EPA-AA-TSS-85-
Technical Report
Cold Starting Spark Ignition Engines with Methanol:
An Analysisof Current Options,
and Their Impact on Air Quality
By
William B. Clemmens
August 1985
NOTICE
Technical Reports do not necessarily represent final EPA
decisions or positions. They are intended to present
technical analysis of issues using data which are
currently available. The purpose in the release of such
reports is to facilitate the exchange of technical
information and to inform the public of technical
developments which may form the basis for a final EPA
decision, position or•regulatory action.
Technical Support Staff
Emission Control Technology Division
Office of Mobile Source Air Pollution Control " •
Office of Air, Noise and Radiation
U. S. Environmental Protection Agency
-------
-2-
I. Introduction
II. Executive Summary
Technical Analysis *******
* * * *
III. Unassisted Vaporization
A. Introduction
B. Required Properties
C. Summary
IV. Physical Vaporization (Heat Addition)
A. Introduction
B. Required Properties
C. Heat Addition Options
D. Summary
V. Chemical Vaporization (Fuel Additives)
A. Introduction
B. Required Properties
C. Summary
VI. Starting Fuels
A. Auxiliary Fuels
B. Fuels made from Methanol
VII. Effect of Cold Starting Approach on Air Quality and
the Certification Process
-------
-3-
I. Introduction
-------
-4-
II. Executive Summary
-------
-5-
III. Unassisted Vaporization
A. Introduction
In order to start a Spark Ignited (SI) engine, it is
generally accepted that a sufficient quantity of fuel must
vaporize to provide a combustible mixture of fuel vapor and
air in the vicinity of the spark plug during the ignition
event (1). Under cold starting conditions, the latent heat
of vaporization and the curve of vapor pressure versus
saturation temperature play important roles in determining
the quantity of vapor available at the spark plug. Gasoline
is a multi-component mixture, and certain components in the
fuel can provide sufficient vapor phase fuel under cold
starting conditions, provided enough gasoline is added to the
air charge. Neat methanol, of course, has no such volatility
additives, and has a fixed volatility curve. Although
volatility components can be added to methanol, in this
section we want to evaluate the requirements for cold
starting neat methanol without additives or other starting
devices.
In this situation, essentially all the heat used to
vaporize the neat methanol must come from the latent heat of
the air and liquid fuel charge, from heat transfer from
warmer engine surfaces (if any), and from the work of the
compression stroke; most will come from the work of
compression. The process can be envisioned as consisting of
five steps: 1) Liquid fuel and air at the prevailing ambient
temperature are mixed. 2) The fuel begins to vaporize and
the temperature of the air and remaining liquid fuel is
depressed until an equilibrium is reached. The amount that
can be vaporized depends on the initial ambient temperature.
3) The depressed mixture temperature leads to heat transfer
from engine surfaces which have remained at the ambient
temperature. 4) The mixture is then compressed and heated,
with some heat loss to the cylinder walls. 5) The heat of
compression is used to vaporize more of the methanol until
equilibrium is reached. In actuality, the second, third,
fourth, and fifth steps overlap in time.
B. Required Properties
Several investigators (2)(3) have used various
approaches to model the vaporization or starting capabilities
Of neat methanol. Some are more sophisticated than others.
Our approach while somewhat simpler than others provides a
different perspective to the issue of cold starting with
methanol.
Our approach was based on the premise that a certain
quantity of vapor is required for the starting of SI engines,
-------
-6-
and that the macroscopic equilibrium conditions to maintain
this vapor must at least be met. Otherwise, some the vapor
would condense and reduce the vapor . fuel-air ratio to a
non-combustible mixture. The temperature of the mixture
after adiabatic ideal gas compression and evaporation serves
as an indication of equilibrium reserve, and as such must be
at or above the saturation or dew temperature for the vapor
phase fuel concentration. Any temperature above the dew
temperature would be a temperature reserve (TR). This
reserve would be utilized for such things as heat transfer to
the cylinder walls, non-ideal process losses, rate effects
(i.e., insufficient time to reach equilibrium), etc. Because
these losses are always present and must be compensated, the
temperature reserve would have to be a positive number in
order to maintain the fuel in a vapor state under equilibrium
conditions. In our evaluation, we compared the calculated
temperature reserve (TR) at known start capability conditions
to the reserve at unknown cold starting conditions. As long
as the reserve at the lower temperature was equal to or
greater than the reserve at the higher temperature (with the
known start capability), we assumed the potential for a
successful cold start existed.
This approach however assumes all of the evaporation
necessary to achieve a combustible mixture occurs during the
compression stroke. Before accepting the reasonableness of
this approach (i.e., no fuel evaporation in the intake
manifold or in the cylinder during the intake stroke when
cranking), we studied the relationship between the vapor
equivalence ratio and the ambient temperature which would be
required to support a given vapor ratio after evaporation.
Because methanol is a single component mixture, a well
defined relationship exists between the partial pressure of
methanol and the saturation or dew temperature for methanol.
The partial pressure of a given fuel mixture is governed by
the concentration of the vaporized .methanol (i.e. the vapor
f/a ratio), and the local pressure. The following equation
from Appendix I in reference (5) was used to determine the
partial pressure of methanol for a range of pre-selected
vapor equivalence ratios.
(III-l) PV = (MAP) (1/[1 + (7.155/VPHI)])
where:
PV = partial p-ressure
MAP = manifold air pressure
VPHI = vapor equivalence ratio
The saturation temperature is, of course, the
temperature that separates the liquid phase from the gas
phase, and in our case it is the temperature which must be
-------
-7-
maintained in order to prevent our pre-selected vapor
equivalence ratio from condensing. Reference (6) describes
the relationship between the partial pressure and dew
temperature as:
(III-2) TDEW = (-1961.8678)/[log,0(PV) - 8.639821)]
where:
TDEW =(°K) = saturation temperature
PV = (mm Hg) = partial pressure from Eq. III-l
Equation III-2, does not consider the fact that in the
case of an intake manifold, in order to achieve a given vapor
equivalence ratio, an equal amount of liquid methanol must be
evaporated. The process of evaporating the liquid fuel would
depress the surrounding ambient temperature. Therefore, in
order to maintain our computed TDEW value for our
pre-selected vapor PHI in equation III-l, our initial ambient
temperature before evaporation must be above the TDEW value
at least by the amount of temperature depression caused by
the evaporation of the fuel. The following equation from
reference (4) was selected to determine the temperature drop
from the evaporating fuel.
(III-3) TDROP = [(x)(F)(HLG) + (Q)]/[(1-F+xF)(CP)](5/9)
where:
TDROP = (°C) = temperature drop
x = portion evaporated, in our case = 1.0
F = fuel-air ratio
HLG = heat of vaporization = 474 BTU/lb (ref. 4)
Cp = is for a mixture of fuel-air vapor,
reference (4) lists Cp as 0.245 for a
stoichiometric mixture of methanol vapor
and air, and a Cp of 0.240 for air
alone.
Q = heat addition, in our case = 0.0
- Heating or cooling of excess liquid fuel
was neglected
- Pressure changes (if any) from cooling
were ignored
By substituting the same pre-selected equivalence ratio
that was used in equation III-l into equation III-3, we can
identify the temperature drop that would be associated with
vaporizing a given amount of fuel corresponding to the dew
temperature calculated in equation III-2. Combining the
results of III-2 and III-3, we have the initial ambient
temperature before vaporization (which is assumed to be the
same as the initial intake manifold air temperature) that
would support the pre-selected vapor equivalence ratio.
-------
-8-
(III-4) TAMB = TDEW + TDROP
Where:
TAMB = ambient temperature.
TDEW = saturation temperature, equation III-2.
TDROP = temperature drop, equation III-3.
Results of equation III-4 for a range of vapor
equivalence ratios and manifold air pressures (MAP), are
plotted in Figure III-l and listed in Table III-l through
III-4. The range of manifold pressures used to derive these
tables covered the range (suitable for modeling purposes) of
observed manifold pressure values for a typical four cylinder
engine under cranking conditions at 99kPa(wet) barometric
pressure and a range of cranking speeds from 200 RPM to 600
RPM. From Figure III-l and the Tables, the influence of
varying the manifold air pressure slightly from the normal
cranking condition (85 kPa) appears to have little effect on
the amount of fuel that can be vaporized in the intake
manifold at temperatures below 0°C. Large changes in
manifold pressure may have an observable affect on the amount
of fuel that can be vaporized, but we will investigate that
prospect later.
For the time being, these equilibrium data suggest that
if we are to consider starting a neat methanol engine at
ambient temperatures of -18°C (0° F) or below (TAMB in Tables
III-l to -4), it seems safe to say that there will be very
little evaporation occurring within the inlet manifold during
cranking, at most maybe a 0.05 to 0.06 vapor equivalence
ratio, and at -35°C, maybe a 0.03 vapor equivalence ratio.
(Lack of time may prevent even these points from being
reached). For a carbureted engine, the lack of fuel
vaporization in the inlet manifold could cause serious
distribution problems, but more important for our study, the
lack of evaporation in the inlet manifold means that the
majority of the evaporation to produce an ignitable vapor at
cold conditions must occur during the compression stroke.
This evaluation, therefore, suggests that our initial
approach (i.e., no fuel evaporation in the intake manifold or
cylinder during the intake stroke when cranking) is a
reasonable assumption to begin our analysis.
We have so far neglected the fact that the temperature
drop caused by evaporation will create a temperature gradient
and encourage heat transfer from the manifold or cylinder
walls, leading in turn to a higher vapor equivalence ratio.
Evaporation of enough fuel to cause a vapor equivalence ratio
-------
VflPOR PHI VS MRNFOLO BIR TEMP
1.0
0.9
0.8
uJ 0.6
o
z
l*J _ _
_j 0. 5
oc
30.4
UJ
g 0.3
a.
cc
» 0.2
0.1
0.0
-60
» VflP PHI H/EVflP
RT 100.0 KPR
= VflP PHI H/EVflp
RT 85.0 KPR
N-29
« VflP PHI H/EVflP
RT 75.0 KPR
N-29
-20 20 60 100 1UO 180
-40 0 40 80 120 160 200
RMBIENT TEMPERRTURE (OEG C)
- /
-------
TABLE III-I
VAPOR EQUIVALENCE RATIO
TABLE III-2
VAPOR EQUIVALENCE RATIO
MAP
(KPA)
100.
too
100
too
100.
100.
100.
too.
100.
100.
100.
100.
100.
too.
100.
100.
100.
100.
100.
100.
100.
100.
too.
100.
100.
too.
too.
too.
100.
.000
.000
.000
.000
.000
.000
.000
.000
.000
000
.000
.000
.000
000
.000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
VAPOR
PHI
0.
0
0
0
0.
0
0.
0
o
0.
o.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1 .
.005
.010
.020
.030
.040
.050
.060
.070
.080
.090
. 100
. 150
.200
.250
.300
350
400
450
500
550
600
650
.700
.750
.800
.850
.900
.950
.000
PART
PRESS
0
0
0
0
0
0
0
0
1
1
1
2
2
3
4
4.
5.
5.
6.
7.
7.
a.
8.
9
to.
10.
11 .
1 1 .
12.
.070
. 140
.279
.418
.556
.694
.832
.969
. 106
.242
.378
.053
.719
.376
.024
.664
.295
917
.532
. 138
.737
.328
.912
.488
.057
.618
. 173
.721
.262
TOEW
(OEG C)
-53
-45
-37
-32
-28
-25
-23
-20
-19.
-17
-15
-9
-5.
-2
0.
3.
5.
7.
9.
10.
12
13
14
15
16
18
18.
19
20
.221
.547
.328
.246
.510
.536
.056
.922
.047
.372
.858
.880
.499
.023
.865
.339
.503
.427
. 160
.735
. 180
.513
.750
.904
.986
.002
.961
.868
.729
TOROP
(DEC C)
0.
1 ,
3.
5
6.
8.
10.
1 1 .
13.
15.
17.
25.
34.
42.
51 .
59.
68.
76.
85.
93.
102.
110.
1 19.
127.
136.
144.
153.
161 .
170.
.850
.701
.401
. 102
.803
.503
.204
.905
.606
.306
.007
.510
.014
.517
.021
524
028
531
035
538
042
545
048
552
055
559
062
.566
.069
TAMB
(OEG C)
-52.
-43.
-33
-27.
-21
-17
-12.
-9.
-5.
-2.
1 .
15.
28.
40.
51.
62.
73.
83.
94.
104.
114.
124.
133.
143.
153.
162
172.
181
190
.371
.847
.926
. 143
.707
.033
.852
.018
.442
.066
. 149
.630
.515
.494
.886
.863
.531
.958
. 194
.273
.221
.058
.799
.456
.041
.561
.023
.434
.798
TABLE I I 1-3
VAPOR EQUIVALENCE RATIO
MAP
(KPA)
75
75
75
75
75
75
75
75.
75.
75
75.
75.
75.
75
75
75
75
75
75
75
75.
75.
75.
75.
75.
75.
75.
75.
75.
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
000
000
.000
000
000
.000
000
VAPOR
PHI
0
0
0
0
0
0
0
0.
0.
0
0.
0.
0.
0.
0.
0
0.
0.
0
0.
0.
0.
0
0
o
0.
0
0
1
.005
.010
.020
.030
.040
.050
.060
.070
.080
.090
. 100
. 150
.200
.250
.300
.350
.400
.450
.500
.550
.600
.650
.700
.750
. BOO
.850
.900
.950
.000
PART
PRESS
0.
0
0
0
0
o
0
0.
0
0
1 .
\
2.
2.
3.
3.
3.
4
4
5
5
6
6
7
7
7
8
8
9
.052
. 105
.209
.313
.417
.520
.624
.727
. 829
.932
.034
.540
.039
.532
.018
.498
.971
.438
.899
.354
. 803
.246
.684
. 1 16
.542
.964
.380
.791
. 197
TDEW
(DEG C)
-56.
-48
-40
-35
-32.
-29
-26
-24.
-23.
-21
-20.
-14.
-9.
-6.
-3.
-1 ,
0
2
4
5
7
a
9
10
1 1
12
13
14
15
.259
.800
.817
.886
.263
.381
.977
.910
.094
.472
.006
.222
.985
.625
.835
.446
.644
.501
. 174
.694
.087
.373
.566
.679
.722
.702
.626
.500
.330
TDROP
(DEG C)
0.
1 ,
3.
5.
6
8
10
1 1 .
13.
15.
17.
25.
34.
42.
51 .
59.
68.
76
85
93
102
1 10
1 19
127
136
144
153
161
170
.850
.701
.401
. 102
.803
.503
.204
.905
.606
. 3O6
.007
.510
.014
.517
.021
.524
.028
.531
.035
.538
.042
.545
.048
.552
.055
.559
.062
.566
.069
TAMB
(DEG C)
-55.
-47.
-37.
-30.
-25.
-20.
-16.
-13.
-9.
-6.
-2.
1 1 .
24.
35.
47.
58.
68.
79.
89.
99
109
1 18
128
138
147
157
166
176
IBS
.409
099
.416
.784
.461
.877
.773
.005
.489
. 166
.999
.288
.029
.892
. 186
.078
.671
.033
.208
.232
. 129
.918
.615
.231
.777
.260
.688
.066
.399
MAP
(KPA)
85
85
85
85
85
85
85
85
85
85
as
85
85
as
as
as.
85.
85.
85.
85.
as.
85.
as.
85.
85.
85.
85.
85.
85.
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
000
000
000
000
000
000
000
000
000
000
ooo
000
000
VAPOR
PHI
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1 .
.005
.010
.020
.030
.040
.050
.060
.070
.080
.090
. 100
.150
.200
.250
.300
.350
.400
.450
.500
.550
.600
. 65O
.700
.750
BOO
.850
900
.950
.000
PART
PRESS
0
0
0
0
0
0.
0.
0.
0.
1 .
1 ,
1 .
2.
2
3.
3.
4.
5.
5.
6.
6.
7.
7.
8.
a.
9.
9.
9.
10.
.059
. 1 19
.237
.355
.473
.590
.707
.824
.940
.056
. 172
.745
.311
.870
.421
964
500
030
552
067
576
079
575
065
548
026
497
963
423
TDEW
(DEC C)
-54
-47
-39
-34.
-30
-27.
-25
-23.
-21 .
-19.
-18.
-12.
-8.
-4
-1 .
0.
2.
4.
6.
7.
9.
10.
1 1 .
12.
13.
14.
15.
16.
17.
.948
.396
.312
.316
.645
.723
.286
.191
.349
.705
.217
.351
.052
.642
.810
615
737
623
321
865
280
586
799
929
988
984
923
811
654.
TDROP
(OEG C)
0
1
3
5
6
8
10
1 1
13
15
17
25
34
42
51
59
68
76
85.
93
102
1 10
119.
127.
136
144.
153.
161 .
170.
.850
.701
.401
. 102
.803
.503
.204
.905
.606
.306
.007
.510
.014
.517
.021
.524
.028
.531
.035
.538
.042
.545
.048
.552
.055
.559
.062
.566
.069
TAMB
(DEG C)
-54
-45
-35
-29
-23
-19
-15
-1 1
-7
-4
-1
13
25
37
49
60
70
81
91 .
101.
Ill
121.
130.
140.
150.
159.
168.
178.
187.
.097
.695
.910
.214
.842
.219
.082
.286
.744
.399
.211
. 160
.962
.875
.21 1
. 139
.764
. 154
.356
.403
.322
. 131
.847
.481
.043
.542
.985
.377
.723
TABLE III-4
VAPOR EQUIVALENCE RATIO
MAP
(KPA)
65.000
65.000
65. OOO
65.000
65.000
65.000
6S.OOO
65.000
65.000
65. OOO
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
65.000
VAPOR
PHI
0.005
O.010
0.020
0.03O
O.O40
0.050
0.060
0.070
0.080
0.090
O. 100
0. 150
0.200
0.250
0.300
0.350
0.400
0.45O
O.SOO
0.550
0.600
0.650
0.700
0.750
0.800
O.B50
O.900
0.950
1 .OOO
PART
PRESS
0.045
O.091
0. 181
O.271
0.361
O.451
0.541
0.630
0.719
0.807
O.896
1 .335
1 .768
2. 194
2.616
3.031
3.441
3.846
4.246
4.640
5.029
5.413
5.792
6. 167
6.537
6.902
7.263
7.619
7.971
.TOEW
(DEG C)
-57.739
-50.383
-42.515
-37.656
-34.088
-31 .249
-28.882
-26.847
-25.059
-23.463
-22.020
-16.328
-12. 161
-8.857
-6. 1 13
-3.765
-1.711
0. 115
1 .759
3.252
4.621
5.885
7.057
8. 150
9. 174
10. 136
1 1 .044
11 .903
12.717
TDROP
(DEG C)
0.850
1 .701
3.401
5. 102
6.803
8.503
10.204
1 1 .905
13.606
15.306
17.0O7
25.510
34.014
42.517
51 .021
59.524
68.028
76.531
85.035
93.538
102.042
1 10.545
1 19.048
127.552
136.055
144.559
153.062
161 .566
170.069
TAMB
(DEG C)
-56.889
-48.682
-39.113
-32.554
-27.285
-22.745
-18.678
-14.942
-1 1 .454
-8. 157
-5.013
9. 182
21.853
33.661
44.907
55.760
66.317
76.646
86.793
96.790
106.663
1 16.429
126. 105
135.702
145.229
154.695
164. 106
173.468
182. 7B6
1
o
I
-------
-11-
of 0.05 at -19°C and 85 kpa MAP implies a mixture temperature
drop of about 8.5°C which seems small enough that the
resulting heat transfer can initially be ignored.
If there is essentially no temperture drop and no heat
transfer, then from an equilibrium point of view, evaporation
before compression does not affect the energy balance which
determines the conditions at the end of compression, so such
evaporation can be ignored in the equilibrium calculations.
With these considerations in mind, in order to evaluate the
effect on equilibrium conditions of the incrementally
increasing pressure that occurs as the piston travels
upwards, we began by reevaluating equation III-2 and equation
III-l. Equation III-2 can be written as
(III-5) TDEW = A/tlogt 0(PV)-B]
where:
A = -1961.8678
B = +8.6339821
Performing some algebra on III-5, we can write the
solution for PV as:
(III-6) PV = 10J
where :
J = (A/TDEW) + B
If we assume for this analysis that the initial pressure
levels in the cylinder under cranking conditions are the same
as the average intake manifold conditions, then we can
substitute for PV in the partial pressure equation (equation
III-l) with the relationship identified in III-6.
(III-7) 10J = (MAP) (1/[1 + (7.155/VPHI)] )
Solving for VPHI, we have
(III-8) VPHI = 7.155/[(MAP/10J) - 1]
Solving III-8 by substituting selected values of TDEW in
the "J" identity results in approximately the same values as
in Tables III-l through III-3. (round-off errors are assumed
to be the cause of the small differences occuring in the
third decimal place). However, equation III-8, is now a
general equation relating pressure (MAP) and saturation
temperature in the "J" term to the vapor equivalence ratio
(VPHI). Further, if we solve equation III-4 for TDEW, and
substitute these results in the expression for "J", we have a
relationship that now includes the effect of the temperature
-------
-12-
drop from the vaporizing fuel, and the ambient
necessary to maintain the vapor equivalence ratio.
temperature
(III-9)
where:
(111-10)
J = [A/(TAMB-TDROP)1 + B
TAMB = (°K)
TDROP = (5/9)(1/Cp)[(f/a)s(VPHI)(HLG)] = (°K)
(f/a)s = 0.155 = stoichiometric fuel air ratio
VPHI = vapor equivalence ratio
HLG = 474 BTU/lb = heat of vaporization (ref. 4)
The relationship for "J" can further be modified to
include the adiabatic temperature which would reflect a new
saturation temperature as the cylinder pressure is increased.
(III-ll)
J = [A/(TAD - TDROP)] + B
where:
TAD = adiabatic temperature (°K) =TAMB (p2/MAP)R
TAMB = (°K) = ambient temperature
p2 = new cylinder pressure
R = (k-l)/k
k = 1.38 = ratio of specific heats for a
methanol and air mixture (4)
MAP = manifold air pressure which is assumed
be the same as the initial conditions
the cylinder during cranking conditions.
Putting everything together, we have
(111-12) VPHI = 7.155/[(pz/10J) - 1]
to
in
where
(111-13)
pz = (mm Hg) = cylinder pressure (divide kPa by
0.133224 to obtain mm Hg)
J = (A/[TAD - (5/9)(l/Cp)(f/a)s (VPHI)(HLG)]) +B
A = -1961.8678
B = +8.6339821
CP = .240
(f/a)s = 0.155
(HLG) = 474 BTU/lb
TAD = (°K) = TAMB (pz/MAP)R = adiabatic temp-
erature resulting from increasing the
cylinder pressure from MAP to pz. For
our analysis, p2 was- incremented from the
MAP value to determine TAD values.
TAMB = (°K) = ambient temperature
R = (k-l)/k
k = 1.38
-------
-13-
With equation 111-12, we now have an equation* that allows
us to increase the cylinder pressure from the manifold
conditions for a given ambient temperature, and observe the
resultant vapor equivalence ratio that would be allowed to occur
under equilibrium conditions. If the vapor ratio goes down as
the pressure is increased, then we can assumed that some of any
pre-vaporized fuel will condense, if it goes up we can assume
that any fuel vaporized in the intake tract (in this analysis)
will remain vaporized.
Figure III-2 shows the results of solving equation 111-12
for four ambient temperature levels. From 85 kPa (the
approximate MAP observed during cranking) to over 3500 kPa
(absolute pressure), we see a steadily increasing vapor
equivalence ratio that is supportable under the equilibrium
conditions given. Equation 111-12 also allows us to explore
the concept of temperature reserve (TR) as it relates to the
potential for predicting the upper limits of cold start
performance with neat methanol. The assumption is that
equilibrium values will define the maximum permissable
performance, and other factors such as heat transfer rates, and
allowable fuel evaporation rates will only serve to reduce the
maximum perfomance. The magnitude of these compromising factors
and their influence is for the time being ignored in order to
determine the upper limit of performance.
Since it is generally accepted that a certain amount of
fuel vapor is necessary to initiate combustion, we can use
equation 111-12 to determine an equilibrium temperature reserve
in the following manner. First, we initially define the
temperature reserve (TRX) as the difference between the
temperature that theoretically exists in the cylinder after
adiabatic compression (TAD) minus fuel evaporation temperature
losses (TDROP) and minus the saturation temperature (TDEW) of
the vapor mixture, without accounting for heat losses to the
cylinder walls.
(111-14) TRX = (TAD - TDROP) - TDEW
Note, that the adiabatic temperature and the temperature
drop are related (in the "J" term) to the vapor equivalence
ratio (VPHI) for a given cylinder pressure by equation 111-12.
The calculated vapor equivalence ratio also defines the
saturation temperature (TDEW) . If we were to use the
* Note that equation 111-12 has VPHI in the "J" term as well.
In order to solve 111-12, we seeded a computer program with VPHI
and iterated 111-12 until the seeded value and the calculated
value were essentially equal.
-------
EO. VflPOR PHI VS COUP PRESS
2.0
1.8
1.6
l.U
1.2
1.0
0.8
0.6
r
a.
o
Q.
cc
OD
_l
CC
0.2
0.0
RMB«15.P U) oSSKPfl.
MVR00.65
N»65 _ _____
RMSOO.P (1) eSSKPR. M
VR«0.65
RM8«-IS,P (1) 08SKPR
.HVR00.65
N*6S
RM80-35.P (1) 085KPR
.MVR00.6S
0 500 1000 1500 2000 2500 3000 3SOO
COMPRESSION PRESSURE -KPfl (R8S)
, Hi- I
-------
-15-
related values for TAD, TDROP, and TDEW corresponding to a
given VPHI, we would find that the temperature reserve would
be essentially zero. But, if we limit the saturation
temperature (TDEW) to the value that occurs at the vapor
equivalence ratio corresponding to minimum vapor ratio (MVR)
required to initiate combustion, we would find that the
temperature reserve would begin to increase once the value
TAD minus TDROP exceeded our .artificially limited TDEW. This
TRX value is essentially a measure of the equivalence ratio
reserve between the maximum vapor equivalence ratio that can
be sustained under the particular pressure and temperature
conditions after evaporation, and the MVR value. If we
truely wish to represent the temperature reserve (TR) above
the MVR, then we must also limit the temperature drop (TDROP)
in equation 111-14 to that which would occur when evaporating
only the amount of fuel necessary to achieve the MVR
selected. Thus equation 111-14 becomes:
(111-15) TR* = TAD - (TDROP @ MVR + TDEW @ MVR)
Figure III-3 is graphic depiction of this phenonon for
several ambient starting temperatures and an assumed minimum
vapor ratio (MVR) of 0.65.
Before we pursue the temperature reserve concept
further, it would be useful to investigate the effect that
the assumed minimum vapor ratio (MVR) has on the conclusions
that may be drawn from the model. For instance the lower
flamability limit (LFL) for methanol at atmospheric pressure
corresponds to a vapor equivalence ratio of 0.455(19).
However, Browning (3) suggests that an MVR of 0.63 is
necessary for initiation of methanol combustion. Bardon (2)
on the other hand, suggests a ratio 0.61 is necessary.
Figure III-4 provides an insight on the effect that the
MVR has on the temperature reserve. Trends are plotted for
MVRs of 0.6 and 0.7, and for two ambient starting
temperatures. Note that the leaner the assumed MVR, the
greater the temperature reserve. Expressed otherwise, a high
*Note, whenever TR was less than or equal to zero, our
computer program automatically set TR equal to zero. The
following figures include this assumption. Future work may
find the amount of negative TR which is not shown in our
figures useful in determining the amount of energy required
to be added to the system to achieve a startable temperature
reserve. However, for our initial understanding, we choose
to set negative TR values equal to zero.
-------
EQ. TEMP RESERVE VS COMP PRESS
400
flM8*15.P moSSKPfl.
HVR00.65
RHBOO.P (1) «85KPfl,M
VRC0.65
flM8«-15.P (1) 085KPR
.MVR00.6S
flHB«-35.P (1) 885KPH
.MVR00.65
N-6S
500 1000 1500 2000 2500 3000 3500
COMPRESSION PRESSURE -KPR (BBS)
F/G. W -
-------
EO. TEMP RESERVE VS COMP PRESS
400
RMBOIS.P (1) .085KPR
.HVR80.60
flM6015,P (1) .085KPR
.MVRC0.70
fi«8e-3S.P (1) ,«85KP
fl.MVR«0.60
N«6S
flM8«-35.P (1) .085KP
fl.MVReO.70
N-6S
0 500 1000 1500 2000 2500 3000 3500
COMPRESSION PRESSURE -KPfl (R8S)
F/6. UL -
-------
-18-
ipor equivalence ratio requirement consumes temperature
'serve. This of course is due to the leaner mixture having
lower saturation temperature, and to less sensible heat
i-inn r-£imi™nrar1 t- r\ Trar^/^iv-iTti 4-ViQ dnallov a m^Mi n 4- o f •Fnol A 1 Q n
vapor
reserve.
a lower saturation temperature, and to less sensible heat
being removed to vaporize the smaller amount of fuel. Also
note that for a given compression pressure, there is
approximately a 21°C difference in temperature reserve
between an MVR of 0.6 and an MVR of 0.7 at either ambient
temperature. However, if we review figure III-3, we observe
approximately a 30°C to 40°C difference in temperature
reserve between different ambient temperatures (+15, 0, -15*)
at a constant MVR (0.65). Therefore, the selection of MVR
can have an appreciable effect on the results, since the
effect of the value selected can be a sizable portion of the
effect due to a change in temperature with a constant MVR.
The MVR selected for modeling cold start performance may
be a function of engine combustion chamber design or it may
be a combination of factors such as the dryness of the
methanol, the chamber design, or the ignition energy. It
appears that one should be careful in selecting an MVR for
modeling purposes, and at least attempt to be consistent when
possible in comparing data from other investigations. For
our analysis we choose an MVR of 0.65 for the reason that at
least two methanol engines with different combustion chamber
designs (Nissan NAPS-Z and Ricardo HRCC) have demonstrated
(19) (22) that they can be operated with reasonable
performance at this overall equivalence ratio in a warmed-up
condition. Also, an MVR of 0.65 is slightly conservative
from the values used by Browning and Bardon.
Now that the concept of what we call "temperature
reserve" has been explained, we can proceed to use this
concept to evaluate the potential limits for cold starting
with neat methanol. Noticing that the temperature reserve in
Figure III-3 is a function of compression pressure, we can
compare the temperature reserve at the maximum compression
pressure with one engine at a given compression ratio and
ambient temperature to the reserve for another engine with a
different compression ratio or at a different ambient
temperature. For example, in Figure III-5 an engine with a
* The difference in temperature reserve between -15°C and
-35°C is approximately 40°C to 50°C.
-------
-19-
peak compression pressure* of 1000 kPa-abs** (130 psi gage)
would have a temperature reserve of 122°C at an ambient
temperature of 15°C. That same engine at an ambient
temperature of -35°C would only have a 23°C temperature
reserve. If it were assumed that a 122°C temperature reserve
were necessary to start a methanol engine, from Figure III-5
the maximum compression pressure would need to be increased
to around 2300 kPa-abs (320 psi gage)*** to obtain a vapor
equivalence ratio of 0.65 under equilibrium conditions and an
ambient temperature of -35°C.
Obviously increased compression is one means to increase
the temperature reserve and improve the fuel evaporation
during the compression stroke. Another means is by
increasing the throttling during cranking. Because the
volume ratio for compression would remain the same when
comparing wide open throttle (WOT) to closed throttle (CT) ,
and if dynamic effects were ignored, the adiabatic pressure
ratio would remain the same for either WOT or CT. Because the
pressure ratio would be the same, the adiabatic temperature
(TAD, Equation 111-12) would remain the same for either WOT
or CT (assuming that Tl is essentially the same for both WOT
and CT cranking conditions (ref. 18)). However, under CT
conditions, the absolute values of p2 and pi would be
lower than in the WOT case. But, because the adiabatic
compression temperature (TAD) remains the same, a re-
computation of equation 111-12 and 111-14 with the lower p2
pressure would result in a higher temperature reserve than
would occur under the original WOT conditions because the
partial pressure of the methanol vapor would be less.
A typical 4 cylinder engine was motored at speeds from
200 RPM to 600 RPM. An analysis of the vapor equivalence
ratio versus ambient temperature (at the rounded MAP values
from Table III-4) was shown in figure III-l, and indicated
relatively little improvement in the ability to pre-vaporized
the fuel prior to the compression stroke. However, if we
were to enter these rounded-off results into equation 111-13
for the "MAP" term, and recompute the temperature reserve by
* Assumes that measured cranking compression adequately
represents the actual "pz" in the P-V cycle.
** This pressure is assumed to be rather typical of low
compression SI engines converted to run on methanol.
Our current test engine, a 2 liter Nissan NAPS-Z engine
with 12.8:1 compression ratio has a cranking compression
of around 1960 kPa-abs (270 psi gage) when motored on a
dynamometer at 200 RPM, note the normal in-vehicle
cranking speed for this engine is specified as 300 RPM.
** *
-------
EQ. TEHP RESERVE VS COHP PRESS
400
RHB«15,P (D085KPR.
MVR00.65
N-65 _ ___
RMBOO.P (1) eSSKPR.M
VR00.65
RHBO-15.P (U Q85KPR
.MVR00.65
N«65
RMBO-35.P (1) «85KPR
.HVReO.65
N-65
500 1000 1500 2000 2500 3000 3500
COMPRESSION PRESSURE -KPR (BBS)
F.,
in- -
-------
-21-
the method in equation 111-15, we would find that the
temperature reserve would increase more markedly at the low
temperatures. Figure III-6, suggests that for an engine with
1000 kPa maximium cranking pressure and 85 kPa MAP , the
temperature reserve at 15°C ambient temperature would be
increased by about 35 percent by lowering the MAP from 85 kPa
to 65 kPa, but at -35°C ambient, the reserve would be
increased by almost 160 percent over the 85 kPa value. In
fact the temperature reserve of 60°C at -35°C ambient and 65
kPa is almost 50 percent of the reserve at 15°C ambient and
85 kPa, a condition at which most neat methanol engines
should start.
Table. III-4
RPM
200
400
600
Cranking MAP'
MAP*
91.72
75.53
62.86
Round-off**
90
75
65
curve fit (R2 = .995), MAP = 105.5567 - (. 072125) (RPM)
300 RPM = 84 kPa rounded to 85 kPA.
Round-off represents arbitrarily picked numbers close to
the observed values for analysis purposes.
Barometric Pressure = 99.17 kPa (wet).
It should be noted, however, that this analysis assumes
that some means is used to maintain the maximum cranking
pressure at a constant level when lowering the MAP. If such
means were not employed, then the maximum compression
pressure would be lowered by the ratio of the new MAP to the
old MAP, unless other factors such as leakage during the
compression stroke would limit the maximum compression
pressure at the higher MAP condition to a value that is below
that which could be achieved at the lower MAP condition with
no leakage. In comparing the measured p2 in our engine at
the different cranking speeds to that at 200 RPM, the actual
p2 is about 7.5 percent higher at 400 RPM and about 12
percent higher at 600 RPM. For a large V-8 gasoline truck
engine these values are about 10 percent at 400 RPM and 18
percent at 600 RPM. Because these values are relatively
small, in this analysis we will simply assume a straight
ratio of the new MAP value divided by the MAP at the normal
cranking speed to be multiplied by the pz at the normal
cranking speed. Those values will then be rounded to the
nearest 50 kPa which will be designated as "MAP corrected
pz" or "corrected p2".
-------
EQ. TEMP RESERVE VS COUP PRESS
400
350
u
UJ
o
300
250
oc
UJ
2200
ec
UJ
3
I-
QC
C
UJ
ISO
100
so
0
RH3015.P (1) 085KPH.
MVR00.65
RM801S.P (1) 06SKPP,
HVR00.65
ftM8o-3S,P (1) 085KPP
.MVR00.6S
N>65
RM8e-35,P (1)
.MVR00.6S
N-69
500 1000 1SOO 2000 2500 3000 3500
COMPRESSION PRESSURE -KPR (BBS)
. nr -
-------
-23-
Reconsidering the lower MAP, if there were a means to
substantially reduce the cranking MAP to the vicinity of
around 30 kPa - absolute (20.5 inches Hg manifold vacuum), we
could essentially raise the temperature reserve from 23°C to
179°C at -35°C ambient if we could maintain a 1000 kPa
maximum cranking pressure (i.e., 57°C higher than the value
for 85 kPa MAP at 15°C ambient - see figure III-7). However,
if we could not maintain the maximum cranking pressure
constant, and the maximum p2 was lowered to 350 kPa (ie.
MAP corrected PZ), the temperature reserve would be only
46°C — 76°C lower than the reserve at the 15°C ambient base
condition. Achieving a 50 kPa cranking MAP might be more
practical.* A 50 kPa cranking MAP with constant peak
pressure would still allow a 97°C temperature reserve at
-35°C ambient, nearly equivalent to the temperature reserve
of 122°C at a normal cranking speed and 15°C ambient.
Reducing the peak pressure by the MAP ratio to 600 kPa, would
lower the reserve to just 38°C.
Compared to the 23°C reserve for 85kPa MAP and -35°C
ambient (see figure III-7), the 38°C reserve at 50 kPa and
the 46°C reserve at 30kPa do show an improvement in the
temperature reserve with lower cranking MAP values.
However, lowering the ambient temperature causes a decrease
in the temperature reserve, and this decrease is not
compensated by the MAP effect. Therefore, a substantial
amount of the normal temperature reserve at 15°C (e.g., the
difference between 122°C and 46°C) would be presumably needed
to be made-up by other means if the engine were to start.
Further, achieving very low cranking MAPs with practical
cranking hardware while maintaining sufficient control of the
fuel-air mixture could be difficult. Also, smooth transition
from a high vacuum crank condition to a normal idle condition
might be tricky. Even so, the effect of lower manifold and
cylinder pressures should not be over looked. For a consumer
acceptable vehicle, the effect of altitude versus lower
cylinder pressure may affect the starting capability, and
should also be given consideration (i.e., the 5 inches Hg
pressure differential due to an altitude change of 1524
meters -5000 ft.- is approximately 17 kPa).
* 50 kPa can be achieved by increasing the cranking speed in
our engine to approximately 770 RPM (normal speed is 300
RPM), or presumably by increasing the throttling at the
normal cranking speed.
-------
EO. TEHP RESERVE VS COMP PRESS
400
RMBOIS.P moSSKPfl.
HVRoO.65
N-65
flMBB-35,P (1)»85KPfi
.MVR00.65
fiM8»-3S.P (1)OSOKPP
.MVR00.65
N = Sil
flMBo-3S.P (1)030KPR
.MVR90.65
N-58
0 500 1000 1500 2000 2500 3000 3500
COMPRESSION PRESSURE -KPflfflBS)
F/6. zzr-7
-------
-25-
In discussing means of increasing the temperature
reserve, we have digressed somewhat from the original
hypothesis of comparing the temperature reserve (TR) at known
starting conditions to those at unknown cold conditions. It
is generally accepted that the lower temperature limit for
unassisted starting of gasoline engines converted to run on
methanol is around 10°C (50°F) (10)(11). Without changing
pistons such engines usually have rather low compression
ratios (in the range of 8:1 to 9:1). A cranking compression
of lOOOkPa-abs (130 psi gage) was earlier assumed as being
reasonably representative of such engines. Under these
conditions (10°C ambient and 1000 kPa compression pressure),
the temperature reserve would be around 112°C for an MVR of
0.65 and a MAP of 85 kPa (see figure III-8).
Gardiner et. al, (10) extended these limits somewhat by
achieving successful cold starts at 0°C. The engine used was
a 2 cylinder air cooled 4 cycle SI engine with a compression
ratio of 5.84:1. The manufacturer suggests that normal
compression pressure on this engine is around 750 kPa-abs
(94psi gage). For this pressure and assuming a 65 kPa MAP
(since Gardiner's cranking speed was 600 RPM*), the
temperature reserve at 0°C ambient with a non-corrected p2
is 96°C, a value not substantially different from the assumed
standard engine at 112°C temperature reserve. The reserve
with a corrected pz of 550 kPa is 60°C. A visual
comparison of Gardiner's engine to the standard engine is
shown in figure III-8 as well.
M.A.N., on the other hand, with a modified diesel (CI)
engine with spark assist has reported instantaneous starts at
-2Q5C— (^)°F-ref. 12) . Informal accounts (15) indicate 3 second
starts at -32°C. The exact cranking compression of the
M.A.N. engine is unknown to us, but from informal
conversations with M.A.N., a range of 25 to 30 bar (2500 to
3000 kPa) was suggested. Typically, CI engines are not
throttled, and therefore, the cranking MAP would be closer to
atmospheric pressure, or approximately 100 kPa for this
analysis. Assuming that the M.A.N. engine has a 100 kPa MAP,
the temperature reserve at -35°C ambient would range between
107°C to 131°C (see Figure III-8) for the range of estimated
compression pressures, values which are still at or above the
value need to start the standard SI engine at +10°C ambient.
* The normal cranking speed for this engine is between 200
and 400 RPM. A nominal value of 300 RPM and 85 kPA MAP was
assumed for this analysis.
-------
-26-
If some throttling were used, for instance to achieve 85
kPA MAP during crank, the corrected MAP temperature reserve
(approx. 2100 kPa-p2)at -35°C ambient for the low range (25
bar) would increase from 107°C to 110°C. The reserve would
increase approximately 2°C for the higher value (approx. 2500
kPa corrected p2) as well - see figure III-9. If we were
to decrease the MAP even further to 65 kPa, we would increase
the temperature reserve another 7°C at the low range and 14°C
at the high range (117°C and 147°C respectively). In fact,
the temperature reserve for -50° ambient of 109°C at the high
compression pressure and 65 kPa MAP is nearly sufficient for
starting at -50°C since the reserve of our standard SI engine
at 10°C ambient is only 112°C.
These results of increasing temperature reserve with
decreasing MAP and corrected p2 match our earlier
observations with one exception. The exception is, in this
case decreasing the MAP causes a more observable change in
the projected starting capabilities at the lower ambient
temperatures. The key to the more observable change lies
simply in the fact that the high compression pressures result
in temperatures reserves that are initially nearer to the
reserve for our standard engine at 10°C.
Table III-5 provides a handy comparison of our analysis
for the three engines that have been discussed, the mythical
standard SI engine, Gardiner's engine, and the M.A.N. CI
spark assisted engine to projections based on the Temperature
Reserve concept. High compression pressures and/or a low MAP
during cranking appear to be avenues to achieve potential
unassisted cold start capabilities in the -30°C to -40°C
range -- the high compression through higher in-cylinder
temperatures during the cranking cycle, and the high speed
cranking through reduction of the partial pressure of the
fuel due to the lower MAP.
The high speed cranking may have an additonal factor
that assists in the cold starting phenomonon. With the high
cranking speed, there exists the potential for more
compression strokes to occur during cranking prior to start.
Since the actual engine cranking is not adiabatic, each
cranking stroke leaves behind some residual fuel vapor at a
temperature .higher than the ambient mixture, potentially
requiring less heat (or lower temperature reserve) on the
next stroke to achieve the MVR (minimum vapor equivalence
ratio) necessary for starting. Gardiner's starting procedure
of cranking the engine at 600 RPM for 10 seconds (i.e., 50
compression strokes) and then resting for 10 seconds followed
by two replications of this cycle, allows up to 150
compression strokes to occur for what Gardiner accepted as a
-------
-27-
EO. TENP RESERVE VS COUP PRESS
100
350
o 300
U4
UJ
250
c
£200
1C
« ISO
h-
(X
5 100
Q.
£
UJ
" so
oi
flriaoio.P (1)
HVRtO.65
N«6S
.P 111 065KPS.M
VR»0.65
H»69
RN8i-3S.P (1) elOCKP
a.MVReo.es
N-62
flM8«-50.P (1) olOOKP
0 500 1000 1500 2000 2SCO 3000 3500
COMPRESSION PRESSURE -KPfl !fl8S)
£0. TEHP RESERVE VS COHP PRESS
<400
350
u
o 300
UJ
a '
UJ
250
-------
-28-
Table III-5
Actual vs. Projected
Starting Capabilities
Condition
Reference
-Standard SI
-Gardiner (600 RPM crank)
-M.A.N. (12)
-M.A.N. (15)
Projected***
-SI engine with high
speed cranking (770 RPM)
-High compression SI engine
0 w/norrnal cranking (300 RPM)
Max Cyl
Press(kPa)*
TR Ambient
>C) Start
1000
550
2500
3000
2500
3000
85
65
100
100
100
100
112
60
142
173
107
131
10°C
0°C
-20°C
-20°C
-32°C
-32°C
600
0 w/high
-39°C
-M.A.N
speed cranking (770 RPM)
-M.A.N.
during
projected capabilities
w/intake throttling
crank
50
112
- 2°C
2100**
2500
3000
1600
2000
85**
1200
100
100
65
65
112
50
112
112
112
112
-34°C
112
-33°C
-43°C
-37°C
-49°C
* Corrected Pz
**Max pressure and MAP data from current test engine,
NAPS-Z with 12.8:1 compression ratio.
***Ambient Start temperatures are linearly interprolated
computer output matrix based on:
AMBN
ter output matrix based on:
= AMBL - ([AMBL-AMBH][(TRS-TRL)/(TRH-TRL)])
.ere:
f R.. = n *=*T*T amHiont- c i- a K f- -f-omnoKaf-iiyd
a Nissan
from the
Where:
AMBN = new ambient start temperature
AMBH = ambient temperature with a TR above TRS
AMBL = ambient temperature with a TR below TRS
TRS = the TR for the standard SI engine at 10°C ambient, 85
kPA-MAP, and 1000 kPa = 111.955°C
TRH = TR corresponding to AMBH
TRL = TR corresponding to AMBL
-------
-29-
successful start. A normal cranking speed (200-300 RPM)
would have only 30 to 50 percent as many compression strokes
for a given time. Thus if the exhaust residuals which in
this case are fuel vapors, build up with successive engine
cycles, the high speed cranking could have an advantage in
this area as well, and may be partially responsible for
Gardiner's ability to start with a lower temperature reserve
than the value predicted for other investigators with a
so-called standard engine (see figure III-8 and Table III-5).
Detroit Diesel Allison (DDAD) utilized a similar
approach with exhaust residuals to develop successful
autoignition in its methanol bus engine (13). The
autoignition temperature of methanol was identified by DDAD,
and the temperature loss due to methanol evaporation was
calculated. DDAD then increased boost pressure and
temperature, and varied exhaust back pressure to increase the
exhaust residuals in the engine in order to recover the
original temperature. Quite possibly a butterfly valve in
the exhaust pipe similar to those used for early fuel
evaporation systems (EFE) on V-8 gasoline engines, when
applied to 4 cylinder methanol engines would, when closed
during cranking, increase the residual fuel vapor retained in
the cylinder from successive cranking cycles. Another
possibility would be to delay the fuel (which is possible on
electronically fuel injected engines) until after several
cranking revolutions in order to have hotter cranking
residuals.
Since Table III-5 suggests the potential starting
capability in the -30 to -40°C range for high compression
ratio engines, and since there is not a plethora of methanol
engines with -30 to -40°C starting capability, other factors
need to be examined (Note also that most of the cold start
studies reported in the literature have generally been with
relatively low mechanical compression or low cranking
compression engines). First the analysis presented is a bulk
equilibrium analysis, and is intended only to show the outer
limits of potential start capabilities (i.e., if equilibrium
conditions cannot support a combustible vapor mixture,
non-equilibrium conditions do not stand much of a chance for
starting). In reality, the evaporation process which occurs
during the compression cycle is a series of non-equilibrium
processes that occur during the finite time of the piston's
upward travel. Time is a factor, in heat transfer to both
the methanol droplets and to the surrounding structure.
Compression heat is continuously transferred to both at a
rate depending on the difference in temperature and the heat
transfer coefficient for each. The maximum allowable
evaporation rate of methanol plays a critical role in
determining if a sufficient amount of fuel can be evaporated
by the end of the compression stroke.
-------
-30-
It is apparent that increasing the compression pressure
increases the temperature reserve which should increase the
potential for successful starting at cold ambient temperatures.
Slightly less apparent is the observation that when the
temperature reserve is the greatest (i.e., normal summer
temperatures, 25° - 40°C), the portion of the initial liquid fuel
that can be vaporized in the inlet manifold (based on equilibrium
conditions) is also the most favorable (see figure III-l) thus
tending to prevent the evaporation rate of methanol from being a
potential limiting factor because the amount that is prevaporized
can be a sizable portion of the MVR (20%-30%) . Both factors
(increased temperature reserve and increased manifold vapor)
enhance starting capabilities. Conversely, when the temperature
reserve is the smallest (i.e, low ambient temperature), the amount
that can be vaporized in the inlet manifold is also minimal, which
indicates low ambient temperature conditions have two factors
working against successful cold starts - (1) reduced ability to
vaporize the fuel during the compression stroke, and (2) reduced
capability to pre-vaporize some fuel in the intake manifold prior
to the compression stroke.
To further explain the failure of real methanol engines to
start at low ambient temperatures for which their theoretical
temperature reserve seems adequate based on the testing of low
compression engines at higher ambient temperatures (10+°C), we
must examine how ambient temperature itself effects the
non-equilibrium processes which are at work.,- One fact which seems
important is that as the ambient temperature decreases, the
difference between the peak compression temperature and the
ambient temperature also decreases. This temperature difference
is roughly the driving force for initial heating of fuel droplets
which began at ambient temperature. Later, when fuel droplets can
be approximated as having been warmed to the boiling point
corresponding to the total pressure in the cylinder, the driving
temperature difference for heat transfer will also be reduced at
lower ambient temperatures since the peak compression temperature
is reduced but total cylinder pressure is not. Thus, heat
transfer and evaporation will be slower for a given temperature
reserve at lower ambient temperatures, impairing starting ability.
In fact, the evaporation rate of methanol may be the only
extremely critical factor in starting a methanol engine under cold
ambient conditions. When . evaluating the M.A.N.
-------
-31-
system* (12), it is interesting to note that the fuel is
injected late in the compression stroke. This provides a
very large temperature difference between the air temperature
and the fuel temperature which enhances heat transfer to the
fuel. Secondly, and possibly more important is that the
engine design does not attempt to evaporate all of the fuel
at once. Instead, the fuel is deposited in a bowl in the
piston. There, only the exposed surface is open for
evaporation, and the evaporation that does occur, cools the
piston, not just the air. After ignition by a spark plug,
the flame provides for controlled evaporation and combustion
of the remaining fuel in a manner that allows an acceptable
pressure rise in the cylinder.
Gardiner, on the other hand, indicated that increasing
the amount of liquid fuel to the engine from an initial
liquid equivalence ratio of about 1.0 to about 3.0 improved
the cold starting performance of a carbureted SI engine.
However, increasing the initial liquid ratio from 3.0 out to
10.0 made little change in the starting performance. One
explanation of this may be that the additional fuel surface
area achieved by an equivalence ratio of 3.0 was enough to
prevent the evaporation rate from being a limiting factor,
such that the work of compression was the limiting factor as
equivalence ratio was further increased. A given amount of
compression work will support only a certain amount of
methanol vapor at equilibrium conditions for a given ambient
temperature (see figure III-2). Another explanation is that
the evaporation rate was limiting throughout Gardiner's range
of equivalence ratios, but because of poor fuel atomimization
and distribution, the fuel surface area in the cylinder did
not increase past an equivalence ratio of 3.0. Gardiner's
own explanation is also plausible, namely that surface area
and evaporation rate did steadily increase with increasing
equivalence ratio, but (expressed in our terms), the
temperature reserve was possibly consumed by the need to heat
large amounts of liquid fuel to the boiling point. Gardiner
does not indicate whether he attempted operation beyond an
initial liquid equivalence ratio of 10.0.
From this information, a hypothesis is put forward that
may be difficult to justify, but none the less may stimulate
further thought in this area. It is hypothesized that under
cold starting conditions that the rate of evaporation of
methanol is the limiting factor. During the compression
stroke, it is hypothesized that the gas temperature remains
*A modified diesel engine with spark assist.
-------
-32-
well above the saturation temperature for the partial
pressure of methanol; but because rate limitations have not
allowed sufficient methanol to evaporate, the temperature has
not been driven down. Figure 5 in Bardon's paper (2) which
shows the measured cylinder temperature after methanol
vaporization to be much greater than the theoretical
temperature after vaporization tends to support this
hypothesis. If sufficient methanol has not evaporated to
drive the temperature down, also sufficient methanol vapor is
not available to form a combustible mixture.
Increasing the overall liquid mixture in order to
potentially increase the surface area of the fuel available
for heat transfer and subsequent evaporation (such that a
combustible mixture is formed) would require that the
remainder of this excess liquid fuel must be consumed by the
flame. However, if the excess fuel to be consumed by the
flame was delivered to the combustion chamber in a manner
such that large droplets remain after the initial
vaporization, there is evidence to suggest that the flame
will not propagate due to insufficient preliminary
volatilization of the large droplets. Hence, the initial
flame front could essentially be quenched by the large
droplets resulting from the excess fuel causing a no-start
condition.
Burgoyne and Cohen (20) conducted an experiment with
tetralin and air mixtures which evaluated the flame
propagation of various sized fuel droplets surrounded by
just-flammable fuel vapor. Their apparatus was configured
such that the droplets flowed downward through a burning zone
at the end of a long tube. The combustion was supported by
the just-flammable fuel vapor, but the vapor flame would not
propagate up the tube. When the droplets were added in
appropriate amounts the flame then propagated up the tube.
They found that for droplets above 300 microns in diameter,
the flame would not propagate upwards against the flow which
they suggest is due to the velocity of the fall being greater
than the burning velocity such that the mass motion of the
droplets prevents propagation. Such mass motion is always
evident to some degree within the cylinder during the
compression stroke. The spark plug could be envisioned as
the location of the just-flamable burning zone and the
droplets would pass through this region due to the swirl in
the cylinder. This might suggest that the more quiescent
combustion chambers need less temperature reserve for
starting than one with high swirl even though the high swirl
may be more desirable once the engine is running.
Conversely, the higher swirl engine may need a larger MVR to
compensate the initial vapor flame speed for the higher local
air velocities.
-------
-33-
Another aspect of Burgoyne's and Cohen's work is the
concept of two regimes of droplet evaporation prior to
combustion. The two methods are evaporation controlled by
diffusion, and evaporation controlled by heat transfer. The
boundary between the two mechanisms is controlled by what
they define as the "stabilized temperature" which is defined
as the " 'wet-bulb' temperature which is below the
boiling-point by reason of the heat carried away by the
vapour". If we compare the heat of vaporization between
methanol (474 BTU/lb or 1.102 J/kg - ref. 4) to that of
tetralin (0.201 J/kg - ref. 23), we see that the amount of
heat that can be absorbed by the heat of vaporization is
about 5.5 times greater for methanol than for tetralin.
Therefore, one would expect the "stabilized temperature" for
methanol to be much higher than that for tetralin because of
the additonal heat transferred away from the droplet by the
evaporating methanol.
Burgoyne and Cohen also point out that aerosols
(droplets under 10 microns) have much higher burning
velocities because the time required to evaporate the
droplets is much shorter than for larger droplets. If we
assume that droplet evaporation will occur much more quickly
in the heat transfer controlled regime than in the diffusion
controlled regime, then because of the higher stabilized
temperature for methanol which keeps the methanol in the
diffusion controlled regime longer, one would expect the
pre-flame evaporation of methanol droplets to take much
longer than tetralin.
Transferring these methanol pre-flame evaporation
assumptions back to Burgoyne's and Cohen's propagation tube
results, we would predict that the 300 micron upper droplet
size limit for propagation with tetralin in air mixtures
would be reduced (possibly substantially) for methanol in air
mixtures because of the longer time needed for pre-flame
volatilization of the methanol droplets. The lack of flame
propagation due to the pre-flame evaporation rate and droplet
size factors could explain the failure of other experimenters
to achieve cold start temperatures as low as predicted here
through purely liquid enrichment. To be accurate, however,
it must be pointed out that not enough experimenters have
reported cold start investigations with high compression
ratios or low MAPs to allow a fair appraisal of the
limitations of unassisted methanol cold starts. Furthermore,
some experimenters have added liquid fuel in a way which may
not have increased the surface area available for evaporation
(the additional fuel may have merely pooled in the manifold
or have merely formed a thicker layer on the cylinder walls),
or added fuel in a way that the fuel droplets may have
congealed forming droplets too large to allow pre-flame
volatilization and subsequent flame propagation.
-------
-34-
Based on this hypothesis, CI engines and modified CI
engines (M.A.N.) are expected to be easier to start than
homogeneous charged SI engines. There are several factors to
support this statement. First, CI engines almost always have
higher compression ratios which have correspondingly higher
compression temperatures. Also, CI engines inject the fuel
late in the compression stroke where there is a much larger
temperature difference between the air and the fuel (which
should increase the vaporization rate). Further, because the
fuel is injected by a high pressure injection system, it
usually has smaller droplets than a homogeneous charge engine
which enhances both diffusion controlled evaporation and
pre-flame heat transfer controlled evaporation. Lastly, the
initial part of the injection cycle creates a very lean f/a
ratio which also improves partial pressure or diffusion
controlled evaporation. All four factors, higher compression
temperature, larger temperature differential, smaller
droplets, and lean f/a ratio enhance the evaporation rate.
Because the potential for the evaporation rate during the
initial part of the injection cycle is much greater under the
CI type conditions than those in an SI engine, less overall
liquid fuel should be required to achieve a combustible
vapor. Therefore, since less overall liquid may be required,
the probability of excess liquid or large droplets quenching
the initial flame front would most likely be reduced.
Finally, even if excess liquid does occur, certain CI
combustion systems (e.g., M.A.N.) allow for controlled
evaporation of the excess liquid in order to prevent
quenching of the pilot flame.
Since many of the features that make cold starting more
probable with a CI or late (near TDC) injecting style of
engine may not be easily transferred to an SI engine, the
outlook for the cold start performance of homogeneous charged
SI engine does not look promising. Items that may be
critical to cold start performance of the SI methanol engine
without starting aids include devising means to increase the
evaporation rates through such mechanisms as higher
compression temperatures through increased compression
ratios, smaller droplet sizes through improved fuel systems
(e.g. EFI, sequential EFI, higher pressure EFI, or other
atomization systems), higher cranking speeds or increased
throttling, reduced heat transfer to the structure during the
compression stroke, increased heat transfer between the air
and the fuel during the compression stroke, and possibly
devising means to better use the cranking residuals. Flash
vaporization might be another possibility as described by Oza
(14), however this technique will be discussed more
thoroughly in the chapter on physical vaporization.
-------
-35-
C. Summary
In summary, we have identified several instances of
successful cold starting of unassisted methanol engines
(Gardiner (10) at 0°C, and M.A.N. officially at -20°C,
unofficially at -32°C). A hypothesis has been developed that
suggests that the limiting factor in unassisted methanol cold
starting is not the equilibrium vapor pressure curve, but the
rate and time sequence of the fuel evaporation. Experimental
work by Gardiner (10), and modeling by Bardon (2) and
Browning (3) tend to support this hypothesis.
This information then suggests that the potential for
cold starting a methanol engine without assist at very low
ambient temperatures is still a possibility. High
compressions ratios and smaller droplets are two possible
factors required to achieve acceptable cold start
capabilities. A list of perceived advantages and
disadvantages is given below.
1. Advantages
a) Cost of Vehicle Fuel System -
cost of a
unassisted
expensive than
associated power
single-fuel system
systems.
Most likely
single-fuel vehicle fuel system
methanol cold starting will be
one with auxiliary heaters
requirements. Certainly
should be less than dual
the
for
less
and
the
fuel
b) Cost of Fuel -
additives should
volatility additives
Methanol
cost less
with no volatility
than methanol with
c) Emissions - With a single fuel, it will most
likely be easier to measure, characterize, and
control both exhaust and evaporative emissions
compared to methanol with volatility additives
or dual fuel systems.
d) Enforcement - With a single fuel, disputes over
using or selling a winter fuel during the summer
or in an unacceptable region of the country will
not occur.
-------
-36-
e) Vehicle Maintenance - A single-fuel vehicle fuel
system does not need the range of authority that
a system would need for methanol with volatility
additives. This could make the single-fuel
system less complex and easier to maintain. The
single- fuel system would almost certainly be
less complex than dual fuel systems.
2. Disadvantages
a) Universal Demonstration - With the exception of
a few situations, universal application of
successful cold start technology has not been
demonstrated. No successful cold start with SI
engines in the -35°C range have been
demonstrated.
------- |