United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-121 July 1981
Project Summary
Recovery of Aluminum from
Municipal Solid Waste at
Recovery 1, New Orleans
Louis P. Soldano
This report summarizes four
technical reports that document a
series of tests (referred to as Test Nos.
5.01, 5.02, 5.03, 5.07) to recover
aluminum from the processing of
municipal refuse at the New Orleans.
Louisiana, Resource Recovery Project
(Recovery 1). The objective of test No.
5.01 was to document the
performance of the Eddy Current
Separator that recovers principally
aluminum cans, for different feed
rates. Test No. 5.02 was conducted to
evaluate the efficiency of the Eddy
Current Separator when the feed rate
was held constant and belt speed
carrying the feed through the separa-
tor was varied. A "zig-zag" vertical air
classifier was added as a cleanup step
in Test No. 5.03. The classifier's
ability to remove aerodynamically
light contaminant from the Eddy Cur-
rent Separator's Product was meas-
ured. Test No. 5.07 evaluated the
ability of a double deck vibrating
screen to separate the "heavy"
product of the air classifier into an
overs stream that is discarded, a mid-
dlings stream that is the feedstock for
the aluminum recovery submodule,
and an unders stream that is the feed-
stock for the glass recovery submod-
ule.
This Project Summary was develop-
ed by EPA's Municipal Environmental
Research Laboratory, Cincinnati, OH,
to announce key findings of the
research projects that are fully docu-
mented in separate reports (see
Project Report ordering information at
back).
Introduction
This report is a summary of tests
performed on a full-scale resource
recovery plant operating at New
Orleans, Louisiana, called Recovery 1
Resource recovery consists of reclaim-
ing for use materials or energy (ferrous
metals, aluminum, glass, paper, refuse
derived fuels, etc.) from processed or
unprocessed municipal solid waste
(MSW). For a flow sheet representation
of Recovery 1, see Figure 1.
At Recovery 1 there is an aluminum
recovery submodule for the removal,
cleanup, and densification of alumi-
num, primarily cans, found in the MSW.
The operating separation equipment
are in order, (1) a large revolving
perforated cylinder called a trommel,
through which MSW is passed The
refuse is separated into greater than 4-%-
m. material that flows through the
trommel, called trommel overs, and less
than 4-%-in. material that falls through
the trommel holes and is called the
trommel unders; (2) a scalping magnet
that collects ferrous metal from the
trommel unders, (3) an air classifier that
blows a portion of light orgamcs away
from the less than 4-%-in. trommel
unders.
Next comes the Eddy Current Separa-
tor, called an Al Mag (Aluminum Mag-
net), which removes the aluminum cans
from the feedstock Aluminum cans
-------
Raw MSW-
• Trommel —
\ \
Trommel Unders
Trommel Overs
-^•Scalping
T
Magnet *• Ferrous
•»• Light Organics
Cyclone
\
Air Knife •*—
& Zig-Zag
Air Classifier
I \
Aluminum Discards
Cans
Glass Feedstock •+
• At Mag •<— Magnet •
Air Classifier —
I
— Screen ^> Discards
I
— Screen
Shredder
Commercial Product
Figure 1. Final upgraded aluminum recovery system at Recovery 1.
usually represent about 3% to 5% by
weight of the Al Mag feed.
The Al Mag operates on the principle
that an electric current is generated in
each conductor passing through an
electromagnetic field. These "Eddy
Currents" have a magnetic field that
interacts with the applied field and
produces a lateral force on the metal.
This force expels the nonferrous metal
from the conveyor.
Downstream of the Al Mag are two
pneumatic separators, an air classifier,
and an air knife whose purposes,
sequentially, are to (1) fly the light
organics from the Al Mag product and
(2) blow the aluminum cans beyond the
heavy organics, inorganics, and non-
ferrous metal that are carried over as
part of the product.
Finally, a shredding process tears the
cans into smaller pieces less than 1 in.,
mainly to increase the bulk density to 15
to 25 Ib/ft3.
The first set of tests, Test Report No.
5.01, was performed to evaluate the
performance of the Eddy Current Sepa-
rator for different feed rates. Al Mag
feed rates of 2, 3, and 4 tons per hour
(tph) were tried.
In the second set of tests. Test Report
No. 5.02, the same equipment was used
but instead the burden was varied. The
burden is the amount of material in the
magnetic field at any one period of time.
This was done by varying the belt
speeds of the Al Mg.
Test Report No. 5.03 was done to
evaluate the performance of a "zig-zag"
vertical air classifierto remove contami-
nant from the aluminum recovered by
the Al Mag.
Finally Test Report No. 5.07 was con-
ducted to document the efficiency of a
double deck vibrating screen in separa-
ting the scalped primary trommel
unders into three streams. The three
streams are an overs stream that is
discarded, a middlings stream that is the
feedstock for the aluminum recovery
submodule, and an unders stream that
is the feedstock for the glass recovery
submodule.
Test No. 5.01—Eddy Current
Separator Performance
In the initial set of runs, three opera-
ting points are described. They differ in
the amount of time an individual piece
of aluminum is inthemagneticfieldand
is subject to become Al Mag product.
The objective is to find out whether this
relatively expensive piece of equipment
can be run at higher throughput levels
without suffering major losses in effi-
ciency. Higher throughput capacity will
provide better processing economics.
The Al Mag is rated to recovery 75% of
essentially whole aluminum can
product. The product should contaii
less than 50% loose contamination fron
3 tph of 8.5 to 11 Ib/ft3 feed material
At runs of approximately 5%, b'
weight, aluminum can concentration
the product recovery rose from approxi
mately 175 Ib/hr to 280 Ib/hr wher
throughput was increased from 2 to ^
tph. Efficiences for the 2, 3, 4 tph runs
were approximately, 81%, 77%, anc
66%, respectively. Factors decreasing
efficiency were inconsistent feedinj
and nonaluminum material that was noi
easily removed from the Al Mag feed-
stock.
For these tests, the feed to the Al Mag
was intentionally seeded with
aluminum cans falling into three shape
categories: flattened cans, minimally
deformed cans, and variously deformed
cans. The percent recovery efficiencies
were:
can shape
flattened
minimally
deformed
deformed
percent
of seed
20
20
60
tph run
2
84
93
77
3
81
89
73
4
77
88
66
Before the tests, it was expected that
the efficiency of aluminum can recovery
would decrease at the higher through-
puts and higher belt speeds, butthatthe
total amount of recovered aluminum
cans (product mass flow rate) would in-
crease. These expectations were true
when the feed rate was increased from
2 to 4 tph. There was only an 18.5%
decrease in efficiency, but the product
rate rose from 175 Ib/hr to 280 Ib/hr—
an increase of 60% of aluminum
recovered in a given time interval. The
average loose contamination in the Al
Mag product for all runs was about 5%.
Test Report No. 5.02—Al Mag
Belt Speed Varied
The next series of tests used the same
equipment as in Test No. 5.01. The feed-
stock consisted of the screened material
(the less than 2-in. light material)
remaining from the feedstock used in
Test No. 5.01. The burden was varied
systematically, being less for faster belt
speeds. The objective of this test was to
measure the differences in Al Mag effi-
ciency, product rate, and product quality
that accompany changes in the
composition of the feedstock, conveyor
-------
belt speed, and consequently, burden
depth.
The composition of the feed cans used
was 18% flattened, 71% deformed and
11 % whole. In each run, the nonalumi-
num can portion of the feed was the
same; the aluminum cans were the
same; andthe massand volumetric flow
rates were the same. The artificial
average burden depths across the 18 in.
of active belt width were calculated
based on belt width, belt speed, mass
flow rates, etc. These calculated depths
were 3/8 in. at 300fpm, 9/32 in. at400
fpm, and 7/32 in. at 500 fpm.
The results indicate only a small
effect on Al Mag efficiency and virtually
no effect on product quality as the belt
speed was varied. The belt speeds used
were 300fpm, 400fpm, and 500fpm. Al
Mag efficiencies for the three belt
speeds were 91% (300 fpm), 85% (400
fpm), and 86% (500 fpm). The product
contamination was 12%, 11 %, and 13%,
respectively.
After adjusting the test results to
account for the greater aluminum
content, product rates were 396 Ib/hr
for the 300 fpm runs and 373 Ib/hr for
the 400 fpm runs. The reason for the
adjustment was that the sample cans
were 12% inthecurrenttestratherthan
the 5.5% of normal feed. It therefore
appears to be a benefit to screen out as
much less-than-2-in. material as
possible.
From the tests it can be concluded
that the effect of reducing burden depth
(by running the belt faster) does not
compensate for the loss of time in the
magnetic field and in the time to leave
the belt. At a constant feed, it is advan-
tageous to slow down the belt.
Test Report No. 5.03—"Zig-
Zag" Air Classifier
The next series of tests was done to
evaluate the performance of a MAC
"zig-zag" vertical air classifier to
remove aerodynamically light nonmetal
contaminant from aluminum cans
recovered by the Al Mag.
The materials that fly in the air classi-
fier are carried to a cyclone, where the
contaminated heavy materials drop out
of the air stream into an airlock feeder
after which they are rejected.
Two measures of effectiveness are
reported. The first, nai, is based on the
degree to which light gauge aluminum
drops in the air classifier. The second,
ni,te indicates the percentage of the light
contaminant in the feed that flies.
At Recovery 1, there are difficulties
with the unit processes that provide the
feedstock for the Al Mag. These up-
stream processes are not as efficient in
removing oversize (greater-than-4-in.)
and undersize (less-than-2-in.) contam-
inants as was anticipated. As a result,
the Al Mag operates on 2% aluminum at
a throughput of 2 tph. This results in the
processing rate of 80 Ib/hr of aluminum
cans.
The upstream processes of the air
classifier are expected to improve so
that Al Mag feed will have a concentra-
tion of about 4% aluminum cans at a
mass flow rate of 4 tph. This would bring
about 300 Ib/hr aluminum into the
feedstock.
To determine proportions to use in
preparing the test feed for the air classi-
fier, actual Al Mag product was col-
lected. Three samples were taken at
different times and separated into
aluminum canstock (85%), less-than-4-
in. and greater-than-2-in. lights (2.0%),
less-than-2-in. and greater-than-l/4-
in. lights (1.5%), and heavies (11.5%),
both organic and metallic.
A total of 12 runs were made on the
air classifier system at a nominal 250
Ib/hr: three with the base line distribu-
tion above; three where the objective
was to double the loading of the less-
than-4-in. and greater-than-2-in.
fraction; three with the loading of the
less-than-2-in. and greater-than-1/2-
in. fraction doubled; and lastly, three
runs with both less-than-4-in. and
greater-than-2-in. and less-than-2-in.
and greater-than-1/4-in. fractions
tripled.
The average for ni,te was 91% for all
tests. In terms of the average of the most
representative run's nai, the percentage
of cans dropped was 97%. Associated
with this figure was a loss of 5.76 Ib/hr
of aluminum that flew into the air classi-
fier. As one would expect, nai fell off
somewhat for the higher velocity runs,
whereas ni,te rose.
After a number of runs, it was deter-
mined that there probably was bias in
the value of nai. The bias resulted from
the method used to reconstitute the
sample of cans after each run.
The air classifier effectively removed
light contamination and appeared to
work satisfactorily at the higher load-
ings even at decreases in air velocity on
the order of 20%. Less than 3% loss in
aluminum occurred. The fraction re-
moved of the large (less-than-4-in. and
greater-than-2-in.) material ranged
from 73% to 100%.
The removal at the "zig-zag" classi-
fier step of about 3 Ib light organics/100
Ib product produced by the Al Mag leads
to approximately a 5% metal recovery
improvement. The net gam from opera-
ting the air classifier is $2.50 to S3.50/
100 Ib product shipped.
Test Report No. 5.07—Double
Deck Vibrating Screen
Test Report No. 5.07 documents a
performance test of the double-deck
vibrating screen. The screen separates
the "heavies" from the air-classified,
scalped, primary trommel unders into
three fractions: an overs stream that is
discarded, a middlings stream that is the
feedstock for the aluminum recovery
submodule, and an unders stream that
is the feedstock for the glass recovery
submodule.
The mam objective of the test was to
determine the effectiveness of the
screen to concentrate recoverable
aluminum and glass into their
respective fractions. The screen oper-
ated at the nominal design condition of
62.5 tph. Three replicate concurrent
samplings of each output stream were
taken. These were analyzed for bulk
density, moisture content, composition,
and particle size distribution. During the
tests, the actual feed flow rate was 56.2
tph and the bulk density of the MSW
processed was 15.0 Ib/ft3.
The double-deck vibrating screen is a
Tyler Ty-Rock Type F-800,* nominally 6
feet wide and 10 feet long. The screen
was designed to process 17-1/2 tph of
15 Ib/ft3feedstock. The screen vibration
is characterized by a frequency of 846
cycles/min and an amplitude of 13/32
in. The top screen deck is a 1 /2-in. thick
steel plate punched with 5-in. diameter
holes located on 5-3/8-in. staggered
centers. There are seven rows of 21
holes and seven rows of 22 holes for a
total of 41.Oft2 of open screen area. The
bottom deck is a 2-in., square-opening
wire cloth. The wire diameter is 0.135
in., which results in an 88% open area.
The bottom deck is 5 ft, 8-3/8 in. wide
and 10 ft long.
The double-deck screen was tested by
sampling each output stream for 30
seconds. The mean feed rate for the
screens was 18.60 tph. The feed to the
vibrating screen separated into three
•Mention of trade names or commercial products
does not constitute endorsement or recommenda-
tion for use by the U.S Environmental Protection
Agency
ft US. GOVERNMENT PRINTING OFFICE: 1W1 .757-012/7222
-------
output streams; 3.0% were overs; 27%,
middlings; and 70%, unders.
As an aluminumconcentrator,thetop
deck of the screen performed satisfac-
torily; 97% of the aluminum cans
correctly reported to the middlings. This
is the feedstock for the aluminum
recovery submodule. In actual practice,
the top deck accomplishes little more, in
terms of preparing feedstock for the
aluminum submodule, than removing
relatively large and inflexible items. The
top screen deck is subject to binding by
flexible organics and requires frequent
removal of accumulated material.
The bottom deck is also expected to
perform the function of removing less-
than-2-in. material from the feedstock
to the aluminum recovery submodule.
The test data showed that only 83% of
the less-than-2-in. material reporting to
the 2-m. bottom deck actually passed
through the 2-in. screen openings. This
compares with the 85% expected to
pass.
The aluminum can concentration was
0.6% in the feed to the top deck. In the
middling stream, it rose to 2.3%. If all
the less-than-2-in. material had passed
through the bottom screen, this could
have risen to 5%. Particle size distribu-
tion for the feed to the 2-in. deck was
95% aluminum cans greater than 2 in.,
94% glass less than 1 in., and 29%
organics between 1 and 2 in.
The screen worked far better as a
glass concentrator as practically all the
glass (99.8%) was less than 2 in. The
amount carried over by the 5-m. deck
was insignificant, and only 3% of the
total glass in the feedstream was lost to
the middlings. The unders, i.e., the feed-
stock for the glass recovery submodule,
was 49% glass.
The full reports were submitted in ful-
fillment of Contract No. 68-01-4423 by
the National Center for Resource
Recovery, Inc , Washington, D.C., under
the sponsorship of the U.S. Environ-
mental Protection Agency.
The EPA author of this Project Summary is L. P. Soldano. who is with the
Municipal Environmental Research Laboratory, Cincinnati, OH 45268.
Donald Oberacker and Carlton Wiles are the EPA Project Officers (see below I.
This Project Summary covers the following reports, prepared by the National
Center for Resource Recovery. Inc., Washington, DC:
"Test of an Eddy Current Separator for the Recovery of Aluminum from
Municipal Waste: Test No. 5.01, Recovery 1, New Orleans," (Order No.
PB 81-217 663; Cost: $6.50, subject to change)
"Further Testing of an Eddy Current Separator for the Recovery of Alumi-
num from Municipal Waste: Test No. 5.02, Recovery 1, New Orleans,"
(Order No. PB 81-217 671; Cost: $6.50, subject to change)
"Performance of an Air Classifier to Remove Light Organic Contamination
from Aluminum Recovered from Municipal Waste by Eddy Current Separa-
tion: Test No. 5.03, Recovery 1, New Orleans," (Order No. PB81-217689;
Cost: $6.50, subject to change)
"Test of a Double-Deck Vibrating Screen Employed as an Aluminum and
Glass Concentrator: Test No. 5.07, Recovery 1, New Orleans," authored by
Perry M. Bagalman and Kelly Runyon (Order No. PB 81-217 697; Cost:
$8.00, subject to change)
The above reports are available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officers can be contacted at:
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
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
PS 0000329
U S
PROTtiCi'IUK
230 S DEARBORN
CHICAGO IT-
------- |