United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-188 Oct. 1981
Project Summary
Densification of Refuse-
Derived Fuels: Preparation,
Properties and Systems for
Small Communities
Jay Campbell, Marc L Renard, and Edward J. Winter
Densified refuse-derived fuel (d-
RDF) is produced by compacting
refuse derived fuel (RDF) into ag-
glomerated pieces sufficiently cohesive
to sustain storage and handling. The
use of this d-RDF product as a
substitute for coal in spreader-stoker
boilers is a developing resource
recovery alternative.
The study summarized here investi-
gated the operation, performance.
and product characteristics of a waste
shredding and densification subsystem
for producing d-RDF from the air-
classified light fraction of municipal
solid waste. The study also provided a
technical and economic evaluation of
d-RDF facilities for small communities.
During the investigation, a pellet
mill was used to produce nearly 1300
Mg (megagram) (1442 tons) of pellets
that were later test burned. Treated in
the project report are the performance
and effects on densification of modifi-
cations to the feed preparation equip-
ment, particularly the secondary
shredder. Densifier operation and
maintenance are discussed as they are
affected by feedstock properties,
feedrate control, equipment wear on
pellet production and properties.
Physical properties of pellets analyzed
over the 1300-Mg (1442-ton) produc-
tion period are summarized. Observa-
tions on handling and storage are
discussed and recommendations are
made on minimizing potential prob-
lems.
This Project Summary was devel-
oped by EPA's Municipal Environmen-
tal Research Laboratory, Cincinnati,
OH, to announce key findings of the
research project that is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back).
Background
The U.S. Environmental Protection
Agency (EPA) assigned the Municipal
Environmental Research Laboratory
(MERL) in Cincinnati, Ohio, major
responsibility for research and develop-
ment in the field of recovery and use of
municipal solid waste. One concept
investigated involves the recovery of
energy from solid waste. Refuse is
combusted either directly for steam
recovery or in combination with fossil
fuels for power generation. The latter
involves processing the refuse to
remove the combustibles for use in a
modified power generation boiler,
usually in combination with coal. The
processed refuse is usually referred to
as refuse derived fuel (RDF).
The RDF concept in the United States
has generally been limited to power
generating facilities that burn pulverized
coal. The use of RDF need not be limited
to large users, however; in fact, RDF
-------
may be more valuable to small power
generating facilities. Small industrial
and institutional boiler owners may find
RDF an attractive and cheaper alterna-
tive to fossil fuels, for which they
receive no quantity discounts, as do the
large users. In addition, small users may
have increased flexibility in negotiating
contracts for RDF (especially with
regard to length of commitment). Many
small power generators are economically
marginal because their boiler facilities
are older, coal-burning models that
require costly air pollution equipment.
The use of RDF may help such facilities
absorb the cost for such controls.
RDF prepared for large utility boilers
is typically the light fraction of shredded
refuse that has been air-classified,
screened, or otherwise processed to
remove the noncombustibles. In this
fluffy form, it can be pneumatically fed
into the suspension utility boiler.
Smaller industrial and institutional
stoker-fired boilers, in the range of
11,000 to 90,000 kg (25,000 to 200,000
Ib) of steam per hour, are particularly
attractive for smaller communities. To
maintain storage and feed system
capacities and to ensure proper dis-
tribution and burning on the grates, an
RDF product with a range of properties
comparable to conventional stoker
(particle) coal is necessary. Rather than
a low-density fluff RDF, a densified form
of refuse-derived fuel (d-RDF) would be
more suitable.
This d-RDF may approximate the
physical characteristics of the stoker
coal fed to the boiler. RDF in this form
offers increased flexibility in transport,
handling, and storage, and it can be
mixed directly with the coal and fed to
the boiler with few, if any, modifications.
Although considerable experience
was available for cofiring RDF and coal,
little information was available on the
production and burning of d-RDF. EPA
therefore implemented parallel pro-
grams to (1) determine the engineering
and economic aspects of preparing d-
RDF and (2) assess the technical and
environmental implications of using d-
RDF as a coal substitute.
In addition to the report summarized
here, the following reports have been
prepared as part of these programs: (1)
"Coal:d-RDF Demonstration Tests in an
Industrial Spreader Stoker Boiler," (2)
"Fundamental Considerations for Pre-
paring Densified Refuse Derived Fuels,"
and (3) "A Field Test Using Coal:d-RDF
Blends in Spreader Stoker-Fired Boilers,"
EPA-600/2-80-095.
Materials and Methods
The work described in this report was
conducted at NCRR's Equipment Test
and Evaluation Facility (ETEF) located in
what was the oversize, bulky, waste-
processing area of the District of
Columbia's Solid Waste Reduction Unit
No. 1, a large, operating incinerator.
Packer trucks, containing municipal
solid waste (MSW), were diverted from
the incinerator tipping floor and dumped
at the test site to provide feedstock for
investigations.
Process Description
Densified refuse derived fuel (d-RDF)
is formed by the mechanical compaction
of a processed waste fraction into
particles. Equipment to compact or
densify the waste includes pelletizers,
cubers, extruders, and briquetters. In
this study, processing of the refuse
before densification included primary
shredding, screening, air classification,
and secondary shredding. Primary
shredding of the refuse was done in a
750 Kw (1000 hp) horizontal hammer-
mill located at the NCRR test facility.
After primary shredding, two process
configurations were Used to produce
the d-RDF. The first configuration
included a screen, a Heil-Tollemache*
shredder (secondary size reduction), the
densifier (California Pellet Mill), and
auxiliary equipment such as a cyclone,
conveyors, and feeders. This first
configuration was later modified to
include a vibratory feeder to the air
classifier and a screen deck added to the
feeder to remove inorganic particulate.
materials. This configuration (Figure 1)
permitted investigating different air
classifiers and other means to improve
the d-RDF product.
The air classified light fraction was
de-entrained in a cyclone, reintroduced
through a rotary air lock into a pneumatic
conveyor, and then discharged onto the
5 mm (3/16 inch) screen or into the
Heil-Tollemache shredder (configura-
tion 2). After the secondary shredder
further reduced the size, the material
was conveyed to the live-bottom bin.
The live-bottom bin helps control surges
and meters the material through two
screw feeders to the densifier. The d-
RDF product was discharged by conveyor
to containers.
These processes were used to investi-
gate the production of d-RDF and to
•Mention of trade names or commercial products
does not constitute endorsement or recommenda-
tion for use
produce the d-RDF for combustion tests
conducted at other facilities. Investiga-
tions conducted included evaluating
screening, air classification, secondary
shredding, pellet mill operation and
maintenance, and product character-
istics.
Observations
The investigations of screening and
air classification were part of a concur-
rent EPA program. Only results ap-
plicable to d-RDF production are
presented.
Screening
After primary shredding, the particle
size of 30.4 percent of the refuse was
less than 6.4 mm (1/4 inch). The
significance of this fraction is that it
contains primarily inorganic particulate
material that is likely to "fly" with the
light fraction when air-classified. These
removable inorganics fines (RIF) will
add to the ash, or noncombustible,
portion of the fuel and lead to increased
abrasion and wear on the secondary
shredder and densifier.
The use of vibrating screens to
remove the noncombustible shredded
waste and the air-classified light
fraction was studied. Flat screening of
municipal solid waste (MSW) cannot be
recommended; a better method was
found to be rotary screening.
Air Classification
Two different air classifiers were
used at ETEF: aTriple/SVibrolutriator™
and a vertical zig-zag-type classifier.
Factors studied included separating air
velocity, light/heavy split, and recovery
of specific seed materials. Both produced
the same light fraction yield, but the zig-
zag classifier required a lower air
velocity, indicating that the carryover of
RIF would be lower than with the
Vibrolutriator™.
Secondary Shredding
To decrease wear on the densifier die
and roller assemblies and reduce size
and milling action, the primary-shredded,
air-classified, light fraction must be
shredded a second time. The secondary
shredder was a vertical-shaft, reversible,
Heil-Tollemache hammermill. When
first operated, large pieces of textiles
and plastics passed through the shredder
and jammed the densifier or wrapped
around the augers. A number of changes
were made in hammer styles and
configurations. Fixed stater bars, which
created a scissor action with the
-------
Shredded MSW
6 mm
(1/4")
Screen
Mostly
Inorganic
Fines
Discharge to
Dust Room
Cyclone
Heil-
Tollemache
Shredder
Figure 1.
d-RDF
Product
Portion of Equipment Test and Evaluation Facility lor preparation of
d-RDF, configuration 2.
hammers, were also added at various
locations in the grinding section. Five
trials varying hammer number (up to 72)
and length were made, and the particle
size of paper and plastics was sig-
nificantly reduced. Textile size reduc-
tion did not follow this trend, however.
An alternative to hammermills for
secondary shredding is a rotary-knife
shredder. Although such a device does
offer the benefits of pure shearing
action as well as grates to control
particle size, it is susceptible to da mage
from tramp metal (tools and hardware)
and high wear from abrasive feedstocks.
In the test, a model 24 Knife Granulator
was used. A single test comparison of
product size from the knife shredder and
hammermill is difficult because of the
effect of throughput and wear and the
natural variability of such test results.
Knife shredders do, however, clearly
limit the larger particles.
Densification
In these investigations, a California
Pellet Mill (CPM) Model B162 ring
extrusion type densifier was used.
Auxiliary equipment included a live-
bottom bin and a screw feed mixer and
conditioner. As material is fed into the
densifier die cavity, it is flattened
against the die by the rollers. Layer after
layer builds up until the material is
forced radially through the die holes. A
tapered hole inlet provides some lateral
compaction, while the resistance from
the taper further compacts the pellet
along its axis.
The pellets are extruded from the
outside of the die and break in random
lengths, either from the centrifugal
force of the die, by contact with other
pellets, or by striking the pellet bar.
Pellet production rates of 7 to 9 Mg
per hour (7.8 to 9.9 tons per hour) were
expected but never realized. There is no
clear explanation for this because a
number of factors influence the per-
formance and capacity of the pellet mill.
These factors include: equipment
configuration, die and roller condition,
stability of feedrate, feedstock properties
(e.g., moisture, density, and particle
size) and presence of oversized textiles.
The interrelation of these and other
factors affected the performance of the
pellet mill to produce d-RDF. Because of
this, the following discussion does not
treat each factor individually, but sum-
marizes general observations about the
densification process.
During early operations, there were
frequent blockages and downtime.
These were evenly attributed to the poor
performance of feed preparation equip-
ment and the densifier. Other factors
such as operator experience also con-
tributed to these problems. Controlling
the flowrate of the feed proved to be an
important factor affecting the pellet mill
performance. Results indicated that an
improved system would include added
surge capacity to control fluctuations
from the shredder and the air classifier
and a system to control volumetric
feedrate. The condition of dies and
rollers proved to be an important factor
affecting throughput. Although new
dies and rollers and reconditioned dies
improved pellet production, the improve-
ment was short lived. The abrasive
nature of the feedstock dulled the sharp
-------
edges after less than 45 Mg (49.6 tons),
and the production capacity decreased.
Feedstocks more consistant and denser
than processed MSW, such as feed and
wood waste, have not caused the
jamming and surging problems experi-
enced with the MSW. This indicates
that process modifications to improve
the homogeneity of the MSW feedstock
would improve the densifier capacity.
Many equipment modifications were
made to improve production. Increasing
motor size of the densifier to reduce
jamming is not a solution. Improve-
ments in the feeder appeared to produce
better results. Continuing equipment
adjustments were necessary, especially
to the roller position.
Power consumption measurements
in these experiments were confused
because of errors in meter connections.
Subsequent investigations indicated a
higher than expected power consump-
tion, perhaps related to differences in
feedstock. Production of 1-inch pellets
(compared to 1 /2 inch pellets) indicated
that increased reliability and decreased
feedstock processing requirements
could be gained. Additional investiga-
tions are required.
Wear on die and roller shells appeared
to be the most significant maintenance
requirement. Wear appeared to be
similar to other feedstocks (e.g., feed)
but occurred much sooner, probably
because of inorganic fines. With respect
to roller wear, techniques available to
reduce abrasive fines in the feedstock
hold promise of extending the roll and
die life significantly.
Die and roller wear significantly
affected machine performance and
resulted in increased blockages, de-
creased pellet densities, and increased
maintenance requirements. Excessive
wear is attributed to the ash content
(inorganic fines) of the feedstock. Based
upon results of this work, it is projected
that a die life of 3,200 Mg (3,550 tons)
cover be expected for a low ash feed-
stock (10 to 12 percent ash). Subsequent
economic analyses are based on these
estimates.
d-RDF (Pellet) Properties
In the study, some physical properties
of pellets chosen for investigation were
moisture content, ash, and pellet
density. These properties were chosen
because of their probable relationship to
the use of d-RDF as a supplement to coal
in stoker-fired boilers. The densification
process results in a 5 percent loss of
moisture in the feed, and preferred
pellet moisture is in the range of 15 to
20 percent. Since the expected ash
content of d-RDF is in the range of 6 to 8
percent, the ash contents in excess of
20 percent experienced in this work was
caused by the inability to remove the
inorganic fines. Low ash contents are
desirable to reduce shredder and
densifier wear and to maximize heat
value. Pellet density is important to
ensure proper flow through storage and
feed systems. Moisture below 15
percent produced higher density pellets,
whereas moisture above 30 percent
results in poorlyformed pellets. Die hole
inlet taper appeared to be the major
controller of pellet density; the more
severe the taper, the more dense the
pellet. Information of pellet bulk density,
dimensions, integrity (ability to maintain
stable dimensions) and similar materials
handling characteristics are discussed
in the full report.
It was necessary to store the pellets
outside for close to a year until shipped
for combustion tests. This afforded the
opportunity to monitor the effects of
long storage and handling on the d-RDF.
Results of the monitoring indicated that
the pelletized form of d-RDF makes it
less attractive as a food source. Although
a variety of crawling insects were found,
they were no greater in numbers or type
than those found in ordinary soil. Odors
generated by the stored d-RDF were
objectionable only when the covers
were removed or the piles excavated.
While covered, dusting from the piles
posed no problem, but during handling,
flakes blown from dry, broken pellets
created minor offsite dusting.
At the end of the first summer, after
10 months of storage, smoldering
seams of pellets with temperatures
above 150°C (300°F), hot enough to
ignite the plastic covers, surfaced
where the pellets had been repiled. Four
times such outbreaks occurred, and 500
tons of pellets were lost. Composting
was believed to trigger an oxidation
reaction that caused the smolderings.
Recommendations based upon these
and other observations were made to
improve storage. These are to use a
ventilated cover, control length of
storage to no more than 8 weeks, and to
restrict handling of the stored pellets.
Cost of Densification
Cost of densification will be the key to
the success of the d-RDF concept. The
study evaluated the technical and
economic considerations of d-RDF for
small communities, processing up to
200 Mg (222 tons) of MSW per day.
Projections for the capital and operation
and maintenance costs were made for
several selected processes. For a
system designed to produce 9 Mg (9.9
tons) of pellets per hour, a cost of $9.20
per Mg ($8.28 per ton) is projected for
the capital and operating costs of the
densification module. Costs of the front
end system, (i.e., the refuse receiving
and processing system) is not included.
Details of the analyses done to
estimate the cost of the plant, front end
processing, and d-RDF module are
provided in the full report. The report
also providesdetails on estimating costs
of producing d-RDF and for estimating if
the process is economical for small
communities.
The full report was submitted in
fulfillment of Grant No. 804150 by the
National Center for Resource Recovery,
Inc., Washington, DC, under the spon-
sorship of the U.S. Environmental
Protection Agency.
-------
Jay Campbell and Marc L. Renard are with the National Center for Resource
Recovery, Inc., Washington, DC 20036; the EPA author Edward J. Winter is
with the Municipal Environmental Research Laboratory, Cincinnati, OH 45268.
Carlton C. Wiles is the EPA Project Officer (see below).
The complete report, entitled "Densification of Refuse-Derived Fuels: Prepara-
tion, Properties and Systems for Small Communities," (Order No. PB82-103 904;
Cost: $14.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
•frU.S. GOVERNMENT PRINTING OFFICE:1981--559-092/3319
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED
U b E-NJVTK
AGENCY
M S t-.LO«^n'TCrT
?30 S DtAKBOKN STKttT
CHICAGO IL -»-"fl
------- |