL SPREADER
oc
                                                                    CENTER CONE
           SWEEP PULL RING
          DRAG BUCKET CHAIN
                             DISCHARGE CONVEYOR TROUGH
SHREDDED
REFUSE
TRANSFER
CONVEYOR
GALLERY
                                                                              KILN FEED CONVEYOR GALLERY
                           Figure  4.   Schematic of  the storage and  recovery unit.

-------
  SIGHT FOR
PNEUMATICALLY
OPERATED
DIVERTER
GATE
                                                                                                      FEED CHUTE
                                                                                                      FLOW SPLITTER
     LARGE RAM
                               Figure 5.   Schematic  of the  ram  feeders.

-------
                                             KILN FLIGHTS

                                     RAM SNOUTS
                 EMERGENCY STACK
 CROSSOVER DUCT
KILN SPIKES
 9" CASTABLE
 REFRACTORY
                                                COMBUSTION AIR
                                                BUSTLE
                                                                                 REFUSE COMBUSTION AIR
                                                                                    FAN INLET
                                                                 KILN LEAD BURNER AND COMBUSTION FAN INLET
                                                                                    KILN HEAT-UP BURNER AND
                                                                                    COMBUSTION FAN INLET

                                                                                        SIGHT PORT
                                                                           OPTICAL PYROMETER
                                                              ACCESS DOOR
                               Figure  6.   Schematic  of the kiln.

-------
                            GAS/PURIFIER PILOT BURNER
                                                                 CROSSOVER DUCT
      GAS PURIFIER START-UP BURNER
      AND FAN INLET
        9" BRICK
        REFACTORY
THERMOWELL
  SLAG HOLE
  BAFFLE WALL
  DUST
COLLECTION
  FAN  INLET
         CROSSOVER
         COMBUSTION
         AIR FAN INLET
 SLOTTED QUENCH
 AIR DAMPER
                                                 BUTTERFLY VALVE
                                                 QUENCH AIR DAMPER
                 Figure 7.   Schematic  of the gas  purifier.
                                       ] 1

-------
reaction.  The combustion takes place under substoichiometric conditions which
are maintained by carefully controlling the airflow into the kiln.  When
approximately 40 percent of the theoretical combustion air is supplied, the
thermal reactions become self-sustaining and therefore do not need the supple-
mental heat from the kiln burners.  Since oxidation occurs in the thermal
reaction zone, the kiln is not a pure pyrolytic reactor, as originally conceived,
but rather a substoichiometric combustion reactor like a starved-air incinerator.

     As the kiln-off gases are drawn out of the kiln feed hood by the induced
draft fan, the kiln-off gas enters the g
     The pyrolysis gas mixture generated and partially burned in the kiln has
a fuel value of 816 J/SCM (90 Btu/SCF) and is completely combusted in the g&S>
puSiifaeJi.  While the gas stream exiting the gcU> puS^i^e/i has a relatively low
particulate content, the amount of particulate is a function of the quantity
and size of the particles in the incoming kiln-off gas.

     Heat recovery from the gases exiting the gas purifier takes place in two
waste heat boilers downstream from the go* puJvi&eA.  These boilers have a
heat recovery efficiency slightly above 80 percent.  Finally, the flue gases
are drawn by the induced draft fan to the emission control device, originally
a wet scrubber which is being replaced by a dry electrostatic precipitator,
and are subsequently discharged to the atmosphere.

     The original design called for a materials recovery system where the
solid residue exiting the kiln would be separated into magnetic metal, glass,
and char fractions.  This separation step used a water flotation tank to
separate the heavies (metal and glass) from the lights (char and ash) and a
magnet to separate the magnetic material from the glass.

     However, the operation of the entire materials recovery system was dis-
continued because of the lack of sufficient manpower to adequately operate the
materials recovery equipment, the lack of financing needed to improve the
process reliability, and the insufficient financial return to justify the
additional manpower and investment.  Consequently, all solid residue is
currently removed for landfill disposal.

Plant Performance

     The Baltimore plant has not operated as designed.  In addition to
numerous engineering, operating, and developmental problems, the plant has
failed to meet two of the standards specified in the agreement between the
City of Baltimore and Monsanto.  First, particulate loadings in the discharged
stack gases were prescribed to be less than the Maryland standard of 0.07
grams per standard cubic meter (g/SCM) or 0.03 grains per standard cubic foot
(gr/SCF) but have been about .28 g/SCM (.12 gr/SCF).  Second, while the plant
was guaranteed to have an 80 percent operational availability during a 60-day
                                      12

-------
test run, the longest run thus far was only 25 days because of miscellaneous
equipment failures.

     Among the operational and equipment difficulties which adversely affected
the plant reliability were kiln temperature control, residue in the kiln
fusing into large slag agglomerations, refractory failures, fan and pump
breakdowns, ram feeder jamming, slag tap hole plugging, and material handling
and retrieval malfunctions.  As a result of the foregoing, the cost of the
facility increased and the plant demonstration sustained major delays.  Many
of the plant shortcomings have been due to the 30 times scaling from the 32 Mg
per day (35 tpd) capacity of the prototype unit to the 907 Mg per day (1000
tpd) capacity of the Baltimore plant.

     Moreover, the structure of the project administration during the demon-
stration hampered plant operations and thereby contributed to limiting the
percentage of on-stream time of the plant.  Under the original contract agree-
ment Monsanto provided turnkey service during the plant demonstration, using
city maintenance and operating personnel during the start-up and shakedown
phases of the project.  Not having the plant staff directly responsible to the
contractor had the expected tendency to cause confusion regarding who should
direct the efforts of the plant staff; the startup personnel, or City manage-
ment personnel.  During the supplemental agreement, Monsanto's role was
changed to that of a contractor to the City not directly involved in plant
operation.  However, technical recommendations were made by Monsanto concern-
ing plant operation during this period.

     In addition to the uncertainties relating to the 30 times scale-up, the
system designer had very little empirical information available on large-scale
shredding, conveying, storing, and transporting of municipal solid waste.
Consequently, the design for the material handling equipment had to be based
primarily on theoretical estimates which proved to be overly optimistic.

     The Baltimore plant was designed to handle a feed rate of 636 kg per
minute (1,402 Ib per minute) and can operate at rates as low as 303 kg per
minute (668 Ib per minute).  While the plant has operated at the design rate,
the total waste throughput has been considerably less than the design capacity
because operational and equipment malfunctions have limited the total runtime
and the average operational feed rate to only 455 to 530 kg per minute (1000
to 1170 Ib per minute).  However, plant performance has improved substantially
subsequent to various modifications made to the plant by the City of Baltimore.
Figure 8 shows the increase in cumulative throughput with time.  In contrast
to the design processing rate of 281,000 Mg per year (310,000 tpy), the actual
annual processing rate has been only 65,000 Mg per year (71,000 tpy).

     The numerous operational and equipment difficulties and deficiencies
occurring during the demonstration have made it extremely difficult to char-
acterize the process.  Nevertheless, the process balances presented in this
report, which are based on the measurements taken by onsite SYSTECH personnel
from November 1976 through August 1977, are representative of the process
stream quantities that prevailed throughout the demonstration.  Of particular
importance to the balance presentation is the variation in the process flows
due to varying waste composition.  Typically in processing systems handling

                                      13

-------
 100,000





  90,000





  80,000





  70,000
60

5 60,000
n
w


w
o
o
rt
PM
50,000





40,000





30,000





20,000





10,000 L
                       1975
                                               1976
1977
              Figure  8.   Plot  of  cumulative refuse processed over time.

-------
mixed municipal solid waste, this variation makes efficient waste processing
very difficult.  Since municipal solid waste is not homogenous and has time-
varying characteristics, the processing equipment must be designed to accom-
modate these variations.

     Figures 9 and 10 show the energy and material balance for the major
process flows with inputs of one kilogram and one pound of feed, respectively.
Although specific flows are shown, the process flow for a fixed feed rate will
vary up to ±15 percent from the average process flow because of variations in
the content of the solid waste.  The energy and material balance was derived
from average field data.  All numbers in the balance were rounded off to
reflect the appropriate accuracy.  As is apparent in the balance, the kiln
process is not a pure pyrolytic reaction but rather a substoichiometric
combustion reaction zone.  The heat transfer equations in a mathematical
simulation of the kiln operation indicate that some j,n k'Ltui combustion of
kiln-off gas is necessary to adequately process the solid waste in the kiln.
(SYSTECH developed this model during its evaluation of the Baltimore plant to
study alternative modes of kiln operation.)

     The major material and energy inputs to the plant process are the solid
waste boiler waters, and the combustion air introduced to various components
in the processing system.  The primary outputs from the process are steam
generated by the waste heat boilers, flue gas discharged at the stack, and
solid residue removed for landfill disposal.  As derived from the energy
balances, the net energy efficiency of the entire plant (energy input in
waste/available energy output in steam) is about 40 percent.  The available
energy generated is the net energy supplied to the customer at the plant
perimeter after in-plant steam use and losses have been accounted for.  Energy
losses from the plant processing equipment are primarily due to skin heat
radiation and conduction and the heat lost in the flue gas discharge through
the stack.  Most of the energy in the electrical power supplied to the plant
is dissipated to the atmosphere.

     The solid, liquid, and gas emissions from the plant were evaluated in
terms of their pollution potential.  Although the plant is currently not
environmentally acceptable, it will be after the equipment modifications and
operational changes detailed in the following section have been made.

     The particulate emissions in the gas discharged from the waste heat
boilers are about 0.6 g/scm (0.26 gr/scf).  This relatively large amount of
emissions is due to the submicron particles emerging during the thermal pro-
cessing.  Although the low energy type wet scrubber employed at the prototype
unit and the Baltimore plant performed at design efficiencies in both facil-
ities, it could not reduce the emissions of the Baltimore plant to the
Maryland limit of 0.07 g/scm (0.03 gr/scf).  Since vendors have successfully
demonstrated in extensive on site testing that a dry electrostatic precipitator
can reduce the stack emissions to the prescribed limit, the wet scrubber is
being replaced with a dry electrostatic precipitator.

     The mass and volume of the incoming waste are reduced by 56 and 96
percent, respectively, which considerably reduces the landfill area required
for the residue disposal.  In addition to recovering energy and reducing the

                                      15

-------
                  SYSTEM LOSSES
                      2.0 MJ
       FEED
      1.0 Kg
      9.3 MJ
COMBUSTION
         AIR
      2.0 Kg

  FUEL OIL &
  ATOMIZING
      STEAM
      <-1 Kg
     -1.0 MJ
   PROCESSING KILN
OFF GAS
3.0 Kg
RESIDUE
   .3 Kg
  |1.5 MJ
   FUEL OIL &
^- ATOMIZING STEAM
   <-1 Kg
 -0.6 MJ
— COMBUSTION AIR
   2.3 Kg
 AFTERBURNER
{SLAG
  <-1 Kg
          AVAILABLE ENERGY
                PRODUCTION
                     5.4 MJ
         _ QUENCH AIR
            2.0 Kg

          FLUE GAS
          7.0 Kg
                                 BOILERS
                                EMISSION
                                CONTROL
                                 DEVICE
                      ATMOSPHERIC
                         DISCHARGE
                             7.0 Kg
                             2.0MJ
       Figure 9. Process heat and material balance (SI units)
                          16

-------
                   SYSTEM LOSSES
                      256 BTU
       FEED
     1.0 LB.
   4000 BTU
COMBUSTION
        AIR
      2.0 LB.

  FUEL OIL &
  ATOMIZING
     STEAM
     < .1 LB.
  -400 BTU
   PROCESSING KILN
OFF GAS
3.0 LB.
RESIDUE J
   .3 LB.
  640 BTU
  FUEL OIL &
 "ATOMIZING STEAM
   <.1 LB.
   ~171 BTU
—COMBUSTION AIR
  2.3 LB.
 AFTERBURNER
f SLAG
<.1 LB.
          AVAILABLE ENERGY
                PRODUCTION
                  2300 BTU
         _ QUENCH AIR
         ~ 2.0LB.

          FLUE GAS
          7.0 LB.
                                 BOILERS
                                 EMISSION
                                 CONTROL
                                  DEVICE
                       ATMOSPHERIC
                         DISCHARGE
                             7.0 LB.
                           850 BTU
     Figure 10. Process heat and material balance (English units),

                          17

-------
quantity of waste to be disposed of, the plant processing reduces the waste to
a relatively inert ash with a low putrescible content.  Although heavy metals
are among the solid residues, their solubility is reduced because of the
alkaline characteristics of any leachate evolved from the residue.  Wherever
disposed of, this residue will have far less detrimental effects on the
environment than unprocessed solid waste.

     Only relatively "weak" wastewater consisting of city water, boiler blow-
down, and water treatment discharge is presently discharged to the sewer.
Most processing waters are in closed-loop lines and are only occasionally
discharged to the ground when drain plugging or other equipment malfunctions
occur.  Such drainage should be treated before its discharge since it does not
meet Federal water discharge quality standards and therefore, is environ-
mentally unacceptable in the raw state.  In addition, rain runoff carrying
wastewater discharged to the ground will adversely effect the surface water
quality in the plant area.  However, if the processing water were discharged
to the sewage system, it would have a negligible effect on the quality of the
final effluent emitted from the wastewater treatment plant.

     Fugitive emissions from the plant consist of dust and carbon monoxide.
Although the effect of the microbes within the dust has not been determined,
the dust concentrations have not risen above the nuisance level.  During
working hours, some areas of the waste receiving building have carbon monoxide
concentrations greater than the threshold limit value (TLV) for continuous
exposure.  However, the building personnel have a limited exposure to these
concentrations.

     Noise throughout the plant is generally less than the OSHA level for an
8-hour exposure.  While there are areas where the noise exceeds that level,
the noise source varies and the duration of and exposure to the noise are
intermittent.

Plant Modifications

     Since the Baltimore plant has not operated as designed, it has been field
modified extensively.  The major modifications include the following:  (1)
replacing and upgrading the refractory linings in the kiln and ductwork;  (2)
installation of an explosion suppression system on the shredders;  (3)
redesign of the residue and slag handling conveyors;  (4)  introduction of a
controlled-air combustion operational mode in the kiln; and (5)  general
improvements in the materials handling equipment.  These modifications will be
detailed in Section 2 of Volume II.

     The cost to modify the Baltimore plant to date is summarized in the
following table as the costs incurred under the "Supplemental Contract."   As
shown in the table, a substantial contingency fund would have been required to
cover the field modification costs.

     In addition to the above-mentioned modifications which were performed
prior to Monsanto's termination, the following modifications are in progress:
(1) redesign of the gat, pu/vifceA;  (2)  replacement of the wet gas scrubber ' *
with a dry electrostatic precipitator;  (3) elimination of the residue


                                     18

-------
          TABLE 2.  DEMONSTATION PROJECT FINANCING (TO FEBRUARY 1977)
                                       Original                   Supplemental
      Source                           Contract                     Contract
Federal Grants                       $ 6,000,000                   $1,000,000

Md. Env. Service*                    $ 4,000,000

Baltimore City                       $ 5,532,000

Monsanto EnviroChem                                                $4,000,000

Subtotal                             $15,532,000                   $5,000,000

TOTAL                                             $20,532,000



*  No interest loan/grant.


separation module;  (4)  elimination of the storage and recovery unit; and
(5)  addition of a 220 foot stack.  Based on the experience gained during the
demonstration, these additional modifications being completed by the City of
Baltimore are intended to bring particulate emissions into compliance with
State regulations, to remedy deficiencies and malfunctions which caused
previous plant downtime, and to simplify the overall plant operation.

     While these modifications have a high probability of being successful, it
is likely that they will reduce somewhat the thermal efficiency of the plant.
However, if the modifications substantially increase the plant reliability,
the plant should prove to be an economical means of disposing of municipal
solid waste.  Insofar as financing is concerned in the plant development, it
must be noted that Baltimore has been in a more favorable financial position
than a nondemonstration plant investor because of the outside agency subsidies
that have defrayed the greater part of the total plant cost.

Cost Evaluation

     The cost evaluation was intended to determine the net cost of operating
the Baltimore plant in terms of the plant processing competitiveness with
alternative systems for municipal solid waste disposal.  Because of the
erratic plant operation during the evaluation period from November 1976
through June 1977, and the extensive equipment modifications with the conse-
quent extreme variation in the cost parameters, a cost evaluation was estab-
lished for each of three scenarios (table 3).  Arthur Young and Company, a
subcontractor to SYSTECH, performed the cost evaluation according to EPA
                                      19

-------
guidelines.*

     Of the three scenarios, the first reflects the plant operating conditions
during the 6 months from November 19, 1976,  to May 1977 the period during the
evaluation of maximum material throughput.t   The second and third scenarios
are based on the projected plant operations  when the proposed equipment
modifications and improved operating procedures will have been implemented.
The second scenario represents reasonably expected conditions, and the third
optimum conditions.  Consequently, the three scenarios provide the means for
making comparisons to assess the potential value of the plant and to determine
which plant changes can be justified from the standpoint of efficiency and
cost.

     Plant capital costs were determined from the City of Baltimore extract
listing and a Monsanto EnviroChem report§ detailing the pertinent expenditure
invoiced in the city Listing.  The capital costs were adjusted to discount the
one-time costs associated with the first-of-a-kind demonstration.  The adjusted
capital cost was $21,960,000.  Then, assuming financing by a Baltimore bond
issue, the cost was depreciated over a 20-year plant life to arrive at a
"nondemonstration" annual cost.

     Net operating and maintenance costs were computed by combining actual
operating data with the projected operating  schedule.  After deriving oper-
ating costs from actual field-measured quantities calculated on a per ton
basis, the costs were applied to the respective operating schedules and
throughput capacities to determine annual operating costs.  Similarly, steam
revenues were derived on an annual basis from operating data applied to each
schedule.  Because of the limited and sporadic on-stream time and consequent
varying maintenance requirements, 5.5 percent of the capital cost was used to
arrive at the annual maintenance costs.   This percentage was based on IRS
guidelines and the experience of similar processing plants.

     Then for each scenario (table 4), the annual capital, operating, and
maintenance costs were divided by the annual throughput rate of the plant to
yield the total operation costs on a dollar  per ton basis.  Given its perfor-
mance to date, the Baltimore plant cannot compete economically with a system
disposing of municipal solid waste by other  methods.  The current net
operating cost, including total depreciation and interest, for the Baltimore
plant exceeds $64 per Mg ($58 per ton) of waste input.  Because of Baltimore's
unique financial arrangement in the plant development, the actual cost to the
city is about $35 per ton.  However, since the proposed plant modifications
and better operating procedures will substantially increase the plant through-
put and correspondingly decrease the cost to process a ton of solid waste,
*  Sussman, D. B., Resource Recovery Plant Implementation Guides For Municipal
   Officals, Accounting Format.  EPA publication SW-157.6.
t  From February 1977 to January 1978, the plant disposed of 68,000 tons of
   solid waste and generated 263,000,000 pounds of steam.

§  Buss, T., Number 9 Report, Monsanto Corporation Unpublished Report.

                                      20

-------
                   TABLE 3.  SCENARIO OPERATING PARAMETERS*
                                 Scenario 1
                   Scenario 2
                 Scenario 3
Operating schedule:
Operating status (days):
  Normal processing, t
  Standby, §
  Heating and cooling, 11
  Downtime, **

Refuse feed rate:
  Mg/hr
  Mg/yr

Steam production (kg/hr):
  Normal processing

  Standby

Staffing:
  Plant manager
  Plant supervisor
  Clerk typist
  Chief operators
  Field operators
  Ram operators
  Equipment operators
  Laborers
  Scalemen
  Engineers
  Laborers/chauffeurs
  Maintenance supervisor
  Electricians
  Mechanics
  Welders
  Oilers
  Instrument technicians
    Total staffing

Fuel consumption (£/hr):
  No. 2 fuel:
  Normal processing
  Standby
  Heating and cooling
  Downtime
  24 hr/day,       24 hr/day,      24 hr/day,
 6 days/week,     7 days/week,    7 days/week,
24 shutdowns/yr  8 shutdowns/yr  4 shutdowns/yr
          104
           56
           48
          157
           27
       67,000
 50,000 (2/3)
 35,000 (1/3)
       35,000
            1
            1
            1
            5
            5
            3
            3
            5
            1

            7
            1
            2
            3
            2
           40
          660
        3,260
        1,960
            0
         264
          21
          16
          64
          32
     203,000
59,000 (2/3)
35,000 (1/3)
      35,000
           1
           1
           1
           4
          12

           4
           7
           1
           1
           5
           1
           2
           6
           1
           1
           1
          49
         660
       3,260
       1,960
           0
    312
     18
      8
     27
     36
270,000
 66,000

 35,000
      1
      1
      1
      4
      8

      4
      5
      1
      1
      4
      1
      1
      4
      1
      1
      1
     39
    660
  3,260
  1,960
      0
                                       21

-------
                             TABLE 3.   (Continued)

Gasoline:
Normal processing
Standby
Heating and cooling
Downtime
Diesel fuel:
Normal processing
Standby
Heating and cooling
Downtime
Electricity consumption (kw) :
Normal processing
Standby
Heating and cooling
Downtime
Water consumption (£/day) :
Normal processing
Standby, tt
Heating and cooling
Downtime
Sewer flow (£/day) :
Normal processing
Standby
Heating and cooling
Downtime
Scenario 1

620
620
93
93

208
208
64
64

2,100
1.109
1,109
142

1,595,380
1,195,780
187,780
187,780

395,380
355,780
187,780
187,780
Scenario 2

620
620
93
93

208
208
64
64

2,10.0.
1,109
1,109
142

1,835,140
1,195,780
187,780
187,780

419,140
355,780
187,780
187,780
Scenario 3

620
620
93
93

208
208
64
64

2,100
1,109
1,109
142

2,021,620
1,195,780
187,780
187,780

437,620
355,780
187,780
187,780

*  Provided by SYSTECH.




t  Constitutes processing of waate.




§  Constitutes onstream with no processing of waste.




1F  Involves startup and cool-down of kiln.




** No activity, plant shut down.




tt 35 Mg/hr of steam.
                                      22

-------
                       TABLE 4.  PROJECTED COST SUMMARY*

Scenario It

Capital costs
Interest
Operating and
maintenance costs
Total Costs
($/Mg)
$22.60
14.10
42.00
78.70
($/ton)
$20.50
12.80
38.10
71.40
Scenario 2§
($/Mg)
$ 7.70
4.80
14.60
27.10
($/ton)
$ 7.00
4.30
13.30
24.60
Scenario 3 IT
($/Mg)
$ 5.90
3.60
11.00
20.50
($/ton)
$ 5.30
3.30
10.00
18.60
Revenues                14.60    13.20      10.70     9.70      12.70    11.50






Net Cost                64.10    58.20      16.40    14.90       7.80     7.10







*  In 1977 dollars




t  Annual throughput is 67,000 Mg/yr (74,000 tpy)




§  Annual throughput is 203,000 Mg/yr (223,000 tpy)




II  Annual throughput is 207,000 Mg/yr (300,000 tpy)
                                      23

-------
reasonable projections indicate that the Baltimore plant can become econom-
ically viable.

SPECIFIC PROBLEM AREAS

     Since the Baltimore plant has not operated as designed, much of the plant
evaluation has been devoted to determining what works and what does not.  When
this review revealed that a system component or procedure definitely failed or
was ineffective, the next step was to analyze the reason for its failure or
deficiency.  All malfunctions, failures, and shutdowns during the demonstration
are detailed in the second volume of this report.  In volume IV of the eval-
uation, an in-depth analysis of some of the more critical design deficiencies
is presented.

     The following paragraphs of this section deal with some of the major
problems that caused plant shutdown or seriously interfered with efficient
operation.  These problems are grouped under four headings:  (1)  Scaling from
Prototype to Large-Scale unit,  (2)  Variation from the Design of Proven
Systems,  (3)  Designing for Heterogeneous Municipal Solid Waste, and  (4)
Project Management.  Among the specific subjects included are kiln performance,
gas purifier reliability, stack emissions control, and material handling.

Scaling from Prototype to Large-Scale Unit

     The following account of the rotary kiln performance and its effects on
the local and downstream equipment and operations highlights the criticalness
of engineering when scaling from a prototype to a large-scale unit.

     In retrospect, the basis for scaling was inappropriate since the combined
aerodynamic and thermodynamic factors changed during the scale-up.  As a
result, the kiln process was unstable, and the kiln temperatures have been
337C (630F) above the design level.  With the instability causing fluctuations
in process temperatures of 150C (300F) over a 20-minute period, it was virtually
impossible to control the process.  Consequently, the residue discharged from
the kiln ranged in quality from fused slag balls 1.2 meters (4 feet) in
diameter to slightly burned paper.  Such extremes in the residue quality
caused the discharge conveyor to be overloaded and the kiln to be shut down on
many occasions.

     In addition to the high and fluctuating process temperatures, faulty
refractory installation techniques and the differential thermal expansions in
the kiln shell caused severe refractory failures in the kiln.  However, a new
method for introducing air into the kiln,  upgraded refractory with a better
installation technique, vessel skin cooling, and expansion slots inserted at
the vessel ends have virtually eliminated the refractory failures.

     The vaporization of metals in the kiln is also attributed to some degree
to the excessive process temperatures.  In the gaA pu/t/CjJ/te/t., the metals in the
kiln-off gas reoxidize and condense, and most of the fine particles remain    *
entrained in the cooled gas discharged to the wet gas scrubber.  Then, since
the scrubber has not efficiently removed the very small metal particles as
well as larger particles from the flue gas, it cannot attain the emission

                                      24

-------
levels of the prototype scrubber.  In addition, since the gas scrubber is
upstream of the induced draft fan, the condensed acids in the flue gas have
corroded the fan blade.  Moreover, solids have so accumulated on the impellers
that the fan has vibrated excessively.  Consequently, the fan was shut down
weekly for the 3 to 5 days needed to sandblast and rebalance the fan.

     The scrubber-related problems have been remedied temporarily by radically
reducing the water flow to the scrubber, and will be solved permanently by
replacing the wet scrubber with a dry electrostatic precipitator .

     Conclusion:  Large magnitude scaling can be dangerous.  Perform sufficient
testing to validate scaling factors, and install sufficient excess capacity
and flexibility to compensate for variations from design.

Variation from the Design of Proven System
     The experience with the Qdi> pu/U-fc&i illustrates the importance of
identifying and analyzing the implications of any variation from the design of
a proven system.  In this instance, the design variation was from the non-
slagging mode of the prototype to the slagging mode of the goi puA/c^eA in the
Baltimore plant.  The following account of the problems with the gcM> puA/Cj
reveals the inadequate understanding and implementation of the provisions
needed for the gcU> puAx^-teA operational mode.
     During the demonstration, the plugging of the slag tap hole at the bottom
of the QOA pU/Lt^'CCA caused extensive downtime.  As the molten slag passed
through the hole, it was chilled by the water quench at the bottom of the
hole.  Then the chilled slag solidified and built up on the edges of the hole
in layers until the entire hole was plugged, shutting down the processing
system to allow clearing the hole which took 7 days on the average.  Conse-
quently, to maintain a continuous flow of slag through the hole, the temper-
ature in the g&!> puHifaeA. had to be increased to 1380C (2500F) which is 260C
(500F) above the design temperature.  But the increased temperature coupled
with the corrosiveness of the molten slag caused frequent refractory failures
in the 306
     Conclusion:  Extreme care must be exercised in making changes from the
tested design.  Anticipate problems and provide flexibility in the design in
areas where changes have been made.

Designing for Heterogeneous Municipal Solid Waste

     The lack of sufficient information and conservative design have caused
many problems and shutdowns in the material handling equipment.  The char-
acteristics of mixed municipal solid waste must be thoroughly researched
before initiating the design for the material handling equipment.  This type
of waste is extremely heterogeneous with time-varying characteristics of
particle size, bulk densities, and moisture content.  Consequently, the material
handling equipment must be overdesigned with provisions to anticipate the
variations in the waste characteristics.  At the Baltimore plant, much of the
material handling equipment had to be abandoned or required redesign as the
result of this condition.

                                      25

-------
     Such variations in the waste characteristics coupled with deficiencies in
the material handling equipment have caused operational difficulties, equip-
ment jams, system shutdowns, excessive wear, and abandonment of some equip-
ment.  For example, compacted waste in the silo storage and recovery system
caused severe operational difficulties and such excessive wear that operation
of that system had to be discontinued.  Compacted waste in the ram hydraulic
feeders frequently became so jammed that the waste could not be extruded into
the kiln until the system was shutdown and the waste cleared away.  Large
agglomerations of fused slag have jammed both the kiln and the gas purifier
residue conveyors to the point where the process had to be shutdown.  Other
equipment jams, such as those on the residue conveyor to the materials
recovery building, have been due to the binding effect of wires and stringy
materials.  In addition, frozen and low density waste sliding on the inclined
conveyors have hampered the operating process.

     Conclusion;  Solid waste is a heterogeneous material having widely vary-
ing characteristics.  To successfully handle or process this material, the
equipment design must allow for the variations in size, moisture content, and
density of the material handled.

Program Management

     The review of the direction and administration of the project stressed
the need for a unified organization when a large-scale scientific-engineering
developmental program is undertaken.  The following summary of the demon-
stration management indicate, that despite the generally common interests and
professional integrity of the participants, the circumstances did not allow
the formation of such an organization.

     Four agencies were involved in the demonstration:  The City of Baltimore,
State of Maryland, Monsanto EnviroChem, and U.S. EPA.  While all worked toward
the successful performance of the demonstration, their particular interests,
responsibilities, and orientation differed widely.  As noted earlier in this
section, any decision-making had to be approved in a varying chain relationship
by the City of Baltimore, Monsanto, and EPA.  Consequently, their varying
interests and perspectives caused delays, especially in the plant shutdown
periods, and hampered plant operation and administrative functions.

     During the startup, organizational and operational complications arose in
defining the roles and responsibilities of personnel.  While the City of
Baltimore provided the personnel for the plant operation, maintenance, and
administration, Monsanto EnviroChem provided the personnel for engineering
support.  Monsanto's services consisted of offering technical recommendations
for the plant operations and supervising the plant operators.  At the outset,
Monsanto had to increase the number of its onsite engineers.  At the conclu-
sion of the demonstration, the entire Monsanto team withdrew from the site
without a similarly qualified group from the city organization to take its
place.  The Baltimore project engineer assigned to the demonstration was not
intended to continue his plant position after the program was completed.  In
addition, three of the management personnel at the plant resigned during the   *
demonstration, and were not be replaced with similarly qualified personnel
because of a city hiring freeze.


                                      26

-------
     During the startup, the plant manager had limited authority over plant
operation because of the assignment of a project engineer who transmitted
information directly between Monsanto and the Head of the Department of Public
Works.  Normally, the downward chain of command from the Head of the Depart-
ment of Public Works is to the Head of the Sanitation Division, to the
Incinerator Chief, to the plant manager.  Consequently, bypassing the Head of
the Sanitation Division and the Incinerator Chief, precluded their normal
function and service as well as their support of the plant manager.

     Still other impediments to the efficient performance of the demonstration
were the city's procurement and accounting systems.  According to established
practice, the city's procurement office required a minimum of 2 weeks to
process a purchase order.  However, after the city took over the operation of
the plant, a more timely processing of the plant's purchase orders occurred.
The city's accounting system was structured for expenditures in typical routine
city operations but not for the unique transactions required for the plant
operation.  Consequently, it was very difficult during the demonstration
evaluation to use the accounting system records in determining 'the cost trade-
offs, the system operational efficiency, and the actual cost of the solid
waste processing.

     Conclusion:  An inadequately integrated organizational structure can be
an impediment to efficient project execution.  Responsibility for plant startup
and demonstration of new resource recovery processes should be assigned to the
system designer, with the transfer to the normal operating staff after
acceptance.

SUMMARY

     From the viewpoint of the Monsanto EnviroChem, the demonstration was
unsuccessful.  But to the U.S. EPA, SYSTECH, and the resource recovery industry
in general, the demonstration achieved most of the desired results.  The
particularly useful results were  (1)  the thorough investigation of the
Landgard process;  (2)  the invaluable trial-and-error experience to prevent
future costly but ineffective innovations; and  (3)  the development of signi-
ficant equipment and operational procedures, notably the rotary kiln.

     The rotary kiln has proved to be an excellent primary reaction vessel
after its evolvement and conversion from a pure pyrolytic reactor to a
substoichiometric combustion reactor.  In addition, the demonstration proved
the feasibility of utilizing the system concept for heat recovery from muni-
cipal solid waste.  Moreover, while the kiln is definitely competitive with
other similar commercial processing systems, it has the potential, when combined
with heat recovery, for being economically competitive with direct raw waste-
to-landfill operations.  The fulfillment of this potential depends primarily
on developing auxiliary system components with the degree of reliability
normally attained in other types of commercial processes.

     In addition to the invaluable experience gained and the various contribu-
tions to the state of the art, the demonstration and its subsequent evaluation
has revealed the critical importance of the following:  (1)  conducting
exhaustive theoretical and empirical research before scaling from a prototype

                                      27

-------
to a large-scale facility;  (2)  identifying and analyzing the effect of any
variations from the design of a proven system;  (3)  designing material
handling and processing equipment with provisions for surge capacities and
equipment-weather contingencies in view of the heterogeneous content of
municipal mixed solid waste with its time-varying characteristics; and  (4)
ensuring that the program management is an integrated organization with the
efficiency and common objectives required for a large-scale scientific-
engineering development.

     Reasonable projections indicate that the proposed modifications and
improved operating procedures will make the Baltimore plant economically
viable.

     In any event, the rotary kiln potential should be further exploited by
conducting additional theoretical studies to characterize and optimize kiln
operating conditions and by replacing malfunctioning auxiliary equipment with
the purpose of  (1)  improving the system reliability,  (2)  simplifying the
overall  process, and  (3)  reducing capital and operating costs.
                                     28

-------
                          EPA  REGIONS
U.S. EPA, Region 1
Solid Waste Program
John F. Kennedy Bldg.
Boston, MA 02203
617-223-5775

U.S. EPA, Region 2
Solid Waste Section
26 Federal Plaza
New York, NY 10007
212-264-0503

U.S. EPA, Region 3
Solid Waste Program
6th and Walnut Sts.
Philadelphia, PA 19106
215-597-9377

U.S. EPA, Region 4
Solid Waste Program
345 Courtland St., N.E.
Altanta, GA 30308
404-881-3016
U.S. EPA, Region 5
Solid Waste Program
230 South Dearborn St.
Chicago, IL 60604
312-353-2197

U.S. EPA, Region 6
Solid Waste Section
1201 Elm St.
Dallas, TX 75270
214-767-2734

U.S. EPA, Region 7
Solid Waste Section
1735 Baltimore Ave.
Kansas City, MO 64108
816-374-3307
U.S. EPA, Region 8
Solid Waste Section
1860 Lincoln St.
Denver, CO 80295
303-837-2221

U.S. EPA, Region 9
Solid Waste Program
215 Fremont St.
San Francisco, CA 94105
415-556-4606

U.S. EPA, Region 10
Solid Waste Program
1200 6th Ave.
Seattle, WA 98101
206-442-1260
                                                              V101712
                                                              SW-719

-------