AN APPROACH TO DETERMINING THE ECONOMIC
FEASIBILITY OF REFUSE-DERIVED FUEL AND
MATERIALS RECOVERY PROCESSING
HARVEY W. GERSHMAN
National Center for Resource Recovery, Inc. (NCRR)
Washington, D.C.
ABSTRACT
This paper presents an approach for deter­
mining the economic feasibility of refuse-derived
fuel production and the recovery of various materi­
als. The information presented here is based largely
upon data developed for [1] the metropolitan
Washington, D.C. area [2] as input for the con­
sideration oCa regional resource recovery program
which would eventually encompass some 4000
tons-per�ay (3,628.8 Mg-per�ay) of municipal
solid waste. The initial facility would process 650
tons-per�ay (589.7 Mg-per�ay).
The facility is designed to recover refuse­
derived fuel (RDF), light and heavy ferrous
metals, aluminum and other nonferrous metals,
flint glass cullet, color-mixed glass cullet, color­
mixed glass fines, and handpicked waste news­
papers and corrugated. For the most part, advance
commitments for the sale of these products have
been obtained. [3]
An early stage in the planning process requires
recovery product revenues to be estimated and the
feasibility of the process to be analyzed. Since
materials revenues can be predicted with a higher
degree of certainty than RDF revenues, it becomes
necessary to determine what revenues will be re­
quired from the sale of the RDF in order that
projected economics can at least be the same as
the alternative disposal practice. A technique is
described here which will assist the decision-maker
to evaluate the economic feasibility of the pro-
1
posed project by determining what is termed here
the RDF "Indifference Value". [4]
INTRODUCTION
Planning for resource recovery should be viewed
as a sequential process whose goal is the implemen­
tation of an economically viable system for a given
,
.
area. There is a definite need, at any early point in
the planning process, to evaluate the economic
feasibility of a particular processing approach. How­
ever, before embarking into the feasibility process,
the first step should be identifying and obtaining
markets where the various recovered products can
be used. The firmer and closer these markets, the
better.
Markets for the various materials [5] products
have traditionally been easier to identify and obtain
than markets for refuse�erived fuel. Prices and
other conditions have been documented, [6] and
the development of consensus specifications and
practices has been undertaken. [7] For certain
materials products, especially aluminum and glass,
the problem� have focused around the technical
uncertainties of being able to produce the target
product. Ongoing research and development in
this area is helping greatly to reduce these un­
certainties. [8]
Different from materials, is the experience of
marketing RDF for use as a supplemental fuel in
coal-fired utility or industrial boilers. Compared to
advance commitments obtained for materials, RDF
commitments are oftentimes open-ended and not
guaranteed to "take-or-pay," even if produced to
the agreed upon specification. This is because of
uncertainty about its long run suitability as a fuel.
Even so, many jurisdictions are quite seriously
looking at the prospects and constructing plants to
produce RDF. [9] The experience with full-scope
materials recovery is more limited.
Materials revenues can be shown in two ways.
First, at the current market levels, determined by
calculating the pricing structure that is tied to
varying commodity quotations. Secondly, certain
materials revenues can be expressed in terms of
their minimum floor price level. Floor prices
become important in determining a "worst-case"
materials revenue level.
The approach discussed here takes the view­
point that an RDF energy product is the chosen
.
energy option. This certainly does not exclude
other energy options such as steam, pyrolysis gas
or oil, or electricity from being considered. But, for
the purposes of presentation, the RDF energy out­
put is used. Along with the RDF product, other
materials products are also recovered. Certain
materials products may not be applicable to a given
•
location because of such factors as (a) variability in
the waste stream, (b) availability of markets, or
(c) capital budgeting constraints.
H can be said that resource recovery will have a
good chance of being implemented where its costs
will be at least equal to or less than present or im­
mediate future costs for the disposal alternative.
With this in mind and the fact that materials reve­
nues can be predicted with a higher degree of
certainty, it becomes necessary to determine what
revenues will be required from the sale of the RDF
product in order that the projected economics can
at least be the same as the alternative disposal prac­
tice. The use of such a technique developed here
will assist the decision-maker in evaluating the eco­
nomic feasibility of the proposed project.
This approach provides the decision-maker on
the processing side a manner in which to establish
his minimum RDF floor price. The RDF user - in
most cases the local utility - will probably do a
similar calculation taking into account the cost of
an equivalent amount of fossil fuel the RDF is to
replace minus the incremental costs of firing RDF.
[10] Neddless to say, such values will vary and
need to be determined on a site-specific basis. Hope­
fully, the "bargain" struck will be above the pro­
ducer's calculated minimum and below the user's
calculated maximum. This is likely to be the only
2
case when a bargain can be struck. Put differently,
an undertaking to produce and utilize RDF would
then be mutually beneficial.
DESCR IPT ION OF THE FACIL ITY
A "full-scope" facility designed to process ac­
cording to the flowsheet depicted in Figure I, is
used as the example. The characteristics of the full­
scope processing is shown in Table L The different
types of output products to be recovered are listed
in Table 11. The capital and gross operating cost
estimates (August 1974 dollars) are listed in Table
III. These estimates represent considerable pre­
liminary engineering detail based on work per­
formed in support of a New Orleans resource
recovery program. [I I] Expected mate rials reve­
nues for the various output materials products are
provided in Table IV. All of these assumptions,
cost estimates and revenue projections are in a
1974 timeframe. Proper updating and adjustment
would be required to bring these projections to
future timeframes.
The facility produces RDF (57 percent by
weight) [12] which is hauled via transfer trailers to
a utility user 20 miles away. Waste newspapers and
corrugated are handpicked prior to processing. The
remainder of the solid waste (36.6 percent by
weight), i.e., the shredded, air classified heavy
fraction, is magnetically separated for recovering
ferrous metals (9.65 percent by weight). An air
knife divides the ferrous fraction into two prod ucts
- light and heavy. Dump trailers transport these
ferrous fractions to market. A rising-current sepa­
rator removes heavy organics while concentrating
the noncombustible materials. A two-stage heavy
media system next concentrates a mixed nonferrous
fraction (0.27 percent by weight) and a glass and
aluminum rich stream. After the aluminum has
been separated by the electrostatic separators, flint
and color-mixed glass cullet (6.3 percent by weight)
are separated by a color-sorting system. Glass fines
(2.5 percent by weight) are recovered from the
various screening underflows by a froth flotation
circuit. These recovered materials are transported
to market via dump trailers. The remaining residue
(I7 percent by weight) is transported via dump
trailer to adjacent landfill operations.
ANALYSIS APPROACH
This analysis determines the minimum value
Refuse-Derived Fuel (RDF) would need to take in
 TAB LE I. C HARACTERIST ICS OF FU L L-SCOPE FACI LITY

Tonnage: 650 tons-per-day nominal, 750 tons-per-day maximum

Operating Hours: two eight-hour shifts, 12 hours operating,

4 hours for maintenance and downtime, 6 days
per week operation, 312 days per year. Glass
cu11et recovery equipment operates 22 hours per
day.

Design Life: 10 years major plant and equipment: 5.years rolling

stock.

Location: Adjacent to existing landfill operations with weigh

scale facility.

Input: Primarily raw municipal solid waste and some bulky waste
as well as commercial.

Operator: Publicly run.

Financing: 7% General Obligation Bonding, 100% Debt.

Recovered Products Transportation to user by truck:

RDF . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 roundtrip miles

Light Ferrous • • • • • • . • • • • • • • • • • • l50 roundtrip miles
Heavy Ferrous • • • • • • • • • • • • • • • . • • 50 roundtrip miles
Other nonferrous • • • • • • • • • • • • • • • 50 roundtrip miles
Glass Cul1et & Fines • • • • • , • • • • • 250 roundtrip miles
Aluminum, waste
newspapers, and ,
waste corrugated . . . . • • . • • . . . • f.o.b. Recovery Facility

Input Comoosition (by weight) moisture included:

Paper 43.8%
Ferrous 10.2%
Aluminum 0.7%
Other Nonferrous 0.3%
Glass 10.5%
Yard Waste 9.7%
,Food Wast
. e 9.9%
Rags and ''lood 5.1% 1 ton = .9072 Mg
Rock and Ash 9.8% 1 mile -
- 1. 069 km
100.0%

3
 TABLE I I. OUTPUT PR ODUCTS F OR T HE RECOVERY FAC I L IT Y

Output Per Cent by Weight

Refuse-Derived Fuel 57.00

waste News 3.20

waste Corrugated 3 .. 20 .

Light Ferrous 8.80

Heavy Ferrous ' 0.85

Aluminum 0.50

Other Nonferrous 0.35 •

Color-Mixed Glass Fines 2.80

Flint Glass CuI let 2.50

Color-Mixed Glass CuI let 3.80

RESIDUE 17.00

TOTAL 100.00

*Revenues are based on 0.27% Nonferrous·Metals.

4
 TABLE III. CAPITAL AND OPERAT I NG COSTS F OR T HE REC O VER Y
FAC I L IT Y

Capital cost*

Processing $ 6,659,000
Materials and Residue
Transportation 156,000
RDF Transportation 125,000
$ 6,940,000

Operating Cost

$ 2,584,000
,

Processing
Materials and Residue
Transportation 358,000
RDF Transportation 228,000
$ 3,170,000

*Includes Indirect Capital Costs for Major Plant

and Equipment of:

Contractor's Overhead 20.5%

Contractor's Profit 5.0%
Engineering 6.5%
Contingency 10.0%
Total 42.0% •
I and

includes Indirect Capital Costs for Rolling Stock of:

Contingency 10.0%
Total 10.0% •

5
 '!
TABLE IV, R E VE NUES FROM FULL SCOPE MATER IA LS REC O VERY FAC I L ITY

•
I
Percentaqe Dec. 2 Revenue Floor Revenue
Mat erials Recover ed Decembe r 2 Price Floor Price ($ r in ut ton) ($ er in ut ton"
Liqht Ferrous 8.80 in excess of $20/lonq ton $4.32 $1. 57
$55/lonq ton2

-
Heavy Ferrous 0.85 in excess of None $ .55
$72/lonq ton
j
,

A1nminum 0.50 in exces � of in excess of $1.50 $1.25

$300/ton

Other Nonferrous 4 -
0.27 in excess of None $ .95
,
$350/ton I
,
I
I
Flint Glass 2.50 $ 1 6/to n : $16/ton I $ .40 $ .40
•
I ,

Color Mixed Glass 3.80 $lS/ton ; $15/ton $ .57 $ .57

0\ Color-Mixed
I
Glass Finas 2.80 $lS/ton $15/ton $ .42 $ .42
, ,
Waste Corruqated 3.20 $20/ton '$20/ton $ .64 $ .64
!
Waste News 3.20
I,
$lO/ton $lO/ton
I $ .32 $ .32

RDF RESIDUE 'Total Revenues $9.67 $5.17

l
These are qross revenues, not including transportation.
21 Long Ton equals 2,240 lbs.
3
The aluminum bid price increases each year from in excess of $300/ton in the first
year to in excess of $400/ton in the fifth year.
48&sis for revenues only. Aotual output is 0.35% which inclUdes non-metallics, upon
which no revenue is received.

1 ton ' .9072 Kg

 TAB LE V. COST A NA LYS IS FOR THE RECO VERY FAC I LITY

Item ($/year)

Gross Operating $3,170,000

Mate�ials Revenues at December 2, 1974 levels

l
Materials Reven�es ($1,961,000)
Landfill Credit ($673,000)
RDF Revenues Needed ($536,000)
Total, Revenues & Offsetting Credits ($3,170)000)
RDF "IV" ($/M BTU) 3
-
.46

Materials Revenues at Floor Price Levels

Materials Reven� esl ($1,048,000)

Landfill Credit ($673,000)
RDF Revenues Needed ($1,449,000)
Total, Revenues & Offsetting Credits ($3,170,000)
RDF "IV" ($/M BTU) 3
-
1. 26

lFrom Table IV

2$4.00 per ton for every ton diverted from disposal

•

35000 BTU/lb., 57% RDF output

1 lb. = .4536 kg
1 ton = .9072 kg

1 BTU = 1. 055 kJ

7
 REFUSE

Receiving Pad Bucket Elevators (7)

Front End Loader Three Deck Screen (2) (0.5%)
Apron Type Pit Conveyor Double Roll Crusher ' Aluminum - - ...
Picking Station Surge & Tote Bins
I (2.5%)
Paper Baler (6.4%) Electrostatic Separators (2 sets) Glass Cullet--_
Large Rejects Bin Transparency Color Sorters (3 sets) (3.8%)
Trommel Color-Mixed --.
Shredders (2) Glass Cullet
•
Shredded Mat'l Conveyors (2) -3/16 inch
I
Fines & Sorter Rejects
I
SHREDDED (0.9%) I
REFUSE I
•
(93.6%)

Fines Belt
Impactor
Surge Bin (2.3%) Bucket Elevators (2)
Air Classifiers and Cyclones (2 sets) IDUE--. Screen
Dual Push Pits (2) Pump and Pump Tank
RDF Transfer Packers (2) Hydrocyclones
Transfer Trailers (5) Wash Tank
Tractors (4) 57% Conditioning Tank
00 Material Reagent Feeder
Flotation Cells
HEAVY Thickener
MATERIAL Tailings Tank (2.8%)
(34.3%) Water Treatment & Lagoon S FINE�_�
Tailings Pond (5.2%)
Glass Fines Silo IDUE- ...

Heavy Materials Screen

Rising Current Separator
Secondary Shredder (8.8%)
Bucket Elevators FERROUS-"
Magnetic Separators (2) (.85%)
Air Knife and Auxiliaries FERROUS-"
Tote Bins (.35%)
Metal Baler and Auxiliaries I >OTHER NONFERROUS •
Heavy Media System (9.5%)
Dryer IDUE-+

-3/16 inch
1 inch = 2.54 cm Fines
(7.1%)
+3/16 inch Nonferrous & Mixed Glass (7.7)

FIG. 1. FU L L-SCOPE E QUIPME N T F L OWSHEET F OR RESOURCE RECO VERY PR OCESS I NG

 TAB LE VI. T HE I ND IFFERENCE PRI NC IP LE

650 Ton-Per-Day
Full-Scope
Sanitary Landfill Resource Recovery

$ per year
Gross Operating 3,170,000

Materials Revenues
($9.67 per input ton) (1,961,000)

Landfill Expenses
($4.00 per ton x 17 percent) 138,000
Operating Costs:
RDF Revenue Needed ($0.46
$4. 00 per-ton x 650 toris­ per-million BTU x 10 million
per-day x 312 days-per­ BTU per-ton RDF, 57 percent
year output) (536,000)

Net Operating Budget $ 811,000 Net Operating Budget $ 811,000

($ per. year) ($ per year)

L---- Annual Cost Calculated To Be Equal-----J

1 ton - • 9072 Mg

1 BTU - 1. 055 kJ

9
order that a facility be competitive with alternative
costs for disposal - here a landfill example is used
as the alternative. This value for RDF is called here
the RDF "Indifference Value" (IV). [13] The
analysis is carried out for a facility processing 650
tons-per-day (589.68 Mg-per-day) and recovering
RDF plus other materials. A credit of $4.00 per
,
ton ($4.41 per-Mg) is assigned for refuse diverted
from the landfill operations. This is discussed fur­
ther later on. The RDF revenues needed can be
deterlllir.ed by the following equation:
A -
-
B + C D = E, where,
- structs a variable "F" [16] which is the credit
A -
-
Gross operating cost;
B -
-
Expected materials revenues;
C -
-
Landfill cost for residue disposal;
D -
-
Revenue needed from RDF;
E -
-
Target net operating cost;
Values for "A" come from the operating cost
for the particular facility configuration as given in
Table III. "B" comes from the expected materials
revenue described in Table IV while "c" is assumed
to be $4.00 per-ton ($4.41 per-Mg) multiplied by
the quantity of residue annually landfilled. "D" is
calculated so that "E" equals $4.00 per-ton ($4.41
per-Mg) times the total annual facility throughput,
such that the target net operating cost for the
facility is equivalent to the operating cost for land­
filling the entire facility input directly. Table VI il­
lustrates this concept from a budgetary perspective.
On the right side of Table VI, gross operating costs
of the facility are shown as $3,170,000 annually.
December2, 1974 exchange prices are assumed for
the sale of recovered materials at $9.67 per ton
($10.66 per-Mg), or $1,961,000 annually. An ad­
ditional expense of $138,000 is shown for land­
filling the process residue. The analysis assumes
that 57 percent by weight of the input refuse
emerges as RDF. The RDF is assumed to have a
heat value of 5000 BTUs-per-pound (11.63 kJ -per­
g) and a moisture content of27.5 percent. [14]
The recovery facility has to bear the cost of its
residue disposal (17 percent of the input) and then
sell its RDF. The resulting operating costs come out
to be $4.00 per ton ($4.41 per Mg), or on a yearly
basis, $811,000. This is the same as not having a
recovery facility and simply landfilling the waste.
This is shown on the left. In order to obtain this
net operating cost of $811,000, RDF would have to
account for $536,000 annually in revenues, or be
worth approximately $0.46 per-million-BTUs
($.44 per-Gj). [15]
10
The equation then means simply that ROF
sales must produce a certain revenue if the sited
facility is to be competitive with the alternative
disposal charge.
In actual practice and as mentioned above, the
equation is solved for "0" the RDF revenue
-
needed to produce the same annual cost as for land­
fill. This is done by transposing "0" to the right side
of the equation and moving "E".to the' left side.
"E" is the target net operating cost for processing,
while "C" is the cost of land filling the residual from
the recovery facility. Merged together one con­
obtained from not having to landfill the recovered
products. Hence, the equation becomes:
. O=A-B-F
"0" is the product of the exchange price (IV)
times the energy content in the RDF delivered to
the user, multiplied times the quantity of RDF
produced and used. This is the approach used to
compute the "IV" values.
A somewhat uncertain area in determining net
costs is the credit used for "F" in the equation
given above. The figure of $4.00 was assigned for
refuse diverted from landfill operations. This has
been used only as an example. This value should be
determined on a case-by-case basis and evaluated
in such a way that actual savings attributed to not
having to landfill portions of the facility input are
estimated. If the processing facility were added to
an existing landfill operation, marginal costs for
not landfilling the recovered products should be
forecasted. Additionally, one may want to consider
the cost benefit for extending the life of a landfill.
This latter factor may be almost impossible to
calculate.
ANALYSIS AND RESULTS
The RDF "IV" 's for the facility have been
determined for both the December 2, 1974 and
the floor price materials revenue levels (see Table
V). At the December 2, 1974 materials revenue
level, the RDF "IV" is calculated to be $0.46
per-million-BUT's ($.44 per-Gj). When materials
revenues are at the floor price level, the ROF "IV"
is $1.26 per-million-BTUs ($1.19 per-Gj).
The outputs of this analysis provide a gauge for
judging the feasibility of the processing described.
These outputs should represent to the municipal
decision-maker target RDF minimum "bargains".
They say what that facility configuration would
 need such that the operating budget for disposal
of the same quantity does not change. An RDF
value above the "IV" will generate benefits to the
annual budget while an RDF value below will
bring about operating budget increases.
The "site·specific" RDF value will natually
fluctuate. The price the RDF user pays for the
alternative fossil fuel must be taken into account.
These costs vary significantly geographically. [17]
The location of the RDF user with respect to the
processing facility, the materials recovery revenues,
the ability to produce the materials to specification
for the expected revenues, also will bear upon the
RDF "IV".
There is roughly a factor of "three" between the
December "IV" and the floor price "IV". The
higher RDF "IV" at the materials floor price may
in some cases be higher than the "bargain" that
may be obtainable.
A caveat concerning the recovery of glass and
aluminum is in order. What is presented here re­
presents, at best, preliminary process design con­
figuration and cost estimates. Refinement in such
numbers will certainly affect the "bottom line" as
to the economic feasibility of their recovery. As
such, the relative values presented here are thought
to be subject to change.
COMMENTS
The analyses and methodology presented use as
a working example what is termed a "full-scope"
resource recovery system which produces a certain
type RDF and other materials outputs. Certainly
one must recognize the wide range of system types
that are being developed, marketed, procured, and
operated. Oftentimes these systems have very dif-
• Engineering Feasibility Study," NCRR; December
ferent types of outputs. While this methodology
can be adapted to serve other types of systems,
naturally one would have to take into account
different ca'pital costs, operating costs, materials
revenues, and alternative'disposal credits. Knowing
these, this methodology could be used to deter­
mine what value the energy product from the sys­
tem would hav'e to assume in ,order to be at least
competitive.
If the value for the energy value can be deter­
mined in the same time frame as the matefials­
all the better. This naturally would be preferable,
but unfortunately not always attainable. If this
were the case, the net economics could be deter­
mined and be compared directly to the alter­
native without using an "Indifference" approach.
11
RE FERENCES
[ 1] "Resource Recovery Engineering and Economic
Feasibility Study for the 1-95 Complex, Lorton, Virginia,"
NCRR, 1975, for the Washington Metropolitan Waste
Management Agency.
[2] City of Alexandria, Virginia; Arlington County,
Virginia; Washington, District of Columbia; Fairfax
County, Virginia.
[3] "1-95 Center Markets for Recoverable Resources
from Mixed Municipal Solid Waste," September 5, 1974
and Draft of No. 1 above, January 7, 1975, NCRR, for
the Washington Metropolitan Waste Management Agency.
[4) Op.Cit. No. 1,
[5] Materials such as ferrous, aluminum and other
nonferrous metals; glass cui let and fines; and waste papers.
[6] Op.Cit. No. 3; "EPA/Specifications For
Materials Recovered From Municipal Refuse," Dr. Harvey
Alter and W, R. Reeves, NCRR, May, 1975; "Solid Waste
Resource Recovery Market Survey and Analysis," Black,
Crow and Eidsness, Inc.Study, Altoona, Pennsylvania,
July, 1975; "Markets and Technology For Recovering
Energy From Solid Waste," EPA by Steven Levy,
Washington, D. C. 1974.
[7] ASTM E. 38 Resource Recovery Committee,
Chairman, Dr. Harvey Alter, NCRR.
[8] Organizations such as the Environmental
Protection Agency, United States Bureau of Mines,
Garrett R&D, Raytheon Service, Combustion Power
Company, Black Clawson, and NCRR.
[9] For example: Chicago, Illinois; Milwaukee,
Wisconsin; Monroe County, New York; Bridgeport,
Connecticut; and the 1-95 Complex in Lorton, Virginia,
[10) Some of the factors affecting this cost are: the
capital costs of the auxiliary equipment for storing,
handling and firing RDF, operating and maintenance
costs, economic dispatch penalty, etc.
[11] "Cost Analysis for the New Orleans Resou rce
Recovery and Disposal Program," NCRR, 1974; "Bid
Specifications for Facilities and Operation of the
Resource Recovery and Disposal Program in New
Orleans, Louisiana," NCRR and the City of New
Orleans, 1974; and "Materials Recovery System
1972.
[ 12] The RDF output (57 percent) has been
chosen this low in order that higher quality fuel would
be obtained in terms of energy and ash content versus
attempting to maximize organic content.
[ 13] Op. Cit. NO. 1.
[ 14] Based on 380 samples taken November 9, 1973
through December 10, 1974 at the Union Electric
Meramec Plant.
As Fired Basis
Moisture Ash BTU/Pound (JIg)
Average 27.5 18.5 5,006 ( 11,643)
Maximum 63.0 53,8 7,593 ( 17,660)
Minimum 3.0 7.6 2,293 ( 5,333)
[ 15) 650 tons-per-day (589.7 Mg-per-day) x 312 days
per year x 0.57 RDF output x 10 million BTU-per-ton
11.63 k J/g RDF x $0.46 (approximately) per-million per-ton ($3.66 per-Mg) credit.
BTU $.44 (approximately) per-GJ equals $536,000 [17] See "Monthly Fuel Cost and Quality Informa­
per year. tion," Federal Power Commission, Washington, D.C.
[16] "F" equals "E" minus "C" or $4.00 per-ton 20426.
($4.41 per-Mg) less 17 percent x $4.00 equals $3.32

Key Words
Cost Estimates
Economic Feasibility
Energy Recovery
Equipment
Full Scope Facility
I ndifference Value
Markets
Materials Recovery
Municipal Solid Waste
Refuse-Derived Fuel
Resource Recovery

12