Production of diesel fuel from safflower oil by a soap-pyrolysis process
by Scott Eric Kufeld
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in
Chemical Engineering
Montana State University
© Copyright by Scott Eric Kufeld (1988)
Abstract:
The purpose of this research was to study the soap-pyrolysis process for making fuel suitable for use in
a diesel engine. The process involved converting safflower oil into a soap and then pyrolyzing the soap
to obtain the diesel range fuel. Commercial high-speed and medium-speed lubrication oils were
investigated with five weight percent soap-pyrolysis fuel contamination relative to polymerization
properties.
The first set of experiments studied pyrolysis process variables. It was observed that: fuel from soap
aged three months had a greater initial kinematic viscosity than fuel from soap aged two days and both
these fuels had about the same kinematic viscosity after 48 hours in the polymerization apparatus;
glycerol in the soap decreased yield; fuels from soap with a 50% excess amount of Ca(OH)2 showed
less thickening after the polymerization tests than fuels from soap with a stoichiometric amount of
Ca(OH)2; calcium soap is superior to magnesium soap since the yield of fuel is greater; the #2 range
distillate has less of an initial kinematic viscosity than the total distillate and both distillates have the
same kinematic viscosity after the polymerization tests; the lighter the fraction of the total distillate the
smaller the initial kinematic viscosity and the greater the viscosity increase; storage time of fuel had
little effect on polymerization properties.
The second set of experiments showed that addition of Ca(OH)2 prior to pyrolysis decreased
thickening while limiting yield.
The third set of experiments found the best soap production method investigated was the fusion method
with no solvent. Other methods tried were the precipitation process and using both toluene and water at
different concentrations as solvents in the fusion method.
The fourth set of experiments showed that it would be desirable for the safflower oil to contain more
oleic acid esters than linoleic or linolenic acid esters.
Optimization of the process showed that addition of a stoichiometric amount of Ca(0H)2 for fusion
soap production, and no Ca(0H)2 added prior to pyrolysis, was most desirable.
The soap-pyrolysis fuels worked better as fuels in a diesel engine than safflower oil. PRODUCTION OF DIESEL FUEL FROM SAFFLOWER
OIL BY A SOAP-PYROLYSIS PROCESS
by
Scott Eric Kufeld
A thesis submitted in partial fulfillment
of the requirements for the degree
of
Master of Science
in
Chemical Engineering
MONTANA STATE UNIVERSITY
Bozeman, Montana
September 1988
/
/<
*r3 s'
ii
APPROVAL
of a thesis submitted by
Scott Eric Kufeld
This thesis has been read by each member of the thesis
committee and has been found to be satisfactory regarding
content, English usage, format, citations, bibliographic
style, and consistency, and is ready for submission to the
College of Graduate Studies.
,4/2/Z./#0
Date
Chairperson, Graduate Committee
Approved for the Major Department
Approved for the College of Graduate Studies
Date
iii
STATEMENT OF PERMISSION TO USE
In presenting this thesis in partial fulfillment of the
requirements
University,
to
for
master's
degree
at
Montana
State
I agree that the Library shall make it available
borrowers
quotations
a
under
from
the
this
rules
of
the
Library.
Brief
thesis are allowable without special
permission, provided that
accurate
acknowledgement
of the
source is ma d e .
Permission for
extensive quotation
or reproduction of
this thesis may be granted by my major professor, or
absence, by
the Dean
of Libraries
either, the proposed use
purposes.
of the
Signature
Date
when, in the opinion of
material is
for scholarly
Any copying or use of the material in this thesis
for financial gain shall not be
permission.
in his
allowed without
my written
iv
ACKNOWLEDGMENTS
The author would like to thank the faculty and staff of
the Chemical Engineering Department
University
for
their
guidance
thanks is extended to my
Scarrah,
author
for
his
would
also
and
research
advice,
like
help,
to
here
at
Montana State
assistance.
advisor,
and
Dr.
Special
Warren P.
encouragement.
acknowledge
Dr.
Robert
The
L.
Nickelson and Dr. Daniel L . Shaffer who served on my guiding
committee.
student,
in
Advice
by
the area
Finally, the financial
Department
of
John
Olson,
a
fellow graduate
of polymerization is also recognized.
support
Natural
gratefully appreciated.
W.
received
Resources
and
from
the Montana
Conservation
is
V
TABLE OF CONTENTS
Page
APPROVAL...................................................
STATEMENT OF PERMISSION TO U S E ......
ii
iii
ACKNOWLEDGMENTS...........................................
iv
TABLE OF CONTENTS.........................................
v
LIST OF TABLES.....................................
LIST OF FIGURES...........................................
vii
ix
ABSTRACT....................
x
INTRODUCTION...........................
I
RESEARCH OBJECTIVES.......................................
6
EXPERIMENTAL DESCRIPTION......
Soap Production................
Pyrolysis.....................................
Polymerization.........................................
Acid Number.............................................
Iodine V a l u e ...........................................
Nuclear Magnetic Resonance............................
Distillation......................................... . .
Yield.................
RESULTS AND DISCUSSION....................................
Pyrolysis Studies......................................
Aged Versus Fresh Soaps............................
Glycerol Content of the Soa p s.....................
Fatty Acid Content of Soa p s...................
Calcium Versus Magnesium Soa p s....................
Distillation Results................
Storage Effects...........................
Results for Pure Linoleic A c i d ....................
Pyrolysis With and Without Calcium Hydroxide Reactant
Soap Production...................................... . .
Unsaturation Studies. ..................................
Soap-Pyrolysis Process Optimization..................
10
10
11
16
21
21
22
23
24
26
26
26
29
32
34
35
38
39
40
43
50
56
vi
TABLE OF CONTENTS— Continued
CONCLUSIONS.......
RECOMMENDATIONS FOR FUTURE
69
RESEARCH.....................
71
APPENDICES.................................................
72
Appendix A-Optimization Results......................
Appendix B-Computer Program Used for Optimization....
REFERENCES CITED
73
77
82
vii
LIST OF TABLES
Table
Page
I.
Fatty Acid Distribution in Safflower Oil..
3
2.
Typical Ranges for Diesel Fuel Types......
23
3.
Fuels from Aged and Fresh Soaps...... .
27
4.
Fuels from Soaps With Different Amounts of
Glycerol...............................
30
5.
Fuels from Soaps With Different Free Fatty Acid
Contents........................
6.
The Total Distillate Versus the #2 Ran g e...... .
36
7.
Viscosity Results for Diesel Fuel Types........
38
8.
Aged Versus Fresh Fuels...... ...................
39
Experiments With/Without Calcium Hydroxide
Reactant....... ..................................
41
Viscosity Results of Fuels With/Without
Reactant........................ ..................
42
Results of Fuels from Different Soaps With
Solvent.... ...... ................................
47
Yields of Fuels from Different Soaps With
Solvent............... ............................
49
13.
Results of Fuels from Different Acid Soaps.....
53
14.
Results for the Complex Method of B o x ...........
61
15.
Results from the 32 Factorial...................
62
16.
Variable Levels Used for Optimization.......... .
64
17.
Final Run Results................................
66
18.
Coded and Uncoded Variables for the Mathematical
Model of the Soap-Pyrolysis Process.........
74
Values for Surface Yield Equation Calculation...
74
9.
10.
11.
12 .
19.
viii
Constants Estimate
ix
LIST OF FIGURES
Figure
Page
1.
Soap-Making Process Apparatus...................
12
2.
Pyrolysis Apparatus With Stainless Steel Vessel.
14
3.
Pyrolysis Apparatus With Glass Vessel...........
15
4.
Polymerization Apparatus.........................
18
5.
Viscometer and Temperature B a t h .................
20
6.
Variable Levels for Sequential Simplex
Optimization Technique...........................
59
Variable Levels for Complex Method of Box
Optimization Technique...........................
60
Variable Levels for the Factorial Experimental
Design............................................
63
Computer Program Used For Optimization..........
78
7.
8.
9.
X
ABSTRACT
The purpose of this research was to study the soappyrolysis process for making fuel
suitable
for use
in a
diesel
engine.
The process involved converting safflower
oil into a soap and then pyrolyzing the soap to obtain the
diesel range fuel.
Commercial high-speed and medium-speed
lubrication oils were investigated with five weight percent
soap-pyrolysis fuel contamination relative to polymerization
properties.
The first set of experiments studied pyrolysis process
variables.
It was observed that: fuel from soap aged three
months had a greater
initial kinematic viscosity than fuel
from soap aged two days and both these fuels had about the
same
kinematic
viscosity
after
48
hours
in
the
polymerization apparatus;
glycerol
in the soap decreased
yield; fuels from soap with a 50% excess amount of C a (OH)2
showed less thickening after the polymerization tests than
fuels from soap with a stoichiometric amount of Ca(OH)2 J
calcium soap is superior to magnesium soap since the yield
of fuel is greater; the #2 range distillate has less of an
initial kinematic viscosity than the total distillate and
both distillates have the same kinematic viscosity after the
polymerization tests; the lighter the fraction of the total
distillate the smaller the
initial kinematic viscosity and
the greater the viscosity increase; storage time of fuel had
little effect on polymerization properties.
The second set of experiments showed that addition of
Ca(OH)2 prior
to . pyrolysis
decreased thickening while
limiting yield.
The third set of experiments
found the best soap
production method investigated was the fusion method with no
solvent.
Other methods tried were the precipitation process
and using both toluene and water at different concentrations
as solvents in the fusion method.
The fourth set of experiments showed that it would be
desirable for the safflower oil to contain more oleic acid
esters than linoleic or linolenic acid esters.
Optimization of the process
showed that addition of a
stoichiometric amount of Ca(OH)2 for fusion soap production,
and no Ca(OH)2 added prior to pyrolysis, was most desirable.
The soap-pyrolysis
fuels worked better as fuels in a
diesel engine than safflower oil.
I
INTRODUCTION
In
the
insecurely
averted
predict.
was
mid-1980 *s
between
and
an
an
energy
energy
The energy
triggered
the
by
United
States
crisis
it
had momentarily
could
not confidently
future
it
crisis that
a
sudden
lasted from
but
brief
was
poised
1973 to 1978
Arab embargo upon
petroleum exported to the United States, quickly followed by
several years of sharply rising world petroleum prices.
United States was forced to recognize it had
dependent
upon
satisfy its
imported
own
existing
domestic
appeared
the
rapidly
rising
energy
United
petroleum within
petroleum
might
decades.
to take
the place
no
demands
longer
from its
More ominously,
face
a
shortage
it
of
Experts remained divided
and uncertain about the future [1].
of energy
become heavily
could
energy
production.
States
a few
and
The
An
alternative source
of fossil fuels might someday
help alleviate this problem.
Vegetable oils show
extenders for
much
diesel fuels.
promise
as
replacements or
Theoretically, a farmer could
plant a fraction of his land in oil producing crops
the
fuel
produced
Seed oils such as
attention
as
from
this to run his entire operation.
safflower oil
direct
and use
have received considerable
substitutes
suitable for use in a diesel engine.
or
additives
Safflower oil
for fuels
has the
2
attraction of
being able to be grown easily in Montana.
would be of great interest to
see
safflower
wouldn't have
prices
and
oil
to
replace
rely
finite
the agriculture
It
community to
diesel fuel since farmers then
on
fossil
fuels
availability,
and
with fluctuating
could
then be more
independent and in charge of their own destiny.
Researchers
estimate agricultural fuel needs for the United States could
be met by planting approximately 10%
with oilseed
crops
[2].
of the
total cropland
As an added benefit, vegetable oil
processing yields meal high in protein for animal
by-product
[3].
Despite
the
diesel fuel there
shows
feed as a
that
advantages
of
are
economic
diesel
still
prices
using
would
vegetable
problems.
oil for
Collins
have to double or triple
before even the best vegetable oils would become competitive
[4].
In
general,
if
a
raw material such a diesel fuel
becomes scarce,
its price will rise.
Since the relationship
between
and
absolute and a sign of
price
supply
is
not
scarcity is not a complete absence of material
but rather a
perception of dwindling supply by the consumer, the price of
diesel fuel is always under close scrutiny [5].
The
vegetable
chemical
oils
difference
is
also
between
responsible
diesel
for
vegetable oils form carbon deposits inside
engines
and
polymerize on
(2)
vegetable
contact with
fuel
and
problems:
(I)
direct injection
oil carried into the crankcase
the lubrication
oil leading to
3
eventual
engine
failure
related to differences in
[6].
These
the chemical
problems
are both
structure of diesel
fuel and vegetable oils.
Vegetable oils
are esters formed by the combination of
glycerol with three fatty
straight-chained,
saturated or
unsaturated.
double
bonds
carbon
double
bonds.
safflower is
and
given in
acids
fatty
be without any
unsaturation is a measure of the
fatty
Table I.
can
engines or
acid
distribution
be
solved
by
is to
change in physical
characteristics
vegetable
a
significant
of
deposits
appears
to
prevent
replace one large
The principal
reduction
the
in
This
formation
viscosity
lowering in
of
carbon
[7].
Table I. Fatty Acid Distribution of Safflower Oil
Fatty Acid
palmitic
stearic
arachidic
oleic
Iinoleic
Iinolenic
The
the transesterified
compared to the unmodified vegetable oil.
viscosity
either using
by transesterification.
triglyceride ester with three smaller esters.
is
of
It has been found that the
of transesterification
oil
acids are
that can be either
Saturated would
The
carbon deposition problem
indirect injection
These
monocarboxylic
carbon
net effect
acids.
Percentage
6.4
3.1
0.2
13.4
76.6-79.0
. 0.04-0.13
Composition
Cl 6Hs 2 O2
Cl 8H 3.6O2
C20H4002
Cl8H 3OO2
Cl 8Hs 2 O2
Cl8H30O2
[8]_______
No. of Double Bonds
0
0
0
I
2
3
4
Transesterification does
previously mentioned
of
polymerization
attacking
a
not
necessarily
polymerization problem.
is
thought
to
occur
carbon
double
bond
acting as a catalyst
in the
reaction [9].
polymerization
problem
should
modification of the chemical
The double
prevent the
The mechanism
by
oxygen first
with the carboxyl groups
be
Therefore, the
able
structure
to be solved by a
of
vegetable oils.
bonds could be removed by hydrogenation but this
results in a fuel
that
Decarboxylation or
is
a
solid
removal of
at
room temperature.
the carboxyl groups could be
accomplished by thermal means with
or
without
the
aid of
catalysts.
A
project
primarily with
his results
found that
that
started
with
decarboxylation and
and continued
work
by Hiebert dealt
this investigation used
work in this area
by converting
the safflower
[10].
oil to
Hiebert
a soap and
■
then
pyrolyzing
the
fuel worked the best
modification
of
soap
of
the
to obtain a liquid diesel range
the
techniques
he
tried.
This
chemical structure was accompanied by
relatively low acid numbers (a
measure
of
the
free fatty
acid content) and significant unsaturation (a measure of the
carbon double bonds).
thickening effects
This procedure
significantly reduced
measured by kinematic viscosity.
increase in kinematic viscosity in 64 hours is
failure in lubrication oil tests
[11] .
A 375%
considered a
This current project
attempted to improve and learn more about the soap-pyrolysis
5
process
for
making
fuel
used in a diesel engine.
from safflower oil that could be
6
RESEARCH OBJECTIVES
The
experiments
performed
for
this
research have been divided into five tasks.
soap-pyrolysis
The purpose of
the first task was to investigate the effects of a number of
pyrolysis process variables that
might have
to
Soap
fuel
yield
and
quality.
addressed in this task; in fact,
safflower
oil,
linoleic
acid
production
rather than
was
not
starting with
(its principal component at
76.6-79.0%) was used to make the
simpler, representative
been important
soap
[12].
' By
using a
soap, the effects of the processing
variables were expected to be more apparent and not confused
by the
acids.
presence of
High
speed
polymerization tests.
soaps from
a number of different fatty
lubrication
The
oil
was
used
for
the
high speed lubrication oil is a
thirty weight oil.
Six specific process
this
first
task.
variables
First,
soaps
were
investigated for
that had been aged were
compared with fresh soaps to see how this affected the fuel.
Second, because
the soap
making could be simplified if the
glycerol did not have to be removed prior to pyrolysis,
without glycerol,
with an
amount added
soap
equivalent to that
produced in the soap making, and with an intermediate amount
consistent with
partial removal
were compared.
fatty acid content of the soaps was varied
to see
Third, the
how this
7
variable affected
the fuel.
soaps were compared since
literature indicates
have similar properties.
probably
have
lower
Fourth, calcium and magnesium
However, the magnesium soaps would
pyrolysis
change the nature of the fuel
fuel fraction of the
distillate
Also,
were
[13].
compared to
in
different
compared
with
which
fractions
one
its
polymerization
properties.
another.
fuels and
of
the
Sixth, any
an effect
Oxidation from contact
with air could have caused polymerization.
on fresh
the total
polymerization
changes in the fuel due to storage might have had
on
could
Fifth, a typical diesel
differences
characteristics.
distillate
temperatures
distillate was
for
these soaps
Tests
were made
after they had been stored for a period
of time.
The experiments completed for the second
pyrolysis
with
reactant.
and
without
Linoleic
safflower oil
acid
the
was
use of calcium hydroxide
again
used
fatty acids.
to compare fuels made
use of
than
soap from
a number
Another purpose of this task was
from pyrolysis
with and
without the
calcium hydroxide reactant with both medium and high
speed lubrication oils.
thirty
rather
to prepare the soaps so as to not confuse the
processing variables by the presence of
of different
task compared
weight
oil
forty weight oil.
and
The
High
speed
medium
addition
lubrication
oil
is a
speed lubrication oil is a
of
calcium
made to the soap just prior to. pyrolysis.
hydroxide■was
8
The experiments
for the third task concentrated on the
soap production process.
the soaps
the fuel
for this
yields
different
Safflower oil was used
task since
obtained
soap
it was important to compare
from
production
to prepare
the
vegetable
methods.
A
oils using
multi-step
precipitation process was compared to the single-step fusion
method for
aqueous
the production
solvents
investigated
of the soaps.
with
with
the
varying
fusion
In addition, non-
concentrations
were
method along with use of no
solvent for the process.
The fourth task attempted to investigate the importance
of the
degree of
research in
safflower
vegetable oil
this
oil
lab
to
a
has
unsaturation.
indicated
oils,
but
more
various
double bonds
products
for
degrees
to
of
soaps
any
the polymerization of
information was necessary [14]
approach used for this task involved
with
converting the
soap prior to pyrolysis significantly
decreases the role of unsaturation in
lube
that
Exploratory
The
converting fatty acids
unsaturation or number of carbon
and
then
differences
comparing
that
the pyrolysis
might be found.
High
speed lubricating oil was used for the polymerization tests.
The fatty acids used in this task would be different only in
the number of carbon double bonds.
Ihe
last
task
was
to
optimize
the
soap-pyrolysis
process for making fuel suitable for use in a diesel engine.
The
results
from
the
first
four
tasks
helped identify
9
important process
optimized.
information
It
variables and
was
now
together
techniques and levels of
process for
and
the
techniques that were to be
purpose
determine
variables
for
to
the
put
most
all
this
desirable
the soap-pyrolysis
making diesel fuel from safflower oil.
10
EXPERIMENTAL DESCRIPTION
Soap Production
Two methods
were used
to prepare
experiments.
The fusion method
these runs.
The
was
the soaps for these
employed
for
most of
process consisted of the following steps.
First, 200 grams of acid (oleic,
linoleic, or linolenic)
safflower
IOOOC.
oil
was
consisting of the
magnesium
heated
alkali
hydroxide)
to
hydroxide
with
Second, a solution
(calcium
or
hydroxide or
without
a
(demineralized water or toluene) was vigorously
Magnesium hydroxide
was only
some of
the
soaps
stirred in.
Toluene was only used
the soap production experiments.
the soaps were allowed to cool
After
solvent
used to prepare the magnesium
soaps in the pyrolysis studies.
solvent in
had
dried
or
which took
as a
Next,
several minutes.
they were ground into uniform
chunks using a hand grinder.
The precipitation method was
soaps.
First, 200
grams (in
acid was heated to IOOOC.
sodium
hydroxide
stirred in.
Third, the
and
This
some cases
Second, a
demineralized
process
sodium soap
also used
formed
mixed
with
a
180) of linoleic
solution consisting of
water
a
solid
was vigorously,
sodium soap.
was separated into two batches each
of which was dissolved in 500 ml of
then
to prepare the
solution
of
water.
calcium
Each batch was
chloride
and
11
demineralized
water
and
exchange precipitated
stirred.
out the
This
process
calcium soap.
of ion
Last, the two
batches of soap were combined and
washed with demineralized
water to
or calcium chloride since
remove any
sodium soap
they are both water soluble and the calcium soap is not.
Soap reactions
were performed
at atmospheric pressure
in a 1000 ml glass batch reactor with a glass top.
The 1000
ml reactor and glass top were both purchased from Ace Glass,
product Nos.
6506—35 and
6485— 16.
A thermocouple measured
the temperature and a powerstat was used to control the heat
supplied
to
mantel.
the
reaction
vessel
by
The setup for the soap making
a
electric heating
process is
given in
Figure I.
Pyrolysis
Pyrolysis
is
the
process
condensing the resulting vapors
of
to obtain
Literature indicates
that pyrolysis
to the
aldehydes (ECHO)
formation of
heating
the soaps and
the diesel fuel.
of calcium soaps leads
and ketones (RCHORi)
shown in the following reactions.
(RCOO)2Ca + (HCOO)2Ca = 2RCH0 + 2CaC03
(RCOO)2Ca + (RiCOO)2Ca = 2RC0R1 + 2CaC03 [15]
Maximum liquid
was
chosen
to
temperatures were
prevent
Pyrolysis reactions
damage
were carried
about 8150F; this maximum
to
the
heating
mantel.
out in a 500 ml stainless
steel reactor vessel and then this was replaced by a 500 ml
12
S tirrin g M o to r
Condenser
Coid
Water
Out
—
Glass Top
Coid
Water
in
-------
Glass Reaction
Vessel
Soap Heater
Figure I. Soap-Making Process Apparatus
13
glass reactor after several
studies.
The glass
project.
Nos.
experiments into
reactor was
used for the rest of this
These were both purchased from Ace
6497—05
and
6927—22.
stainless steel and
glass
the pyrolysis
The
Glass, product
apparatus used with the
pyrolysis
vessels
is
given on
Figure 2 and Figure 3.
The reason the stainless steel vessel was chosen in the
first place was the ease of
vessel.
The change
to the
implemented because the
lower with
with the
(DNRC
Grant
vessel
Agreement
study.
The possible
yields
steel vessel
during
the
greater
reasons for
discussed in the Aged
Versus
were significantly
than those obtained
previous investigation
RAE-84-1041)
resulted in a significantly
to the glass
glass vessel in this study was
product
the stainless
glass
cleaning compared
[16].
This
yield
in
change
the current
this yield difference is
Fresh
Soaps
section
of the
Pyrolysis Studies.
The soap
was first put into either the stainless steel
or glass pyrolysis apparatus and a glass connector
the top
then
and was
emptied
insulation
was
attached to
into
a
packed
the condenser.
separatory
funnel.
around
top
apparatus, to minimize heat loss.
was a two-necked flask.
the
The 500
of
came out
The condenser
Fiberglass
the pyrolysis
ml glass reactor
The first neck led to the condenser
where the vapor temperature was measured and the second neck
was used to insert a thermocouple which measured the liquid
14
S tirrin g Motor
^ V a p o r Temp
Cold Water
OuK
Condenser
Stainless Steel
Reaction Vessel
„
Pyrolysis
Heater — -
Cold
Water
Collection
Vessel —
Figure 2. Pyrolysis Apparatus With Stainless Steel Vessel
15
Vapor Temp
Cold Water
Condenser
H eater
Glass
Reaction
Vessel
Cold
Water
In — "
Collection
Flask —
Figure 3. Pyrolysis Apparatus With Glass Vessel
16
temperature.
A
quarter-inch,
glass
covered
magnetic
stirring bar was used at the bottom of the flask to keep the
soap well mixed.
glass top
was also
three necks.
from
The 500 ml stainless steel reactor with a
the
pyrolysis.
The first neck was where
motor,
thermocouple to
third neck
used for
the
second
measure
the
attached to
condenser where the
the stirrer
neck was
used
went in
to insert a
liquid temperature, and
a glass
vapor
The glass top had
the
connector which joined the
temperature
was
also measured.
Heat was supplied by a different heating mantle than the one
used for soap production.
Polymerization
The
simulate
purpose
of
crankcase
polymerization
oil
conditions
determining thickening effects
percent concentration
measured
in
1500C
in
100
ml
(cupric
catalyst and
The.
Ten
Unocal
to
in a diesel engine for
to
adding
a
5 weight
The kinematic viscosity
used as
the
measure
of
batch, reactors
using
50
ml of
parts per millon of a soluble copper
or
CuAcAc)
was
used
as a
oxygen was. bubbled through at the rate of 3 cc
per 15 seconds.
was
was
reactions were carried out at
acetylacetonate
.
Super HDll
was
glass
lubrication oil.
form
of the fuel.
centistokes
thickening effects.
due
experiments
.
The. high
SAE 30W
SAE
.
40.
....
speed lubrication
while the
These
oil was Amoco
medium speed lubrication oil
commercial
lubrication oils
17
contain
lubrication
polymerization.
oil
Since
effects
commercial lubrication
no attempt
was
mechanisms
of
were put
made
additives
of
a
vegetable
affect
oil
fuel on
oil were the focus of this research;
to
study
how
polymerization.
into
that
additives
affect the
The 100 ml glass reactors
custom-built
constant
temperature bath.
Eight different samples could be tested simultaneously.
apparatus was arranged so
The polymerization
bath was
under a hood.
apparatus is shown in Figure 4.
pieces of equipment were
bath and
that the
used for
all were from Ace Glass.
The
Several
the constant temperature
These w e r e :
adapter #11
THD 24/40, product No. 5261-37; test tube, product No. 875202; 300
mm pore
c tubes,
product No. 7202-16;
bushing, product No. 7506-02;
24/40 bearing
11 mm nylon
c, product No.
8042-115; adapter inner, product No. 5028-30.
Evaporation
of
vegetable
oil
fuels
polymerization apparatus was considered.
in
the
Evaporation of the
soap-pyrolysis fuels would leave the lubrication oil left to
be measured for thickening.
Many
the
oil
high-speed
viscosity
lubrication
approximately
polymerization
tests.
the
The
contained the lubrication oil
fuel.
Evaporation
previous work [17].
was
not
of the
experiments with
ended up with a kinematic
same
as
the
control
control
polymerization tests
and
added soap—pyrolysis
no
considered likely because of
While current work had measurements for
kinematic viscosity initially and after 48 hours in the
C i r c u la t o r
Sample
Holder
Valves
Thermometer
Figure A. Polymerization Apparatus
19
polymerization
apparatus,
previous
work took measurements
every 10-12 hours up to 72 hours.
The earlier
work showed
the kinematic viscosity continued to increase above the neat
lubrication oil after the
evaporation
of
possibility.
A
soap-pyrolysis
viscosity at
fuels
was
400 C .
used
to
in
was
only
measure
a
the
remote
kinematic
The technique used was to add 8 ml of sample to
seconds
was
for 10
recorded
on
minutes.
multiplied
by
a
calibration
kinematic viscosity
Canon-Fenske
This viscometer
The initial
number
in centistokes.
viscometer,
No.
to
The
L799,
This was then
convert
it into
viscometer was a
size
was calibrated
viscosity was
Then the
a stopwatch as the oil
traveled between two marks on the viscometer.
21859.
Therefore,
A drawing of the viscometer is given in
the viscometer and let in stand
time
measurement.
Future work in this area should be considered.
viscometer
Figure 5.
48 hour
350, report No.
at 40«c
and IOOOC.
measured and then the sample was
put into the 100 ml glass reactors which were placed in the
constant temperature bath for 48 hours.
time the
final viscosity reading was taken.
work done on this
Natural
Resources
Processing
[18],
it
At the end
of
was
project
and
the
Montana
Under previous
Department of
Conservation (RAE-84-1041, Chemical
Vegetable
shown
for
of this
Oils
to
Prevent Polymerization)
the thickening properties of the fuel
were apparent after 48 hours; the
kinematic viscosity using
straight safflower oil had increased 110% from about 95-200
20
C irc u la tin g
Motor
•Thermometer
Viscometer
figure 5. Viscometer and Temperature Bath
21
centistokes
while, using
the
safflower soap decomposition
product had led to approximately a
22% increase
from about
90-110 centistokes.
Acid Number
The
oil
product is the
relative measure of free fatty acid content.
The procedure
used
acid
is
number
described
dissolving the
heating until
of
in
diesel
ASTM
range
boiling.
The
base
oil
D
or
466-78.
product
this
in
95%
involved
ethanol and
Then a titration procedure was used
with phenothalin as the
found.
an
indicator
used
for
until
the
the
endpoint was
titration was potassium
hydroxide with a normality of close to 0.1.
Iodine Value
The
iodine
unsaturation or
value
ASTM
D
the
carbon double
The method used was
to
is
of
bonds present in the sample.
procedure performed according
This
method has been shown to give
for oils
conjugated double bonds.
and their
derivatives with non—
These are the only type present in
■safflower oil and its derivatives.
20 ml
measure
the Wijs
1959— 69.
accurate results
relative
The procedure was to add
of carbon tetrachloride and 25 ml of Wijs solution to
the diesel sample.
one hour.
After
This was
this hour,
to be
stored in
the dark for
20 ml of 15% potassium iodide
and 100 ml of demineralized water were added.
This was then
22
titrated
with
sodium,
the indicator.
that were
thiosulfate solution using starch as
Simultaneously,
made up
of 25
blanks had
to be titrated
ml of Wijs solution, 20 ml of the
15% potassium iodide solution,
and 100
ml of demineralized
water.
Nuclear Magnetic Resonance
The
Carbon-13
Nuclear Magnetic Resonance Spectrometer
used was a Bruker WM-250 with a Aspect
NMR
spectra
provided
qualitative
2000 computer.
information
about
functional groups attached to the carbon molecules.
parameters
had
to
be
tested
in
appropriate set to use for these
order
The
the
Several
to determine the
oil products.
These were
receiver gain, receiver delay, and the number of scans.
The
receiver gain varies with the concentration differences.
An
appropriate receiver gain would be set so the free induction
decay is two centimeters above and below the screen midline.
A typical
receiver gain
for these
samples was
receiver delay is the time between scans.
set by
choice of
the other
delay was around two
scans should
be taken
seconds
parameter.
for
these
The
This was already
A typical receiver
samples.
Enough
to give a good signal to noise ratio
which varies by the square root of the number of
the concentration
1600.
scans.
As
of the sample is increased, the lower the
number of scans that are needed.
The
1.5 ml
sample of oil
was mixed with 0.5 ml of solvent (deutero chloroform).
This
23
fairly high concentration required only 500 scans which gave
good results.
There
were
several
areas
of
interest
spectrum.
The unsaturation region occurred
parts per
m i I Ion.
Ketones
Esters
were
from
were found
NMR
between 110-140
mil Ion.
All peaks
at 170
parts per
205 to 208 parts per m i I Ion.
Finally, solvent had three peaks with
parts per
the
The carboxylic acid region was found at
180 parts per mil Ion.
mil Ion.
in
the middle
one at 77
were done in comparison with
the solvent.
DistiIlation
Distillations
characteristics
were
of
one
done
set
to
of
determine
products in the pyrolysis
tests.
The method used is described in
Davis,
a
Chemical
Engineer
at
volatility
ASTM D
86-78.
Tom
the Farmers Union Central
Exchange in Billings, MT, provided these
typical ranges for
diesel fuel:
I. Typical Ranges for Diesel Fuel Types
Diesel Fuel Type
heavier-#3
regular-#2
lighter-#!
Initial(OF)
Endpoint(OF)
500
. 430
315
720
625
520
It should be noted these ranges vary with the season.
The
apparatus
used
included
reactor used in pyrolysis, a vapor
the
same
500 ml glass
temperature thermometer,
24
a water
cooled condenser,
and a glass recovery flask.
procedure used was to heat the organic product
vapor would
condense and
product was distilled in
the
temperature
was
temperature was
brown.
slowly.
The
collect in a recovery flask.
The
order of
lightest tp
increased.
material to come off
was
a
The
very
increased the
a green
tint.
heaviest as
first and lightest
light
yellow.
As the
color of the product went to
The heaviest diesel that came off
color with
The
was a
dark brown
The differently colored products
did not stay separate as they mixed in the recovery flask.
Yield
Yield was calculated two wa y s .
were used
which
When the
various acids
to make the soap and in the soap production study
used
computed per
safflower
oil
to
make
of organic
the weight of soap
second way
charged
it
yields
starting
the
This
minus
was done
the
yield was
was charged since
was
important
water
by taking
and dividing
this number
it by
by taking
product.
A
used when safflower oil
to
compare
vegetable oils.
the fuel
This yield was
calculated per 100 grams of total safflower oil
was done
yield was
product collected and dividing it by
to calculate
from
the
100 grams of dry soap since the effects of the
process variables were important.
the weight
soap,
used.
This
the weight of organic product collected
the soap
was multiplied
charged for
pyrolysis.
Then
by the total weight of soap and
25
divided by the weight of safflower charged.
26
RESULTS AND DISCUSSION
Pyrolysis Studies
Aged Versus Fresh Soaps
The
first
set
of experiments compared polymerization
properties of fuels prepared from soaps
for two
months prior
that had
to pyrolysis with fuels prepared from
fresh soaps aged three days before pyrolysis.
prepared by
the fusion
hydroxide mixed
High
two
different
numbers for
with 71
speed lubrication
for the polymerization tests.
these experiments.
The soap was
method with linoIeic acid and a 50%
excess amount of calcium,
demineralized water.
been aged
Table 3 gives
grams of
oil was used
the results of
The fuels obtained from these soaps aged
times
had
similar
properties.
The acid
the fuels dried two months were both 0.7.
This
compares to acid numbers of 1.0 and 1.4 for fuels from soaps
dried
®
three
days.
The
iodine
ficantIy from one another.
values
did
The fuels from
not
differ
soaps dried
two months
had iodine values of 136 and 141 while the fuels
from soaps
dried three
150..
days had
iodine values
of 149 and
Therefore, no significant differences existed relative
to acid numbers and
results of
iodine values.
spectra confirmed
small acid numbers and significant unsaturation.
The yields of the fuel from
was not
NMR
comparable since
the soaps
the pyrolysis
aged different times
apparatus had been
27
changed from stainless steel to glass and will
later in
this section.
soap and pyrolysis
apparatus
for
The
was
the
be discussed
yield is per 100 grams of dry
performed
in
the
stainless steel
soaps : dried three days (fresh) and the
glass apparatus was used
for
fuels
from
soaps
dried two
months (aged).
Table 3. Fuels from Aged and Fresh Soaps_____________ _______
Viscosity
Initial
Final
Soap % Y ieId Iodine V a l . Acid No.
(cSt)
(cSt)
fresh
16.7
149
1.0
65.8
68.1
109.1
110.1
fresh
17.4
150
1.4
70.1
69.9
105.8
107.4
aged
49.5
136
0.7
82.8
80.1
115.3
108.8
aged
45.0
141
0.7
78.5
77.5
105.1
107.4
The
vapor
temperatures
observed were much lower with
the stainless steel compared to the glass apparatus used for
pyrolysis.
When the
stainless steel
used the maximum vapor temperatures
200-2IOOF.
pyrolysis vessel was
were
in
the
range of
With the glass vessel the vapor temperatures got
as high as 4000F.
The liquid product recovery
was
considered
complete
when
formed several
it fell to about 10 seconds
between drops of product . into the
lightest and
first of
layers and
the product
separatory funnel.
The
to come out was a clear
28
water
layer.
The
organic
layer
was
next.
When the
stainless steel pyrolysis vessel was used this organic layer
took the
The
form of
layers
were
two distinctly
a
light
different colored layers.
yellow and a dark brown.
These
organic layers did not stay separated when they mixed in the
separatory
funnel.
When
the
stainless
apparatus was used, but the rate at which
was
reduced,
the
organic
steel pyrolysis
heat was supplied
product took another form.
The
product was then formed in one distinct yellow—brown mixture
with the
color yellow
was different again
used.
The
more apparent.
when
product
the
then
glass
This organic product
pyrolysis
vessel was
usually took the form of a more
uniformly colored brown liquid product with a green
tint on
the last of the product.
There
is
difference in
a
possible
yield between
pyrolysis vessels.
explanation
the glass
of
As the soaps were heated they vaporized
take place
a longer
top increased
the
longer
to
have
a
the volume causing
holding
time.
polymerization was possibly catalyzed • by components
stainless steel-
First,
retention time with the stainless steel
apparatus since the glass,
vapor
in the
This could be more drastic when the stainless
steel pyrolysis apparatus was used for two reasons.
there was
extreme
and stainless steel
in both vessels and polymerization could
vapor phase.
this
vessel.
The polymers
Second,
in the
could then condense
and, since they, would be much less volatile
than the soaps,
29
they would char rather than vaporize.
The thickening
soaps
dried
kinematic
two
different
viscosity
months increased
increase was
dried three
when
effects of
the
an
were
fuels
average
of
the
after
are
48
and 106.3
from
centistokes)
the average
examined
hours
in
were
Table 3.
very
close
The
and
soaps aged
two months
The average initial viscosities of
soaps
were
while
centistokes compared to 109.6 and
three days.
fuels
soaps dried two
observation takes place
106.6 centistokes for the fuels from
compared to
from
The
48 hours for the fuels from soaps
numbers
taken
different.
36.9%
An interesting
actual
averaged 112.0
times
the
57.8% after
days.
viscosities
for
the fuels obtained from the
dried
much
two
greater
months
than
(81.5
and 78.0
those of fuels from
soaps dried three days
(67.0 and
shows that
soaps aged the longer period of time
gave
fuels from
fuels
These
with
fuels
greater
from
initial
soaps
aged
approximately the same kinematic
Some of
70.0 centistokes).
kinematic
different
This
viscosities.
times
viscosity after
were
at
48 hou r s .
this variation in initial viscosity might have been
due to the switch that occurred in the pyrolysis apparatus.
Glycerol Content of the Soaps
The second set of experiments for the pyrolysis studies
investigated the
the fuel.
with
effect of glycerol content of the soaps on
The soap was again prepared by the
linoleic
acid
and
a
50%
excess
fusion method
amount of calcium
30
hydroxide mixed with 71 grams of demineralized water.
High
speed lubrication oil was used for the polymerization tests.
The results are given
in Table
4.
Fuels made
from soaps
without added glycerol were compared against fuels made from
soaps with 23 and 46 grams of glycerol added.
was added
prior to
pyrolysis.
glycerol that would
every three
glycerol
be
moles of
is
the
produced
from
calcium soap.
total
Correspondingly, 23
There is
amount
The glycerol
only one mole of
vegetable
oils for
Therefore, 46 grams of
that
would
be
produced.
grams represents half the amount in the
soap.
Table 4. Fuels from Soaps With Different Amounts of Glycerol
Viscosity.
Glycerol
Initial
Final
Added
%Yield
Iodine V a l . Acid No.
(cSt)
(cst)
none
none
23g
23g
46g
46 g
12.6
13.7
11.6
10.7
5.8
4.5
134
124
129
131
—
—
—
—
—
—
1.3
1.0
1.9
1.9
1.6
1*7
The glycerol content of
properties of
the soaps
the fuel significantly.
the fuels from soaps without
1.0.
Acid, numbers
glycerol
1.9.
100.3
107.0
110.9
109.8
108.5
107.1
did not
change the
The acid numbers for
added
were
1.3 and
for fuels made from soaps with 46 grams
of glycerol added were 1.7 and
fuels from
81.8
72.7
74.9
75.7
77.8
76.9
soaps with
1.6.
Acid numbers
for the
23 grams of glycerol added were both
Iodine values showed
fuels that could be tested.
this
same
consistency
for the
Since the yields were only 5.8%
31
and 4.5% for the
added not
fuels obtained
enough product
determination.
Iodine
with 46
grams of glycerol
was available for an iodine value
values
for
the
fuels
without added glycerol were 134 and 124.
from soaps
Iodine values from
soaps with 23 grams of added glycerol were 129 and 131.
NMR spectra
The
once again confirmed the findings of small acid
numbers and showed a significant unsaturation region.
Thickening effects
the addition
of the
of glycerol
fuels were
to the
not affected by
soaps prior to pyrolysis.
The kinematic viscosity for the fuel from soap with no added
glycerol increased
Likewise»
the
an
average
kinematic
of
3.4.2%
46.6% and
that glycerol content
performance of
48 hours.
viscosity increased for the fuels
made with 23 and 46 grams of glycerol
average of
after
added to
the soap an
39.4%.
This leads to the conclusion
of
soaps
the
does
not
affect the
the fuel in a diesel engine with regard to a
tendency to polymerize.
Although the
glycerol
significantly affect
content
the polymerization
decrease the yield calculated per
Yield was
steel
low because
apparatus.
glycerol added,
of
100
pyrolysis was
Yields
with adding
obtained
the
did not
properties, it did
grams
of
dry soap.
done in the stainless
from
23 grams,
soaps
with
no
and adding 46 grams
were averaged to be 13.2%, 11.2%, and 5.1%.
decreasing yield
soap
The
pattern of
with increasing glycerol added to the soap
leads to the conclusion
that glycerol
is not
desirable in
32
the soap
prior to pyrolysis.
Fatty Acid Content of Soaps
Next in the pyrolysis
studies was
the fatty acid content of the soaps.
linoleic
acid
polymerization
This
was
used
tests
presented
a
an investigation of
The fusion method with
to , produce
used
the
soaps
and
the
the high speed lubrication oil.
problem
when
an
attempt
increase the fatty acid concentration.
was
made to
The first try was to
heat the soap to turn it to a liquid form and add
the fatty
acid
Then
which
would
have
corresponding acid
was
the
soap
been
linoleic
number could
would
starting to pyrolyze.
not
be measured.
return
Nex t ,
it
acid.
to
was
linoleic acid to prepare the soaps.
the
The problem
a liquid form before
tried
to
add excess
Several runs were tried
using excess linoleic acid to complete
the reaction.
This
.
technique proved
to be a failure.
it just remained in
the
liquid
Solid soap never formed,
state.
Finally,
it was
determined how much excess calcium hydroxide was required to
produce a solid soap.
that solid
it was discovered
soap required at least stoichiometric amounts of
calcium hydroxide.
50% excess
After several runs
The two soaps compared in this study had
and stoichiometric amounts of calcium hydroxide.
This was not the initial comparison that was to be
made but
it was the best that could be accomplished.
The results
fuel does seem to
are given
be
in Table
changed
5.
slightly
The nature of the
by
the
amount of
33
excess calcium
hydroxide used.
The acid numbers for fuels
from soaps with a stoichiometric amount of calcium hydroxide
were 6.1
and 8.1.
Acid numbers for fuels from soaps with a
50% excess amount of calcium hydroxide were both
0.6.
This
showed
were
also
a
changed:
slight
181
and
stoichiometric
difference.
184
amount
for
of
Iodine
fuels
values
from
calcium
soaps
hydroxide
with
compared
a
to
iodine values of 132 and 139 for fuels from soaps with a 50%
excess amount
of calcium
hydroxide.
This means fuels made
from soaps with less calcium hydroxide had more unsaturation
and more
free fatty acid content.
Yields were not affected
by the calcium hydroxide in the soaps.
of
dry
soap
were
stoichiometric
hydroxide.
and
averaged
the
The glass
50%
Yield per
100 grams
to be 42.6% and 49.1% for the
excess
amount
of
calcium
pyrolysis apparatus was used for all
these experiments.
Table 5. Fuels from Soaps with Different Free Fatty Acid
Contents
Viscosity
Initial
Final
Ca(OH)? %Y ieId Iodine V a l . Acid No.
(cSt)
(cSt)
50%xs
43. 8
132
0.6
76.7
77.7
106,3
101.9
50%xs
54.5
139
0.6
75.9
77.6
105.1
107.4
0%xs
41.2
181
6.1
76.5
76.7
117.8
118. I
0%xs
43.9
184
8.1
76.9
76.9
120.0
121.4
34
The thickening
the free
of the
fatty acid
fuel did seem to be affected by
content of
viscosities increased
the soaps.
The kinematic
an average of 54.5% and 36.9% for the
fuels from soaps with the higher free fatty acid content and
the lower
free fatty
difference in
this
acid content respectively.
increase
viscosities of the fuels.
was
higher
free
fatty
polymerize faster.
significant if
the
final kinematic
The initial kinematic viscosities
were all within 1.8 centistokes.
the
in
The major
acid
This
Fuels made from soaps with
content showed a tendency to
probably
higher free
would
fatty acid
have
been more
contents could have
been achieved.
Calcium Versus Magnesium Soaps
The fourth
set
compared calcium
of
tests
for
hydroxide
and
The results from the
were used
for the
pyrolysis studies
soaps with magnesium soaps.
soap was prepared by the fusion
magnesium
the
59
process with
The magnesium
33.1 grams of
grams of demineralized water.
other tests
in the
calcium soap.
pyrolysis studies
Magnesium soaps gave such
low yields not enough product was
available for
The stainless steel pyrolysis apparatus was used.
apparatus a typical yield.from other
studies for
were
preferred
to
the
0%.
, Thus,
magnesium
calcium soaps gave a greater yield.
With this
the pyrolysis
the calcium soaps was around 10-15%.
soaps gave yields that approached
soaps
runs in
any tests.
Magnesium
the calcium
soaps since
the
35
For a possible explanation of this
the
reader
is
difference in
pyrolysis
first
referred
yield between
vessels
given
the
magnesium
in
the
discussion of the
the stainless
in
the
section of the Pyrolysis Studies.
that
to
difference in yield
the
Aged
It is
soaps
steel and glass
Versus
Fresh Soap
possible, however,
promotes
vapor
phase
polymerization more than the calcium found in the soaps.
this is
true it
would cause
decreased yield
If
for the same
reason as discussed earlier.
Literature indicates
pyrolysis temperatures
the 1920’s Sato in
of calcium
magnesium soaps
might have lower
than their calcium counterparts.
Japan investigated
and magnesium
that the magnesium soaps
the dry distillation
soaps of soybean Oil.
decomposed
In
at
Sato found
lower temperatures
and gave heavier products than the calcium soaps
[19].
Distillation Results
The pyrolysis
studies next
fractions of the diesel fuel.
fuel corresponding
to the
for polymerization
tests.
Soap
prepared by the fusion method using linoleic acid
and a 50% excess amount of
calcium hydroxide
grams of demineralized water.
up of extremely
liquid
Fuel, and the fractions of the
#2 diesel range were tested with
high speed lubrication oil
was again
investigated the different
diesel
light,
#1,
products.
distillate was 51 weight
mixed withx 71
The total distillate was made
#2,
The
#3,
#2
percent of
and
extremely heavy
fraction
the total
of the total
product and
36
the
fractions
range fuel
product was
that
lighter
fraction were
properties
and heavier than the #2
approximately equal.
compared to
given in Table 6.
the
were
the #2
range and
acid
of
numbers
the
of
Iodine values for the
while iodine
the results are
This comparison proved the differences in
fuel
were
not significant.
numbers for the total distillate were
to
This total
0.8
0.8 compared
and 0.9 for the #2 diesel range.
total
values for
0.6 and
Acid
distillate
the #2
were
136
diesel range
and 130
were 141 and
138.
Table 6. The Total Distillate Versus the #2 Diesel Ranee
Viscosity
Initial
Final
Fuel
Iodine Val . Acid No.
(cSt)
(cSt)
total
136
0.6
82.0
110.1
total
130
0.8
81.0
106.7
#2
141
0.8
62.3
62.2
108.4
108.3
#2
138
0.9
63.6
63.7
106.5
106.5
Thickening
distillate and
effects
the #2
were
different
diesel range
The total distillate had
initial
when
the
total
product were compared.
viscosities
of
82.6 and
81.0 centistokes and final viscosities after 48 hours in the
polymerization apparatus
of
110.1
and
106.7 centistokes.
The #2 diesel range had initial viscosities of 62.2 and 63.7
centistokes and final
viscosities
after
48
hours
in the
37
polymerization
apparatus
of
108.3
and 106.5 centistokes.
The total distillate experienced a 32.5% average increase in
kinematic viscosity after 48 hours while the #2 diesel range
had a 70.6% average increase.
This
shows
the
#2 diesel
range experiences
a greater increase in kinematic viscosity
after 48 hours in
the polymerization
the
total
viscosities
compared
to
distillate
were
the
much
#2
because
greater
diesel
apparatus compared to
the
for
range.
initial
the
kinematic
total distillate
After 48 hours the #2
diesel product gave about the same viscosities as
the total
distillate.
Next,
the
total
distillate
was separated into three
fractions: lighter than #2, #2, and heavier than
These three fractions were then compared.
in Table 7.
fraction had
#2 diesel.
Results are given
Thickening effects were different.
The lighter
an 93.0% increase in kinematic viscosity.
The
regular VA2 diesel increased an average of 70.6% in kinematic
viscosity.
The
heavy
kinematic viscosity.
fraction
Al I
these
had an 27.2% increase in
increases
hours in the polymerization apparatus.
were
after 48
38
Table 7.
Fuel
Viscosity Results for Diesel Fuel Types
Viscosity
Initial
Final
(cSt)
(cSt)
light
56.4
107.8
regular
(#2)
62.3
62.2
63.6
63.7
108.4
108.3
106.5
106.5
heavy
77.9
99.1
These results indicate that for hot
should be
made up
diesel range.
of diesel
This is what
weather regular #2
on the heavier side of the #2
Tom Davis,
a chemical engineer
at the Farmers Central Union Exchange just outside Billings,
Montana, said is the case.
Storage Effects
Next, storage effects on the fuel were studied.
Fuel fresh from pyrolysis
was tested
pyrolyzed and let stand for ten weeks.
and compared
Polymerization tests
used high speed lubrication oil and the soaps
by the
fusion method
amount
of
calcium
demineralized water.
Table 8.
hydroxide
mixed
were 5.1 and 1.3.
were prepared
acid and
with
a 50% excess
71
grams
of
Data for these experiments is given in
Properties of the fuel were
storage process.
numbers of
with linoleic
to fuel
not affected
by this
For fuel fresh from pyrolysis acid numbers
This
4.2 and 0.6.
compared
to
the
aged
fuels acid
Iodine values for fresh fuels were
39
160 and 147.
The iodine values for the aged
fuels were 143
and 150.
Table 8. Aged Versus Fresh Fue Is
Fuel
Iodine V a l .
fresh
fresh
aged
aged
Viscosity
Initial
Final
CcSt)
(cSt)
Acid No.
160
147
143
150
5.1
1.3
0.6
4.2
77.5
71.7
71.0
85.6
111.0
112.0
108.3
108.0
Thickening effects were also similar.
from pyrolysis
the kinematic viscosity increased an average
of 49.5% after 48
For
fuel
that
For fuel fresh
hours
had
in
been
the
aged
polymerization apparatus.
ten
weeks
the
kinematic
viscosity increased an average
of 39.1%
the polymerization
In conclusion, the fuel aged
apparatus.
for ten week behaved
similarly
in
the
after 48
hours in
simulation
of the
diesel engine.
Results for Pure Linoleic Acid
Next, the
pure linoleic acid was compared to the other
results in the Pyrolysis Studies.
linoleic
acid
was
153
The iodine
which was comparable to the iodine
values found for the soap-pyrblysis fuels.
of 15.5
for the
value for the
The
acid number
linoleic acid was significantly higher than
that of the acid numbers of
the soap-pyrolysis
fuels.
The
polymerization results showed an initial kinematic viscosity
of
96.4
centistokes
and
after
48
hours
one
centistokes with the high speed lubrication oil.
of
178.0
This is an
40
85% increase in thickening of
significantly
more
than
the
linoleic
greater.
which is
that of the soap—pyrolysis fuels.
In addition, the initial viscosity of
much
acid
the linoleic
acid is
Therefore, soap-pyrolysis fuels would work
better than pure
linoleic
acid
in
a
diesel
engine with
regard to tendency to polymerize.
Pyrolysis With and Without Calcium Hydroxide Reactant
The
next
step
after
the
pyrolysis
completed was to investigate pyrolysis with
use
of
calcium
hydroxide
as
a
just prior
the
the
A
review
pyrolysis with the use
of
product in
the
[20].
following
The calcium
to the
heating of
literature indicated that
of calcium
may produce hydrocarbons
were
and without the
reactant.
hydroxide reactant was added
soaps.
studies
hydroxide as
a reactant
Hydrocarbons (RH) would be a
reaction
of
calcium
soap with
calcium hydroxide:
(RCOO)2Ca + Ca(OH)2 = 2RH + 2CaC03
A
charge
of
12.37
stoichiometric amount
water.
grams
of
required
The precipitation
calcium hydroxide was the
if
the
soap making
contained no
process was used to
prepare the soaps in this
group
insure no
hydroxide would be present in the
soaps.
amount
excess calcium
The
of
precipitation
sodium
demineralized water
method
hydroxide
and
a
of
soap
50%
experiments
used
mixed
excess
so
as to
a stoichiometric
with
86
amount
grams
of
of calcium
41
chloride mixed
with 176
grams of demineralized water.
amount of calcium hydroxide
required on
a wet
The
basis could
not be calculated until after the pyrolysis with the calcium
hydroxide
reactant
content of
was
the soaps
performed
because
was unknown.
the
moisture
It was determined after
pyrolysis with the calcium hydroxide reactant
was performed
that the amount of calcium hydroxide required on a dry basis
for the runs with the
Therefore, 12.37
reactant
was
7.16
and
6.36 grams.
grams was a 72.8% and 94.5% excess for the
two runs with the reactant.
The results are giver*
numbers and
iodine values
differences whether
Acid numbers
calcium
for fuels
were 3.2 and 2.3
in Tables
of the
hydroxide
was
numbers were
unsaturation.
No
small acid
significant
The acid
added
or not.
the calcium hydroxide,
fuels when the calcium hydroxide was added,
support the findings of
10.
fuel show no significant
made without
while acid
9 and
4.2 and
0.6 for
NMR spectra did
numbers and significant
differences
could
be
determined from these results.
Table 9. Experiments With/Without Calcium Hydroxide Reactant
Run______
1-w/o
2-with
3-w/o
4-with
Soap
% Y ieId
Iodine V a l .
Acid No.
I
I
2
2
79.7
58.8
73.1
61.4
148
138
145
140
3.2
4.2
2.3
0.6
42
Table 10. V iscosity Results of Fuels With/Without RAmrtant
High Speed Lubricating Oil
Viscosity
Initial
Final
Run
(cSt)
(cSt)
I
76.9
115.7
2
76.6
109.3
3
72.8
115.1
4
77.4
109.2
neat
lub.
oi I
109.3
108.2
Yield
Medium Speed Lubricating Oi I
Viscosity
Initial
Final
Run
(cSt)
(cSt)
I
118.9
200.6
2
118.6
190.9
3
117.2
255.9
4
121.5
197.7
112.5
113.0
was
neat
lub.
oil
significantly
hydroxide was added.
175.0
179.0
decreased
234.2
233.0
when
the calcium
Yields per 100 grams of dry
soap were
79.7% and 73.1% when calcium hydroxide was not used prior to
pyrolysis compared to yields of 58,8% and 51.4% when calcium
hydroxide was
used.
Therefore, yield is adversely affected
by the addition of calcium hydroxide prior to pyrolysis.
Polymerization tests were
high
speed
kinematic
lubrication
viscosity
polymerization
oil.
increase
apparatus
was
without the calcium hydroxide
the addition.
done
with
These
after
greater
both
mediiim and
results
showed the
48
hours
in
the
for
the fuels made
addition than
for fuels with
With the high speed lubricating oil or thirty
weight oil the viscosity increase average^ 54.2% without the
addition of
calcium hydroxide
When the medium speed
was
used
the
and 41.8% with the addition.
lubrication oil
kinematic
or forty
weight oil
viscosity increase averaged 68.7%
43
(ignoring
addition
run
of
viscosity
3)
without
calcium
readings
for
error.
addition
hydroxide.
all
centistokes in run 3 was the
probably in
the
and
Looking
61.8%
at
the
with
final
eight runs the value of 255.9
only one
out of
line and was
These results indicate the addition of
calcium hydroxide prior to pyrolysis is
somewhat beneficial
in reducing thickening effects.
Control
polymerization
lubricating oil
Table 10
without
data indicate
runs
added
were also made using the
soap—pyrolysis
the high
the kinematic
the medium speed
viscosity of the former increased 3.7%
and the latter increased 32.0% in 48 hours.
addition of
The
speed lubricating oil was
much more resistant to polymerization than
oil;
fuel.
soap-pyrolysis fuels
Therefore, the
had comparable effects on
both lubricating oils.
Soap Production
The next set of experiments investigated the effects of
the
soap
production
processes
with
lubricating oil polymerization tests.
had
been
*ne^-h°ds.
only
produced
by
both
the
Until
to
investigate
vegetable
oil
production studies.
was
speed
this time soap
However, linoleic acid had been used
important
high
precipitation and fusion
since it was
certain process variables.
Now it was time to use safflower oil
the
the
since the
yield using
an important variable for the soap
The oil used for these
experiments was
44
mill
run
safflower
Culbertson, MT.
from a
The
number of
raw and was
safflower
oil
safflower
in
was
is
made
up
This was
freshly extracted
The safflower oil was
refrigerated
storage.
expected
to
be
of
safflower
oil
calcium soap and glycerol.
fusion
exactly the
because the reaction of linoleic acid with
calcium hydroxide produced calcium soap and
reaction
Even though
of primarily linoleic acid the
soap-making procedure was not
same.
oil
common varieties.
kept
oil
from Continental Grain Company in
soap-making
Toluene and water at
and calcium hydroxide produced
These
process
water while the
experiments compared the
with the precipitation method.
different concentrations
were used as
solvents with the calcium hydroxide in the fusion procedure.
In addition, use of
no solvent
with the
calcium hydroxide
was investigated.
The
results
guide for
of
deciding the
decided that
the
pyrolysis studies were used as a
techniques to
no advantage
was gained
Second, glycerol was removed by pouring
then
washing
with
magnesium hydroxide
hydroxide
was
use.
by drying the soaps.
off any
demineralized . water.
dramatically
chosen
as
the
First, it was
excess and
Third, since the
decreased
yield, calcium
alkali hydroxide.
The soap
production method is outlined for both the precipitation and
fusion methods in the Experimental Description section.
batches
of
soap
were
prepared
using
the
Two
precipitation
procedure and 180 grams of safflower oil; a solution of 25.7
45
grams
of
sodium
hydroxide
demineralized water
mixed
was then
precipitation procedure
with
added.
added a
77
grams
The next step in the
solution of
26.4 grams of
calcium chloride and 79.2 grams of demineralized water.
sodium hydroxide
was
in
of
stoichiometric
quantity
The
and the
calcium chloride corresponded to a 50% excess.
The remaining
fusion method.
used for
excess.
batches of
soap were prepared using the
Twenty-nine grams of
calcium hydroxide were
all the fusion runs and this corresponded to a 10%
First, two batches of
grams of
safflower oil
prepared with 200
and a slurry containing 29 grams of
calcium
hydroxide
and
Second,
two
batches
more
soap were
52
grams
of
of
demineralized water.
soap were prepared with 200
grams of safflower oil and a slurry with 29 grams of calcium
hydroxide and
132 grams of demineralized water.
batches of soap were
oil and
prepared with
200 grams
Third, two
of safflower
only the 29 grams of calcium hydroxide.
The most apparent difference when the safflower oil was
used instead
of the
linoleic acid
required for the reaction
process.
I.5-2.0
to
take
was the
place
length of time
with
the fusion
With the safflower oil agitation was required for
hours
before
the
reaction
was
complete.
When
linoleic acid was used the reaction was complete in a matter
of minutes.
seemed to
If
this agitation
want to
was not
sustained the soap
separate out into two phases that looked
like safflower oil and calcium hydroxide solution.
46
Soap was then made by
hydrocarbon
as
the
the
solvent
fusion
method
instead of water.
this was tried was that a hydrocarbon such
provide good
soap were
slurry
contact between phases.
prepared with
of
29
grams
with a
The reason
as toluene might
First, two batches of
200 grams
of
but
of safflower
oil and a
calcium hydroxide and 11 grams of
toluene.
Second, two batches of soap were prepared with 200
grams of
safflower oil,
22 grams of toluene.
29 grams of calcium hydroxide, and
Several different approaches
had been
initially tried for the set of runs that used toluene as the
solvent.
The 29 grams of calcium hydroxide were first tried
with 50
grams of
with about 90 ml
Next, 50
toluene.
This produced very little soap
of unreacted
safflower oil
and glycerol.
grams of toluene and 50 grams of calcium hydroxide
were used to produce the soap.
the first
method but
This produced more soap than
it still
left approximately 35 ml of
unreacted safflower oil and glycerol.
The soap
produced by
these two tactics was not enough for pyrolysis.
It was then
determined that toluene in
worked
satisfactorily
hydroxide and
glycerol.
the concentrations
along
left very
with
the 29 grams of calcium
little unreacted
Therefore, there
finally used
safflower oil or
was enough soap for pyrolysis.
The time required for this reaction to take
place was about
2 hours.
The test
results on
the fuels
are given in Table 11.
As in previous tests, all the acid numbers and iodine values
47
for the
fuels produced
were quite similar.
between 0.9
155-170.
amount
by the
Acid
and 2.7.
spectra
unsaturation
were
different
all
soaps
carboxylic
and
the
or
the
values were
free
for
the
the
and
spectra
fuels ranged
in the range of
the fuels
in the
fatty acid content.
spectra
peak
NMR
for
existed between
similar
acid
Therefore,
numbers
Iodine
No differences
of
various soap-making methods
fuels
showed
produced
only
significant
a
NMR
from
slight
unsaturation.
confirmed the findings of the
acid numbers and iodine values.
Table 11. Results of Fuels from Different Soaps With Solvent
Viscosity
Initial
Final
Soap Prep.
Iodine Val . Acid No.
(cSt)
( cSt)
158
162
151
162
166
159
170
161
160
155
157
164
precipitation
precipitation
fusion 52g Hg0
fusion 52g Ha0
fusion 132g HaO
fusion 132g HaO
fusion H g to I .
fusion IIg tol.
fusion 22g t o l .
fusion 22g t o l .
fusion
fusion
The thickening
1.5
2.4
1.9
1.9
2.3
0.9
2.7
2.5
2.1
2.1
1.7
2.4
effects
for
77.3
76.7
78.9
78.4
82.2
81.8
76.0
76.9
74.9
80.3
78.8
79.9
the
fuels
120.2
108.9
109.8
114.3
117.7
116.0
115.4
110.0
112.5
114.8
114.3
116.3
obtained from
soaps made
by different procedures did not show significant
deviation.
The kinematic viscosity increase
in 48
hours for
the fuel
averaged 48.8%
the fuel from the precipitation soap.
from
the
fusion
concentrations
for
the
soap
solvent
with
of
For
demineralized water
52
and 132 grams the
48
kinematic viscosity increased an average of
in 48
hours.
For the fuel from fusion soap with toluene as
the solvent in amounts of
viscosity
42.5% and 42.5%
increased
46.5%.
The final
with no
solvent and
11
in
and
48
22
hours
soap was
grams
the kinematic
an average of 47.4% and
produced by
the fusion method
only the 29 grams of calcium hydroxide
had a kinematic viscosity increase after 48 hours an average
of 45.4%.
Therefore, in respect to thickening effects in a
diesel engine
procedures
the
would
fuels
made
behave
from
similarly
different soap-snaking
in a time period of 48
hours.
Two types of yields were calculated as described in the
Experimental Description
in Table 12.
of
dry
section and these yields are given
The first yield was
soap.
The
data
calculated per
in this table indicate that the
precipitation and the fusion soap production
solvent gave
soap gave
produced
the greatest
a average
with
no
fuel yields.
yield of
solvent
greatest yield
method with no
The precipitation
41.9% while
the fusion soap
gave an average yield of 40.2%.
Pyrolysis Run TO used soap with 22 grams
had the
100 grams
of toluene solvent
of 66.1% but this was not confirmed
by its duplicate run 9 with a yield of 34.3%.
49
Table 12. Yields of Fuels from Different Soaps With Solvent
Yield (g fuel/
IOOg dry soap)
Soap Prep.
precipitation
precipitation
fusion 52g H 2O
fusion 52g Ha0
fusion 132g HaO
fusion 132g HaO
fusion H g tol.
fusion H g tol.
fusion 22g t o l .
fusion 22g t o l .
fusion
fusion
Yield ( g fuel/
IOOg safflower oil)
44.8
38.9
24.2
26.2
26.9
26.4
26.3
29.2
34.3
66.1
42.6
37.7
45.9
23.4
16.6
30.0
33;8
43.6
39.1
Second, a yield was calculated per
safflower oil
used.
Some of
calculation was inadvertently
value for
each of
100 grams
the duplicate
destroyed
but
the soaps was saved.
of total
data for this
at
least one
These values along
with the first set of yields provided enough information for
a pattern
to develop
and a recommendation to be made.
The
yields on Table 12 show the precipitation soap process had a
45.9% yield
yields of
and the fusion soap method with no solvent gave
43.6%
and
39.1%.
experimental section
more
steps,
prepared
with
and
the
In
reference
under soap-making
ultimately
more
precipitation
simpler fusion process.
to the
procedures it shows
time,
method
Even though
back
for
soap
rather
the fusion
to be
than the
method is
complicated slightly by the use of safflower oil rather than
linoleic acid,
it
is
still
easier
to
perform
than the
50
precipitation method.
From
the
recommended.
data
a
First, the
soaps demonstrated
48
fuels
hours.
regard
Second,
process and the
fusion
greatest yields
and these
fusion
with
no
to
with
can
these different
tendency
to polymerize
with
no
were about
is
a
multi-step one
solvent
the same.
one-step
which is
gave the
Since the
process
and the
harder to perform
and takes more time, the fusion process is preferred.
use of
be
the precipitation soap-making
method
solvent
precipitation a
made
procedure
only a negligible difference would exist
in a diesel engine with
within
soap-making
a solvent
Also,
of any kind is not recommended to achieve
the greatest yield.
Unsaturation Studies
The mechanism by which polymerization proceeds has been
described by Rheimich and Austin as follows
1. An
the
induction period
oxidative
chain
[21]:
occurs, preceding the initiation of
reaction,
during
which
no
visible
physical or chemical properties change.
2.
Oxygen
interacts
hydroperoxides.
with
carbon
A considerable
double
uptake of
bonds
to
form
oxygen coincides
with the beginning of a perceptible polymerization reaction.
3. The polyunsaturated species undergo conjugation of double
bonds arid isomerization of cis to trans forms.
4. The hydroperoxides decompose
resulting in
free radicals
X
51
which in turn contribute to autocatalysis.
5. Production of high-molecular weight cross-linked polymers
and low-molecular
occurs
via
a
weight
free
carbonyl
radical
and
hydroxyl compounds
polymerization
and
scission
reactions.
The number
affect
of
fatty
reactivity.
In
acid
double
addition,
bonds
Miyashita
recently proposed that higher oxidative rates
acids than
should tjien
and
Takagi
of free fatty
those of their methyl esters could be due to the
catalytic effect of the carboxyl groups on the
free radicals
by the
decomposition of hydroperoxides
It is a possibility then that number and/or
carboxyl groups
formation of
relative to
the carbon
[22].
position of the
double bonds would
effect the polymerization reaction.
Previous research in
this
lab
(DNRC
grant agreement
RAE-84-1041, Chemical Processing of Vegetable Oil to Prevent
Polymerization,
properties of
[23])
indicated
was
the
polymerization
the fuel produced by a soap-pyrolysis process
may not be very sensitive to
It
that
found
that
sensitive to the free
the unsaturation
the
the quantity
of unsaturation.
polymerization character was more
fatty acid
content.
studies conducted
The purpose of
here was to investigate
the effect of unsaturation on polymerization with high speed
lubrication oil.
This was done by preparing soaps that were
made from fatty acids with varying
The
fatty
acids
all
had
18
degrees of unsaturation.
carbon atoms with different
52
numbers of carbon double
oleic
acid
(one
bonds:
double
stearic
bond),
acid (saturated),
linoleic acid (two double
bonds), and linolenic acid (three double bonds).
of
unsaturation
can
be determined more specifically using
the technical grade fatty acids than
using a
composed
made
of
a
mixture
of
esters
different degrees of unsaturation
carbon atoms.
from
acids with
and different
numbers of
fatty acids.
the soaps
An
from each
interesting observation
fatty acid had a distinctive
color: off-white for the stearic acid
the oleic
acid soap,
and medium tan for
all
the
soaps
soap, light
the linolenic
except
acid soap.
Pyrolysis of
that of stearic acid produced fuels
room temperature,
Because
the fuel
stearic
acid
fatty
acids.
When
produced from stearic acid,
when
cooled
in
the
pyrolysis
the
condenser
is a
produced from it also
had a high melting temperature compared to those
other
gray for
light tan for the linoleic acid soap,
suitable for further testing.
solid at
safflower oil
Soap was prepared by the fusion method using
each of the four
was that
The effect
fuels from
was run on the soap
resulting
fuel solidified
leading
to
collection vessel and plugged the apparatus.
the
product
Since
a solid
fuel is not at all attractive for use in a diesel engine, it
was decided to eliminate the stearic acid soap
from further
tests.
The
results
are
numbers of the fuels are
presented
all
in
Table
relatively
low
13.
The acid
and indicate
53
little free
fatty acid
content.
The average acid numbers
were 0.2 for the oleic acid soaps, 0.6 for the linoleic acid
soaps,
and
0.5
for
the
starting acids were all
linolenic acid soaps.
monocarboxylic , i .e .,
Since the
they contain
only one carboxylic or acid group, the acid numbers were all
similar: 157
for
oleic,
155
for
linoleic,
and
162 for
Iinolenic.
Table 13. Results of Fuels From Different Acid Soaps
Viscosity
Initial
Final
Fuel
%Yield
Iodine V a l . Acid No.
(cSt)
(cSt)
oleic
acid
soap
69.2
132
0.2
65.4
141
0.2
147
0.6
148
0.6
143
0.6
150
0.4
linoleic 59.4
acid
soap
62.7
linolenic 59.1
acid
soap
57.4
The iodine
those
of
values of
the
starting
unsaturation.
The
79.0
77.6
76.2
75.2
97.7
99.0
95.9
99.6
79.2
78.6
78.7
77.3
109.3
109.9
106.7
107.7
79.1
78.2
78.5
78.8
104.8
103.1
107.0
106.3
the fuels were all similar while
acids
values
increased
of
the
with
starting
increasing
acids
are
understandable (106 for oleic, 153 for linoleic, and 190 for
Iinolenic) since
iodine value is a measure of the number of
carbon double bonds present and oleic
linoleic
has
double bonds.
two
double
bonds,
and
has one
double bond,
linolenic has three
However, the similarity of the
iodine values
54
of the
fuels made
surprising.
be
from soaps
produced from these acids is
Literature has indicated that
formed
during
soap-pyrolysis,
hydroxide is present.
unsaturation
in
This would
the
fuel
promote
vegetable
oxidation
oils.
and
This
particularly if calcium
explain the
increase in
using the oleic acid soap [24].
Further literature indicates that
to
double bonds can
soaps are
used as driers
polymerization
would
lead
to
of
a
unsaturated
decrease
in
unsaturation as
some double bonds are eliminated in forming
larger molecules
[25].
the
findings
of
The carbon 13 NMR
essentially
no
spectra confirmed
carboxylic
acid
groups
present and a significant amount of unsaturation.
The yield data calculated per
indicate
yield
decreased
acid has increased.
as
100
grams
of
dry soap
unsaturation in the original
The average yields of
fuels were 67.3%
from the oleic acid soap, 61.0% from the linoleic acid soap,
and 58.2% from the linolenic acid soap.
Mehta
in
India
cracked
vegetable
In 1939
oils.
Dalai and
They found that
liquid product yields decreased with increasing unsaturation
[26].
This
unsaturation
pattern
has
of
decreasing
therefore
been
yield
with increasing
documented
before
in
vegetable oils.
Polymerization tests were performed with the high speed
lubrication oil.
The values from the table indicate initial
kinematic viscosities for each
soaps were
similar.
of the
fuels from different
These viscosities were 77.0, 78.4 and
55
78.6
centistokes for the fuel from the oleic,
linolenic acid
soaps.
After 48 hours in the polymerization
apparatus the kinematic viscosities
increased.
The increase
for the
from the oleic
acid
soap,
linoleic
acid
soap,
and
linolenic acid
soap.
values
of
the
iodine values of
proportional.
34.0%
viscosities
of
all
the
fuels had
fuels was 27.3% for fuel
for
for
the
fuel
the
fuel
interesting to
taken
starting
For both
order of oleic,
38.3%
It is
the
linoleic, and
after
fuels
from the
from
the
note that the
48 hours and the
are
almost directly
properties the values increased in
linoleic,
and
linolenic
(values
for the
linoleic and linolenic fuels were very close however).
The purpose
of
unsaturation
funded by
Of this task was to investigate the effect
on
the DNRC
polymerization.
[2.7] had
The
previous grant
suggested that
unsaturation on polymerization might be less
the
free
fatty
acid
content
present.
the effect of
important than
The unsaturation
studies conducted here cannot distinguish between
variables
since
all
these two
the starting soap-pyrolysis fuels had
little free fatty content as evidenced
by the
acid numbers
and qualitatively confirmed by the NMR spectra.
Several
conclusions
can
be
drawn
from
First, soap made from stearic acid is undesirable
fuel was a solid at ambient conditions.
desirable for the vegetable oil to
esters than
either linoleic
this study.
since the
Second,
it might be
contain more
oleic acid
or linolenic acid esters.
The
56
fuel yields would probably be greater and there
tendency to
polymerize with
may be less
the lubrication oil.
Finally,
when virtually no free fatty acids are present iodine values
serve
as
a
good
indication
of
polymerization
characteristics.
Soap-Pyrolysis Process Optimization
This
task
optimized
the
soap-pyrolysis
process for
making fuel potentially suitable for use in a diesel engine.
The
results
from
the previous experiments helped identify
important process variables and techniques that were used in
this investigation.
The
purpose of the optimization study
was to put all the prior results together
combination
of
most
desirable
and determine the
techniques
and
levels of
variables for the process.
Several
variables
experiments that
were
observed
from
the
previous
affected the fuel and were optimized h e r e .
First, the Pyrolysis Studies tests addressing the Fatty Acid
Content
of
the
Soap
indicated that the amount of calcium
hydroxide added to the linoleic acid
might
affect
polymerization
during soap production
characteristics.
optimization used safflower oil instead of
produce
the
yields from
soaps
the
Therefore, the
since
vegetable
This
linoleic acid to
it was important to compare fuel
oil
first variable
and
not
just properties.
optimized was
the amount of
excess calcium hydroxide added to the safflower oil
to make
I
57
the soaps.
The
lower constraint
on the amount of calcium
i
hydroxide added to 200 grams of safflower oil was 0 grams of
excess (stoichiometric) or 26.3 grams.
when pyrolysis
was performed
Second, it was shown
with the
addition of calcium
hydroxide reactant that the addition of calcium hydroxide to
the linoleic acid prior
while
limiting
optimized.
to
yield.
pyrolysis
This
decreased thickening
was
the
second
variable
The lower constraint on this variable was to add
no (0 grams) calcium hydroxide prior to pyrolysis.
Several techniques
were also found to affect the fuel.
First, the Pyrolysis Studies
the Soaps
found that
of the soap.
that the
removed
by
washing
the soaps
water and pouring off any on the surface
Second, the
soap should
no solvent.
Glycerol Content of
glycerol in the soap decreased yield.
Therefore, the glycerol was
with demineralized
under the
Soap Production
Studies indicate
be prepared by the fusion method with
This procedure
was
used
to
prepare
all the
soaps for this milestone.
Optimization
The yield of the
polymerization
of
the
fuel was
process
included several steps.
■ ,
selected for optimization.
The
characteristics
were
considered to be less
important because all the soap-pyrolysis fuels were good and
quite
consistent
in
this
respect.
First, the sequential
simplex optimization procedure was used as
[28].
long as possible
The number of variables to be optimized was 2, so the
number of vertices in a simplex was 3.
The simplex vertices
:■
1
I
I
58
corresponded to
the various levels of the variables.
comparing the three
experiments,
discard
experiment.
the
worst
experiment was calculated by
resulted
in
a
new
violated,
i .e .,
decision
The
using
simplex.
when, it was found that the
a
was
After
made to
!
location of the new
simple
rules
and this
This technique was abandoned
constraints were
starting to be
it would require using less than the amount
of calcium hydroxide required
negative
(impossible)
prior to
pyrolysis.
to
amount
produce
of
This made
the
soaps
or a
calcium
hydroxide added
it futile
to maintain the
'
,
original shape of the simplex (See Figure 6 ).
The Complex
Method of Box [29] was next implemented to
locate the optimum.
to the
The Complex
simplex method
in that
Method of
the shape of the pattern of
experimental runs can be changed; this
experimentation
without
This technique used 4
allows for continued
violating any variable constraint.
vertices which
again corresponded to
the levels of variables (See Figure 7).
next experiment was
Box is superior
found
again
by
The location of the
discarding
the least
desirable point and using simple rules to find the new point
in the complex.
From this method the optimum appeared to be
at values
excess grams
of 0
of calcium hydroxide for soap
production and 0 grams of calcium
pyrolysis.
hydroxide added
prior to
The results of this procedure are given in Table
■
14 and it does indicate the
yields
at
or
very
near this
point were the greatest while the kinematic viscosities of
^
i
'
Excess Grams of Calcium Hydroxide for Soap
59
Grams Calcium Hydroxide for Pyrolysis
Figure 6. Variable Levels for Sequential Simplex
Optimization Technique
Excess Grams of Calcium Hydroxide for Soap
6.0
I.S r
X
X
2 -25.5 %
4 -23.3 %
1.0
0,5
3 -36.8 %
1- 40.2 %
5 -50.2 %
XC------------0
6-24.0 %
J __ X
1.0
____
,
7-24.8 %
2.0
Grams Calcium Hydroxide for Pyrolysis
Figure 7. Variable Levels for Complex Method of Box
Optimization Technique
,
3.0
61
all the
experiments were
of each other.
Run 7
looked suspicious.
good.
Run 3 and 7 are duplicates
was performed
The
since the
duplicate Run
Run 3 yield
7 showed that Run 3
was out of line and probably in error.
Table 14. Results for the Complex Method of Box___________
Run
I
2
3
4
5
6
7
XliSl
X 2Igl
% Y ieId
Viscosity (HS)
(cSt)
Initial
Final
0.0
1.45
0.39
1.50
0.03
0.03
0.39
0.00
0.78
2.90
3.00
0.00
1.26
2.90
40.2
25.5
36.8
23.3
50.2
24.0
24.8
79.2
82.3
81.7
76.9
79.8
80.1
77.3
Viscosity (MS)
(cSt)
Initial Final
125.6
131.7
110.5
109.1
115.3
122.7
112.1
138.9
133.6
143.9
119.3
126.3
136.2
136.8
189.8
198.8
212.5
200.6
207.5
202.2
184.8
yield is calculated per 100 grams of safflower oil
Xi=excess of Ca(OH)2 used to make the soap
X2=Ca(OH)2 added prior to pyrolysis
H S = M g h speed lubrication oil
MS=medium speed lubrication oil
Once the optimum
experimental
design
appeared
was
surface of the yield as
[30].
The
reason
for
used
a
to
be
to
fit
function
obtaining
found,
of
this
a factorial
an equation to the
the
two variables
equation
was to
represent all the possible yields over the ranges of the two
variables
rather
experimental runs.
used to
then
be
find the
verified
conditions.
than
just
the yields for the particular
An optimization technique
maximum yield
with
an
could then be
on the surface; this would
experimental
run
at
optimum
This factorial required nine experimental runs
62
of which three of the previous
experiments were
runs were
made (See Figure 8 ).
experiments was determined randomly.
used and
six new
The order of the new
Table 15 shows the
experiments for the 32 factorial and the corresponding
yields and kinematic viscosities at each level.
Table 15. Results from the 32 Factorial
Xi
0.015*$
0.000
0.000
0.195
0.195
0.195
0.368
0.368
0.390&
X2
% Y ieId
Viscosity (HS)
(cSt)
Initial
Final
0.00
1.45
2.90
0.00
1.45
2.90
0.00
1.45
2.90
45.2*@
24.0(3
23.5
37.7
26.5
26.8
29.8
14.7
24.8(3
7 9.5 *@
80.1(3
78.1
77.8
79.0
79.2
82.0
73.2
77.3(3
120.5*@
122.7(3
114.2
114.2
108.6
115.8
116.2
119.3
112.1(3
Viscosity (MS)
(cSt)
Initial
Final
132.6*@
138.20
126.5
132.7
133.1
130.0
129.0
108.0
136.80
198.7*0
202.20
190.2
188.6
188.8
185.0
184.3
199:4
184.80
* averages of duplicate runs
@ taken from previous work in this study
$ although this number should have been 0 .0 , it was close
enough and could be used in the factorial experimental
design
& although this number should have been 0.368,
it was close
enough and could be used in the factorial experimental
design
The
mathematical
equation
that
describes
the yield
surface w a s :
y = B q X o + Bi Xi + B2X2 + BiiXi2 + B2 2 x22 + Bi 2X1X2
where:
y = yield per 100 grams of safflower oil
xi as excess grams of Ca(0H )2 in soap
x2 = grams of Ca(OH)2 added prior to pyrolysis
63
29.8 %
Excess Grams of Calcium Hydroxide for Soap
0.368-X
37.7 %
0.195-X
45.3 %
O x
14.7 %
X
26.5 %
X
24.0 %
24.8 %
X
26.8 %
X
23.5 %
- f —
1.45
2.40
Grams Calcium Hydroxide for Pyrolysis
Figure 8. Variable Levels for the Factorial Experimental
Design
64
The
procedure
Davies
[31].
constants is
used
to
A more
complete
given in
the variables
were
calculations.
find the constants is described by
Appendix A.
coded
The
discussion
and
levels
finding the
As outlined by Davies,
later
of
for
decoded
to simplify
the variables are given in
Table 16.
Table 16. Variable Levels Used for Optimization_____________
Variable
Xl
Level (grams)
0.000
0.195
0.368
0.000
1.450
2.900
X2
The equation was calculated to be:
y = 44.6 - 7.8xi - 2 0 .4x2 - 98.2xi 2 + 4.6x2 2 + 15.7x i X2
AlI the constants were determined to be significant.
There are many optimization techniques that can be used
with
an
equation
such
representing possible
Box is
reliable and
this technique.
as
this
one
fuel yields.
a computer
The program is
The
for
the
surface
Complex Method of
program was written to use
given in
Appendix B.
The
program includes several subroutines unique to this problem.
These are a subroutine for the objective to be maximized and
a
subroutine
for
maximized was the
constraints
were:
the
constraints.
equation
0<xi< 2
for
and
purpose of the constraints was to
the
The objective to be
yield
0<X2 <5,
surface.
all in grams.
limit the
The
The
search for the
65
optimum yield ranges of the variable levels.
xi and x2 again correspond to
hydroxide used
to make
the soap
hydroxide added prior to
program
uses
function.
and
is
a
subroutine
already
that
the
maximum
The last
maximizes
uses the
programmed
of excess calcium
and the amount of calcium
pyrolysis.
This subroutine
determined
the amount
The values for
part of the
the objective
Complex Method
in the computer.
yield
occurred
of Box
This program
at
values
0.315735577E-6 for X 1 and 0.151638811E-6 for X 2 .
of
Therefore,
the maximum yield can be said to occur at the point Xi=O and
X2= O .
This was used as a check on the optimum value of the
objective function.
One
last
run
stoichiometric
was
amount
done
of
at
the
calcium
production (0 grams excess) and no
(0
grams)
added
to
addition, duplicate
with an
acid number
the
soap
hydroxide
yield.
44.6%
prior
polymerization
This yield of 39.8%
experimental
by
for
to
the
variability
a
soap
tests
pyrolysis.
were
this was
can be
. The
In
done along
and iodine value for this fuel.
indeed indicate
predicted
using
calcium hydroxide
17 gives the results for this final run.
table does
optimum
Table
data on this
the point for maximum
compared to
equation.
This
present.
The
a value of
shows there is
variability
further evidenced by runs I arid 5 in Table 14.
is
Performed at
almost identical conditions the yield differed from 40.2% to
50.2%.
66
Table 17. Final Run Results
%Y ieId
Iodine V a l .
Acid No.
39.8
158
1.7
Viscosities (HS)
(cSt)
Initial Final
76.9
77.3
Viscosities (MS)
(cSt)
Initial Final
113.3
114.7
The
120.0
120.6
thickening
characteristics
previous results in the
viscosities
for
210.7
207.2
the
fuels
in
the
point given
run
increased
and medium speed
while fuel
the polymerization
earlier in
there is
used with the
increased an average of 73.7%
apparatus.
this thesis
kinematic viscosity in 64 hours.
1.7 shows
to
Fuel used with the high speed lubrication
lubrication oil
in 48 hours in
similar
The kinematic
final
high speed
oil increased an average of 47.6%
medium speed
all
Optimization Study.
different percentages with the
lubrication oil.
are
little free
The failure
was a 375% increase in
Also, the
acid number of
fatty acid
content and an
iodine value of 158 displays significant unsaturation.
Several
studies.
things
First,
were
yield
learned
was
differences between the fuels
properties
values
for
were
the
not
as
variables
from
the
optimization
effective for evaluating the
since
changes
significant.
tested
in thickening
Second, the optimum
occurred
with
(I)
a
67
stoichiometric
amount
of
calcium
hydroxide
added to the
safflower oil for soap production and with (2 )
no amount of
calcium hydroxide added to the soap prior to pyrolysis since
the yield was
increased
a
less
apparatus when
compared to
reference
maximum
after
the
48
high
the medium
to
the
Calcium Hydroxide
he r e .
The
hours
speed
in
kinematic viscosity
the
polymerization
lubrication
oil
speed lubrication oil.
results
in
the
Reactant results
was used
However, in
Pyrolysis With/Without
section, the high speed
lubrication oil is more resistant to polymerization than the
medium
speed
lubrication
oil.
The control polymerization
tests that were run with lubrication
soap-pyrolysis
fuel
polymerization
showed
apparatus
the
oil without
after
high
48
any added
hours
speed
in
the
lubrication oil
experienced a 3.7% increase in kinematic viscosity while the
medium
speed
had
differences in
a
32.0%
increase.
Therefore,
the
viscosity increases were probably due to the
characteristics of the two lubricating oils
rather than due
to the addition of soap-pyrolysis fuels.
A final
comparison is
the soap-pyrolysis fuel.
made between
safflower oil and
In previous work done
in this lab
for the soap-pyrolysis technique, straight safflower oil was
tested [32].
The
previous
polymerization tests
lubricating
oil
concentration
for the
was Phillips 66 HD II SAE 30W (similar
to the present high speed lubrication oil) and
percent
used
of
safflower
oil
was
a 5.0 weight
used.
The
68
safflower
oil
viscosity in
experienced
48 hours
a
•170.
fuels
increase
In addition,
the iodine value
safflower oil was 149 and a typical acid number was
These
in
results
this
in
study
comparison
subjected to
to
the soap—pyrolysis
indicate safflower oil has a greater
tendency to polymerize than
the
soap-pyrolysis
much
soap-pyrolysis
greater
safflower
oil
unsaturation.
fuels when
conditions of a simulation of a diesel engine.
The most apparent chemical difference between
and
in kinematic
in the polymerization apparatus from
about 95 to 200 centistokes.
for the
110%
acid
and
fuels
is
number.
safflower oil
that the safflower oil has a
Iodine
values
of
both the
soap-pyrolysis fuels showed significant
69
CONCLUSIONS
I.
An
advantage
opposed to old or
fresh soap
might
be gained by using fresh soap as
dried soap
would have
since the
a lower
fuel produced from
initial kinematic viscosity
than that of the fuel from the old soap.
2•
Glycerol is not desirable in the soap prior to pyrolysis
®ihce it
decreased yield
of the fuel in the soap—pyrolysis
process.
3.
The greater the
produce the
amount
soap in
of
calcium
hydroxide
used to
the fusion process the less thickening
that will occur in the fuel.
4.
Calcium soap
calcium
soap
is superior
gives
a
to magnesium
greater
soap since the
yield of fuel in the soap-
pyrolysis process.
5.
The lighter
fraction of
pyrolysis process
The lighter
process
the
was
had a
in the soap-
lower initial kinematic viscosity.
fraction
subject
product obtained
of
to
fuel
in
greater
the soap-pyrolysis
thickening
in
the
polymerization apparatus.
6.
The total product obtained by the soap—pyrolysis process
is thicker
starting out
and experiences less thickening in
the polymerization apparatus.
7.
Aging of soap
pyrolysis
thickening properties.
fuels
does
not
affect their
70
8.
The addition
of calcium hydroxide to the soap prior to
pyrolysis decreases thickening while limiting yield.
9.
The fusion soap-making process
with no
solvent and the
precipitation soap production method give the greatest yield
of fuel of the methods
investigated.
The
fusion process
with no solvent is recommended since it is easier to perform
than the precipitation method.
10.
more
It might be desirable for the safflower oil
oleic
acid
esters
than
esters since the yields would
linoleic
probably
to contain
or linolenic acid
be
higher
and the
thickening not as severe.
11.
Optimization
of the soap—pyrolysis process indicates
that calcium hydroxide used in a stoichiometric quantity for
soap production gives the best results.
12.
No great
advantage will
be gained
by adding calcium
hydroxide to the soap prior to pyrolysis.
13.
Soap-pyrolysis fuels are better than straight safflower
oil
in
the
regard
of
tendency
polymerization apparatus at 1500F
engine.
to
polymerize
which simulated
in
the
a diesel
RECOMMENDATIONS FOR FUTURE RESEARCH
1.
Investigate
pyrolysis
with the same dimensions as
in a stainless, steel apparatus
the
glass
apparatus
used for
pyrolysis to determine any effect in yield.
2.
Study the
relationship between free fatty acid content
and unsaturation more thoroughly.
3.
Investigate the
percentage of
soap-pyrolysis fuel that
could be added to diesel fuel.
A.
Investigate the evaporation of vegetable oil fuel in the
polymerization tests.
APPENDICES
APPENDIX A
OPTIMIZATION RESULTS
74
Table 18 gives the values of the variables used to find
the constants for the mathematical model of fuel
the
soap-pyrolysis
process.
yield from
The variables were coded to
simplify the calculations.
The variables
grams of calcium hydroxide.
This following procedure is
given in Davies
are expressed in
[33].
Model of the Soap-Pyrolysis Process
Coded
-I
0
+1
Uncoded xi
0
0.195
0.368
Table 19
Uncoded x?
0
1.45
2.90
gives the values of the independent variables
needed in the equation for yield.
The equation is shown in
the Sample Calculations section of this Appendix A.
Table 19. Values for Surface Yield Equation Calculation
Coded Variables
%Y ieId
Xl
X.2
-I
-I
-I
0
0
0
+1
+1
+1
-I
0
+1
-I
0
+1
-I
0
+1
xi 2-2/3
45.2
1/3
24.0
1/3
23.5
1/ 3
37.7
-2/3
26.5
-2/3
26.8
-2/3
29.8
1/3
14.74
1/3
24.8
1/3
sum y2=7735.0
X2.2-2/3
1/3
-2/3
1/3
1/3
-2/3
1/3
1/3
-2/3
1/3
Xl X2
XO
+1
0
-I
•4-1
0
-i
+i
0
-I
I
I
I
I
I
I
I
I
I
Table 20 gives the estimates of the constants.
75
Table 20. Constants Estimate
constant
estimated
n0
Bi
B2
Bh
B2 2
Bi 2
(2 )
sum x2
9
6
6
2
2
4
(3)
sum yx
253.1
-23.5
-37.5
-6.7
19.1
16.7
(4)
estimate 3/2
28. I
-3.9
-6.3
-3.3
9.6
4.2
(5)
32/2
7116.6
91.9
234.1
22.2
182.4
69.9
sum=7717.I
Sample Calculations
The equation for yield fits form of:
y ■= B q X o + BiXi + B2X2 + Bi 1xi 2 + B 22X22 + B 12X1X2
The above equation when expressed in different form is:
y = n0X0 + B iXi + B 2X2 + B n (xi2-.2/3) + B 22 (x22-2/3) +
Bi 2 Xix 2
The relationship to return to the first equation is:
■bo
2/3bii — 2/3b22
b0 = 28.1 + 2/3(3.3) - 2/3(9.6 ) = 24.0
The estimated constants with the ranges a r e :
bo = . 24.0 +/- 1.8
-3.9 +/- 1.0
bi
-6.3 +/- 1.0
b2
b n S -3.3 +/- 1.7
b 2 2 S1 9.6 +/- 1.7
b n S 4.2 +/- 1.2
The numbers needed to calculate the range are:
7735.0 - 7717.1 = 17.9 = deviations from regression
17.9/3 = 6.0
V (bo) = V(Y) + (2/3)2V (hi 1 ) + (2/3)2V(b2 2 ) = 0.05555var2
The ranges for the constants a r e :
b0 : sqrt((0.5555)(6.0)) = 1.8
bi,b2 : sqrt(6.0/6) = 1.0
b n , b 22 : sqrt(6.0/2) = 1.7
bi2 : sqrt(6.0/4) = 1 . 2
76
The next step was to decode the equation and this was done
by finding the following relationships between the coded and
uncoded variables:
X2 (coded) = (X2 (uncoded)- I .45)/ I .45
xi(coded) = (xi(uncbded)-O.184)/0.184
The final step was to substitute the right side of the above
equations into the left and simplify:
y = 44.6 - 7.8xi - 20.4X2 - 98.2xi2 + 4 .6x2% + 15.7x i x 2
APPENDIX B
COMPUTER PROGRAM USED FOR OPTIMIZATION
78
Figure
9. Computer
Program Used For O p t i m ization
MILESTONE 5 OPTIMIZATION
IMPLICIT REAL*8(A-H9O-Z)
DIMENSION X(2,4),XL(2),XH(2),XC(2),XX(2),FUNC(4),XP(2,4),
1XXP(2)
1
READ*,M,N,KK,KPRINT,EPSI
READ*,(X(I,1),I=1,N)
READ*,(XL(I),I=1,N)
READ*,(XH(I),1=1,N)
CALL CM80X(M,N,KK,KPRINT,X,XL,XH,XC,XX,XP,XXP,FUNC,EPSI)
GOTO I
END
SUBROUTINE 0BJ(X,N,XP,0B,K0B)
IMPLICIT REAL*8(A-HsO-Z)
DIMENSION X(N)sXP(N)
KOB=KOB+!
Yl=23.97-3.91/0.184*(X(l)-0.184)-6.25/1.45*(X(2)-1.45)
Y2=-3.33*((X(1)-0.184)/0.184)**2+9.55*((X(2)-1.45)/1.45)**2
Y3=4.18/0.184/1.45*(X(1)-0.184)*(X(2)-1.45)
Y=Y1+Y2+Y3
OB=Y
DO 2 I=IsN
XP(I)=X(I)
2
CONTINUE
RETURN
END
SUBROUTINE CONSTR(XsNsIVI)
IMPLICIT REAL*8(A-HsO-Z)
DIMENSION X(N)
IVI=O
Gl=X(I)
IF(Gl.LT.O.O) THEN
IVI=I
RETURN
ENDIF
IF(G1.GT.2.0) THEN
IVI=I
RETURN
ENDIF
G2=X(2)
IF(G2.LT.0.0) THEN
IVI=I
RETURN
ENDIF
IF(G2.GT.5.0) IVI=I
.
.
79
Figure 9— Continued
RETURN
END
SUBROUTINE CMBOX(M,N,KK,KPRINT,X,XL,XH,XC,XX,XP,XXP
1FUNC,EPSI)
IMPLICIT REALa B(A-H5O-Z)
DIMENSION X(N5KK),XL(N),XH(N)5XC(N),XX(N),FUNC(KK)5
IXP(N5KK)5XXP(N)
K=I
KI=I
NUM =7621
KCOUNT=O
KRCOUNT=O
INIT=O
IF(EPSI.LT.l.D-6) EPSI=I.D-6C
KPR =100
KOB = O
GO TO 10
4
IF(INIT.EQ.l) GO TO 42
IF(K.EQ.KK) GO TO 30
DO I T=I5N
XC(I)=O.
DO 5 J=I5K
5
XC(I)=XC(I)+X(I,J)
1
XC(I)=XC(I)ZK
K=Kfl
KI=KIfl •
DO 2 I=I5N
NUM=241*NUM f 5
NUM=MOD (NUM565536)
ANUM=NUM/65536.
2
X(I,K)=XL(I)fANUM*(XH(I)-XL(I))
10
DO 6 I=I5N
6
XX(I)=X(I5K)
IF(M.EQ.O) GO TO 4
CALL CONSTR(XX5N5IVI)
IF((KI.EQ.l).AND.(IVI.EQ.I)) GO TO 80
GO TO 81
80
WRITE(2,400)
WRITE(6,400)
RETURN
81
IF(IVI.EQ.l) GO TO 20
GO TO 4
20
DO 3 I=I5 N
3
X(I5K)= (XCd) f X(I5K) )/2.
GO TO 10
30
DO 31 K=I5KK
80
.Figure 9 — Continued
DO 32 1=1,
32
XX(I)=X(I,K)
CALL 0BJ(XX,N,XXP,0B,K0B)
DO 29 1=1,N
29
XP(IfK)=XXP(I)
31
FUNC(K)=OB
36
IF((KPRINT.EQ.O).AND.(KCOUNT.NE.KPR)) GO TO 37
35
WRITE(2,200)(FUNC(K),K=1,KK)
WRITE(6,200)(FUNC(K),K=1,KK)
DO 51 1=1,N
WRITE(2,200)(XP(I,K),K=1,KK)
51
WRITE(6,200)(XP(I,K),K=1,KK)
WRITE(2,100)
WRITE(6,100)
IF(KC0UNT.EQ.6) GO TO 70
IF(KPRINT.EQ.l) GO TO 37
IF((KPRINT.EQ.O).AND.(KCOUNT.NE.KPR)) RETURN
KPR=KPR + 100
37
small =f u n c (1)
IR=I
DO 33 K=2,KK
IF(SMALL-FUNC(K))33,33,34
34
SMALL = FUNC(K)
IR=K
33
CONTINUE
BIG=FUNC(I)
DO 38 K=2,KK
IFfBIG-FUNC(K))39,38,38
39
BIG=FUNC(K)
38
CONTINUE
KCOUNT=KCOUNT+!
EPS=DABS((BIG-SMALL)ZBIG)
IF(EPS.LT.EPSI) GO TO 60
IF(KCOUNT.GE.999) GO TO 60
DO 40 1=1,N
XC(I)=O.
DO 41 K=IfKK
41
XC(I)=XC(I)+X(I,K)
XC(I)=(XC(I)-X(I,IR))/(KK-1.)
40
X(I,IR)=2.3*XC(I)-1.3*X(I,IR)
K=IR
INIT=I
GO TO 10
42
CALL OBJ(XX,N,XXP,OBfKOB)
KRCOUNT=KRCOUNT+!
DO 45 I=IfN
45
XP(IfK)=XXP(I)
81
Figure 9— Continued
FUN=OB
IF (KRC0UNT.GT.49) THEN
WRITE(2,700)KRCOUNT
WRITE(6,200)KRCOUNT
GO TO 60
END IF
IF (FUN.GT.FUNC(K)) THEN
KRCOUNT=O
GO TO 50
END IF
DO 43 1=1,N
43
X(I,K)=(X(I,k)+XC(I))/2.
DO 44 1=1,N
44
XX(I)=Xd,K)
GO TO 42
50
FUNC(K)=FUN
GO TO 36
60
WRITE(2,300)EPS,KC0UNT,KK,KOB
WRITE(6,300)EPS,KCOUNT,KK,KOB
IF(KPRINT.EQ.O) GO TO 35
RETURN
70
WRITE(2,500)
WRITE(6,500)
READ(5,600) KPRINT
GO TO 37
100
FORMAT ()
200
FORMAT(4G18.9)
300
FORMATd FUNCTIONAL VALUES WITHIN FRACTIONAL
I DIFFERENCE ='
1,614.5/,15,' CYCLES WITH ',13,' VERTICES HAVE BEEN'
2' COMPUTED'/' OBJECTIVE HAS BEEN EVALUATED',15,
3' TIMES')
400
FORMAT(' YOUR STARTING POINT DOES NOT SATISFY THE'
I' CONSTRAINTS')
500
FORMAT(' ENTER A VALUE FOR KPRINT')
600
FORMAT(Il)
700 FORMAT(' PROGRESS NOT MADE AFTER ',12,'CONTRACTIONS.'/
I'START OVER WITH CURRENT BEST X(I)S AND/OR NEW EPSI')
END
REFERENCES CITED
83
1. Rosenbaun, Walter A . , Environmental Politics and Policy,
Congressional Quarterly Inc., Washington D .C . , 1.985 ,
p. 220.
2.
D u k e , J .A . , and M.O. Bag b y , "Comparison of Oilseed
Yields: A Preliminary Review", Proceedings of the
International Conference on Plant and Vegetable Oils as
Fuels, Far g o , North Dakota, 1982.
3.
Quick,
G.R.,
"An In-Depth Look at Farm Fuel
Alternatives", Power Farming Magazine, 59(2), 1980.
4.
Collins,
G.S.,
R.C.
Griffin,
and R.D. Lacewel I ,
"National Economic Implications of Substituting Plant
Oils for Diesel Fuel", Proceedings of the International
Conference on Plant and Vegetable Oils as Fuels, Fargo,
North Dakota, 1982.
5.
Simon, Julian L., "The Scarcity of Raw Materials", The
Atlantic Monthly, June 1981.
6 . Van der Walt, A.N., and F.J.C. Hu g o , "Diesel Engine
Tests with Sunflower Oil as an Alternative Fuel",
The
Third International Conference of Energy Use Management,
Berlin, 1981, Permagon Press, 1981.
7.
Hiebert, Dwight Randall, "Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels",
Montana State University, Bozeman, Montana,
May 1985, p. 2.
8 . Windholtz, M., et. a l ., The Merck Index, IOth ed . , Merck
and Co., Inc., Rahway, N.J., 1197, 1293, 1234, 1903.
9.
Miyashita, Kaz u o , and Toru Takagi, "Study on the
Oxidative Rate and the Prooxidant Activity of Free Fatty
Acids", J . Am. Che m . Soc., Vol 63, no. 10., October 1986.
10.
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels",
Montana State University, Bozeman, Montana,
May 1985.
11.
CRC Handbook of Lubrication, Vol I, The American Rubber
Company, Cleveland, Ohio, 1983.
12.
CRC Handbook of Lubrication, Vol I, The American Rubber
Company, Cleveland, Ohio, 1983.
13.
14.
Sato, M.J., Che m . Ind., Jap a n , 25, 1922.
Hiebert,
Dwight
Randall,
"Decarboxylation
and
84
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels",
Montana State University, Bozeman, Montana,
May 1985.
15.
Markley, K.S ., Fatty A c i d s , 2nd, 1961.
16 .
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels",
Montana State University, Bozeman, Montana,
May 1985.
17.
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels", Montana State University, Bozeman, Montana,
May 1985.
18•
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels", Montana State University, Bozeman, Montana,
May 1985.
19.
Sato, M.J., Che m . Ind., Japan,
25, 1922.
20.
Markley, K.S., Fatty Acids, 2nd e d., 1961.
.
21.
Reinbeck,
A.E.,
and R.O.
Austin,
"Treatise
on
Coatings", Myers, R.R., and J.S. Long, EDs., V o l . I, Part 2,
Ch.4, Marcel Dekker, New York, 1968.
22.
Miyashita, Kazou, and Toru Takagi, "Study on the
Oxidative Rate and the Prooxidant Activity of Free Fatty
Acids", J . Am. Chem. Soc., V o l . 63, no. 10, October 1986.
23.
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels", Montana State University, Bozeman, Montana,
May 1985.
24.
Scarrah, W . P . , Milestone
Studies,
Montana
Department
Conservation.
25.
Report #4,
Unsaturation
of
Natural
Resource
Markley, K.S., Fatty A c ids, 2nd ed., 1961.
26.
Dalai, N.M., and T .N . Mehta, "Cracking of Vegetable Oil
Fuels" J . Am. Soc., Ind. News E d . , 2, 1939.
27.
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels",
Montana State University, Bozeman, Montana,
May 1985.
85
28.
Beveridge, Gordon
Optimization Theory and
367-383.
S.G.,
and Robert S . Schechter,
Practice, McGraw-Hill, 1970, pp.
29.
Beveridge,
Gordon
Optimization Theory and
353-356.
S.G.,
and Robert S . Schechter,
Practice, McGraw-Hi11, 1970, p p .
30.
Davies , Owen
Experiments, Hafner
518-522.
L ., Design
Analysis
of Industrial
Publishing C o . , New York, 1967, pp.
31.
Davies , Owen
Experiments, Hafner
518-522.
L . , Design
Analysis
of Industrial
Publishing C o ., New York, 1967, pp.
32.
Hiebert,
Dwight
Randall,
"Decarboxylation and
Hydrogenation of Safflower and Rapeseed Oils to Produce
Diesel Fuels",
Montana State University, Bozeman, Montana,
May 1985.
33.
Davies,
Owen L., Design Analysis of
Experiments, Hafner Publishing Co.,
New York,
518-522.
Industrial
1967,"pp.
MONTANA STATE UNIVERSITY LIBRARIES
3
I
762 101 26973 4