600977034
EPA-600/9-77-034 December 1977
THE PREPARATION OF
FUELS AND FEEDSTOCKS FROM
MUNICIPAL SOLID WASTE
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and applicat-on of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9 Miscellaneous Reports
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/9-77-034
December 1977
THE PREPARATION OF FUELS AND FEEDSTOCKS
FROM MUNICIPAL SOLID WASTE
Proceedings of a Workshop held at
The Braniff Place Hotel, New Orleans, Louisiana
February 8-10, 1977, and Sponsored by the
U.S. Environmental Protection Agency
Solid and Hazardous Waste Research Division and by
Technology Transfer
Edited by Marjorie A. Franklin
Franklin Associates, Ltd.
Prairie Village, Kansas 66206
Project Officer
Donald A. Oberacker
Solid and Hazardous Waste Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
FOREWORD
The Environmental Protection Agency was created because of increas-
ing public and government concern about the dangers of pollution to the
health and welfare of the American people. Noxious air, foul water, and
spoiled land are tragic testimony to the deterioration of our natural
environment. The complexity of that environment and the interplay between
its components require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention,
treatment, and management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources, for the
preservation and treatment of public drinking water supplies, and to
minimize the adverse economic, social, health, and aesthetic effects of
pollution. This publication is one of the products of that research; a
most vital communications link between the researcher and the user
community.
This report is intended to record the proceedings from a workshop
in which important matters relating to solid waste processing were
discussed. Attending the workshop were a large number of people who
represented the nation's leading expertise in the field of converting
solid waste from the as-produced state into useable fuels and feedstocks
for recycling. It is anticipated that each comment, conclusion, and
recommendation by these experts from government, industry, and the
public sector will be of immense value to anyone planning or engaged in
the selection, research and development, design, or procurement of such
equipment and systems.
Francis T. Mayo
Director
Municipal Environmental Research Laboratory
iii
-------
PREFACE
Municipal solid wastes (MSW) contain almost everything imaginable in
terms of materials and resources which could, but for the correct tech-
nology, be recycled or used again for a good purpose. Much of the wastes
is burnable, therefore some of the first uses explored have been to use
MSW as a fuel for heating and/or other power needs. For the non-burnable
fractions of MSW, it has been recognized that these portions are rich in
metals and glass. The iron, steel, aluminum, brass, and other metals can
be separated out and reused. The glass can also be reused with a net
benefit to all concerned.
All of the above mentioned recycling schemes are dependent upon
finding the right processing and handling technology to effect the desired
separation steps or to produce the right form of the end products. The
disciplines of engineering, science, and economics are heavily involved.
Above all, the environmental effects must be carefully watched so that
man does not further pollute the environment while in the process of
saving valuable materials or energy from this waste stream which is
otherwise destined to be disposed of by burial or incineration.
This report should help shed light on a number of possible process-
ing techniques whereby recycling might occur. The technology discussed
is the latest "state-of-the-art" available, as presented by some of the
leading professionals and experts currently working in this field of
interest.
IV
-------
ABSTRACT
This report is a record of the proceedings from a workshop held in
New Orleans, Louisiana on February 8-10, 1977, among a group of experts
in the field of processing municipal solid waste into useable fuels and
other recyclable materials. The body of the report consists of a
verbatim discussion among meeting participants. These sections were
transcribed from the tape-recorded meeting sessions. In addition, a
summary of all discussions and conclusions is included. The report
serves as a comprehensive, up to date, state-of-the-art summary for
municipal solid waste processing technology which should be useful to
designers and researchers in the field of solid waste management.
-------
CONTENTS
Disclaimer ii
Foreword iii
Preface iv
Abstract v
Contents vii
Acknowledgement viii
SUMMARY 1
VERBATIM TRANSCRIPT 31
SESSION 1: SYSTEMS PROCESS DESIGN 33
Moderator: Dr. Albert J. Klee, U.S. Environmental
Protection Agency
SESSION 2: SELECTION OF EQUIPMENT 70
Moderator: Stephen Lingle, U.S. Environmental Protection
Agency
SESSION 3: TECHNICAL OBSTACLES 130
Moderator: Dr. Harvey Alter, National Center for Resource
Recovery
SESSION 4: APPROACH TO R/D&D PROGRAMS 174
Moderator: John Burckle, U.S. Environmental Protection
Agency
SESSION 5: SUMMARY AND CONCLUSIONS 207
Moderator: David Berg, U.S. Environmental Protection
Agency
APPENDIX A - Agenda 224
APPENDIX B - Participants 228
vii
-------
ACKNOWLEDGEMENT
The author, on behalf of the U.S. 'Environmental Protection Agency,
wishes to express his appreciation for all of the efforts made by the
entire group of people who participated in the development of this report.
In particular, Mr. John 0. Burckle and Dr. Albert J. Klee who
originated and implemented the workshop; Mr. Carlton C. Wiles who devoted
considerable efforts toward its success; Dr. James E. Smith Jr for the
conference accomodations and organization; and finally all of the
participants who contributed so much are to be acknowledged.
Mrs. Marjorie A. Franklin of Franklin Associates and her secretary
Mrs. Shari Brown performed exceedingly well on the difficult task of
technical interpretation, transcribing, summarizing and typing the
manuscript. Mr. Steven C. Lees of EPA, co-op student from the University
of Cincinnati, handled the job of overall coordination and management
of EPA's requirements, editorial changes, and final proof reading with
professional skill.
Finally, Mr. Herbert I. Hollander is commended for his thorough
efforts in reviewing and commenting on the entire volume of transcribed
dialog, not just his own meeting contributions.
Vlll
-------
THE PREPARATION OF FUELS AND FEEDSTOCKS
FROM MUNICIPAL SOLID WASTE
SUMMARY
The workshop on preprocessing and preparation of fuels and feedstocks
from solid waste was called to order by John Burckle of the Municipal
Environmental Research Laboratories, Office of Research and Development,
USHPA. He introduced David Berg, Office of Energy and Minerals Industry,
Office of Research and Development, USEPA, who gave the opening remarks.
Berg explained that the purpose of the workshop was to bring together
people most involved in resource recovery—architects/engineers, cities,
government agencies, and companies—to talk together. He mentioned that
the Office of Solid Waste and ORD had cooperated on the meeting because of
:;hair interest in solid waste. EPA is looking for mechanisms to facilitate
debate, discussion, and resolution of issues in solid waste. Berg outlined
7.1-e agenda: Session 1, System Process Design; Session 2, Selection of
Equipment; Session 3, Technical Obstacles; Session 4, Approach to R/D&D
'Programs; and Session 5, Summary and Conclusions. He mentioned the site
visit to Recovery 1 scheduled for February 10.
SESSION 1: SYSTEM PROCESS DESIGN
Dr. Albert J. Klee moderated Session 1. He outlined two objectives for
the session. The first was to determine what information on preprocessing
is wanted or needed by those involved to select a system given the equip-
ment available today, and what information is needed to design more ef-
ficient systems than those available today. The second objective was to
identify the critical aspects of design in the area.
Information Needs
A strong theme running throughout the session (and other sessions as
we]l) was the lack of information available to designers, manufacturers,
and users of the equipment and systems. Harvey Alter described it that
there was no state of the art recorded. On numerous occasions members
spoke for or against the desirability or even possibility of writing a
handbook or "cookbook."
The discussion on information began when Dave Bendersky presented a
summary of a report by Midwest Research Institute on the state of the art
of preprocessing. MRI had interviewed three groups—system designers,
manufacturers, and users/operators. All of these groups agreed on the
need for more data. The designers need information on equipment and
arrangement alternatives, how the equipment actually works on municipal
solid waste, and cost effectiveness data. The manufacturers need in-
-------
formation on actual operating conditions and an exchange of information
between manufacturers and system designers and users. Bendersky remarked
that adequate data in the open literature is not now available, just
personal exchange of information. The users are not instructed how to
operate new plants efficiently, and must learn from experience.
Bendersky and Alter stressed the need for good planning before making
tests in a plant. Bendersky mentioned that the ultimate purpose of the
data should be determined before testing to avoid getting random inform-
ation. Alter suggested determining the forirat and range of parameters
before taking data. Bendersky said that we should develop tests with
commonality, so that we can evolve comparative data from different plants.
It was generally agreed that the withholding of proprietary Inform-
ation by private companies is a serious impediment to the flow of inform-
ation. Those speaking on this point included Floyd Hasselriis, Dave
Spencer, Steve Lingle, Art Purcell, Tom Lamb, and Peter Ware. Hasselriis
felt that a possible solution is to use government facilities, such as the
Bureau of Mines, to test fuels and get standardized information which can
be compared with other fuels such as coal. Lingle pointed out that private
companies are reluctant to let a government contractor come into their
facilities for the purpose of collecting data. Spencer suggested that a
solution is for the government to fund basic unit operation development and
demonstration programs, with the results available to all. He felt this
would be more economical than the various industries spending development
money over and over. Hollander suggested developing data-gathering programs
and installing monitoring equipment at all the full-size operating plants
being installed in the public sector, i.e., Chicago and Rochester. Any
new facility should incorporate a Federally-funded data-gathering system.
With regard to feedback of information from users to manufacturers,
some members felt that manufacturers do not actively seek this information
on their equipment. Speaking of their experience on this point were Dr.
George Trezek and Bill Ryder. Ryder felt that this sort of thing fell
into the category of R&D and competition being what it is today, in order
to keep prices down manufacturers are reluctant to expend money for research
on a piece of equipment once it has been built.
Bob Powers said that since manufacturers do not make information on
their equipment available, it is up to the plant operators to make the
equipment work. He felt there can never be a "cookbook" plant because a
plant must be designed for local conditions.
Hollander felt that operating instructions were not possible for a
new plant, but guides should be developed as operation is experienced.
Unfortunately, experience is not always passed on to new people coming
into the plant, therefore a concerted "on-going" orientation and training
program should have high priority.
-------
Speaking about proprietary information, Ware said that since Occi-
dental's plant is partially funded by the government, all testing there
would become public knowledge. He felt that rather than losing their
competitive edge, Occidental would benefit from the publicity.
Dr. Greg Rigo felt that you should not shut down a plant while col-
lecting data. Spencer felt that testing an operation would always disturb
the operation.
Modeling of Alternative Systems
Rigo described a modular design which his firm had used to simulate
performance. He said that everything cannot be accounted for, and the
results depend on the quality of the data going in. Rigo felt that with
good models and good planning, data collected can be used to validate models.
Otherwise a great deal of data can be collected, but have no value.
Burckle stated that there could never be a standard plant, but the
principles of equipment models, unit operation models, and systems models
could give optional alternatives.
Sequencing of Unit Operations
The discussion of sequencing was opened by Alter, who said that de-
signers now follow the sequencing of existing plants. He suggested that
instead they should look at their objectives first, then sequence unit
operations accordingly. Later Alter suggested comparing different oper-
ational sequences on paper by writing down the performance of the various
components, then determining product quality to see if it meets speci-
fications.
Ware said that in Occidental's experience, their process and end
specifications were already known. They visited other plants, then de-
signed their sequence to make the feedstock for pyrolysis.
Lamb commented that there is no one right answer on sequencing.
Spencer felt that unit operations must be evaluated before systems are
evaluated.
Materials Handling
Early in the session, Alter stated that materials handling is an
important problem. He contrasted the cleanliness of European plants with
the dust and other hazards of many plants here. Bendersky agreed that
materials handling has been ignored.
Ware described a method they had used where air was circulated back
into the shredder to eliminate dust. A vacuum system was applied to the
shredder.
-------
Scale-up
Considerable attention was given to scale-up of pilot plants and
whether testing could best be done in pilot plants or full-scale operating
plants.
Roger De Cesare stated that a first plant is always a pilot plant what-
ever its size. He felt that there must be flexibility in the design of a
plant to take care of unforeseen things that happen. Different problems
arise from scale-up of a pilot plant; therefore design is empirical.
Trezek felt it important to use a realistic scale in testing.
During the discussion of testing at full-scale, Bendersky stated that
full-scale testing can obtain some data, but other data requires special
test facilities. Ryder felt that testing on a full-scale plant is more
productive because the problems associated with a L&B Model or pilot plant
differ radically from those in the full-scale plant and stated that "the
mere act of increasing the size changes the problem." Ryder also felt
that Federal grants should include any expense of downtime during testing.
There was some discussion of testing unit operations rather than
systems. Rigo thought that someone (probably the Federal government)
should build a pilot plant so controlled experiments could be run on
unit operations. Spencer saw a need to test operations rather than build
a pilot plant.
Quantity/Composition of Raw Refuse
Much debate was generated over the need for data on quantity and/or
composition of the waste stream. Hollander and Don Walter felt it important
to know what is going into a system in order to design properly. Hollander
pointed out that the cellulosic and mineral fractions are needed to do the
economics. Burckle agreed that composition is important to economics.
It was generally agreed, as Arnold Chantland pointed out, that there
is great variability in the incoming material. Alter suggested that a
system should be designed for a bandwidth on composition, and that the
contractual tip fee should be readjusted based on operating experience.
Dr. A. W. Joensen stated that it is very expensive to look at compo-
sition going into the system. He suggested looking at each piece of
equipment within a range of composition, and designing subsystems.
Lamb stated that quantity of waste is much more important than compo-
sition, since composition varies day-to-day. He said the quantity to be
delivered is important for a contract with a community. He felt that the
economics should not be dependent on a certain mineral fraction.
-------
Specifications and Requirements of Fuel
Lingle asked how much processing is necessary to produce a marketable
RDF. Rigo felt that the fuel should be as poor quality as would be ac-
ceptable, thus saving processing costs. Hollander pointed out that coal
has been used for years, yet it varies greatly in quality.
C. W. Fay disagreed with the "poor quality" idea, stating that another
utility failure would hurt the industry. He said the utilities don't want
to take the risks. The fuel should be tailored for the customer.
Hasselriis pointed out that the industry had gone from an unprocessed
fuel burned on a grate to a dry powder fuel which is more uniform than coal.
He said that processing equipment must be designed to accommodate a plus-
or-minus 25 percent variation, or more, in any particular characteristic
or constituent. The variability of the final fuel or other final product
depends on the amount and effectiveness of the processing done.
F. E. Wisely observed that the quality of the refuse-derived fuel was
more a function of the potential fuel user than of some predetermined
standard. Some users specify a high degree of refinement of the fuel;
others are far less demanding. In the interest of practicality, some
compromise often is necessary.
Walter suggested that it is a mistake to assume the utility is the
best customer for RDF. The utilities have long-term contracts for low
cost fuel, and they are conservative. On the other hand, the smaller
industrial boilers with spreader stokers have higher fuel costs now, and
are less conservative. There are also more potential customers among the
industrial boilers. Hasselriis agreed with this point. Hollander pointed
out that the industrials are being considered low priority fuel users and
have limited leverage in obtaining fuels.
-------
SESSION 2: SELECTION OF EQUIPMENT
The moderator, Stephan Lingle, opened the session by breaking unit
operations into materials handling, separation, and size reduction. He
listed numerous topics as possibilities for discussion. The moderator
then called on Dave Bendersky of Midwest Research Institute, who pre-
sented a summary of a report prepared by MRI. The report lists 12 pre-
processing equipment categories. Bendersky said they found research
needs in all categories. Shredders have received the most research at-
tention, but many questions are still unanswered. Full-scale operating
experience is not generally reported in the literature; people are getting
experience on their own. Bendersky discussed shredder problems such as
maintenance schedules, safety, and operating costs. In areas other than
shredders even less is known, and experience is not reported in the lit-
erature. Process control and dust control are not covered at all.
Materials Handling
General. It was generally agreed that more research and exchange of
information is needed. Dr. Harvey Alter mentioned the problems of materials
bridging and hanging up. He said we learn through a process of art, and
experiences are not recorded. Steve Hathaway pointed out that a theory of
design and flow for flow and storage faciliaties was developed at the
University of Utah in the 1950s. However, this is known by very few.
Hathaway felt that research is needed on the mechanical properties of RDF.
Harvey Funk felt there are better ways of storage and retrieving than
those in use now. He felt research could identify conversion processes and
feed rates. New systems could be compared to the Atlas systems in use.
Funk thought a design handbook would be of benefit, and noted that a few
manufacturers have published handbooks.
Bob Powers felt a problem is that the equipment is not designed for
garbage handling but is borrowed from other industries. He felt that manu-
facturers should design specifically for the refuse processing. Bendersky
noted that industry does not get the information needed to improve equipment
design. He suggested a conference of users and manufacturers to exchange
information.
Atlas Storage Systems. The moderator stimulated much discussion by
suggesting that the Atlas storage system is not working. He mentioned
Baltimore where, perhaps because of material size, the material has solidi-
fied, and there have been wear problems on the floor.
Arnold Chantland said that the Atlas system is not a total failure.
He said the system will handle material less than 2 inches, but will not
handle larger material. To prevent spontaneous combustion and Btu loss,
the bins should be emptied once a week.
-------
Funk also felt that the Atlas system does work, but he said it is a
high energy user. He said the problem of floor wear can be handled in
design. Funk felt the Atlas system is still the best available.
F. E. Wisely stated that the Atlas bin should be used as a distrib-
utor, not for storage. He also felt the Atlas is the best available now.
Wisely mentioned that vehicles can be used for storage up to a certain
point.
At Chicago, Bill Ryder said the City of Chicago ended up with two Atlas
bins as a result of preparing specifications for competitive bidding. The
system has not yet been tried with refuse. They do have a performance
specification the contractor must meet.
Dr. A. W. Joensen mentioned that an Atlas bin sweats inside and that
therefore ice may form. This can be avoided by running unused conveyors
every three or four days. Joensen said a mass weighing system for control
purposes into the boiler should be part of a storage and retrieval system.
He recommended a belt system using the Atlas bin.
Conveyors. Peter Ware described the Occidental system. Refuse is
received on the floor from trucks. It is fed into the shredder, then
conveyed to the floor of the building. The shredded refuse can be stored
on the floor for up to 8 weeks with no problems. A front end loader then
feeds the refuse into a doffing roll bin; it then goes to the air classifier.
He felt the doffing roll bin is very effective and the capital cost is about
half that of an equivalent Atlas bin.
A number of problems with conveyors were discussed. Tom Lamb mentioned
spillage and belt cleaning. He said MSW does not flow—diverge, not con-
verge. Alter described problems in the NCRR pilot plant when they first
had self-cleaning head pulleys on all conveyors but even these sometimes
get clogged. There is some now-obscure relation between the nature of what
is clogging and the size of the pulley. When the material is like gravel,
it "dances." Rags hang-up everywhere. Design of materials handling equip-
ment for a resource recovery plant is often a process of trial and error
and it is important to record experiences if others are to learn.
Another problem described by Dr. Greg Rigo was that a dry load of
material would not go up an inclined conveyor. At one installation they
sprayed the material with water to make it go up. Rigo stated that compo-
sitional differences must be accounted for.
Wisely recounted an experience at St. Louis, where a vibrating con-
veyor was put in a pit. The frequency was a direct multiple of the building
height, and the problem was accentuated by the pit.
Chantland mentioned the dust problem caused by the drop of material from
one conveyor to another. Lamb added that you will find dust at every trans-
fer point. To control dust at Ames, Chantland used a pneumatic system.
Replaceable liners were used at the elbows to solve the wear problem. To
solve the wear problem on air lock feeders, replaceable blades are used.
-------
Size Reduction
Investigations at Berkeley. Dr. George Trezek showed slides and
described his investigations on shredders for EPA. He said that about
1969 shredders were designed for brittle materials. Since refuse is about
25 percent brittle, further investigation was needed. He first investi-
gated existing machines, then set up a machine at the University to run
controlled experiments.
The equipment includes a Gruendler hammermill modified to operate at
variable speeds. Feed rates on the conveyors can also be varied. The
equipment also includes an air classifier.
Trezek stated the subject of investigation is: Given an input size
distribution, can the output size distribution be predicted for a particular
set of conditions, using a heterogeneous material like solid waste. He
summarized two theories of size reduction. The continuous theory works for
brittle materials. The discontinuous theory, which Trezek used, is a
statistical treatment. He also described a breakage function and a selec-
tion function. The breakage function describes how material is broken
according to size distribution. The selection function describes the
fraction of material selected to be broken during a given task.
Trezek showed slides summarizing the relationships developed in his
work. He correlated energy consumption and other parameters such as speed
and moisture content with size distribution. One can therefore determine
how energy savings might be obtained. Given the product size distribution,
the model can predict primary, secondary, and tertiary grinding. The re-
sults show that running the machine at lower speeds produces less wear of
the hammers. Trezek measured size distribution of all components of the
waste stream and looked at system arrangement of shredders, screens, etc.
Energy expenditures were plotted to obtain a characteristic size with
various arrangements. Moisture in the waste stream can also be related to
energy content.
Alter described work where they had looked at particle size distri-
butions of material from 10 shredders around the country. They found that
the PSD fit the Rosin-Rammler distribution. They also found that the Bond
Work Index is within about plus or minus 20 percent of 400 for all shredders.
If this value is verified, then the relationship between power input and
particle size can be predicted for shredders.
Alter also mentioned that the hammers in the secondary shredder oper-
ating on air classifier light fraction did not shred textiles well enough
to avoid plugging the d-RDF pellet mill. When the hammers in this vertical
shredder were replaced with rotor blades and stator blades, the textiles
were cut. Alter said that their work generally agreed with the results
described by Trezek.
-------
Trezek added that the next stage of his work will be to look at wear
characteristics versus particle size distribution.
Moisture Loss. Following Trezek's presentation, there was some dis-
cussion of moisture loss through the shredder. Floyd Hasselriis pointed
out that in Trezek1s test the moisture range was 15 percent up. He said
that at lower moisture content, the material is very tough and hard to
pull apart, then suddenly it becomes brittle.
Doug Fiscus observed that the moisture loss through the shredder is
apparently about 6 percent, and this must be accounted for in the economic
model of a plant. Roger De Cesare said that the moisture loss in the Bureau
of Mines pilot plant is 10 percent (including air classification and shred-
ding) . Herb Hollander said that under certain atmospheric conditions the
moisture released in a shredder can actually be seen as a vapor plume
leaving the shredder dust cyclone (i.e., St. Louis).
Hasselriis stated that the horsepower into a mill heats the material,
then drys it. Alter ended the discussion on moisture loss by observing that
moisture is not evenly distributed in refuse, therefore a moisture balance
is meaningless.
Other Research Needs. There was some general discussion of other re-
search needs in the size reduction area. Lamb felt that types of size
reduction other than shredders should be explored. He reported success with
ball mills. Ware asked why we shred and stated it was for convenience of
storing and handling. He felt that detailed research is fine for optimi-
zation, but this can come later. Ware observed that shredders have vast
differences in performance for the same horsepower, and would like to know
why. Alter said that we do not really understand size reduction, and there-
fore we have no basis for invention. He suggested that a sizing circuit to
achieve a homogeneous material may be a good approach.
Rigo observed that power surges through the shredder must be accounted
for in design. He also suggested presegregating wood items, which could be
put into a secondary shredder rather than primary size reduction. This
would avoid losing the heating value of the wood.
The topic was concluded with a discussion of size reduction for small-
scale operations. Hollander observed that size reduction goes back many
years, and shredders in most early waste processing installations were too
small. Rigo felt that not much was now available for small-scale operations.
Alter disagreed, saying the flail mill works well at 5 tons per hour. The
large items can be picked out first. Rigo said you can use a trommel pre-
breaker to "open up" the refuse bags. He suggested using a magnetometer
to stop the feed conveyor if something massive comes through, to minimize
damage.
-------
Separation
General. There was general agreement that sufficient information on
separation is not available. Alter and Bendersky spoke on the need for a
sharing of information among those having operating experience. Alter
said we need orderly experiments in existing plants and to make comparative
measurements on the same basis. Lingle said that we don't want to vary
parameters in existing plants and need separate research facilities.
De Cesare felt there is a need for basic research on unit operations,
while Ryder saw a need for basic research to improve systems. Lamb sug-
gested setting up a matrix of data-defining variables such as particle
size and moisture content. He felt the research should test unit operations,
not define systems. People can optimize their own systems. Dave Spencer
suggested that since testing is very expensive, it should be done properly
or not at all.
Hathaway and Rigo spoke on the need for classification for small sys-
tems. They noted that military waste is different from civilian waste,
however, so research there may not help others.
Rigo pointed out two problems in air classifier research. The first
is the basic fundamentals of how an air classifier works. The second is
research on the performance of specific equipment. Rigo felt the government
should not do the latter.
Tuning of Systems. There was some discussion of tuning the equipment.
Bendersky said that individual companies tune their equipment and should
make the information available to others. Spencer observed that tuning
equipment depends on the material to be put in and performance therefore
varies. Peter Cambourelis added that it also depends on the fuel require-
ments of the customer. John Burckle felt that rangeability of the fuel is
not so wide and a tune point is needed.
C. W. Fay felt a need for flexibility in design so changes can be made
as the waste stream changes. He said there could be trouble in the future
if we fine tune for conditions today.
Hasselriis pointed out that an air classifier is not a simple device;
different functions take place inside, each of which may take separate
tuning to get the desired result. He felt we need fundamental research on
the unit operations inside the device.
Types of Separation. Some members of the workshop described specific
types of separation. Hasselriis said that magnetic separation needs a
second step, similar to winnowing, to separate other materials that go
along with the magnetics.
Don Walter described the use of a drum conveyor which takes a first cut
of the magnetic material, then a second magenetic belt conveyor separates
the light from the heavy ferrous fraction.
10
-------
Alter described the use of a ballistic drum at Warren Spring, England,
to separate food waste by bouncing the light material.
Process Control
The moderator outlined some areas of interest: Dust control, explosion
control, and fire control. He mentioned problems at early solid waste
plants.
Dust Control. Hasselriis said that it is necessary to draw air through
shredders, since they act as fans, and to keep the shredder outlets under
negative pressure to prevent dust and lint from blowing throughout the
building. When enclosing the air system, fire hazards can be reduced by
recirculation of air, and maintaining fairly high levels of water vapor.
A Fenwal rate-of-rise sensor can detect the fast pressure buildup associated
with fires.
Lamb observed that equipment is available to control dust. He felt
information is needed on the properties of dust. Safety is noncontroversial,
he said, and people should be willing to share experiences. Safety standards
need to be set.
Hathaway said they have done some testing on the properties of dust,
and it is both highly fibrous and highly volatile. It would be desirable
to reinject the dust into the fuel production process because of its
volatility.
The health implications of dust were discussed at some length. It was
agreed that this is a highly sensitive area. Chantland said that limited
research only creates a problem in the mind of the public, without really
answering the questions. Hasselriis observed that the investigation creates
the problem by suggestion, and therefore we must be very careful in in-
vestigation.
Trezek said they have done health research in their laboratory, but
have only scratched the surface. Alter described the program at NCRR's
pilot plant. Employees are given an annual physical exam, and are in-
structed to report any job-related illnesses. He said if others would do
the same, information could be pooled. NCRR is also doing dust sampling.
Alter also told the group that ASTM Subcommittee e 38.07 was working on
standards in the health area. This is a sub-committee of the Resource
Recovery Committee E 38.
Walter said it would be necessary to extend health investigations to
include garbage collectors. Bendersky remarked that New York City has
done this. The findings generally were that the garbage collector is not
a greater health risk from respiratory diseases than other street service
people.
Lamb told the group that puncture injuries can be prevented with steel
soled safety shoes.
11
-------
Explosion Control. Alter said that a design manual on explosion pre-
vention is needed, but Bendersky felt that shredder manufacturers do not
have the information to produce a manual.
Ryder described the experience in Chicago. They had a shredder with
a Fenwal system for explosions and a sprinkler system to hold down dust.
Ryder felt that they had avoided an explosion in over 5 years of operation
because of this. However, last year a partially filled tank of acetylene
gas found its way into the stream and caused an explosion which damaged the
conveyor at the bottom of the shredder, the side wall of the building, and
relief vents on the roof. Contributing to the damage was the fact that
the relief vents had been sealed tight as they had been leaking water.
Ryder felt that it was impossible to prevent explosions when items of
this nature get into the refuse stream.
Ware said that they expect explosions at their plant. They have de-
signed their feed hopper so it goes straight through the roof; therefore
an explosion will blow through the roof. They keep the shredder isolated
to prevent injuries from detonations. They do not use a Fenwal system.
Fire Control. Ware said that steam can be used for snuffing fires if
it is available. However, water will produce steam, giving the same effect.
12
-------
SESSION 3: TECHNICAL OBSTACLES
John Burckle opened the session by saying that he hoped to hear about
technical obstacles from the perspectives of the public sector, the archi-
tects/engineers, and the consultants. This would help the Federal govern-
ment to direct their resources in R&D. Dr. Harvey Alter, moderator, out-
lined three areas for discussion: judgment and risk assessment, development
of a research agenda, and who might do the research.
Information and Research Needs
Dr. Alter first asked for comments on information needs from repre-
sentatives of municipalities, A&E firms, and the military.
Municipal Viewpoint. Richard Bush, representing Connecticut Resources
Recovery, mentioned scale-up problems, location of equipment, defining the
operation of individual items of equipment, sizing of systems, composition
and quantity of material recovered, and trommel operation as areas of
interest to him.
Ron Lalka of Los Angeles County said he was looking for: 1) relia-
bility to take all refuse delivered; 2) environmental acceptability; and
3) economics.
Economics. Economics was a subject of interest to many. Dr. Greg Rigo
stated that without good accounting systems, you cannot pull out reliability
and cost information. Dave Bendersky pointed out that present accounting
systems vary greatly, and you cannot now generally tell what it costs to
operate any one piece of equipment. He said that there must be some
standardization if comparative analysis is to be done. Arnold Chantland of
the City of Ames felt that economics is the uppermost consideration. Dr.
Alter said that NCRR is developing an accounting manual for Recovery 1, and
would be glad to share. Steve Lingle pointed out that EPA has an accounting
guide, which probably should be made more detailed.
Duplicate Lines. Chantland brought up the idea of having duplicate
lines to provide continuity of operation and avoid the cost of downtime.
Bill Ryder of Chicago agreed with the need for duplicate lines to meet the
commitment of fuel to the utility. He gave the example of the EPA shutting
down an incinerator in Chicago, which lacked air pollution control devices.
As a consequence, the City's new RDF plant must have the capability and
reliability to make up for the plant which had been shut down. Ryder stated
that in the quest for a reliable RDF plant, it should not be forgotten that
one of the primary purposes of an RDF plant is disposal of refuse. This is
insured with a standby line. However, should the utility be unable to
receive the refuse for any reason, a compactor will be provided in the
City's RDF plant so that the refuse, shredded or not shredded, can be
trucked to another disposal site. Ryder also mentioned that if more fuel
could be sold, the redundancy would be needed.
13
-------
F. E. Wisely commented that whether duplicate lines are economical
depends on local circumstances. They cannot be justified for some small
facilities. He also said you must have alternate disposal. Steve Levy
used the example of Franklin, Ohio, where they have not had to divert
waste because they have front end and intermediate surge storage to use
if the plant is down. He said you can oversize the parts of the plant
you know are going to be down more often. Wisely replied that that
system is adequate for a small system, but not for a large one.
Monroe County Experience. Dave Spencer described the Monroe County
plant experience concerning storage and redundant lines. He said they
have storage for one day's feed in the tipping area. The plant can operate
as a transfer station to the landfill if the power lines are down. They
have intermediate storage bins to allow time to fix problems, and also have
bypasses to remove feed to the landfill. They have additional machinery to
densify material for handling if they have to move it by truck. Also they
have redundant processing lines. Spencer said a problem in doing relia-
bility availability analysis is sufficient unit operations information.
They looked at lead times on spare parts and repair times to determine
inventory policies. Spencer felt that the government should not spend
money on availability reliability analysis because it is system oriented.
A&E Viewpoint. Harvey Funk opened this part of the discussion with a
comment that it is difficult to find something that has been demonstrated
when evaluating equipment. He said that developers of systems may not be
willing to sell an individual piece of equipment. Peter Cambourelis re-
sponded that at the time Funk was looking for an air classifier, Raytheon
did not have anything. He said it takes time to develop a demonstrated
product.
Performance Specifications and Bonds. The discussion initiated by the
A&Es turned mainly to writing performance versus hardware specifications,
and requiring a performance bond from the vendor. Wisely thought that a
specification should be a combination of good performance and hardware,
otherwise it is impossible to evaluate on a comparative basis. The spec
may allow alternates, but must attempt to define hardware.
Floyd Hasselriis said that if a contractor takes total responsibility
to design, build and operate the plant, he can accept equipment which has
a limited guarantee and which may not have been fully proven beforehand,
and make it work. He suggested allowing for substantial retrofit or
improvement in the project budget, a practice not readily permitted by
municipal bidding procedures.
Peter Ware described Occidental's situation. They are a private
developer but also have government money, so must go to the low bidder.
He has found a spectrum of response from manufacturers. Some will make
their equipment work, others have no interest. Ware suggested looking at
the responsiveness of manufacturers as well as their other qualities.
They have performance guarantees, but cannot always enforce them. In
answer to a question, Ware said they did not have performance bonds, but
retain a percentage of the purchase price.
14
-------
Ryder was asked about performance bonds at Chicago. He said that in
addition to a performance bond for the full value of the contract, they
hold back 10 percent of payment until the equipment is accepted. They
also have liquidated damages for late delivery.
There was some discussion of whether requiring a performance bond
adds to the project cost. Ryder felt it must add something, but believed
that a lot of grief was eliminated because a lot of vendors would not wish
to bid with this sort of a threat hanging over their head.
Tom Lamb stated that performance guarantees are good if they can be
effected. However, the owner cannot afford to lose money while he waits
for a manufacturer to make his equipment work, therefore the owner must
sometimes take action himself. Lamb said that the equipment manufacturer
cannot afford to take on large liabilities for the sake of a sale.
Spencer described the liability system for Monroe County. The owner
selected a system manager to provide professional services, rather than a
turnkey or full service contractor. The system manager will buy service on
a non-competitive basis, given design and construction management, and 5
years of operation. After 5 years the owner could bid operations to other
people. The Monroe County system is similar to that of the Bureau of Mines,
although unit operations are sometimes different. Raytheon gave a per-
formance guarantee for the total system and will market the output products
to specific specs. They will enter into long-term agreements with buyers
and guarantee to meet product specs. Raytheon passed the specs on to the
equipment manufacturers. They wrote performance as well as hardware specs.
Each vendor is only responsible for his own unit operation or subsystem,
and has complete liability for his equipment to the extent of the purchase
price. Raytheon assumes the liquidated damages, except for a few equipment
items which assume a reasonable share.
Dr. Alter observed that the performance guarantee is a way of dumping
the entire R&D responsibility on the equipment vendors.
Military Viewpoint. Steve Hurley felt that he must look at getting
fuel with the least investment in equipment. He asked what can people do,
what can machines do?
Captain R. F. Olfenbuttel stated the military is looking for the lowest
capital investment: this means source separation. However, they may have
to use alternatives. He felt they are faced with a lack of information on
what to do.
Full-scale vs. Unit Operation Testing. There was much discussion on
the usefulness of full-scale testing versus unit operations research. The
discussion also touched on full-scale testing versus the pilot plant ap-
proach. A general consensus was not achieved.
15
-------
Lalka slated thai he wanted to look at information from full-scale
operating plants. He felt that the feedstock at pilot plants is sometimes
synthesized. Alter disagreed, saying that the Bureau of Mines, NCRR, and
George Trezek ail vise real refuse. Lalka replied that every city has dif-
ferent refuse, and we should look at operations around the country. Com-
menting on Lalka's remarks, Burckle said we must ''eal with the factor of
scale because not all communities are the same size. Site specificity
also must be dealt with. He asked if we should look at a band width.
In another comment, Burckle suggested that a designer first draws a
process diagram and sets the performance needed from each unit operation to
get the desired quantity and amount of fuel. The designer then fills in the
blocks with specific equipment. Arlie Capps agreed with Burckle on the
systems approach.
Speaking on unit operations research, Bendersky felt we should do work
on air classifiers, etc., similar to that done by Trezek on shredders. By
controlling variables basic information can be gathered. Trezek remarked
that his next work would be on densifiers.
Alter said that to understand a unit operation, orderly change is made
and related to changes in performance. This makes it almost impossible to
make such investigations in an operating plant. For example, NCRR is build-
ing an air classifier which can easily be changed for research. The
changes - and their investigation - take time, hence are not the sort of
thing to be done in a production facility. Alter suggested a spectrum of
research, from the university to occasional measurements in a plant. NCRR
falls in between.
Rigo said there is a need for a statistical description of waste and a
statistical description of unit operations. He suggested looking at new
unit operations not now applied. He said the laboratory can be used to get
the shape of a curve, then field testing can get data points to validate.
Lamb saw a need for fundamental research on a unit operations basis.
If the basic data is available, the designer can select the pieces to
optimize his plant. Burckle pointed out that the performance of a unit
operation depends on the feedstock. He felt the need to look at complete
subsystems. Lamb replied that we can test individual pieces and predict
subsystem performance. Whether done all at once or by pieces, the basic
design data is still lacking. Spencer agreed with the latter point.
Lingle summarized by stating a need for a whole range of data col-
lection. First, we should gather data on the overall performance of total
plants, total systems. This would be of interest to the municipal sector.
Second, we need research on unit operations. This would be of interest to
the design engineer. Third, we need test procedures at a pilot plant to get
really basic data. This also would benefit design engineers. Fourth, we
need to develop new equipment. Lingle emphasized that different data is
useful to different people.
16
-------
Design Handbook. Repeating remarks of previous sessions, a design
handbook was again discussed. Roger De Cesare mentioned the Taggart book
on mineral dressing, which is based on tests of operating machinery. He
suggested EPA could have a mobile research lab to get similar data for
shredders. Fay felt that shredder manufacturers might not let EPA test
and publish data. Spencer also wanted a handbook like Taggart with oper-
ating information on various equipment, i.e., energy requirements, replace-
ment parts. Spencer said we are not even starting to obtain this infor-
mation; it is not in the research budget.
Other Needs. Some other research needs were touched upon throughout
the session. Rigo suggested looking at source segregation. Harry Freeman
felt that was outside the scope of the meeting. Rigo said that with good
unit operation data, solving the mixed waste problem will give a data base
to solve the segregated waste problem. Alter said that the decision whether
or not to institute source separation in a city is often a political one.
Hasselriis said that Trezek's mathematical model method is very power-
ful for interpreting test data on our equipment, and to evaluate the po-
tential for improvement.
Again there was some discussion on compositional data. Alter said
compositional data only gives one point in the year. We must have the
ability to adjust a processing plant to changes in composition. Lamb said
compositional data is useless because of daily variations.
Harry Freeman asked if we just need better information on what we have;
are we drawing the lines too narrow? Hurley replied that the Navy is trying
to apply what's available. Freeman wanted to look at research, not just
evaluation of existing equipment. Don Walter agreed that we should develop
new equipment for the solid waste feed source, not only use existing equip-
ment. Wisely also felt we should look at equipment that may have potential
for the future. He said we have been trying to get our information too
precise. We need ranges of information for various parameters.
Capps suggested setting up a field test laboratory to define R&D
problems.
Setting Priorities. In order to set priorities on research needs,
the moderator asked the group to turn to the list of research needs prepared
by Midwest Research Institute, and to choose five items of highest priority
and five that should not be investigated. Some members felt the list was
restricted to existing equipment, so Alter suggested that other priorities
could be added.
During Session 5, Dave Berg made a preliminary report on the results
of the voting. Receiving high priority votes were:
General
Determine Optimal Arrangement(s) of Unit Equipment
Study Emissions from Processing Equipment
Determine Effects of MSW Characteristics on Processing
Evaluate Other Potential1y Applicable Equipment
17
-------
Shredders
Compare the Performance of Various Types of Shredders
Screens
Determine the Effects of Pretrommeling
Economics
Develop Effective Accounting Method(s)
Determine Equipment Operating and Maintenance Costs
Receiving low priority votes were:
General
Determine Effects of MSV Characteristics on Processing
Magnetic Separators
Determine Effect of Magnetic Separation
Dryers
Determine the Effect of Drying on Bacteria and 7iiv.s
Evaluate the Effect of Drying on the Combustion Charac .>.-?"" tics
of RDF
Receiving Facilities
Evaluate Receiving Facilities
Evaluate MSW Segregation Frier to Processing
Ad Hoc Committees
The moderator suggested that workshop members should indicate their
interest in serving on committees to develop methods of evaluating unit
processes. Four areas were suggested: size reduction, separation,
materials handling, and drying.
A great deal of discussion ensued. Chantland suggested the need for a
matrix of boundaries to work within. Hasselriis felt the matrix would pro-
vide a framework which information could put in, even if it were proprietary.
Doug Fiscus called for a standard method of investigation: 1) the
methods have to be fairly specific; 2) the methods must be minimum or base
line and leave room for unusual occurrences.
Berg and Lingle suggested adding process control to the list, but Alter
said methods of evaluation were already available for process control.
There was a good deal of discussion on how to treat economics. Ben-
dersky suggested a separate economics group. Walter said economics is
really a part of each category, and Bendersky accepted that idea. Lingle
objected that that would mean economics would be done five times; it would
be better to do it all at once. Dan Keyes felt that you could not divorce
economic and technical problems. Lalka said the A&Es would be interested
in design information, and the municipalities in economics. Burckle said
we need a standard economic analysis for the overall plant and also a
measurement methodology for each unit operation, which would feed into the
total analysis.
-------
Alter again mentioned a design handbook. It should discuss the manner
in which data are taken. We need to use the same range of parameters so
data will be comparable.
Rigo said that the problem of intensity of data is critical. Alter felt
each subcommittee should handle this. He also said while the groups would
evolve the matrix, each individual investigator would judge how much data
to take. Alter also described the ASTM recommended practice, where an exact
format is followed. He indicated the groups might work toward that.
Burckle said we were talking about a commonality in taking measurements,
which an ASTM committee is working on now.
(During Session 5, Alter said he had had some discussions with other
members of the workshop, and had concluded that the ASME and ASTM provided
a suitable way to continue committee activity. He suggested workshop
members use the existing committees rather than form new ad hoc committees.
This seemed to be the consensus of the group.)
Technical Obstacles
Toward the end of the session, the group listed some technical ob-
stacles, as opposed to social, institutional, marketing, and economic
obstacles.
Chantland mentioned whether everyone can use the end product. For
instance, oil or gas boilers have to be retrofitted to use RDF. Cambourelis
agreed.
Capps mentioned the concern over air pollution in California. This
caused people to go to things like pyrolysis, which is yet unproved. Trezek
added that boilers in California are not equipped with ash handling equip-
ment, and therefore cannot take RDF.
There was considerable discussion about use of RDF in industrial
boilers versus utility boilers. Rigo saw a shift from gas and oil fired
industrial boilers to coal and solid fuel, at least in the Mid West. He
said new construction could be tailored for RDF. Lingle said EPA has looked
at industrial boilers. Nationally their total fuel use is one-half that of
utilities. Their average capacity is 30 tons per day of RDF. Only 50
existing industrial boilers could handle the output from a 500 ton per day
plant. This means multiple contracts for industrial boilers, which is not
as stable financially as a single utility contract.
Herb Hollander pointed out that since the energy crisis, oddly enough,
very few new industrial coal fired boilers have been contracted. It will
not happen overnight. A community would have to find several industrial
users for RDF, as opposed to a utility which signs a long-term contract.
However, it is the non-utility fuel users who have the greatest need for
a new local fuel source. Since they cannot realistically commit to long-
term fuel supply contracts, and only require relatively modest quantities of
fuel, they do not have the leverage to command fuel supply.
19
-------
Wisely said that industries which are really good markets for RDF have
a rather constant year-round plant factor. Most industries are seasonal in
their steam demand. The industries must have financial incentives to use
RDF, although there are other incentives such as being a good neighbor and
the environmental image.
Cambourelis said the utilities must make money. They have a unique
position as a buyer, because they are organizationally stable. He said we
must find a way to make RDF economically attractive to the utility, and
perhaps Federal tax incentives might make it worthwhile.
Fay suggested that the resource recovery plant should produce some-
thing like pyrolysis oil that anyone can use and get a good price for.
20
-------
SESSION 4: APPROACH TO R/D&D PROGRAMS
John Burckle, moderator, summarized various strategies for R&D: pilot
plant, testing existing systems in the field, testing subsystems on a
demonstration scale, and full-scale'demonstration. He asked members of the
workshop to pay particular attention to scale as they discussed.
Scale-up Problems
Ron Lalka opened the discussion on scale-up by mentioning Monsanto's
scale-up from a small pilot plant to a 1,000 ton per day plant. This was
a tremendously large scale-up, and he felt it more realistic to do less.
Lalka said that Union Carbide is going to a concept of 350 ton modules,
which can serve small communities as well as large. He felt efforts should
go in that direction. Lalka felt that data from one pyrolysis system might
not be applicable to another.
Burckle asked at what scale can research data be used in plant design?
Steve Lingle felt the scale depends on the objective, whether you are trying
to prove a system or develop design information on a component. Don Walter
agreed that it depends on the situation. Any research program must work in
an orderly manner to a full-scale demonstration. There is no given answer
on how much you can scale up. Walter said that the real problem is that
the operator makes a full-scale plant work, but the information does not
flow back to the designer.
Herb Hollander commented that equipment suppliers look at the market
size and the size of unit with the greatest sales potential. He said a
plant made up of modules would give the greatest flexibility. Hollander
asked what is scale-up to commercial size? Later he asked what is the
smallest commercial size plant that should be tested? Hollander asked
Arnold Chantland if he would have put in equipment of less capacity at
Ames if he knew the flow rate would be less. Chantland replied that he
would not, because of the severe abuse caused by the material. Hollander
then asked if a test program should be on equipment of the size at Ames.
Chantland said yes. Hollander remarked that a program must be on-going to
get both peaks and valleys, and Chantland said that he agreed with testing
of units within a process. Burckle asked whether this testing should be on
existing plants or a large scale test facility, and Chantland said either.
He suggested establishing parameters and ranges.
Bill Ryder stated that once a full-scale plant is built, it becomes the
pilot plant for the next generation of plants, and therefore there should be
no need to construct another pilot plant. Peter Ware replied that a pilot
plant would make the next big plants better. Ryder said that after a large
plant gets on-line he would evaluate it for possible improvements before
building another plant exactly the same way. Burckle said you could tune
a big plant to optimum performance, but not determine if a different sequence
of operations would be better. Ryder replied that he would be trying dif-
ferent things in his plant, and make on-going improvements wherever possible.
21
-------
G. C. Chisamore remarked that scale-up also goes on in other industries,
such as pulp and paper. Lalka, however, said that the chemical or pulp
industry works with a particular feed material, while solid waste changes
every day. He felt that is why a small-scale pilot plant cannot easily be
scaled up. Spencer felt that information from a system below 30 to 50 tons
per hour might not be applicable to a larger system.
Dave Bendersky said that we do not have systematic data to show what
equipment is scalable. The Baltimore experience showed that certain things
are not. Dr. Harvey Alter felt that we can learn from the minerals pro-
cessing industry. He said that screens, froth flotation cells, and trommels
can be scaled.
In defense of pilot plants, Dr. George Trezek said they give a feeling
for what to measure. Also they give flexibility to change things and try
new ideas. Spencer added that you have to go to the next step after the
small-scale plant. Dr. Greg Rigo said that hopefully the larger scale is
a validation of pilot scale data.
Sampling
Another major topic of discussion was sampling. Bill Parker of Recovery
1 remearked that it is easy to sample in a small-scale pilot plant, but in a
large plant built for profit, it is difficult to sample unit processes or
systems without perturbing the process. He asked for others' experience on
how large a sample should be and where to take it. He felt the best data
would be from large plants, but it's hard to get and expensive. Parker
said he needs testing to draw up maintenance manuals and set limits of ac-
ceptable performance. Ted Sjoberg said you have to evaluate data within a
time frame because of variations in the stream. Batch sampling will not
give the desired information. Parker agreed.
Speaking of their plans at Americology, Sjoberg said they will evaluate
efficiency on a weight basis and set up bookkeeping to evaluate maintenance
downtime relative to performance downtime and costs. Parker agreed that an
accounting manual is vital to determine the operating effectiveness of the
plant. They are working on a manual now, and designing it to pull data out.
However, the accounting manual does not see when aluminum, glass, etc., are
not meeting specifications.
Sjoberg said they will be sampling material every day to determine the
quality of the aluminum, glass, steel, and RDF. Their revenues depend on
quality.
Rigo said that at the Landgard plant, they are looking at outfalls and
hope to be able to "walk back" into the process. They will establish re-
lationships, and when a relationship changes, it will mean something has
gone wrong. Parker pointed out that going back from an outfall point is
difficult because you soon reach a common point. Rigo agreed that it is a
research problem, but felt that the relationships can be used as a warning
system.
22
-------
There was some discussion of sampling fuel. C. W. Fay said Wisconsin
Electric will do sampling of fuel in their own laboratory. Over a year's
time they will arrive at a variation agreement with Americology. The one-
year testing program is written into the contract.
Pilot Plant Operations
Burckle then introduced various speakers who summarized operations at
pilot plant facilities.
NCRR. Dr. Alter showed slides and described the NCRR pilot plant in
the District of Columbia. The building is leased at no cost from the
District in return for maintenance. The waste is received on the floor and
pushed with a front end loader into a conveyor pit. It then goes to a
Williams shredder. The shredder refuse discharge goes to a Triple/S air
classifier. The air classifier light fraction goes through a blower and is
pneumatically conveyed to the incinerator pit, where it is disposed, or it
goes to the d-RDF module. The heavy fraction goes past an Eriez belt
magnet to a magnetic head pulley, then through a trommel. The minus 2 inch
goes to a glass recovery module, the plus 4 inch is rejected as waste, and
the 4x2 inch middling goes to a CPC eddy current separator. The Al Mag
product is collected on the side. Dr. Alter mentioned that much of the
equipment is loaned to the Center.
Dr. Alter said that the problem of blowback at the shredder was solved
by hooding the input conveyor, so that it keeps coming back in through the
shredder. The Triple/S air classifier will be replaced by a vibratory
feeder to the research air classifier mentioned in an earlier session. In
answer to a question Dr. Alter said that evaluating air classifier per-
formance depends on your objective. One air classifier may be fine for
maximizing yield of fuel; for other applications, you may want to maximize
quality of fuel.
The Al Mag product contains aluminum cans, foil, and some other ma-
terial. It needs to be cleaned up to meet reasonable specifications. The
product is sent to a prototype air knife, followed by a vacuum hood. They
have been experimenting with the blowers to first suck off the organics,
then blow off cans and foil from diecast. They have been able to produce
Alcoa Grade 1 can scrap. In other tests they have found city-to-city
differences.
Dr. Alter then described glass recovery. The minus 1/2 inch goes from
the trommel to a screen, then on a conveyor to a jig. From the jig the
glass goes to an impact mill, where it is crushed to minus 20 mesh. The
slurry is pumped to the froth float arrangement. They have dispersion float
followed by a rougher. The product goes into a spiral classifier and the
gangue goes to a settling tank. The glass is dried.
23
-------
The light fraction is de-entrained and pneumatically conveyed to a
Heil secondary shredder, then to a surge feeder and pellet mill densifier.
Dr. Alter described some bridging problems with the cyclone, where the
light fraction hung up at the bottom. They built their own configuration
that did not hang up. Also, a hood was added to the Heil shredder, and
some air is removed and the dust recirculated. They found that at first
the hammers would not reduce the textiles, which jammed the densifier. The
hammers have been replaced with rotors and stators. The material is shredded
to 3/4 inch. Some textiles still accumulate at the bottom of the machine;
these are cleaned out each morning.
In answer to a question, Dr. Alter said that most of their equipment
is commercial size. The trommel is undersized because of space constraints.
Dr. Alter then described the densifier. It is a 30-inch die with about
1300 one-inch holes. The pellets are extruded through the die. He described
some problems. The machine has to be started very slowly or it jams. They
will try another diameter rotor to try to avoid jamming.
Dr. Alter mentioned the large amount of maintenance required in a plant
of this sort. It is time and money consuming.
Bureau of Mines. The Bureau of Mines pilot plant was shown and de-
scribed by Roger De Cesare. Their program began in 1965. Early research
centered on municipal incinerator residue, and the residue proved amenable
to conventional mineral dressing techniques. The pilot plant has been
operating since 1969, using conventional off-the-shelf equipment. As the
number of municipal incinerators decreased, the Bureau began to work on
separating fractions from raw municipal wastes.
At the pilot plant, material larger than 2 foot is picked, smaller is
processed. A flail mill efficiently breaks open bags of refuse. Massive
pieces do not damage the flail mill. The mill required an exhaust hood, so
this was utilized as an air classifier to remove about 25 percent of the
light fraction. They use a series of air classifiers developed by the
Bureau.
De Cesare briefly described the early incinerator residue plant. It
uses a wet process, with washing, screening, and shredding operations. Mag-
netic separation is used, and nonferrous metals are concentrated on a screen.
Aluminum is separated from copper/brass with heavy media or by jigging.
De Cesare then described the raw refuse plant in some detail. Most of
the machinery is scaled for 5 tons per hour, although the flail mill is
larger. Material goes from the air classifier mentioned earlier to a mag-
netic separator, where they get only around two percent paper. The material
then goes through an air stream, where lights go into a cyclone. Heavy
nonferrous items drop into a trough. There is a medium fraction—rags,
cardboard, glass, aluminum, wood, food, etc.—which is put through a trommel.
Glass is removed through 3/4 inch holes. The organics in the glass are
24
-------
separated by jigging. The ceramic and stone are removed by a fioth flotation
process, so that mixed glass cullet is removed. Material larger than 3/4
inch passes from the trommel through a secondary shredder and into the third
air classifier, a three-stage aspirator. This classifier is tuned to drop
out aluminum cans. They are recovering 95 percent of the can stock and 90
percent of the foil. The heavies from the three-stage aspirator are dried
and put through an electrostatic separator. They are also testing Raytheon's
nonferrous metal separator.
De Cesare pointed out that their research is low budget. They test
refuse for municipalities planning resource recovery operations. They will
also treat wastes from industrial facilities. The Bureau's policy has been
to develop pilot plants, then assist the private sector to develop full-
scale demonstrations.
University of California. Dr. George Trezek then presented work being
done at the University of California. The project was started for EPA,
first doing shredding experiments only. The refuse is received from the
University (mostly paper) and from a local sanitary service. Refuse is fed
into the shredder with a simple front end loader. They have a plywood air
classification system which they built themselves. The heavies come out
underneath the air classifier and are conveyed past a magnet which pulls out
the ferrous scrap. That stream then goes through a trommel and the under-
size goes to an air table. A drum magnet removes small screws, nuts, etc.,
from the air table. By going back through once more, they get a 99 percent
clean glass fraction, which has been used for building blocks. The light
stream comes through the bottom of the cyclone. They have had some problems
with the air lock feeder, which they have attempted to solve themselves.
They also have experimented with screening. The screened material goes into
a baling machine. Roofing companies have used this material.
Dr. Trezek described experiments with the minus fraction from the
trommel, which normally would be landfilled. They have digested the ma-
terial with sludge. In another experiment they have made pulp from the
screened light fraction by using a plastic removal dilution operation. The
plastic removal operation involves a traveling screen; jets of water flush
the pulp through the screen, leaving the plastic on top. From the pulp they
did produce paper. Also, its heating value is 7300 Btu per pound.
A new project at Dr. Trezek's laboratory is studying combustion aspects
of densified fuel.
U.S. Navy. Steve Hurley of the U.S. Navy spoke about his work under the
DoD program. They are planning to follow the EPA Guidelines and utilize the
fuel value of solid waste. They have many options under the regional con-
cept, in that some installations are near large municipalities they might
join; others must go it alone. They are trying to bring about cooperation
between the Navy Environmental Protection Program and the Energy Conser-
vation Program.
25
-------
IIUTJL;;' ,<;•<] ,v ;-, 1 j.ns Co build a full-scale experimental unit in the
Norfolk area to determine the problems, costs, etc., to make an RDF. He
said rult ,; have been chan;;fd so that an installation which practices re-
source a..:c-' 'i v can cttrieve the funds; this will give some incentive.
The Defense ; \: '-vies agency has responsibility for doing the marketing.
There frilowed some discussion of the Federal Facilities Guidelines.
Robert Freeman explained that if all Federal facilities in an SMSA generate
over 100 cons a aay oi; solid waste, the largest Federal facility would be-
come the Lead a&ei-cv for administering the solid waste. They may utilize
an existing resou.1:;:.- recovery facility or develop a new one. Hurley added
that at Norfolk the option is being kept open for others to use the facility.
Dr. Greg Rigo commented that the Norfolk area was selected for the test
because the Navy does not have an unmanaged solid waste problem there, there-
fore they can test and not be forced into a production mode.
Toronto, Ontario. G. C. Chisamore made a presentation on an experi-
mental plant being built by the Ministry of Environment in Ontario, Canada.
The Ministry is a regulatory agency and also a funding agency. Three years
ago a cominic.ment was made to build a solid waste plant. They are using off-
the-shelf equipment. The. plant can be shut down to determine efficiency or
make changes. They have a 600 ton per day transfer station to handle the
waste if the plant is shut down.
At the pL'erieiit time they are exploring both energy and material re-
covery. Many industries have expressed interest in RDF, and they will try
to provide samples. They have an agreement with a private cement company to
use the light traction, Also the plant will be used as a training facility
for operating scarfs.
The plo.ut ha.<-> a small Consumat system for energy recovery. In addition
there is a 5U ton per day mechanical composting module.
The plant is ;=et up on a modular basis, and any manufacturer can test
his equipment vhtrt.; the information will be public knowledge. The plant
is designed for public f~->ars.
In answer to a question, Chisamore said that there will be extensive
monitoring o -.missions f"oin -he plant.
Cliisamore was asked now his plant differed from others. He said it is
basicaJi\ ";he same. However, they hope the private sector will take streams
they Dru1'-1 i-_- and reoGYKi; Llie fractions of value. This would minimize the
Provi ucv;':. r tnanc i.t i invo I Yement.
26
-------
SESSION 5: SUMMARY AND CONCLUSIONS
David Berg, moderator, opened the session by asking, why are we here?
He said that we have a fuel crisis and at the same time environmental
problems. An expanded use of RDF and materials recovery could help. We
are trying to telescope the birth of an industry to perhaps 10 or 20 years.
Berg noted an interplay during the discussions of near term problems
and longer term optimization. A lack of data at all levels is impeding
implementation in the near future and development of new unit processes and
systems. We need comparative design data. We need design handbooks for the
short term, orderly R&D for the longer term.
Session 1 - Systems Process Design
Berg felt that an outstanding point was that there is no state of the
art. An extreme lack of information was noted, especially in four areas:
materials handling; scale-up of equipment; quality and composition of raw
refuse; and specifications and requirements of fuel produced. Sequencing
of unit operations and modeling of alternative systems were also important
topics. Trade secrecy and proprietary information were felt to be barriers
to information flow.
At that point Dr. Harvey Alter said that trade secrets and proprietary
information do not have to be an impediment. Comparative data, at least,
could be advanced. Berg partially agreed, but said there would be reluc-
tance to expose systems to comparative testing. Alter felt that as de-
velopers gain confidence, they will share information.
Session 2 - Selection of Equipment
Materials handling, size reduction, separation, and process control
were discussed. Again, scarcity of data was noted. There was no general
agreement on how much data is needed, or how to get it, in what form.
Equipment has been adapted from other industries and is suboptimal. No
major problems that would prevent plants from being built were identified.
People recognized a need to understand basic parameters, as in George
Trezek's work on shredders.
Berg briefly summarized comments in the four areas mentioned. He said
there was agreement that waste handling is difficult, it is an art. Equip-
ment choices for storage and retrieval are limited and are adapted from
other fields. For size reduction, there is more experience with shredders
than any other type. The state of knowledge is poor. With respect to
separation, there was uncertainty about air classification. In the areas
of process control, dust was of great concern. There is a need for a design
manual for shredder explosion control.
Dr. Greg Rigo pointed out a need for reflective work to come up with
something new.
27
-------
Session 3 - Technical Obstacles
Here Berg again pointed out a duality between the need for better
information on existing equipment and the need for better unit operations
and systems configurations. Anything we can do to reduce technical ob-
stacles and risks will be helpful.
Berg summarized some points made during the session. The type of data
needed on performance differs according to the audience. Data may be ob-
tained from full-scale plants or separate unit operations, and there was no
clear consensus which is best. Responsibility for obtaining data may be
relegated to the suppliers through performance bonds. There is a need to
coordinate the gathering of data for comparability later. There was a
desire for better design handbooks.
Berg then reported the results of the workshop members' vote on pri-
ority research work. The item which received the highest number of first
place votes was Determine Optimal Arrangement(s) of Unit Equipment. The
topic with the largest number of total high priority votes was Compare the
Performance of Various Types of Shredders. Other high priority items were:
Study Emissions from Processing Equipment, Evaluate Other Potentially Ap-
plicable Equipment, Determine the Effects of Pretrommeling, Develop Effec-
tive Accounting Method(s), Determine Equipment Operating and Maintenance
Costs, and Determine the Effects of MSW Characteristics on Processing. The
last-named also was a low priority item.
Most often named as of low priority was Determine Effect of Magnetic
Separation, Others of low priority were: Determine Effects of MSW
Characteristics on Processing, Evaluate MSW Segregation Prior to Processing,
Evaluate Receiving Facilities, Evaluate the Effect of Drying on the Com-
bustion Characteristics of RDF, and Determine the Effect of Drying on
Bacteria and Virus.
The category with the highest priority rating was General, followed by
Economics, Shredders, Air Classifiers, and Storage and Retrieval. Attracting
the least interest were Receiving Facilities, Dryers, Magnetic Separators,
and Conveyors.
Berg mentioned that immediate and short term needs dominated the dis-
cussion, but there was recognition of the need for basic parametric research
on theory, unit operations, and systems configurations. Finally, it was
suggested that we form four ad hoc committes to explore methods for eval-
uating unit processes.
Session 4 - Approach to R/D&D Programs
Berg made five points about this session. First, scale factors are ex-
tremely important to those using research data. There is a need to exercise
all modes of R/D&D from pilot scale through full-scale demonstration.
28
-------
A third point was the duality of preference on scale for gathering
process data. Both large scale and small-scale operations were supported.
A need was expressed to look at industrial boilers as a market for RDF.
Another point was that many options should be explored and none cut off.
Possibly even some older systems should be looked at again. Finally, Berg
said that there was no clear consensus on how the government should conduct
its research programs.
Peter Ware brought up the subject of motivation to do research, which
he said was money. He felt the large companies which have the resources and
the profit motive will do most of the research. Ware felt the Federal
government should fund research in areas like disease or other potential
problems in the plants. As for forming committees to develop comparative
methods, he agreed, but pointed out that the companies cannot be forced to
comply. Ware felt there should be more workshops with specific goals.
The moderator asked for more comments on what role of the government in
research should be. Floyd Hasselriis felt the government should help with
government-created problems. Examples would be pollution and health issues.
Berg replied that the government did not create the problems, but rather
identified them and suggested remedies. He agreed it was an option for
research.
Ted Sjoberg said the government should provide incentives for the re-
source recovery industry, municipalities, and power companies. He mentioned
tax incentives specifically.
Don Walter pointed out that resource recovery is marginally economic.
The government should not finance something that is going to make money.
Sjoberg replied that environmental impacts may require resource recovery,
but it is a speculative business. Walter suggested that the market is too
small, but Sjoberg felt that rather it may be too big and speculative.
Concluding Remarks
Berg again brought up the discussion of ad hoc committees. Dr. Alter
said that he had been talking with some other members of the workshop, and
they felt that ASME and ASTM provide a suitable mechanism for this activity.
Dr. Alter suggested using these committees which already exist. Workshop
members were especially invited to participate. The group concurred with
this suggestion.
Dave Spencer expressed regret that the resource recovery field was not
moving ahead faster. He felt that government money was being wasted, with
no cohesive effort. The companies are called upon to invest a lot of money
with no guarantee of a high return. Dr. A. W. Joensen added that money for
research really comes from Congress. He suggested working through members
of Congress.
29
-------
Bill Ryder said that the government had not helped with the Chicago
plant because it was not experimental. He felt that Federal money should
go for constructing known on-going processes, as well as research.
Sjoberg suggested that the government help with financing in ways like
loan guarantees. He also said that the government agencies should get to-
gether and present a united front on recommendations to legislators.
Walter asked how many cities felt a loan or bond guarantee would be use-
ful. No hands were raised. Spencer commented that grants were better than
loan guarantees. He felt no policy decision had been made to move resource
recovery ahead quickly and provide the resources.
Dr. Alter said that we had not considered the demand-pull financial in-
centives that might be used.
C. W. Fay said that it was more important to protect utilities from risk
than to assure that they would make money.
Dr. Rigo said that the pressure of getting rid of solid waste as land-
fills are outlawed, incinerators shut down, will provide an incentive for
resource recovery. Cost is immaterial if something has to be done.
Tom Lamb spoke against artificial stimulation of the industry by govern-
ment. He said that the system will work when it is the right thing to do,
and that time is almost here. Tax credits or loan guarantees may possibly
be necessary.
Berg concluded the session by thanking those who had organized the
Workshop and those who participated.
30
-------
VERBATIM TRANSCRIPT
BURCKLE: David Berg of the Office of Energy and Minerals Industry, Office
of Research and Development, will give us the opening remarks.
BERG: I have been working in the solid waste business for a little over
three years now. Although we have meetings all around the country
sponsored by various organizations, we really never get the chance to talk
to the people who are most centrally involved in the research and develop-
ment in the architectural and engineering community, and in the cities, and
among the companies that are most involved together to talk about what
their problems are and what their perspectives are. I hope that today is
going to be the first of a series of meetings like this. I am very
confident about resource recovery. I think that all of us at EPA recognize
the need for technical gatherings in situations like this. Where we will
begin I don't know. We can talk about that at the end. The Office of Solid
Waste and the Office of Research and Development cooperated to put together
this meeting and I would mention that because together they represent (with
ERDA and other agencies) the entire Federal spectrum of activities in the re-
lationship of solid waste. The research and development people, of course,
are looking at it through the perspective of development systems and unit
operations and what have you. With respect to resource recovery and the
solid waste offices, we are groups inside the Federal government helping com-
munities with their solid waste management problems. Today with this group of
people we have here, we hope we have all the perspectives in the room and can
bring out the issues that are related to solid waste and processing of solid
waste. The last thing that we should do is try to be nice to one another, I
think that the more [unintelligible] to be made the better we will all feel
at the end of the day. Why is EPA doing this, besides to get everybody to-
gether in a room to fight? Besides this, we are responsible for looking at
things from a Federal perspective. I think we can see that there are a lot
of needs, as I said in the beginning, that aren't being satisfied, although
we have several publications now in which papers are being presented. It may
be that we need to have additional publications. For example, we need to get
specifications together for community review. If you want to put in a sys-
tem, what kind of shredder do you need or how do you decide the thing? If
Union Electric is going to put some RDF in their boilers, they need to be
able to work with somebody who knows specifications [unintelligible] to go
into their boilers. The people in the public community systems or privately
developed systems need to ask for guidance in terms of what there is de-
veloped or what's needed. I hope that we can touch on those topics and
others like that and perhaps lead to some sort of example of what other
facilitating mechanisms need to be put together so that the country will
have a community that facilitates debate and discussions and resolutions of
these various issues. The agenda is straightforward. I think you all got
that in the last letter that was mailed. The first topic is system process
31
-------
design. The second is selection of equipment this afternoon. Tomorrow we
have technical obstacles on the agenda, and the fourth, approach to R/D&D
programs. At the end we will spend about an hour summarizing what we talked
about, maybe some sort of notes for the future which will appear in the
proceedings that will be published at the end of this thing. Where we have
found room for agreement, maybe we can put together some conclusions.
Thursday we have a site visit to Resource Recovery 1. NCRR is here to host
us, as well as the City of New Orleans and Waste Management. I would like
to say in closing my personal thanks for everybody for being here. I hope
that the discussion will be as good as the response. In very short order
39 [unintelligible] became 44 people present. Only five people invited
were unable to attend, including people from England, and they had a long
way to go. The last thing I would like to say is that there is one big
complaint that I have about preprocessing as a topic, and that is that we
don't have a good term or title. Preprocessing leaves something to be de-
sired. It implies that it's really a secondary or primary way-stop before
something else. It seems that we could come up with a term that would be
a little more representative and make this topic be something less than a
stepsister in its title. I'd like to announce a contest. Free drinks to
the winning contributor of the new name for preprocessing—the award to be
provided at the end of the meeting if we can reach an agreement at the end
of the summary and conclusion session. Thank you.
BURCKLE: It seems that we have an entire technology in search for a word.
We are going to put together a complete proceedings and, of course, this
is going to take a lot of cooperation on the part of the workshop members.
In doing this I would like to say that we are talking on the record, and
we are also making an effort to provide comments off the record. If you
are going to say something which you don't want to appear on the record
please so state. The purpose of producing the proceedings is so that we
have a record that we can use to help us formulate R/D&D programs. Also
the proceedings will be distributed to all members of the workshop. It's
not going to be something that is going to be tucked away and disappears in
some dusty archives. That's why we WHnt to have the opportunity to make off
the record comments and remarks. Our first session will deal with the sys-
tem. I have to use the word preprocessing, I don't have a better word as
yet. Now as best I can tell, the word preprocessing is a contraction. At
one time, I think back in the early seventies, we used the word preconversion
processing, which meant the processing that came before the conversion to
energy as we have used postconversion processing, which is the conversion
after energy recovery, and it got contracted down. So much for my word
science today. I would like to introduce Al Klee who will be our moderator
for the first session. Al Klee is the chief of the Processing Branch of
the Solid and Hazardous Waste Research Division of the Municipal Environ-
mental Research Laboratory of the Office of Air, Water and Land Use of the
Office of Research and Development of the USEPA. Which is why it takes us
so long to get any work done because the titles are so long. Thank you.
32
-------
SESSION 1: SYSTEM PROCESS DESIGN
Dr. Albert J. Klee, Moderator
KLEE: We have two major objectives today. The first one is an attempt to
determine what information is needed or wanted. Given the equipment we have
today, how do we select the system? What information do we need to design
more efficient systems? Well, this is a matter of what information do we
need to do those two things. That's the first objective. The second ob-
jective is an attempt to identify the critical aspects of design. Now in
preparing for this session, John has asked several questions he thought
would be at least nuclei for attempting to approach these objectives. The
question is really three questions, and one of them can be broken down into
two. One question involves the end product. Can we produce something use-
ful? Can we produce something useful with what we have today (we are talk-
ing about systems now)? A second question: is there active information
available for users of systems, and can one go out and select from what is
available from the industrial offerings that are on the market today? A very
related question has to do with those who design systems. Is there enough
information now to design systems with some stated degree of reliability?
What sort of guarantees can we place on design? And the fourth question, a
very interesting one, involves the alternatives; can we take what is there
and rearrange them; can we change the configuration; is there something
coming down the pike—some new piece of equipment that might enter into con-
sideration? These are just thought questions. If we address ourselves to
these questions it might enable us to better approach the objective. The
objective is primarily to determine what information we need to use what we
already have. What information do we need to design more efficient systems,
and what in general are the critical aspects for process design? The topic
suggested, then, is specifically a summary of what I've said. What are the
requirements of the product that affects the system design? What about the
system liability guarantees? What about the selection and sequencing of the
operations? Incidentally, we're talking both about the effect on quality
and cost as well; we don't want to forget about cost. Energy efficiency is
an important aspect here too. We will mention economics later when we talk
about systems process design. Let's discuss our ability to do heat and
material balances for design. And let's finally summarize what we know about
present designs and what the present information is for selecting these
systems. With that in mind, we have stated our objectives and in general
have provided each and every one with some lead questions. It is up to you
whether you want to address yourselves to those questions to get to the ob-
jectives. They are just an aid, i.e., some key topics. We would like to
open this section to the conference members. It makes no difference who
starts, but let's just proceed on an informal basis, seeking out any one of
these question areas or a new question area that you feel might contribute
to these objectives. Well gentlemen, do we know everything we need to know
about the systems process design? I doubt that very much. Greg, why don't
you start?
33
-------
RIGO: We just finished a parametric tradeoff design study for the Navy.
Steve Hurley is the ultimate sponsor. The study approach is interesting.
We took the problem, what do you do with 20 tons of garbage a day? Then
we said, let's assume we want to preprocess, and broke the processing down
into nine classes of modules or functions and assigned each of these modules
anywhere from three to five different approaches for accomplishing a function.
Examples of modules are size reduction or dirt removal or separation of com-
bustible and inert fractions, etc. We then went ahead and developed a very
rudimentary type separation model. We took each module using the best data
we could lay our hands or from nauv of you in trr's room plus several vendor's
representatives and tried to sin.uJ.i're performance in a primary manner. Now
what we did was, for example, go to an air classifier and :i-?«, where does the
paper report? Well, it goes basically in two places excluding the f-jitive
dust - some of it flies, some of it sinks. What are the spjJt cra/:= used
the Midwest Research dcta. We used .some data we've beon collecting in the
field. When alJ was said '-ind done, it turned out Lhac } cm can model, to
some extent, what happens in systems variations. We l:oui;-; i.riat today it is
impossible to come up wit;i (I) a coruci stcnt data set; (2) to account for
interactions between various coiuj. ontnc,.,, '•,''"'' it is very difficult co account
for what happens to certain components wn:~: you go through a module. A
prize example of this is glass. We all kuov when you hit glass with a
hammer, some of it goes to sand, some of it goes to cullet, and some of it
comes out in big pieces Well, we cannot yet account for, in any data base
we can lay our hands on, where the glass goes, and we have not been able to
account for what happens to it when you get to the air classifier. The
modules are strictly primary models, very rough models. What came out of this
study was a very, I think, interesting find;n;;. It turns out, as you would
expect, that ?0 tons a day is not too terribly cost-effective to process
unless you don't have any choice, e.g., Federal facility guidelines forcing
you to do it. However, if you can get some source segregation and split
the metals and glass from everything else, we came up with a departure ap-
proach proc€;ss1n,; .-..;•, r. eg aft d streams, which appears to be, I am going to say,
cost-effective i/a about: '-he 50 ton plus range. Essentially what I am saying
is that we just finished attempting to model systems. We have just attempted
to finish a parametric study. We came out with some numbers, but accuracy of
the entire system was so poor, that I'm not certain what the numbers mean.
KLEE: Gre,j,, cue asv/ect of modeling which has long been an interest of mine
is a verificatioaal modnj . How In the world can you verify? Anyone can come
up with a modeL but the question is, is this model useful? One has to verify.
How do you verify dueh a model?
RIGO: Essen':i,"' • ..i>'e going to develop comparatively complex design
procedures ?u a:_ i Lindy 01 unknowns. These procedures, after you've got them
laid out. ;.. ?.u t.'-ti; b^ -=i r
-------
models, which are the ones that are used in related industries. Then you
use the available data to develop the parameters that fit these modules.
If you cannot get the data to fit these models, you know your model is no
good and you start over again. I think it's possible, with some of the data
that's out there now, if we've got a good design procedure development pro-
gram, to use the available data to validate or reject these procedures. I
think that if we just simply go out and measure things and not have the
collateral design development program, we are wasting money and time, and
we're certainly getting a lot more data than we probably need. In fact, it
may not be the right data.
KLEE: I think it is interesting to open the discussion talking about model-
ing processing systems because one thing about modeling is that you soon
find out what information is missing. I think that I'd like to ask MRI to
say a few words about their study of preprocessing as it applies to systems,
just to further set the basis for discussion. Dave, I would appreciate it
if you would say some words about the study, particularly as it applies to
system design. This will help set a background.
BENDERSKY: We started this project for MERL Cincinnati in March 1976. The
purpose of this program is to first of all determine the state of the art
of preprocessing and to lead us into other phases of our project which deal
with the collection of missing data. The first phase of our project, which
we have recently completed, was to assess the present state of the art and
to identify the research needs in this area. The results, incidentally,
were recently put together in a rather thick report. We condensed some of
the information in this brown folder for you to use as reference material.
As far as the systems themselves are concerned, first of all we are con-
cerning ourselves with waste-to-energy systems and those waste-to-energy
systems which process the waste in some manner before its conversion into
an energy product. We are not talking about waterwall furnaces, etc.,
which do not process the waste before use. The first thing we did was make
a survey of those kinds of systems that exist now and those that are planned.
What we found is that approximately 50 systems—that is waste energy systems
with preprocessing—are either now existing or in the planning stages.
That's the base that we have to work with. Of these 54 systems, approxi-
mately 12 to 14 are scheduled for operation this year, in other words by
January 1978. As you are all aware the dates are the best that we've gotten
from the people who are responsible for these systems. There are three
groups of people in this industry that we're trying to satisfy. First of all
we are concerned with the system designers—the people who are designing
these systems. Secondly, we are concerned with the manufacturers of equip-
ment; and finally we are concerned with the user/operators of this industry.
We have interviewed representative groups from these three basic sources. As
for an overall picture of what we have done so far as to the state of the
art and needs for research, we find that all three of these groups need
additional data beyond that they now have. The system designer needs more
information on equipment alternatives and arrangement alternatives, not so
much from the manufacturer of the equipment but how it will actually work
in the field on MSW (municipal solid waste). As we all know this is a
young industry. There is as yet comparatively little data based upon actual
35
-------
performance in the field under full-scale conditions, so the industry in
general is operating with lack of data simply because the industry just
hasn't progressed that far yet. More field test data is required under
full operating conditions. This is a generalization. Specifically, there
are problems here that evolve into what we call cost effectiveness. In
other words, you need more than just how well will a piece of equipment
perform, plus under what price tag. As we all know there are tradeoffs
in this as well as other industries. The old saying is you get what you
pay for. Unfortunately right now there isn't enough data for comparative
analysis if you want to put together a system and do a detailed cost analy-
sis and feel real confident that you are designing an optimum cost-effective
system. These are some of the problems that the systems designers face.
There isn't enough data. Now they are using all that they have available,
but unfortunately there isn't enough yet. We are finding that the manu-
facturers of the equipment are handicapped. They are trying to provide
equipment to the systems designers and the systems users, but they are
handicapped because they aren't sure first of all under what conditions
their equipment has to operate in the plant. Secondly, there isn't as yet
enough intercourse between the manufacturers and the users and systems de-
signers. There are a lot of personal contacts going on, but as an industry
it hasn't yet started to mix enough so that there is open literature—open
data—that everybody can use. Everybody is exchanging this kind of infor-
mation on a personal level now. There has to be more of this kind of in-
formation gathered for the whole industry. Manufacturers are interested in
how their equipment is working in the field; they don't always get this
information. I have asked a number of them "how is your particular piece
of equipment at such and such a plant working?" Well, some manufacturers
make more of an effort to get this information than others. Some of them
say we really don't know because things change from day to day out there.
There is still not a good exchange of information. The manufacturers are
handicapped basically two ways. One, they are not sure what their equip-
ment is supposed to do under what precise conditions. They don't know how
their equipment relates to an allied piece of equipment—a shredder versus
something else down the line, or other piece of equipment with regard to
another. The interrelationship or the arrangement of equipment and the
effects of arrangements are not available yet. So these are some of the
problems that the manufacturers are dealing with. The users, we find, or
the operators of these systems, are handicapped because they don't really
know when they are given the keys to the plant how to operate that plant
under optimum conditions. There's a lot of worrying going on and has to go
on when any new plant comes into being. What it means is that a lot of
self-education is going on. The operators of the plant have to learn them-
selves how to operate the individual pieces of equipment and the total
system efficiently. They wish they had more instruction books or knowledge
so that they wouldn't have to go through this extensive, costly learning
program. Many of them have to change things as they go along in order to
make their systems more optimum than it was originally designed. These
are some of the problems we have found by looking into this field as far
as systems design and operation is concerned. I will go into more detail
later, but Al, that is a sort of summary of what we found as far as the
system is concerned.
36
-------
KLEE: Thanks, Dave. We found that coordination was needed on equipment
alternatives, and coordination of equipment manufacturers with system
designers. People operating these plants certainly need more information,
not only on the equipment but on the system as a whole for more optimal
conditions.
ALTER: Let me suggest a little reorganization of what Dave has said. There
now exists the literature in other fields, particularly minerals processing.
Given a piece of equipment to operate over a range of particle sizes, with
a variety of feedstocks, to perform in a certain way, you can interpolate
and come up with some projected performance. Appropriate assembled data
do not exist. I suggest that this organization meeting will lead to pro-
viding data, but on what pieces of equipment? Dave organized the materials
handling into size reduction and into separation. It's interesting to me
the materials handling seems to have been ignored. When the operator goes
into the plant, and strives for optimum performance is when he gets hung up
on the materials handling problems. He never gets around to research
aspects, or if he does, he's too busy trying to make things operate that the
measurements aren't made or aren't reported. Information is transferred by
conversation not by the written word. I'm appalled at the lack of infor-
mation—a "no state of the art" if you will, or "poor state of the art"—
of materials handling in refuse processing, the dust and the blowing and
the dribble, and the consequent housekeeping problems and fire hazard. I'm
appalled by the fact that it seems that many plants do not know how to mate
a conveyor to a shredder, or a conveyor to a bin. I raise this to the
group because I've seen overseas beautiful, clean plants, no blowing, no
dribble, no dust, clean floors without large crews with brooms. Those of
us who have been in plants throughout the country see quite the converse.
I think that before we can learn something about size reduction and separ-
ation on the operating mode, we will first have to learn how to handle the
material so as to avoid startup problems, to avoid housekeeping problems,
and avoid having to account for the portion of the materials balance flying
all over the room. I suggest that we try to hold the discussion to maybe
these three categories—or others which people may want to suggest.
KLEE: Thank you, Harvey. You're identifying one of the critical aspects
of the system design—materials handling. To my mind, we can separate size
reduction and separation into three categories—mechanical, pneumatic, and
an "other" category.
BENDERSKY: We in our work organized the equipment into two basic categories.
First of all there is what we call unit operations, which are those pieces
of equipment which actually change the character of the waste stream. I'm
talking about shredders, air classifiers, magnetic separators, etc. Those
we call unit equipment or unit operation. Then the second classification
that we have is supplementary equipment, which includes the conveyors, the
storage, things that do not affect the character of the waste but are very
important from a standpoint of handling the waste stream. Incidentally, on
page 10 of this handout we have a list of the basic pieces of equipment we
have been looking into and they are identified by these classifications.
37
-------
There's a dozen areas we looked into—all the way from receiving facilities
down to fire control systems, and they roughly spread into about six unit
operations and about six supplementary, which are all-important to the
operation of the entire system. There's been a tendency to ignore the
handling equipment and some of the very important aspects. It doesn't do
you any good to have a shredder unless you can get the material to it.
LEVY: I would like to get back to Harvey's comments with a couple of my
own comments and maybe we can follow throughout the whole two-day session,
and that is to get back to what Harvey suggested. I would like to raise the
question how could we find out this information. Maybe we could have a
discussion on what kinds of steps we should take to answer these questions
and evaluate the existing systems, or paper studies or comparisons of other
industries or laboratory comparisons. What specific steps could we take?
I think we get into that and elaborate at this point and get into some
specific tasks that we can actually sink our teeth into.
BURCKLE: I'd really like to reserve that sort of discussion and type of
approaches to developing the information and satisfying the informational
needs.
KLEE: Does everyone agree with Harvey that materials handling is the
critical area?
CHANTLAND: I would have to agree with Harvey that materials handling pro-
bably is one of the major items in any processing system that has to be
given great attention and probably is one, certainly from the aspects of
investigation so far, that is extremely critical. It's critical from man-
power, it's very critical simply from running an operation. I think, how-
ever, when we started out here I have to back up to what has been said,
what is the end product. I would have to say that question probably is the
most important and must be answered if you are going to talk about the
selection of systems. That's the one where the economics is; that's the
one that determines basically what kind of equipment and so forth you are
going to put into a facility; that's going to determine whether you are
going to have some pre-separation that we talked about here before. It
goes clear back from one end clear to the other. If you don't satisfy what
the end product is going to be, as an example, you may be able to put up
with some glass or some rubber in some cases, maybe you can even put up
with wood particles - large or small. In other processing you can't put up
with this type of material, so there is an importance there in terms of
equipment selection and the selection of the system from that point on back.
LINGLE: Let me follow up on Arnold's comments and pose the following ques-
tion relative to producing a marketable product. The question is, how much
processing do we have to do to produce a marketable RDF? The early attempts
at producing RDF, and I guess the attempts that are still going on now, is
basically one- or two-stage shredding, plus probably single-stage air
classification. When we look at the composition of the RDF, we can look at
its performance in a boiler, we see that we don't have a very sophisticated
38
-------
material and that may be one of the reasons that we find the utilities very
reluctant to purchase that material. For example, one of the things we find
is fairly high moisture content in RDF, and that moisture content causes in-
creased gas flow to the ESP, and that in turn results in increased emis-
sions. That was one of the conclusions and results of the testing in St.
Louis. We also find that there is a significant increase in bottom ash
when you burn RDF in the boiler. That's partly because the RDF has a high
ash content, higher than coal, and partly because the particles simply
don't all burn. Now, if we want utilities to burn this material, we may
have to do something to improve the product. How can we do that? Dry RDF
to eliminate some of the moisture; use additional screening or classifi-
cation or other processes to reduce the ash content somewhat. In short,
we're going to have to move to processing systems that will produce a
higher quality RDF.
WILES: I would like to ask a question on the other side of that point of
view, Steve. Perhaps the utility industry could use a lesser quality RDF
rather than what you are talking about with some somewhat minor modifi-
cations of the system. Maybe that's a more cost effective way to go. I
would like to have some input from participants in that area.
RIGO: I think the question cannot be how high a quality of fuel do we have
to produce to sell. I think it's got to be how poor a fuel can we get away
with making and use it successfully. Because every time you process, your
cost goes up, your reliability goes down, just backing up what Carlton said.
So those of you who are playing around with low quality fuels, maybe we can
look at it from that point of view, how low quality can we go? We're burn-
ing pellets at Hagerstown right now and we are getting deteriorated fuel on
occasion—not very bad, it's about 16-17 percent ash. I have seen pellets
produced with 12 percent ash. The fraction fines range from 15 to 50 per-
cent. The pellets still go in, and guess what, we can run a boiler on them.
We have burned 100 percent pellets, and experienced some derating, but no
major problems. We burn less, we get no derating. We don't have a lot of
time burning d-RDF yet, nowhere near what Ames has. But to date, we're
fairly encouraged, we are working with Harvey Alter now and getting ready
to start with him trying to answer the question—how low quality a fuel can
we get away with?
HOLLANDER: Mr. Lingle mentioned the report from Washington; he put it in
the context of the utilities. North Dakota lignite, Wyoming lignite, the
subbituminous coal, all have more moisture than most RDF generated anywhere.
And yet, there is this urge to embrace these lower Btu, high moisture,
higher ash coals, merely because of their lower sulfur content. Actually,
when we look at a fuel product there are two basic aspects instinctively
considered. What is it you would like to have and what is it that you really
need? When we look at coal burning, does anyone really know what is being
burned at any point in time? What are the variations in coal quality that
you must be able to cope with? There were periods recently at a station
that gray coal was being consumed or just because it appeared black it is
burned as a fuel. The system design and operation must have the capability
of contending with variations in coal moisture, ash content, including
39
-------
extraneous material as well as calorific value. Very little of the coal
used for steaming is beneficiated. Why refine the material—why massage
the material any more than you absolutely have to—for a particular use?
However, we must recognize we have various degrees of sensitivity to the
quality of the fuel material for the application and genuine need is very
site-specific. But in our passion to come up with the ultimate in product
quality, perhaps we are overreaching, which may not be universally necessary.
We should consider what Mr. Chantland implied—let's take a look at what we
can get away with or what we can tolerate, and not impose upon ourselves the
over burden of seeking the penultimate. Perhaps in that way we can have
higher availability and a more consistent reproducible product, which pro-
bably also is most cost effective. Now we can't write a book of operating
instructions for any new plant, either in the power industry, or industrial
utility. Many times it takes six months or a year to get a unit operational
and on-stream; and we are addressing an area of process activity where
supposedly we have all the answers. To fully expect that this is going to
happen in any waste processing plant—walking in and pressing a button, or
reading some sort of conceptual operating book, we are just deceiving our-
selves. We have to go through the effort of trying to get an operational
plant that is satisfactory for the particular purpose at this particular
location. In time you will develop specific normal and extraordinary oper-
ating and maintenance procedures. Unfortunately, if you go back to one of
these plants considered operational—go back six months or a year later—in
too many cases you have to try to find the same operating people you talked
to six months ago or a year ago. It is this on-going process of reeducating
and indoctrinating the new people coming on-board that is also a real serious
problem. There are very few plants where there is a deliberate, concerted
program for making certain that a new man who comes on-board is aware of all
the trials and tribulations experienced and the procedures to be followed
to operate the system in the manner which satisfies the current objectives.
FAY: I think, depending on which side of the coin you look, how bad can
you make it or how good can you make it—I think that at this stage in our
development we should not be pushing too hard on how bad we can make it. I
think what we don't need today is a couple of failures, or maybe I should
say a couple more failures. If the Milwaukee project should turn out to
be a lemon and the Commonwealth Edison project turn out to be a lemon,
there's going to be a terrible time trying to get utilities involved. I
think most people agree that the utility with large boilers is the logical
consumer of the fuel, and I think from the utility aspect there are zero
incentives to get into this business with exceptions. If the utility is
municipality owned, then they've got a mission I guess. But in our case—
in most investor owned utilities—it is not our problem. To get into it,
we are taking on problems and we're taking on risks and they could be
serious risks. We don't really know how bad those risks are today. We get
enticed into this business for various reasons, I guess, and what we really
need to do is have some real success stories. In determining what kind of
fuel you should be making, I think you will have to take a real close look
at where you are going to put it. If you are going to put it on a grate
stoker someplace, I think you can do a lot different than if you are going
to try to burn this in a small utility boiler, and a small utility boiler,
40
-------
I suspect, is going to be a lot different than a large utility boiler. I
think you are talking today about a customized kind of system that you've
got to be looking at, and I'm a little doubtful that you're going to get
all this nice data that we talked about a few minutes ago readily available.
Assuming that some of these turnkey operations that are going on are suc-
cessful in that they do turn out a product that is consumable, there are
probably some tricks of the trade, some little gimmicks they use along the
way that for competitive reasons they may not want to flush out for the
public eye and their competitors. So I don't think you should hope too
much that you're going to get all of the data from every project to make the
next one you build real easy. I think it's going to be a tough competitive
struggle.
KLEE: Between the last two speakers we have the classical difference be-
tween the buyer and the seller, or the producer and consumer, laid out on
the table.
BERG: I was just wondering—maybe what you need is a success, but what is
it you really need? A success tells you that you can do it. What would
make you more comfortable?
FAY: I think that from the utility aspect, success means that we are able
to use this fuel; that it's not destroying our boilers, it's not causing
pollution problems such that we have to backfit a plant at great expense or
shut down because we can't meet air regulations or water regulations. If
burning this material should reduce the reliability of our major boilers—
which in our case we are using the largest ones on our system—if we take
those off the line because of this process very often, not only Wisconsin
Electric but the Public Service Commission of Wisconsin and probably the
people of Wisconsin are going to be not very happy because somebody obvi-
ously pays for that in the long run. So we need to have a thing that hope-
fully would operate like we hardly even know we are burning a different
fuel—and maybe even a better fuel. If I can point five years from now and
say the slagging on the boilers is a lot less than it used to be, there's
an inducement.
HASSELRIIS: There was one philosophy many years ago which was the only
way to burn refuse is on a grate, and that's with no processing. We've come
down the road a long distance since that time and that type of processing
is almost forgotten today. The question would be—what's the other end of
the scale, and if it may be appropriate to make major jumps ahead. We are
doing this basically by jumping entirely across the whole span to a fuel
which is a dry powder of extremely uniform characteristics, which would
meet this request of a utility for a fuel is actually much more uniform than
the coal, which they are well equipped to handle. It remains to be proven
exactly if the economics of the processing goes that far, because there are
more steps in it. On the other hand, the moment that you make one major
jump, which is to go into the drying phase, you're immediately in a dif-
ferent type of fuel of very uniform characteristics. There is initial
processing and the result is a fuel of a very uniform characteristic derived
-------
from an input of extremely nonuniforra characteristics. One of the major
things I wanted to contribute at this point in terms of preprocessing is
that number one, if you want to make a general rule, the manufacturer can
deliver about half of what you promised us based on the current technology.
If he says that the mill is good for 50 tons, it is probably good for 25.
That is probably fair. If he says the conveyor is good for 50 tons, it is
probably good for 25. Basically this has gone right down the line because
our methods of obtaining bids and getting back information are idealized to
a point to disregard the variations of flow that take place for many
reasons—the variations and constituents. So what happens is the manufac-
turer says, well the mill, that's average what I told you, but it won't take
the peaks. The peak becomes the average for the processor. If the motor
will trip out at 50 tons, then that is the peak, but the average relating to
50 tons is 25. So day in and day out you get 25 out, but you bought a 50
ton machine. So that's something that I'm sure all the people who built
these plants have come to learn. I think it's time for us to address that
fact. The variations we have to live with within a diverse fuel are extreme
in comparison with other types of material. Some studies I have seen or
made myself, trying to find out what the variation of any particular material
might be. If you say plus or minus 25 percent in the characteristic and
nature of a material of any kind or any piece of equipment, that's probably
a fair number. It's all of that. In other words, whatever you designed it
for, it can be less or more 25 percent of that characteristic. That would
be in bulk or weight, in density, in glass composition or any other
characteristic you want to apply. Therefore what we are really designing
to is not a point like in a simple system that carries water, something
that you know the properties of. It is a wide band, it's a very thick wide
band with a plus or minus 25 percent, sometimes plus or minus 100 percent,
200 percent in the case of cans or glass or small constituents of the
system. You may have a huge variation in one constituent, and we have to
understand that all of our equipment has to absorb, like a great big bal-
loon, the residuals that result from their usual variations. How anybody
is going to know what the variation is going to be in a plant is, of course,
our big question. We know that industrial waste is totally different from
residential and you have to consider in what size units the variations come
in. Yet one complete truck load of roofing shingles, and that does not
average in any way. It's that type of a variation in taking your bulk
refuse which has to be considered. Yet in the end by sufficient steps of
classification, you can end up with something that is almost pure cellulose
and a good clean fuel product. The question is the value of that product
and how much machinery you are willing to spend to produce it reliably.
WISELY: I thought that it might be of some interest to the people here to
learn a little bit about the philosophy that prevailed during the very
initial St. Louis work. Both Steve Levy and Bob Holloway, I'm sure, were
involved to some degree in certain of the discussions that took place rela-
tive to the quality of the material and the degree to which it had to be
processed for burning in the Union Electric boilers. The first reaction
thac we. 3,ot prom UE was purely and simply this (and this is in complete
substantiation to what Herb Hollander said) that the variability of the
42
-------
coal which fired through their facilities was such that it really made
little difference as to what quality we were actually able to inject into
the boilers. In the first discussions we had, the general reaction was,
why even remove magnetic metal? We concluded that this was such a simple
procedure that it was ridiculous not to do so. You may remember that the
initial facilities did not even include an air classifier, and this was
purely and simply because we didn't even know whether the stuff was going
to burn when we blew it into the furnace. We thought it was going to, but
we weren't completely sure. The next step, of course, more of a necessity
than of a desire to clean up the fuel, was to install the air classifier
which was installed purely and simply as a means of making the operation
more reliable. The only reason that that had to take place was that the
pneumatic feeders would not operate when they got chuncks of material that
wouldn't readily burn, and it wouldn't actually go through the feeder. In
contrast to that, in various subsequent investigations, it has been quite
apparent that there are facilities that could take no material of very much
a lesser quality than that that Mr. Hasselriis has discussed. In certain
cases it may be very well that is the only type of material that might be
utilized as some sort of a supplemental fuel. There have been cases also
that we've run across, one very current one, in which the boilers have a
very limited bottom ash handling capability. In this case, in order to
make the material function as a fuel within the boiler, it is going to be
necessary to remove as much of the ash—probably more glass than anything
else—as we possibly can from the material. In short, as Mr. Fay has said,
it's going to be necessary to consider in every single facility that is
really given attention, the individual circumstances surrounding that indi-
vidual application. In many of the larger boilers, sure, you can fire 3/4
inch stuff with some reasonable ash content without the moisture content
getting so blasted high that you can't handle it. This gets back in the
materials handling problem, of course, and from that point on it's purely
and simply a matter of getting an agreeable, rational sort of a compromise,
really, as to what quality material you can actually burn. For the most
part the utilities, I think, have been very realistic and really not too
demanding as to the quality of material that they are willing to take.
Obviously they want the best they can get. They also recognize, of course,
that it is not always possible to get the quality they would like.
KLEE: We've made a number of points this morning; two seem to stand out.
Of course, the big one was the matter of looking at the final product. In
this particular case looking at the fuel and realizing that you just can't
design a system unless you have some end product in mind. The other, and I
think we discussed that to an extent, that some of these other issues ought
to be looked at also. Harvey brought up the matter of materials handling.
Everyone seemed to agree that this was a problem. [Unintelligible]. After-
wards air classification was added. Back in those days you pretty much
could sequence. You knew what a preprocessing sequence was like. I doubt
if you can do that anymore because there have been other things added;
there has been two-stage shredding now. The use of trommels now, for
example, has come into the fore to a point where some systems—the systems
43
-------
over in Italy, for example, that hardly uses any shredding at all—they rely
to a great extent on trommeling. Now I've looked at all of these various
reports generated about processing systems and they differ. People make
different selections about which item comes first. Here's a good question:
what have we learned about sequencing? We get in there and we grind first,
or do we do some trommeling first, do we try some sort of separation first?
What have we learned about this? What don't we know about this particular
area? If you have some contribution to make in either the area of materials
handling alone or the sequencing operations, I think it would be valuable at
least to say something on that, albeit if we think that the fuel question is
more important we could go back to that. I would not like to see some of
these other important areas overlooked either. Harvey?
ALTER: Let me address almost all of your points. I like your point about
the sequencing because it is the first time it focuses on the need for
customers' specifications for the fuel or whatever else the product is. I
would like to suggest that the recent plants—many of the recent RDF plants
(to dwell as Steve did) copy from other people. I do not denigrate the
designers for doing this. The municipal customer and the bond counsels like
it if it's like everybody else's. They want to go into the showroom and
kick the tires, say paint mine green, and we wind up pretty much replicating
the historical development in St. Louis. I noticed our conversation this
morning was following a similar line. We started talking about plants as
they now exist and mostly exist—the shredding and the air classification
sequence, etc. We ought to go back, though, to customer specifications and
objectives, processing objectives, and then talk about sequence of oper-
ations. Example: It is certainly an objective in making refuse-derived
fuel to reduce the ash. We have one base line level of ash as a result of
the early work in St. Louis, and now Ames, and soon elsewhere. It's his-
torical and a lot of people said it's too high. Let's examine the refuse.
I think there are three different types of ash in the refuse. There are
the large items, the stones, the cans, that won't burn. They have to be
removed. There's the second type of ash, which is if the material is
shredded and some of the non-combustibles are crushed and thus coat the
combustibles, a second kind of ash. By the way, in either category I will
include hunks of rubber, leather, or heavy plastic which don't burn, not
in the sense they wouldn't burn in a laboratory, but their rate of burn is
too low in an operating boiler. In the third category of ash is the ash
inherent in the materials we are dealing with. The tree grows and develops
some ash, and there's no way we can get that out. So we say if the objec-
tive is to reduce the ash, now let's use that objective to determine the
sequence of events—the sequence of processing to take out those first two
categories. Return to your comment about trommels. The trommel, it appears,
will take out some non-combustibles, the glass and the rocks, before they
ever get crushed to become the second category of ash. I use that as an
example of the objectives driving the sequence. I'd like to see some of our
conversation this morning repeated, but from the point of view of objectives
to produce a specification fuel and thus drive design of appropriate
sequences of operation—not existing sequences.
WARE: I would like to address the whole system design problem from our
point of view. We already had a process. It had been developed for coal
44
-------
liquefaction and gasification, and we decided to modify it or utilize it
for the processing of municipal solid waste. In this way, we already had
a set of end specifications that we had to meet. The pyrolysis system
that we have is quite sensitive to particle size, to ash content, to alu-
minum and glass content, and we had to design a front end that would satisfy
an existing unit. When we started to get into the design of the front end,
obviously we started with classification and characterization of the streams
and how we ought to handle these, to end up with a feedstock for our pyroly-
sis system. The further we got into the design computer model, the more we
realized we knew nothing about it, and that any amount of characterizing
wasn't going to give us a system. And this reflects the comments made
earlier in this session, that first of all the data is very hard to
generate, and secondly when you get it, it's not worth that much. So to
be quite honest we came up with a position where we couldn't design a front
end process. So our first approach was to beg, borrow and steal. We
prowled around the country and visited every other facility, not with the
object of copying them; more to find out what they had done wrong. And we
certainly learned plenty of that. By talking to the people right down on
the plant floor, the maintenance superintendents in most cases, we learned
a few of the things they had done right. Then we came back and sat down
and saw how we could shuffle it together, again reflecting the comments
just made, that we had to characterize it in some way by having a sequence,
by getting out the materials that we didn't need first, prior to shredding,
removing materials after shredding, again which we didn't need. Our ob-
jective was still to make feedstock for pyrolysis. It wasn't to make any
of the other secondary streams. Then we settled on the design, as you ul-
timately have to do, and decided to make it work. In this way we are ad-
mitting that what we came up with is probably not the best way to go, we
wouldn't claim it was, but we designed in enough flexibility that it would
work and that we would be able to generate data from it. What we are saying
is that we virtually built a pilot plant. It's a commercial pilot plant,
whose objective is to find better ways of doing what we probably have done
wrong already. Now in the chemical industry, the normal way of going to
a capital plant is through years of research and pilot plant work. You
really have to generate a hell of a package before the board of directors
will approve expenditure. It seems to me in the solid waste business this
has been vastly accelerated. We've jumped in with both feet in a lot of
cases and built a plant first and tried to make it work. We have done the
same thing, to be quite honest, and the only thing that we would claim is
that in our design we have left enough flexibility in it that we can operate
within a very wide range of variables, and that we can generate the data.
Our plant is highly instrumented for the sole purpose of generating data, so
we can do it better the next time. Coming back to the question of system
process design, this whole business has been so accelerated in the past
three years and the pressures of the environment, the people, and the
government, have been such that some of the elementary steps of process
design have been completely overlooked or bypassed. Now we are at a stage
where we are approaching commercialization in a lot of areas, and we have
the opportunity to collect the data. Now here's the thing that I fear most
of all—in the generation of these plant designs of all of our various
45
-------
systems, we've learned a lot of things that we're not prepared to part
with. That's the way we feel, and that's what I find when I visit other
plants. It takes a lot of socializing and a lot of exchange of informa-
tion, on a personal basis, before people are ready to divulge what they
have learned. Very little of it gets back to the manufacturer, and under-
standably, because there is some resentment on the part of the customer.
He didn't get the information he wanted in the first place from the vendor.
I wonder whether today we are going to be honest enough or friendly enough
or free enough to part with the information we have learned. That will for
me be the major benefit of this session we are having.
POWERS: I know Peter quite well, he was out to my plant. I second most of
the things he says. In Recovery 1 and Waste Management, Inc., our parent
firm, we designed, speced, and built the front end of Recovery 1. Waste
Management and the National Center are co-designers of the back end of
Recovery 1, which is the recovery portion of the project. In going through
just the primary parts of the operation—the specifications of the con-
veyors, specifications of motors, specifications of the shredders—trying to
evaluate this equipment from the vendor's data, their handbooks, their bro-
chures, it's a virtually impossible task. What you finally wind up with is
a situation where the buyer beware. Guarantees, warranties—in refuse you
can forget it. They aren't there. One of the gentlemen over here said if
you buy a 50 ton shredder you can expect 25 tons. We have been a little
lucky, we bought a 60 ton shredder and we get 75. As part of our doing,
not necessarily the manufacturer's. It's an operational situation which
Peter can attest to; you find different things that go wrong and you fix
them. There's nothing in the book that tells you how to fix them. It's
a situation of day-to-day living and how to make the plant operate. Flexi-
bility of design is very important. I don't know where we can go from this
point as far as RDF. In New Orleans we do have a big problem because we
don't have any grates to burn anything on, everything is coal and gas, so
as far as that particular phase of the operation in this area it's almost
nil. We would have to look to new small businesses to burn in this area.
I don't know where they are going to come from—we need more help from
the city, I guess. That's my comments and reinforcing Peter's stand.
KLEE: On this business of equipment sequencing, George has been working
with secondary shredding. George, am I misstating some of your conclusions
when I say that you found that the energy requirements to grind to a par-
ticular average size or size distribution is about the same whether you
use one shredder or make it in two passes with two shredders?
TREZEK: I have some slides of that I am going to show this afternoon in
which we laid out different sequences of equipment and how much energy it
takes to make the final product.
KLEE: We'll talk about this in the equipment session. Still, from an over-
all process design standpoint, things like this are important. I bring
this up because this is something specific, and Peter had said that a lot of
information isn't getting around; it's being held back for proprietary
46
-------
reasons, lack of communication or mistrust in some cases, and it is diffi-
cult to really come up with some concrete statements. Should you trommel
first before you grind? A couple of years ago everyone said sure, grind
it first before you did anything else. They were "positive" about that
then, but where are our positive statements now about these other alterna-
tives that have come down the pike? We can't seem to get a consensus from
the group.
LAMB: I don't think we can get a consensus and there isn't a right answer.
Maybe you want to shred before you trommel; if you do why? Maybe you want
to do the other way around and, I don't think we should be looking at spe-
cific answers because there aren't any. It really depends upon how the
system is going to be applied and even one manufacturer may want to apply
his system differently, for example, in one case he may want primary trommel
after shredding. The thing that would be helpful is to know in making these
choices, what impact does that have on what you are trying to accomplish.
That's where you need the data. I see the real dilemma as coming and de-
pending upon who generates that data, what is the availability going to be.
What is the incentive of the private sector to share data among themselves
or with the public sector, particularly if it's selling plants competitively
or offering services competitively. A company goes and spends a lot of
money in developing this information, frankly guards it kind of jealously.
Will only tell what he thinks everybody knows anyway, and really is not go-
ing to tell the answers that he really thinks are the gut issues. I sit
here and listen to people talk about trommeling and say, oh boy, we get so
many pounds per square foot, or square foot per hour. I listen to them
and they run a test for 15 or 20 minutes or maybe as long as 8 hours. You
want to trommel for a month and see if you get the same data. Everybody
says they can solve the problem of blinding. Who has done it? The only
time you go and see a trommel is after somebody cleaned it and they put on
a demonstration. We do the same thing. But are we setting our objectives
too high, maybe we should so when we fall short we've accomplished something
good. But the real meaty questions which are being asked here—I frankly
think most people who have much insight into the problem may not be willing
to share it.
KLEE: Are you pessimistic about this? Do you think this information is
going to take years, if ever, to become disseminated?
LAMB: I think that the information that the designers would really like to
have, the complete package to be certain when they build a plant, that it
has some opportunity to meet the reliability criteria whatever they be. I
think expecting to find that information from handbooks or publications or
to be generously shared by others in the business is very optimistic. That
doesn't mean that we shouldn't have meetings like this and do everything we
can to encourage discussion and interchange, but I am concerned particularly
in the areas where people think they have an edge, that there's going to be
little incentive and their managements would be concerned if we the tech-
nologists went out and completely shared with everybody everything we know.
It's been in most cases one of very very large costs.
47
-------
TREZEK: We've had .* lot of people come to our lab, a similar kind of thing,
I find sort of a one-way conversation. They come in and want to see what
we're doing; we get out our publications and we start telling them, but we
don't get this kind of thing back. I can see it's a proprietary thing. We
found one interesting area in the shredding work th.-t we did. When we got
into it I felt the equipment manufacturers were just really enthusiastic
about finding out how their machine works with refuse, but I found it quite
the contrary. It seems as though people really didn't want to know that
much about it.
KLEE: Herb, did you have a comment?
HOLLANDER: Yes, I did. Whenever I get a little discouraged I sort of step
back out of the picture and look at it, and ask myself—look how far we've
come! Five years ago would you have had a meeting like this? Five years
ago how many people in this room were even aware that this concept was at
all viable? I am very encouraged! When I look at what happened to just one
piece of equipment, the shredder itself—the evolution in design, the number
of people that have become involved and have addressed the problem. Five
years ago no one ever heard of a trommel, although a trommel's form has
been around for many, many years. The idea of burning waste reduced to a
controlled size was startling when St. Louis was first emerging. Frankly,
it's a conditioning process. You have to massage these things, you have to
talk about them, you exchange experiences, especially the catastrophic ones,
or what was catastrophic to you was par for the course for me, or things of
that nature. So I am very encouraged, and I see down the road this entire
concept maturing. We will better define our needs, we will try to avoid
complexity, we'll try not striving for getting to the moon in one year or
two years. The philosophy that was related to us by Occidential Petroleum
is a good one. They recognize they are not going to attain their goals
immediately and that it is a growing process. I think meetings like this
are essential. They are necessary. Hopefully there will be expressions
from everyone who attends the meetings and not merely absorbers. You have
to have this interchange of views, recognizing we all have our own vantage
points and are maybe naive in some areas but knowledgeable in others. So
I am encouraged, Dr. Klee, I think that we have made great progress in the
last 10 years, especially in a democratic process.
DE CESARE: This follows what Herb was saying. Six years ago when the
Bureau of Mines first put in their trommel and shredder, we found out that
you've got a lot of other variables which everyone who puts in a full-scale
plant or a pilot plant has to realize. No matter of the scale of the plant
he's going to have a pilot plant no matter how big it is. The first one is
going to be a pilot plant. When you make differences in designs, make
changes, whether you're doing it the way they are doing here in New Orleans,
or the way you are doing it in Americology, it's going to change the system
enough that you've got to put in the flexibility and take care of the dif-
ferences that happen. For one example, take a hammermill. First thing that
happens when you put material through a hammermill, it blows out the bot-
tom, so you need an exhaust system. Well, you go to other hammermill in-
48
-------
stallations and you'll see it go straight down without exhaust system. What
determines what's going to happen? You change the refuse, change the mois-
ture content, it's going to affect your air classification drastically.
Every single condition all along the way—there's no definites anywhere.
When you make changes, unless you build identical plants and put the refuse
through, you are not going to come up with any concrete decisions. They
are going to have to need flexibility and have variations all along the
line. So any plant they're talking about—like Occidental—they shouldn't
feel alone that they are building a gigantic pilot plant. Teledyne has a
gigantic pilot plant in Baltimore and anyone who builds it is going to have
the same thing.
KLEE: I'm going to be a dissident here. This seems like a cop-out, saying,
well you have too many variables, therefore you can never make a statement.
You've got to use Occam's Razor or the Law of Parsimony to bring some order
out of chaos. It seems to me that at least one can generalize some basic
fundamental cases—three, four—and make some statement about this business
of sequencing or equipment selection, instead of just saying well, there's
just too many variables and we give up. If we use tha't for a criterion,
we'd never build an airplane, for example.
DE CESARE: That's OK for a pilot plant, but when you try to scale up, you
get entirely different problems. When we first built the plant there were
no air classifiers, so we built our own. How do you scale up an air clas-
sifier; how do you scale up all of the equipment? They're all designed
empirically.
KLEE: That's another good question too.
DE CESARE: When you have empirical design, there's no data base. Your
materials change so drastically that there is no good way of scaling up.
So what good is it to optimize the pilot plant when it doesn't do you any
good because you are going to have entirely different problems when you
scale up.
KLEE: Good point. I'm glad you brought up scale-up; that's going to be
something else on the list.
RYDER: I would like to comment on some of the statements that have been
made here. The thing that I keep hearing here is about these manufacturers
who just can't seem to get data to improve their product. They get their
equipment installed and you would believe that they pound doors down trying
to get data as to how well their equipment is working. Nobody will give it
to them. I've never seen that. I've seen a lot of pounding on my door to
bring potential customers in and to sell them more of the thing that they
sold me. But in general except for the "research" that they went into in
trying to get their equipment to work for acceptance purposes, no real
effort seems to have been made to get any further data. I certainly wouldn't
hold anything back. Secondly, as far as an interchange of information is
concerned, we should understand that these people are in a very competitive
49
-------
market. They've got a product to sell and why give out information which
may benefit the.it competitors. So where are you going to get this in-
formation? You are. going to get it from people like those in this room—
people like this; engineers, and professional people who are just going to
have to see what's around, make some judgments and move from that point.
Now even within th.. my opinion, moving from the lab to c. full-scale
plant is another -„--,-. l^ • ., -inly this time it is full-scale. What am I
going to do with a; i i',,.1;-, i.x-fus*. i.i the equipment becomes an R&D project
with its consequent l«!i.-.( -,iiic. uni 3 of downtime? If the. rcJi-ral government
wants to pay for full-scol , .-il.-f nimts, I'm all in f^vcr of it, but they
also should pay when the ;>~l>i\. i-iarrs pet- bogged down ar>d che refuse piles
up, 5)."- far, the oni> Feder^j ;;h,:;t, ;' ?.7 aight line!
RICO: ''is > [••'- su'jje-rt of analyzing plants, I agree with you that it is going
to cost nicney to arrilyze, for example, your plant up in Chicago. I do not
believe, aijwever, thai this analysis cost interrupts your processing. A
properly designed experimental analysis package has got to recognize that
the plant ':.;i.bl i,uer,i,tt. in an uncontrolled manner, as it normally operates.
It must iT't 'jit.trtf.-,:••„• , i believe personally that this is quite possible.
If you de.-,-i>jf an t-xpt-rurertal program which requires interruption, you are
going to he, 'r,&^:;,i- yr.sr data so badly you might as well not have bothered
in the. fi •-•-" !.; t- /.'!ir!"i -fie analyzed Franklin [Ohio], we did not interfere
with c..-.et . .. • '.'• .; >,-.: '!"=.- trying to do with Baltimore (if that plant
will «vi * !uu> , . ^ . . ..;••. jacetfere with operations. I don't think that
is IK-• »,':,;,.,„ : ;;' >. :. .. i,:. orn. It calls for coordination, it calls for
c:ci,peJ.-• t i-'..:, ' - . '-'• •- call; for external funding of the analysis, but you
sho'.:!d <,a>,'.-- ;>. • '^r ^'• \\r interfering with your day-to-day operations. If
ii rljffa. • ,• '•'!•" i :.-.,. ,. .j co iprrate, and you shouldn't cooperate.
50
-------
BENDERSKY: I just want to say that our program is going to go into full-
scale field testing, and we are very much aware of the need to avoid dis-
rupting normal production operations. Generally speaking, what we have
found in our analysis so far of this problem in testing in full-scale
plants is that certain information that you want is available and can be
obtained in a full-scale normally operating plant; other information that
the industry needs cannot be obtained under these conditions. You have to
have special test facilities to get this kind of flexibility of conditions
and flexibility of control. We are becoming very much aware of the limi-
tations of testing under full-scale field conditions, but much good in-
formation can be obtained this way.
RIGO: The key to getting the type analysis I'm talking about is good models.
BENDERSKY: Good planning.
RIGO: If you've got good planning and good models, the data you collect
can be used to validate those models. And as a result you need much less
data and in many fewer points throughout the plant. But if you are just
going in and blindly taking data—running around monitoring everything and
its brother in the plant—you come up at the end with a data set that is of
no value to anybody.
BENDERSKY: Agreed. You must start off your test program knowing ahead of
time how you are going to use this data or for what purpose the data is to
be collected as an ultimate end product, if you will, and not just tests
for the sake of general information. When you are first starting any in-
formation you gather tends to be useful simply because nobody's gotten any
information. This industry is now getting a little more sophisticated,
there's more need for specific information, which means that the people who
are planning tests and field test programs, in particular, have to give
more attention to planning these tests so that the ultimate data is
meaningful.
KLEE: Harvey, did you have a comment here?
ALTER: Yes, two please. One, let me make a proposal I was going to raise
tomorrow. Maybe it's appropriate now. In agreement with what Dave is say-
ing—we go out and make tests and I think that we ought to make them in the
same general format. When we take the data, we ought to within the industry
have some kind of agreement, that for example, if you are taking certain data
on shredders, that it ought to be taken over the same range of parameters.
If we take the data on trommels, it ought to be for the same range of through
puts, rotation speeds, hole size, etc.; also, efficiency of separation ought
to be measured the same way. Again I would like to raise this tomorrow when
it's more appropriate, but what I would like to see come out of this work-
shop is some follow-up, if maybe it's just ad hoc committees to design for-
mats and tests procedures. In New Orleans, Bill Parker is going to be very
active in making these measurements. We would like to share the data with
people who follow the same procedures for the measurements. In fact, I'll
51
-------
even, if necessary, volunteer secretarial effort to make this happen. But
there's another way of doing what we are talking about, in addition to field
testing. We can scope out on paper, let's say a minimum of three operational
sequences. And for these we can start to look for what is the likely related
performance characteristics. We can even go into what is the likely cost, if
only in pounds of steel required. We can look at the likely power require-
ment, which you know if you have a shredder for a certain throughput; you
may not size the motor exactly right, but you are going to come close to it
on judgment. For these we can go further and say what is the likely mass or
materials balance through that sequence. I think there are enough data in
the literature to make some pretty good guesses, especially if people agree
to the same criteria. This sort of approach leads, then, to what are the
performance objectives of the components? For example, if the sequence says
we are going to trommel and then shred and then air classify, we can guess
at what's going into the air classifier and what ought to come out. We can
now write down that an air classifier, drawn as a box on paper, has to meet
certain performance requirements. We have learned something. For another
sequence, there's maybe a different performance objective, hence a different
air classifier. This, then, leads to the likely product quality that comes
out on the shipping dock. We say, if that's the likely product quality for
that sequence, how does it compare with the customer's specifications—which
is the ultimate test. If it can't compare, we throw out that sequence. If
it comes close, there is a feedback loop as to what has to be changed. What
I'm looking for is to fill in the matrix of comparative attributes. I think
a lot of that can be done on paper. Now it doesn't necessarily arrive at
what might be termed the best sequence. I don't seek superlatives. But
what it comes out to is at least some information in the literature—I have
thought about this, and he has thought about this, and for you who haven't
yet, here maybe is a guide. I think if this type of exercise were started
today, it would be added to by experience. It would be added to by other
investigators, and we'll build a body of knowledge on which to make some
decisions.
KLEE: It seems to me that that would be a start at bringing about some order
out of chaos. We talked this morning before the break about the end product.
With regard to process design, how about the raw material? Do we need more
information on quantity and composition? How critical is knowledge of quan-
tity and composition on process design? Is quantity more important than
composition? Do they affect particular subsets of system design more than
others? Does anyone care to comment on the raw material side of this pro-
cess design?
HOLLANDER: Harvey just told us about testing all of the subsystems for their
performance. Unless you relate it to the feedstock coming in, that infor-
mation is of no value at all. And the data must be reproducible. Having a
snap piece of data here and a snap piece of data there is what we've been
suffering with because of the EPA programs when they first started, whether
it's air pollution or water pollution or whatever. We don't have enough
data for them to establish criteria as to what is acceptable or what is un-
acceptable. Now the ASTM group is trying to develop a set of standards
based on reproducible data. One of the biggest gripes is that so much
52
-------
uncorrelated data that has been published—is it credible merely because
it's a government publication? Therefore, if we are going to embark on a
test program, we have to at least come up with a semblence of steady state
conditions and say o.k. under these conditions, with this feedstock, this
is what happens. Just testing the back end of a classifier without knowing
what the raw material was to begin with is not telling us very much. If we
are striving for a cookbook, we are never going to get one. The object of
this whole meeting, as I understand it, is research and development and
demonstration, not generating a cookbook of design. Is that correct sir?
KLEE: Yes, this is an EPA OKD-sponsored meeting.
ALTER: There are two aspects here. Of course Herb is right. Any materials
balance starts with what's going in the first step. What I was talking
about before is that I think we can start on paper before getting to the
actual measurement. For those of you who haven't made these sorts of sepa-
ration measurements on refuse, it is lengthy, time consuming, and back
breaking. It's 20 minutes of work operating and two days of hand picking,
at the minimum. You operate a pilot plant for one day and you are down for—
let's guess—a week, two weeks, trying to organize your thoughts as to what
you did. But some of this can start on paper. If we only get some scattered
points, hopefully better than that as Herb says, we can at least test them
against the judgment and refine the judgment. We also, I think, can make
first judgments on paper for some reasonable ranges of feedstock. What's
reasonable? You go look to these same government reports, and you say, well,
the steel varies from x to y so we will make our judgments from x to y. And
hope that somewhere there is an average household with an average housewife
and an average garbage.
WALTER: Let me tell you that as an engineer who does some structural design,
one of the first things I must know before I design anything is what is going
into it. We are completing an assessment of European systems that illus-
trates the point. There are some towns in the world that designed plants
based on an assumed quantity of trash. In one instance they designed three
furnaces for 8-1/2 tons per hour. However, they find that they can put 10
tons per hour through each furnace, and they find the third furnace isn't
even necessary, in fact the third furnace is making their process uneco-
nomical. Another city designed a plant for a heating value that allegedly
was lower than what they are actually receiving, and they find they never
can get all the trash they've got through that specific furnace. It leads
back to one thing, that if you don't know what you are putting into it, you
can never design it. I think we cannot overemphasize the need for people to
know what they are going to put in these things, and I submit to you that
right now we don't know.
JOENSEN: Listening to a lot of the comments in regard to the performance of
the process plant, likewise the power plant, I guess I would have to draw an
analogy. When you design an electronic system and match components together,
you look for the frequency response of your overall system. I think that we
have the same objective here. There is no way without a large expenditure of
53
-------
capital that we're going to identify the material going in, so I think coining
back to your outline in Session 1, that what we have to do is really look at
the response of each piece of equipment, and say within the scope of the
composition of that material, how does that particular piece of equipment
perform? Then the system designer says, based on what I have, I can predict
the performance over a range, because I think designing for one material spec
may not be the right approach and I think perhaps we should discuss it. We
look at the end product and say, what is the range of composition that the
boiler can handle if we wish to perform it? I think the main point here is
we look for one composition of material, looking at the Ames plant, when
Arnie Charitland puts in a set of hammers he gets a fine quality material,
but two months later we have a degradation. So the cost tradeoff is, do we
want to change it and maybe we can sustain the performance over that wide
range of composition material because the end product still satisfies it.
Otherwise, coming back to this idea of material composition, that will re-
quire a tremendous amount of testing, and so it's the subsystem that has to
give this performance over a spectrum of material. I think that is the only
way you can design a system.
KLEE: What was Ames' experience, Al? You know they designed for something
like 200 tons, and I presume some sort of composition. Was Ames disappoint-
ed? Maybe Arnie Chantland could also comment.
JOENSEN: I think I'd rather let Arnie discuss that first.
KLEE: Were you surprised? Were there some expectations that didn't measure
up?
CHANTLAND: In dealing with solid waste, one of the aspects of this type of
material is that it is the most non-homogeneous material. This whole aspect
of variability, it ranges from day to day, and from season to season par-
ticularly. In spring of the year, you have high moisture content and you're
dealing with all these parameters of change, the output material or end pro-
duct, also changes to some degree. It may be finer, coarser, or it may be
higher in moisture content, it may be one which will compact more or less,
this type of thing. Did we have some surprises? Obviously we had some
surprises in the operation, but I would have to say that we had a lot of
good experiences too. Nothing has been mentioned about the firing on spread-
er stoker type boilers, or some of the problems in dealing with the vari-
ability in the sizing. Of course, some of our problems are the equipment
itself. Let me give you an example. We found ourselves in the situation
where we reverted from double stage shredding to single stage shredding. We
were burning on spreader stokers, and we knew this was not as sensitive to
fire as the suspension powered boilers. We decided that here was an oppor-
tunity for a larger material. They had always told us that they could stand
coarser mater Lai, so we would send them some coarser material. At the same
time we quickly found out that other systems were not capable of handling
coarser shredded materials, such as the air classifier and the pneumatic
conveyors, air locks, and so forth. So we quickly had to revert back to
the smaller type material. I think there are a lot of good things going on
54
-------
and there is a lot of need for improvement. I think we can find this im-
provement in the systems. Something that has already been mentioned here,
I'd have to take exception that there's no difference in the dual shredding.
The reason that I'd take the exception is that I think the reasons you'll go
to more than one shreding stage is that you're going to do something in be-
tween stages. If you're not going to do anything in the middle, then I'd
agree that probably you're spending the same amount of energy,but if you are
going to do some breakdown shredding, you'll have an opportunity to do some
operations before fine shredding.
KLEE: I simply summarized some of these findings, but I don't want to mis-
represent. All I talked about was the same energy output to get to either
a given particle size or size distribution. Obviously with two shredders
you've got an in-between there that you can do quite a bit with. I'd like
to ask one more question. Arnie, you've had some experience now at Ames.
You are looking at composition from day to day, and you're looking at
quantity day to day. Would you, if you knew then what you know now about
quantity and composition, have changed the process design in any substan-
tive way?
CHANTLAND: I think probably the greatest thing that we would have worked
harder on—we worked on it some but finally gave it up—was extracting more
of the objectionable materials between stages because what we are looking
for is an end product—that is, non-combustible materials as early in the
stage as possible and practical. I guess in our particular system I would
have to say that more would have been done between the two stages of
shredding.
KLEE: Thank you. Don?
WALTER: I'd like to ask Arnie if he would share with the group their ex-
perience firing the suspension boiler, and then I'd like to make a point.
CHANTLAND: Don, I'd like to refer to the people who have been doing the
testings on those boilers, either Al Joensen or Doug Fiscus.
JOENSEN: The question addresses itself to the firing of RDF in suspension
fired boilers. Under the EPA grant administered by Carlton Wiles on my
right, we have not done any formal testing on a suspension fired unit. We
did do a practice burn last week to learn what happens as far as how much
material falls down at the bottom hopper. The only comment I could make—
again it's a single burn—is that you do get material falling down. Right
now we have people drying it and analyzing it for composition and size. At
that time we can identify a little more as to what the problem is concerned.
Now the city last year tried to burn refuse in the suspension fired unit.
Their own experience as given to me is the sluice water necessary to remove
it from the bottom hopper increases because the sluicing time is doubled.
As far as the quality of the product, exactly what's happening, we don't
know because we haven't tested it. So these are really the only remarks I
can make at this time.
55
-------
KLEE: Maybe Arnold would have a rejoinder here.
CHANTLAND: I thought the question also pertained to what was firing on the
other boiler. Are you interested, Don, in the spreader stokers too?
WALTER: You see very much earlier there was a comment made that the most
attractive person to use this fuel is the utility. I submit to you he is
not. He's superficially attractive because he's big, and he's there, and
he can take all of the product, but he's firing in a suspension fired boiler.
The heavier combustibles are going through that suspension fired boiler into
the ash pit, they're not being utilized. Second, he is very intelligent, so
he secured long term, low cost fuel supplies, and therefore the price you
can receive for your product from him is lower. Third, he is very conserva-
tive. His sole purpose for being is to make electricity, and even where we
have found people who control both the electrical production and the trash
disposal, the guys making the electricity don't want that fuel. So I begin
to wonder if maybe we aren't looking at the wrong target, that the right
target is the smaller industrial boiler who has the spreader stoker, who
has higher fuel cost now, who therefore will pay more for RDF, who probably
is a little more daring, a little less conservative. I just wonder if,
really, the statement that the ultimate user is always the utility is not
really wrong. Then there are also a lot more industrial boilers, so you can
start now having competition for your fuel.
KLEE: That's a very good point. Bud Fay actually made that statement this
morning; so Bud, you will have to take the heat for that. I was wondering
when someone was going to remonstrate with that, because a lot of our own
research in EPA deals with the small stoker-fired boiler that may be better
for industrial use than utility use. Floyd, do you have something germane
about this particular subject?
HASSELRIIS: I want to agree with that point, for the moment, that that's
the big problem. The utility is always telling us how they buy the coal
for nothing and they want to compare our fuel product on that level. The
small consumer, definitely, the small boiler, industrial boiler, has to pay
full price for fuel—the regular market price—and he is the best customer
from that standpoint. We can even pretty well estimate the future fuel
costs for him, because that is part of a known process, as much as we
know about it, so we can project and capitalize to sell fuel to that type
of customer with much greater reliability. The point I wanted to bring up
before—the question of sharing information and type of research that's
needed and what's the right way to access to it. I want to take the po-
sition of a producer who has a proprietary process and doesn't want too
many people coming in his plant to see the proprietary part. Yet he has
information which could be useful to other people. He's perfectly willing
to divulge it; however, he doesn't want to bring in a government agency to
obtain that, nor a research agency which might spread some other information
around that he wouldn't like. He might like to generate that information
himself and offer it to the outside, provided there's a vehicle for it by
which he can retain the proprietary nature of his business and at the same
time make that information useful to other people on a standard basis.
56
-------
Harvey mentioned the question of how to put it on a standard basis so that
when we interpret the information we all mean the same thing by it. Which
means that it goes through a standard laboratory procedure, and therefore we
understand what the outcome is. Now we have made some use of the Bureau of
Mines facilities for testing fuels, the combustion of fuels, and the conse-
quences of combustion. This has made it possible for us to look at the
total chemical constituents of refuse-derived fuels—all the minerals, and
compare all of the minerals in a refuse-derived fuel with, for instance,
coal or lignite. Now by using a government testing laboratory as the place
we go to get some information, we find that the information we got is now
compared on a universal standard. That's very, very useful. In other words,
their method of testing with other fuels as compared to our fuel shows how
we related on a universal standard. And that's something we are willing to
do and it's valuable to everybody. So this is one avenue I think, that can
be greatly—where we can help everybody—is to encourage some of the govern-
ment laboratories that do standard work, and of course that can be private
laboratories too, I don't want to offend all of the private laboratories here.
What we are looking at is a standard procedure, whereby if you have tested
for mineral content everybody knows it is the same procedure and the mineral
contents are reproducible. This kind of information as to the mineral con-
tents of these fuels is important all the way through the process. The pro-
cess of making, preparing the fuel has some influence on the outcome. The
input, where the refuse came from, has a great deal of influence on the out-
come and ultimately the combustion, what it does to a boiler, and what and
how difficult it is to collect it is very, very important and relates to
the mineral content. The mineral content, some minerals, are fine, are
inherent, and have certain characteristics. There has been a great deal of
research done on this subject in coal, and I think that a parallel, many
parallel studies can be made. The same methods that you use to test coal
can be used to test refuse-derived fuel and you have a universal comparison.
That has the utility of pointing out to people that they've been living for
centuries with coal with all kinds of hidden and surprising elements in it,
and refuse-derived fuel is not that different. Therefore they don't have to
be afraid of the smoke that comes from a refuse-derived fuel boiler, be-
cause they've always had coal, and we're going to have a lot of coal in
the future. That is one area that I think the type of cooperation and
sharing of information can be very valuable.
PURCELL: I'd like to follow up on that very important point. I think
either the corporate secrecy question or proprietary information areas
certainly are quite vital. As we all know, we are in the midst now of a
new energy crisis and I suppose it's all so much with us that we haven't even
mentioned it explicitly this morning. Of course this conference's effort
can go about resolving some of our energy problems. Certainly the whole
problem of proprietary information, holding back of data, trade secrecy has
certainly beleaguered the whole energy and environmental problems. There
are accusations of do we have shortages, if we do, what are they, etc. Is
the technology there in the environmental field of toxic substances?
Problems have been greatly aggravated by inhibition of information flow
57
-------
in proprietary information work. The Kepone example is probably the most
outstanding in that respect of where the conflict between withholding the
proprietary information and who should know what brought about a very
serious and tragic problem. So I would certainly like to emphasize that
if this proprietary information flow problem is as significant as it seems
to be from the discussion this morning, one of the things that should come
out of this meeting might be some sort of resolution as to how to go about
improving mechanisms for sharing information or at least an ad hoc group
that might work on this at a later time. I think it's a very important
problem.
WARE: I would like, if I may, to steer this session back to a topic which
was brought up earlier, and that is of specific component design, most
particularly materials handling, because it's something that interests me
most of all. I was responsible for the design and the construction and
the startup of the San Diego plant. It was the area where I found we were
most hurting and that most other people didn't have a whole lot individually
to offer. Within that area of materials handling I'd like to start by dis-
cussing the shredding system. Roger brought up earlier something which
quite amazed me, in that he had found that shredders blow air down. The
City of San Diego, before us, found quite the opposite and in our unit we
have indeed found the same thing—that shredders in fact blow back towards
the feed. We explained this by the fact that the hammers going down are
carrying material with them, the hammers coming up have nothing between
them, and therefore bring air back. Whatever the cause, and whatever the
end result, obviously between the Bureau of Mines and ourselves there is
a difference. I think it is something that is worth discussing, when you
start talking about shredders. Going back to the manufacturer, we identi-
fied the problem, we went to American Pulverizer and said, we think the
shredder will blow air, tell us how much. They knew nothing about it, they
were unable to help us, they were unable to tell us how much air. So we
again had to come up with a design which we could live with and use it to
generate data at a later date. We have a large recirculating system on our
shredder, which dumps something like 15,000 cubic feet a minute against the
flow. We circulate it back into the shredder to eliminate dust flowing
everywhere, I think this is something I found in almost every other plant I
visited; dust in the shredding area is probably the most obnoxious, the
least pleasing, the most significant problem found in other plants. By
recirculating the dust and by applying a vacuum system to the shredder
whereby we draw off some of the air, we've been able to maintain a very
clean shredding area. So I would like to suggest that we would proceed from
now on discussing specific components, and the problems that have arisen in
our experience, and our anticipated ones.
SPENCER: Regarding that blowback problem, I think Roger's shredder is very
different than your shredder in that he has a flail mill with two rotors
that's drawing the material down, and so the blow-through problem that he
incurred that caused him to add a skim classifier is very different than the
blowback problem that you have in a down-running shredder. Following up on
the same point that you made, I think that every company would probably have
58
-------
the same feelings that we do regarding proprietary aspects. You spend a lot
of money developing information that you don't particularly want to share
with others. I think some of the work, for example, that Waste Management
and NCRR is doing is very worthwhile. It's very disappointing to me that
the Federal government doesn't support that kind of activity, and my hat's
off to private companies who, for example, will sacrifice the operation by
adding two different kinds of shredders, making their inventory problems
more complex, knowing, certainly, that one's going to be worse than the
other, but not knowing which is best. Frankly I can't understand why the
government or why somebody isn't funding basic unit operations development.
Greg Rigo was talking earlier about trying to develop a system model. Well,
we can't even model a single unit operation, so how can we even proceed to
even look at how to model an entire system when we don't know anything about
the elements of it. I look at air classification for a moment, we've got
about 10 different kinds of air classifiers. I have yet to see some good
data that when you pull 60 percent as of lights, how clean is your light
fraction, how dirty is your heavy fraction? Everybody talks about 80 per-
cent lights. Well, we can pull 100 percent lights, we can pull zero percent
lights. How good is it? It's very disappointing that there isn't compara-
tive data being developed, like Waste Management is developing. I can't
understand why a private company will necessarily spend its assets to develop
that kind of information and then divulge that information to the public with
no apparent benefit to itself. Frankly, I think that a government testing
facility should be developed that allows comparison of one unit operation
against another. Let's take air classification where we shred the different
sizes, where you shred it 3 inches, 6 inches, 8 inches, and 12 inches, and
you classify at 60 percent lights, 50 percent, 40 percent, and look at the
light fraction to see what's in it. Measure the operating data, measure the
power consumed, and so forth and then try a different air classifier and see
what that is. Then you go back to evaluating the system, and you say o.k.,
in Wisconsin Electric or in St. Louis where we've got huge 300 megawatt
boilers, we can use a low quality air classified product and we'll take just
about anything we want. Dave Klumb's very happy to take anything you want
to feed into it, and as Ed Wisely was saying, for a while they didn't care
if they had the cans or not. In our facility at Monroe County we are deal-
ing with 40 megawatt units where we need a particularly clean combustible
product, and there is no basic information that allows a comparison of one
air classifier to another. So a company spends hundreds and hundreds of
thousands of dollars developing things to meet its particular problems, and
yet the next time you come up against a new problem, and you wind up having
to spend all that money over again to look at a new way to solve some dif-
ferent system problem. I think it's a responsibility that should be a
shared responsibility through government funding. I think these demon-
stration programs have been pooh-poohed by EPA. I think they are very
worthwhile; the fact that there are failures is very beneficial. I think
it's important that before a community makes a major investment into a
system that they don't walk into it blind. I think these problems that have
come up have been beneficial. Just to say that companies ought to take all
the responsibility here, or individual companies should, and not share the
59
-------
responsibility among many companies or many communities through Federal
funding, 1 think, is a mistake. Frankly, I think that the Federal govern-
ment has really dropped the ball here. When you compare the development
with, let's say a missile system, to the development of resource recovery,
you are really operating in the Dark Ages. In developing missile systems,
the government funds the research and development and then forces their
people into that information, collects the data themselves, and then forces
it on to the competition in order for competitive bidding. They found that
that makes the lowest cost system to the eventual user, which is the tax-
payer. I think the system that we are looking at right now of having pri-
vate companies invest at much higher interest rates than the Federal
government can and the waste that goes with developing the same information
through private companies over and over again and then passing that off to
taxpayers through tipping fees is a mistake, and I think you've really
dropped the ball.
KLEE: I'd like to get a hundred copies of those remarks! Carlton, did you
have something to say?
WILES: Would he like to send those comments to the congressional committees
and so forth that control appropriations?
KLEE: I think everyone knows that even though he's criticizing the govern-
ment decision here, what he's really saying is that EPA should be better
funded; more funds should be provided by government for these purposes.
SPENCER: I've seen the requests, and EPA hasn't requested any outlandish
amounts of money from the government. You can't do the job if you haven't
got the money and you haven't got the resources. They don't seem to be re-
questing it.
SPENCER: Look at the last congressional hearings and look at the people who
were selected. I don't know who selects or who goes to those things, but
certainly, for example, somebody interested in mass burning is surely going
to say that technology is here because his competitive edge is the fact that
he is old and you are new. If you look at university professors, you have to
scratch your head sometimes and wonder if they are tied down to this earth.
You bring these people into a congressional hearing, and you put some titles
behind them and some nice universities, and people become impressed by what
they are saying when they are not even associated with the realities of the
problem. I have the feeling it's an education problem, but it's one that
we shouldn't be sitting around here trying to get Raytheon to tell Arthur D.
Little what we are doing, and vice versa, because it's not going to happen.
On the other hand, if you make it attractive to us we might do it. You've
got to find the method to make it attractive. In the meantime, we have to
be profitable and we have to justify the expenses that we are making which
have to be made in order to continue development. The only way you can do
that is to show a payoff somewhere down the road. You're not going to get
that when you are giving the information that you just paid for to your
competition.
60
-------
KLEE: Steve, we haven't heard from you for a while.
LINGLE: I would like to respond to the problem that Dave Spencer just dis-
cussed in terms of private firms giving up information they've developed
with their own funds and also regarding the comments made about evaluating
new facilities which are coming on-line. We recognize the value of evalu-
ating new commercial plants that are coming on-line, and EPA is developing
a program to do that. Based on discussions that we've had already with
some of the companies who are involved in the building of these systems (and
in some cases they own them as well), it's clear that they are very reluc-
tant to have a private contractor funded by the Federal government come in
there and collect a lot of detailed information on their system. The reason
is very simple. Sure, there's a need to get information out, and I'm sure
that most companies would like to share information to the extent possible,
but there is that very real business issue of giving up competitive edge.
Now there's been a lot of discussion about the value of evaluating existing
systems, but the extent to which that can be useful is going to depend on
the extent to which the firms can allow us to collect detailed information
on systems. I understand the problem. If I were in the private sector I'd
be reluctant myself to give up information on how a proprietary air clas-
sifier or some other piece of equipment performs. Does anybody have any
suggestions as to how we can overcome this situation?
SPENCER: First off, I think it's a mistake to try to evaluate systems at
this stage of the game where we haven't even got the information with re-
spect to evaluation of the unit operations, and therefore anything that you
do with respect to evaluating the system will be useless. The second thing
is, by its very nature you can't evaluate a unit operation unless you lit-
erally destroy the operation in a sense. You can't not disturb an operation
and still test the range of five different feed sizes and six different light
and heavy fractions. It has to be a test and evaluation program. You can't
walk into a running system that its lifeblood is moving waste through a site,
and start screwing around with the equipment and yet if you don't screw
around with the equipment, you'll never get the essential data that's neces-
sary to understand the unit operation. If you don't understand the unit
operation, you'll never understand the system; you'll never be able to model
that system.
WARE: I'd like to address myself to the point that has been brought up. To
give the EPA some credit, our particular plant at San Diego is indeed par-
tially funded by the government to the extent of four million dollars. Above
and beyond that they're letting, or just have let, another contract for some
half million dollars for an independent contractor to evaluate the plant
throughout the demonstration year. The net result of that will be that
everything we've done during the test, everything we've done during the
year, will be public knowledge. There's the tradeoff, we've put some of
our money into it, some eight million dollars, which really for us is
expensive advertising. The County of San Diego has put some money in it,
as a gesture I guess, and they will benefit from it to some extent, possibly
61
-------
getting their money back. The Federal government has put money into it,
they have assisted us in our development, and in return all the information
thai we generate will become public. We don't feel we are losing our com-
petitive edge because of it; in fact we feel it may well enhance our po-
sition, because everything we've done will be publicized—the things we've
done right and the things we've done wrong and most important of all what
will be publicized is how we went about it. We feel that a year from now,
regardless of the competitive technologies, the things that municipalities
are going to buy will be expertise, management skills; they won't be buying
a process, because by that time, a year or two from now, most of the pro-
cesses will be highly refined; they will be most competitive; consumers and
customers will be buying the company with its management abilities and its
development skills. As I say, the EPA programs do exist. I don't know how
many more are envisaged, but they've certainly assisted us, and the end re-
sult will be public knowledge and sharing of information. The contractor
will come in, he will have access to everything we do. Now there may be
things in my head and the head of our maintenance superintendent that we
don't talk about, but a sharp contractor will be able to pick out every-
thing we have ever done and why, and publicize it, so we end up divulging
everything we do, to the benefit of everybody, but not to lessening our
competitive edge. I believe to answer Raytheon's problem, it is happening
in some cases. Now whether it should be extended or not, that's another
question altogether.
BENDERSKY: T think this discussion is very valid with respect to the
future of ttiifa industry as far as the various test programs that are going
on. I think we have to differentiate, and I think I see the pattern coming
about here. Testing an individual system (a la El Cajon or St. Louis or
something like that) is valid for that system and that system only. You'll
get ail the information you want about that particular system. Where I
think it's been lacking so far, and I would like to see and I hope that the
direction will go that way, is that these tests will have a commonality to
them, so t.uat we can evolve comparative data which will be useful to some-
body who wants to design a new system—not the one at El Cajon, and not the
one in St. Louis, and not the one at Ames. What we need to do is plan some
of these plant test programs so that the data then becomes comparable and
ustlul for future systems, and not just the evaluation of that particular
plant alone,
R1GO: Dave, back on the concept of how you go about getting this data, I
agree with you wholeheartedly that if we could set up individual unit
operations, control the inputs to them on a series of controlled experiments,
that we are going to get a fantastic amount of data in a very straight-
forwai'd manner. That is the ideal, and I feel that if the EPA or DoD or
somebody comes up with a good pilot plant, we can get a tremendous amount
of information this way. In fact, probably this is the only way to do it.
However, 1 believe there is a second approach which makes a lot of sense,
and this is with systems analysis coupled with design procedures, or models.
YGU rely u;..on '.:ue dynamics of the normal variability of the input refuse and
62
-------
the normal changes in just operating that plant to establish the range of
plant operations. If you are monitoring all the outputs you can then back
into what the transfer functions for the individual components have to look
like. That is the second approach for getting this type of data. It is
one that is not widely used because it is much more difficult. It requires
a tremendous amount of analysis, as opposed to a tremendous amount of data
acquisition. Given that, there are plants available today that we can pro-
bably get our hands on if we promise to stay on the outside and not inter-
fere with operations, compared to waiting even if Dave Berg could come up
with money to build this massive pilot plant next week, we're looking at
three years, four years before it is on-stream. We can do a lot of work
from the plant we're working in today and get some handle on the transfer
functions now. We should validate them with the type of experiment you're
calling for in the future.
KLEE: Dave brought up the matter of a pilot plant, and we haven't heard
from representatives of the armed services today. I know that Steve Hurley
of the Navy has some thoughts about a pilot plant. Would I be embarrassing
you, Steve, to say a few remarks about this?
HURLEY: I'd like to make some comments along those lines and they're
general comments, but I was thinking they fit more into Session 4 than they
do at this point. However, I come to this table with a certain experience,
and as every moment goes on I'm more happy that I'm here, because I think
there are some people here who have been struggling with the problem for a
number of years now and have reached a certain frustration, while there are
others of us who are new to the problem and would like to utilize your ex-
perience—or play leap frog research, if you will—so that we can bring
about changes to the existing solid waste systems that I think we could
agree to are necessary. I think that if there is a premise to the workshop,
it seems to be jelling in my mind that it goes something like this: That
changes to the existing system, to existing systems that we operate today
must take place, and we haven't defined yet or we're dealing with the
problem of how those changes should come about. We need to know more, first
of all, about what that waste stream is, because we didn't need to know that
much to operating existing systems. Second of all, I think it's important
that we do indeed get the funds that we are able to justify, working at what
I call the field level or at the place where the problems exist. To that
extent I think it's interesting to note that DoD put out an instruction
called resource recovery three years ago in which the word was that we were
all going to proceed vigorously. It didn't take long for people to realize
that you couldn't proceed vigorously based on the information and state of
the art. Today the same instruction has been reissued, and I have to speak
for the Navy, which I represent. We've been able to muster a million dol-
lars, and it's our intent to get the money spent as close to the problem
as possible. That means to me that you spend your money at a field activ-
ity working on an actual waste stream in order to demonstrate how you can
go from the existing system to whatever it is you propose to do. And that
to me has the distinct advantage of drawing anyone who wants to come down
and see the advantages and the costs of changing—the advantages of pro-
63
-------
duelag a usable macerial oi recovering whatever it ic, fuel or otherwise.
R<- that to me is part of the situation as I see it. I think the discussion
is gettipg better every moment I hear.
RYDER: That comment, if I'm reading Mr. Rigo right and Mr. Spencer perhaps
or the other gentlemen over there, about possibly getting some 25 million
dollars or any number, whatever number you pick, and build a pilot plant as
Mr. Rigo correctly stated—I should say understated—in getting in the line
of three years, it will take four to five (that's only up to the testing
stage) and so, the pilot plant of what? What system are you going to make
a pilot plant of? We've got a number of different systems, a number of dif-
ferent ideas, that's my point, you've got pilot plants all over the country,
so build up from that point.
KLEE: Well, maybe we ought to let Dave Spencer reply to that since this is
rather specific.
SPENCER: I don't think the objective would be to build a pilot plant or a
system. That's exactly what we've been doing, which is not the way to go.
I think that the information that is lacking is the unit operation in-
formation, and what we've got now are too many systems that we can't compare
to each other on any kind of uniform basis. That's not to underestimate the
problems associated with getting, with three years or five years, but at
some point in time somebody is going to have to do this if we are going to
make sensible decisions regarding the selection of unit operations for any
system, whatever one it is. The other thing that concerns me here is, when
I hear about the Navy putting up a million dollars, I kind of say, big deal.
That's very little compared to what communities are putting up in terms of
50 million. If that's the best the government can do, good luck.
KLEE: In a sense this discussion is really more germane to Session 4 to-
morrow afternoon. Now back to the process, the system process design, if
we can. Is there someone who would like to get back to the subject here?
POWERS: I'd like to make a comment on that. Herb, I think, hit it on the
head when he said we are not really trying to design a cookbook plant; each
plant has to be designed specifically for its requirements. The waste that
is generated in New Orleans is entirely different than the waste that is
generated in Buffalo, New York—seasonally, geographically, and whatever
way you want to evaluate it. Whether we can do an RDF operation at our
particular place as to some other place in the country, or whether we
should go methane or what, we don't really know. So you can't design a
cookbook plant. It has to be in the location and it has to be as a local-
operation. So that's particularly what I have to say, as far as this goes.
BURCKLE: I don't think the objective is to come up with any kind of a cook-
book plant or a standard plant that will fit everyone in the nation. I
think what we are really concerned about is to determine the underlying
principles of this design, which would include how the equipment performs,
64
-------
the equipment model, the unit operation model, and the system models that
are optional alternatives to getting to different types of end products
based on the types of feed material coming into the plant. Now, cellulose
is cellulose, and that's basically what we're looking at when we are talking
about a fuel. Every community generates varying quantities of this material
seasonally, and it's contaminated with different kinds of inerts, which we
want to pull out in recovery. When we talk about composition of solid waste
from the municipal sector, we are really talking about a quantity of cellu-
lose coming in, we're talking about a quantity of metals coming in, we're
talking about a quantity of glass coming in, that's where the composition
enters in. We're talking about all of these things as though we can look
at them as separate streams, although they are really and truly mixed at
the inlet side, and the question is how do you separate these things? I
think the underlying principles are going to be pretty much the same from
city to city, but translating those into use in a given city is where you
get into the differences.
LAMB: We're talking about quantity and composition. My own thoughts on
that are quite strongly that I don't really care what the composition of
refuse is, because even if one does a lot of studies on it and comes up
with averages, the day-to-day variations are so large as to pale them into
insignificance. It is very important to know the quantity. There have been
some disappointing surprises therein. For example, at East Bridgewater we
got less than we thought from the communities, if you look at the historical
statistical data on generation rates. I think, however, the principal
reason that the quantity is so vital, is that when you enter into a con-
tract with one of these municipalities, you want a "put or pay" contract.
There will be a lot of discussion over what amount is going to be put or
paid for, and it's really with that information that you can then decide
what size the plant has to be and what the costs are going to be. So I
would put a much higher priority on quantity generation information than I
would on composition.
HOLLANDER: Where the economics on other than the fuel fraction is contin-
gent for the viability of the plant, you have to know something more than
just what the cellulosic fraction is. You should know what the cellulosic
fraction is and the mineral fraction.
LAMB: But I would say, Herb, if your basic economics are, for example, a
specific composition of metal, you're in trouble already. That's an ex-
tremely high risk venture.
HOLLANDER: I agree, Tom.
BURCKLE: Based on those two comments, that is why we have to look at waste,
not as a mixed fraction total amount coming in, because on the other end
your economics say you're signing contracts to produce a certain amount of
product in each of the categories on the outlet end of the system, which
says that what you are really doing is, what is my average or base quantity
65
-------
of cellulose coming into the plant that I can recover as a fuel product.
What is my average or base quantity of aluminum coming into this plant
that I can recover and sell as a product on the output? I think those
from the compositional standpoint are the important points that you have
to consider on your economic side, because your economics are dependent
as much on your products that you are selling as it is on the tip charge
that you are charging with total quantity coming into the plant.
H. FREEMAN: I'd like to ask some of the systems designers—those of you who
represent that element of our group—how much do you really use these gene-
rational studies that we put out so many of, as far as across the board
generation of refuse in the average town when it comes down to the design
of these systems you are bidding on. I guess I'll direct that to Mi. Lamb
of ADL. Do you use them at all, or do you go out and run another study?
LAMB: On the work we've done, we haven't really paid a lot of attention to
composition data. We use the national averages to more or less forecast the
financial model. Because the compositional data is going to be variable any
way, I don't know what kind of a program one would propose to get a repre-
sentative sample. Am I going to sample all of the wastes for a year; ani I
going to spot sample it for a week, and if it's a week, is it going to be
the summer or the winter or spring or the fall? I think we are more con-
cerned about generation rates, and there I guess, we rely principally on
historical data from the communities and that's frequently lacking. Where
this weakness in our experience has ultimately shown up is when it comes
time to draw up the contract and to state the quantities that are going to
be put or paid you'll see a lot of backstepping, saying, we know we are
going to give you 1,500 tons and we want the dumping fee based on 1,500
tons, but we'll guarantee you eight. We know there's going to be a percent
and a half of aluminum, we've heard aluminum's worth $500 a ton, and we
know that it's going to be worth $9 in the refuse, and we want that for a
credit, but if there's only 3/10 of one percent aluminum in the refuse,
well, we are going to be disappointed. We the systems suppliers accept
that risk, not the community. So I think it gets difficult in the compo-
sitional thing, because you have no control over it. For example, legis-
lation might become active like a bottle bill or container bill that would
influence the composition. You have to deal with these issues in the con-
tract and allow for ranges, because I think it's unreasonable for either
the systems supplier or for the municipality to make concrete promises which
they can't meet because that's either going to drive the system supplier in-
to bankruptcy or make the contract at great expense. There has to be flex-
ibility there. I don't know what other systems suppliers do, but we don't
do a lot of compositional work; we focus principally on quantities rather
than composition.
ALTER: We are all talking about a bandwidth problem. We have an average—
EPA publishes a national average, your city publishes it—they freeze one
day of two weeks in a year and say this is the composition, this is the
quantity. We all know it's wrong, but we all like the warm feeling in the
breast by having a single number. I think that the only outlet, the only
66
-------
solution to this problem, is not to sit around here and talk about it and
worry about how much it is, but at least for the time being, handle it
contractually; it becomes a legal-institutional matter and not technical.
The engineer still has to design for the bandwidth, not for the single
point. One example, here in New Orleans the contract talks about the plant
and the tip fees are based on an asserted single composition, a "baseline
composition." But the contract also says by the end of the year the parties
will sit down and adjust the tip fee up or down based on the operating ex-
perience of the plant. The year following, they will readjust and so forth.
This is the only fair way of handling changes in composition because the
operator—even if it's the municipality has to avoid bankruptcy. We can't
predict very well. We can go to all of the generation studies and it
doesn't tell what will happen two years hence if Campbell's soup is no
longer packed in a steel can but in a plastic pouch. I think what we are
talking about here is the ability or the needs or the demand that a system
be designed for an average plus or minus on the bandwidth, and that con-
tractually the tip fee (or whatever the method of payment) is adjusted ac-
cording to experience.
KLEE: Gentlemen, this session ends at 12:30 and I intend to end on time.
As the chairman's prerogative, I'll take the last 3-1/2 minutes myself to
sum up briefly. We've had a lot of discussions on a lot of different items.
There seem to be two items receiving the most importance, and they were at
the antipodes of any process design; more specifically, what goes in and
what goes out. We had most of our discussion on the specification for
fuels—how tight it should be or how loose. We've had as much or more dis-
cussion on the quantity and/or composition of the incoming material (es-
pecially the quantity) which affects process design as well. Now there
were some items that came up that were discussed and identified as problem
areas, areas in which information needed to be developed more; certainly
materials handling and scale-up were mentioned. On more than one occasion
it was reiterated that there are many, many variables in system design,
too many to produce a simple handbook. One can't paint by the numbers here,
although there was some suggestion that perhaps we could generalize into a
smaller subset to create some order from the chaos. There was some dis-
cussion about the ultimate use of the fuel—utility versus the industrial
or smaller user. There was quite a bit of discussion about proprietary
information and funding. We started the discussion with a consideration
of four questions. Dave Berg summarized our answers this morning. How-
ever, it's not a particularly positive list, because there are a lot of
"No's" here. Let's look at them one by one. The first question was: With
the technology available today, can we produce a suitable product which the
consumer considers a useful and desirable fuel? That was Part A of that
question. It appears that the consensus is mostly "No" at this point in
time. The second part of that question was: Are specifications for the
fuel product available? Again, in the main, there's a lot more work to
be done too. We'd have to answer somewhat negative on that. The second
question was: Is adequate information available to the implementer or
67
-------
investor for selection from competing commercial and developmental systems?
It looks like kind of a "No" on that one too! The third question was: Is
there an alternative process concept offering a potential for significant
improvements in product quality, economic performance or both over the
present practice of primary shred and air classification? Well, certainly
no one suggested that there was a panacea—that there was a better flow
chart. Evidently we just don't understand the problem well enough to make
a blanket statement like that; there are so many variables and we recognize
that there's no way in the world you are going to come out with the answer
with regard to such sequencing. The fourth question is: Is adequate in-
formation available to the systems supplier or consulting engineer to per-
form the process of design with sufficient confidence to offer performance
guarantees? Definitely there was a big "No" here. Although in general
we've answered these questions negatively, it doesn't mean to say that
there isn't some information available in regard to each one of these
questions. It does seem to me from the discussion we had this morning,
that we do need more information to design more efficient systems, and
we do need more information to make selections of what we already have.
Although to keep on time I'll have to declare this session closed, we can
continue our private conversations during lunch. Thank you very much.
68
-------
SESSION 2: SELECTION OF EQUIPMENT
Stephan Lingle, Moderator
LINGLE: Session 2 deals with selection of equipment. I think this gives
us an opportunity to move from the discussion we had this morning, which I
think was beginning to yield some good results, into discussions of spe-
cific problems. Maybe we can start to get some more specific directions
in terms of what EPA or others should be doing in the way of R/D&D to solve
these problems. We're going to focus this discussion on individual equip-
ment items. I'd like to use as a starter the breakdown that Harvey Alter
suggested this morning in terms of breaking the different materials unit
operations up into materials handling, separation, and size reduction. In
discussing each of these, I think we should keep in mind the discussion
items which are listed in the agenda for the session, which was sent out
earlier. The suggested issues we should be thinking about regarding indi-
vidual equipment items are such things as the equipment performance data;
the experience that we have with existing equipment; capital and operating
costs; guarantees; scale-up factors; effects of changes in the incoming
waste stream (we touched on that a little bit this morning); different types
of waste, besides municipal wastes, that we might include; scale of oper-
ations in terms of large versus small cities; and so forth. Keep those
factors in mind in discussing the individual equipment items, and I think
we can focus on some of the problems that we have with these types of equip-
ment, and some of the real needs for R/D&D. I think to start this off I'd
like to ask Dave Bendersky to give us a rundown of the study that they did
and some of the suggestions that they came up with for research development
and demonstration, and then I'd like to go back and start on the materials
handling. Dave, can you comment?
BENDERSKY: In our program, as I mentioned earlier this morning, we broke the
various so-called preprocessing equipment down into 12 categories, which are
listed on page 10 of this handout that we passed out this morning. [See page
69]. Starting, if you will, at the beginning of the plant with receiving
facilities and listing a few of the items that are involved in receiving fac-
ilities and going on down to conveyors, shredders, magnetic separators, air
classifiers, and so on. What we were looking for in our study was the mis-
sing information in these areas. What are the research needs? I must tell
you that we found research needs in all of these areas. Generally speaking,
the area that has received the most attention so far is shredders, and in
spite of the effort today on shredders, there's still a number of unanswered
questions. There's a number of basic missing data. People like Trezek and
other people have done laboratory tests on shredders. Quite a few shredders
are out in the field (the last count that I had was over a hundred shredders)
that are actually operating on MSW. What we are not getting yet is full-
scale operating experience reported in the literature. Everybody is essen-
tially learning through their own efforts. The basic data that's needed by
people who have to either design shredders or design systems—there is
still quite a bit of missing information. I don't know if this is the
time to be very specific about the types of missing information, but basic-
69
-------
ally technical and economic data is still missing as far as we can deter-
mine, for example, the proper maintenance schedule for a shredder. Most
people are retipping the hammers when they wear, but they retip them at
different time scales. [It ranges] all the way from every day retipping,
some every week, some a month, and some just let them wear out and change
the hammers. The question is: What is the most economical procedure,
what should be done from a cost-effectiveness standpoint? We don't find
this data yet.
PREPROCESSING EQUIPMENT USED IN WASTE-TO-ENERGY SYSTEMS
1. Receiving facilities (S)
a. Scales
b. Building
c. Loaders
d. Ancillary equipment
2. Conveyors (S)
a. Flight
b. Belt
c. Vibratory
d. Pneumatic
3. Shredders (U)
a. Harnmermills - vertical, horizontal
b. Grinders - roller, disc-mill
c. Flail mill
d. Wet pulp
e. Ball mill
f. Knife mill
4. Magnetic Separators (U)
a. Belt type
b. Drum type
5. Air Classifiers (U)
a. Straight
b. Zigzag
c. Vibrating
d. Drum
e. Concentric
6. Screens (U)
a. Trommel
b. Vibrating-reciprocating and gyrating
c. Disc
7. Dryers (U)
a. Drum-type
b. Fluid-bed
8. TK-msifiers (U)
a. PeJletizers
b. Briquetters
c. Cubers
d. Extruders
70
-------
9. Storage and Retrieval Systems (S)
10. Process Control Systems (S)
11. Dust Control Systems (S)
a. Cyclones
b. Baghouses
12. Fire and Explosion Control Systems (S)
(U) = Unit equipment.
(S) = Supplemental equipment.
Source: "Study of Preprocessing Equipment for Waste-to-Energy Systems:
Summary Material and Research Needs," Midwest Research Institute
for Environmental Protection Agency, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, February 1977.
We have serious problems in so far as the safety of operation of shredders,
which hasn't been addressed properly, at least in the literature, yet. All
of the shredder manufacturers that we contacted are very concerned about the
hazards of running shredders on MSW. Frankly, they are keeping their fin-
gers crossed, hoping that nobody gets killed before real solutions to these
problems are found. People are trying to avoid the problem, trying to pick
out the dynamite and the other obvious things that might be in there. But
things will and do get through. There have been fires, there have been
explosions, and there hasn't been enough attention paid to this particular
area. Operating costs—there are some scattered data here and there; it's
hard to compare one set of costs against another because accounting pro-
cedures are different. In many cases people are keeping cost data on
systems and not on individual pieces of equipment, so it's hard to ascer-
tain whether one type of shredder is more economic than another. Or even
basic questions, like is a horizontal hammermill better or worse than a
vertical hammermill. Basic questions like this haven't really been nailed
down yet. Now that's in an area that (as I said earlier) most people have
done the most work in, and it leaves the rest of them with even less basic
information. For example, receiving facilities—what is the best layout
for a receiving facility? What's the truck traffic pattern; what's the
best way to go? Should it be dumped on the floor and then pushed by a
front end loader into a conveyor, should it be put into a pit, should it
have hydraulic rams in it? These are questions that haven't really been
addressed on a comparative basis; individually yes, each system and each
system designer has considered these problems, but it's not in the liter-
ature, it's not really available to the industry. Conveyors have been
around a long time, used for all sorts of things, but MSW is a unique
problem. Some of the characteristics, for example, of MSW that have been
studied—composition, etc.—are fine, but it doesn't answer questions that
the conveyor people ask. Compaction, angle of repose, and other char-
acteristics of MSW are not readily available. Magnetic separators have
been around a long time. Simple units were used initially; they were
71
-------
found to be somewhat lacking; they do pull out some of the magnetic mater-
ial, but the quality of the product has been questionable. There has been
some redesign by the manufacturers of magnetic separators in order to de-
sign their equipment more specifically for MSW. However, there's only
been a couple of years experience so far, and again the field data isn't
as complete as it should be. For example, although you get reports on the
percentage of metal being recovered, they don't usually give you the amount
of metal that was in the raw waste, so you don't get an efficiency factor
of the magnetic separator. Furthermore, you don't often get the quality
of the recovered ferrous metal, in other words, how much nonferrous metal
is in there. Air classifiers were already discussed to some extent this
morning. There is no good comparative data between the various types of
air classifiers—which is best, which is most efficient? We don't have
the data, at least we were not able to find it. I can go on down through
the list that way indicating specific areas of need. I would like to point
out a couple of areas that have almost totally been ignored in the liter-
ature, and that is process control. This has sort of been a black box that
each of the systems have, but they haven't been reported. Some of them are
very simple, some of them are very complex. Should the system have a very
sophisticated control system, should it be automatic, or should it be mostly
manual? These questions are not answered readily right now. Everybody has
their own opinion, but there isn't good comparative data yet. Incidentally,
a lot of this information is considered proprietary by the designers of the
control systems. Dust control is something that hasn't been reported. It's
being used more, I'm glad to say; it has been sort of ignored in the past.
There are more dust controls now on systems, but their effectiveness has not
yet been determined, again, primarily because there just hasn't been enough
experience. So these are some of the areas that we have found very lacking
in so far as basic data that almost everybody in the industry would like to
have, but it just doesn't exist yet.
LINGLE: Thank you, Dave. If you look on page 10 of the MRI report [page
69], let me just make the following sort of organizational suggestion.
I'd like to break these different unit operations down into the areas that
I mentioned before, and maybe you can just put a little note by each one
of these. I think that the first two are materials handling issues, and
Number 9 is a materials handling issue. I think Numbers 3 and 12 are size
reduction issues. [Numbers] 4, 5, and 6 are separation questions. Ten and
11 and 12 we can classify possibly as process control. Then 7 and 8 are
additional processing or something along those lines. I'm bringing this up
just so we can organize the discussion. I would like to start with the
materials handling issue. Within that I'd like to pose a question: Do we
have a single adequate means of storing and retrieving shredded solid waste?
I would postulate that we don't. I think the means that has been designed
in most systems, which is the Atlas storage system, frankly is not working
well in some cases. I think we need some improvements, and I'd like to
start off by asking for some comments on that.
72
-------
CHANTLAND: I might say that I don't believe that the Atlas system is a
total failure. It does provide a storage facility; it does provide a means
of unloading the facility with equipment that is not buried within the waste
stream. There are times when mechanical problems do exist, but I think you
are going to have that in almost any type of equipment. I will say one
thing about its limitations, from our experience at least. It is limited
to handling only the fine ground material. This was another part of our
little test. We are going to send over the larger material to the spreader
stoker boilers and it will not go through an Atlas bin. Tt really isn't
designed to handle larger material.
LINGLE: When you say larger material, what size range are you talking
about?
CHANTLAND: Maybe 3 to 4 inch. I would say in general it has to be less
than 2 inches in size.
RIGO: Steve, as people are addressing the storage question, let me get
Arnold to shed some light onto the problems of spontaneous combustion when
storing chopped garbage. Have you been able to develop any rules for de-
termining detention time, depth of storage, moisture, or anything else
where you start looking at it and being concerned?
CHANTLAND: I don't believe we have any real data on this area. We do have
some procedures. One of the procedures is that we expect if at all possible
to empty the bin once per week. We have, however, had the material in the
bin for as high as three weeks, without any spontaneous combustion. There's
only two things that we are concerned about there. One, of course, was
possibility of some spontaneous combustion; the other thing is, we know that
due to the moisture and oxygen present, there is going to be a breakdown of
that material and there's going to be a Btu loss as it sits there. So it
is important that this material be used in an early stage and as it is
produced.
FUNK: I worked with Arnold in the design of the Ames plant, and we faced
a decision of coming up with a storage and retrieval system quite a few
years ago. Your remark that it just doesn't work, I disagree with totally.
There had been some cases where there had been some problems with it, but
based from the designer's standpoint, when it comes to the selection of
equipment, we had this decision to make and we had the opportunity to view
all of the Atlas systems as well as all of the other systems that were in
operation, and we selected the Atlas. For what it's designed for it works
very well. If I had to make the decision again today from what I've been
able to observe, it would probably be that one. I think that's further
enforced by the fact that Americology made that decision, the designers of
the Chicago system made that decision. I would just like to turn it around;
where has there been one besides the Hamilton, Ontario plant that doesn't
work?
UNCLE: Baltimore.
73
-------
FUNK: Now has that been the Atlas system itself? Now all of these stories
about explosions and having a D8 Cat hook onto the chain drag and drag it
out—they are not the fault of the Atlas system.
LINGLE: Well, no, it's been the Atlas itself that's failed in Baltimore.
Maybe the distinction we need to make there—and I really pose the state-
ment that it doesn't work to stimulate comment—is maybe it's the size of
the material. In Baltimore you have a larger size material; it's not a
small material, and Arnold mentioned that it perhaps does work on smaller
size material= In Baltimore the material is larger. The feed rate has not
been uniform, the material has solidifed in the Atlas, and this is a prob-
lem that I've heard other people talk about. Regardless of size of ma-
terial, you just can't store for any length of time in an Atlas or any
storage system. They've had significant wear problems in the floor of the
Atlas in Baltimore. The concrete is just wearing out, and there's got to
be some solution to that. Perhaps other people would like to comment in
more detail on some of the problems or successes that have been experienced
with the Atlas system or other systems.
WISELY: I think that there's a misconception, really, of the function or
at least of the combined function of the so-called Atlas system. One of the
main advantages of a system of that type is that it can both retrieve in
various streams as well as perform a certain degree of storage. One of the
principles which has become fairly obvious in essentially every application
of shredded waste is that storage, if possible, should be avoided. This is
just a statement of fact. The Atlas system, I think, is being used merely
because it's the best of any of the available systems at the moment. I
don't personally regard it as a failure. I agree with Harvey on that; that
it's just a matter of the lesser the various available evils. The principle
that I think some of the smarter designers are trying to get into now is to
store the material in such a way that it can be unloaded from vehicles, more
than anything else—actually store it in vehicles. Up to a certain limit,
this is a more economical method of storage than anything else. There is
also the possibility that you get a certain amount of metering out of the
vehicles, not very accurate, but to a certain degree. The idea of using
an Atlas bin, I think, has been somewhat misused, because instead of trying
to use it as a distributor they have been using it for storage. I think
in some of the newer installations, I believe in Milwaukee, for example,
that the same principle is being incorporated as intended in St. Louis, to
use the material only for a very limited amount of storage, for only say a
few hours, with the idea that the principle function of the Atlas bin would
be purely as a distributor,
RIGO: In the study we just finished up for the Navy Civil Engineering
Laboratory, while performing tradeoffs we asked the question: What happens
if we want storage and we go to d-RDF plus some kind of conventional bunker
or silo system versus going with an Atlas bin that is capable of giving you
two or three days surge (looking at weekend capacity principally)? Based
74
-------
upon extremely shaky data on the cost of producing the d-RDF, and based
upon a series of assumptions on how you can handle the material, it looks
like the d-RDF plus conventional silo storage is roughly the same cost as
storing it loose because of the huge cost of the shredded waste storage bin.
I am throwing that observation out for what it's worth, recognizing that
it's preliminary.
FUNK: The biggest areas of concern I see with the Atlas bin: one, it's a
high energy user, and I think that's something we should address in all of
these storage systems, is look at how much energy we are actually using for
each function in all of these things we are looking at here. I think that's
a disadvantage. This floor wearing out—that's something to be handled in
design. As Mr. Wisely was pointing out, I think there's a lot of merit in
using this as a retrieval system or distributing system—a metering system—
which is essential for feeding certain types of burning equipment. That's
where we always get into our flap. I've got roughly in the neighborhood of
about 15 guys working on various designs, and every one of them come up with
a new and unique area in the way of handling this solid waste into storage
and out of storage. When it gets down to the final decision it gets back
to: Has it been demonstrated? Our clients are not necessarily in research
and development, and has it been demonstrated? The Atlas system certainly
has had more successful demonstrations than not, and of all of the avail-
able systems, I would say that it has to be ranked at the top. As far as
economics, we've gone through many evaluations where we put it into bulk
storage and used some kind of a bulk or front loader to move it into a
smaller distributing system; and when you shank out the economics, once
again it gets to be about a toss-up. What you've done is, you've tied up
usually more man-power, and that's the thing most of our clients would like
to avoid and stay away from that man-power usage.
WARE: I think it's appropriate to describe our system for the benefit of
those of you who aren't familiar with it and the way we came about deciding
upon it. We receive our material on the floor directly dumped from trucks,
and feed it as fast as we can and as soon as it's received into the shred-
der, and from there it's distributed onto the floor of the building—simply
dumped onto a concrete floor with a rotating conveyor. We store about 450
tons there. We have stored it as long as 8 weeks with no problems, no
severe odor, no fire hazard. From storage, obviously, there is a need to
meter into the rest of the plant. We store, incidentally, because the rest
of our plant runs 24 hours, where we receive only 8 hours. So we recover
from our pile of garbage with a front end loader and feed it into a unique
piece of equipment, which is the heart of our front end process, called a
doffing roll bin. This bin has sufficient capacity to absorb the surges
resulting from using a front end loader. It holds around 1,000 cubic feet.
From that bin the material is then metered by variable flow control into
our air classifier. We went with the bulk storage on the floor because the
material we were handling is shredded, very similar to Baltimore. At the
time when we were designing, they were having troubles and we didn't want
to commit to an Atlas. So storage on the floor posed no problems for us,
recovery is easy using a Bobcat or similar machine, and as I say, this is
75
-------
all complemented by using this doffing roll bin. The bin itself is borrowed
from the paper industry. It's a rectangular bin with a live bottom. The
material is fed in at the top, and it travels horizontally. The pile builds
up to a certain height, where it's leveled with screws, and the conveyors
across the bottom of the bin move the entire pile together. There you have
no feeding problems, normally associated with live bottoms; the entire pile
moves horizontally into a series of rollers with teeth on them. These rol-
lers are arranged in a near vertical row (actually diagonal), and as the
pile moves into the rollers, the teeth on the rollers pick the material
off of the piles, roll it overhead and down into a conveyor. We have only
demonstrated this thing for a thousand tons but we have found no discernable
problems with it. So by a combination of the refuse storage on the floor
and the metering bin, we are able to store, reclaim, and meter, without any
appreciable problems. The capital cost, I would venture to say, was about
half of the equivalent Atlas bin.
RYDER: I'd like to throw in my experience with the Atlas bin. Some gentle-
man over there said the City of Chicago had selected Atlas because we
thought it was best. Well, that isn't quite true. We don't select people;
we don't look over the whole spectrum of people and say, we like you. We
have to go out for bids and all other things being equal, must select the low
bidder. We had heard the same horror stories about Atlas bins—they get
plugged up and they have to dynamite the refuse out of it, and other kinds
of stories, but we did not disqualify Atlas from bidding as we felt that
they had one of the most workable systems on the market and we were pro-
tected, as our specifications included a full contract value performance
bond. We had to reject the bids twice before we got a successful bidder.
Either the bids far exceeded our estimates or the equipment did not meet
our specifications. The third time around, the low bidder was sufficiently
close to our estimate and agreed to meet our specifications. At this mo-
ment, the construction of the bins is nearing completion. We have two Atlas
bins—one about 60,000 cubic feet and the other about 90,000 cubic feet, or
about 300 tons capacity in one and 450 tons in the other. Tonnage, by the
way, is difficult to figure because of the varying unit weight of the ref-
use. Part of the difficulty with other manufacturers meeting our specifi-
cations was that we required that the feed rate from each conveyor leaving
the silo to be variable. This will allow the utility to regulate the flow
of RDF to each of the four boiler ports. Only Atlas seemed to be able to
come up with a viable system to do this. In any event, there is a per-
formance specification that the contractor must meet. They must not only
be able to retrieve the RDF at the specified rates, they must, be able to
store it. So if there's any problem with storage or operation, it's going
to be with the contractor. He is going to have to answer for it.
SPENCER: Is your performance contract directly with the construction con-
tractor, or is it with Atlas?
RYDER: It's with the general contractor; Atlas is the subcontractor.
76
-------
SPENCER: And your general has the responsibility for just that particular
item of equipment, or does he have responsibility for foundations?
RYDER: The entire contract runs about 4-1/2 million, and he is bonded to
the tune of 4-1/2 million.
SPENCER: O.K. so the whole receiving facility is 4-1/2 million with the
single general contractor?
RYDER: Yes.
SPENCER: And all the others like electrical, structural are subs to him?
RYDER: Yes, that's correct.
JOENSEN: One comment I would like to add (Bill referred to it now) is that
when we talk about a storage and retrieval system, I feel some adequate mass
weighing system for control purposes into the boiler should be considered
part of that system. Ames has a volume flow, the City of Chicago has an
impact plate position below the bottom of the in-feed conveyor. I don't
know what Milwaukee has, but in terms of the operation of the boiler, not
only am I concerned about the specification of the quality of the product,
but an adequate retrieval to indicate the flow rate. Doug Fiscus and I
have looked at a belt system used on Atlas bin, and we feel that should
be strongly considered. It's a more positive weighing system for control
purposes.
LINGLE: Let me see if there's a consensus here, in terms of the need, par-
ticularly for the Federal government to conduct development on storage
systems. Is the feeling that the systems we have are at least adequate and
that it really isn't high priority for the Federal government to either try
to improve them or better them, or do you think that there is a need for
some specific research programs in this area, and if so what are they?
FUNK: While we're talking about these storage and retrieval systems, let
me just throw my comments in on this. We're talking about people that are
ready and willing to commit today; we're talking about this as a long-term
venture, all the things that we are doing here. I definitely think there
are better ways of storing and retrieving. I just unfortunately haven't been
able to think of them myself, or I'd probably go in business. I think there
are a couple of suppliers that are offering new types of systems, and they
are a combination of different things. What Occidental just described is
something that's been demonstrated in several different cases, and I think
that has a lot of promise. We can't forget also that we're talking about
a specific need, and that's firing in a combustion unit that has to have a
controlled rate of firing. There are other applications, such as intro-
pyrolysis, where this fine control or supposedly fine control isn't neces-
sary. In that case you can consider some other. So there's a place I
77
-------
think that research would help: identify the combustion processes or con-
version processes, I should say, that and what their feed rate requirements
are. That could be something that could be identified for the industry,
and there certainly could be some research or at least some study given to
that point. Then perhaps some type of program to monitor the cost, the
actual operating cost of the existing installations, and especially if some
of these new ones get installed, so they can be compared against the Atlas.
In the business that we are in, we have to rely on the equipment suppliers
for information if it isn't available from any other source. Now we had the
job of sorting out what is true and what isn't true and then you have to
make comparisons, and usually they put it on such a basis you can't make
comparisons. You realize that going to look at things that are in question
and talking to the operators and piece that with the suppliers' information
to come up with your answers. One of the things that we've been toying with
within our firm is to put together a designer's handbook, that tells about
a lot of the things that are on this page of MRl's report. In other words,
getting into conveyors or anything else, tell them what angle of repose you
can have, or what angle of incline you can have on them, how fast they can
run, what type works better than others. The same is true in storage sys-
tems—dicuss the pros and cons. This type of information could go into a
handbook that I think could be of merit to the entire industry.
ALTER: You asked about research. Those of us who have worked with various
processes and fractions of refuse, particularly the organic fraction, have
all experienced the material bridges and hang-up easily. It doesn't flow
down chutes all the time; it stores oddly, as has been discussed. A good
deal of research is needed in the materials handling. We know through art,
we learn through art, that a chute or a bin has to be shaped a certain way
and then maybe the material won't hang-up. But there is no guidance doing
this; it is all empirical. There is not even a systematic approach to
solve the problem through art. You try it, it works, and you stop work.
There may be a better way of doing it, but you stop work because the par-
tial solution may be good enough. I think this has to be taken the next
few steps, people's experiences have to be recorded, or somebody has to
approach the learning process in a more sophisticated way. Before doing
so, you have to learn to describe the material and its flow properties. We
don't know how to do so.
HATHAWAY: I disagree with Harvey on one standpoint. The rigorous theory
of design and flow for the design of storage and flow facilities for pumped
solids was developed in the fifties at the University of Utah. The pro-
cedures for designing storage facilities, feeders, and some certain types
of conveying apparatus are relatively straightforward, and unfortunately
known by a select few. Our lab is looking into this. The things that we
don't know are the mechanical properties of RDF coming out of the various
unit operations. Some research ought to be done in this area.
ALTER: So we don't disagree. As I said, we need measurements of these
properties. We don't even have to know how to describe the materials so
we don't know how to apply the theory. I'm sorry to hear that the art and
the science of materials handling has been kept such a big secret. Maybe
some of the research is finding out where it's been kept a secret.
78
-------
LINGLE: O.K., does anyone have any other comments on materials handling
problems, conveyors, receiving systems, or anything of the like?
LAMB: On the practical side, the biggest problems I see with the conveyors
are the spillage and belt cleaning problems. We need good skirts. We need
ways to clean the belts to make sure that we can control spillage because
this results first, in a housekeeping problem, and then in a fire and ex-
plosion control problem. And we have several types of conveyors in the
plant that we are now operating and none of them have adequate skirts and
spillage is a very large problem. In describing its flow properties, I
think I can describe it in a word, it won't. That's a very good rule to
learn about municipal solid waste, particularly the shredded organic
fractions thereof: they simply do not flow, and you must diverge, don't
converge.
ALTER: Little story out of our pilot plant. For a while as a general
rule, all conveyors had self-cleaning head pulleys—something we learned
from the scrap industry. But then we found two things. One, if the ma-
terial was too much like gravel (I'll call it that), but even organic, we
found that the self-cleaning head pulley would cause the material to dance,
and then the performance of the entire conveyor belt design changes. We
replaced pulleys. Two, we found that a self-cleaning head pulley is not
good enough in that there appears to be some sort of relationship between
its diameter and the self-cleaning ability for a given feedstock. This is
particularly true, we found, on air classifier heavy fraction, because it
contains rags which hang-up on some sizes of pulleys and not others.
There's no design guidance here; it's a question of trial and error. These
are the sorts of things I think have to be recorded so that the experiences
can be shared. The design handbook, Harvey, I think is going to be a nar-
rative for the first edition.
RIGO: On those problems you are pointing to, if you get a particularly dry
load of material, for example, a pure shot of cardboard or wood chips, what
would normally go up a conveyor at a fairly shallow angle of incline, will
not go up it any more. Vista Fiber and Chemical Company had an experiment
using a tree sprayer that pumped water out at the rate of about a gallon
for every 10 tons. The dry material tended to go up. Just coming out with
some design guidance on mixed municipal waste is not enough if you have a
plant with the probability or likelihood of getting industrial waste, which
comes back to compositional guidance. You are going to have conveyors that
work for 90 percent of your material but won't work for 10 percent. There
has to be some way of accounting for this, either a general design or some
kind of attemperating system that can be cut in and out.
WISELY: Just for the benefit of those of you who have not heard of this
phenomenon, I think that you should be aware that if you install a vi-
brating conveyor with a certain frequency and stroke in a pit, you are go-
ing to have a little trouble. As some of you may know, in the initial in-
stallation at St. Louis we had a really magnificent piece of equipment that
79
-------
was designed for a variable stroke but a constant frequency, that was de-
signed to accommodate from zero up to about 60 tons an hour of raw waste.
It really worked, except that the frequency was a direct multiple of the
natural frequency of the building height and the building width, and this
was accentuated by the fact that it was within a pit. There was no way of
predicting this. We had this analyzed afterwards, but. there was no way of
predicting this in advance. For goodness sake, don' jut a thing like that
in a pit.
CHANTLAND: One of the problems mentioned was the problem of skirts. I
would like to add one other problem, and that is the drop of materials from
one conveyor to the other. When it drops out of the shredder onto the
conveyor, and the conveyor drops on to another conveyor, a certain component
of the grind is going to be a dust fraction, and whenever it drops, the
air picks it up from the hopper, and as material goes down, air has to come
out.
SPENCER: After air classification, do you find that that also is a problem?
CHANTLAND: In our particular case the light particle is no longer exposed,
so we are dealing with the heavy fraction after air classification, and
there is only a small amount of dust, and in our case I would say most of
it is associated with the bypass screen on the air classifier. You do get
some fine particles coming through this screen, which is trying to take out
the fine glass on the air classifier.
LAMB: Dave, we find dust at every transfer point, either before or after.
You don't elutriate the dust, if that's the question to which you are al-
luding in the air classifier, at least that hasn't been our experience.
We have not elutriated the dust, and then had a dust-free operation at sub-
sequent transfer points.
POWERS: I'd like to make one observation we had in building Recovery 1.
The thing that I've found, in almost every instance every piece of equipment
that we use in refuse handling has been bastardized from some other piece of
equipment. The shredders were originally designed for something else; the
conveyors were not designed for MSW; all these pieces of equipment are
retrofitted, massaged around, and put on the shelf and called a standard
piece of garbage equipment by the industry. I think that if we can get the
industry to start designing equipment specifically for refuse processing,
that would be a big step.
SENDERSKY: The industry, in some cases by our contacts with them, are
willing, in some cases anxious, to improve their equipment. Unfortunately,
however, they don't always get all of the information back that they need
to improve their equipment. I said earlier this morning I think there has
to be more of an interchange of information between users of equipment and
manufacturers of the equipment, etc. I would, incidentally, propose for
80
-------
EPA and perhaps for some others a suggestion for a conference of some kind
in which equipment manufacturers, system designers, system users, have a
general get-together, so that there's generally more of an exchange of in-
formation. I found a number of equipment manufacturers interested in im-
proving their equipment, and a number already have, specifically for MSW
app]ications.
FUNK: A couple of the manufacturers have started to respond to the solid
waste area and have published handbooks or design manuals that are so
common to the rest of our industry that have addressed and have [unintel-
ligible] solid waste handbooks. Mayfran is one that has put a book out
like this, and they have some rules of thumb about designing. They went
around to the industry and looked to see what had been done right, what
had been done wrong, and got suggestions from a lot of different designers,
and they have listed these. I think that's a step in the right direction.
It certainly isn't an answer, and I'm not saying they didn't just take some
other piece of equipment they already had and now call it a solid waste—
I'm not saying that, but they have started to put some of this in writing
for general distribution to anyone that might want to get a hold of it.
They're not the only one. We've got, I think, about three books now and
catalogs in our files that have this type of information in it. So I think
they are starting to gear up to this thing.
FISCUS: Bob Power's comment was very interesting. It seems that what we've
got to date is either pneumatic conveying or belt conveyors. So if you look
at starting from scratch designing a system to handle x tons per hour of
some sort of shredded material from Point A to Point B, what would you use
specifically other than the belt conveyor or a pneumatic conveying line?
The reason I asked that, it's a leading question, I get this kind of ques-
tion bounced off me quite often. People say there is a need for this, and
I haven't yet heard some good suggestions. What are some other types of
conveying mechanisms that can be used?
POWERS: I can't give you an answer there. I do know that every piece of
equipment we have in the plant, we modified. Every one. I can't think of
a piece of equipment that has come into that plant that we haven't massaged
around, rebuilt, redesigned, sped up, slowed down, we've done something to
it to make it work. This is something you can't put in a text book, or in
a handbook, you just have to feel it and you have to work it. I think that
if we can get the manufacturers to be aware that there is a large, large
industry out here that needs good equipment, that needs rugged equipment,
they'll start to turn it out. They have apparently made [unintelligible],
that's the conveyor. Now the shredders, you know they are interested in
automobiles, and rocks, and things like that and not necessarily garbage.
RYDER: I will agree with Bob on that. I think if you were to write a spec
today and consult all of the manufacturers as to what you should put in
that spec, they would tell you all you need is tell me how many tons per
hour you want— 25, 50, 75 —and we'll give you the equipment to move it,
shred it, classify it, or whatever.
81
-------
LINGLE: Does anyone have any comments on pneumatic systems? There hasn't
been too much said about that. That was at least one of the problems that
seemed to occur in early RDF systems, the wearing out of the pipes and so
forth. I know there have been some pluggage problems, I believe there have
at Ames. Would anybody care to comment on pneumatic conveying systems?
CHANTLAND: I would like to make the comment that we do have one pneumatic
system for the plant—the Atlas bin—and four pneumatic systems from the
Atlas bin to the boilers. Probably one of the big reasons we selected this
type of system, was from the dust control standpoint of transporting this
from point to point. It's one thing to put up with dust inside of the
plant; it's another thing to put up with dust you breathe outside the
plant, as we had observed in some installations. Since we sit right in the
downtown area and we just couldn't have the type of an operation that wasn't-
relatively free of dust. We found from the St. Louis experience a^out
wearing out of the elbows. We put in an elbow with a replaceable liner.
The liners, of course, will still wear out; that's what they are made to
do. The question is really finding an optimum point; how much money you
want to spend for a particular liner versus the time span that it will take
to wear out. I'm not sure that we totally have that answer yet at this
time, but the material itself will move essentially in suspension, and it's
only at the corners and the elbows where the material really comes in con-
tact and wears the pipeline. Since this is a pneumatic system, wear also
takes place at the air lock feeders. You have to lock this material through
one point into the end of the line, and as mentioned before, oversized ma-
terial can jam air lock feeders, and this is also a point of wear. A cer-
tain percentage of the material has to be cut, or separated, as the air lock
moves around. This is a wear point, and there are replaceable blades and a
certain amount of maintenance that is needed.
SPENCER: Have you considered soft rubber air locks as opposed to the harder
Ny-Hard type?
CHANTLAND: We don't have any of that type for the pneumatic system. Now
we do have that type on the air classifier and it works very well. We
would be very interested in anybody who has the experience with the use of
the soft system. I do understand that the newer air locks come out with not
only replaceable blades, but they also have a replaceable diagonal vane on
the rotor. It also has a replaceable shield or liner.
LINGLE: O.K., one more comment on this, and then I think we ought to move
on to the next subject.
JOENSEN: There's been some sweating inside the Atlas bin, and the municipal
utility makes it a practice every Monday morning to run the two conveyors
that have not been operated the previous week. This cleans out and pre-
vents any solidification. The sweating caused some drifting down into the
bottom of the conveyor, and since the grit has a tendency to accumulate on
the bottom, there was a sheet of ice formed. By running the two unused
82
-------
conveyors about once every three or four days, this was a minor operating
problem that was solved right away.
LINGLE: Why don't we move on to size reduction now? Perhaps we could start
this off by having George Trezek present some of the slides that he has on
the work that he's done on shredding.
TREZEK: Before I get into the slides, I would like to preface my remarks
by giving a little bit of history of how the project got started. About
1969 there were a few shredder installations which were just being put
together. EPA was interested in trying to learn something about how
shredders operate; specifically, what governs their performance. I think
that all of these machines were really designed for brittle materials,
that is, for crushing rocks and these types of things. When you review in
the brittle material literature on comminution (a fancy word for size re-
duction) , it turns out that there's a whole host of laws and theories, but
when you start dealing with refuse, which is only about 25 percent brittle,
you have an entirely different situation. One of the first things I did
was to survey the shredding installations at the direction of the project
officer at that time. The purpose was to try to obtain comminution data
from the existing machines. However, it was very difficult to accomplish
this in the field. We decided then to set up one machine in one place in
which you could monitor all of the things going in and out, and do a very
controlled study. First, it was necessary to ascertain what kind of
things we should be trying to measure. Secondly, conduct a series of
measurements under controlled input and output conditions. What I'd like
to show you today are some of the results that we've obtained over a period
of time, and then talk a little bit about some of the kinds of energy loops
that you get into when you start arranging processes in different orders.
This is the laboratory as it's set up. This, of course, was really not
intended to be a size reduction laboratory when we started out. I con-
vinced the University administration to let me use it in order to set up
the facilities. The shredder is a 10 ton per hour Gruendler hammermill
which according to today's standards is a fairly small machine, but at the
time it seemed like it was a fairly reasonable size machine to begin to
study. We changed the drive system of this machine to operate at a variable
speed, because one of the things we wanted to study was what effect the
rotor speed would have on the size distribution of the product that comes
out. A belt drive system is used. Normally the motor would be connected
to a flexible coupling directly on the shredder shaft. Refuse is loaded
with a front end loader. The conveyors are also designed so that we can
vary their speed and study the effect of feed rates. This machine, by the
way, used to be at the Johnson City Compost Plant, and when that project
was phasing out, the EPA decided that we should use it for our studies.
Consequently we set it up to run these comminution tests. I might just
point out that when we received the machine it was in reasonable shape, but
when we tried to replace the hammers, I called the factory one day to ask
them how to get the hammer pins out. It wasn't clear how that should hap-
pen. We were told you sort of burn them out and then purchase a new set.
83
-------
Here's basically a conveyor discharge of shredded material. The air clas-
sifier is in back; I won't get into that at the moment. Let me summarize
what it is we're looking for. What we want to know is: Given an input
size distribution, can you predict what the output size distribution will
be for a particular set of conditions? Because if you can control the
output size distribution, you can then design some other follow on pro-
cessing equipment. We want this sort of general theory which will describe
first the comminution mechanism. Then we'd like to verify the theory with
experimental data. We want to be able to predict size distribution. The
current literature (this was three years ago) dealt primarily with homo-
geneous brittle materials, and did not deal with heterogeneous materials
like solid waste. Traditionally when you looked at theories that described
size reduction, they were of two types, the so-called continuous theories
and the discontinuous theories. The continuous theories involve a relation
between energy and size, i.e., dE _ , n
~l— — — C/x
dx
where E is the net energy required per unit weight in a given process of
comminution, x is a factor denoting size, n an exponent, and C an index
which depends mathematically upon the unit chosen and physically upon
material properties and mill operating variables. However, it is difficult
to express "size" in a mathematical form. In any actual process the mater-
ial being comminuted has a considerable range of size on entering and
leaving the process and also it is not possible to denote this size by a
single index because of the variations in the shapes of the particles. The
three most frequently quoted laws are derived from the above general energy
relation. These are attributed to Rittinger, Kirk and Bond, when the ex-
ponent n is equal to 2, 1, and 1.5 respectively.
The discontinuous theories, which are the ones that we eventually
adopted,are essentially a statistical treatment and involve two things
which I will get into a little bit later—things called a breakage function
and a selection function. What those come down to are—in any given pass
when the material enters the machine, a certain fraction is selected to be
broken during that pass; that's the selection function. The breakage
function says it will be broken according to some size distribution. This
summarizes some of the size distribution relations. These are the Gaudin-
Schuhmann equation Y(x) = (x/k)04 which relates the fraction of undersize
Y(x) (% passing) to the relative size x/k. In this relationship Y(x) is
the cumulative fraction by weight finer than the stated size x, k is a
quantity called the size modulus (theoretical maximum size), which together
with the slope parameter<*• characterizes the product. The Gaudin-Meloy
size distribution in relation for single fracture has the form
Y(x) = 1 - (1 - x/x )r
o
where XQ is a characteristic dimension of the feed particle before fracture,
84
-------
r the size ratio which is a measure of the number of breaks in the particle.
The most useful size distribution selection for refuse is the Rosin-Rammler
equation, i.e.,
Y(x) = 1 - exp[-(x/x )n]
o
where n is a numerical constant and XQ is a characteristic particle size
(the size at 63.2 percent passing). The quantity x is an extremely useful
parameter in correlating the comminution parameters. It is important to
note that the 63.2 percent passing quantity is obtained by substituting
above relation, namely
Y = (1 - 1/e) = 0.632
x for xo in the above relation, namely,
Here for example, are some typical size distributions that were obtained for
different types of machines with a raw refuse input (Figure 1). The size
distribution for raw refuse is shown on the right. Shown are some data
taken from our lab which are accumulated at different speeds. They fall
into this general band. Here's some data taken out of the Eidel shredder
which produces essentially a finer size distribution. Now the idea is to
see what we can do with this kind of information. I want you to observe
one thing. If you take an input distribution and go through a comminution
series of events, this would be the average size distribution that you
would see with the machine running at 1200 rpm. We take that same size,
and use it as the feed, and run it through the machine one more time to
produce this secondary size distribution (Figure 2), which represents a
movement toward smaller sizes. The characteristic particle size is smaller,
x0, which becomes the feed for the secondary shred. We then move to this
tertiary system, and you can see there is a diminishing size with each
pass. Now the interesting thing here we need to remember is the difference
in power consumption which was 12.5 kilowatt-hours for the primary shred
(Figure 2). Getting from there to there and regrinding and achieving a
smaller XQ, we expend about 5.6 kilowatt-hours per ton, and so on down the
line. That's what I'd like you to remember about this particular data. Here
is the same information (Figure 3). What we wanted to see were how some of
the variables affect the kilowatt-hours per ton for power consumption.
I must admit there is some scatter in the data, but nonetheless, it was
what we found. After a lot of thinking about it, it turns out there does
appear to be a minimum power consumption at about 35 percent moisture
content. That is interesting because you can optimize where you should
run the system. Higher moistures gave greater power consumption. The
thing that I think is important to realize here, is that if you are running
the machine at 1200 rpm, you would tend to follow the dashed curve (Figure
3). If you are running at 790, you would tend to be following this lower
solid curve. For example, at 5 tons per hour, the power consumption dif-
ference is approaching 50 percent. You ask, well, that's great but what
is the consequence of that? The consequence is that you have produced a
particle which has a slightly coarser x or slightly larger x . However,
85
-------
CO
LU
X
o
LU
M
CO
LU
LU
tr
o
CO
01
S-l
oo
ONISSVd
86
-------
o o o o o o o
O O O O O CT>
CM CM CM CM CM r- c-
IO IO IO IO tO CO CD
IO lo ID — — — —
a a
o
o
o
ro
9 w
— UJ
X
o
UJ
N
cn
UJ
UJ
cc
o
cn
CN
-------
IN30a3d
-------
if that material is going to be used in some other process down the line,
it doesn't matter, except you can achieve a fairly substantial savings in
energy at this point. This (Figure 4) is the specific energy in kilowatt-
hours per ton versus the characteristic particle size, x0> and interest-
ingly enough the machines tend to fall along a curve. You get some types
of shredders giving finer sizes up here; traditional hammermill type machines
fall in here; and another type of grinder called the Rabco system falling
down here. This kind of information can be correlated against grate spac-
ing, so we get a relation between energy and grate spacing. I might point
out, interestingly enough, that when MRI did a study a while back, they took
the data from Ames, I believe, and plotted it on this curve and it fell on
the curve, which was comforting.
Now let's get into a little bit of modeling that produces the size
distribution curves. Basically, all the model says is that you select a
certain fraction of the particles to be broken and that represents the
selection function. The key is the product that you operate on. There
are various sophistications of the model. The simplest is called IT break-
age, which says the material goes in, a certain fraction is going to be
broken in each pass, i.e., in matrix terms, S = irl and the equation de-
scribing the process is
p = ]irB+(l-ir)l| f
Here S(y) = IT for all sizes y; that is, a proportion TT of the particles
independent of particle size is broken according to the breakage function
B(x,y). The parameter TT completely describes the process if B is assumed
to be known and is a useful measure of the breakage effected by the process.
The repeated breakage cycle model is also useful. This model depicted
by
P = { rr (BS, + i - s.)} f
k_, h h
describes the situation where the feed and product sixe distributions do not
cover the same sizes, and the size ranges differ considerably. In effect
the breakage process is considered as a series of cycles of mild breakage.
Here h = 1, 2, N refers to the hth cycle in which the probability of
breakage is S, .
h
Two forms of the breakage function were selected as being appropriate
for refuse communition. They are a modified form of Broadbent-Collcolt.
B(x,x ) = {l - exp [-(- )n] } /[l-exp(-l)]
where n is some positive index. Experimental results indicated that certain
products of the size reduction conformed to the Rosin-Rammler distribution
89
-------
IS3HON1)
11X3 dO 9NI3VdS
in sr ro oo
i1 r^r
y- n | • j | p
0
Q
i
LU
O
UJ
D
Q
or
<
Q
2
O
CD
i
_J
UJ
X
<]
K
Id
_J
O
T^
C3
ZD
DC
0
O
O
O
CO
<
Q:
*
0*
0
rO
i
85
UO
CO
h-
LU
l«,
"P*
O
C.)
UJ
(T
H3
00
b
s
UJ
CO
ID
U_
U.1
c:
0
(T
LJ
M
UJ
!-—
P
en
0
— 9
fc—
o
Q.
CO
LJ
or.
o:
o
c.^
o
t£>
it
IU
o
z:
LU
c:
UJ
u..
UJ
rv-
U_
>~
U)
o:
UJ
2
UJ
0
X
\
Q\
\
1
q x°
cu "-'
q
CD
0
(D
6
sf
o
ai
|SJ
en
LU
_J
o
H
tr
<
CL
0
(__
w
tr
UJ
H
o
<
-------
with an n index varying between 0.845 and unity. The other form was pat-
terned after the Gaudin-Meloy distribution equation, i.e.,
B(x,XQ) - (B(x/xo) = 1 - (l-x/XQ)r
The relation was necessary since the first relation for B did not simulate
primary grinding.
We've done a lot of work trying to apply these models and let's sum-
marize. Given a product size distribution, you can in fact using IT breakage,
predict fairly reasonably primary, secondary, and tertiary grinding (Fig-
ures 5 and 6). I will leave it at that and not try to go through all the
curve fitting. In some cases we don't predict at all. Some models ac-
tually do not work at all (Figure 7). We've gone through a lot of studies
of this nature, and deduced that we can, with some modeling, predict what's
going to happen. We also monitored carefully the wear of the hammers.
Basically the thing I want to point out to you is that when you start run-
ning the machine at lower speeds, you do, in fact, get a considerable de-
crease in the wear that's going to occur. I think that is a very important
point which is overlooked. Along with this size reduction, we've done some
work at looking at different kinds of screens and how they affect screening
efficiency. We've looked at flat vibrating screens, rotary screens and
we've looked at trommel screens. Here is a plot (Figure 8) of screening
efficiency as a function of feed rate. For example, trommels tend to fall
into this group, and flat bed screens tend to fall in this group.
We've also taken this ground material and looked at the size distri-
bution of all the components of the materials and the incoming size dis-
tribution, sorted them all out, and measured the size distribution of all
of the components in the waste stream (Figures 9 and 10). There's a number
of ways you can look at it. Here essentially is a histogram (Figure 11)
sort of plot versus different size spectrums. This was a breakdown of
material that was in there to start out with—heavies and lights. We have
also looked at the light fraction, and have broken it down into a size dis-
tribution (Figure 12).
What this is leading to is the work that we've gotten to now on finer
secondary shredding. Experiments that we are now conducting relate to a
fixed hammer machine, which is a high speed, fixed hammer machine with
grates. Actually this is a light fraction put through that fixed hammermill
and extruded in a California pellet mill (pelletizer). Here are some pel-
lets which were produced from this fine secondary ground material. Our
attention has been focused on examining how to arrange these systems. I
would like to call your attention to the three basic kinds of things we have
examined so far (Figure 13). One possibility is to say, suppose you just go
and try to make a product finer and finer, down to one inch, and go to an
air classifier, cyclone, screen and then a product. The other is to go
through a primary shredder, air classifier, screen, and then fine grind and
make a product. Another way of doing it is to put two shredders back to
91
-------
\
\
°^
w **
= ss
(/> Z! W
— O UJ
z z
o o
h- p
N uj u
" ju2
X
z
_
o
f-
^
n
UJ
01
0004
K>
a»
o
»-
«
^
UJ
y
^,
O
-4
K
o
c
O)
CO
o
^
^
•/I
S
Uj
>
^
5
0
O
0
o
^
CO
o
m
0025
(0
in
W)
O
o
*-
~~z~
o
UJ
3
^.
Q
O
v
i
^
o
o
o
tg
o
0
o
it
^
o
o
w C
"^
>-
J
2
a
1
•
1- ^
*• *" n
tr o !JiJ
5 et ~
° ~J o
£„
—
H
>
tr
a
a.
o
CTJ
z
13
- _
O
rO
*n
o
M
>?
o
OJ
CJ
^_
i'
o
j?
*•
a:
>
^>
a
—
\t~.
— -
in
O
tn
to
r-
-?
s»
4-
O
a
o
it
a
V
or
<
cr
ut
ir
CO
in
<0
^
O"
O
tM
Q
Ui
u*
tr
4> w
||
U. 3
92
-------
\
•— o ui
V} It
§5
fc u) <->
"" ^ §
x
0
Z
z
o
l_
o
00328
—
0
0
-
0
u
2
2
5
3
ID
0
O
CJ
r>
o
c
Q
.1
£v
DO
D -I
CM
O
0
II
c
a
O
t- w
mo
QO
•a -i
D_J
DO
—
O
0
0
II
o
X
Q
o
00073
O
w
o
o
c
c.
O
s
y.\-
OJO
oo
no
GO
6
o
o
~
o
0
0
-
o
S
z
si
2 3
— .0
id •%
«« a
8?
.X
T3
a
K
be
n
3ZIS 031V.LS NVH1 M3NIJ NOI1DVHJ 3AIIVinWfO
93
-------
OOlr-
>
K
<
0001
. D FEED, RAW PACKER TRUCK REFUSE
/ -•- EXPERIMENTAL, PRIMARY GRINDING
1 I I
EQUATION
6ROADBENT-
C/ILLCOTT'S
INDEX
n : 0 345
K = 1
K -2
K=3
1 1 1,1
(ir)
SELECT'CN
FuNCT OH
1 Z7
iSC)
SOUlfiED
RESIDUAL
07563|
098 JOOBZ9
0 90 j 0 2 1 G 4
I
L i ,_1. „
001
10
01 10
PARTICLE: SIZE, INCHES
Figure 7 Comparison between computed and experimental prod-
uc' -,ize distributions for primarv grinding using the ;r-k Breakage
Process Model.
94
-------
(00
u a
90 -
80 -
70-
60 -
50 -
40
ROTATING CYLINDRICAL SCREEN
INCLINED FLAT DECK
VIBRATING SCREEN
15 20 25 30
f 8 The effect of mass flow rate on the screening efficiency of the air classified
frai tion.
95
-------
iDO (—
I 0 -
0 254
(001)
254
(0 I)
254
(I 0)
254
(10)
SCREEN SIZE.m x 10 ,(IN)
Fig.9 Component size distribution of shredded municipal solid waste.
100
o 10 -
O.I
OZ54
(001)
2 54
(01)
254
(I 0)
254
(10)
SCREEN SIZE.mX |Cf3(|N )
Fig]_QComponent size distribution of air classified heavy fraction.
96
-------
K
Q
O
o?
HEAVY FRACTION |
LIGHT FRACTION [j
jJlJj
'16 -095 '051 '028 '013 -Q 15
(«%) t'3/8> (» 020) (.01 I) (.005)1-005)
SCREEN SIZE, m« I0^2 , (IN )
Fig. IJComposition of air classified shredded solid waste, 70/30, split.
100,
'r
10-
O..L
J
0254
(001)
2 54
(01)
354
(I 0)
254
(10)
SCREEN SIZE.mx 10"^ (IN)
Fig,12Component size distribution of air classified light fraction.
97
-------
A
LU
h
(O
O)
o
in
ui
O
O
c:
a.
U.
Q
t
l_
o
-------
back which have the same grate spacing, and then going into product. So
I'd like to show you what happens when you do that. The dash line around
the screen indicates that we'd like to put the screen in and take the
screen out and see what happens. In trying to figure out how to represent
this kind of information, I drew these graphs (Figures 14, 15). We plot xo>
that characteristic particle size, against the cumulative energy in kilo-
watt-hours per ton. This first one would be the solid line, then we have
the dashed line and then the crosses. Here, for example, if we took a
primary shredder with one inch grates, you would essentially move along
this bottom solid line, and reach an xo of about 0.22 inches, the charac-
teristic particle size, for a machine with one inch grates. This is now
air classified, so then we'd move through the air classifier up to a larger
xo because we've been dropping out material in the size distribution. Some
energy is expended in going to a larger XQ. We then go through the next
stage of grinding. It depends now on what final xo we want to achieve. At
this particular xo (0.4 inch) we would follow this path; likewise at this
path a smaller XQ (0.3 inch). You can start to get the feel for the rela-
tive amounts of energy which one has to expend to achieve various XQ.
The interesting thing about this is that when you connect up these end
points, they don't fall into a straight line (Figure 16); in fact, what
they do is tend to go up in a parabola, so as you get past Point 1 the
energy that you need to prepare this material starts to really increase.
What you are asking is, what minimum can I go to to achieve what I need
without further expending energy. Here is a similar kind of chart if you
do the screening (Figure 15). Essentially you come out with the same kind
of thing. You follow along until you get to the point after primary shred-
ding and work your way up the curves. The interesting thing about this is
that even though you have to expend maybe 100 to 200 kilowatt-hours per
ton to get this material, this is only about 5 to 10 percent of the energy
content of the material that you're working with, so in that sense though
it looks bad, it's not quite that bad. Now here's the correlation I want
you to see—the correlation between xo and xgo% (Figure 17). The people I
talked to in the industry like to think in terms of the size at which 90
percent passes, and I think that's a good way of doing it also, except that
we find that the data and power consumption tend to correlate better with xo
at 63.2 percent passing. It's not too bad down here at the small sizes but
when you're here (the larger sizes), the data tends to be scattered. One of
the things that came up here this morning was the idea of cleaning up the
fuel, making it as bad as you can tolerate. It's interesting because when
you look at this, it turns out that moisture content does not separate as
material content separates, so by doing air classifying and screening you
are in fact using the moisture content as a screen. And if you take a
look at coming in screened or unscreened, there is a difference (Figures
18 and 19). When you relate this back to the energy that it takes to
vaporize this material, then have a comparison of energy lost by trying to
not screen the material. Another way of looking at it is, take a look at
the heating values of unscreened material and screened material, then you
can also look at the ash content between unscreened and screened—these
are for pellets we may burn. You do get a significant change in the ash
produced. Another way of looking at it is basically on the dry weight
heating value, and you can get a feel for basically a change in the energy
content of the stream and that's pretty much related to how much moisture is
in the waste stream.
99
-------
180
160
140
120
100
80
GO
40
20
ENERGY CONSUMPTION OF VARIOUS
PROCESSING SCHEMES TO OBTAIN
A GIVEN AIR CLASSIFED LIGHT FRACTION
PARTICLE SIZE
0
O.I
1.0
X0, INCHES
Figure 14
10
100
-------
700
600
5pO
400
CO
X
o
o
LJ
Z
UJ
UJ
tr
o
Crt
u-
o
z
o
300
200
o
a:
UJ
z
UJ
100
ENERGY CONSUMPTION OF VARIOUS
PROCESSING SCHEMES TO OBTAIN
A GIVEN SCREENED LIGHT
FRACTION PARTICLE SIZE
PRI. I
SEC. 2^
O.I
1.0
INCHES
10
Figure 15
101
-------
70
60
o
01
o:
UJ
40
30
o
u.
o
UJ
o-
IO
ENERGY REQUIREMENTS FOR
FINE GRINDING
LIGHTS
-SCREENED
LIGHTS
X0, INCHES
1.0
Figure 16
102
-------
RELATIONSHIP OF
AND X
GRUENOLER HAMMERMILL
2.0 -
in
LU
x
o
.. 1.0 -
o
O)
X
0.0
AVG. 2-jj- GRATES
X90= 2J45
AVG. I" GRATES
°-096
I i I I
I I
1.0
X0, INCHES
Figure 17
103
-------
Comparison of Unscreened and Screened Fuels
Heating Value,
Dry Basis
Heating Value,
As-Received Basis
Unrecovered Heat
in Exhausted Water
Vapor, As-Received Basis
Net Recoverable
Energy on an As-Received
Weight Basis
Unscreened
jBTU/1 b) _
6,733
5,218
254
4,964
Screened
(BTU/lb)
8,000
6,696
184
6,512
Ash Production of Coal and RDF Fuels
Heating Value (BTU/lb)
Ash Content
Ash Production
(Ibs/MMBTU)
Unscreened
Light Fraction
5,218
20.9%
40.0
Orient
Coal
11,579
7.6%
6.6
Screened
Light Fraction
6,696
12.1%
18.1
Figure 18
104
-------
Heating Value
BTU/lb dry solids
Moisture Content.
BTU/lb, (as-received)
% decrease
Decrease in Heating Value
Due to Presence of Moisture
Unscreened
Light Fraction.
6,733
22.5%
5,218
22.5
Screened
Light fraction
8,000
16.3%
6,695
16.3
Energy Loss Due to Unrecoverable
Heat Associated with Moisture Content
Vaporization of two Refuse-Derived Fuels
Moisture Content
Wate*-/Solids, weight basis
Energy loss, BTU/lb water
Energy lost in Exhausted
Water Vapor, BTU/lb dry solid
Energy lost in Exhausted
Water Vapor, As-Received Basis,
BTU/lb
Unscreened
Light Fraction
22.5%
0.290
1,131
328
254
Screened
Light Fraction
16.3%
0.195
1,131
221
184
Figure 19
105
-------
ALTER: People may wonder about the applicability of what George is talking
about on shred size to the real world. We've looked at the particle size
distributions of 10 shredders around the country. (One was originally de-
signed for OBW, but we put packer truck refuse through at the time; another
was designed for automobiles but we put packer truck refuse through it.)
For these 10, we measured particle size distribution. (We had as few as
seven data points for one shredder, but as many as 55 for another.) The
statistical fit of the particle size distribution was very good to the Rosin-
Rammler distribution function, as George said. We were able to compute the
Rosin-Rammler parameters, which means now we have predictors for the PSD.
There's a lot more to this, but one very significant point—there is a num-
ber, which George alluded to, called the Bond Work Index, which is the pro-
portionality constant between the energy to shred and a function of particle
size. This was developed for friable materials, and we were bold enough to
test it for refuse. Within about plus or minus 20 percent we get the same
number for the Bond Work Index for the refuse shredded in all 10 shredders.
The number, by the way, for those who are interested is about 400, which
is huge, it's tremendous, and we don't quite understand it. We've also
measured the Bond Work Index using a small Heil-Tollemache machine with
hammers; the importance of the hammers will be evident in a moment. We
measured the Bond Work Index for some of the components in the refuse.
and you do this by measuring energy input and particle size of the product
and of the feed. We ran aluminum through a couple of times and looked at
the primary shred, the secondary shred, and the tertiary—glass, steel and
paper—and the highest value we got was for paper, which is about 200. All
the other materials were lower. I don't understand why we got 200 and the
refuse was 400. Thus, combining the particle size and the Bond Work Index,
and if you can apply the theories for hard materials to refuse, you predict
a relationship between power input and what George calls xo. We are start-
ing to try to investigate this from the same point of view on secondary
shred of light fraction, which is the material feeding our pelletizer. So
far, (and here's where practicality may override all of the theories) the
Heil-Tollemache doesn't shred textiles. They all went through and plugged
the pellet mill. This shredder has since been changed (and maybe tomorrow
I'll show you a picture of it) so that rather than hammers, the shredder
has stators and rotors. It is more like a scissor and does a fine job of
cutting the light fraction, including the textiles. Finally, regarding the
information from Berkeley on the light fraction, its heat content, the
effect of screening, etc., I'm delighted to see it. We have found the same
things including reducing the ash below 12 percent, more like 9 percent. I
think potentially it can be reduced even further. The effect of moisture
in the fuel and the issue of drying we predicted in a paper (that has since
appeared) where, for example, we computed the amount of RDF that would be
necessary to dry the remainder. It is staggering. What the message really
is, is you don't have to dry it. I'm glad we're in such good agreement.
TREZEK: I'd like to point out one other thing I forgot to say. When you
look at those energy pathways, it would appear that what you want to do is
have one machine and get down to where you want to go in one pass. Now
that may be all well and good, except that we don't know what the wear
106
-------
characteristics are going to be if you do that kind of procedure. Although
from an energy expenditure point of view it looks very promising, from a
wear point of view, what we need to correlate next is the wear versus this
x0, and then we'll have kilowatt-hours versus wear. Then we'll really have
the whole ball of wax completely put together. We're almost there but
there's still other things we should look at.
HASSELRIIS: I wanted to make some comments on Dr. Trezek's presentation if
it's appropriate now. Can you show that previous slide of kilowatts per
particle size, could you backtrack on that? The major observation that I
would make about these tests, is that they are all basically for a range
of moisture content from, say, 15 percent up. Now when you start to shred
materials with a lower moisture content, you run into an area where the
material gets tough and very, very hard to pull apart, and then you turn
over the hump, and you get to the other side of the world, where all of a
sudden the material gets brittle and you are on the opposite side of it.
But actually, I notice that in extrapolating this curve out to here is about
where we find the kilowatts per ton is for brittle material. Somewhere over
here not way out there, it's over here somewhere, like the 80 percent line
is 200 mesh. In other words, you are talking here of using a shredder,
hammermill, or any type device like that. I'm raising the question, the
possibility, of jumping right through that machine, not going that way,
and ending up with the total consumption for the whole plant in the vicinity
of, say 40, 50, something like that, the whole operation, say, 30 or 40
kilowatts per ton. This we know from a lot of test data can be done, but
it requires a process which gets into the other side of the hump. You get
into that region—it gets very, very tough. There's another curve you have
that's got the moisture content, and that also doesn't show what would
happen if you went the other direction. It shows it going up here, but then
when we go further back [unintelligible]. You shouldn't fail to know
there's another part of the world you're working in, and that comes up under
the subject of embrittlement, which it's necessary to accomplish particle
embrittlement to do that.
TREZEK: You don't get that without an expenditure of energy, right?
HASSELRIIS: Yes, but in the vicinity of 30 kilowatt-hours per ton, is what
you end up with. That's why I say it's really an extrapolation on the low
end of your scale. What do we have here? 20 or 30? Extrapolate this,
there's 15, 10, 5. That's not a bad extrapolation, as well as you can tell
those points.
[Unintelligible question].
HASSELRIIS: Yes, you have to use a different operation. This curve doesn't
show the particle size you are creating here; that's a variable that's mis-
sing, and it would have helped to explain this.
TREZEK: You have to go through this curve here. [Unintelligible],
107
-------
HASSELRIIS: You have several points of moisture content here. You can
correlate them as you go across here and you can extrapolate the relation-
ship of the moisture content to the horsepower [unintelligible]. It does
show a very strange characteristic, that's what I wanted to point out.
SPENCER: Was that the hammermill?
HASSELRIIS: No, it was a ball mill.
SPENCER: So it would really be a somewhat different phenomenon, perhaps,
with the ball mill than with the hammermill?
HASSELRIIS: Yes. But there's preparation required (which in this case is
a flail mill) is what I was basing that on as the first step, not a severe
working like a hammermill or a secondary shredder.
SPENCER: Have you done tests with the hammermill or secondary shredder at
low dryness to indicate what the kilowatt-hours per ton would be there?
HASSELRIIS: It happens that those machines are very unsatisfactory for
really dry material or almost dry material. The dry paper just takes so
much volume that the machine doesn't handle any of it. That's actually why
you have to abandon that kind of machine if you try to go for dry material.
TREZEK: We tried ball milling light fraction [unintelligible]. You just
couldn't get the volume in the machine that you needed. We tried ball
milling light fraction coal, and we have gone through a whole set of dif-
ferent things, but you just can't get enough in there.
FISCUS: A look at the material loss through a hammermill. At St. Louis
over a whole year of study totaling up the weight that went into the plant
versus the weight that came out of the plant, you end up with a 6 percent
material loss that's unexplained. We assume that this is moisture loss
through the shredder. There are a couple of other people that experienced
the same order of magnitude. At the plant at Madison, Wisconsin they found
(this is not published but we talked with the people) that you could look
for a moisture loss through a shredder of somewhere of around 2 to 5 percent.
This is 2 to 5 percent of the total material weight. In the study conducted
by the Bureau of Solid Waste, in a small shredder installation they reported
something on the order of the magnitude of 4 to 7 percent. So I'm wonder-
ing, has anyone else looked at the material loss through a shredder?
DE CESARE: We've not looked at a shredder specifically, but through our
pilot plant we get as much as 10 percent loss as best we can measure, which
we've attributed to moisture loss from air classification and shredding.
HASSELRIIS: May I point out one obvious thing, and that is whatever horse-
power you put into the mill goes directly into first, heating the material
and then, drying it. Most of the time you try to correlate how much drying
108
-------
you accomplish in the mill. The better part of it is created by the horse-
power itself.
HOLLANDER: I took a photograph at St. Louis—of the shredder and the dust
purge cyclone—which shows the moisture driven off right within the shred-
der. Under the right weather conditions, you can see a vapor plume coming
off that cyclone right above the shredder. It's a very dramatic illus-
tration of the phenomenon.
ALTER: This problem of how much is driven off, Doug, you are talking about
averages. How do you measure if you shred and air classify? [Unintelli-
gible] also been interested in how much moisture is "lost" in air classi-
fication; you think it would be pretty good drying step. Really, what it
is, is that moisture is not evenly distributed in refuse. Sometimes, as
you know, when you hammermill refuse you get a pool of water on the bottom,
you free it, you liberate it. Yet if you took the raw refuse before it was
shredded, then somehow homogenized it and there was a pool of water, it
would go into the computation. I suggest that the simple statement of how
much water is lost in shredding or air classification—the simple statement
is meaningless. If you took paper, for argument's sake, or lawn waste, yard
waste, some homogeneous matter, and said, all right now, I know the moisture
content of this and then hammermilled, and then looked for the moisture
balance, then I think that you'd see again a steam plume, but raw waste, I
think it's meaningless. You are fighting the concept of averages.
FISCUS: We did measure the relative humidity of the air in the ADS exhaust,
and also in the hammermill dust collection system, which was the steam plume
that they were referring to, and also the particulate loss in those systems.
The total material balance loss at St. Louis was 7-1/2 percent, and you lose
about 1-1/2 percent, which you can account for to moisture gain in the air
exhaust from the dust collection system to ADS, the particulate loss which
you can measure there, and you are still left with about a 6 percent loss,
which means that if you have a hundred tons coming in, you're only going to
have 94 tons coming out. This has to be accounted for in the economic model
of a plant.
LINGLE: I'd like to make a suggestion here. I think we are getting into
very interesting discussion about some of the operating experiences of
shredders, but what about the issue of what sorts of research or development
or demonstration needs there are with respect to shredders? I think the
reaction to the data presented by George Trezek was very positive, that
these are the sorts of things we need to do. Should we continue this sort
of research; if so how? Where should the Federal government be going in
this area?
LAMB: I do think that other types of size reduction equipment ought to be
explored. Everybody here speaks about shredders and now it's as if it's a
fait accompli. It may well be, but obviously we have a slightly different
point of view and have had some very outstanding success with ball mills.
109
-------
Ball mills, at least conceptually, can present many advantages. I really
wonder if the group is wise in closing the issue with shredders, because I
really don't think that's the only kind of size reduction equipment that has
been known to man, nor is it the one, to me, that is obviously the far su-
perior of all the size reduction equipment. It may well be, but there isn't
enough data to suggest that. We almost become self-fulfilling prophets by
every time we put in a plant and put in size reduction equipment, they are
shredders, and therefore they are the only equipment supplied. I wonder by
what rationale that conclusion was reached. Certainly if you look at the
power consumptions of shredders, you can see that if you try to correlate
this to Bond or anybody else's theories on size reduction, if you look at
the new surface area created, there is simply no justification for the
power consumption. So when you are trying to spread fibrous materials, you
don't follow the Bond equations for a less fibrous material, which is the
point that you made. Is it worthwhile, for example, to try to make the
materials more friable? Can you size reduce them better?
BENDERSKY: When we evaluated the shredder, we didn't confine ourselves to
hammermills per se. We had classified all methods of size reduction under
the generic name of shredders or size reduction. We, at least from our re-
search vantage point, did not exclude other means or other methods of re-
ducing the size.
LAMB: I still think it's well to keep the book open on that score.
RIGO: A major unknown is the power supply that you have to install in these
plants for these shredders. We all know that current surges through the
shredders are just incredible. They've got a locked rotor maximum, they've
got idling power, but what do you have to install when you include the di-
versity of surge and load throughout the entire plant. This can be a very
critical question. You're having to install a large new major substation.
Point number two, as I mentioned earlier today, I'm afraid that if I put
wood into my primary, by the time we get to the air classifier I will lose
it all. If I'm dealing with a facility or location that has a lot of wood,
such as in industrial waste, I'm afraid I might lose an awful lot of my
heating value. Something that we are looking at doing now for a Navy pilot
plant is presegregating the large wood, the pallets and boards, and running
it through a pallet breaker, and then introducing essentially large tooth-
picks into the secondary shredder thus bypassing the initial screening,
size reduction, and air classification steps to eliminate this problem.
There's a concept you might want to consider—handling pallets separately.
ALTER: Steve, you asked about whether this type of research should continue.
I think absolutely. There's so little known, as I hope people will con-
clude listening to all of this discussion. I think there is so little known
about size reduction of refuse it's pitiful. We could argue that maybe the
friable theory shouldn't fit, but really it does. Why does it fit? I was
joking to Dave Spencer, and expect this is true, that if we took the dis-
tribution of ages and weights in this room, we'd find a Rosin-Rammler
110
-------
function it would fit. What we're doing really is describing what we al-
ready know, but we're describing it in a more concise manner. It's not
the same as understanding it. I think that if we understood size reduction
then, Tom, we could talk about all different kinds of devices, of size
reducers, maybe of types not yet conceived. We do not yet have a basis for
invention. Further, we haven't discussed at all and I'd like to just throw
it in as a point not to be buried, the idea of a sizing circuit. We just
talked about size reduction. There may be much merit in having a sizing
circuit, so as to prepare homogeneous material for the next recovery step.
Interestingly, this is done in some of the European pilot plants. If we
want to air classify, for example, we just might be a lot better off having
a homogeneously sized material, instead of a broad distribution. The sizing
circuit is one approach to it.
TREZEK: One of the very first objectives of our work was to essentially
find the grinder or the shredder for solid waste. That was one of the pie
in the sky kind of things. It took us a while to try to understand what
we were doing.
LAMB: Did you find the shredder? Don't keep it a secret.
TREZEK: I completely agree with your idea of not closing the books, because
we have ideas of making the material self-destruct itself, you know.
LAMB: That was the only thought I wanted to bring out. It was not a criti-
cism of anybody, it was simply an observation. If one were to sit in this
room and listen, the only thing we heard talked was one, shredders, and two,
hammermills. The two names generically imply the same kind of size reduc-
tion philosophy, and that is the high speed shaft with some kind of a
massive device in it. I really question if the score card is all in, and
that is really the way that all of us should be believing that this business
should go.
TREZEK: First of all, it's clear that we probably shouldn't run these ma-
chines at as high a speed as they are being run. Secondly, I think we
haven't even thought about all of the possible ways that you could, in fact,
achieve this size reduction.
LINGLE: We'll entertain more questions, but something that hasn't been
brought up is the issue of size reduction equipment in terms of scale. In
other words, do we have suitable size reduction equipment now for small
scale operations? Would anyone like to comment on that?
RIGO: Being intimately involved in trying to find something right now
smaller than 20 tons a day, from what I have been able to find there isn't
much out there that will take the abuse of the things that are in the waste.
I think that we've got to spend a lot of time or at least some effort look-
ing for small scale equipment that won't be destroyed when Fred is working
on his Volkswagen and throws out an engine block.
Ill
-------
ALTER: I disagree, Greg. I've seen pilot plants at 5 tons per hour using
a flail mill that just seemed fine. Roger has one. What it is, is that at
5 tons an hour you have a chance to tip on the floor and spread out the re-
fuse with a front end loader and pick out the engine block. There's a 5 ton
an hour pilot plant in Sweden using a flail mill through a sizing circuit—
a trommel sizing circuit—it's beautiful. They also cip on the floor and
load carefully. Now I think in your case, when you start talking about a
small plant, 20 ton a day maybe, for the Navy with all the dunnage and other
material, perhaps it's a different ball game, or maybe it's a different
sorting basis. You've already said you are going to sort and go through a
pallet breaker, maybe. I worry more about the 200 ton a day, where Arnold,
for example, winds up with a 50 ton per hour system and can take everything.
You're too big, perhaps, to do a good job of sorting on the floor, but
you're too small to take the chance that the VW chassis and engine block
together will go through the equipment.
RIGO: The position that I have been forced to take because of what I have
seen happen to some equipment, especially handling military waste, is even
with a picker, even with spreading material, something's going to get
through, and when that happens I don't care if it rips out the hammers, I am
very concerned if it warps the rotor. At that point I'm not talking a mat-
ter of repair and replacement that is local, I'm talking a major renovation.
The flail mill opening device is obvious, it will pass these hazardous
things. If the plant is designed properly to take that, that's also great.
You can accomplish basically the same thing with a trommel pre-breaker,
where you open the stuff up and you get a change to spot objectionables. I
think that we can protect things quite a bit if we start following the min-
ing industry's lead, and put magnetometers in front of the shredder and just
stop the feed conveyor when something massive gets in there. I think that's
another safety factor. My concern is not reducing the frequency of in-
jesting objectionables, but when it does occur, I want minimum damage.
LINGLE: There seems to be a lot of interest in going ahead on this, but we
are going to run out of time on the other things. Maybe we should cut this
off and move on to separation. We'll take two more comments and then we are
going to go on.
WARE: The first thing I don't think we have covered, and it may not be that
important, but it is to me, is why shred in the first place? We have two
objectives in San Diego. Our secondary shredders are sizing to meet the
requirements of our pyrolysis unit. That's a process requirement and we
can't do much about it. It is necessary for conveying the material around
the plant, for storing and reclaiming, getting rid of oversized particles,
and for the efficiency of our air classifier. That hasn't really been
covered much in this discussion, as to why we shred in the first place.
Another side benefit, which we are very happy about, is that the homogeniz-
ing of the material makes it very storable, insofar as it doesn't appear to
give off obnoxious odors, it doesn't support combustion, it doesn't attract
112
-------
rats and flies. That is why we shred, for convenience of storing, and
convenience of handling. These detailed research activities that give us
horsepower and particle size distribution may become interesting to us in
a few years when we come into optimization, but right now, quite frankly,
they don't mean anything to us as far as design goes. We got a shredder
to beat the hell out of it, to homogenize it so we can store it and move it
around, and that's frankly all we care about. The size of the machine is
predicated on the surge loading. We have to put a certain amount of material
through that plant in a day to meet our requirements, and it has to be able
to handle the worst thing we can expect; engine blocks and the like. This
detailed information only comes in years from now, in optimization, cutting
down that 1,000 horsepower machine to maybe 500 or 200, and the way it af-
fects our heating load later in the plant. That's the first comment I'd
like to make. The second thing I'd like to comment on, as far as the EPA
goes, is what does the Federal government want to be able to contribute to
this? Well, I found travelling around the country looking at other shred-
ders, there is a vast difference in performance for the same horsepower.
Let's take horsepower as a base. Bob Power's machine does better than three
times our productivity with the same horsepower. The kind of information
we need is why. Whether it's a matter of grate size, RPM, hammer size, the
nature of his feed, whether that's a significant difference. We are doing
experiments on our own, we are changing the grate size of our machine, but
I'd like to know why shredders do what they do. I don't think Dr. George's
information tells us that yet. I don't think the manufacturers could begin
to tell us. When they offer a machine to do a certain job, they expect it
to, but in a lot of cases it doesn't. That's the type of research we need—
relating energy consumption to particle size reduction and differences in
the machine, the differences between the Eidal and the Heil and the American
Pulverizer and the Gruendler. Visibly there isn't much difference between
our American Pulverizer and Bob's Gruendler but the productivity is there,
the factor of three. So that's the type of research we would like to see
and it could certainly benefit us in later designs.
HOLLANDER: The idea of size reduction is really not all that new. If you
go back to the early forties, and back to the thirties, many shredders were
installed in waste reduction plants and incinerator plants. In fact, in
Louisville in 1949 they had a shredder for oversize wood wastes and corru-
gated. Most of the early shredder installations were installed too small
and underpowered. They couldn't take the impact loading—they would
stall, and when a shredder is stalled and loaded with material, it's a lot
of fun. If the Navy is going to put in an installation, they want produc-
tivity. To put in a Mickey Mouse shredder for the sake of saving a few
dollars, one that cannot receive all of the material that is fed to it, is
poor economy and a poor resort to expediency. You have to look back at what
has happened in the past, and all the trials and tribulations that were
experienced. This whole idea is not a new one, it has come up before. When
the St. Louis plant was being built, these aspects were explored and a ro-
bust machine was installed in the hope and expectation that it would not
become the bottle neck.
113
-------
LINGLE: Why don't we move on to the next subject, which is separation. I
think tomorrow there's a discussion on technical obstacles. I'm not per-
sonally sure exactly how that differs from this discussion, so maybe we can
continue some of these comments during that session tomorrow. If you still
have questions or comments, just write them down, and we can bring them up
tomorrow. There's a lot of fertile ground in separation. We can talk about
magnetic separation, air classifiers, screening. We can get into the trom-
mel issue that was brought up this morning. Would anyone like to start off
with any salient comments on the issue of separation of wastes, and some of
the problems and research needs in that area?
ALTER: I think there is a problem common to all separators, wich is—we
don't know. Let's take the oldest and simplest—the magnetic separator. I
don't know about anybody else; I have not found quantitative data or defau.£u
guides, beyond general brochures, on magnetic separators. There's vary
little in the literature as to what you might expect if you put in a separ-
ator of a certain type, in a certain location, for a certain feedstock, and
you install it in a given angle, belt speed, and so on and so forth. There
are few data on what's the composition of the product and its cleanliness.
There's no way of making a comparison between Manufacturer A and Manufacturer
B.
DE CESARE: The research that is needed is in unit ops. If EPA wants to put
in money, it would be well invested in a lot of low cost (relatively low
cost considering some other EPA projects) as far as basic research and air
classification, nonferrous metal separation, Harvey said magnetic separa-
tion, this kind of research should go on. We need basic research and there
is absolutely none. No one knows what parameters we are working with.
LINGLE: The obvious question I would want to ask is: How would you suggest
that such research be carried out? That's really for another session, so we
won't get into that here. I got the impression from Harvey's comment that
perhaps some of the data, for example in magnetic separation, can be obtained
by simply collecting data on what's in the field because there are a lot of
magnetic separation units out there. It seems what Harvey was saying is:
It would be nice if we had good data on the performance of all of those
units.
ALTER: That's part of it. I would like to make a point of how it might be
obtained. There are a number—a few now—reports in the literature of going
into a plant and "evaluating" it, which is measuring the output of the mag-
netic separator while the evaluators were there. I suggest that's the wrong
way to do it. Rather it has to be done by first going in and measuring for
a given configuration, and then playing with the machine a bit saying, well,
it isn't working quite right, and changing, trying to (I won't say opti-
mizing) but at least making it better or even making it worse provided you
record how this is done. This can be done in existing plants, I believe,
but must be done in some orderly experimental manner.
114
-------
LINGLE: All right, I believe that's a point that was made earlier this
morning, too, that if you are going to carry out measurements, you've got
to be able to vary your parameters, and that's apparently a problem in col-
lecting data in existing facilities. They are interested in throughput and
you can't really play around with those systems very much. I suggest that
you need research facilities to do this.
HASSELRIIS: I think one generalization that everybody would probably agree
to is that the magnetic separator may take all of the magnetic material,
that may be possible, but it takes an awful lot of other material with it.
The real problem is not what that takes, but how to get the other material
out of it after the magnetic has taken it. This requires another step of
classification which musn't be overlooked. I've seen a simple practice of
winnowing work very well. You drop it into a container with a good wind and
it cleans it very well. We'd like to control that kind of performance and
blow all of the other materials out of the metals systematically. That is
something which people do neglect to install; that principle in the plant
usually has to be corrected afterwards.
WALTER: I suppose anybody who has visited Ivry in Paris has watched their
drum separator take a gross cut at magnetic materials, and put it on one
conveyor and drop the ash off into the space, and then watch that second
belt conveyor (magnetic belt conveyor), which very carefully then took the
smaller stream and separated light ferrous from the heavy ferrous. They do
a very nice job of separation, in fact it is fascinating to watch because
you can see a spoon come up the conveyor and actually levitate, sit there
6 inches above the belt trying to decide whether it's going to go with the
good stuff or the bad stuff. So it can be done and it can be very cheaply
and really quite simply.
BERG: Warren Spring Laboratory is having the same experience. Maybe some-
body who has visited it could give a little more to the conversation about
their use of the drum separator.
ALTER: No, that's not a magnetic drum at Warren Spring; it's a ballistic
drum.
BERG: It's a ballistic drum there. They do achieve basic separation,
though, using that type of system at Warren Spring.
ALTER: Yes and no. I'm sorry that the Warren Spring people weren't able to
come today or at least send the movie. If I knew they weren't coming I
would have brought slides everyone could see—it's a delightful place. They
use a large drum, about 3 or 4 feet in diameter, rotating very fast d
don't know the speed), and they drop material on it. The idea is to sepa-
rate by bouncing light material from what they call dense material. Es-
sentially it's a food waste separation from stones and rocks and small bits
of nonferrous metal. The latter all go into the froth flow circuit for
glass recovery, and the food waste falls off into a waste bin. They also
have another type of ballistic separator where they just take material off
115
-------
a high speed conveyor. They've sometimes augmented with a blast of air,
and again, it's a low density, high density kind of separation. (Warren
Spring, for those not familiar with it, is operated by the Department of
Industry of the UK government. It's kind of our equivalent to the Bureau
of Mines. Their function is in the investigation of resource recovery unit
operations, not necessarily a system.)
LINGLE: My perception, as perhaps people have been suggesting, is that air
classification at this point in time is to a large degree an art, and that
we know relatively little about what we can expect to get out for different
types of inputs. What sorts of specific problems are there there, what
kind of a program can we carry out to try to clear some of those questions
up?
LAMB: What is the function of moisture content? What are the scale-up
plans that everybody seems to feel it's the [unintelligible] how many zigs
and how many zags do you need? It seems to me that if you could define the
variables (I think fundamentally they are particle size and mositure con-
tent) , one would then be prepared to decide when he sets up his primary
shredder, does he want to shred to 12 inches, does he want to shred to
6 inches, to 2 inches, and maybe he would want to shred to one of those
particle sizes depending on what he was trying to accomplish in the next
step. Was he trying to remove lights from heavies or heavies from lights,
or fines. None of that information is available, and that is fundamental
information like was presented here a little earlier. That's the kind of
information that everybody would benefit from, and it doesn't really pre-
judice one person's position or another to make the data publicly available.
When you start to put all these things together in systems, then you are
making judgments for people that perhaps the government or not really any
public party or anybody other than the people who want to install this sys-
tem are prepared to let be made public. It's the same way with a primary
shredder. How many, what size product will be wanted, and if you look at
that graph say gee, do I want to go to 2-1/2 inch or 6 inch? What does that
mean, what is the implication of it? Do you want to trommel before or
after? Well, if you trommel before some things don't even go to the shred-
der. So maybe the shredder is half the size required if you didn't trommel.
You've solved some feeding problems. What things have you given up? It
really depends on what shape you want the glass in before you recover it.
Do you want it in big pieces or little pieces? That depends on what people
are interested in. This is fundamental data that it seems would benefit
everybody in the business. Are they going to put in a secondary trommel?
If you put in a secondary trommel, what's the performance of it? What is
it as a function of the treatment that goes before it? You can set down a
matrix of data I believe, which people would like to have and it doesn't
have to get into proprietary equipment, and it doesn't have to get into
proprietary processes, and it doesn't require the expenditure of 25 million
dollars. It's actually beneficial.
LINGLE: What you are talking about is basically a series of controlled
experiments and tests on different unit operations for a whole range of
preprocessing equipment.
116
-------
LAMB: You just have to talk to people who have historically developed and
understood [unintelligible].
LINGLE: Right.
LAMB: I don't think we have to reinvent science, and Newton's law is still
applied, by and large. We seem to be [unintelligible] to talk about it.
LINGLE: Right.
BENDERSKY: Tomorrow we are going to be going over some 40 specific recom-
mended research areas that we have identified and amongst them are two as
far as air classifiers. First of all, the optimum operating condition for
existing air classifiers; in other words, how to run them most efficiently.
LAMB: What does that mean?
BENDERSKY: Well, it means what effect does air flow have? What effect does
feed rate have? What effect do the various parameters that you can control
have on air separation?
LAMB: To me it's not so important to design how to run it "most efficiently"
because depending upon the person's needs, efficiency may be something very
different. But if you present a range of data, a matrix of experiments with-
out trying to decide what conclusion you want to come to, if you simply lay
the data on the table, then let people take the data and use that data to
optimize their own system because they have so many variables to play with
there—changes in composition, changes in feed rate, changes in product spe-
cifications. What we need is a matrix of data to be able to allow us to
forecast what we think a. change might have cascading through the system.
BENDERSKY: Agreed. Amongst our recommendations is the need for special
test facilities at which you can manipulate these conditions. However, what
we are also finding is that the present operators of plants are having to
manipulate their own equipment and determine efficiency of operation. Manu-
facturers really don't know how to operate their equipment most efficiently
under actual plant conditions and the operators are having to find this out
themselves. It would be helpful to the operators if they had some basic in-
formation as to optimal operational conditions. The manufacturer puts it
in, installs it, he tunes to what he feels is satisfactory, and generally
it's left up to the operator to really operate the thing properly after he's
gone. Another area that we are recommending is that there be some compara-
tive tests done between the various basic air classifiers under similar
operating conditions, so you can determine which actually is most efficient.
LAMB: Which most meets your needs.
ALTER: I think we first have to sit down and agree what is to be measured.
Let's say you want to recover all the aluminum cans. 1 can drop 100 per-
117
-------
cent of the waste and so recover all of the aluminum cans. Then we find
that there's a difference at the same split with different air classifiers.
For example, say at a 70/30 split, there may be a "dirty" heavy fraction or
a "clean" heavy fraction, dirty or clean according to the amount of paper
and other organic material included. The point is, how are we to make such
measurements and comparisons? The terms and methodologies should be con-
sistent. This group should meet to draft such methods, and I suggest per-
haps we can reconvene in some sort of ad hoc group to do so. We have some
procedures in draft and would like to get the input of others. If these
comparative measurements can be made on the same base, we can better
communicate. I agree with Tom; what you measure depends on why you air
classify.
SPENCER: I agree that it makes sense to do that, and you can do that. It
takes a lot of time, and it's hard to get out of committees, but you can
arrive at how you measure the performance of an air classifier. The thing
that I am very concerned about is that it can be very expensive to do this
properly at the proper scale. You cannot underestimate the amount of dollars
associated with the testing of one air classifier, let alone the amount of
dollars involved in comparing several different air classifiers. I agree
with what Tom says: When you are all done you just have to make the data
available and allow the designers to decide what's the best for their
particular application given the data, whatever it is, that you gather out
of that. You can make measurements of performance, and some do perform
better than others. But it's really expensive and you're not talking about
some small potatoes in terms of dollars here to test this thing. If you
want to test five different commercially available air classifiers, you are
well up into the millions of dollars for a testing program—several million.
Considering the cost of a single system, it's low. Considering the amount
of waste with 50 systems, it might result from wrong engineering decisions,
the overall cost again to the public is very low for such a testing program.
To do it halfheartedly—you just look at somebody's air classifier and
another guy's air classifier and get a couple of data points and try to make
a comparison—is just a waste of money and probably money better not spent.
I think you either have to decide it's something that ought to be done,
we'll gather the appropriate amount of researchers and do it properly, or
just don't waste your time.
BENDERSKY: One of the sources of information that is already available, but
unfortunately only to individual people and individual companies, is the ex-
perience they have to go through to tune their own equipment. That in-
formation is available, and what I would like to see is more of that kind of
information published if possible.
SPENCER: This idea of tuning equipment I don't understand very well, be-
cause the typical air classifier: Somebody says, well, my user, Utility X,
can take a dirty fuel product and therefore because we don't have any bottom
ash problems, I'll take 80 percent as a light, even though it's got all
kinds of junk in it. Buyer Number 2 says, I need the cleanest doggone fuel
118
-------
product that came down the pike; I've got all kinds of problems with bottom
ash, and I need a very clean product that's not going to create bottom ash
problems for me. In any event, probably most air classifiers work pretty
much the same if you just take a particle at a time and throw it into the
air classifier, and you tune it for that particular particle. Where they
start to operate very differently is when you have a broad spectrum of
particles going into it, and you have all the interference problems, and
the problems of keeping the material liberated and free. That's where some
air classifiers start to perform one heck of a lot better than others.
CAMBOURELIS: What complicates this kind of thing even further is, we really
don't start with the true requirement. In the true requirement—we're speak-
ing generally here about a production of a fuel for some place, a utility in
most cases—you really are confronted with an adversary situation. The com-
munity, the user, is really an adversary to the buyer in a sense, because
the buyer wants the best possible fuel he can get, and the seller wants to
make the cheapest, lousiest stuff he can make and get away with. We really
have to address that question as well here, and we are ignoring that a little
bit, I think.
BENDERSKY: This comes back to what we were talking about this morning and
that is specifications. What you actually have to produce, what quality of
product.
CAMBOURELIS: There is no RDF, single RDF, there's a whole slew of them, and
they'll be different in any city. Each plant will have a different kind of
RDF, depending on the efficiency of "dicker" with the utility if nothing
else.
RIGO: I think that we are looking at two separate problems in the air
classifier. Number 1 is attackable through basic research, and that is, how
does an air classifier work? The problem comes when you start looking at
multi-phase systems. You've got some fundamental laws, and you want to look
at how to get the coefficients to apply these laws to give us a handle on
what one can expect to do with an air classifier. You have the other pro-
blem, Number 2, and that is, what does a particular piece of equipment do
as opposed to the generic thing called an air classifier? The generic air
classifier's function is to split lights and heavies and to de-amalgamate or
de-agglomerate anything that's become an agglomerate, so you can effect the
separation. That's one problem. The other problem is to quantify what a
specific unit will do, how well it separates, how well it de-agglomerates.
I can see a Federal role, and I'm speaking as a consultant, in doing the
first—the fundamentals, what we can expect this thing to do. I can't be
really excited about spending Federal dollars developing essentially a sales
data base for a specific piece of hardware. What concerns me is that if you
develop performance curves, effectively, for a particular air classifier,
you are opening yourself up whereby you must do it for every air classifier,
or face suit. No, I think you are going to have to be very cautious about
doing vendor type sales data support research.
119
-------
BURCKLE: I would like to digress a moment back to this concept of tuning a
piece of equipment, and the value of that kind of data. I would assume that
in an engineering sense, when a person has selected a process and the pieces
of equipment that comprise that process, there are certain benefits in oper-
ating that process to produce the best fuel he can because of the tradeoffs
on the other end. Even if a customer is willing to accept dirty fuel, there
are problems that arise in handling that dirty fuel in terms of Number 1,
getting it to the boiler by pneumatic transportation; you've seen that in
St. Louis. That's a very expensive proposition. Although he can accept it
in his boiler, he has maintenance costs in the pneumatic transport area that
must be traded off against operating the air classifier at a more optimum
point. So when all is said and done, I don't think there's that wide a
rangeability that's going to be available after the process is selected. I
think you've restricted your point or restricted yourself when you select
the process to a particular type of a product that you are going to make;
I'm just talking about the fuel fraction. The problem being is that you are
going to want to produce the best fuel that your process is going to allow
you to produce at a reasonably economic rate. I think that when it boils
down to a specific piece of equipment like the air classifier, you're not
going to have that wide of rangeability; you're going to be looking for a
tune point.
FAY: I am getting the impression that we are looking for more data so that
we can optimize the design of the equipment. I'm concerned that we don't go
in the direction that a lot of us utilities went, say 10 years ago, where we
designed our boilers to burn a particular kind of fuel, which, after a few
regulations came down the pike, is no longer available to us. I think that
if we get more data it ought to be used to design whatever piece of equip-
ment we are talking about such that we know how to modify it in the future
or to design in adjustments, so you can change it as the waste stream
changes. I don't know what the waste stream is going to look like in 1985,
and I don't think anyone does, but if we try to fine tune these things for
today, you're not going to have everybody in trouble 10 years from now. I
think you've got to build in some flexibility with this data you want to
get.
SPENCER: I'm not sure that it's a matter of so much of fine tuning as much
as it is trying to get some kind of a handle on information that's just not
available, that's necessary for design. If you have all of the data avail-
able you can say well, this air classifier performs in this way over this
range of input; this one only works well on this very narrow range of input.
Then somebody can make a decision and say well, do I want to pay more and
buy flexibility or do I want to pay less and buy something for my specific
need? The problem is that the information isn't even available so that you
can consciously make that decision today, or in any knowledgeable fashion in
any event. We try to apply principles in some sort of gut feeling and what
seems to intuitively make sense, but the real hard information that would
allow us to support that decision doesn't really exist.
120
-------
RYDER: This sounds like a good area for R&D (I'm not an R&D man), but
merely testing existing equipment is going to tell you only what existing
equipment is capable of doing, not how to improve it.
HASSELRIIS: If you think of all the things we are trying to do in this de-
vice, then you would realize you can't have one device that does all these
things we are trying to do—not one simple device. Actually there're a num-
ber of different steps that these classifiers have to accomplish and it in
itself is a series of unit operations, if it becomes a successful device
with more than one function in it, maybe five or six. Then it's the number
of steps that are inside of it that will be determined how well it does
each of them. Each of them might need separate tuning, so this is what the
device maybe really requires. The question is, how much of a compromise on
that can you really tolerate. You can always make these things to pick up
everything, and you can make them drop everything, you can make them os-
cillate between those limits, but you can't get a fine distinction between
different types of things without using the principle that makes the se-
lection. So you have aerodynamics, you have trajectories, you have accel-
erations of particles, you have the problem of making heavies go over here
and the lights go over here and not have to cross through each other. It's
a very complicated device and I think we mostly tended to grossly over-
simplify what we are trying to accomplish in it. That's why the fundamental
research on the several unit operations that are inside the device are pro-
bably necessary for somebody either to build it in the first place, to con-
trol it when you are running it, and get the results we want. We probably
have to consider these devices have to be tuned as they are operated to some
extent. As the shift of materials takes place there's an adjustment that
is actually an active adjustment, not fixed, but has to be tuned. I think
we have to anticipate that it's really a more complex device.
LINGLE: Let me ask this question, then we'll move on to the next area.
Again in terms of scale, is there an issue in terms of classification equip-
ment for smaller scale applications and also in terms of types of wastes.
We've been talking mostly about municipal waste, but if you're going to ap-
ply these to possibly other types of inputs, do we know anything at all
about that?
HATHAWAY: We've had a lot of experience working with feasibility of re-
covering energy from waste at different military installations, principally
Army and Navy. Our ballgame is in the area of 30 to 60 tons a day. If we
are real lucky, it might be 100, sometimes in the more regional systems.
We're definitely talking a scale that is very far below the principal orien-
tation of this conference. Consequently we are very interested in anything
that can be resolved regarding scaling factors, not just of individual
pieces of equipment, but of systems. Keeping in mind that our ballgame
would also be marginally economic, and therefore we have an additional con-
sideration to the problem.
LINGLE: I think that in response to one of the points he made, the scope of
this discussion should include all sizes. We shouldn't only be talking
about large systems, so please comment on any size ranges.
121
-------
RIGO: Let me support what Steve has said, but add an observation. One
thing that makes our life a little bit easier is a "tinker toy" game—that's
what lots of people have called pilot plants—and prototype developmental
units tend to be sitting in the size range we want. Now the problem is,
that these things have also been baling wired together, and as was observed
by the gentlemen from Occidental, we really don't know why they work, and as
a result, if we go to duplicate it, I really don't like the thought of
moving all that baling wire, but I don't know that I have much choice. I
think there is need to develop small systems, and I think Steve will confirm
this. The entire processing approach to small systems has got to be radi-
cally different. We cannot be capital intensive, because you just can't pay
for it. We have to do a lot of tradeoff on the efficiency of separation.
We also have to recognize, at least when we are tangling with the military,
we're dealing with small-scale spreader stoker users, or at least stoker
users if they are going to have coal. We are dealing with people that have
steam demands, base loaded steam demands. We are dealing with an entirely
different set of problems that you deal with in the civilian sector. In the
small-scale civilian sector a lot of what is going to come out of the Navy
work and the rest of the DoD work may hopefully spill over. I don't think
you can hold out a lot of hope for the DoD work making a significant con-
tribution to the rest of the country's problems. It's going to be an in-
dicator that their waste is different, because their industrial to family
housing mix is radically different than what you deal with in a normal
community, and this is going to change how those unit operations work.
LINGLE: I would like to suggest that we move on. The last area of prepro-
cessing equipment that we have not yet discussed relates to the subject of
process control. It can involve dust control systems, shredder explosion
control, fire control. I think this is an area where we know we've experi-
enced some problems at some of the early plants, shredder explosions are
occurring, dust has been a problem in some systems. Would anyone like to
make some comments on some of the problem areas here and the directions we
should be going in trying to solve them?
HASSELRIIS: I think that many plants started out with open shredders and
find that the shredders are very powerful fans and blow the dust all over
the place. We found this out at our early plants of that sort, but we have
a fan, an air classifier, associated with the mill. The result is that the
air classifier associated with the mill was a vacuum cleaning device and
then the material was handled by air classifying by an air system, a ducted
air system, into the bin. The result is that we had a. clean plant by having
that type of system. In other words, it is basically totally enclosed.
There are many drawbacks to that, and we ourselves got away from it and got
back into belt conveyors and back into the dirt. However, right where the
shredding takes place it is apparently necessary to remove enough air and
pull it through the shredder in most cases so that there is an influx and
not an exflux, and the dust doesn't go all over the building and create
health hazards and explosion hazards and dust collecting on all of the beams
so you'll have a flash fire if it ever falls down. Basically, it appears
that having a fan system will take care of keeping it sanitary. Now the fan
122
-------
system develops another problem which is that now you have a closed duct
and you'll have a possibility of a different type of fire hazard. Now the
question comes, how do you handle that? Partly it can be handled by recirc-
ulation, which keeps the water vapor level high and reduces the chance of
any dry materials being present in that system. I've observed in general
that that type of recirculation of the moist environment is quite safe.
However, the explosion hazard of somebody's can of gasoline is greatly in-
creased because you've confined your system. What I'm saying is that the
way to have a sanitary system or a clean system is to have it under suction
with a bag filter at the end of the system somewhere, and then you've cre-
ated a closed duct system which creates a need for an explosion proof or
safe system. You can use a Fenwal type system to extinguish a fast pres-
sure buildup, provided it is a ducted system. In other words, these ducted
systems do lend themselves to a Fenwal safety approach, because they are
closed and can be closed. I just bring this up as one of the directions
that we may all find that we're heading in. We can't leave these plants
open with all the processing exposed. We actually have to withdraw a large
amount of air in order to pull the dust into the system and keep it out of
the working spaces.
LINGLE: Is the dust control area one where basically we think we know
enough to control it in terms of installing bag houses or whatever, or is
it an area where we really do need to do some more R&D type work? Are the
control techniques essentially at hand and just need to be applied, or is
there still a lot of uncertainty about dust control or how to control dust?
LAMB: I think the equipment's available to be able to control the dust.
I think it would be well to have a lot more information on the properties
of the dust. For example,what is the safe range in which you can work with
the dust, what's the lower explosive limit, what's the higher explosive
limit? How do you want to design the duct system, do you want to specify
that they be twice their maximum rate of pressurizing in the event there
is an ignition. When you design a mill, do you want to design it to vent,
do you want to design it to suppress? If you are going to suppress, well,
let's suppress all materials. For example, gasoline, which requires oxygen,
that's one thing, if you have dynamite or something which has a contained
oxidant, that's another problem in terms of explosion suppression. It seems
to me that another place that's noncontroversial and where people are pre-
pared to share data, is in the area of safety. How should we fight fires?
Should we use sprinklers, should we use hoses? Should the water contain
wetting agents? Where do you sprinkle—inside equipment, outside equipment?
Do you sprinkle the building? Do you design the building to be washed down?
Or should you use vacuum systems?
LINGLE: You say that you think that information is available.
LAMB: I don't think that information is available. At least I don't think
standards have been set for the industry which would say we consider these
kinds of room Class 1, Group B, or we consider these kinds of rooms Class 2,
Group 2, Division G, or something.
123
-------
HATHAWAY: We have done some initial testing on properties of dust, and we
are currently undergoing some chemical analysis of trace elements, toxicity.
We found that the material is both highly fibrous and highly volatile. As a
matter of fact, its volatilization rate is greater than the RDF coming out
of the process. Its fibrosity presents interesting capturing problems. For
example, if [unintelligible] adheres, you get a static field, and it's very
difficult to get it off of conventional fabric-filled medium, short of
knocking it with a baseball bat. Mechanical separators would be ideal, to
capture the material that we don't know we can. Because of the volatility
we would like to reinject it into the fuel production process. Again, these
are ideas yet to be worked out in a pilot plant.
LINGLE: . What about the area of the health implications of simply breathing
the dust in plants? This is something that EPA has gotten into. The ques-
tion is: How much research should we really do on the health problems
associated with dust generated in solid waste processing plants? I don't
know how familiar you are with some of the testing that's been done, but
tests were run in St. Louis. Additional tests have recently been completed
there. They found certain levels of bacteria and virus in there, but not
necessarily levels which are higher than occur in other kinds of industrial
processing plants or maybe sewage treatment plants and the like. Whatever
absolute levels may be found, we still don't really know what the health
implications are. That's the kind of work that could really go on for years
and could eat up a lot of research dollars. Is this an area that you feel
is important for additional work? Maybe MRI should comment since they are
involved in some of this work now.
CHANTLAND: This general area, I believe from the public standpoint, is a
highly sensitive area—extremely sensitive—so sensitive that I think it
can mean the difference between your operating the facility or not operating
the facility. The problem that I've seen in this whole area of reasearch
is nobody is willing to put up the money and go all the way. I think that
without going all the way, it's a waste of money and it's only going to
create problems in the minds of the public and are going to really get on
your back. You've got to answer the question of what is the effect upon the
human beings, the public, and the people in whatever area is exposed; what
is the true effect? This is more than just gathering up how many viruses
or bugs or bacteria or something you've got in the area, or what's its con-
centration and so forth. You know that human beings have certain abilities
to ward off diseases, and these diseases are amongst us all the time. The
type of things that we are getting into our plant are the same things that
come from the housewife, the grocery store operator, and others in society—
they put it there, and we end up with it. Supposedly the truly hazardous
waste out of hospitals and others are supposed to end up in incinerations
or destroyed by other means; they are not supposed to come to this plant. I
think they are reasonably controlled. My only caution is that I think it's
dynamite unless the real answers are given down to the point of its effect
on human beings.
124
-------
TREZEK: We had a research project which lasted a year at our lab, which we
conducted with the School of Public Health, in which we had everybody instru-
mented and we measured what they were breathing and things like that. I
think we've only actually begun to scratch the surface of this problem. As
far as seeing things like hospital wastes, we've seen packer trucks roll in,
and the first thing that comes out are a bunch of needles and blood bags and
all this because the last stop was at a hospital. So I agree, I think we
need more.
LAMB: You know the trouble is, those wastes are not pathogenic, and there-
fore they can legitimately deliver them to our facilities. We see them
routinely—hypodermic needles, plasma bags. We don't receive any amputated
limbs or anything like that, but short of that, we get everything.
LINGLE: Our perception at EPA has been that this is a very sensitive area.
Since the initial tests were done at St. Louis and reported, we've gotten
an awful lot of questions and comments on this area. I'd just like to know
if there is pretty much a general feeling that this is a highly sensitive
area to the public and may be very important to the future of any type of
plant where shredding, for example, is employed and therefore is an import-
ant area for additional work by EPA in terms of research. Any yes or no
comments?
HASSELRIIS: It seems that the big difficulty in this area is as soon as you
start to investigate it, that creates the problem. If I ask you if you get
sick from smoke, all of a sudden you'll start coughing, right? The sugges-
tion brings it up. So it takes a very subtle investigation to find out
whether people are subject to contagious diseases and so forth. You find
your employees will suddenly realize that they have a cold and they are
entitled to the full sick leave. So, it's extremely sensitive, it's got to
be measured in a very careful way if we're talking about pathogens in the
air. At the same time, we also have to measure a lot of background situ-
ations which have been going on all the time and nobody got sick. They were
making out very well under these very bad circumstances, and make a compari-
son between the backgrounds. Those situations are also typical. As soon as
you start to make the measurements, the people will suddenly be reporting
all sorts of minor illnesses, and that's how they caught the cold. So
that's why it's so delicate, I guess, that the measurement creates the dis-
turbance, which now you can't turn off.
ALTER: We are dealing with people whom we are exposing to unknowns. At our
pilot plant, to our knowledge, nobody has had a job-related illness, but we
are still concerned that they don't get them. We have a program on-going
with a small group of mostly professionals. I'd like to share this program
with you and invite anybody to join in. We have also had contact with some
of the overseas laboratories, and I think the only way we are going to ad-
vance the knowledge of worker health is to pool our data. I doubt there
will be any plant with enough employees to do meaningful epidemiological
125
-------
investigations. We are asking our employees to take an annual physical
exam, at company expense. Also, they are instructed that if they have
illness which they think is job-related, such as diarrhea, to go to their
physician or the company physician. A form is filled out reporting the
illness and its details. Names, of course, are kept confidential to guard
the individual's rights, but the data can be used anonymously. If other
people followed a similar protocol, we could pool health data at some future
time, should it be necessary. There is also a new activity in ASTM Committee
38, the Resource Recovery Committee, within a subcommittee on health and
safety in resource recovery plants. This subcommittee is working toward
standards, definitions and standard methods of measurement, and will be act-
ing as a clearing house.
BERG: I have more to say before the last few comments, which I appreciate
for their candor and depth. Now maybe I can encourage MRI to comment on
this some more. To put it in the most potentially negative light, we are
talking about hundreds of communities around the country that may very well
be doing large scale resource recovery within the city limits, near popu-
lation concentrations. If the government gets involved in a situation or
starts advocating a program that's going to subject people to problems that
they haven't had before (not saying they are or that they will) but if that
happens, you know I would feel very badly in retrospect. I would hope that
we can all follow up this with some sort of data sharing at the very least
effort. MRI is now doing work for us on that subject. ERDA is also doing
some work on it at Ames, and as others have pointed out, there's other work
going on. I at least would advocate very strongly that we talk about this
a little bit more and see what we can do.
WALTER: I'm taking off my ERDA hat and putting on my Arnold Chantland, di-
rector of public works hat, and I'm saying to you, if you are going to look
at these effects in a given processing plant, you had better be very prepared
to extend them to the garbage collector. He right now thinks that he is
exposed to a higher risk of illness because of his job than other people.
So I'm just cautioning you, if you are going to do this don't stop, go to
the trash truck. Better start looking at this, because he's going to be
back down upon us.
HOLLANDER: What about back to the trash can? When I open or dump that
trash can, what am I exposing myself to?
LINGLE: Did somebody once say you are what you throw away or something like
that?
BENDERSKY: The City of New York did make a study of their garbage collec-
tors' health, and some of their other public service people, and those re-
sults have been published. They are not very definitive. It does not
generally show that a garbage collector is in a significantly greater res-
piratory disease category than other street service people.
126
-------
WALTER: You're speaking of illness. Let me point that they are a higher
risk for injuries.
BENDERSKY: Yes, yes, indeed. I'm talking of respiratory diseases.
CHANTLAND: I just want to say that a little information is extremely danger-
ous. As a recent example, some of the work that is being done on the exami-
nation of the emissions from our boiler systems. In the reporting data it
was mentioned that among the many, many things that were being examined and
so forth was mercury. At least one newspaper in this country reported al-
ready that we are finding substantial amounts of mercury as discharges from
the emissions from the Ames solid waste plant. So that's what I mean by a
little information is very dangerous, but I also appreciate very much what
Harvey Alter is suggesting here, and we'd be most happy to cooperate.
LAMB: At the plant that we have we found that we were plagued with a lot of
puncture injuries. Full steel-soled safety shoes are very effective against
those punctures.
LINGLE: Thank you. Well, we've got about 15 minutes left, and maybe we can
end the day with a big bang by talking about shredder explosions. Many of
you are familiar with the Factory Mutual work that's been done on shredder
explosions. I think they've learned a lot. I think a question which I have
is, where do we go from here in terms of work on controlling shredder ex-
plosions or at least minimizing damage from explosions? Is the work that's
been done sufficient, do you feel that you can design a system with at least
reasonable control or protection or suppression of explosions, or is there
a need to do significantly more work in this area?
ALTER: I think that it would do well to have a design manual of some sort
on explosions, and I'm surprised that it hasn't come out already, consider-
ing that so much of this is known in the chemical process industry. High
pressure reactors in buildings are designed with blow-out discs, venting,
and so forth. Some shredders, as we know, have explosion doors, or other
pressure relief devices. I would hope that there is some thought given to
the response time versus a detonation. I'm also appalled that so many
shredders don't have such explosion protection. I would advocate that some-
body fund the proper design manual for shredders to avoid fire and explo-
sions. I would include in there sprinkler systems as well.
BENDERSKY: I can tell you from having talked to a number of shredder manu-
facturers that they are not in a position to provide this kind of in-
formation.
POWERS: Who's had an explosion in their shredder? In the room?
WISELY: I would suspect that every shredder, almost every shredder in the
country, has had an explosion of some kind or other.
POWERS: I'd like to know what the damage is.
127
-------
WISELY: To the shredder itself?
POWERS: Damage to the building, to the shredder.
WISELY: Have you read the Factory Mutual work?
POWERS: It doesn't do much for me.
WISELY: Well, that's probably the most up-to-date information relative to
at least recorded incidents.
RYDER: In Chicago, we have a 25 ton per hour bulk refuse shredder. This
machine handles only bulky trash such as chairs, tables, refrigerators, bed
springs, automobile tires, and so on. This machine is equipped with a Fen-
wal explosion suppression system and a sprinkler system to hold down the
dust. So for the first 5 years, we never had an explosion. We attribute
that purely to the sprinkler system holding down the dust as the Fenwal
system was found to be inoperative. One day a portion of an acetylene tank
partially filled with gas got in there, and they didn't catch it. It went
through and blew up inside of the machine. In answer to your question, it
blew out the relief vents in the roof and pushed out siding on the wall of
the building. The relief vents, which were supposed to open if and when
there was a pressure increase inside the building, were inoperative as over
the years somebody went up there because they were leaking and fastened them
down. There was not considerable damage to the building—nothing inside the
shredder and only the conveyor at the bottom of the shredder required work.
WARE: We haven't operated very long, but we've had one explosion. I think
the whole question here is a matter of magnitude, how you define an ex-
plosion, how big an explosion, or small an explosion. We fully expected
them. We expect more, many more. What have we done about it? We've de-
signed the feed hopper so it goes straight through the roof, with a membrane
on the top, mostly to keep the rain out, and to maintain our vacuum system
to some extent. Explosions we can handle. We expect them to blow straight
through the roof. The last time it happened it did just exactly that.
Detonations—I don't think we or anyone else can handle. Fenwal won't
handle detonations. If it's big enough, neither will the wall of the
shredder. So we isolate the whole installation, we have nobody within about
a hundred feet of it, and we say that if it goes, it goes. If it's a mortar
shell, it's going to blow the shredder to pieces. If we don't catch it in
the initial separation, there's nothing we can do except make sure it
doesn't hurt people. So we have not used Fenwal systems in our designs,
and probably never will. The explosions we can let go harmlessly; the
detonations we can't handle anyway.
LINGLE: Is it a matter of applying, in terms of the design manual, some of
the knowledge that we have now about control of shredder explosions, or do
we need to try to find some new techniques, second generation Fenwal, or
whatever, to control explosions?
[Unintelligible].
128
-------
LINGLE: So you are saying you don't think that's really a good opportunity?
SPENCER: You don't have a vertical hammermill, do you?
BENDERSKY: One of the findings of the FMRC report is that there are
techniques available for fire and explosion developed for other industries
that have not been applied in this industry. Where they have been applied,
they have not been applied according to other standards. That would seem
to be the first area that should be investigated and applied.
CHANTLAND: I've heard of some use of steam as a fire fighting device in the
area. I wonder if anybody has any comments, or if there's anything happen-
ing in this area. I suggest that it is something to look at.
WARE: Steam is a well established snuffing device in the chemical industry.
It's largely a matter of having it available in your plant. We simply don't
have it anyway. Water in that respect works quite adequately if there is a
fire. The generation of steam from the water more than adequately puts out
the fire, not the water itself. I would not have thought it was a design
criterion which you could incorporate unless you independently generated
steam in the plant, and had it readily available in large quantities.
LINGLE: Thank you for those comments. I was going to end on the explosion
issue but it was pointed out to me that one important subject has not been
discussed, so maybe we can take 5 or 10 minutes and squeeze in a discussion
of densification. Densification is something that is now gaining increased
attention. The issue of industrial boilers as a market is one of the rea-
sons for that. Does anyone have any particular comments on what we know
and don't know about densification? What we should be doing? I know at
least one person has some comments. Harvey, can you respond on this?
ALTER: Maybe tomorrow afternoon, if we have the time. I have some slides
of the pilot plant and the densifier and I can make some comments.
LINGLE: Alright, we'll reserve that until tomorrow. At this point I should
try to summarize, but that isn't going to be easy. Clearly the feeling is
that in terms of unit processes, there's a huge data gap. There's a lot of
data that needs to be collected, and we need to figure out a way to do it.
This seems to apply to almost all types of equipment, but I must say that I
was personally a little bit surprised at the first discussion that we had,
which was on storage and handling, because I didn't get the feeling that
there was a consensus that this was a major problem area, and frankly, I
thought otherwise. If there is a feeling that's a major problem area, then
maybe tomorrow there should be more discussion of that. I certainly didn't
get the impression that there was a need for a big Federal effort in the
area of storing and handling waste. It seems more problematical than any-
thing else.
129
-------
BERG: Just one other thing, at least at points during the discussion today,
it seemed that there were more comments that people wanted to make than
there was time available. If people would like to add addenda to what
they've said, or introduce additional comments when you get the draft trans-
script to go over, then please feel free to do that. If you want to write
them up tonight, then those could be handed in before you leave. Thank you.
SESSION 3: TECHNICAL OBSTACLES
Dr. Harvey Alter, Moderator
BURCKLE: The subject of this morning's first session is technical ob-
stacles, which is another way of saying research needs, I think. The ob-
stacles that we are interested in are those obstacles or barriers to the
successful implementation of the resource recovery systems, as seen from
the perspectives of the public sector, from the A&E's viewpoint, from the
consultant's standpoint, in terms of helping us, the Federal government, to
better understand how you perceive things and how to direct our resources
in R&D to aid overcoming these barriers. As part of the study on prepro-
cessing equipment, waste-to-energy systems, that Midwest is conducting for
us, we have compiled a list of research needs based on their work. Now
these research needs obviously (there are some 40 of them, quite a list)
are based simply on looking at problems that exist, without assigning any
priorities to them, without saying these are the problems that are important
to the A&Es for such and such reasons, or these are the problems that are
important to the people in the public sector, or the people in the utility
business, so we don't have that flavor here. Now obviously we are going to
come out of this with much more, hopefully, than information just for this
particular study. This is just part, a very small part, of what we are
trying to accomplish with the meeting. What we are trying to accomplish
with the meeting, as we stated in the second letter that was sent to all
the members, is a broad based understanding of these problems to enable us
to better handle our overall R&D program. I'm hoping that this workshop will
be the first of several in the preprocessing area. I hope that we will be
able to continue these over the next few years. This depends a lot on how
you, the members of the workshop, perceive the usefulness of this sort of
thing. It also depends on how well we interpret what you have said and act
on it in our R&D planning. So with that in mind, I would like to turn the
meeting over to Dr. Alter, our session chairman for Session 3, and to the
workshop members to discuss what you feel are the major technical obstacles
to the introduction and use of resource recovery systems on a nation-wide
basis. Thank you.
ALTER: Before we start I would like to add one or two things to what John
said, giving some thought as to how this session might proceed. I view the
130
-------
session as discussing how to proceed in the implementation of resource
recovery in the face of technical uncertainty. I think that's what we
really mean by obstacles. What do the implementers need to proceed.
During the discussion I'll try to exercise the chair to make the distinction
between what would be nice to have and what we may really need to have.
It's always nice to have more and more information but we can't afford this
either in time or money or people. I think this discussion might go in
three parts. One, is what I call judgment and risk assessment, and we have
to involve, as John said, the different actors here: the city who will be
the customer, the A&E, and the systems designer. The second is the de-
velopment of what I prefer to call the research agenda. We have MRI's
research needs, and I think we should avoid the temptation to sit here and
nit pick, whereas they have spent a lot of time and effort and careful
thought into preparing this document. We might use it as our guide for
discussion, or might not, but we should come up with some sort of research
agenda. We can talk more about that when we get to it. The third part,
which I think is essential, is who might do the research. We might talk
about methods and format and inter-comparisons, and how we as a group might
proceed, or how as individuals we might proceed, and keep the doors of
communication open so that the initiative of this meeting is not lost. To
return to my list, judgment and risk assessment, how to proceed in the face
of technical uncertainty, what sorts of information do we need to proceed?
I'd like to, if I may, have the discussion opened by first perhaps hearing
from the municipal representatives present. What sorts of information would
they like to have to diminish uncertainty? Inasmuch as there are no volun-
teers to start, how about Dick Bush. You've had some experience in this.
BUSH: We were concerned, for one thing, about the scale-up problems from
the pilot or bench model up to a research and development program, and then
from a research and development setup to a full working program. We all
discussed this yesterday as one thing. Then there are other aspects that
we are concerned with, such as the location of equipment as well as defining
the operation of each individual piece of equipment, such as the magnetic
separators. Should they go before or after an air classifier? That would
change its function, for instance. We would like to compare, for instance,
mini systems versus larger systems. We also touched yesterday on the vari-
ations of as-received material. We also touched upon quantities. Quanti-
ties that's generated in each municipality is really a problem, because when
we tried to get a commitment from towns, as was brought up yesterday, even
though they expect to give you say, 600 tons a day, they'll only commit to
say, 400. So this is a real problem in determining how much they'll gene-
rate. Of course, I'm most interested in trommel operation, whether that
should go before or after a shredder. These are some of the things that
we are concerned with.
ALTER: Are there other comments from the cities?
LALKA: I think the things that we are looking for are basically three
things. Number 1, reliability. If one were to implement the plant, there-
131
-------
fore you've made a commitment to that facility, and presumably you should
take all the refuse that comes in there and not turn any trucks away. Now
if a plant is inoperative, then you have a problem. You have to direct the
flow to a landfill or somewhere else, and this presents a tremendous problem.
We don't really have good numbers on reliability due to the fact that most
of the plants are small-scale, and when you are talking a 1,000 ton per day
plant, we really don't know anything about how frequently it would be down
or what kind of contingencies must be applied. So that's one major concern.
The second concern would be the environmental acceptability of a system. We
have very little good air emissions data on most of these energy recovery
systems. Some Federal tests have been done on these facilities, but as far
as Los Angeles County goes, the Federal testing methodology does not follow
the procedures used by our air pollution people; therefore, they find it
very difficult to interpret the data. Now manufacturers have made tremen-
dous claims. I've heard things said like, there are no emissions from a
plant, which is completely absurd. Somewhere in between there must be some
level which we can monitor the system on an even basis, so at least Los
Angeles County will have faith that what's happening in Baltimore, for
example, could perhaps be investigated with regard to our air pollution
codes. At this time we just have no faith in the numbers that have been
presented. As I said, there are just so many different people making dif-
ferent claims we just don't know where to look. The third concern that we
would have would be the economics. We have not touched on economics at this
particular meeting here. Again, manufacturers are making all kinds of
claims. We've seen numbers all over the place from 2 cents a ton to 20 dol-
lars a ton. Different consultants look at the same process, and come up
with different numbers. We really don't know what it costs to run an air
classifier or any of the individual components of the system. We are
pretty much at the mercy of the manufacturers. They come in and say an air
classifier costs $250,000, and we have to accept a figure like that. We
don't feel too comfortable accepting what manufacturers tell us, because in
the past we've had some bad experiences along those lines. So I think those
would be our three major points of concern.
.ALTER: Unless this group feels differently, I think economics is not an
agenda item. Maybe it's a subject for a different day. You mentioned re-
liability, operating reliability, and certainly there is a paucity of oper-
ating data. How do we get it, how do we fill the gap in the interim? Is
it possible to find out anything about reliability in the absence of many
operating years experience in the plants?
RIGO: Even with many years of operating experience in the plants, unless
the accounting systems were established and used to allow you to pull out
costs and reliability information, you aren't going to get anything anyway.
We found that out in taking apart Franklin. We just don't have a good handle
on historical costs. I suspect that the Army study to assess the relia-
bility of gas turbine generator sets from a 20-year historical record set
down in the Kwajalein Islands for the Army is representative, and that was
hopeless. The only way we could reduce the data was literally put it on
microfiche, and that way we could say we reduced it by a factor of 20 to 1.
So we are going to have to put some thought into how we want to maintain our
records.
132
-------
ALTER: Would it be useful to develop for general dissemination, uniform
accounting systems in resource recovery plants? Do you think it is possible
to develop a uniform accounting system?
BENDERSKY: We have looked into the present techniques that people are using
in their accounting systems at these plants, and I can tell you that they
vary all over the map. It is very difficult because of the variation in
the existing accounting practices to make a comparative economic analysis.
You really don't know how to compare, certainly not on an individual unit
equipment basis. In other words, you can't ascertain from most of the
existing records what it is costing them to actually operate any one par-
ticular piece of equipment within their line. People are assigned in the
plant and they roam around from one piece of equipment to another, so you
can't get manpower, you can't get basic information. One of our recom-
mendations is that if comparative analysis is to be done, and indeed as
it's pointed out here, if you are going to ascertain reliability versus
costs, then there's going to have to be some standardization. It is going
to have to be simple, something that's easily done in a plant, because you
are going to upset normal accounting procedures, which is a difficult thing
to accomplish. It's one of our recommendations that, if possible, some
attempt be made at standardization of accounting systems.
ALTER: Let me tell you that the Center is developing an accounting manual
for Recovery 1. If any of you are interested in having it when it's de-
veloped, and more important in reviewing it now, please let me know.
LINGLE: Harvey, you may be familiar, or may not be, we have developed an
accounting guide for resource recovery systems, and I think it's at least a
good first step toward what you are talking about, it may be that you are
talking about something in a little bit more detail. I think the basic
concept of being able to evaluate the cost of systems on an equipment
basis, we do have a start in an accounting guide which we've developed. I
think we'd probably be interested in expanding it, and making it more
detailed.
ALTER: Other city representatives? Arnold, you had to proceed in the face
of uncertainty.
CHANTLAND: I'm sorry that economics isn't a larger item on the agenda, be-
cause I would say that this is uppermost in the minds of the communities.
Probably the smaller the community, the greater impact it has. I would like
to relate to the operation of the equipment that we are talking about today.
We're concerned about continuous operation, that's a part of economics. Any-
time you are down, you are going to be losing revenue in all probability,
and it's costing you a substantial amount of money just to sit there. When
something does absolutely nothing, it's still being depreciated; those are
real costs. In line with that, I would hope that everybody would be think-
ing of duplication. We weren't so fortunate to have duplicate lines, but
133
-------
I certainly hope that others in the business will have duplicate lines,
so that they can be working on one line while the other one continues to
operate, even if you have to extend your operating schedule. Maintenance
costs. There are certain wearing factors in almost every moving part,
elements of almost any piece of equipment. A good look at these would be
well to reduce the costs and economical impacts. Probably all of you are
aware the shredder is the so-called self-destruct piece of equipment in
this operation. It is certainly the larger in terms of maintenance cost
and cost to operate. You literally wear this thing out from one end to the
other; whether it's the breaker plates, the grate bars, the hammers them-
selves. All of these things are intended to be worn out at some period in
time. The key, of course, is to get a longer life in this operation. I
guess I'd have to say I'm a little surprised that the manufacturers don't
have a longer track record. Hopefully there's becoming more of a record.
We see some encouraging signs along this line. I hope that everyone will
continue to explore new ideas or something that will reduce these costs.
ALTER: You raise an interesting point of duplication or replicate lines
or redundant lines. Redundancy is expensive and ought to be, I think, cal-
culated off against the alternative of diverting to landfill. Depending on
costs of land, it might be cheaper to be down than redundant. Has anyone
gone through this kind of tradeoff analysis and deciding whether or not in
a site specific case, it pays or not? Mr. Ryder, did Chicago do this, or
did you just put in duplicate lines?
RYDER: We put in duplicate lines to assure that we would meet our commit-
ment to the utility. Our commitment is 3,500 tons a week and we can do that
with one line. Of course should one line be down for two or three days we
can shift over to the second line in order to meet our commitment.
ALTER: If the line were down and you didn't meet the commitment, is it
really more expensive than having installed and operated a duplicate line?
I'm really asking, was that kind of analysis done?
RYDER: No. We felt that there wouldn't be much need in building a plant
that couldn't meet its commitment.
WISELY: We have done analyses similar to what you are talking about, Har-
vey, and the results will vary depending upon the local circumstances.
There are some cases for relatively small facilities where there is no way
you can justify a replication of the basic line. I think that one thing
that many people forget is that there are a number of areas of required
redundancy in any system like this, not the least of which is the place to
which you are going to deliver whatever you are making. In many cases, this
may be a more important element of redundancy than the actual processing
facilities. For large facilities, I don't think there is too much of a
question, if there is enough waste available to warrant it, that duplicate
lines are quite desirable. The one thing that I think that always bugs me
a little bit is, though, that people are continually talking about tons
134
-------
per day; they should be talking about tons per week, at the very least.
The rationale behind that, of course, is the daily input of waste in any
normal municipal corporation is going to be something like 25 percent of
the total weekly total, say on a peak day, and it may be as little as 10
percent of the weekly total on another day. It all depends on their pro-
cedure. Any facility of this type, I think, has got to take swings of at
least that much.
ALTER: Has anyone else given thought to the probability of both lines being
down, and the subsequent contingency plan. In other words, do you use
storage space, or purchase enough trucks available to divert?
WISELY: There is no extant process for processing waste that cannot get by
without some form of alternate disposal. I think it's going to be a long
time before that condition can be reached.
ALTER: Are we leading now to a research need which is a probability analy-
sis or to obtain the data in order to determine the probability, or relia-
bility of a plant.
LEVY: A couple of points, one in response to what Ed said. I don't know
what the current status is, but I know Franklin, Ohio, after it went on-line
in April '74, up until at least recently, had never had to divert any waste.
They had designed into the system enough, not necessarily redundancy, but
other provisions to insure reliability. One that you mentioned was front
end storage, the ability to store waste if you are down. Another one that
hasn't been mentioned, which is the point I wanted to make, was one of in-
termediate storage, surge storage within a system. A lot of times a system
goes down because of one individual piece of equipment. Maybe there is one
piece of equipment prone to a lot of maintenance. In most RDF systems it's
probably the shredder, but in a system like Franklin maybe it's some other
piece of equipment more internal to the system. The way to overcome that
was to put in surge storage within the system, so that you could be oper-
ating the front end, and you could be operating the back end, but somewhere
inside the system you could be piling up, say, the crude pulp while you
are working on the screen. You have 2 or 3 hours storage capacity and you
can clean that out. You could be surge storing while you are doing repairs
and go back on-line, and you've never really upset the overall throughput of
the plant. So that's a factor, oversizing certain pieces of equipment
relative to others that you know are going to be down more often than the
others. Those things all go into an equation to determine how to design
the system.
WISELY: This, I think, Steve, is perfectly true for a small facility such
as Franklin, when you have a relatively few tons per hour that you are talk-
ing about. For any major facility, if you are talking in terms of say, any-
where from 70 to 80 to 100 or more tons per hour, then you'd pose a real
problem in trying to put that intermediate storage into the facility. I
don't think in most cases that you can be justified at all. I never even
looked at that particular aspect, and I would doubt it very much.
135
-------
SPENCER: In the Monroe County plant, we did reliability and availability
analysis, and did a lot of the things that Steve was just talking about.
In the tipping area we have storage for about one day's feed, which is
2,000 tons per day. We have about 2,000 tons storage in the tipping area.
We looked at the probability of complete electrical outages and things like
that, and determined that since we are right next to a substation, over the
last five years there were only two outages, all of which were less than 3
hours over the past 5 years, and decided it wasn't worth spending a lot of
money for. What we are worrying about weren't power outages, but we do have
a capability of operating as a transfer station in the event that those
lines are down, so that we have an auxiliary hookup to a compactor in the
tipping area, which has operated on the order of 20 hours a day. It can
move all of the material from the west side of the city out directly to a
landfill through transfer trailers. We also have inside the plant, and
spent a fair amount of money doing, putting in extra conveyors for bypasses
directly from intermediate stages within the process out into bins and di-
rectly into trucks for transporting. Our transportation analysis allowed,
then, for moving variable types of materials through given bins in order
to keep material. Let's say in the event that the wet processing stage is
down, for example, there's an intermediate bin that allows about 4 hours
of storage intermediate to fix a particular problem. In the event that
that exceeds the 4-hour period, we can then operate a bypass from there
that directly removes the feed to that particular module out to landfill.
Of course, with some of the products there are problems with respect to the
materials handling in terms of densifying this product and so forth. We've
added additional machinery and equipment in order to be sure that we can
handle the volume as well as what we are talking about, in the event that we
have to call for additional trucks, which we won't own at the time, or that
we will be able to densify the material to the point that we will be able
to move it at realistic costs. We also have redundant processing lines, and
we did attempt to do reliability availability analysis. As you know, the
availability of this string of equipment is the multiplication of the avail-
abilities of the individual items. The biggest problem is that you haven't
got the unit operation information that you need to do a proper analysis.
If you want to know what the reliability of a transistor is, you go to a
catalog and you look it up and you know that out of a thousand transistors
that one is going to fail every so many years. If you look at the availa-
bility of shredders, the basic information doesn't exist and so it's not
possible to do any kind of a rigorous analysis. As a result you are, there-
fore, left really to try to evaluate on the basis of judgment, engineering
judgment. With what operating experience we do have, we try to assess what
might happen. We looked at lead times of all the various spare parts for
each individual unit operation, and what the repair times were, and went
through rigorous analysis to determine what our inventory policies ought to
be in terms of should we spare an extra rotor, should we spare an extra
motor? In preparing our specifications, for example, since we had several
shredders in the plant we did use single frame sizes for our motors so that
a 1,000, 800 and 122 horsepower motor would all fit on the same frames. So
if we spare one 1200 horsepower motor, in the event that we lose one of the
other three, we can at least until we get a new one on-line, pull that out
136
-------
of our inventory and put it directly into the plant. On the other hand, I
think that while it's clear that it does cost money to inventory, it becomes
more difficult when you look at individual manufacturers, in that the manu-
facturers have their own inventory control policies. Williams, for ex-
ample, does not have any spare rotors available for any of their shredders
or any spare motors or anything like that, and let's say lead time on a
shredder rotor might be 8 months. Newell, on the other hand, will spare
all of these things and will have them within 3 days. The exact analysis
is really a function of the equipment manufacturers whom you are purchasing
from, which even complicates it further. My own feeling is that it doesn't
make sense to spend government research dollars on that particular analysis,
because it's very system oriented, it's very manufacturer oriented. The
general principles of how to do it are well known, but the basic unit oper-
ation information is missing, so that I think that it's really futile to
try to generalize anything there. I think you just have to look at each
situation, each design problem, and how much capital is available, and make
judgment decisions based on that particular situation. In any event my
recommendation would not be to spend design funds on availability analysis.
RYDER: Generally I would agree with that, but from a slightly different
viewpoint. We have built-in redundancy because really, we have no choice.
Those wonderful folks in the EPA are forcing us to shut down one of our
incinerators, which has one of the best burnouts in the city. Unfortunately
it doesn't have any air pollution controls on it to speak of, other than
water sprays, and therefore we are going to have to shut that plant down.
That's 1,000 tons of refuse a day that's going to have to go somewhere. Now
if we had a single line RDF plant, and that single line goes down, where do
we go with that 1,000 tons a day? We've got to be able to dispose of it
somehow and therefore we need that second line, as landfills within the City
limits do not exist. Also there's a little more to that. There are two
ends to this plant. At one end we are producing a fuel, which if the
utility doesn't get, they don't really care much because they can just add
coal to replace the fuel that is not available. However, at the front end
of the plant, there is refuse which has to go somewhere. At this end of
the plant we don't have a choice. An additional advantage of redundancy
is that there is a possibility we may find that perhaps 10 percent is not
the proper ratio of refuse to coal. Maybe it ought to be 12 or 13 or 15.
Also, the market might get better. We might find another customer. I don't
know how all those could be cranked into any kind of an equation with all
of these unknowns.
ALTER: Let's turn now from this emphasis on the city's viewpoint on risk
and uncertainty, to the A&Es. What sort of information are they looking
for, feel they must have to make the judgments for their customers, the
cities?
FUNK: What if you establish some guidelines? This is a subject we can
talk about from one end to the other. Why don't you get specific?
137
-------
ALTER: I don't know if I have to get very specific. You have a number of
clients and potential clients. You have to advise them, swing them to one
system or another, one means of processing or another. There is a desire
to move out and advance the state of the art or to do it'safely like every-
one else. What sorts of information are lacking, other than an existing
plant, to advance the state of the art?
FUNK: Some of the problems that are quite frustrating in our business is
evaluating pieces of equipment that may or may not be made available to a
municipality. Let's just pick air classification, for instance. When we
were developing and working with the City of Ames in putting their pro-
cessing together, we took bids on air classifiers. We got two bids. The
one outfit was going to scale up from about a 10 or 15 ton per hour plant.
It looked like a pretty good outfit, but it had never had any experience
operating on a large scale. The other bidder had one that was going to go
on-line, and we could evaluate that performance before we made our final
commitment. So we really didn't have a lot of choice. There was one avail-
able. That was four years ago. I'm not sure but what the exact same con-
dition doesn't exist today. The people that do have air classifiers, or
something that will work in that function, most of them are proprietary
systems, and most of them we've consulted and asked them if they would bid
an air classifier in a system, and a lot of them were reluctant to do so,
and understandably so. Raytheon is one of them. Boeing has an air clas-
sifier; Allis-Chalmers, I guess might market one, but they've never demon-
strated one on a large scale. You can go right down the line, there's a MAC
equipment offering the system now that they claim is the greatest thing
since sliced bread, but they have never demonstrated it on a large scale,'
so we are still stuck with the same position. Triple/S (I'm waiting for
Chicago to get going)—they are offering systems, but until they have demon-
strated, sure you can write performance specs, but the fact remains that if
that piece of equipment does not work, you've got a white elephant sitting
there until you get that thing squared away and get it operating. So
there's quite a few areas. That's just one of the frustrations that comes
in, being able to get your hands on what technology is available other than
going the proprietary route.
CAMBOURELIS: I think you are describing as far as Raytheon is concerned,
an accident of time. We perceived early on the necessity for coarse shred-
ding to minimize glass content. We couldn't find a good kind of an air
classifier, and ended up having to select an appropriate technique, and
this resulted in the rotary drum. Now you folks got started 2-1/2 years ago
making a selection process, and at that time there really wasn't anything
for us to give you. Being a conservative sort of organization that re-
quires of itself the demonstrated design, there wasn't anything there at
that time demonstrated full-scale, which may be available right now, which
would be a different kind of situation. We in fact do have a full-scale
rotary drum, and they are getting now full-scale results, so we are in a
different situation. I think that's one of the things that the conference
in general has to recognize, that these things have never happened before,
138
-------
many of them. It's the first time around. You have to be a little patient.
In particular ASTM activities, where specifications are required, but you
can't develop them until you've done it, and until you've done it you can't
do it.
ALTER: You both hint at something that I would like to hear from the A&Es
and others on—the idea in this field of incomplete information and ex-
perience, going out on a performance spec, Number 1, and Number 2, being
faced by a "buy-in," where a manufacturer bids low. He wants his first
sale so he can get the experience. Especially for the A&E but also for
everyone else, what's your reception to such offers? Do you think it is
possible to write the performance spec in a way to accept the buy-in and
thus move along the deal?
WISELY: I personally do not think so. I think that any rational specifi-
cation, particularly for a municipal corporation, has to be a combination of
good performance and hardware. I think that otherwise, in general, the
kinds of responses you are going to get for a request for bids might well be
impossible to evaluate, that is on a comparative basis. This doesn't neces-
sarily mean that you have to confine your request for bids for one specific
generic type of equipment. There's no reason why you can't have an alternate
in there. Even at that, I think that you must make an attempt to define the
hardware element of it along with the actual performance.
RYDER: The alternate to that, Ed, is a performance spec. Somebody's got to
design it, either it's a manufacturer or an A&E. Is there any A&E in this
room who can stand up and say, I'll design a classifier and have somebody
build it and guarantee that it's going to work? I don't think there is.
Therefore, who is going to be responsible for performance? It's got to be
the manufacturer.
WISELY: I did not mean to imply that an A&E should put himself in the po-
sition of designing proprietary equipment. No, but you certainly can define
things to a point that would permit you to get comparable bids. I don't
think that based on performance alone, you can get any kind of bids that are
going to be worth a damn, really.
HASSELRIIS: The advantage of a contractor taking total responsibility for a
plant (that's just the way it works) has made it possible for us to accept
whatever type of guarantee a manufacturer might offer, but not really have
to depend on it too much. It made it possible for us to buy a piece of
machinery, get it in the plant quickly, and find out what's wrong with it,
and fix it ourselves, and simply not expect to have to see something proven
beforehand. That's a risky procedure, and it's the only short cut to every-
body standing around and waiting for somebody to produce something that's
guaranteed and never have it happen. We have been able to put our plant
through about four generations of improvement in 4 years simply by going
ahead, failing sometimes, succeeding sometimes, but always managing to work
and find out what the faults are. Even today, we received one of these—
like a feed bin—and it works in principle. It wasn't designed for the
139
-------
purpose, it is an available piece of equipment, and we find we have to re-
build it. We find that in certain places the welds are wrong, the shear
pins that they were protected from major failure don't exist. There are
many, many faults in it, and we end up rebuilding it. What does this say,
is that you can't hold manufacturers or anybody else really responsible for
a lot of things, but the simplest thing to do is to allow in your budget a
major retrofit budget. In other words, whatever you do, figure that you are
going to have to spend a substantial amount of money changing the equipment
and improving the equipment, and you're going to build it on the job. Now
you can appreciate that municipal bidding doesn't permit this very readily,
but I have heard one municipal operator say that if he had 20 percent extra
in the price of his total plant available for retrofit it would be about
right. If he had had that money to spend himself, it would have been possi-
ble, but if he had to do nothing but blame the manufacturers and get them to
do their job, it never worked. So that's a recommendation that people think
along those lines to allow the money to go ahead and make these improvements
on the equipment as they receive it, and it is one way out.
WARE: In our position we find ourselves constrained in two ways. We are a
private developer. We are trying to come up with our own technology, new
ideas, we stick our neck out. At the same time we are expending government
money, which means we are required to go to the lowest bidder in most cases.
In both circumstances we have found a full spectrum of response from manu-
facturers. When we write a spec as well as we can, and the equipment that
we receive doesn't perform, we've found the full spectrum of responses,
from the manufacturer being willing to do anything it takes to get his ma-
chine to work, to absolutely no interest. The responses haven't been a
function of the size of the vendor, or the nature of their business; I
think it is very much a personal thing. Some vendors are selling us pieces
of equipment that they can't hope to make money on in the future in similar
businesses. Others, like shredder manufacturers, one would think they had
a lot more potential in their future markets for these machines. All I can
say is is that you have to investigate a vendor for his ability to make a
machine, the availability of the parts, the performance of a similar piece
of equipment in the past, and finally, and it's quite an important point,
his responsiveness. His responsiveness you can generally only find from
other users, but I think it's worthwhile when you are comparing bids, and
when you are selecting a vendor, to have gone to a previous owner of
similar equipment or somebody who has dealt with that vendor and find out
what kind of response you have. Specifically, in the case of our shredder,
I mentioned yesterday it isn't performing the way we expected it to, it's
certainly performing worse than other similar machines, the vendor is most
happy to supply $10,000 worth of interchangeable parts to make it work
better, and at no cost to us. This is in the performance guarantee. We
have performance guarantees on every single piece of equipment we've ever
bought, but enforcing them is a totally different world. I would again
recommend that anybody here who plans to buy a piece of equipment should in-
vestigate previous owners of that vendor's material.
140
-------
SPENCER: Pete, did you have performance bonds on all of your equipment as
well?
WARE: We didn't have performance bonds, simply a specification requirement
that we would retain a percentage of the purchase price.
SPENCER: Did you get away with this?
WARE: Only in a few cases where the delivery was such that we were still
within the warranty period by the time the machine started up. We still
had retained 10 or 15 percent.
ALTER: Was the warranty more than materials and workmanship?
WARE: No, it was variable with the machine, because a lot of cases the
vendor came with a proposal that for some reason or other didn't incorporate
or claimed exceptions. But in most cases they were related to performance.
In other words, we gave them extreme characteristics and required a certain
product out the end of the machine, and have required them to stand by that.
SPENCER: To Bill Ryder. You have performance bonds, certainly, on your RDF
receiving facility which you mentioned yesterday. Do you expect to have any
problems in terms of exercising the vendor or supplier to meet his perform-
ance obligations?
RYDER: Not in the least. Number 1, our specs call for 90 percent payment
on delivery. When he delivers that item to the site to be installed by some-
body else, he gets 90 percent and we hold back 10. He doesn't get the other
10 until we accept it, so we've got that for leverage. Number 2, we have
liquidated damages, depending on the contract price, varying anywhere from
$500 to $1000 a day if he doesn't meet his delivery date so that's something
else. The third is a full performance bond for the entire value of the con-
tract. If the manufacturer cannot meet performance or the company goes un-
der, we call in his bonding company who must complete the terms of the con-
tract at their expense. We've done this before and it has worked.
ALTER: Did you find this cost-increasing? For example, we once asked for
a shredder bin on the basis of if it doesn't work and doesn't handle the
feedstock to the performance spec, then the manufacturer will pull it. He
agree i and increased his price 25 percent to self-insure. I'm over-simpli-
fying, but did you find that these requirements were indeed cost-increasing?
RYDER: I really don't know; I'm sure it must add something.
ALTER: If it were 25 percent, it may be pretty near a tradeoff.
RYDER: If I understand what you are saying, in our contracts which have
liquidated damages amounting to,let's say $1000 a day, the contractor, know-
141
-------
ing full well that he cannot meet the delivery date, will merely crank
enough money into the bid to cover $1000 a day for each day that he figures
he will be late. However, that has its limitations, to exaggerate the case,
for example, he can't double his bid price, or he'll no longer be com-
petitive.
SPENCER: We had the same experience. 1 found that we went to municipal
bidding, but our process contractor certainly had the responsibility to do
the same kind of things that Bill was talking about. In fact through compe-
tition, the vendor's first reaction was we'll just crank 25 percent into the
price, and we said we'll just talk to somebody else. And after a while,
when you are pricing orders for seven shredders, where the order is a pretty
large order, and there are some shredder manufacturers who are interested in
doing business, basically the guys who took that position are just thrown out
of the office until we got people in who were willing to sit down and wanted
the business. That came back to the thing that you were talking about. The
guys who were willing to take the risk were the guys that wanted the busi-
ness, and the ones that were interested in staying in resource recovery and
making a good showing. They were willing to add these things in, more at
their cost rather than at their risk. Those games were played, I guess I'm
saying, but we played them right back, and by and large have gotten now
performance guarantees from everyone.
LAMB: I think good buying and performance guarantees are very good proce-
dures wherever they can be effected. I think it is extremely unrealistic
to expect that you are going to let your plant startup go to a series of
equipment suppliers who have performance guarantees. First, you won't be
able to spend the time, because when the equipment doesn't perform it's not
always clear why, and you can't afford to have the conferences, and if you've
got three pieces of equipment in a row, sit around the table and play ring
around the rosy—if mine was working, yours would be, or because yours is
not mine isn't. The owner or operator of the system is going to have to be
prepared to take precipitous action. He's not always going to have the
luxury to insist on the contractor's performance, and contractor's not al-
ways going to be performing on a timely basis. For example, in Connecticut,
after a certain date, it is required that the bond debt be retired whether
or not the plant is operating. So the owner of that plant incurs a cost of
about $10,000 per day when the plant's down. He can't be sitting around
with a rinky dink contractor who has got a $100,000 contract, and think gee,
should I spend $5,000 to replace this part or that part. I really don't
think that's very realistic. If you look at the contingent liabilities and
costs and damages of not running a plant, there's simply no way that you can
take much satisfaction from the kind of guarantees that you can expect from
equipment suppliers; they simply can't do it, gentlemen. They can't take on
millions of dollars of contingent liability in the sale of a few hundred
thousand or million dollars. Their profit margin just isn't there. The
owners don't like to take those responsibilities on; I think it's equally
unrealistic to expect suppliers to.
142
-------
ALTER: I'd like to hear from the military on these points, and I haven't
at all. I see the Air Force pointing to the Navy, or to the Army, which is
it? Who is the senior service here?
HURLEY: I'm not sure that I could address the questions asked with respect
to guarantees, but I think it is recognized today the question of can we
continue to operate the systems that we are operating and at the same time,
what if, indeed, there is some fuel there, fuel value in that solid waste
system, how can we best utilize it? So I look at it as a question of what
are you willing to do to change the way you are operating today? And to
what extent is this a people/machine system? Most of the discussion is
about the equipment, but I think you buy equipment when you don't want to do
it with people. If we can think that while we want to buy the least amount
of process equipment (I'm looking at MRI's chart here), we'd like to get the
fuel with the least investment in equipment, which means to me that we have
to at the same time look at what people are willing to do at the head end of
the system. That may be begging the question, but we have to try, I think,
to determine how to get what fuel we can with the least amount of equipment,
and try at the same time to see what you can get people to do at the front
end—source segregate and other things.
OLFENBUTTEL: What we are essentially looking at in the military, of course,
is the lowest capital investment alternative, which means source separation.
However, I think we have to recognize that for the military situation we may
have to go to other alternatives because of economics, lack of markets, etc.
Some of the alternatives may be in regional areas where we have to consider
large (what we'll call large) 100 ton a day and larger resource recovery
systems. I don't think that is realistic except in a few areas which have
been designated by EPA for Federal lead in regional resource recovery be-
cause of waste generation quantities, lack of disposal sites, sanitary land-
fills, and such. We are trying to look at waste heat incinerators, but
we're talking about small waste heat incinerators—20 tons per day or some-
thing like that. There's just a lack of information on the operating and
maintenance characteristics and costs of these systems. I'm just reinforc-
ing what we all have said about all of the other equipment: You can't get a
handle on any operating experiences and associated system reliability per-
formance. I'm not saying anything new. I saw a report that came out of
EPA very recently for three incinerators which were investigated by a con-
tractor, only one of which was a waste heat recovery incinerator, in Ar-
kansas. The data is really lacking there, also, as to how much it is cost-
ing and what the reliability is. The report used four months' worth of data,
which is really insufficient to make projections for planning purposes.
(I'm sure that EPA will recognize that that's not a really good length of
time for collecting reliable data). So, we in the military have to consider
all of these options, and no matter where we look, there's just a lack of
reliable information.
ALTER: I think everybody is leading down a single path in research needs.
Perhaps it can be pulled together. There is perhaps a way of dumping the
entire R&D responsibility on equipment vendors. The way of doing it is
143
-------
called the performance guarantee. Indeed, if everyone had the same lia-
bility, if the only way you could play the game, the only way you could
sell equipment, was to succumb to a performance guarantee and recognizing
this would increase cost, it would be a cost that presumably the equipment
suppliers would have to spend in doing their own R&D work so that they could
have some assurances about their equipment fulfilling the performance guar-
antee. I'd like to lead into this alternative to R&D and then into our R&D
agenda, by first asking Dave Spencer to go into a little more detail on how
they used performance guarantees in Monroe County.
SPENCER: I don't know how many of you are familiar with the Monroe County
project, which is not as far along as some of the others like Chicago and
Ames and so forth. We're into construction now. We're about 5 percent into
construction, with the bad weather in Buffalo and strikes, we are a little
slower than what we had hoped. We're maybe a couple of weeks behind sched-
ule. Monroe County took a sort of typical architectural engineering ap-
proach, but with a funny twist, in that they wanted to control the job at
every stage, which I think makes sense for an owner to do. They also wanted
to some extent to minimize the risk on the system contractor, not so much be-
cause they cared about what risk he was taking necessarily, but that they
felt that that would minimize their costs, which I think tends to be true,
that cost and risk are somewhat associated. So basically what they did is
they kind of stretched the rules of New York State and selected what they
called a system manager who would provide professional services rather than
a turnkey or full service contractor who would own, operate, and so forth.
They twisted the rules a little bit further and said not only will it be a
professional service, therefore buying service on a non-competitive basis,
but that as part of the professional service they wanted design and con-
struction management and 5 years of operation. At the end of the 5 years,
they'd have the right to bid the operations further to other people. They
wanted that contractor to have a stake in the job, but in all phases of the
job, so that we'd be worried, scared during the design and during construc-
tion to make sure that everything worked out for them, even though they
wouldn't be paying us that much for the effort that we were making. We had
worked with the Bureau of Mines, I guess we've had a cooperative agreement
now about 5 years. The system that was developed for Monroe County was by
and large along the lines of the U.S. Bureau of Mines pilot plant in terms
of process configuration, although the equipment to perform each of those
unit operations was in some cases different. For example, the horizontal
air classifier at the Bureau; we weren't particularly wild about in terms
of its ability to handle surges and so forth, but yet we had a gut feeling
that we should be trying to only do a very coarse shred in the first step of
operations and that we should only try to, say, do the total size reduction
after we got down. When we knew we had paper, then we'd try to reduce it
to a fine size, not when we had run of the mine kind of material. We took
the position that, based on the data that we had, we would give a perform-
ance guarantee for the total system, that we would market the output products
to specific specifications which we had developed based on the work of the
144
-------
Bureau of Mines, and enter into long-term agreements with buyers of the out-
put products, and then have guaranteed for the county that the process, the
total process, would meet those output product specifications. In the event
that it didn't, we wanted them to have a stake in it as well, but beyond a
certain amount, about 10 percent of the construction cost, that we would to
an unlimited extent guarantee making all modifications necessary at our own
expense, that we would meet those specifications, and that extended not only
to, for example, the RDF product, which has very low ash content and so
forth, to meet Rochester Gas and Electric specifications, but even the glass
products, where the inorganic content has to be particularly low. So basic-
ally that's the approach we took. Then we developed specifications that
would pass down the requirements for an individual unit operation directly
to the equipment manufacturers. We weren't about to stand in the middle
of how an equipment supplier should design his piece of equipment. We did
do as Ed Wisely said: We wrote performance as well as hardware specifi-
cations, so that there was a performance requirement with bonds, with all
kinds of sledge hammers that you could wield, and other liabilities that
manufacturers will be responsible for, but not things that they can't be
responsible for. We only try to make them responsible for their individual
items of equipment, and have passed all that responsibility down for each
unit operation to an equipment supplier. In the case of the rotary drum
air classifier, for example, if there's a problem with that, Iowa Manufac-
turing, who is the supplier of it, will have to repair that at their own
expense, not from where I sit in the program office, not at our expense.
The same thing would be true of Williams, and we tried to tie together sub-
systems, like infeed conveyors, shredders, and discharge conveyors, so that
a vendor would supply a subsystem, which he could be held responsible for.
LAMB: Excuse me, Dave, what was the nature of some of these performance
specs, such as the air classifier? Other than mechanically moving and bear-
ings not burning, and motors turning, what other kinds of performance were
you able to pass down?
SPENCER: Well, we are passing down that it will take a certain percentage
of light fraction, and that it will have no more than x percent free inert
material in it.
LAMB: What's the liability if it doesn't meet that?
SPENCER: It's complete with respect to that item of equipment, that if it
doesn't meet that, that they would have to replace it totally, at their own
expense, refund your money, or to the extent that it doesn't meet the total
capacity requirements, there we have an option, or I think we'll have an
option, we are still working out the certificates of compliance now. We
might just accept a refund of a certain percentage of the equipment price,
if that added performance which is specified is determined by experience not
to be required.
LAMB: In the worst cases, are you saying that the maximum extent to which
the contractor is liable is the purchase price of the equipment? Didn't
you say, for example, he had the right to remove it and refund the money?
145
-------
SPENCER: That's the most that you can reasonably pass down to a manu-
facturer.
LAMB: That's right, so that he really has no obligation as far as running
the plant is concerned, or really any contingent liability.
SPENCER: No, but that's not to say that he doesn't have an incentive.
LAMB: I appreciate that, but again, it's the moral, ethical, professional,
reputational incentive.
SPENCER: No, it's not that. It's the withholding, and it's the performance
bond, as well as the materials and supply bond. I think with those bonds
and with the withholding, you have a sledge hammer. There are certain i^.
quirements in there, that if he doesn't take certain actions within given
time frames and otherwise, you notify his surety. You let either the surety
come in and work it out or the surety [unintelligible].
LAMB: You say in the worst case, the maximum extent of the liability is the
refund purchase price of the equipment.
SPENCER: I don't think you can expect any more than that.
LAMB: I'm agreeing with you that that's right, but regardless of what sure-
ty says, or the lawyers or anybody else, the maximum extent of the liability
which they incur is, in the worst case, to pull their equipment out and re-
fund, so you have a process with a piece missing.
SPENCER: Yes, except what you really have is not that. What you really
have is a vendor who supplied a piece of equipment that he's so sure will
work, in other words, he's competing pricewise also and lots of other ways.
The way I would look at that, Tom, would be that he has been conservative
enough in developing the design of his equipment that he is sure that he
won't have to do that. Because of the risk that he incurs if there's any
probability that he's going to have to pull out the equipment and take it
away as scrap, I think that he'll be very conservative in what he supplies
and he'll be very fast in terms of making what repairs he can. It's true
that if you get to the end point, and you reach a point of impossibility,
then it's true you are holding the bag, and I guess in this case Raytheon
is holding the bag.
ALTER: Let me have just one more comment on this topic, because we have to
get into our research agendas.
CHANTLAND: Who assumes the cost of the delay in downtime? Is that Raytheon?
SPENCER: It depends. In terms of completing the construction, we have a
liquidated damages of $5,000 a day. In terms of equipment items you can't
expect a guy supplying a hundred dollar pump to accept a $5,000 a day liqui-
146
-------
dated damages charge. We do have some equipment items, where we are par-
ticularly concerned liquidated damages, but they are reasonable based upon
the price of that equipment. Again, we are not trying to get the equipment
suppliers to pay for problems which they can't control; we're only trying to
give them a strong financial incentive to rapidly maintain, repair, or do
whatever they have to to their equipment. We will also do it for them if
they choose.
ALTER: This leads into our research agenda. Now we can do this in one of
two ways. We can attempt to go through the MRI booklet that was handed out
yesterday, and I am disinclined to do so. The MRI staff has put a lot of
time, effort, and thought into the book, and I think it presumptuous of us
to sit around and nit pick and criticize. However, it is a starting point.
When I went through the MR! book, I was impressed that the objectives were
stated and an approach was outlined, where the reader is directed toward
what a project might look like. I also think that in a group of this size
and sort, we may not be able to do likewise. For this reason, I would like
to differentiate between what I call the research agenda and a research pro-
gram. I think it unlikely that we can design the program today, and we
certainly should avoid trying to design the experiments. What I think we
can talk about are the categories of items and items within the categories.
Yesterday we talked in two categories, you will recall. One is systems pro-
cess design and the second was selection of equipment/unit operations. I
would like to see if we can structure the discussion at this point to re-
search agenda in those two categories. While we are doing this, I ask each
of you to take a piece of paper and write your notes as to what you think
is important. If you will, we are taking a vote. It's really a consensus.
I think when your notes are handed in, they will assist our EPA hosts to
distill some sort of final consensus, which I presume will be included in
the proceedings and which you will see and can react to. Write your notes
as to what you think is important, or if the discussion doesn't get around
to your pet subject, write it down. I think if you put your name on the
piece of paper, you will also give an opportunity to the EPA staff to get
back to you if they have a question. With that, let's try to address an
agenda of systems process design, the sorts of information which are neces-
sary to reduce this technical uncertainty, and hence risk for the actors in
the field; the municipal official, the A&E, and the systems process pro-
vider, and those combinations in between. Let's open the discussion with
research items for systems process design. What subjects need to be at-
tacked on a high priority basis, in order to make the judgments?
RIGO: I think that there is something that even comes before those, I think
it's a legitimate research need. Some of the work we've done implies that
you can almost halve the cost of processing if you can get the households to
do a rough sort for you. One that we looked at would split the glass and
metals from everything else. This does some very interesting things to your
processing line and cost. I think that it's very critical that some good
work be done on source separation and degrees of compliance, looking not at
147
-------
recovering ledger paper, not at recovering newsprint, but instead looking at
a segregation technique that makes sense with mechanical processing as a
secondary step. I think this is critical.
ALTER: Now wait, before we proceed further everyone ought to be aware of
one, the OSW-sponsored work at Marblehead and Somerville along these lines;
txro, the Swedes started this five years ago, and they have been working on
it.
RIGO: They did it in Cleveland up until 1954.
ALTER: They tried something like this in Florida with different colored
bags, specifically with aluminum. LA was different, LA was like Philadel-
phia, where they separate the garbage from the trash. Are we talking about
behavioral or technological problems? In Sweden, when they tried to recover
the glass, the material needed more processing which might require equipment
the same as being installed in Monroe, Hempstead, and New Orleans. Do we
discuss, do we start this way? What is the consensus? You haven't touched
on "pre-sort" as a first step.
H. FREEMAN: I feel it's a little outside the scope that I would like to
hear. I would prefer to hear the research needs directed more towards the
machines, and save that for another meeting.
BURCKLE: I'd just like to comment. We are well aware of this need, and the
Office of Solid Waste is sponsoring studies in this area. We hoped that
based on their study we can compare, get some idea of looking at collection
and transportation costs as a trade-off against processing costs. Our basic
problem is, we have to fill up the gap on the processing side on the mixed
waste, and the information that we get on the mixed waste and processing of
mixed waste. Hopefully we can make those kind of cross comparisons.
RIGO: One quick comment. The type of source separation we want to do,
when you are looking at simplifying mechanical processing, is different
than trying to retrieve components that are individually marketable as re-
ceived. I'm not talking about that. But the interesting thing is that if
we get good unit operations data, then I agree with you, solving the mixed
waste problem will give us the data base to solve the segregated waste
problem. But if we are not making a strong attack on unit operations, then
you're going to have to set up a different type of R&D program, because your
processing lines do look substantially different. They use different com-
ponents, different component sequences.
ALTER: Because of that, I think we ought to take the tack that Harry Free-
man suggests. However, the decision by a community to institute source
segregation (with or without some processing) may turn out to be a political
decision. For example, how do you enforce the requirements that people (and
especially old people) separate their refuse? Return to systems design and
what information is necessary. We went through many alternatives yesterday.
148
-------
Let's try to assign priorities. For example, George Trezek presented an
attempt to compute the energy of various combinations of unit operations.
la this the sort of task that needs to be continued? Let's start with the
prejudiced viewpoint, George?
TREZEK: No comment.
BENDERSKY: I'm going to put George on the spot here. I think the work that
George did on shredders is indicative of what can be done in other unit oper-
ations equipment. I think it's directly applicable, for example, for the
study on air classification to do the same kind of manipulation George did
on shredders, do it on air classifiers and perhaps some other basic unit
equipment. In other words there are variables that are controllable in
these pieces of equipment. There's fundamental information to be gathered.
Sometimes you can't put it into a nice, neat equation, but even when you
can't basic information is needed on equipment like shredders and magnetic
separators and other items. I suggest that more thought be given to similar
work on unit operations at a laboratory level as well as full-scale and
other tests.
TREZEK: We are going to do some similar kinds of studies on densifiers—
pellet mills, densifying, this kind of thing—and probably we will evolve
similar kinds of curves so that we can predict the final step in the process
of what the densifying step adds to the energy.
ALTER: Let me comment a bit on the air classifier. We attempted such
things, as David said, and found we were frustrated by the inability to
easily change design of the air classifier at a reasonable throughput.
Reasonable, I'll say, is 10 tons an hour, but really higher. As was dis-
cussed yesterday, in order to understand the unit operation, you almost
have to destroy it. You have to examine orderly change even if it results
in a detrimental performance, and this is different from going into a plant
and trying to tune to optimum performance. Indeed the kind of data that
George has been able to obtain in shredding is difficult if not impossible
to get in a plant. We are proceeding with an air classifier by building
our own. It's a vertical zig-zag that can be easily changed in terms of
width, depth, angle of zig-zags, number of zig-zags, feedpoint entry, and
the point of reintroduction of recycled air. It's a research 10 to 20 tons
per hour unit. I hope we have the time and resources to fully utilize this
device. In researching this unit operation (or others), there are different
approaches representing a spectrum of research philosophy. Perhaps this
spectrum extends from plant operator through designers and equipment ven-
dors, government laboratories, and universities with their need to educate.
TREZEK: I think that our work has been useful related to reality.
ALTER: It's absolutely useful. I want to construct the spectrum. I'm not
going to say what's on the left and what's on the right, or what's right
and what's wrong. One part of the spectrum is university. Let me assert
149
-------
that another part of the spectrum is the occasional measurement in the
plant, or somebody writes a paper in Public Works Magazine, or such, where
they purposely do not go into a lot of detail, and maybe they don't have the
detail to go into. What some of us have tried to do is something in between.
What we've tried to do is take enough data to relieve anxiety and show that
something can be done without getting a lot of detail, a sort of middle
ground, which I hope is useful. When we start constructing these research
agendas, how much detail do we go, where is this spectrum? I don't mean to
imply you can't even go further in the sort of detail, as George did. Let
me go back to systems process design. We talk about alternatives. Given a
number of unit operations, A, B, C, and D, how far do you go in looking at
the combinations and permutations of how you might arrange these? Do we go
to one end of the spectrum, to the other end, which is most useful?
H. FREEMAN: I wanted to ask a question, Harvey. Just from the conversation
that I've heard this morning, I haven't heard a whole lot of technical ob-
stacles brought up. We seem to have gotten off of that rather quickly and
into other things. What I'm hearing now is the need to evaluate these dif-
ferent pieces of equipment. I guess my first question is: Am I hearing
this right, am I hearing the users say there are really no technical ob-
stacles, we just need better information on what we have? My second ques-
tion is: In setting up a research agenda, are we drawing our lines too
narrow to just talk about going out and evaluating what is there? I guess
I would like to hear that tossed around a little bit.
HURLEY: I think we are trying to apply what we have been told is available.
It's other people's experience, it's EPA knowledge, it's College Park exper-
ience, and with a guideline now that says we must develop plans for resource
recovery. What have we got in what we call the small-scale, or suitable for
a small-scale operation. I think the Navy program is one that says let's
apply what is known now. That's what we hope to do, with a field experiment
working on actual waste, not in the laboratory, but get out there on a
[unintelligible] operation and see what it costs and what the problems are
and whether you should give up the ghost and just join all regional systems,
or whether you can in fact utilize it economically.
H. FREEMAN: I can appreciate that's probably what most people here are in-
terested in, but I would hope you wouldn't lose sight of the fact that we
are talking about a research program. At least here in this room we have
the luxury of being somewhat outside the lines and this may indeed be nar-
rowed when David Berg's group gets restrictive. They restrict us and they
in turn are restricted by a bigger restrictor, and this goes on. I think
here we should keep in mind that we have a big world to talk about. Let's
don't just talk about evaluating the XYZ shredders, unless of course that is
the optimum that we can ever expect from all shredding operations.
WISELY: There appears to have been at least a tendency in this group to at
least refer to the desirability of deriving rather precise knowledge in a
150
-------
very, very imprecise field. From a practical point of view, from a stand-
point of applying processes such as we are discussing here, this precise
knowledge really isn't worth much. Some of the more pragmatic people on
the other side of the table referred to this to some degree yesterday, but
I just wanted to bring it up again. There are types of equipment that have
some potential desirability in application—small ones such as yours, Mr.
Hurley. To the best of my knowledge, there are not data as to what kinds
of information can be derived from those. I was going to suggest that
rather than concentrate on the XYZ shredders as Mr. Freeman talked about,
that you address yourself to some of the kinds of equipment that may have
potential and may be in the long run far superior to anything we know about
right now. We know what some of this equipment is, but we don't know how
it performs. I'd like to talk to you about that some time, Harvey.
ALTER: Thank you. You raised a very good point. You talk about over-pre-
ciseness, too much detailed information. Then what is the type of informa-
tion that we need?
WISELY: As the gentleman over here mentioned yesterday, what we need are
ranges of information for various applicable parameters.
LALKA: A lot of data has been accumulated on pilot plant operations, small-
scale shredders, that type of information. Listening to Peter Ware, he indi-
cated that his full-scale shredder is not doing the job of another unit that
apparently is comparable, plus he feels that his air classifier is more a
state of the art thing rather than something you can write equations around.
I think rather than concentrate efforts on gathering data on pilot plant
units and smaller units that have a scale-up problem we should concentrate
efforts on units that are in full-scale plants, and take a look at what kind
of information we can see from an actual operating plant. On a pilot plant
level people have a tendency to synthesize the feed material. It's not like
a packer truck is backed up to a pilot plant, and you put in real refuse;
sometimes it's kind of a synthesized feed stock.
ALTER: You'll get a lot of disagreement on that.
LALKA: Well, there could be a lot of discussion I'm sure.
ALTER: Roger uses real packer truck refuse. We use real packer truck re-
fuse. George uses it. We don't know how to simulate garbage.
LALKA: Right.
ALTER: We don't know what small-scale garbage looks like.
LALKA: That's one of the problems of scaling up from a pilot plant. Every
city has a different type of refuse, and presumably the operation of this
equipment varies tremendously with the composition of the refuse. I think
you could get some decent information just by looking at various parts of
151
-------
the country, that, for example, have full-scale shredders in it. If you
could just compare all of the information on those units, you'd probably
have a pretty good idea of what these units can do.
HASSELRIIS: I would go back to George Trezek's presentation as being a
good example of what can be done by putting together a mathematical model
of what's happening and of the utility of doing that. By looking at that
presentation, you can see why and how different shredders work for the first
time, once that clarification is made. But what none of us understood,
probably, in the beginning was what statistical distributions of particles
really mean, and how to work with them. What that presentation shows is
that the statistical curve is just an equation, and the equation can be
applied with another equation, and they come up with answers and correla-
tions and solutions. That procedure is very powerful, and it eliminates
this need for preciseness which we've been brought up with. If you wanted
to apply the same thing to an air classifier, I think you could find again
the same things would apply, that there's a mathematical model which would
give you some indication of what the implications of scale-up are. If you
find that a particle has to pass through a 2-foot bed instead of a 6-inch
bed, obviously it's a very, very different thing and I think it has some
relationship to the fact it was a 2-foot or 6-foot, or whatever. I think
this is something we can get a very great benefit from the university type
approach of a real study of the mathematical principles, the theoretical
principles that we are really working with. If that model is put together,
then each of us as users can use it to test our own information and to see
how close to perfect we are. For instance, if you did it with the shredder,
we find out that if one shredder is actually doing three times as good as
another for some reason, you'd know it right in the equations that it was
three times as good or three times as bad, and if you say that's my poten-
tial for improvement. If we already know there's no further potential for
improvement because it's inherent in the operation, we wouldn't try to
improve it. That's very important to know what our entropy level is or the
point of what is the most possible performance that could be achieved from
a piece of equipment and how far we are from it. We'd know how far we are
from it, we'd know how much we are willing to invest and how much improve-
ment we could expect. So I think that's a very important aspect—funda-
mental research that really comes up with the model that we can use to put
various inputs into and get some kind of an output, a useful output.
DE CESARE: Harvey, I believe many people here know of Taggart's book on
mineral dressing. They go around to existing plants without data now and
take machinery that is in operation and put down the full range, what that
machinery is doing in its installation. If EPA has money I don't know why
they can't get a mobile research lab set up to go out and monitor existing
shredding stations. Whether they have to have funds to pay each shredder's
installation for cooperation or not I don't know. Go there and find a
method of attaching equipment to monitor kilowatt-hours, so they can chart
kilowatt-hours versus feed rate, versus size of shred, versus moisture
152
-------
content, if possible. Perhaps at certain shredder installations they can
even vary sheave sizes to get speed changes. You have to work with the
people that are running the shredder to get them to cooperate, get good
information on maintenance and reliability, this type of thing. This would
be an intensive program that should be done on existing installations and
you'd be able to get quite a display of information this way.
ALTER: We are going to discuss the Taggart approach, I promise you, but
let's finish this conversation first.
BURCKLE: I'd like to go back and speak to the comments that Ron Lalka made a
few minutes ago, because I'm very concerned about these two points. One is
scale and the other is the nature of the site specific characteristics of
the waste. Because these so influence any kind of research program one sets
up, be it a pilot plant or field test situations, in terms of transplanting
that data for use to another site. If our data is site specific and cannot
be translated, there's no sense in spending the money to do the work, in my
opinion. Now, in terms of scale, I think there is a need to deal with the
scale factor problem, because not all plants are built at the same size. It
depends on the community needs. That's the one aspect. The other aspect is
site specificity. If we're stuck with saying that the data that comes out
won't give us the optimum design information for a given plant, what do we
do about this? How do we attack this problem? Is it really a problem? If
we operate on the band width concept, if we properly structure our band
width by first going into a number of communities and looking at how similar
or dissimilar that waste is to establish our band width, is that a legiti-
mate approach to structuring our research program, or is it even more im-
portant that we do that? I'd like some of our comments at this point from
the A&Es to deal with this problem, if that's possible.
FAY: I just had one comment. I'm not sure that the shredder people would
look with great favor on EPA coining out and testing their shredder and
publishing all the data. The only way you got automobile mileage is pass a
law saying they got to do it. While I can understand the desirability of
this from this group's aspect, I don't have much hope for it.
ALTER: I don't think we should dwell on it because George presented his
information, and we are going to publish ours. We went around, as I said
yesterday, to 10 operating commercial shredders; that data will be pub-
lished. I think there are other unit ops that need more of this kind of
attention. Shredders are only an example of how it might be done.
RIGO: I understand statements that we really don't need precise data on
what the components do. Today that's true. Looking at what we would like
to do in selecting plants 5 to 7 years from now, and that means we have to
start today getting data, it isn't true'. If we can characterize the waste,
and I don't mean a point, I mean get a good statistical description of what
that waste is and get a good statistical description of what the unit oper-
ations are doiag, it is possible to make an intelligent design using
153
-------
stoichiocastic design procedures that will come out with the "optimum"
plant. That plant will not be right for any condition, but on average over
the year and including projected changes in waste composition, it will be
best. There is a very strong need to get good data and not just something
that says well, we are going to fly 70 percent of the stuff that's thrown
into the air classifier. We also need to be looking at new unit operations—
ones that are not now applied, and we need to obtain data so that when we do
go out in the field (and this is lab data) that we only need to get a limit-
ed number of data points on the full-scale operation to verify the results
which use the lab data (where it's relatively inexpensive and convenient to
play) to get shapes of curves. When you go out in the field, you validate
the shape of the curve. So you need a laboratory effort in conjunction with
the full-scale application.
ALTER: We talked about this yesterday—about compositional data, good data
and bad, and so on and so forth. I have a personal prejudice that says this
is a cop-out. We can do all of the detailed compositional analysis we want
to do today, and we've frozen one day out of the year of this year. I don't
know how to project future waste composition; rather, we must have the abil-
ity in a processing plant to make adjustments to change. Am I wrong? Do
we take Greg's approach, or do we take the approach of a general sort of
compositional data, you need the band width plus or minus. He wants to nar-
row the band width.
RIGO: No.
ALTER: Is there then a research need to get better data or not?
BURCKLE: Harvey, I think this breaks down into the question that we are
dealing with here of Greg's point that the designer needs some sort of
model, if you will, so he can predict the performance of the equipment. It
seems to me (and please correct me if I'm inaccurate) the first thing a
designer needs is to be able to sit down and draw a process diagram, which
says I need this kind of performance out of each unit operation to get to
this kind of quality fuel and this production rate of fuel. Then he needs
to be able to fill in those blocks with specific pieces of equipment that
will give him the performance that he needs, taking into account the pro-
blem of the variability of the waste, both from community to community and
also in time. I think that that's the overall envelope of the problem that
we are dealing with, putting a preprocessing system in the field ready to
operate. Now on top of that we have to impress the fact that we've gone
perhaps 80 percent of the way to a totally operating system, and then we
have the shakedown period during plant startup. Please correct me if I am
inaccurate or have only stated the partial case. I'd really like to hear
from the experience of the A&Es and the people that have been involved in
the past in starting up these kind of plants. Is this or is this not the
kind of procedure that's in general followed? Because I think that bears
on our informational needs.
LAMB: My plea is for more fundamental data and on a unit operation basis.
You might make an analogy to a barrel of crude oil in a refinery. You don't
154
-------
tell a refiner what the mix of end products ought to be from that barrel of
crude oil he has. He needs the tools, distillation curves, hydrogenation
curves, and things like that, and he can make his own business decisions as
to how he wants to put the pieces of the plant together to optimize what he
thinks is his product mix. I think we need exactly the same data. I agree
with Harvey, I think the compositional thing is a cop-out, because there is
no way to deal with it anyway. Your day-to-day variations are far going to
exceed any precision that you are going to be able to achieve in the data.
We ought to know if we primary trommel what kinds of separations we can get,
and what's going to be the composition of the materials that pass through
and what over. We ought to know that maybe as a function of rotational
speed or hole size or are deck screens better than rotary screens. You can
go down—if you want to dry, how do you dry, do you dry counter-currently or
co-currently, do you dry while you aspirate or do you dry in a rotary dryer?
What are the tradeoffs? Where's the kind of basic data where a designer can
sit down and hypothesize the design and say if I get this heat transfer rate
in this particle size at least I have a feeling that the thing's going to be
about 10 feet in diameter and about 40 feet long. Fine, then I can go to
the system supplier if I want to buy one and he's going to have a model in
his plant and we're going to ship him some material and we're going to moni-
tor the performance on his piece of equipment. But we've done a lot of
thinking and a lot of screening before we've gotten to that point, and that's
the kind of data we don't have. If we flail with a flail mill, is the glass
a different particle size and if that is, does that simplify or complicate
a glass recovery scheme? I hear the concern from the pragmatists that we
don't get too theoretical, but I think one thing we could all agree on in
this business to date is we haven't erred on the side of being too theo-
retical.
BURCKLE: Tom, yes, I agree with what you said, but I think maybe I don't
quite understand one part of it. It seems to me that the performance of a
given unit operation is going to be somewhat subject to the feed material
that's coming into it.
LAMB: It is.
BURCKLE: What I'm driving at here is, I don't see how we can get over the
process part, for example, where we pre-trommel before we primary shred. I
think we have to look at complete sub-systems, if you will, or the process
type research in addition to the unit operations.
LAMB: What do you mean by a sub-system?
BURCKLE: Take for example sub-systems, I would say that one sub-system
might be to look at pre-trommeling shredders and some sort of size classi-
fication system on the outlet of the shredder with a recycle. That might
be a complete sub-system.
LAMB: You don't have to test that all together, I might argue, you could
test that by piece and pretty well predict closely enough for your needs
what's going to happen when you tie the components together.
155
-------
BURCKLE: Only if you can control your inlet feed. What I'm saying is, I
suspect the performance of a primary shredder and a classification system
might be considerably different than if you pre-trommel, primary shred, and
classify.
LAMB: I suspect it will be considerably different.
BURCKLE: I don't think that if we test just trommeling separately from the
shredding that you are going to be able to marry the data.
LAMB: Well, my suggestion would be, for example, if you designed your re-
search program properly, when you are evaluating the shredder you would
hypothesize first. You might want the trommel to hypothesize, then the
range of conditions you want to test for the shredder. Maybe we are argu-
ing about the technique, that you do it all at once or by pieces. I guess
I don't care, but it does seem to me there's a lack of basic design data.
Evaluating equipment, we're all doing that now. That's not going to benefit
any one of us who's out running a plant for the next 5 years because you're
not going to have any data for it.
CAPPS: I'm working with Steve Hurley in the Navy program trying to sort out
R&D problems that we can tackle. I share Harry Freeman's view. I haven't
heard any technical obstacles that would be approachable in any system that
I have thought about to date. What we have come up with as a conclusion is
that there is so much uncertainty in this whole system that we have to set
up a field laboratory, a field test evaluation, so that we can define the
R&D problems so that we can then get at them. I think this is somewhat the
approach that NCRR is taking, and I would hope that we are paralleling or
at least complementing, what NCRR is attempting to do. I think that John
Burckle's views of the systems approach is absolutely true. Our view, and
I think that we are evidencing it right here in this meeting, we are having
great difficulty identifying those technical obstacles. The strategies that
we hope to do is to identify them in the laboratory of the field proving
ground, if you wish.
SPENCER: I think your problem is a little different, because you have the
opportunity to do a certain amount of hand sorting, so that you can probably
work very effectively at a smaller scale. I think with the large scale,
other people are doing design and systems supply. I want to get back to one
of the things Roger De Cesare said, and I know it's a problem Harvey Funk
and I talked about before, and Tom Lamb is saying the same thing. We don't
have basic design information. Frankly, my own opinion is: You know I'd
like to see a plant that had in the front of the line had a large vertical
shredder, and next to that had a horizontal hammermill, and next to that
one of these slow speed grinders like they have in Europe with different
conveyors that can feed the four different types of air classifiers, one
being a rotary drum, one being a variable zig-zag, one being something dif-
ferent, and then having behind that, let's say, trommels, and have enough
space and enough flexibility that you can then interchange things, try a
156
-------
coarse shred material with various size grates, and feed that into the vari-
ous air classifiers, and develop a handbook like Taggart that says, when you
do this you can anticipate this kind of a result. You don't know it for a
certainty, but when we did it this way this is what happened, and you have
that kind of data and you have the operating data. There's no basis for any-
body to do a really rigorous design. Even a bin prefers cans. You don't
even know exactly what the angle should be, or you don't know how you dis-
charge from the bin, you don't know what the gate should be, and you're not
sure whether your're going to have a bridging problem. Those kind of basic
fundamental things don't exist. What's worse is we don't even—here we are
5 years down the pike and we're still not even starting to do something
positive to get that basic information. It's true, we can look around at
various systems and say (and Taggart does this also) well, so and so's sys-
tem, when they did this they got this result, and in so and so's system,
when they did this they got that result. That doesn't always help you,
though, as much as the more basic information that says when you do these
kinds of things using this piece of equipment, this is what you'll get.
You can expect to use so many balls per hundred tons of material; you can
expect to use this amount of electricity; and when you shred and you are
shredding to this size, you can expect to spend so much material on retip-
ping your hammers; you can expect to use so much power per ton shredded,
when you are shredding to this size range; when you change the size range
you can expect the following. You need that information to do a sensible
design, and we don't even have the mechanism by which to obtain the in-
formation. What's worse is we're not even starting down the path yet to
develop that mechanism, that research tool, that will allow you to do it.
The thing that scares me worse than anything else is that the total budgets
that I see being made available to resource recovery are not sufficient to
ever obtain that information. We'll look at that one, we'll look at that
one, and we'll try to steal little bits and pieces. When you're all done
you haven't really evaluated the thing. Evaluating composition doesn't
help me one bit. A lot of the things that you are doing don't help us de-
sign systems and I don't think they help the A&Es design systems either.
They are nice information but in terms of doing real things at fair scale,
it doesn't help us in terms of what we have to do when we have to buy some-
thing now, and it's going to be at least 5 years to get this information
anyhow, at least, and it might even be a little bit longer. You've got to
start somewhere. The first place to start is with a realistic expectation
of huv much it's going to cost to do something that's worthwhile.
LINGLK: I think what I hear from the various comments in the room regard-
ing how we collect data, what type of data to collect, at what level of
detail, and on what size facility is that at this point in time there is
so little information available, that we really need a whole range of dif-
ferent types of data collection. The different mechanisms for collecting
the data and the different kinds of analysis that you can do are going to
be useful to different people to varying degrees. Maybe there are three
or four steps. One step might be simply to gather data on the overall per-
formance of total plants or total systems, new systems coming on-line.
157
-------
That sort of data is probably particularly of interest to the municipal
sector. They want to know are these plants working, what do they cost,
what is their overall reliability, what sort of emissions do they produce,
do they meet product specs. This sort of information is needed. It may be
of some use to design engineers, but probably more limited use than some
other types of information. As a lot of people have said, there is not
much flexibility for really varying parameters to get individual performance
data on unit operations. Perhaps another level of data would be to investi-
gate particular unit operations at various installations. For example, you
could pick out ten shredder operations in the country and collect perform-
ance data. That's at a slightly more detailed level, and it starts to be-
come more interesting to the design engineer. A third level of data is to
set up some test procedures, whether it's in a pilot plant or whatever, and
get the kind of data that George Trezek has gotten on unit operations—
really develop some basic data. That's certainly of interest to the design
engineer. There may be a fourth step, and I think Harry Freeman was al-
luding to this, which is that maybe we also need to forget about the exist-
ing equipment to an extent, or say there's some need to just develop some
new items of equipment. I think we are trying to find an answer, and there
are really a lot of answers. I think we need all of this kind of data; we
need to develop programs to get it; we have to realize that different data
is useful to different people.
WALTER: I'm glad that Steve knows what he's hearing because I'm not quite
sure I know what I'm hearing. I hear Harry Freeman take off on one of my
pet subjects, which is the thing that Steve ended up with, that really
shouldn't we step back and take a look at this very heterogeneous mass that
we are trying to process, and apply some thought to what the various things
we want to do to it are, and then try to develop equipment specific to that
mess that we are trying to play with. Particularly I think that we should
(at least I remind myself when I read some of the original Bureau of Mines
reports) remember they very candidly said and I think very properly did,
took existing equipment and tried to modify it to a new feed source. May-
be we have to look at the feed source and build the equipment to fit the
feed source. Not saying that we should not continue where we are, that is
continue to look at the equipment that's now being used.
ALTER: I'd like to postpone this discussion just a moment and try to get
to it through a different route. Also, I want to alert everybody to keep
to the research agenda and not a wish list. There is a very important dif-
ference. Let me suggest that we try to approach this a little differently.
I'd like to try an exercise. Turn to pages 13 and 14 in the MRI book [see
pages 158 and 159]. MRI has outlined the research needs by category. I'd
like to suggest an exercise. Number the general categories (shredders,
magnetic shredders) Roman numerals I, II, III, and so forth all the way
through Economics on page 14. Then, I'm going to ask that you choose five,
your first, second, third, fourth, and fifth choices of research needs, and
write them down on a piece of paper. Put them in order that your first
choice is what you regard as most important. Let's delineate what the
158
-------
items on pages 13 and 14 might be the five highest priorities. But let
me also tell you where I think I'd like to see this discussion lead to.
I want to return to what Roger said in terms of the equivalent of a Taggart,
Taggart's Handbook of Mineral Dressing. It's a thick book and a lot of it
is still up to date,even though the first edition was 40 years ago or so.
It's a type of handbook like a Perry or a Marks. It's the equivalent of a
design handbook, a where we're going or what we might expect handbook.
What are the most important things to go into the equivalent handbook of
resource recovery? Of the 40 titles [pp. 158-159] where do you think we
should start? Put down your fist five choices of what should be attacked
first, then put down your first five choices of the things that shouldn't
be looked at, what do you not like?
LISTING OF RESEARCH NEEDS, BY CATEGORY
I. General
1. Determine Optimal Arrangement(s) of Unit Equipment.
2. Study Emissions from Processing Equipment.
3. Determine Effects of MSW Characteristics on Processing.
4. Evaluate Other Potentially Applicable Equipment.
II. Shredders
5. Compare the Performance of Various Types of Shredders.
6. Determine Maintenance Requirements of MSW Shredders.
7. Determine Optimal Hammer Design(s) for MSW Hammermills.
8. Establish Shred Size Requirements.
9. Evaluate the Cost Effectiveness of Single Versus Multiple
Shredders.
III. Magnetic Separators
10. Determine Effect of Magnetic Separation.
IV. Air Classifiers
11. Determine the Optimal Operating Conditions for Air Classifiers.
12. Compare the Performance of Different Types of Air Classifiers.
V. Screens
13. Compare the Performance of Coarse Screens.
14. Compare the Performance of Fine Screens.
15. Determine the Effects of Pretrommeling.
VI. Dryers
16. Determine the Effect of Drying on Bacteria and Virus.
17. Evaluate the Effect of Drying on the Efficiency of Air
Classification.
18. Evaluate the Effect of Drying on the Storability of RDF.
159
-------
19. Evaluate the Effect of Drying on the Combustion Character-
istics of RDF.
VII. Densities
20. Determine the Optimal Operating Conditions for Densifiers.
21. Determine the Specifications for DensLlied RDF.
VIII, Conveyors
22. Determine the Optimal Operating Conditions for Vibrating and
Pneumatic Conveyors.
23. Compare the Performance of Conveyors.
24. Compare Air Transport Systems.
IX. Storage and Retrieval
2.5. Compare the Performance of Various Storage and Retrieval
Systems.
X. Receiving Facilities
26. Evaluate Receiving Facilities.
27. Evaluate MSW Segregation Prior to Processing.
XI. Controls
28, Characterize Present Control Systems.
29. Determine the Effectiveness of Present Control Systems.
XII. Fire Protection and Explosions
30. Study the Incidence of Fires in Solid Waste Plants.
31. Determine Fire Resistance Characteristics of MSW Processing
Equipment.
32. Determine the Effectiveness of Fire Protection Systems in
MSW Plants.
33. Study the Incidence of Explosions in MSW Plants.
34. Determine the Effectiveness of Explosion Protection Systems
in MSW Plants.
35. Evaluate Spillage and Dust Controls.
XIII. Economics
36. Develop Effective Accounting Method(s).
37. Establish Capital Costs of Processing Equipment.
38. Determine Economic Life of Processing Equipment.
39, Determine Equipment Operating and Maintenance Costs.
40. Perform Cost Effectiveness Analysis.
Source: "Study of Preprocessing Equipment for Waste-to-Energy Systems:
Summary Material and Research Needs," Midwest Research Institute
for Environmental Protection Agency, Municipal Environmental Re-
Search Laboratory, Cincinnati, Ohio, February 1977.
160
-------
H. FREEMAN: To exercise my prerogative as a member of this group, I would
like to have heard what Don had to say. Maybe it was because he sounded
like he was getting ready to agree with me [unintelligible]. The second
thing is, I think in just the exercise you are going through here, it's
kind of the same thing. I don't see a big category called "Other" or "New
Ideas" or "New Studies" or whatever, so I feel you are restricting us to
exactly what he was saying at the beginning.
ALTER: You can do it with a third category.
H. FREEMAN: I would have liked to hear him make the point.
ALTER: I agree with you. You jumped me completely, because that was my
next step to stimulate the study. Has everybody got these 10 down? I sug-
gest 1 through 5, which you think are high priority, the highest 5 priority
in your judgment. Then 1 through 5 of which you think are the lowest pri-
ority, where 1 would be the "pits."
WALTER: These are in terms of the general categories, of general shredders,
magnetic separation.
ALTER: If you like, but I would hope that maybe you can do it of the 40
categories.
BERG: You are talking 40, instead of 13 now.
ALTER: Or if you feel like you can only do 13 in this time period. I think
that the EPA staff can analyze the consensus for that. John Burckle has as-
sured me that he can. All right, now to get to the third category. Don?
WALTER: The third category contained number 1 on my first category, which
was number 41 [unintelligible].
ALTER: My third category is as Harry was talking about. What's not on the
list? Now let's do this in terms of a need and not a wish list. What
might MRI have left out? What's your pet peeve?
SJOBERG: This group has been looking at this basically as a plant. Every-
body who is involved in this business has to approach it as a business unit
which is from the time you receive the material, the transportation at the
transfer station, transportation, the plant, the economics of the plant, and
what you do with the products that come out the other end. We have not
found it very difficult to assess what plant operations and the equipment
and the costs should be. I find the biggest problem we have found to be
able to assess is when you go throughout the country to different prospec-
tive projects, is what and how is the power company going to be able to
utilize this material? Because in order to make this industry viable, the
economics of it is such that you have to have, as we have in our case, a
[unintelligible] and they have a dearth of information that they are lack-
ing. I think if you really want to apply where the priorities are going to
be for research, to make this industry viable, you have to put it there.
161
-------
ALTER: That's the next work shop: Talking about mechanical processing up
to that interface. John tells me the next workshop will talk about that
need. You have three categories now, and as soon as everybody is finished
we'll talk about how to get the information; how to translate this con-
sensus into an action plan. Now one other requirement which would be help-
ful. If you don't wish to sign these things (I hope you will), please indi-
cate your viewpoint. Are you municipal, A&E systems designer, interested
observer, however you'd like to characterize yourself? I'd like to talk
about one aspect of how these data may be taken. We've talked about design;
we used the term design handbook yesterday. The subject of Taggart's Hand-
book of Mineral Dressing came up today. In either case we're talking about
a design book for resource recovery. Of course in the next fow years it's
going to be thin and loose-leafed, but a book that we could use for Oi
-------
organization is a subcommittee on shredders. Another on air classifiers,
and so forth.
OLEXSEY: I think that's an excellent idea, because one of the things that
was brought up very much in the course of the meeting was a lot of people
have proprietary type information and are not willing to share it and you
can understand that. It seems like a committee to touch across various
companies and/or industries and somebody may have some industry funding like
EPRI or NCRR, outfits like that, so that in areas where the government hasn't
been able to provide information, where these companies can get together and
even fund, because we are all looking at the same machines, the same pro-
cesses, and we're duplicating efforts internally.
ALTER: Any other comments on this concept?
HASSELRIIS: I think it's a very sound idea to set up this kind of communi-
cation whereby, while my information may be proprietary internally, I know
perfectly well what I can put into a framework that's been agreed to by
everybody. I can see why, for instance, in the case of a trommel system,
there's the possibility of our feeding that type of information, everybody
feeding it, and the whole thing making some sense. It's the type of ma-
chinery where I think that it's possible to do that. The only difficulty is
that, again, we are confounded with a total different nature of the ma-
terial, and we may not have the money to analyze the material on the basis
of which the data was produced. That would mean that whoever had the money
for that type of information would add it, and then we maybe would be able
to guess at what our analysis was, close enough. It's good enough for our
purposes, so we can trade something useful to other people at no loss to us
and at the same time we'll get back a return for our efforts to try to under-
stand it. In other words, it's not strictly a one-way mirror. It's a valu-
able exchange of information, but it requires an external matrix that gives
meaning to our information and gives meaning to what we get back. That's
what we're trading.
ALTER: That is something which has to be raised in these subcommittees,
which I'll return to in a moment.
FISCUS: One thing I wanted to mention—a fairly common sense sort of thing,
but I did want to bring it out—if you are going to develop standard methods
of investigation, I think there's two points you have to consider. One is
that the methods have to be fairly specific. If you keep them too general,
then there is enough variability in a general method that it doesn't get
equally applied in all cases. The other danger is that you have to be sure
that these methods are minimum recommendations or base line recommendations,
leaving room for investigations of unusual occurrences. If you don't do
that and make sure everyone understands that, then maybe you're evaluating
refuse processing plants and everybody says o.k., I've done what the com-
mittee recommended, and that's all that we have to do. That does not neces-
sarily tell the whole story. It's the design philosophy, I guess, is what
I'm trying to communicate here, of promulgating recommendations for
investigation.
163
-------
ALTER: We suggest to break into such committees. They might be organized
into the following categories: I think we may need only three or so such
ad hoc committees, maybe more. One on size reduction, one on separation,
and perhaps we divide the separation committee further into air separators,
mechanical separators, i.e., screens, etc. Others might be nonferrous
separators or magnetic separators or glass separators. Bill, maybe it
would be good if you could write that on the chalk board. [See Table 1].
Table 1
PROPOSED COMMITTEES
I II III IV
Size Reduction Separators Materials Handling Drying
(Densification)
1) Air
2) Screen
3) Other
RIGO: Why is size reduction singular?
ALTER: It's generic.
RIGO: Separation's generic.
ALTER: Suggestions for other ways of organization. Materials handling has
been suggested. We heard densification, what else? Drying has been put
down.
LINGLE: How about process control?
BURCKLE: Process control is what you do, once you've got these answers.
ALTER: We're talking now of methods leading to evaluation. Storage is in
the materials handling category, and Dave Bendersky has included it that
way in the MRI list. We are talking about methods of evaluation. There's
room for distinctions here, I want to make certain they are distinctions
with a difference. There are methods of fire protection and explosion
protection; perhaps they haven't been applied; perhaps we don't know how to
apply them. Now if we learn how to apply them, should they be evaluated the
same way? I'm referring to means of evaluation and measurements of per-
formance, not toward invention.
SJOBERG: You're going to have to itemize what techniques are available.
One of the most critical parts you have relative to these plant operations
is not to lose it.
RIGO: Harvey, we just finished up something from the Navy. We had to take
garbage processing systems all the way through materials recovery, and break
164
-------
them out into nine modules: Groups of things that did roughly the same
thing. Maybe that would be a good way of helping the breakdown. The
module systems that we came up with were receiving, size reduction, com-
bustible separation (you might argue with the word combustibles). From
there we went down into what we called an RDF preparation step (upgrading
of some form) densification, and storage. We also looked then at magnetic
separation, non-magnetic separation, and glass. Now,this gives us nine
groups of things, each of which are going to have characteristic evaluation
requirements and procedures in reporting requirements. Maybe that's better
than trying to go to this type of breakout. I know it's more detailed.
ALTER: I have to tell you that I fear too many. If we are going to break
into committees, the people who will want to attend and participate will
have too large a drain on their time. If there were, say, one committee
for materials handling, they in themselves could give priority order.
Whereas if there were one committee on materials handling and one on stor-
age, etc., these committees would involve all the same people interested in
the general category of materials handling. Where does densification fit
in? Let's make an arbitrary decision: in size reduction?
BERG: There's one whole set of things I think you are missing, and that's
your process controls, including the emission control questions and what
have you there.
ALTER: No, as I said, I think these lists refer to methods of evaluation,
not existing methods of process control or measurement, so do not include
existing methods of emission control. We're talking only about methods of
evaluation, what measurements are to be taken. For example, in emission
control, I think the measurements which have to be taken are clear, be-
cause of law and regulation.
BENDERSKY: Harvey, I know you didn't want to get into this economics, but
I would suggest that before too long consideration be given to the combi-
nation of cost along with technical performance. I think the evaluation of
equipment performance always comes down eventually to dollars per ton.
ALTER: I think that's part of the matrix.
BENDERSKY: The only thing I was going to suggest was the possibility of an
economics group.
ALTER: All right, what does the group feel? Do we get into uniform cost
accounting procedures; that would be the method of evaluation of economics?
BENDERSKY: That's just one of many things that need to be done.
LALKA: I think the A&E people have different interests than the municipal-
ities. I think if I were an A&E, I'd be interested in getting all the
design information I could on a piece of equipment, but from the municipal
165
-------
standpoint, the thing that would interest me most would be the cost, the
O&M. Once the plant has been built, and it's turned over to me, then I
have to start worrying about dollars per ton. So there's quite a dif-
ference between the two interests here. I would be concerned about the
environmental considerations once I have the plant, plus all the costs
and power consumption, etc.
WALTER: I would submit to you that economics is a matter within the cate-
gory, because if you can sell something for a buck a ton whole and it costs
you $2 a ton to make it small, and you only beneficiate by a dollar a ton
when you make it small, you don't want to make it small. So you really
ought to look at economics in terms of size reduction. Similarly you should
look at economics in terms of separation. Do you get anything back for your
separation? Your economics relates to each one of those phases and should
be looked at that way, so you make discreet judgment as to whether you should
even be doing this stage or not.
BENDERSKY: I accept that if you will include economics within these groups.
WALTER: Sure, they're essential in the end.
BENDERSKY: Not just technical.
WILES: I want to know where we are headed as far as these subcommittees are
concerned and this workshop. After this are we going to have another work-
shop that approaches economics and the institutional problems? If we do, I
think we are going to have to have a representation that crosses the two
areas, because restating what Harry said, I've heard no real technical
inhibition against resource recovery. What we find is that there's a combi-
nation of technical, social, economic inhibitions that all lump together
that caused one resource recovery facility not to work, another one not to
work. We are concentrating here on the technical aspects, and we are just
forgetting or we are ignoring those things. If we go into these subcommittees
and don't have some tie-ins to these other factors, I think we are beating a
dead horse.
ALTER: Let's return to that in a moment. It's on my agenda.
LINGLE: I can throw in a vote for keeping the economic accounting issues
separate. Rather than have the unit process committees develop methods of
economic evaluation or accounting formats four or five different times, I
think we ought to develop it once and try to follow those procedures. You
might have to adapt it a little bit to different specific unit operations.
I would throw in a vote for doing it all at once. With regard to what
Carlton said, is the purpose of these groups to develop methods of evalu-
ation?
ALTER: I would like to cover one thing first, and then return to that
issue.
166
-------
KEYES: The one thing that I can see against having the economics as a sep-
arate subcommittee is the fact that I don't think you can divorce the two as
far as economic and technical problems. I would think that the people in
that economics section, if you had one, would be re-doing a lot of the work
that the other people are doing in order to find out exactly what type of
information they wanted to gather. So it might be a duplication of effort
on that part, because they are dependent on each other, they are not inde-
pendent.
ALTER: There appears to be a consensus not to separate out the economics;
correct me if I am wrong.
ANONYMOUS: You're wrong. I think there were two votes for and two against.
FUNK: I'd suggest this, I have under my "Other" list here to establish
standards for data taking, and I've also listed as a priority Item Number
36—Develop Effective Accounting method. I think we should have a standard
group that could start establishing these standards both for economics and
for taking data from any performances that might be taken.
BURCKLE: I see the need both for the column and the row in the economics
area. Number one, your economics framework is not going to be what I would
call a measurement method. It tells you what you need in terms of data to
perform an economic analysis. But in the row, you've developed the method
for examining the economic aspects of each of the four areas we listed, and
that feeds in to the column, if you will, of the overall economic analysis.
I think both points of view are correct, if you are willing to look at it
from that direction. In other words, we need a standard way of doing eco-
nomic analysis, if you will, of the overall system of the plant, but we
also need the measurement methodology under each of the types of unit oper-
ations we'd be dealing with.
RIGO: Harvey, it's not clear to me that we are yet in a position to talk
about developing standard methods. I've got a feeling that that's pro-
bably a relatively straightforward exercise. One person can draft and get
two or three people to agree that on trommel, we want to evaluate it in
this manner. What we have not yet established, and what I think is the key
to everything here, as Steve Lingle pointed out, we've got different data
needs by different people. What intensity of data must we get, and do we
need to collect data three times, can we do it once? The basic problem of
what data on a particular piece of equipment you get is trivial. The gen-
eral problem of what intensity of data do you get I think is critical, and
is the key. That's not being addressed at all.
ALTER: That would differ with each unit operation in part, because only
through experience can statements of precision and accuracy be developed.
I would suggest that intensity of data is something that each subcommittee
has to consider and make a recommendation. I think it would be wrong to
try to solve those problems now; recognize them yes, as you did, but not
try to solve them in this group at this time.
167
-------
RIGO: I agree with you wholeheartedly that we cannot establish precision
and accuracy. My question is one of what type of data will we get. Do we,
for example, go after something that tells us the unit operations gives a
70 percent split, or do we recognize that if we are going to get a meaning-
ful answer, we have to know what is there, and we have to have some defi-
nition of do we measure the total pounds in the stream? If we are measuring
power, do we want to just get total kilowatts, or do we want to get the de-
mand power, or surges? This is a top level decision that has to be made
before you can even begin to do that.
ALTER: There is a difference here, I think, between groups evolving the
matrix of what needs to be done and the individual judgment of the investi-
gator, within his constraints of budget for example, as to how much data is
taken. The subcommittees might seek to develop what we call in ASTM a
recommended practice. If you state that you followed the recommended
practice (and there is a movement to even change the name to a standard
practice at ASTM), then you followed that exact format. What I am suggest-
ing here is to start out, because this entire subject is in its infancy,
with the matrix and hopefully this will lead to the equivalent of the
standard practice. Certainly, within our own constraints of budgets and
objectives, we are going to be taking data somewhat differently until there
is a standard practice to follow. Let me ask now that you put one other
thing down on your sheet of paper before we collect it. That is, given your
"druthers" and permission of management, which of these four groups, if they
met, would you like to participate in?
FUNK: Before you jump into that, let me offer one other suggestion. I've
recognized one thing here that I'd like to point out. I believe the biggest
problem has been in these two days that we are trying to be all things to
all people. We've got too many special interest groups involved, and you
can hear it coming from every direction. I recommend, just suggest this for
a minute, why don't we break into groups, these special interest groups,
let's get the A&Es in one group; the operators, and in that category there
will be the small ones, the people like the Navy pointed out that are looking
at small quantities, and the large ones. You've got the L.A. Sanitary
District, Chicago, and that type of thing. We have system suppliers, that
would be the CEAs; and the researchers, that would be the professors and
people who are involved in research and development type things, and would
include the MRI group and that type of thing. With that kind of breakdown,
I think we could come back with what we feel the research needs are. Some
program could be put together out of that, and then a workshop with a se-
lected group of volunteers could endorse it or whatever and then the EPA or
whoever wants to carry this forward would have something that has general
consensus.
ALTER: That's why we asked you to put down on your sheet of paper either
your name or your category, because I think your background does influence
your choices. Now I have to tell you that I don't know how EPA will choose
to use these sheets, and do not presume to tell them, but I think they will
168
-------
be valuable to evolving an input to their thinking. If you will first put
down in which group you might like to serve then I'd like to turn the dis-
cussion into where this exercise might lead.
FUNK: I don't want to get into one of those groups, because I am interested
in all of them. I want to get into a group with the rest of the A&Es and
similarly related people so that we can talk about what we want out of this
whole thing and what's necessary to us to perform our function. I want the
Arnold Chantlands to get together and say from the guys, just like the L.A.
pointed out, the designers were there for so long, and then they leave and
then they are stuck with operating it. But we want to hear that from their
standpoint.
ALTER: Let's separate two things, please, on your sheet. There is one, the
needs as you perceive them, and there is two, of these four groups, as a
method of accumulating data so different sources would be comparable. Now
data on what? If, for example, the A&Es say size reduction is the highest
priority, and it turns out that that's going to be investigated, then the
ad hoc group might recommend data on size reduction ought to be accumulated
in a particular way. There are two separate objectives here. I thought
this was all discussed. The groups we are now talking about—unless there
is a consensus or decision to accumulate the data, then the matrix would
not be used. There are two things on your sheet. One is the ranking of the
research needs and suggestions for others. The second is an expressed in-
terest in working towards the method by which the data should be evaluated.
Now on the latter, I return to the idea that even those of you who are in-
volved in proprietary work, I do not think it excludes you from saying how
the data might be accumulated, so that the data when published, even if not
your own, would be more valuable to you.
BENDERSKY: Let me try a compromise. I want to suggest that if you maintain
this technological list this way, it would seem to me to be well to have at
least one representative from each of these major groups in each one of
these so that these groups can be represented. I understand the need for
groupings.
ALTER: Is this little conversation one of the things coming out, as David
points out, is that in each of these groups on how the data might be accu-
mulated, there indeed be representation of the various viewpoints. In
other words, it would be a mistake to not have an A&E in the groups; it
would be a mistake not to have a municipal man or systems person; and su
forth.
BURCKLE: I think I see what Harvey's driving at now, talking really about
two completely unrelated matters from the aspect of planning. One is de-
veloping a commonality in how we take measurements to describe system and
economic performance as opposed to determining what things we would end up
concentrating our research efforts on as Harvey Funk has suggested. We are
not going to, of course, be able to accomplish that within the framework of
this meeting at this time. I think that the methodology for making measure-
169
-------
merits is being attacked by the ASTM; I know that we in EPA are supporting
it. Perhaps that's really a more appropriate place to attack that sort of
need. Hopefully the ASTM Committees would consider structuring themselves
to include the various interests.
ALTER: As we do.
BURCKLE: Harvey Alter says they do now. Perhaps if there are those of you
here that would be interested in working on those committees, this could be
done, Harvey.
ALTER: We are getting ahead as to what to do now with all of these pieces
of paper. One, if you are finished, I think they ought to be collected now
so they don't get lost. I'm making an open suggestion now to EPA. I think
a next step would be for ad hoc committees to meet to work toward evaluation
methods and to reconvene this workshop to continue the discussion at some
future date. To do so, the ASTM Resource Recovery Committee is a possible
mechanism. The work of the ASTM subcommittees right now (for those of you
who are not active and I wish you would be) has not been for methods of
evaluating equipment; it has been for specifications for recovered materials
and methods of tests for those materials. Almost all of the technical sub-
committees have taken this upon themselves as a first priority job. I think
an evaluation matrix as we're talking about now would be a leg up; you could
move ahead in the field sooner. I would make the open suggestion to EPA
that based on the pages that have been turned in, to decide upon and even
call to order these subcommittees and invite people to participate. I would
also suggest for this body to buttonhole our EPA hosts and give them your
views as to when it should happen and where it might be held. I'll say right
now that if there's any hesitation as to where to hold the meeting, I'll be
happy to host the subcommittee meetings in our office or possibly elsewhere,
provide the secretarial and typing services to get it going. I don't think
that the matrix subcommittees need meet perhaps more than once. Now I'd
like to exercise the prerogative of the chairman and return to a subject
which Carl and Steve Lingle really raised, and that's the entire issue of
are there technical obstacles or are they intertwined with the social,
institutional, marketing, economic, etc. To do it, can I challenge you to
say, we have a technical obstacle, maybe not well defined or articulated
yet, which is preventing us today from moving ahead faster. In other words,
if we could remove that obstacle, would there then be a sudden increase in
the number of plants contracted for. By inference, if there are not, then
as Carlton points out, we have to devote attention to the other aspects. So
I challenge you: Is there a technical obstacle in our path to moving ahead
further. Would the A&Es and systems designers be able to "sell more plants"?
Would the municipalities be able to accept plants sooner, if this obstacle
were not there?
CHANTLAND: There are many obstacles. I just point out one that I've heard
others mention. You must look at the end product. I would say that you
have to start at the end and work backwards. We've got to have somebody to
170
-------
receive this material, somebody who's willing to pay money for the product.
It was mentioned here that there are several boilers in this particular area
that they can't sell the material to simply because they were designed for
gas or oil, or something like this. Retrofitting these boilers would cer-
tainly open up a new area of interest to a lot more buyers. I'm sure that's
perhaps true of many other things such as metals. I think this is the point
that is very important, if we can't make some sales, our costs are going to
be rather prohibitive.
CAMBOURELIS: Those are fundamentals, and I agree wholeheartedly. We're ig-
noring that, all of us.
CAPPS: I would like to make one comment concerning what's happening in Cal-
ifornia at some of the meetings I've been attending out there. The concern
over air pollution is driving many people to try to guarantee meeting air
control requirements. It's driving people to look for pyrolysis, and yet as
far as I'm able to find out, there is no pyrolysis system that's ready for
anyone to accept at the moment. So that's a technical obstacle related to
a concern over a political problem in San Diego, the Bay area, and so forth.
TREZEK: I think the California problem stems more from the fact that boil-
ers are not equipped with ash handling equipment, and therefore no boiler in
California can really take RDF. So that's the problem.
RIGO: Harvey, it looks also that if we shift our focus from utilities back
down to industrial scale, and if we can come up with some reasonable require-
ments for emission control, at least in the Mid West I'm relatively certain
we are going to see a major shift from gas-fired and oil-fired industrial
and small commercial scale units back over to coal and solid fuels. So if
we shift our focus from the utility to the industrial scale, and we look at
new construction as opposed to retrofit, there's going to be a lot of it.
We suddenly open up a whole new market, one where we are going into boilers
we can tailor to what we want, and one where we are going to be able to
charge two or three times the fuel prices we can get from the utility. In
Dayton, for example, Dayton Power and Light is willing to pay 23 cents a
mega-Btu for RDF. That's all they are paying for coal. If we go to the
industrial users, Delco-Moraine, for example, right now is buying emergency
propane well in excess of $5 a mega-Btu, and they can afford to do almost
anything to get off of that.
LINGLE: A quick point in response to the issue of the smaller-scale indus-
trial boilers. I guess this also partially answers the question of the ex-
tent which technical issues are compounded with other institutional type
issues. Clearly, the industrial boiler market is a very interesting one for
the reasons that have been presented. We took a look in a gross national
sense at the magnitude of this market, and found that there are a lot of
industrial boilers out there, but total national capacity is about half that
of the utility boilers. That's not the concern, the concern is that the
average capacity of industrial boilers is sufficient to handle about 30 tons'
171
-------
per day of RDF. In fact, we found that nationally there are something like
50 industrial boilers that could handle the output from a 500 ton per day
plant, and that's assuming 40 percent RDF firing and 100 percent utiliza-
tion. Now clearly, some people have multiple boilers and so they are going
to fire the waste in two or three boilers. Also, you can market the multi-
ple users. But, that raises the problem of financing the system, because
now you've got to have contracts for the series of different industrial
users. The financial community may not view this whole series of contracts
as nearly as stable in terms of projected revenues as they would view the
revenues from a single utility that they knew would be there for years. I'd
be interested in any comments on that from the standpoint of the A&Es who
are maybe looking at this market.
HOLLANDER: I'd like to make a comment, if I may. Certainly an industrial
market would be an attractive one. They don't have the luxury to buy fuels
that the utilities do. Consequently, they are in a somewhat weakened po-
sition. They are also the first one that's cut off when we have a fuel
scarcity in one form or another. Historically the industrials do not con-
tract for fuel for 5 or 10 years; in fact, most industrials rarely contract
for fuel even for a year. They buy spot fuel ordinarily; they may have a
one year contract, but we have to rely on having long-term contracts from
industrials and the probability that this will ever come to pass is ex-
tremely remote. Strange as it may seem, since the energy crisis—what was
it, 1973, the oil situation—the number of coal-fired industrial boilers
that have been contracted for are invisible. There has been virtually no
activity in new construction on coal-fired boilers. It's been very, very
spotty around the country. The ABMA, the equipment manufacturers are abso-
lutely flabbergasted; they expected to be inundated with orders. Right now
they are scratching around for an order for an industrial type coal-fired
unit. Now, when will it come to pass—you should be thinking of two-three-
four years from now—it's not going to happen overnight. With the new FEA
approach on MFBI, and I'm sure you're all familiar with that, they are going
to try to at least control what kind of fuel is going to be burned in units
requiring 100,000,000 Btu per hour. The industrials are going to have to
justify why it's impossible for him to do anything but burn coal. It's
going to take a long time before this will have any real effect. If a
community decides they want to approach the industrial market, they are go-
ing to have to find more than one user, probably half a dozen users, and
hope that they have such a desirable product that they will all vie for it.
That happens to be why it will be difficult to get firm contracts to justify
the construction of such a facility. The big attraction of the utilities is
you think you've got them or can get them by the short hair and sign them up
for 15 years, and this justifies building the plant.
FAY: I think the big attraction of a utility is people on a refuse proces-
sing process think they can minimize their own capital investment and throw
the burden at us. No, I think that's a very key issue, because if you guys
want to have a product that you get more money for, a real easy one to pro-
duce,, theoretically, is oil. You can sell that any place, but you—the
cities that is>—are going to have tc. put up the front end money for a
172
-------
pyrolysis or whatever it is, which I don't think is quite here today, but
maybe it will be shortly. It's probably going to cost more, hopefully you'll
sell your product for a higher value when you've got to trade off those
things. If Milwaukee is any example, I think it is, the cities, no matter
how highly motivated, are looking for somebody else to pick up the bucks.
In this field, it seems to me, what held it back is that they seem to be al-
ways wanting to get some gimmick that the Federal government will throw into
it for them. The real easy answer is turn out a product that you can sell
more than just to the guy with a big boiler sitting there.
WISELY: In our own experience along those lines, the only industries which
appear to be a really good market for this type of material are those which
have a rather constant year-round plant factor. Most industries are somewhat
seasonal in their steam demand. Other factor that enters into this is that
an industry has got to have a financial incentive, usually, in order to use
a fuel of this kind. That seems that in our own experience in a very re-
cent project in which all the agreements between the municipality and the
industry were already executed, the net (I wouldn't say profit) but the net
project cost, which happens to be a negative all the way through, will be
shared by the industry and the municipality. It's kind of funny here in
this case where the industry because of the tax situation, feels that it
really has to be a three-way split, one to the community, one to the indus-
try, and one to Uncle Sam. All right, now, if you think in terms of the
utility, some utilities of course are taking a rather different view on
what they are willing to pay for fuel of this kind than others. Some of
them are taking the attitude that sure, we'll do this as a good neighbor
as long as this does not cost our rate payers, as long as it is not adverse
to our stockholders, we'll pay what the comparable value would be in dollars
per million Btus. Other industries take the attitude that well, they have
to have a little more incentive than that. So I think you have to recognize
that there are other incentives though, which certainly the industries and
the utilities bear in mind and one of them, of course, is the matter of be-
ing a good neighbor. Others are the national image they have in the envir-
onmental, resource recovery field and that sort of thing. Although it's
quite possible that you can't really allocate a dollar sign to that value,
it is there, and I know darn well they recognize it.
CAMBOURELIS: I sympathize with the utilities and the pressure they are sub-
jected to (to use RDF). I don't think they should be expected to function
as eleemosynary institutions. They are there to make a profit like most
private companies, like my own is. On the other hand, I think they occupy
a unique position as regards to being potential refuse-derived fuel buyers
almost everywhere. The central issue is that 70 to 80 percent of municipal
solid waste can be extracted as refuse-derived fuel. The flow is continuous
and waste cannot be stored for any significant period of time. Nor can the
extracted RDF be stored for a significant period of time because of the
quantity and volume of material involved. It must go to a stable consumer,
if possible, one as stable as the community generating the waste. Organi-
zationally, there's nothing, absolutely nothing as stable as the utilities
173
-------
(as an RDF buyer). There just isn't any other class of organization that
can consume RDF as reliably as they. However, the utility operates uni-
quely; they are enfranchised as a monopoly in their service area. They
have to make money to provide the service because private investments are
involved. The complicating factor is their fuel adjustment clauses. They
are necessary, nobody argues this, certainly I don't. I think as important
as all of these problems that we've talked about here is that of finding a
way to make it economically attractive to the utility to buy RDF. A bargain
isn't good enough, and while it's something that we should all be appalled
at, we should recognize that the utility's first responsibility is stability
as an energy supplier. If we could find a way to make it worth their while
to use RDF in the context of the monopoly and fuel adjustment clauses that
they have, and should have, we would accomplish as significant progress as
solving purely technical problems. I think that the (resource recovery)
technology is going to work out; I think the technology is here mostly;
it's now a question of worrying it through routinely. However, we have to
figure out some way to make the utilities anxious to buy RDF for self-serving
financial reasons. We don't have that, and I don't know what the answer is
short of the kind of tax incentives that the Federal government might enact,
providing them (the utilities) with the potential to increase their profits.
ALTER: If I may sum up before we break for lunch, I think there is one
message that we are left with, which is: Anything we might do to reduce the
technical risk and uncertainty in people's minds even a little at a time,
will move us ahead. What I don't hear from anybody is a panacea. With that
we will adjourn for lunch.
SESSION 4: APPROACH TO R/D&D PROGRAMS
John Burckle, Moderator
BURCKLE: We now have in our hands a list of R&D needs, and we kind of
danced around the problem of what kind of strategy to employ in conducting
research and development. As I see it, we really have three or possibly
four ways to go about this. We talked about a pilot plant, for example,
and we talked about testing existing systems in the field. There's a pos-
sibility of testing a subsystem on a demonstration scale, large-scale sub-
systems. For example, say there is an existing fuel preparation system
somewhere in the United States and we want to find out how a particular
pyrolysis unit will work. We could have the ability to fund the construc-
tion of a pyrolysis unit in proximity to that fuel processing system so that
we don't have to completely build a total system just to test the pyrolysis.
We also have the mechanism of the full-scale demo, where the system is
actually integrated into the community's solid waste management system.
These various strategies give us different types of qualities of levels of
information, and I think that we talked about pilot plant testing and the
problems there. We talked about field testing and the problems there to an
extent. I'd like to open this area up to the panel to explore in more detail
174
-------
the areas of pilot plant or field testing versus demonstration either of
the subsystem or the complete system, and welcome questions at this point
or comments. I think one thing we ougnt to pay particular attention to is
scaling in our discussion here. I think that strongly influences the ap-
proach to conducting research and development, and I ask the people in the
area that furnish technology or technological systems to consider their
experience both with their own plant design efforts and utilizing EPA data
and the data of others. What kind of problems did they have with scaling?
How believable was the data generated in government-sponsored work? How
useful was it to you when one considers the factors of scale? What kind of
impact does data from a full-scale demonstration project have relative to
pilot plant work? What is a reasonable scale at which to perform R&D for,
say, a large municipal system, or an intermediate size municipal system?
What unit processes is it felt can be adequately scaled with the existing
information or techniques? What unit operations are essentially impossible
to scale based on today's technology?
LALKA: On the Monsanto project you sponsored, we went all the way up to a
1,000 ton a day plant from a small pilot plant unit, and I think that that
stuff's a tremendously large scale-up from a pilot plant to a 1,000 ton a
day plant. A 1,000 ton a day plant would probably handle a population of,
say, 400,000 people. I think it would be more reasonable to attempt to
scale up to something somewhat less than a 1,000 tons a day. In fact Union
Carbide, for example, has now gone to the modular concept, where they are
talking 350 ton modules, something like that, and that way you might be able
to develop a system that would serve a small community as well as a large
municipality. I think your effort would be better spent in that direction.
This problem of scale-up is really tremendous, and any data that you would
get from a larger plant, I'm pretty sure it would not be applicable to any
other type of pyrolysis system. You'd be developing data for just one
system. None of them are exactly the same. So overall, I'm not sure how
useful this information would be with regard to pyrolysis. It would be just
maybe a statement of one system, not really any other type of unit.
HOLLANDER: I agree.
PARKER: I have a problem. I'm up to my hips in 700 tons of data every day,
and my problem is sampling it. I think that the question of data extraction
is diametrically opposed to pilot plant and demonstration, small demonstra-
tion plants. If you have a small pilot plant, you can stop and start the
conveyor as often as you want. You can clean the conveyor, you can divert
the waste here and there. When you build a plant for profit, like Recovery
1 is, that's exactly what you have to do, make a profit. And when you start
getting in and sampling, it wasn't designed to do that, it was intended and
designed to process waste. When you perturb that by trying to take data
samples, you run into difficulties with running the plant. Now the problem
I have is, how often do you sample, how large a sample do you take, where
do you take it, and how do you take it without perturbing the plant process.
I've been wrestling with this for several months now, and I haven't come up
175
-------
with any an^vars. I don't know of any place that there's a book that says
a 250 pound sample of ferrous material is adequate to sample the ferrous
recovery system. A 250 pound sample on the tipping floor is adequate. I
throw 250 pounds out, because that's the number I found, but I don't have
any confidence in it. I'd like to know what other people have found in
regards to sampling process flow and efficiencies on plants of this size.
Harvey has the luxury of [unintelligible] where I don't. That's why I say
that there's an opposition here, where the best data comes from the big
plants, because the sample size is larger, but that data is harder to get
than in the pilot plants where you can stop and start it. So you have to
compromise somewhere, it would seem to me. I think it is going to be very
expensive to get the data from a large operating plant because you are not
going to be able to do it when the plant's on-line. You are going to find
yourself working on the third shift, you're going to pay the operator at
the plant extra to have the shredder operator there, and to have the Man-
power, Inc. people come out with rakes and the sorting tools for you. Data
obtainment on the big operating plants is going to be a very expensive
proposition. I just hope it's going to be worth it when we get it done.
HOLLANDER: We have here today a gentleman from Wisconsin Electric and also
from the City of Chicago. On the one hand we have a producer, and on the
other hand we have a consumer or user. The user is concerned with what he's
going to be buying, what is it that he's paying for. Similarly the producer
is going to be providing something to his customer. How do you propose to
monitor the value, the quality, how do you intend to sample, do you intend
to sample? Tell us what your approach is to this very problem that was just
raised.
FAY: We are very concerned about this because we are paying for Btus and
Mr. Sjoberg thinks 80 cents a million is too low, but nevertheless that's
about the going price. We're very concerned about getting a representative
sample, but nobody knows what that is, including the ASTM committees who
are trying to say what that is. We are going to do a lot of sampling initi-
al] y. We have our own analysis lab that we analyze all our coal in, and
hopefully over a year's period of time, American Can and Wisconsin Electric
will arrive at some sort of variation agreement, and determine from that
point do we really have to sample every other truck, or can we do it once
a day or twice a week or whatever. What we did not do two years ago when
we agreed on a contract, was try to determine all of these things without
any operating experience. We built into our contract a one year program to
do lots of testing, including sampling and analysis, and then we are going
to have to sit down and agree on what we have. That's the way we are ap-
proaching the subject. We don't think the answers are here today. The man
on this in iny shop has been to Ames, he's talked with as many people as he
can talk to. We think Ames is a different ballgame than ours because (and
I may be stating it incorrectly) I guess what they lose on one side they
make up on the other, what have you, it's all the municipality. That's not
the case when you have a utility like ours buying fuel.
BURCKLE: Is that on a dry basis, 80 cents per million?
176
-------
FAY: As received, same way we analyze coal.
BURCKLE: As received, the same way you analyze coal.
CHISAMORE: We are working currently with Ontario Hydro in the design stage
of a plant similar to what Bud has described, and we hope to get some of the
answers from what we call our experimental plant. It is about a 300 ton a
day plant, and it is being designed in such a way that we can shut it down
at will, because no municipality is depending on it. The transfer station
will start up next month and the processing part of it in June. Certainly
the agreement with Ontario Hydro and the municipality is such that they are
paying for Btus. We wrote in the agreement the same thing, it's the equiva-
lent of Btus, but the problem of how you come to an agreement on how many
Btus we deliver has not been resolved yet.
PARKER: I was misunderstood. I'm not concerned with sampling the end pro-
duct. I'm referring to the discussion we had this morning about sampling
unit processes or systems. How do you determine how efficient your aluminum
recovery system is, how do you determine how efficient your ferrous recovery
system is, x^ithout stopping the plant?
BURCKLE: Bill, do you see a need in your plant operation to do some of this
kind of sampling on a periodic basis to check out how the equipment is oper-
ating, to find out perhaps if maintenance is required in certain unit
operation?
PARKER: Yes, there's that need, but there's also an obligation we have to
test the plant. I'm concerned with that right now. I'm concerned with data
gathering just to see how well the plant will function and to make it func-
tion better. This is a 12-year plant, now down the line and not too far
down the line, we're going to have to draw up maintenance manuals, as you
referred to, we're going to have to determine what is the minimum perform-
ance that the drum magnet will operate at before we have to start to think
about changing it or readjusting the air gap, or changing the belt speed or
something. Those are corridors within which you will normally want to oper-
ate the plant. Before that you want to find out how well it's going to
operate and how well it can operate. Then you set the limits of acceptable
performance. The problem I have is testing it to get those limits of ac-
ceptable performance, and at the same time doing some R&D. What happens
when you speed the trommel up? What happens when you change the air gap on
the aluminum magnet? Or when you change the reagent in the froth flow cells?
All those things are easy to say you are going to do them. They are more
difficult to do in a plant that runs 700 tons a day.
HOLLANDER: Burckle, there should be someone here from Americology. They
intend monitoring the effectiveness of their line, and he addressed that
question. Is he here?
SJOBERG: We are currently in the process of shaking down the plant and have
built into plant operations the subject relative to evaluation. I think one
177
-------
thing you have to evaluate is the time frame over which you are going to
gather your data, because of the inconsistencies in the stream that you are
dealing with, plus the seasonal and cyclic effects that you are going to
have. I don't think any batch sampling is really going to give you any in-
formation.
PARKER: I think you are right.
SJOBERG: I think what you have to do is establish your evaluation program
over some reasonable time-frames, which goes along with what you presume the
cyclic effects are going to be. During that period of time you not only
evaluate your efficiency, which we will be doing on a weight hasis, but then
at the same time set up a system of bookkeeping within the plant to ^valuate'
your maintenance downtime relative to performance downtime, relative to
maintenance costs—normal, relative to emergency maintenance. All of these
types of things are being built into our accounting system, if you will., or
plant operations system, so that not within the first month can we presume
to tell you how the plant is operating, but within at least the first six
months. I don't think any batch system is going to prove anything.
PARKER: I agree. The accounting manual is a vital ingredient in determin-
ing the operating effectiveness of the plant. We are building an accounting
manual now that the company that runs the plant and owns it will use to re-
port their expenses and expenditures on. Beyond that there's the academic
interest of data, and the accounting manual comes in there because we have
designed it so that we can pull that sort of information out. There are
other things the accounting manual isn't going to take into account (that's
not intended to be a pun); there are some things it won't see. It won't see.
when the aluminum magnet becomes less efficient. It won't see when the glass
isn't as pure as it ought to be. That's another whole ballgame. How do you
test 50 tons of glass that we're going to generate in one day to see if it
passes the GPI spec on glass, without getting rid of the whole batch of
glass? Turnaround time is like two days on a test sample.
BURCKLE: Could you all perhaps speak to the difference on the sampling; you
mentioned batch sampling not being satisfactory. What is the alternative to
batch sample?
SJOBERG: If I read his initial statement correctly, or heard it correctly,
it was a function of shutting down the plant and operating it for specifi-
cally short periods of time to evaluate performance, and to me that's batch
sampling, which is not going to be of any value to you. I think that rela-
tive to the performance of the plant you have to then differentiate what you
mean you are trying to establish the performance of. We go to the plant
overall and you break into the cost centers and things of that nature; you
end up with a performance based upon manpower allotted or working at a
specific station of the plant. Downtime. Why the downtime? This gives you
an equipment performance trail that you can follow. We are going to obvi-
ously sell the material from the plant. Consequently our sampling relative
178
-------
to the products we put out the back door of the plant are going to have to
be an on-going, everyday situation of which we are in the process of de-
veloping specific techniques required to evaluate that product as if it
came out of any other manufacturing plant on a QC basis. We will have to
know every day exactly what the quality of the aluminum is, the quality of
the glass, the quality of the material we send to Mr. Fay, the quality of
the steel, because our revenue is dependent upon that quality.
RIGO: In trying to take apart the Landgard plant, our design calls for not
biting back inside the process too far but simply getting our hands on all
of the outfall data. Now what does this tell us on the plant? We will be
able to walk from the outfalls back into the process, hopefully, to some
extent. For your problem of determining when you are getting a quality
change, what you can do is, when the plant is running properly, establish
relationships, for example, on percent aluminum contamination in your mag-
netics versus some other point in your outfall. You'll be able to get some
kind of relationship, and when this relationship changes, you'll know that
something has gone wrong. You might be able to use that as a warning or
detection system.
PARKER: I have trouble backing into the plant from one of the outfall
points, because you don't go very far before you come to a common point.
RIGO: That's right.
PARKER: The cyclone collects the organics from the ferrous concentrator,
and from the air classifier. Which part goes where?
RIGO: That's a separate problem; that is a research problem.
PARKER: But it's a typical problem.
RIGO: Right, but your problem or your second problem, of maintaining quality
control I think you might be able to address by monitoring all of your out-
falls and establishing some relationships for proper operation, and estab-
lishing what the relationships look like under certain types of upsets. You
then know when a relationship changes, you've got a problem, and you might
be able to detect it before you have catastrophic failure—like you can't
sell your product.
PARKER: The product quality is one of my concerns. The way the plant oper-
ates and how well it operates is the concern I have right now. Before we
make a pound of iron, I'd like to be able to say how well the ferrous re-
covery system is supposed to work, and what makes it work good or bad.
BURCKLE: To get us back on the track here a little bit, I would like to
pose a question to the workshop. I think the basic question before us is
in terms of generating data. At what scale does data from the research
effort become believable to the community that will use this data for their
179
-------
plant design? Are there advantages in not trying to go the pilot plant
route but just simply field testing? For example, say the Americology
people were kind enough to let us come in and test their system out. Is
this sort of thing possible? How does the community look upon this kind
of testing? We have the option of a subsystem demonstration or a complete
full-scale demonstration such as San Diego, Baltimore, or St. Louis in terms
of trying to generate data for the general community use. I'm really inter-
ested in trying to get your opinions and feelings as to what approach is
perhaps more appropriate for generating data for your use and particularly
the concept of the scale in which research should be done or carried to be-
fore you feel it's usable to you.
LINGLE: I was just going to suggest, John, that again I think you have to
put the end use of this information into the equation of at what scale of
operation do we want to collect data. You have to also decide, are you do-
ing this to try to prove or disprove performance of a system or are you do-
ing it to try to develop design information on a component? I think that
the scale of operation at which you are going to carry out your testing is
going to be different depending on which one of those objectives you have.
So again, I don't think there is a single answer; I think it depends on
your objective.
WALTER: I'll agree, and I'll give you the military answer, which says it
depends upon the situation, depends upon what specifically you are trying
to do. Any given research program has to start from a basis and to work
itself in an orderly manner through the entire spectrum until you finally
hit the big crunch, which is called full-scale demonstration to see if it
will really work—potentially Seattle, or Saugus, or Nashville, or what
have you. There is no given answer. You may be very willing, because of
other data that you know, to scale up a thousand times. There may be other
times when doubling the size of the thing may be too far to go. There's
just no answer. You have to look at each situation and analyze each one;
there's no way to do it in general terms. I get a feeling when you are
talking about this sort of equipment from what I've heard here in the last
day and a half, that it's a particularly difficult answer, since everyone
seems to believe that there is no way to predict scale-up figures for
plants. If you build a bench model then you might as well forget you have
built it, except you proved what's going to happen with that size and what-
ever you put in it. You go out and build a full-scale and operate and tinker
with it. The only real problem I have is that it seems to me that what
people have been saying in general is that the operator, if you will, is the
tinkerer. I had a lot of experience in building maintenance, and as a
building maintenance person I was a tinkerer. My only real problem was,
there was never a flowback of information to the designer or the builder of
the piece of equipment, for him to know how I tinkered to make that thing
work. Therefore his next one looked like his first one, and the next guy
started to tinker with it too, until I bought the second one. Then I told
them, make this work, I want you to change it this way. There's got to be
a better backflow of information and I have no magic solution on how that
occurs.
180
-------
HOLLANDER: There's another aspect. If you were an equipment supplier, you
would be concerned with how many units you can sell in a year. You would
address what the size of the market is and what the size of the components
would be that would give you the greatest potential in this area. If you
were a systems builder, like Monsanto attempted to be—they determine the
single most economic size for them that would be attractive to the market-
place. The idea that was expressed earlier from our friend from California,
of a plant made up of one or more modules of intermediate size, would pro-
vide the manufacturer of the various components and the hardware with the
greatest flexibility for replication. We have to recognize somebody has to
fabricate this equipment, and what drives them, and what are the incentives.
So when we talk about scale-up, are we talking about scale-up from a bench,
are we talking about scale-up to commercial size, what is commercial size?
Is a single module, a single line of equipment a commercial size? When you
address that problem, then you have another perspective as to what you are
addressing.
RYDER: I have a slightly different viewpoint, I hope, from a real world
approach. I always thought that when a new idea came out and before mil-
lions of dollars were invested, a pilot plant was built to see if the idea
was sound. After the pilot plant was built, conclusions were drawn as to
what was needed for the full-scale model and then the money was spent.
Therefore, after people spend millions of dollars and essentially use the
large plants as oversize pilot plants knowing that the idea is basically
sound, why go back and build a small-scale model to see if we did it right?
WARE: It's easier to work with pilot plants.
SPENCER: If I could add to that, I think that it's so that the next time
that you build a big one, you do it better.
RYDER: We'll know that when our present plant gets on-line, and we get
operating experience. It's much more reliable to judge from the real thing
than from a small-scale model under laboratory conditions. We've got the
big one right here. A year from now we'll know many things that we could
never learn from a small-scale model in the lab.
SPENCER: You might be able to identify how that big one works, but you may
not be able to identify what to do to make it all work better.
RYDER: If we didn't we would be remiss in our duty. We wouldn't just run
the plant without making some kind of evaluation as to how it's working be-
fore we built another one.
BURCKLE: I think what Dave was getting at is what if you were to use a
different sequencing of unit operations, or insert a new unit operation,
for example. I don't know whether you are intending to use pre-trommeling
or not, but suppose you were not, and the question came up, could I produce
181
-------
the same quality fuel for less cost if I use a different type of process?
Pre-trommeling is an example. I think that's the point Mr. Spencer was
getting at. Surely we would know in testing an existing system over the
band in which we conducted our test, how we could tune that system to
operate at its optimum performance, or if there was a part of it that didn't
perform its job properly, what kind of fix to utilize so that that partic-
ular design could be changed on the next iteration.
RYDER: We will be doing that in the next year and the next two years, and
probably as long as the plant is operating. We will be trying different
approaches all the time to improve the efficiency of the plant.
CHISAMORE: It seems to me that the scale-up features are not unique to
this industry. The pulp and paper industry has done this in terms of
oxygen bleaching, for example. They simply went from small-scale to the
full-scale working plant and obviously having committed a major investment
to it, having to make it work. We may be in the same position.
LALKA: I think the chemical processing industry, the pulp industry, they
are dealing with one particular feed material. We're dealing with a type of
feed material that changes every single day. It varies all over the place.
It's tremendously difficult to handle. It's like no other industry. For
that reason, I think the small-scale pilot plants cannot be easily scaled
up to large units. On the pilot plant level people are running, say, one
load of waste. That's fine, one load of waste doesn't in any way represent
what you'll see over the life of that plant. So much of the data developed
is meaningless, and we've seen all of the problems in the past that have
occurred in this attempt to scale up refuse plants. It's very easy to work
with these things in your backyard, but when you go all the way for the big
time it doesn't seem to work that easy.
RYDER: I might add a little more to what I said. We are already doing
that. For example, we are already looking at improving our magnetic sepa-
ration system. Because our system was designed three to four years ago,
we find new ideas have developed since then - such as an improved magnetic
separation system. And we'll look at other things such as the trommel or
whatever else appears in the next few years. We're not sitting still in
other words, we have an on-going laboratory within limits as long as our
capability to dispose of refuse is not jeopardized.
BURCKLE: Any other comments on the point of scalability or research
approach?
SPENCER: I'd be a little skeptical of, for example, shredding information,
where the shredder was at a capacity much below 30 to 50 tons per hour in
terms of our problems. I'm sure other people have different problems, and
would be happy to look at information on a smaller system like the one you
are looking at, George. I think there is some data you've developed there,
but I think there is other data that you can't develop because you are talk-
ing about much smaller units of a different design than what you are en-
182
-------
countered with when you get to larger systems. Certainly in air classifi-
cation, I'd think, it would also have to be up around that capacity—30 to
50 tons per hour. I think the same on magnetic separation. I would think
that when you got into glass recovery that you could probably in a test
facility split down to a much smaller size, because your material is more
homogeneous, that's not to say it doesn't vary from day to day, but I would
think that from a large bin you wouldn't have to process all of the material.
You could split it out and have a continuous operation taking some percent-
age of the total glass produced at a 50 ton per hour rate and just be
[unintelligible] and how big the sample is. How much can you afford to test?
Once you've made glass, at least by froth flotation, you sure can't hand
pick it. The real problem there is, of course, the nonfusable materials.
That test can only be done on one pound samples, even though it might be a
split out of a 50 pound sample. Then you still come up with the question of
how do you take a 50 pound sample out of a 20 ton or 100 ton railcar. Do
you take off a very small feed stream continuously, or do you take every so
many minutes a sample out of a continuous feed stream, or do you dig your
shovel into the top of the car? I don't think there are any real answers
to that, except as that testing method shows up in your marketing agreements.
I think that the glass industry's probably pushing that in one particular di-
rection more so than a systems supplier. In any event I think that there are
some operations that could be tested at much less than 500 tons a day, or
whatever you want to talk about, and there's some that anything that was
less than that I wouldn't trust, especially where the material is very ran-
dom and variable, like at the beginning of the feed stream. After you've
made those classifications of magnetic separations and screenings, I guess
I would trust data that was substantially less than 500 tons a day.
HOLLANDER: Are you saying that there is a minimum size plant, too?
SPENCER: No, I'm only saying that in my particular design problem, which
might be different than a lot of others, if I see a lot of data on some of
the earlier stages, certainly on our air classifier we were testing it, I'm
not making a pitch for it, but we didn't believe our own data and wouldn't
really believe it until we tested it full scale. I'm sure Americology feels
that way, and Garrett does, even when you do it at small scale you don't
really believe that you can do it full scale until you do it.
HOLLANDER: Mr. Burckle, earlier I mentioned a commercial size plant that
should be tested. What is the smallest commercial size plant that is prac-
tical? It may be that you might have to put in a 40 ton per hour shredder
even for the smallest size plant that would have half that capacity. We
might use as an illustration the experience at Ames. Arnold, would you put
in less robust equipment than you have now, if you knew that all you would
have to contend with was half the flow rate through that plant than what it
was designed for and what it's really capable of?
CHANTLAND: I would have to say that I believe I would not, particularly as
most of this equipment is subject to severe abuse by the material that is
183
-------
being fed into it and it's being asked to receive. If you know our system,
we do very little at the front end, and unless you can guarantee some sort
of front end separation that's going to pull out objectionable things which
we talked about before, it gets very difficult to, you might say, scale
down.
HOLLANDER: So if we are going to have any kind of test program, it has to
be at least on equipment of the size such as you have at Ames.
CHANTLAND: I think so.
HOLLANDER: It has to be an on-going program monitored over a long period of
time so that the peaks and the valleys and unusual things will show up, so
you can be aware of the limitations and constraints of the system.
CHANTLAND: Yes, only I have to support very heartily Mr. Lamb on the test-
ing of units within a process. The pilot plant, as I see it, is a totally
new system you want to subject to evaluation; whereas, if we are simply
looking at problems and what we are facing today, it seems to me that we'd
better back off and test the various units available. We might find out
that even a plant that is now built or one that's in the design stage has
an option to move into these other units, because they do a little better
job or some other criteria. Just remember, I'm involved in problem solving,
I'm not a deep researcher. There are problems and I'd like to find
solutions.
BURCKLE: Arnold, I'd like to ask you, when you say you'd like to see more
information developed through testing on alternate units. In what frame-
work do you see these tests being done? At existing production plants? Or
do you see a large scale test facility dedicated to this sort of thing, or
just how would you see this being accomplished?
CHANTLAND: I think it could be in either case, if you get back to what we
were saying this morning. We've got to establish some parameters, some
ranges that you are going to go with in testing any particular unit. Every-
body talking about the same ballgame in testing. I can see it in either
existing plants and existing facilities, special research facilities, or any
way you can get the job done. The important thing, I think, is to get the
j ob done.
BENDERSKY: We've been looking at some of the test data that's been recorded
primarily on a research basis. We haven't found good data which shows a sys-
tematic approach to determining the question of what equipment is scalable
and what isn't. Until that is done, we know from some hard experience on
Baltimore and other places that certain things appear not be be scalable.
We really do not have systematic data that answer the question: How small
can you test a piece of equipment and use the data on a larger piece of
equipment? Until we have that kind of data, it seems to me we have to take
the cautious approach and test at full scale where we possibly can, so the
data is more meaningful.
184
-------
ALTER: Yes, David, but, and I think ttv t we can learn from the minerals
processing industries. Some of our equipment scales from there. The out-
standing examples are screens and froth float cells. I include in screens,
with fingers crossed, trommels. The trommel we'll see tomorrow was scaled
from pilot tests using the scaling laws for trommels.
BURCKLE: Are there any more comments on this particular subject?
TREZEK: Just to answer these other, I think the pilot scale gives you some
feeling for what it is you're supposed to measure in the first place. I
think it's a very important thing to get your head screwed on right before
you go out in a big plant. Secondly, it gives you the flexibility of easily
changing things around and trying new ideas, and things like that which you
can't do. I think that's the real value of the pilot thing.
SPENCER: You'd have to agree also that the job's not done when you've
finished with the small scale but you then have to go to the next step.
RIGO: Hopefully the pilot scale gives you enough data that the larger unit
is a validation problem instead of redoing everything, where your cost has
also scaled up.
BURCKLE: If there are no more comments on this section, I think we will
move along. We have several other people who are involved with pilot plant
operations. I've asked them to give a brief 10 minute summary of their
pilot plant capabilities, the type of work that is being done at these
pilot plants. I'll turn that over to Dr. Alter for his presentation at
this point.
ALTER: We started out five years ago to get a few numbers and relieve the
technical uncertainty I spoke about so much. What we found in working on
the trucks and by air shipping refuse here and there, and then moving
equipment hither and yon, was that it is very expensive to ship garbage to
the equipment or equipment to the garbage. Thus, there is a need for a
pilot facility. Then we got lucky and found that in the District of Colum-
bia, in the incinerator, there was an unused OBW facility. We signed a no
cost lease with the District, where they provide the facility and we provide
the maintenance and pay for the power. We moved in a few years ago and the
pilot plant has grown like Topsy. The majority of the process flow is shown
in the upper part of the slide. The waste is received on the floor, pushed
with a front end loader into a conveyor pit, and up to the existing sLredder,
which is a Williams 780, 1,000 horsepower, vastly oversized for this pur-
pose. However, the conveyor is undersized; it was sized for OBW and that
sort of limits us using MSW. At the present time, the shredded refuse dis-
charge goes to a Triple/S air classifier, one which was the prototype for
Chicago and New Orleans. It is loaned to us by Triple/S and we have been
working with them. This is one of a number of instances where we have been
able to borrow equipment rather than having to buy it. For example, the
heavy fraction goes past an Eriez belt magnet loaned by Eriez, to a con-
185
-------
veyor, by a magnetic head pulley, and then through a trommel with 2-inch
and 4-inch holes. The minus 2 goes to a glass recovery module, which I'll
show you. The plus 4 goes into a chute as reject, as a waste, and the 4x2
middling goes through a CPC Almag eddy current separator. The Al Mag pro-
duct is collected on the side, and then it has to be cleaned, which I will
also show you. The Al Mag system was purchased from CPC by Alcoa and loaned
to us. The air classifier light fraction goes through a blower and is de-
entrained in a cyclone, through the rotary valve and is pneumatically con-
veyed, either to the incinerator pit, where it is disposed of for us by the
City, or it goes to the d-RDF module, which I'll also show. The District's
incinerator is a 1,000 ton a day, no heat recovery, and clean. The nice
thing for us is their cooperation that we can get the refuse and they can
dispose of unrecovered or recovered fraction. We do not sell anything from
this plant. We're willing to share samples for investigative purposes and
when possible, share the resultant data. This is the receiving area,
originally installed for OBW. Within the plant is the ability to move equip-
ment in and out and try this and that. Here you see commercial waste on
the in-feed conveyor belt. I point out the hood that is built over the
belt. It was mentioned yesterday that shredders "blow-back," and this one
is no exception. We tried venting the shredder, and it didn't work, so we
solved the problem of the blowback by hooding the input conveyor, and it
just blows back on the input conveyor and keeps coming back in through the
shredder. It is a rather inexpensive solution and I wonder about its broad
applicability. This view, for those who haven't seen one, is of the Triple/
S air classifier. The new air classifier to be installed in the spring,
will go outside, and this one will be replaced by a vibratory feeder Triple/
S is lending us. We will feed the air classifier that way.
WISELY: [Unintelligible] comment on its performance?
ALTER: Well, this gets to what I tried to raise yesterday. What is the
objective of air classifying in a particular location? If it is to maximize
a yield of fuel, then this air classifier is fine. If it is to maximize the
quality of the fuel, no. Related, if it is to maximize, say, aluminum re-
covery, questionable. You have to define such objectives and have leeway in
changing the operation and adjusting the device. Performance for each ob-
jective is not easy to measure. The diagonal conveyor in the background
carries the heavy fraction. The primary magnet can not be seen in the
shadow. This view shows the trommel feeding the eddy current separator.
This arrangement has since been changed and no longer goes to chute, but to
a conveyor. The minus 2 goes to the glass plant, the middling, the 2x4,
to the eddy current separator and the plus 4 to a waste conveyor on the
other side. The Al Mag product comes out here, and on the other side, bi-
laterally, and on the left is the 4 x 2 Al Mag reject. The Al Mag product
contains aluminum cans, foil, some dicast, occasionally a piece of red
metal, and quite a bit of organic material which gets swept off the belt or
otherwise clings to the metal. In order for the product to meet any sort
of reasonable specification, it needs a cleanup.
186
-------
LALKA: Is this Al Mag a commercial size unit?
ALTER: Yes.
LALKA: What other pieces in your installation are commercial size?
ALTER: The Triple/S air classifier. By loading up the belt, we've put
through much more than 20 tons an hour, but we've run it at 20-25 tons an
hour. The belt magnet, the head pulley (the magnetic head pulley), the
entire Al Mag system are commercial size with one possible exception. This
trommel may be undersized, but it was sized as a compromise to fit into the
space. A difficulty that we have, a constraint, is space. We have to fit
into the building and the building wasn't designed for this purpose. As I
show you the other modules, I'll point out scale. The shredder is vastly
oversized, as I said; that was part of the existing installation. Just to
give you an idea, we now have conveyors going hither and yon in order to
carry off the waste fraction. This again is a function of having to be put
into an existing building. To clean the Al Mag product I referred to, to
separate diecast alloy from the can stock and organic contaminants, we have
been playing with what I call a "breadboard prototype air knife," which you
see in this slide. It has a vertical zig-zag air classifier, followed by
a vacuum hood. The diagonal pipes are to blow off the cans and foils from
the diecast. I don't know the answers yet, but certainly at a rate of may-
be, real pilot scale, say 100 pounds an hour, we have been able to produce
what is commonly referred to as Alcoa Grade 1, is the can stock grade scrap.
I should also say, before Arnold Chantland asks me the question, we have
also played with the Ames aluminum product here, and the answers are dif-
ferent. The Ames Al Mag product is much different in character from ours.
Here's a possible effect of city-to-city differences. The brochure from
Harvey Funk says Ames waste contains 0.3 percent aluminum. The Bureau of
Mines found more than 1 percent aluminum in Washington. The result is we
see less wood, less diecast, more cans, more foil, more paper in the Wash-
ington product than Ames. I think the settings—the manner of operation—
of this device may be suitable for one city and not the other. To the glass,
Under the trommel at the present time there is a screen. The minus half-
inch is fed from this screen to a conveyor, from the conveyor to a jig. We
are using a Bendilari jig we bought from the baryte mines of Utah. We be-
lieve the jig to be oversize in full commercial scale. From the jig the
glass product is chuted to a Penn Crusher impact mill. Again this is a
commercial piece of equipment, and if this one should be small for a full-
size plant, the manufacturer knows how to scale up. The material is crushed
to minus 20 mesh, and slurry pumped to the other end of the building, where
we had more room, to the froth float cells. We have used Denver cells,
again, full-scale equipment or easily scalable. At the present time, there
is a dispersion float follcwed by a rougher, two float stages and one
scavenger. At least, that's the way it is piped now. The product goes to
a spiral classifier, the gangue is slurry-pumped back to the other end of
the building, to a Henry settling tank, which was originally part of a wet
dust collecting system. This is a close-up of the four cells. The spiral
187
-------
classifier, for those of you who haven't seen one, is shown here. The dryer,
which is not completely installed yet, is shown here. It is a rotary type,
oil-fired. You see there is much in here that doesn't need research, such
as a spiral classifier. But I think things like the froth cells themselves
need research. Equipment from other technologies must still be installed,
shaken-down and maintained. Its performance on this new feedstock—from
MSW—must be determined. To those of you who have never seen froth float
glass, this is feedstock. On the scale on the bottom of the slide, each
mark is one millimeter. You can see the little black specs and other fine
materials, and a piece of brick, etc., that are mixed in with the glass.
When this material is floated, as you might expect, you see more glass, even
a piece of opal and you usually see a piece of blue. It is clean glass,
which shows why the interest in the froth float techniques. Moving to the
light fraction, we de-entrain the light fraction and pneumatically convey it
to another room. What is not shown in this little drawing is that there is
a cyclone feeding the screen on the left. It is a vibrating screen, pro-
bably undersized, but it had two other good attributes. One, it fit, and
two, we owned it. That screen may stay as a feeder because with our new
air classifier, we are going to screen before air classifying, thus hoping
to increase the glass yield and reduce the ash in the light fraction. The
light fraction is fed into a Heil secondary shredder, which I will show you
in a moment (it has been modified considerably for the light fraction), then
into a live bottom surge feeder to a California pellet mill densifier. This
is the outside, where the light fraction comes out of the air classifier.
At this meeting we have talked about materials handling problems, such as
bridging. This cyclone illustrates some of these problems. That cyclone
was sized working with one of the large cyclone manufacturers. After
analysis, we bought one the next size larger and yet it is woefully too
small for the capacity. I question if anyone knows how to size a cyclone
in this business yet. You notice a spinout on the bottom of the cyclone
of an odd shape. We found the light fraction bridging at the bottom of the
cyclone. So we rebuilt it with 2 x 4s and plastic film, and found a con-
figuration that didn't bridge and duplicated it in metal. Since then it
doesn't hang up. Why that particular configuration and whether another one
would be just as good, I don't know, which emphasizes why there must be
research in materials handling. The duct work conveys the light fraction to
the pit, or alternatively into the d-RDF room. We have not yet seen wear in
this chute. This view is of the d-RDF room; the Heil-Tollemache shredder is
in the background, the CPM in the foreground. I'll show both in closeup.
Again, quarters are cramped because of the space limitations. We also found
that the Heil moves a lot of air, and it really blew the product through
cracks in the covered conveyors. We have since added a hood and a vent and
exhaust to remove some air, and recirculate the dust. We tried a number of
hammer arrays in the Heil in an attempt to reduce the size of pieces of
textile. We did not know what particle size we needed to put through the
densifier, but we found, more importantly, that the textiles hung up in the
densifier and caused jamming. The result was not hammers, but rotors and
stators. If this device continues to work well, I will start to call it a
light fraction shredder. It is certainly not a hammermill. There are a
188
-------
few hammers in the bottom as sweepers. The material is shredded to minus
3/4 of an inch, so the textiles are fairly well macerated. We still get
some accumulation of textiles on the bottom machine.
SPENCER: [Unintelligible question].
ALTER: These are free-swinging.
ANONYMOUS: [Unintelligible question].
ALTER: We are screening to attempt to take out some of the inorganic fines.
Apparently we do, judging from the ash measurements. This is looking at the
densifier. For those of you who haven't seen it, it's a 33-inch die with
one-half inch holes. We are presently making pellets one-half inch diameter
by nominal one inch long. There is a breaker bar to control the length. In
this closeup, you can see the pellets extruded through the die. The d-RDF
in this hopper weighs about a thousand pounds.
CAMBOURELIS: Harvey, do you have any idea of how much that unit operation
would cost in a production situation?
ALTER: No. It depends very much on the operating availability of the equip-
ment. We are going through that now, and don't yet know.
CAMBOURELIS: Can you estimate it?
ALTER: Not ready to. In the end we certainly shall.
ANONYMOUS: What's the horsepower?
ALTER: One hundred fifty. You are now looking at about 9 tons of d-RDF
pellets. We ship them in roll-off containers, each container about 9 tons
of pellets. We are presently shipping the pellets to a power plant and
Greg Rigo, under EPA contract, is burning them. The way we have had to
put together the plant, there is a surprisingly large amount of demand
maintenance. Also, it is not always a clean operation. I'll also tell you
that we've consistently underestimated the amount of time and money it takes
to do research in this plant. There are always surprises here. If you want
to do R&D, our experience is, have a large bank account. Also plan on more
people than you figure. We use engineers, mechanics, co-ops are very good
form of labor, and day-laborers. Thank you.
BURCKLE: The next one I have slides available on is the Bureau of Mines,
Roger De Cesare.
DE CESARE: The majority of you are familiar with the Bureau of Mines his-
tory, but for those of you who aren't, briefly, let me say that when the
Solid Waste Act was passed in 1965, the Bureau was asked to evaluate the
potential of recovering minerals from refuse. They recognized that the
easiest way to concentrate the minerals from the fraction of municipal waste
189
-------
was to burn off the organic fraction; therefore, early research centered on
municipal incineration residue. Residue was in many ways as rich as many
ores being mined today, and it proved to be amenable to conventional mineral
dressing techniques. We set up a pilot plant which has been operating since
1969, using conventional off-the-shelf equipment, and developed a system for
recovering the minerals from incinerator residue. Separations are achieved
in a relatively simple manner, and projected economics for large scale oper-
ations are very attractive. With the reduction in number of municipal in-
cinerators in the country, the Bureau realized that residue recovery would
not solve the solid waste problem in the majority of the cities. We then
set out on a program to develop methods of separating various fractions con-
tained in raw municipal wastes. We also centered on the problems of the
East coast—large cities who are running out of landfill space. Unlike
incinerator residue, raw refuse could not be treated as normal ore, and is
not amenable to conventional mineral dressing machinery. Even though the
machinery we found to do the job was from agricultural and construction
industries and not the minerals industry, the approach taken to separate the
refuse was the mineral dresser's technique. In other words, we recover the
minerals in as large a particle as possible, separate into fractions which
can be further processed as needed. We arbitrarily decided to pick material
larger than two foot in size, and to process materials two foot or smaller.
The first thing we felt was needed to be done was break open bags of refuse
and to liberate all the material within. We found the flail mill to do this
job very economically. It would take everything that we would feed it, and
while it did not shred any massive items, at least the massive pieces caused
no damage to the shredder. As I mentioned yesterday, we found that the
flail mill required an exhaust hood, so we utilized this exhaust hood as an
air classifier to remove about 25 percent of the light paper and plastic
fraction. This fraction was a relatively clean fraction, because it was
separated before it was mixed too thoroughly with the rest of the refuse.
Since no separation process is 100 percent efficient, we decided to use a
•series of air classifiers rather than a single air classification stage.
We did purchase and test a few air classifiers which were used in the agri-
cultural industry, but were not able to apply them to our needs. Thus we
were forced to develop our own classifiers for our plant. I'll now show you
some slides of our equipment. I'll show you a few pieces of equipment from
the incinerator residue plant; I don't intend to dwell on it. Our process
for incinerator residue is a wet process. We have washing and screening and
shredding operations. Here is the first screening step, washing of the ash
from the incinerator residue. We then go to shredders, more washing and
screening, and then to magnetic separation. The non-magnetics are then
further crushed, and this crushing operation breaks the glass up to minus
quarter inch in size, and a vibrating screen concentrates the nonferrous
metals. If you have a good burn out on the incinerator residue, the non-
ferrous metal product from this screen is better than 95 percent metal.
Heavy media or jigging can be used to separate the aluminum from the
copper/brass. The rest of the operations deal with the minus quarter inch.
That's more crushing and washing steps. It gets rather tedious to try to
explain in a short period of time, but it's a similar type of equipment,
190
-------
and I don't think I'll go any further into it at this point. In the raw
refuse pilot this is the flail mill we were talking about. We bought this
unit back in about 1970. It cost about $13,000. It will handle a lot more
than we can put through it. We've never been able to test the full capacity,
I'm sure it will handle 10-15 tons per hour. The rest of the machinery is
scaled at 5 tons an hour. This test unit we bought with two variable speed
40 horsepower motors. We are able to vary the center distance on the shafts
and all of these variables were included in the total purchase price. We've
tried several different designs on this light air classifier as I mentioned
before, and it works very effectively for the purpose we are trying to
achieve. Next along the line is the magnetic separator. We have found by
having a light air classifier where it is, that a lot of the paper is re-
moved right off the top of the conveyor; thus our magnetic product is run-
ning very clean. We find that the magnetics contain less than 5 percent
paper, other than labels. In fact, over the past year, we've been getting
around 2 percent paper. This classifier is one which the Bureau built based
on one designed in our Salt Lake City research center used for separating
nonmagnetic auto rejects. The material is fed in at the upper right, and
there's an air stream going from right to left across the horizontal clas-
sifier. Lights enter the air stream and go into a cyclone. If you have
any massive metal that was not picked up by the magnetic separator, a
bronze pump housing or something like that, it's going to drop out on this
first conveyor as it will not be affected by an air stream. We needed this
as a protection device for our secondary shredder, and it has worked as
such. We get a medium fraction which contains all the rags and cardboard
and heavy plastics, glass, aluminum cans, wood, food wastes, etc. We thus
get a three-way separation. This middle fraction which I have mentioned,
the food wastes and the aluminum cans and glass, we put through a trommel.
This picture shows a trommel with one-inch square openings. We have since
standardized on 3/4 inch diameter holes. We have found that the way we are
running our flail mill, that all glass bottles are broken to minus 3/4 inch
in size, and we remove all of the glass at this point. In addition to the
glass, minus 3/4 inch ceramics, metals, organics, plastic and wood are all
removed here. We use a mineral jig to recover a glass concentrate from the
rest of the trommel undersize and we crush this material down to 20 mesh in
size and go to froth flotation to recover a mixed colored glass cullet. The
trommel oversize, this is material that is larger than 3/4 inch in size,
passes through our secondary shredder and into the third air classifier,
which is a three-stage aspirator. This is a commercial unit made by Fors-
berg. We've modified it, but it is a piece of available equipment. The box
under the three-stage aspirator is where we recover the aluminum concen-
trate. We have tuned this air classifier to drop out the majority of the
aluminum cans. We are able to tune it to remove all paper from the stream.
In other words, we can treat all paper as lights if we want, but we do lose
aluminum. Consequently we accept a little bit of paper in with the heavies,
but we are recovering currently 95 percent of the can stock that's fed to
the shredder, and believe it or not, we are recovering about 90 percent of
the foil. We then take the heavies from the three-stage aspirator, dry
them, and put them through an electrostatic separator. The electrostatic
191
-------
separator we have is a commercially available unit. The electrodes are
charged with 40,000 volts dc, the drum is grounded, and the dried material
is fed over the rotating drum. As the material enters this ionization
field which is set up, all the material picks up a charge. The aluminum
and the other nonferrous metals, being conductors, lose their charge im-
mediately and fall on one side of the deflector plate. The combustibles
fall down on the other side of the plate. We get a very good recovery.
However, the economics of it are high because of the drying cost of the
three-stage aspirator heavies. We are also currently testing Raytheon's
nonferrous metal separator, and they have been doing some work in our lab
also. Our research is conducted on a relatively modest budget. Our current
plans include treating refuse from municipalities which are planning to
implement resource recovery operations. In addition to refuse from Monroe
County, New York, and Baltimore County, Maryland, we have processed refuse
from Tampa Bay Authority in Florida, Montgomery County, Maryland, and
several other areas also. We have a small in-house program to upgrade our
recovered products and to optimize the equipment in the pilot plant. We
also want to improve the nonferrous metal separation. As was mentioned
several times, these past two days, we do have a cooperative arrangement
with Raytheon and Teledyne, and if any problems arise in their plants once
they get them on-line, our facility is available if needed to help them
solve their problems. We also use our equipment and everything that we
have in both pilot plants for treating specialized wastes from industrial
facilities which we feel could possibly be processed with this type of
machinery. Traditionally, it has been the Bureau's policy in our metal-
lurgy research programs, to conduct minerals recovery research in areas
where technology has not been demonstrated. We develop pilot plants, per-
form economic analysis, and if the developed process appears viable, we
assist in implementing full-scale demonstrations. Once the process is
demonstrated and the private sector has a handle on the problems and are
operating these plants, the Bureau then stops their research and goes on to
other work. Consequently, at the present time the Bureau is not planning
to expand their solid waste research program. As I mentioned, several
companies are implementing much of the Bureau-developed technology in their
plants. Often these plants are utilizing the technology without giving
credit where credit is due. When we started our research program, no one
else was doing this work. When the full-scale facilities are operating, the
Bureau of Mines will have accomplished its goal; our mission will have been
achieved. Solving these unit operation problems that we've been talking
about yesterday and today should be approached in a fashion not unlike the
manner in which the Bureau has solved similar problems in years gone by.
BURCKLE: Thank you, Roger. The next one on my list here is George Trezek,
University of California.
TREZEK: Let me begin by saying how our plant got going. As I pointed out
yesterday, we first started out as a facility which was set up just to do
shredding experiments, and then little by little we added more things on to
it. When the EPA laboratory at Center Hill was phased out or disbanded, I
192
-------
was called upon to receive some of the equipment from this lab, and try to
put it together and make it work. Let me run through a little bit of what
we are doing. This is a picture of the facility as it now stands (Figure
20). The building back here used to be an old munitions test center, and
that's the reason for all of the thick walls because those were blast
chambers. The solid waste facility is the part in the back. Packer trucks
can back into this facility (Figure 21), and essentially outflow comes out
of this part. The cyclone discharge from the classifier is shown up at the
top. We receive refuse from two types of places. One, University refuse,
which is pretty much 90 percent paper fiber type of thing, and we also have
an arrangement with Richmond Sanitary Service, who bring refuse to the lab
and take it away when we are finished processing. This particular load
happens to be a University batch, which has, as you can see, a fairly high
fiber paper content. This material would be typical of the kind of com-
mercial batches that we see, very high cardboard corrugated fiber types of
things. Over the 4 or 5 years that we've run this, we now pretty much have
the City of Richmond mapped as to what kind of materials you can get from
what location in the city. When a truck pulls in (Figure 22) we can pretty
much tell where he came from, by what kind of liquor bottles come out,
things of this nature. As I pointed out yesterday, we feed the shredder
with just a simple front end loader, (Figure 23) and I don't think this
would be the way I would do it if I were running a commercial plant, but
this is what we had available. We have built and studied air classifi-
cation, and that is the plywood system that you see in the back (Figure 24).
It is something that we put together and have been doing some air flow aero-
dynamics type of tests on. The ductwork up at the top is the ductwork which
leads into the cyclone, which I'll show you later. This is a picture of the
back of the classifier. We built this out of plywood for a total cost of
about 50 dollars. It happens to work fairly well, gives us about 6 to 7, 8
tons per hour throughput through this. We built it so that the back wall
could be easily movable by just removing just a few bolts. This way we can
begin to try to look at some of the scaling parameters as related to cross
sectional area distribution of material within the column, air distribution,
and things of this nature. This is pretty much a straight column except for
one baffle, which we observed through these windows was needed to induce
some circulation in the material. The heavies come out from underneath the
air classifier (Figure 24), come out on this heavies conveyor, and then have
a magnet which pulls out the ferrous scrap into this bin. We then take that
stream through a trommel and the undersize then goes into a system which is
essentially an air table. I found this to be an extremely effective way of
cleaning up glass. We come in here, we go one pass through the air table
and remove an essentially clean glass fraction. We have a small rotating
drum magnet which takes out any of the screws and nuts and bolts which
happen to separate out on this air table. We then go back through this
air table one more time and are able to get something like 99 percent clean
glass fraction. We have taken this material and sent it to a company called
Ecoceram, and they have taken that and ball milled it and have made what
they call Ecoceram building blocks out of it. So there is apparently some
commercial use for this type of material. The light stream coming up from
193
-------
•rl
O
Cfl
4-1
CO
0)
•p
0)
(U
PQ
O
csi
QJ
M
s,
•rl
Pn
194
-------
4-1
•r)
O
«d
CO
OJ
o
(-1
JS
O
CO
P-i
t>0
•H
PM
195
-------
0)
CO
a
IH
QJ
i-i
•H
CD
60
c
•H
M
0)
CM
CM
60
•H
196
-------
0)
t)
13
0)
t>0
c
•H
t)
0)
0)
Cfl
o
id
a)
C
o
ro
-------
light fraction air
classifier discharge
discharge
conveyor
ferrous recovery
heavy fraction
trommel screen
ass fraction
recovery
Figure 24. Air classifier.
bottom cone
of cyclone
light fraction
rommel screen
1 rotary air-lock
feeder
Figure 25. Air lock feeder, trommel screen, baler,
198
-------
the top of the classifier up through the roof comes down through the bottom
of the cyclone. We have experienced a number of problems with air-lock
feeders and things of this nature (Figure 25). This is an air-lock feeder
which we made ourself, and we have studied different kinds of blades in this
air-lock, and try to make it work so that it handles light fraction without
jamming. We also, as I said yesterday, went through some screening experi-
ments. We used to have a flat horizontal screen, which was one that was
sent over from Center Hill. We've since replaced that with a trommel
(Figure 25), and as you can see, have suffered the usual kind of problems.
The trommel screen that we made was built so that we could easily remove
the screen surface, experiment with different screen sizes, and categorize
and study the effect of the parameters governing screening. The screened
material then goes into a baling machine. We have looked at some of the
uses for mixed paper wastes. We had a project with Consolidated Fibers
and Certain-Teed Products in which Consolidated Fibers would come pick up
this material, take it to Certain-Teed Products, and they would pulp the
baled material and use it directly in the production of roofing materials
and other products which they manufacture. The cyclone discharge is ar-
ranged so that no material goes through the fan. The air lock drops the
material onto a conveyor to the trommel screen, the air flow goes through
the fan and is exhausted through a filter.
This is an example of one of the bales which came out (Figure 25). As I
said, these have gone to Certain-Teed Products. We have done tests to
instrument these to measure temperature rise of the bale over some period
of time. That gives you a fairly good indication of how well the screening
processes have been carried out. I'd like to point out that we have done
experiments to see what could be done with the minus fraction from the
trommel, which would normally go into a drop box and get taken back to the
landfill. We've gone through a number of digestion experiments in which
we digested this kind of material with sludge. Here is a small-scale unit
(Figure 26) which we started, in order to get some basic ideas about the
behavior of the various controlling parameters. In this set of experiments
the control was a 100 percent sludge digester along with the refuse sludge
system. After this we went to a 400 gallon size unit in which refuse mix-
tures are fed in through this feeder. This is a doable wall concrete
cylinder which is heated by a hot water tank heater. This is shown now
with the cover plate off. There's a mixture which goes into the top of
this, and we've done some work on what kind of mixing problems you get into
when you start getting into sludge refuse digestion experiments. At one
point we were very curious to see what you could do with fiber recovery if
you had a good front end system ahead of the fiber recovery operation. So,
one of the things we've set up is to take the screened "accepts," which we
call them, or essentially screened light fraction, and subject it to a
pulping process. You need to essentially slurry the pulp, but not the
plastic. We worked on that problem. We then found you could do that,
then go through what we call a plastic removal dilution operation, followed
by some cleaning operations and then eventually come out with pulp. I'll
show you some slides of that (Figure 27). We've also looked at the kind of'
199
-------
Figure 26. Incubator containing digesters,
200
-------
or
o
tr
LU •
e
51 X
> UJ
co
t-
a
~>
UJ
CC
j2
Q
s
O
CO
o
U)
<
o
Nl
to
UJ
_J
IjJLU
CC "
{2 >-
i3 z
K U,J z
1 -
J llJ t~~
Iz 1 a -J
«" O to to ^
<-> ;p P
_J H |10 >- 0°-
< *- to «>H """ -J^^o-
> *s f° ° -i
O _J 5 cc <
2 0. 0 >-
UJ (- 0 O.
UJ
,| 0
CC <3j y
Z L \-^
LJ M-"
tf L
o ^
CJ CC
to
H
r- -• O
2 . UJ
ii I *~ ~~>
UJ , - K W
-*- or [ col is> tr
<_> i — — •-
W wu^ !f?
lg «
H-Q^ -^
PRIMARY
!
<
i
t
i
1
<
WATER
o
_j
h-
1-
UJ
to
r
5
r
jj
4
•>
SETTLIN
e
Cert
KO
* V.
SLUDGE
i.
1
I
SUPERNATANT
1 °-OUJa)HUJQ:
UJ
co z
00
-JUJH
--Jo
Q c^Ij
ZcoO-
y
CL
Q_
fc
cot-
UJ
o
Q§
uH
^8
99
to.j
J a. <
2!
UJ
UJ
CO
o
— UJ
OJ
O
O
«t
o
OUI
ao:
bO
n)
•H
13
I
CO
CO
0)
o
o
0)
n
e>0
•H
201
-------
water treatment problems that you get into if you had a good front end
system or if you didn't. This is a picture of the top of the pulper,
which has now slurried the pulp or the light fraction. 'You can see large
pieces of plastic and styrofoam and all sorts of other things which are now
liberated and float. Here's a picture of the plastic removal operation; a
traveling screen which the pulp slurry comes down on. Jets of water flush
the pulp through the screen, the plastic remains on top, and the clean
slurry is now diluted to the right consistency so that you can go through
a centrifugal cleaner and essentially clean up the fines and other contami-
nants. We've grown out of the building we were in. This is a plastic re-
moval screen, the pipes going in and out are to be hydrocyclones which are
inside the building. Here's a sample of a hand sheet which was prepared
from a light fraction material. We've sent this to a number of paper mills
and so on to see what could be done. By the way, the heating value of this
material is about 7300 Btu per pound, and it's very clean. We've also
looked at this as an energy source. I might just point out we sent this as
a feedstock to some of the people that are doing fermentation experiments,
and they found it to be a reasonably good feedstock for that kind of pro-
cessing. This is just a close up view of the same hand sheet showing that
it is possible to really clean fiber.
Now one of the things that we are just getting into in the lab is studying
the burning characteristics and acquiring a fundamental feeling for some of
the combustion aspects of densified fuel. We've made this small combustor
in which we have a way of feeding in pellets. This combustor can be co-
burned with either gas or oil. This grate vibrates, and therefore we can
flow the material down and begin to study some of the combustion character-
istics. We are now changing this configuration to go through a rotating
basket in which the pellets will drop out and be rotated through the flame
zone. I think that gives some overview of what we are doing.
BURCKLE: The Department of Defense, Navy, is going to give us a brief over-
view of the DoD program in resource recovery.
HURLEY: I don't know, John, if I can address that. That is a large topic.
I would like to at least convey some of the thinking that has gone into
what we call an experimental operation. We picked the Norfolk area. We are
about to advertise in Commerce Business Daily for expressions of interest.
By way of background as to how we came to this, I'd like to say that there
is definitely thinking and agreement of consensus, backed by the Navy organ-
ization responsible for planning and construction of shore facilities for
the Navy, that we can utilize the fuel value of the solid waste we have been
throwing away. With respect to the rising prices of fuel, with respect to
the EPA Guidelines, (which some people say is pressure; I would like to
think of it in terms of an incentive), and certainly a DoD position that
says we will indeed follow the Guidelines, and henceforth, or at this moment,
start planning to be in compliance with those Guidelines. It's a combi-
nation of circumstances that, I think, causes an organization to look at
what they are doing and how they are operating and making a decision to do
202
-------
something about it. In looking at the number of organizations and the re-
gional concepts, it seems like a rather endless number of options are avail-
able. With some activities in close proximity with rather large municipal-
ities, and the option of joining in a municipal system very promising, very
likely, and other locations, small systems, look like they have to go it
alone, and other locations are faced with being in the proximity with other
military installations, and the idea of inter-agency cooperation becomes a
difficult aspect of arriving at how you introduce projects and how you in-
troduce a cooperative venture. So it seems prudent that both the Navy En-
vironmental Protection Program and what is now called the Energy Conserva-
tion Program should join forces and bring together an action that would
accomplish the objectives set out in both of those programs, so that has
been a bit of a task to bring that cooperation and understanding about. At
this point we have two programs joining forces and we intend to go with a
full-scale, I'll call it prototype, experimental unit in the Norfolk area,
to try and determine what degree, and with what problems, and at what cost,
we should be interested in or should be trying to make an RDF. I'd like to
answer whatever questions I can at this time, and that's about the extent of
where the project is right now. There is interest in terms of putting it on
a DoD basis, which means to what degree it can be related to the other
services, and to the responsibility that belongs to an organization called
Defense Logistics Agency, who have the assigned responsibility of doing all
of the DoD marketing. The situation today is that the trash at the curb on
the military installations belongs to Uncle Sam, and the NAVCOM, the DoD,
control rules have been changed to accommodate or create an incentive so
that the installation that wants to practice this resource recovery oper-
ation can indeed retrieve the funds that it generates; they don't go back
to the U.S. treasury as formerly was the case. So it is with some incentive,
the ground rules have been changed, that I think has precipitated some of
the action over the past year. To what degree the commanding of fleers will
recognize and get information that helps them make decisions, what options
they have locally, are yet to be seen. But as I say, the DLA, the Defense
Logistics Agency, has the assigned responsibility for handling all the
marketing, which means that all materials that are recovered and are to be
sold are handled through that organization. They have people called Defense
Property Disposal officers at most installations, who handle this, who de-
termine what the market is, and how you recover resale material. I think,
John, that in essence is what we are about to launch into, that's relative
to this group and their purpose.
BURCKLE: Thank you very much, Steve.
HOLLANDER: About nine months ago, maybe it was a year ago, there was some-
thing in the Federal Register mandating that a program like this be under-
taken for military facilities adjacent to population centers or within popu-
lation centers, having a certain minimum waste generation rate at the mili-
tary facility, that they should erect the facility for resource recovery and
also size it to accommodate some if not all of the community wastes.
HURLEY: I think you are referring to the EPA Guidelines.
203
-------
HOLLANDER: Well, it was in the Federal Register; it referred to military
installations. It seems to me that it was directed at Federal facilities.
HURLEY: [Unintelligible] option of going in a region or going it alone.
HOLLANDER: If the waste generation rate at the Federal facility was beyond
a certain point, 50 tons a day or 100 tons a day, something like that, if
they had that much waste, singly or in concert with other Federal facilities
in the area, that whatever they built should also try to accommodate the sup-
porting community's waste.
LEVY: You are referring to the Federal Facilities Guidelines, and I don't
have copies of them with me, but I think you are misinterpreting it. There
is a mention of the joint cooperation, but I think it was more intended to
be the other way too, that the facilities would cooperate with the communi-
ties and their systems, or maybe it opens it for either, going in either
direction. It is the Federal Facilities Guidelines which are available
which you are referring to.
HOLLANDER: Is there something more current that this directive about a year
ago?
LEVY: No.
R. FREEMAN: I'm Robert Freeman from Stanford Research, and we did the work
for Steve. What you are referring to is the Guidelines that said that any
one Federal facility within an SMSA that generates at least 100 tons a day
and all of the Federal facilities from that area at least 100 tons a day, at
that time the largest Federal facility would become the lead agency for
other Federal facilities to take care of and administrate the garbage. They
could choose to either utilize a resource recovery facility, or establish a
resource recovery for gaining their energy, materials, or both. EPA promul-
gated the Guidelines, the Guidelines came back, they broke out the areas of
responsibility, the Navy had nine areas of responsibility, the Air Force has
one, the Army four, and so on [unintelligible] 100 tons per day scale, and
this information would then fall into the Guidelines that you're talking
about.
FUNK: How does this request for interest tie in with the Naval operating
base and the existing waterwall incinerator utilizing the plant waste from
there right now?
HURLEY: There's no connection. We are going to that area because we have
a variety of military installations in the area, and we hope some market,
too, for the materials.
FUNK: Are we looking for outside wastes coming into this facility or are
you looking at a 100 tons a day facility?
HURLEY: We are not looking for outsj.de wastes, but I think the opportunity
exists, and if other people can see a need for joining in or using the
204
-------
facility, I think the option ought to be kept open. So I don't think it
should be restricted, and I certainly haven't heard anybody's intent that
it should be restricted to just military operations.
FUNK: Well, the reason is that we are actively involved with the planning
district there and the Navy, and there's quite a program that's launched and
there's quite a bit of momentum to be utilizing nearly all the waste in the
area.
HURLEY: I understand the question, and I can tell you that our organization,
and the geographical split in our organization, and that is our [unintelli-
gible] office who handles that location. They are aware of the project, and
they are trying to keep people involved in the planning clear so that local
confusions do not exist, in which an experimental setup would come into a
region and it would be misconstrued as a project to solve a local problem.
RIGO: The key to selecting Norfolk was that the plant was not needed. There
is no problem of solid waste management for the Navy at Norfolk. The Norfolk
incinerator and the Portsmouth incinerator, both of which are under-utilized,
provide excess capacity. That is the reason the site was selected, because
we were afraid if we went to a place that needed the plant or had a solid
waste problem, the plant could not be used as a test bed, but would be forced
into a production mode.
CHISAMORE: In Ontario the Ministry of Environment is, I guess, a bit unique
in that it's a regulatory agency and over the years has also been a funding
agency in terms of providing capital funds for sewage and water works, and
three years ago the same funding was extended into the solid waste side of
it. At that time a commitment was made to build what we have termed an ex-
perimental plant. It's experimental, basically, in name only in that we did
decide at that time to go to the off-the-shelf equipment. You've got the
smaller scale in terms of shredding, 40 tons per hour. It is being set up
in such a way that we can shut it down at will to determine efficiency or
make changes. Certainly no municipality is depending on it for disposal
purposes. We have incorporated as part of the receiving area a 600 ton per
day transfer station so that if the plant is shut down the waste can be
transferred. The brochure which I brought with me provides a description
of the equipment which has been installed. As a highlight there are two
stages of air separation, one of which is patented by Vic Brown, who was
acting as a sub-consultant to our main consultant. In designing the plant
layout, we wanted to maintain as much flexibility as possible to explore both
energy and material recovery. The plant will be used to demonstrate tech-
nology and also for market development purposes. Many industries have ex-
pressed an interest in using RDF; however, the obvious problem is how to pro-
vide a sample, which in some cases will exceed several hundred tons. The
Experimental Plant will give us the capacity to provide necessary samples.
The Province has also made available a capital subsidy for front-end plants.
Under this subsidy program the Province would finance and construct the
plant and recover 50 percent of the cost from the municipality as a user
205
-------
charge. This offer was made initially to six municipalities, but to date,
because of many of the uncertainties which were described yesterday and to-
day, these plants are not proceeding at this time. In addition to the Ex-
perimental Plant we are encouraging demonstration projects. Yesterday I
mentioned the project with Ontario Hydro, which is a public utility. Right
now we are waiting for environmental approvals. Ontario strongly favors en-
vironmental assessment and even though the Province has some funds in the
demonstration project, we are experiencing the same problems of obtaining
environmental approvals that the private sector is. We also have an agree-
ment with a cement company to use what we term the light fraction, which is
primarily paper and plastic film. This autumn we will be doing some test
burns in their kiln. In the Experimental Plant we have a small Consumat
system for energy recovery to provide some information which may assist small
municipalities which are considering this type of system. In addition, the
plant has a 50 ton per day mechanical composting module. The plant is set
up on a modular basis and facilities will be made available to manufacturers
who wish to test their equipment. Any information gathered from this plant
is public information. The plant is designed such that you can see as much
as possible and will be available for tours. The transfer station will
start up next month with plant operation contracted to Browning-Ferris In-
dustries (Toronto) Limited. The remaining part of the plant should start
operation in late summer, so it's a little premature to schedule any tours
yet. If there are any questions I would certainly be glad to answer them.
FISCUS: Have you made any plans to monitor or measure in any way the environ-
mental emissions from the various processes that are incorporated in this
plant?
CHISAMORE: Yes, there will be extensive monitoring done throughout the
plant on unit processes as well as emissions to atmosphere from the Consumat
unit and dust collectors.
FUNK: When did you say you were going to be operational?
CHISAMORE: The transfer station will be operational next month; the pro-
cessing side of it in June this year.
FUNK: June of '77. Tours when?
CHISAMORE: Anytime after that. Actually we have had some tours. People
have been going through already, but certainly at this stage it is rather
difficult, you know, in the final stages of completion.
HOLLANDER: Is this going to be in Toronto?
CHISAMORE: The plant is located in North York, which is one of the boroughs
of Metropolitan Toronto.
PURCELL: If somebody asked you what is particularly different from your
plant from what we've been talking about, what, in a word or two, would you
say?
206
-------
CHISAMORE: The basic train is similar to that which would be used for pro-
ducing RDF. It does have additional features, which I mentioned, to provide
flexibility for both material recovery as well as energy recovery. In the
early stages we have not included froth flotation or aluminum recovery in
the plant layout. We are hopeful that the private sector can recover valu-
able materials from plant residue on a commercial basis without the need for
financial involvement on the part of the Province.
BURCKLE: Any other questions? In view of the time and our desire to get on
into the summary session and the nebulous nature of Area 5 here, I think for
the purposes of the workshop we'll skip over this particular question and
suggest that people give it some thought after they've returned to their
respective offices and homes and determine then if they would like to com-
ment on Item 5 when they review their comments, before publication and per-
haps send in something in written form on this particular aspect. I feel
personally that these systems are going to be built at some point in time,
depending on the cost of fuels in the future, and that left to its own, it's
a matter of time. I think the basic question rolls around to, do we want to
furnish some sort of incentive as in 5 or another type of incentive that
would accelerate the employment of resource recovery systems on a relatively
nation-wide basis? If so, perhaps what is really the best way to go about
this? I would be very interested in your opinion and comment.
HURLEY: I think the thing that I heard and have experienced is that there
is little interest in the organization I'm with when you talk about national
problems or big city problems, or 1,000 ton unit, but as soon as it became
a realization that small activities had to do something and could possibly
do something, and I think that's when I got a little attention with respect
to what can you do for the organization. I think the project we are ven-
turing into is one that it has something hopefully in it, it has an objec-
tive that returns something to the organization, whether it's fuel or some-
thing else. That to me is the thing that triggers the people to make a
decision. So we hope that our project will indeed deal not only with small
military installations and other small communities like it, but also many
of the problems that have been discussed here in the last day or so.
SESSION 5: SUMMARY AND CONCLUSIONS
David Berg, Moderator
BERG: This is Session 5: the final session. I feel a little bit like a
desperate man right now. I never did like trying to do summaries, so I
guess at the outset I'd ask you all to bear with me as I try to do the best
I can. I guess the best way to start is to go back to the beginning of why
we are here. One way of looking at it is that we are in a period of fuel
crisis, we have at the same time an environmental problem which is really
affecting many cities around the country, which in the estimates of all the
207
-------
experts will only get worse as time goes on. There's a generally recognized
agreement that greatly expanded use of RDF and of materials recovery could
greatly help, or at least to some extent help, ease this crisis, and cer-
tainly to reduce the impacts of future crises. What we are talking about,
then, is trying to telescope the birth of an industry from 50 years to 75
years to maybe 10, or 10 more. As I said, this is partly because of the
energy situation that we might want to do that, but it is also because in-
dustrial society has reached a state of development such that we don't have
time the way we used to have time to develop the secondary systems and the
analytical data to operate. We have work-place standards and environmental
standards and competitive uses, and this is not the same ball of wax that
we used to have before; the system is much tighter. There's much less room
to flex. Having said those general remarks, where's that leave us? I sup-
pose that as I sat here the thing that struck me the most about the meeting
was the very strong interplay that dominated the discussion. It seems to me
that on the one hand we were talking about near term problems, and on the
other hand we were talking about longer term optimization. The group was
much more comfortable talking about the near term problems, naturally, but I
must say that it seems to me there was a commendable struggle that we all
put up to address the pathways that are going to hasten optimization of
resource recovery and implementation of optimized systems. Maybe I could
bring this observation home by illustrating. Immediate issues seem to be
raised most frequently where this observation that immediate ease issues
seem to be raised most frequently. The municipal implementers continually
stress their need for operation and maintenance costs, revenue estimates,
reliability data in planning systems and improving ones they have. The
planning engineers continually stress that they also have data needs. They
need to understand equipment operating experience, and they also could use
sets of performance specifications. The system designers expressed their
desire for data to predict unit operation performance, and then all that
recycles back to develop improved unit operations and the data is a key to
developing improved unit operations. So lack of data at all levels seems
to be impeding both implementation today and in the near term, and on the
other hand, development of new systems and new unit processes. The last
point on that is that the need for comparative design data was one clear
point that was raised. A design handbook for the short term, for the longer
term, R&D in an orderly fashion seems to be recognized as being needed on
the parameters of processes, again so that we can develop systems that will
be better in the future. Thanks to the session chairmen I have several pages,
which I will try to skim through quickly, of summaries of each of the ses-
sions. I took the liberty to mark each of them up, and I'll try to run
through these as quickly as I can. Maybe what I will do is pause at the end
of each one and if people would like to challenge or modify things that have
been said, or add additional remarks that they feel are particularly import-
ant with respect to those sessions, this would be a good time to do it.
Session 1 was the Systems Process Design discussion. Here I heard Harvey
Alter mention an outstanding point; he said there is no state of the art.
During this session workshop participants seemed to me to note an extreme
208
-------
lack of information on virtually all aspects. There were several areas
which were identified as being most important, areas where more information
or research is required. Four stood out. One, materials handling systems;
second, scale-up of equipment; the third, quality and composition of raw
refuse; fourth, specifications and requirements of fuel produced. There
was a strong interplay between the third and the fourth, the quantity and
composition of refuse and the specifications and requirements of fuel pro-
duced. Process design, of course, depends largely on the specifications of
the fuels, and parenthetically I would note that little is known about what
these specifications should be. On the other hand, system process design
is also highly dependent upon incoming refuse quantity and composition.
The accuracy and reliability of this information affects the economic per-
formance of the system, as well as the technical performance, so the trade-
off is posed between the quality required by specifications, the user's per-
spective, and the economics of production, the municipal perspective. Of
course, caught in the middle are A&E firms and EPA, ERDA, the Bureau of
Mines, and the Defense Department, which are trying to be facilitators.
Two other important points or items which were singled out were sequencing
oi unit operations and modeling of alternative systems, which seem to be
generally considered to be important topics. On those the consensus seemed
to be that there's insufficient information available to make specific
statements about optimal sequencing, and there's insufficient information
Lo implement or verify models that would be of much use at this point in
time. A] so there seemed to be no good comparison of the alternatives; this
point came out strongly throughout. Another outstanding consensus was that
there are a lot of variables. (I told you it is a disaster to try to
summarize.) One point which will win no friends is that trade secrecy and
proprietary information was felt by many people to be a real barrier to
information flow, and this is a real counterpoint to the concern that peo-
ple seem to be expressing that we should be increasing the information
flow. This impediment will affect the ability of the players to improve
design and develop new unit operations as well as systems. That's what I
had on Session 1. Is there anybody who would like to comment on that or
add points?
ALTER: I'm not sure that trade secrets and proprietary information have to
be an impediment. I think we would move along faster if there were more
room for invention in a proprietary position. I think we can advance, as
we talked about this morning, at least advance the methodology for ob-
taining comparative data, even working with the people with proprietary
systems. We can share each other's experience and help the proprietary
positions lead to profit. The profit leads to incentive, and we will move
faster. I don't like the word impediment.
BERG: I think there is merit to what you are saying. That debate is re-
flected pretty clearly within the government itself in the various rules
that apply to government support and patent rules which apply, where the
energy, particularly the atomic energy program, used to have a much more
liberal clause in there permitting much more proprietary information to be
protected than did almost all of the other R&D areas. I think that with
209
-------
the advent of the new legislation in solid waste, we are moving more in the
direction of the Atomic Energy Commission approach, which is now ERDA's
approach, where there will be more protection. I suppose you are right.
On the other hand you have the problem that there is a reluctance to expose
systems to comparative testing. That's natural in a competitive situation,
but it does make more difficult the data collection and sharing.
ALTER: I think that as proprietary systems are developed, and the develop-
ers gain experience and confidence, they will make more information public.
They want customers to be aware of their systems, and they want the custom-
ers' consultants to be in a position to pass judgment on the systems. The
only way they can do that is to make the information public. At the same
time, I sympathize with them for not wanting to release information until
they have confidence in it.
BERG: Any other comments on Session 1?
CAPPS: Dave, what about the site specific market or users? Wasn't that
the thrust?
BERG: Yes, I think so. Session 2 was on Unit Operations. Within that
session, the general format was broken out in materials handling, size
reduction, separation, and process control. Again in an observer role, I
would say that much of what was said with respect to the system design
questions or topic applied to the equipment session. First, it was widely
noted and restated in each of those four aspects of the discussion or parts
of the discussion, that there is an overall scarcity of performance data and
an overall scarcity of parametric data, especially with respect to compara-
tive performance data, performance over a wide range of operating conditions,
and the relationship of input and output parameters. Looking at it from the
short term versus the long term perspective, there was no general agreement
on how much data is needed, what form it should take, or how to get it.
There was some comment on the proprietary problem again during that ses-
sion, although within the constraints of what Harvey was pointing out.
There was also a general feeling that equipment works, but I would say that
people recognized that what equipment exists today is clearly suboptimal
equipment, primarily equipment that's been adapted from other industries.
There were no specific major equipment problems identified to prevent plants
from being built; I think that statement would be accurate. Steve Lingle
put out a trial balloon that Atlas storage bins were an impediment, but
there seemed to be a feeling that that was not the case. Materials handling
equipment seems to work, although there are problems. I think that that was
the thrust of each of the areas that we talked about. On the other hand,
looking at this thing from the longer term perspective, there was a need to
understand, or people recognize the need to understand, fundamental para-
meters. George Trezek's paper or discussion gave us a pretty good snapshot
of how that kind of data is useful and can be made useful, and I think re-
inforces the thought that we need to develop more data like that. As Peter
Ware pointed out, we need to extend the state of the art of resource re-
210
-------
covery from an art more towards a science, and it's that kind of data that
will allow us to get there. Running quickly through the specific comments
on each of the four parts of the discussion: With respect to materials
handling, it seemed that people agreed that it is difficult to handle wastes;
again to repeat, waste handling is an art. Secondly, equipment choices for
storage and retrieval are limited. They are largely an adaptation of equip-
ment designed for other materials. Size reduction: There's more experience
with field operation of shredders than from other unit operations. We can
shred refuse, but as Harvey Alter pointed out, the state of the knowledge is
pathetic. Again, there's not good information on comparative performance,
nor is there on the relationship of input and output parameters. Third,
other types of size reduction equipment need to be explored. That was a
clear conclusion. With respect to separation, there was one major point I
think: That air classification was recognized as an area where there is a
great deal of uncertainty. This particularly was pointed out as being an
area where we are dealing with a subject that's an art form today. I had
to feel somewhat amused when Harvey was describing the Triple/S classifier
(I guess it was the Triple/S) that you are just putting in where you can
vary just about every single parameter in the entire unit.
ALTER: No.
BERG: It was the experimental one that NCRR is putting together, where you
can vary virtually every parameter from the angle of the baffles right on
through to the air flow and what have you. In the process control part of
the discussion, important considerations were raised with respect to dust
control. The feeling seemed to be that the existing technology could be
applied to solve these problems. There was a very gingerly discussion on
the health implications of dust. People seemed to not want to touch that
one with a 10-foot pole, although Arnold Chantland pointed out very aptly
that there is a tremendous sensitivity, and others as well pointed out,
there's a tremendous sensitivity to this problem. There seemed to be an
agreement that we should get thorough information, but at the same time
people are aware of the tremendous expense involved, on the tune of, I
would guess, in excess of or nearly 10 million dollars, by the time you
finish doing epidemiological studies. A very dismaying dollar figure.
With respect to fires and explosions, the need for a design manual for
shredder explosion control was pointed out. That's the notes that I have
for that session. Are there any comments here?
ALTER: Where did the 10 million dollar figure come from?
BERG: Out of my head; nobody said that. I just made a guess. We'd been
talking about that a little bit in the office. It just seems like it would
require a fairly enormous amount of monitoring, and that individual epide-
miological studies are just expensive to do. MRI is gone, or at least Larry
Shannon and Maurie Schrag are, so they are off the hook, and I won't even
look at you. Are there any other comments on that session?
RIGO: I think one thing that you missed that was mentioned in passing was
that there is need and room for some blue sky work, probably low dollar
limit, but where people are inclined to sit back and reflect on something
new.
211
-------
BERG: I think that point was brought out. Maybe we should take a quick
break from the sessions and announce that there is no winner in the pre-
processing renaming sweepstakes. Don Walter submitted the worst suggestion
of all, that was mecachemical or chemimeca or some combination of chemical
and mechanical. Don, you owe me a drink. Steve Hathaway had the other
alternative. Steve Hathaway's suggested title was beneficiation (I'm chok-
ing on the word here), which is a term that has been applied to coal bene-
ficiation, and naturally has some applicability here. It is just that,
Steve, it lacks a little bit of zing. So I guess that what I should do is
extend the sweepstakes for another 20 minutes.
Session 3: Technical Obstacles. Here a big duality was pointed out. Again
this goes back to the summary remarks that were made in the beginning.
There's a clear need for data on existing equipment, yet there's a need for
better unit operations, better systems configurations. I think that at that
point we came out very strongly during that session. As Harvey put it, any-
thing we can do to reduce technical obstacles and risks will be helpful, and
that seems to me to sum up that session fairly well. It was notable that
until the end of the session, people were really a little bit unwilling to
list the major technical obstacles. I think that they came out by impli^-
cation during the course of the meeting, or during the course of that ses-
sion. Then at the end they came out in spades when we did that exercise.
There may be about eight to ten points; I'll go through these quickly. The
first, again, the need to accumulate data on process and unit operations
performance to better inform implementers as to the risks and uncertainties
of resource recovery came out very strongly. To go back to a point I made
in the beginning, the type of data differs for each of the different audi-
ences. The municipal manager has one perspective, the A&E firms another,
and the system designer a third. From that we have a bit of difficulty in
putting together an approach to get together the data that will be useful to
all parties. A second point is that the necessary information may be ob-
tained by studying either full-scale plants or separate unit operations, and
recognizing the concomitant difficulties of each, we were unwilling to say
which was the most beneficial approach, although I think that there may have
been a slight leaning towards separate unit operations. If I recall back to
the first session, there seemed to be some comment that EPA might have had a
better bigger bang for the buck if we had shaped our expenditures different-
ly over the years. Part of the responsibility of obtaining the data may be
relegated to suppliers through performance bonds and related mechanisms.
This was another point that was brought out. I guess that that would be as
an outgrowth of the bonding requirements. The fourth point, there is a need
to coordinate the gathering of data so that it may be interrelated and com-
pared later. This may be done by working toward some sort of standard
method of evaluation for each unit operation, or each class of unit opera-
tions. Another point is that once evaluative data are accumulated on a
comparative basis, these may be assembled into design handbooks. There was
a bit of discussion about design handbooks. There were at least two or
three design handbooks that were offered, one by EPA, I don't recall the
212
-------
others at this point. NCRR's, and there was a third. I think that people
felt that at this point we don't have an adequate design handbook that can
be used by everybody. Developing a better design handbook seemed to be an
objective that people would like to see pursued. Then Harvey ran his
Delphi exercise.
ALTER: Do you have any results?
BERG: I have some quick results here. We didn't have our computer working,
but thanks to Harry Freeman and Jim Smith, we came up with a quick summary
here. The number one item in terms of first place votes for high priority
work was Number 1—Determine Optimal Arrangements of Unit Equipment. The
topic that had the largest number of total votes for high priority was
Comparing the Performance of Various Types of Shredders. Other research
needs that received, I would say, the higher mentions were Emissions from
Processing Equipment, Evaluation of Other Potentially Applicable Equipment,
Determining the Effects of Pretrommeling (Number 15). So that's Numbers 1,
2, 4, 5, 15, and under economics, Number 36, Developing Effective Accounting
Methods; and Number 39, Determining Equipment Operating and Maintenance
Costs. There was one fence hanger, and that was Number 3, Determining Ef-
fects of MSW Characteristics on Processing. That had a moderately high
number of votes, putting it in the high priority category, and it was tied
for second place as being most mentioned for low priority. What can you
expect from 48 people? The most unpopular items were: In first place,
Determine the Effects of Magnetic Separation. As I said, tied for second
was The Effects of MSW Characteristics on Processing, Number 3, and also
Number 27, Evaluating MSW Segregation Prior to Processing. Other low lows
were Number 26, Evaluating Receiving Facilities; Number 19, Evaluating the
Effect of Drying on Combustion Characteristics of RDF; Number 16, Determin-
ing the Effects of Drving on Bacteria and Virus. Now if you looked at it
by the Roman numeral groupings, numbers I through XIII (this is an incomplete
compilation based on the ones that were easiest to pull out), there were 30
high priority ratings in Category I, the General area. Category XIII,
Economics, had the second largest number of votes. Category II, Shredders,
was mentioned third most frequently. Category IV, Air Classifiers, was
mentioned third most frequently, and Category IX, Storage and Retrieval, had
a high average, although not a high large number of total votes, perhaps be-
cause there was only one item in there. There were only five mentions, but
that was an average of five, which would put it up there. In Category V,
the Effects of Pretrommeling was a very high rated one, but the overall
topic, Screening, was not. Now there were three categories that stood out
as attracting the least interest in terms of research. Those were Category
X, Receiving Facilities; Category VI, which was actually mentioned more fre-
quently, that's Dryers; Category III, Magnetic Separators, which had the
highest average of all and the highest low rating of all; and Category VIII,
Conveyors. One other point on this session is that the immediate and short
term needs for data and unit operation comparisons seemed to dominate the
discussion, but wide recognition was given to the need for basic parametric
research with respect to theory, unit operations, and systems configurations.
213
-------
There was a major action suggested, perhaps it was the strongest action or
most strongly advocated action in terms of a priority for follow-up, and
that is the suggestion that we have four ad hoc groups to explore methods for
evaluating unit processes. Maybe we can come back to that one. That con-
cludes my summary of that session. Are there any comments? Harvey, you
can't say anything about my calling it a Delphi exercise.
ALTER: I thought all oracles of Delphi were false.
BERG: That's the reason why you weren't supposed to comment.
Session 4 scarcely needs summary, so I'll limit myself to five quick points.
The first is that there is a need to exercise all of the modes of R/D&D,
from pilot scale through full-scale demonstration; this seemed to be an
accepted point. Scale factors seemed to be extremely important to those
using the data produced by any research or test program. Third, there was
a dual preference expressed for gathering the process data for operations,
and that duality concerned the scale. There was a strong set of opinion
that we need to collect data on large-scale operations—30 to 50 tons per
hour at pilot. On the other hand there was also a strong suggestion that
we should be working more in the small-scale area. This gets back to a
point which was brought out strongly by Herb Hollander, and that is the
need, perhaps, to start looking at stokers and industrial size boilers as
a potentially large market for using the refuse-derived fuel. I think that
that was a view that was challenged, particularly by Mr. Fay. My feeling
is that refuse-derived fuel in densified form is new enough that people
just haven't really made up their minds about it; it will just take a while
before people are willing to say that the small-scale systems are the ulti-
mate target for refuse-derived fuel. It seemed that there was an advocacy
for more research in that area. The fourth point would be that many options
are being explored and none should be cut off. That seemed to be the feel-
ing. As Roger De Cesare pointed out, we might even want to resurrect some
older systems that haven't been explored as much recently to see whether
there is something we can learn from that. That goes back to developing
basic parametric data and also trying to continue our optimization work.
I guess the bottom line comment was the most disquieting of all to me. That
is that you didn't make my job easier or that of Don or the other government
representatives here. There just wasn't a clear consensus for how the gov-
ernment should go about conducting its research programs. So about the best
thing I could do is weakly plead for you to continue thinking about it and
to get in your written thoughts.
WARE: I think I would like to make a comment or two in this summary section
that has been overlooked, and that is the motivations behind pursuing any or
all of what we've discussed in these last two days, because there are bene-
fits for all, but not sufficient in a lot of cases to activate any of these
programs whether they be R&D definition, system development, equipment
improvement, or design. Ultimately, subject to the American way and human
motivation, there has to be dollars at the end of it. There are going to be
214
-------
different benefits for different people. The cities in a lot of cases
have to have these systems; they may well get revenue from them; the
architectural and engineering contractors will make their 10 percent by
pulling together technology which they have begged, borrowed, stolen, or
invented. The manufacturers of equipment will make a profit ultimately.
They have to be motivated in the first place to design the stuff. The per-
son or organization right now that stands to gain the most is the system
developer, to mention but a few Monsanto, Union Carbide, ourselves at Oxy;
and they are the only ones for now who are going to commit a tremendous
amount of corporate money to the development of processes. That's not to
say that cities don't have to come up with front money when they build
plants, but as far as pure R&D is concerned, the bulk of it is going to be
performed by the larger companies who have got the resources and the moti-
vation to do the work. So it's all very well and good for us to sit here
and define the things we would like to know more about. The fact remains
that the only people who are going to do anything are those that really want
it and will gain from it. That reflects our position and those of the other
major companies involved. There are some areas where I think the Federal
government should invest money, not to our benefit or to other companies'
benefit, but on general subjects. One of those is the discussion of disease
problems or potential problems associated with these plants. That wouldn't
invade proprietary areas, it wouldn't threaten or even interfere with
marketability of these systems, so that the Federal government can inde-
pendently be involved in that type of work and should be so. Ultimately,
systems design, unit operation, efficiencies of pieces of equipment will
rest, at least for the present, for the next few years, with the system
developers. I don't think you can overlook that. I recognize the value of
setting up ad hoc committees to develop comparative methods. ASTM type pro-
cedures would be fine. I wouldn't be surprised if the major system de-
velopers will comply. On the other hand, you can't make them, and they
don't have the motivation. You must understand that why should they define
how well their air classifier works now, when first of all they don't know,
and secondly they sure as hell don't want you to buy the other person's.
So there will remain a mystique, and until the system developers come up
with their own methods of evaluating them, you can't hope to do it as an
independent body. So getting back to what happens with R&D, which is, I
think the essential ingredient of this session, it will remain with the
people that have the money, the people that stand to gain the most from it.
That may in some cases be the cities; I would like to see more, but I recog-
nize the money really isn't there. It will principally be with the big
companies, and I hope the Federal government will recognize the advantage
of investing their funds in general subjects, particularly disease. My
last comment is that I had come to this session expecting something quite
different. I expected a lot of "hands on" practical people to throw out
their ideas on specific technologies on how well conveyors work, whether a
pan is better than a belt is better than a vibratory, etc., whether a
hammermill is better than a ring crusher, etc. It didn't happen. In that
respect I was disappointed; I think I may have misconceived what this whole
thing was about. I did enjoy meeting a number of people who have similar
215
-------
interests. I think it may be to our mutual advantage to set up specific
workshops where the goals are well recognized, where you have plant managers
and maintenance superintendents thrown together and forced to discuss
things, where you have policy makers, city government, corporate management
together to discuss the direction in which the industry needs to move.
There needs to be these types of interfaces. Right now they rest more in
the form of requests for proposals and on an individual basis. I think
that's where we ought to aim ourselves in the form of committees and further
meetings. Otherwise 1 have enjoyed meeting all of you; I've benefited to
those limited extents.
BERG: Are there other comments at this point? You got ahead of me a little;
I was going to get into future steps in just a moment. Maybe I could just
back up for a second and open a question, not offer any summary. We've
kicked around just about every aspect of the R/D&D business. Maybe we could
just go around the table quickly (of course everybody doesn't have to stick
in their two cents worth) but it would be very helpful, I think, to all of
us to get your bottom line statement: What should the government role be?
Peter's given us one view; it certainly has merit. If government is the
facilitator, what should that role be?
HASSELRIIS: A thought occurred to me which follows directly with your com-
ment. The one area that I think a lot of us feel we could use government
help is the government-created problem. If the government agency creates a
problem for industry, maybe it should help solve the problem that it created.
Some of the things that EPA brings up which relate to environmental problems
and health problems are created for the good of the public, the investiga-
tion is created for the good of the general public, and the solution is
perhaps a responsibility of whoever creates the problem for industry. For
instance, if you want a company that operates a boiler to stop polluting,
then you can also help that outfit to find a way to stop the polluting.
They're very complementary. If you raise the question of health hazards,
it's also an obligation to help solve the problem. You mentioned that the
health hazards is one of the stickiest problems, and that is probably why a
lot of us feel that it would be very helpful, but don't make the problem
worse at the same time as solving it. That is just a comment I couldn't
help.
BERG: Well, now, you throw out an option there that the government could
get involved in solving some of these problems. When the government helped
pull together the research that has really been going on outside of the
government and identified sulfur dioxide as a problem or sulfur oxides as
a problem, the government's response was to pour probably 50 million dollars
over the last 8 or 10 years into developing a flue gas cleaning system. One
way that the government could serve in the resource recovery area would be
to develop or to pour a fair amount of money into developing a base line
system for resource recovery. Now that wouldn't be a big hit with companies
like Oxy or others that have put a lot of money in so far, I don't think we
are at exactly in the same stage that we were 10 years ago in flue gas
216
-------
cleaning. That would certainly be one way. Another way to go would be for
the government just to do a holding action and try to help solve the spe-
cific problems that corae up in the course of developing various processes or
implementing them like the health research. I don't really think the govern-
ment creates the problems, I think the government is the group with the
social responsibility to help identify the problem and suggest social reme-
dies for those problems, or remedies for those social problems. I tend to
disagree with some of what you said there, but maybe that's one way that the
government could go. Suppose I lay that out on the table, that's a way of
being divisive. Maybe what the government should do is build a base line
resource recovery system that's better than anything that's out there. What
do you think of that? Do you want to try to comment? We should do this very
quickly.
LAMB: I think [unintelligible] evaluate, it probably would succeed.
SJOBERG: Environmental, energy, and raw material national problems have
created a resource recovery industry, and the government has to address it-
self to that problem. If you are going to have a resource recovery industry,
the industry that is going to be there requires an incentive as established
by the government. One of the things that the government has to address it-
self very diligently to, is what type of incentives are you going to provide
the municipalities, for power companies, for companies like American Can if
you stay in this business.
BERG: What incentives would you suggest?
SJOBERG: There are many. There are tax incentives; you have a whole spec-
trum of things,
BERG: Do you have a favorite though?
SJOBERG: Not specifically. I think here you have an open option because
it's not only an incentive that's applicable to American Can, but there are
incentives you have to apply to this municipality, to equipment vendors,
whomever. This is the kind of thing that's going to be necessary to moti-
vate this industry, because it takes money and it all can't be government
money. We've invested a lot of money already, all major companies have; it
takes a lot more, particularly if you are going to make it grow. A city
with several major plants now in work and there are many, many millions of
dollars invested in those, and for this industry to go there is going to
have to be a tremendous amount of money invested. You have to provide in-
centives for us to do it, for the cities to do it, the states to do it, for
the authorities to do.
WALTER: Well, I guess I should say that if we are not lying in our teeth
that these systems are marginally economic, i.e., there will be an income
to a city which will help reduce its cost of landfill, why should the Fed-
eral government provide money to do something that is going to make money?
Well, I'm just being divisive also. Really we are saying these systems are
not economic.
217
-------
SJOBERG: Maybe they are marginal. Maybe in situations they are not eco-
nomically feasible, but the environmental impacts require a mandate that we
go in this direction. Every industry that goes into this is gambling and
speculating on the cost of scrap, on the cost of fuel. It's a speculative
business; there is no guarantee by the municipalities or the government,
that we will get refunded for our investment.
WALTER: Well, I appreciate your comment and I was interested and pleased to
receive and reinforce something that we've been saying, that the market is
really not big enough to provide the incentive for business to go into these
systems. That's really where it's at. The market is too small.
SJOBERG: Not too small, it's speculative. In fact, it may be too big and
it takes too much money for any one company to go into it in a big way.
WALTER: Well, we would now have to spend a lot of time to define what terms
we are using.
SJOBERG: This is right.
BERG: Before I move into the last part here, Art Purcell has another new
name to throw into the sweepstakes. I'm not sure that I can pronounce it,
Art. Chemacaficiation. To go back to the thing that I just touched on
when I reviewed Session 3, it seems that perhaps the strongest recommendation
to come out of the meeting is that we should have all people here and others
that we may want to call in or who may volunteer become actively involved in
cooperative work designed to develop methodologies and mechanisms for shar-
ing technical information and compiling technical information, methodologies
for data gathering and actually gathering the data. Am I misstating,
Harvey?
ALTER: Close enough. Since this morning I have been talking to some of the
participants, particularly George Trezek and Herb Hollander. I think a
suitable mechanism for continuing this activity might be through the ASME.
and ASTM. I offer the good offices of the ASTM Committee to do this, and
indeed we might even call or write you for a meeting on this subject during
the next meeting of the E-38 Committee, which is the week of April 20 in
Philadelphia. The committee can provide the meeting room, visual aids, the
coffee, good fellowship, and there's no registration fee.
BERG: Are you tending to abandon the idea of the ad hoc committees to fol-
low up on this meeting?
ALTER: Well, some people have convinced me, we need the umbrella of a pro-
fessional-technical organization, to do this.
BERG: How do people react to that? I guess some of you expressed your
opinions to Harvey. Does that seem to be the best way to go? O.K. Are
there other recommendations that you'd like to make at this time?
218
-------
SPENCER: I am a little disappointed in some of the comments I heard people
say. It bothers me that there is no cohesive effort on the part of the
government to do a lot of this work, and at the same time companies are going
to invest a lot of money. I have to say that at least from where I sit (I'm
not necessarily speaking for my company,) but at least as an individual,
frankly, I don't think the field is developing as fast as I certainly
thought it was going to. I know myself that if it were my money I would be
a little bit hesitant to go after a lot of jobs where the cost of preparing
requests for proposals is very high, the probability of proceeding with the
given project is very low. Somebody in government ought to set the policies.
I see driblets of money going down lots of paths with much of it being
wasted. I see that on the one hand we've got an employment problem, we've
got public works funds that we are talking about, we've got money in ERDA,
we've got money in the Navy, some at the Bureau of Mines, some in EPA, and
it's all being done in a sort of small way, all being spent, but I'm not
sure that it is being spent very effectively. It's very disappointing to me
that there isn't any kind of a major effort like we had in other system de-
velopment within the government, where the government takes a major step and
really does something positive. I know that sounds kind of general, but
frankly, I think you have to start with the general policy as to are we go-
ing to try to conserve our resources, conserve energy, employ people, and
not pollute the environment. All of those things, if properly put together,
can move the industry forward rapidly. What I see happening is a government
effort, a congressional effort, whatever you want to call it, it is very
disappointing. I see a lack of cohesiveness, a lot of small efforts being
done, many of which I don't think have much value whatsover to attacking the
real problem, and an unwillingness to move the thing forward quickly. I'm
sure that resource recovery will all take care of itself in due time. I'm
sure that when it becomes economical that we will do additional things when
we can see it beginning to pay off in short term. Meanwhile, as far as I
can see, there are lots of other things that a company can invest in with a
lot higher returns on the investment and a lot lower risk than resource re-
covery. As far as I can tell, a lot of these things don't make sense and
it's not moving very quickly.
JOENSEN: I guess I would like to refine Mr. Spencer's comments, in that
money for research and energy development really comes from Congress, and
right now Congress is enthused with solar energy as a panacea of the
country's problems, or solutions to the country's problems. I think some-
how working through our Congress, which at least we are trying to do in
Iowa, that resource recovery has to be really advertised more and given more
emphasis. So when we talk about government funding, I think we've come back
to Congress and not government agencies, because they submit budgets like
everybody else and someone goes down and arbitrarily ticks off and strikes
out line items. So we have to work through our members of Congress. They're
the people who ultimately say where we should spend our money.
BERG: Are there any other recommendations people would like to bring up at
this point?
219
-------
RYDER: From our standpoint when we went to build our plant (somewhere
around 20 million dollars), we thought perhaps this was an area that the
Federal government would help us out on. However, we never did submit a
formal request for a grant because we never got that far. We were told
that funds were not available for a process that's already been demonstrated
and that works and therefore "We're not going to help you build a plant.
However if you experiment on something unproven, we'll be happy to work with
you." Our plant was considered a scaled-up version of the St. Louis plant
and therefore funds were not available because the initial idea had already
been demonstrated. We had to float a bond issue to get the money for our
own plant. I think maybe Chicago is one of the few cities perhaps that can
do that. There are many, many cities that can't do that. This is where the
Federal people ought to spend some money; more for ongoing processes. It
appears that once the subject is researched nobody has money to put it into
practice. In this report, if one has to choose, money for research should
be minimal.
BERG: Let me just ask, Bill, one question. Suppose the government doesn't
have enough money to do a demonstration grant in every city around the
country, what percentage support or what other instrument would have been
most useful in your eyes for the government to do?
RYDER: Well actually, if you are going to build a plant that is going to
cost 20 million dollars, I would like to have had the Federal people give me
20 million dollars.
BERG: All right, short of that?
RYDER: It would depend on the circumstances. Chicago obviously was able to
get by with nothing. Other communities have nothing, but would need the en-
tire 20 million dollars or perhaps something in between.
SJOBERG: Loan guarantees and things of this nature where the government
will guarantee the bonding of the municipalities which have a nebulus bond-
ing capability is one way. There's many economic facets that you can take
relative to this. I think one of the things we found is that there are
divergent viewpoints between EPA and the energy people as to whether this is
viable or not. I think somehow or other you have to attack not just that
specific issue, but all of the different potential ways of doing it so that
the government agencies, when the legislators come to you and ask you what
do you recommend, you will at least have a concentrated effort behind it,
realizing it's mandatory that you have some stand in order to get this in-
dustry off the dime. You can research all you want, but if the financing
isn't there to make it move ahead, there will be no movement in the industry.
BERG: Then your preferred financing option would be bond guarantee?
SJOBERG: 1 don't know what it is, but it is one of the options. What I am
saying is, though, there should be a uniform front between the government
agencies that you've looked at. Then you can advise your legislators because
they come to you. Because it's your bills that are put up—from ERDA and EPA.
220
-------
WALTER: Just by a show of hands, how many of you feel that a loan or bond
guarantee is valuable and useful to your city, assuming that it has any kind
of thing on it you want, the bonds are still municipal bonds, they are still
tax free, and anything else you want? How many of you think a loan guaran-
tee is useful to you?
CHANTLAND: What's the interest rate?
WALTER: At whatever interest rate falls out. Presumably when you get the
full faith and credit of the Federal government behind it, the interest rate
is going to be fairly low, the low fives, maybe even the high fours. I see
no hands from any city.
SPENCER: I don't think that loan guarantees are such a good idea. I think
grants are great. In the case of Monroe County, the State chipped in 15
million dollars and sweetened the pot with an additional 6 to 8 million,
which out of a 50 million dollar program, is a pretty good lump of money,
probably more than the Federal government put out. I think that in terms
of policy again, nobody has made the policy decision that we want to move
in this direction and set the resources to do it quickly. Furthermore, I
think that there's an attitude on the part of the agencies that this is a
local problem, that it's not a Federal problem, and therefore it doesn't
make sense to invest Federal dollars to solve a local problem. I think
that you are now reaping what you have been preaching, and that's exactly
right, now that you've said that there are no Federal dollars available to
help municipalities in solving their problems. There is no question that
the problem in one municipality, while the solution may be different, the
problems are pretty much the same with respect to solid waste, and that
there's some commonality and some savings that all communities can gain by
a certain amount of Federal investment. When you are talking about dollars
where the total Federal expenditure is less than what this one customer has
to put out, the City of Chicago and Monroe County, New York, when the Fed-
eral level of investment is lower than one single project, what have you
got?
ALTER: Listening to all this, I hear no return in the discussions to the
demand-pull; what are we doing is pushing on a string. Consider for a
moment—-and I'm really putting this as a question, particularly to the
utilities—what if there were a demand-pull such as being indemnified
against damage to a boiler. Would this increase demand-pull? The utility
says, all right, I can breathe easier now; the risk to my capital plant is
insured. Now I want the fuel. Or is there some other demand-pull neces-
sary, such a a favorable return on an investment, for using this material?
Or both? Or a third factor? If then, the market was shored up some way,
would not it be easier for a city to raise capital and wouldn't it be
easier for the private sector to have a business?
FAY: I think there are some things that could be done. I have preached
this to EPA every time they give me an opening, and that is all you got to
221
-------
do is tell the utility that you will waive all initial limits if they fire
garbage in that boiler, and I think you will have them beating off your
door. Of course the reply is a lot of them ought to do that. Your point
on idemnifying boiler risks, I think could ease a lot of the utilities into
the position of accepting it easier than they will today. There are pro-
bably some other things, but as soon as you start saying that you're going
to make money on this, it doesn't quite happen that way very easily due to
fuel clauses and some of these other things. I don't think you really need
to tell the utility they are going to make money on it, protect them from
the risk.
RIGO: The reason that the Navy, actually DoD, is principally interested in
solving a 100 ton a day problem, is they have a minor difficulty called
garbage disposal. They don't have any place to put it on the bases, the
surrounding community normally or frequently does not have a large garbage
processor, they want nothing to do with your garbage, they've also got a
problem called they've got to get rid of the garbage. At that point,
whether it costs you $5 a ton, or $10, even $15 a ton becomes quite im-
material. Now this is the situation that's developing nationwide. Land-
fills are being outlawed, incinerators are being shut down, something is
going to have to be done. Economics may not be there for private industry
today, but can anyone really say, especially near large urban areas, since
Ohio won't let you landfill in their strip mines from the East Coast, there
isn't going to be an incentive out there in the not too terribly distant
future. We are looking at some really incredible tipping fees. They are
beginning to out in Alameda, California. They are paying something like
23 dollars a ton to unload their garbage into a municipal system. Now with
that kind of incentive, you can really get excited about doing something.
Do we need any other incentive?
LAMB: I'm really against artificial stimulation by cash injection. I think
grant programs consume lots of money and really show marginal benefits.
This system won't work until it is the right thing to do, and when it is the
right thing to do it will work fine, and we are pretty close to it being the
right thing to do. So I'm very strongly against things like grant programs
and demonstration programs and the building of large demonstration facili-
ties. I really don't think they are necessary, and I think they are already
too late. That is, there will be no useful data generated for the people
who in the next two years will shake this thing out. So I think that tax
credits or loan guarantees may be ways that may be necessary, I'm not even
sure they are necessary. I really would like to strike a vote against
large government, artificial stimulation of the market, with large govern-
ment checks.
BERG: I want to say just one last thing. I think that there are quite a
few people who deserve a vote of thanks. First of all, John Burckle, who
put in the most work of all, I think, in organizing this whole thing, in
getting in contact with you people; Midwest Research for putting together
222
-------
that fine backup document; and the Franklins for all of their help and
support here. Each of the session chairmen also deserve a hand with
the possible exception of myself; and New Orleans has been an excellent
host. In advance I'd like to thank NCRR for the trip to the facility
tomorrow, and the City, and Waste Management. Most of all, I'd like to
thank everybody who came, took the interest, and put in a good bit of
work, I think. I appreciate it and I think everybody here deserves to
applaud one another.
223
-------
APPENDIX A
WORKSHOP
ON
THE PREPARATION OF FUELS AND FEEDSTOCKS
FROM MUNICIPAL SOLID WASTE
AGENDA
February 8
February 9
February 10
Assembly
Welcome Address
Ratification of Agenda
SESSION 1: Systems Process Design
Lunch Break
SESSION 2: Selection of Equipment
Assembly
SESSION 3: Technical Obstacles
Lunch Break
SESSION 4: Approach to R/D&D Programs
SESSION 5: Summary and Conclusions
Assembly
Departure
Site Visit
Expected Arrival at Airport
9:00
9:05
9:15
9:30
12:30
2:05
9:00
9:05
12:30
2:05
5:00
8:15
8:30
9:15
1:00
- 9:15
- 9:30
- 12:30
-2:00
-6:00
- 12:30
- 2:00
- 5:00
-6:00
- 12:00
224
-------
SESSION 1: Systems Process Design
Objective
The objective is to identify the critical aspects of process design and
determine what information is needed to permit 1) more effective selection
of systems and, 2) the design of more efficient systems.
Discussion
The following questions are considered of prime importance in judging the
adequacy of our abilities to select and design preprocessing systems.
1. With the technology available today, can we produce a suitable
product which the consumer considers a useful and desirable fuel?
Are specifications for the fuel product available?
2. Is adequate information available to the implementor/investor for
selection from competing commercial and developmental systems?
3. Is there an alternative process concept offering a potential of
significant improvements in product quality, economic performance,
or both over the present practice of primary shred and air
classification?
4. Is adequate information available to the system supplier or con-
sulting engineer to perform the process design with sufficient
confidence to offer performance guarantees?
We suggest discussion of these issues with respect to the following topics:
a. Product quality requirements and quality control, performance
guarantees
b. System reliability guarantees
c. Selection and sequencing of unit operations, including interactions
on product quality and cost
d. Energy efficiency
e. Economics, including the importance of materials recovery
f. Ability to calculate adequate heat and material balances for
process design
g. Adequacy of present designs and information base for selection of
competing systems
-------
SESSION 2: Selection of Equipment
Objective
The objective is to identify the critical aspects of equipment selection and
determine what additional information is needed for the selection process.
Discussion
Identify and discuss the adequacy of equipment and information for equipment
selection when applied to refuse processing such as: 1) equipment perform-
ance data; 2) experience with existing equipment; 3) capital and operating
(including maintenance) costs; 4) guarantees; 5) factors of scale (i.e.,
scale-up from research to full-scale systems; 6) effects of changes in in-
coming wastes, i.e., municipal, agricultural, and forestry; 7) large city vs.
small town/rural applications. Where is there potential for improvements in
equipment performance, and where are improvements most needed?
1. Waste receiving, storage and retrieval
2. Size reduction and size classification (hammermills, trommels,
screens, etc.)
3. Light/heavy fraction separation (air classifiers)
4. Thermal and chemical treatment of cellulose for fuel production
5. Fuel densification (pellet mills, cubetters, briquetters)
6. Solids handling equipment
7. Fuel handling, storage, retrieval, and transport
8. Environmental control systems
9. Process control systems
SESSION 3: Technical Obstacles
Objective
The objective is to identify the technical areas of greatest need for re-
search, development, and demonstration.
Discussion
Based upon the discussions in Sessions 1 and 2, the workshop members will
prepare a listing of the research and development required to overcome
technical barriers to implementation.
226
-------
SESSION 4: Approach to Research, Development, and Demonstration Programs
Objective
The objective is to identify valid approaches to conducting R/D&D, to identi-
fy the appropriate role of the Federal government in sponsoring R/D&D, and
to discuss the existing resources and capabilities of organizations con-
ducting R/D&D.
Discussion
Given the R/D&D needs identified in Session 3, address the following aspects
of past and future R/D&D programs:
1. The utility of information from government-sponsored R/D&D
activities
2. The suitability of approaches for, a) process and equipment de-
velopment and for, b) process and equipment performance evaluation
considering the pilot plant, the field test of existing systems,
development and demonstration of sub-systems at full or reduced
scale, development and demonstration of a complete system inte-
grated into a community solid waste management system
3. Capabilities of the private sector
a. NCRR pilot plant
b. Others
4. Capabilities of existing publicly-sponsored pilot plants
a. Bureau of Mines
b. University of California - Berkeley
c. Department of Defense (planned)
d. Department of Environment - Ontario
e. Warren Springs Laboratory - England
5. Potential for R/D&D with limited Federal action or support, con-
sidering rate of implementation, for four types of actions:
a. a regulatory program based upon environmental protection and
resource conservation with no financial support
b. full-scale development and demonstration of systems integrated
into the community waste management system
c. limited scale development and demonstration of subsystems
d. R/D support only
227
-------
APPENDIX B
WORKSHOP
ON
THE PREPARATION OF FUEL AND FEEDSTOCKS
FROM MUNICIPAL SOLID WASTE
PARTICIPANTS
Dr. Harvey Alter
National Center for Resource Recovery
1211 Connecticut Avenue, N.W.
Washington, D.C. 20036
Mr. David Bendersky
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Mr. David Berg - RD-681
401 M Street, S.W.
Washington, D.C. 20460
Mr. John Burckle
U.S. Environmental Protection Agency
26 West St. Clair Street
Cincinnati, Ohio 45268
Mr. Richard Bush
Vice President for Engineering & Operations
Connecticut Resources Recovery Authority
Hartford, Connecticut 06106
Mr. Peter J. Cambourelis
Raytheon Service Company
12 Second Avenue
Burlington, Massachusetts 01803
Mr. Arlie Capps
Stanford Research Institute
Menlo Park, California 94025
Mr. Arnold 0. Chantland, P.E.
Public Works Department
City of Ames
Ames, Iowa 50010
228
-------
Mr. G. C. Chisamore, P. Eng.
Manager, Program Development Section
Ontario Centre for Resource Recovery
Ministry of the Environment
4375 Chesswood Drive
Toronto, Ontario, M3J2C2, Canada
Mr. Roger S. De Cesare
College Park Metallurgy Research Center
Bureau of Mines
College Park, Maryland 20740
Mr. C. W. Fay
Wisconsin Electric Power Company
231 West Michigan Street
Milwaukee, Wisconsin 53201
Mr. Douglas E. Fiscus
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Mr. William E. Franklin
Franklin Associates, Ltd.
8340 Mission Road - Suite 101
Prairie Village, Kansas 66206
Mrs. Marjorie A. Franklin
Franklin Associates, Ltd.
8340 Mission Road - Suite 101
Prairie Village, Kansas 66206
Mr. Harry Freeman
U.S. Environmental Protection Agency
5555 Ridge Avenue
Cincinnati, Ohio 45213
Mr. Robert Freeman
Stanford Research Institute
Menlo Park, California 94025
Mr. Harvey Funk
Henningson, Durham & Richardson
8404 Indian Hills Drive
Omaha, Nebraska 68114
Mr. Floyd Hasselriis
Combustion Equipment Associates, Inc.
555 Madison Avenue
New York, New York 10022
229
-------
Mr. Steve Hathaway
Office of Chief Engineer
Department of the Army
DAEN-ROM
Washington, D.C. 20314
Mr. Herbert Hollander
Gilbert Associates
P.O. Box 1498
Reading, Pennsylvania 19603
Mr. Robert Holloway - AW-462
401 M Street, S.W.
Washington, D.C. 20460
Mr. Steve Hurley
U.S. Navy
Facilities Engineering Command
Alexandria, Virginia 22332
Dr. Alfred Joensen
Professor, Mechanical Engineering
Iowa State University
Ames, Iowa 50010
Mr. Dan Keyes
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Dr. Albert J. Klee
U.S. Environmental Protection Agency
26 West St. Clair Street
Cincinnati, Ohio 45268
Mr. Patrick J. Koloski, Secretary
Department of Sanitation
City Hall Room 2W13
1300 Perdido
New Orleans, Louisiana 70112
Mr. Ron Lalka
County Sanitation Districts of Los Angeles County
P.O. Box 4998
Whittier, California 90607
Mr. Tom Lamb
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02104
230
-------
Mr. Steven J. Levy - AW-462
401 M Street, S.W.
Washington, D.C. 20460
Mr. Stephan Lingle - AW-463
Chief, Technology & Markets Branch
401 M Street, S.W.
Washington, D.C. 20460
Mr. Robert A. Olexsey
Industrial Environmental Research Laboratory
26 West St. Clair Street
Cincinnati, Ohio 45268
Captain R. F. Olfenbuttel
Department of the Air Force
Pentagon, AF/PREVP
Washington, D.C. 20330
Dr. Thomas Padden - RD-682
401 M Street, S.W.
Washington, D.C. 20460
Mr. Bill Parker
National Center for Resource Recovery
1211 Connecticut Avenue, N.W.
Washington, D.C, 20036
Mr. Robert Powers
Recovery 1
17000 Chef Menteur Highway
New Orleans, Louisiana 70129
Dr. Arthur H. Purcell
Technical Information Project
1346 Connecticut Avenue, N.W., Suite 217
Washington, D.C. 20036
Dr. Greg Rigo
Systems Technology Corporation
245 North Valley Road
Xenia, Ohio 45385
Mr. William C. Ryder
Chief Environmental Design Engineer
Department of Engineering
City of Chicago
320 North Clark Street
Chicago, Illinois 60610
231
-------
Mr. M. P. Schrag
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Dr. L. T. Shannon
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Mr. H. T. D. Sjoberg
Americology
American Can Company
American Lane
Greenwich, Connecticut 06830
Dr. James E. Smith, Jr.
Environmental Research Center
U.S. Environmental Protection Agency
Technology Transfer
26 West St. Glair Street
Cincinnai, Ohio 45268
Mr. David Spencer
Raytheon Service Company
12 Second Avenue
Burlington, Massachusetts 01803
Dr. George J. Trezek, M.E.
College of Engineering
University of California
Berkeley, California 94720
Mr. Donald K. Walter
Chief, Urban Waste Technology Branch
Office of Conservation
Energy Research and Development Administration
Washington, D.C. 20545
Mr. Peter J. Ware
Occidental Research Corporation
1855 Carrion Road
La Verne, California 91750
Mr. Carlton Wiles
U.S. Environmental Protection Agency
26 West St. Glair Street
Cincinnati, Ohio 45268
Mr. F. E. Wisely
Horner & Shifrin, Inc.
5200 Oakland Avenue
St. Louis, Missouri 63110
232
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPCR r I\O
EPA-6CC/9-77-034
4. TIT! & /-i\C S JB1 ' TLf
THL PREPARATION OF FUELS AND FEEDSTOCKS FROM
MUNICIPAL COL ID WASTE
6. PERFORMING ORGANIZATION CODE
•S)
8, PERFORMING ORGANIZATION REPORT NO.
Marjorie A. Franklin, Editor
1MINC _. . _ . ...
hranklin Associates, Ltd.
8340 Mission Road
Prairie Village, Kansas 66206
12. SPONSORING AGt'-jC1' MAME AND ADDRESS
Municipal Environmental Research Laboratory--Cin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
3. RECIPIENT'S ACCESSI ON-NO.
5 REPORT DATE
December 1977 (Issuing Date)
10. PROGRAM ELEMENT NO.
EHE624C
11. CONTRACT/GRANT NO.
CA-7-2431-A
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Proceedings of U.S. EPA Workshop held at the Braniff Place Hotel, New Orleans,
_Louisiaria,_ on February 8-10, 1977
16. ABSTRACT
This report is a record of the proceedings from a workshop held in New Orleans
Louisiana, on February 8-10, 1977, among a group of experts in the field of
processing municipal solid waste into useable fuels and other recyclable
materials. The bojv of the report consists of a verbatim discussion among
meeting participants. These sections were transcribed from the tape-recorded
meeting sessions, in addition, a summary of all discussions and conclusions
is included. 1 he report serves as a comprehensive, up to date, state-of-the-art
summary for municipal soiid waste processing technology which should be useful
to designers and researchers in the field of solid waste management.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Waste Disposal
Waste Treatment
Natural Resources
Processing
Circulation
Recovery
b.IDENTIFIERS/OPEN ENDED TERMS
Solid Waste Processing
Resource Recovery
Shredding
Components Separation
Recycling
c. COSATI Field/Group
13 B
18. DISTRIBUTION STATEMEN1
Release to Public
19 SECURITY CLASS (This Report)
unclassified
21. NO. OF PAGES
241
20 SECURITY CLASS (This page)
unclassified
22. PRICE
EPA Form 2220-1 (9-73)
233
tf US GOVERNMENT PRINTS OFFICE 1978— 757-140/6655
------- |