LIBRARY
OF THE
UNIVERSITY OF CALIFORNIA.
RECEIVED BY EXCHANGE
Class
THE
HEAT ENGINE PROBLEM
CHARLES EDWARD LUCRE, M.S.
SUK.MITTKD IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY IN THE
FACULTY OF APPLIED SCIENCE, COLUMBIA UNIVERSITY.
NEW YORK
1902
THE
HEAT ENGINE PROBLEM
BY
CHARLES EDWARD LUCKE, M.S.
SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY IN THE
FACULTY OF APPLIED SCIENCE, COLUMBIA UNIVERSITY.
t /
NEW YORK
1902
V
'
CONTENTS.
PAGE.
THE HEAT ENGINE PROBLEM ....... i
Introduction ......... i
Resume of Work and Results . . .... 3
A METHOD OF CYCLIC ANALYSIS OF HEAT ENGINES . . 9
Heat Engine Cycles Analyzed . . . ... 9
The Atmospheric or Vacuum Cycles . . . . I9
Comparison of Cycles . . . . . . . 82
k Temperatures after Addition of H v B. T. U. . . 82
Pressures after Addition of #j, B. T. U. . . .84
Volumes after Heating by ff v B. T. U. . . .85
Temperature after Expansion . . . . .86
Pressures after Expansion . . . . . .87
Volumes after Expansion . . . . . .89
Heat Discharged or Abstracted . . . . .90
General Propositions . . . . . . .94
Efficiencies ........ 05
Temperatures ........ 96
Pressures . . . . . . . . .08
Volumes . . . . . . . . 99
THE HEAT ENGINE PROBLEM (No. 926) . . . . l
Coal -burning Engines ...... .20
Oil-burning Engines . . . . . . . .29
Gas-burning Engines . . . . . . . ^r
LIQUID FUEL COMBUSTION (No. 0934) . . . . I
PHYSICAL PROPERTIES OF EXPLOSIVE MIXTURES . . . . i
Rate of Propagation ........ 5
Temperature of Combustion ...... 7
Resume of Temperature of Combustion . . 8
218506
2 CONTENTS.
PAGE.
SOME NEW WORK ON PROPERTIES OF EXPLOSIVE MIXTURES . .12
The Apparatus ....... .12
Neutral Gas Generator ..... .15
Constant Pressure Combustion Gas Calorimeter . . .16
Constant Volume Combustion Pressure Ratio Chamber . .17
Constant Pressure Combustion Volume Ratio Apparatus . .18
Results Obtained with Apparatus. P'lames in Atmosphere of
Different Air-Gas Mixtures . . . . .20
Constant Pressure Combustion Calorimeter . . .22
Volume-ratios During Constant Pressure Combustion . 26
Pressure-ratios for Constant Volume Combustion . .27
Conclusion . . . . . . . -3
THE HEAT ENGINE PROBLEM.
INTRODUCTION.
IT is now a good many years since the first proposition was
made for obtaining work by the heat-transforming action of a per-
fect gas and though each process as it appeared has been more or
less completely worked out by those interested in it to show the
possibilities of the system and compare it with others yet no inves-
tigation of all systems with their mutual relations has ever been
made by a general method. This is desirable because no compari-
sons can be justly drawn otherwise and it is, unfortunately, true
that invariably in the past the best conditions of one system have
been v selected for comparison with some other system working
under indifferent conditions. This may not have been done inten-
tionally, in fact it appears that in many cases the comparison
seemed perfectly just to the author but the results are almost
valueless as bases of generalization for the purpose of reaching
clear notions of comparative value.
A commonly used mode of comparison considers different cycles
working through the same temperature range whereas equal quanti-
ties of heat for each case will result in different temperature ranges
and it is pretty clear that comparisons should be made on the basis
of some initial conditions one of which is the heat supplied.
A perfect gas will transform heat into work and considering the
gas alone without reference to any engine the fraction of the heat
that is transformed is dependent on the relation both in sequence
and extent of the operations of heating, cooling, expansion and
contraction in short is dependent on the cycle first and the extent
of each cyclic phase secondly. It is first required to find out just
how much heat energy will be transformed by each cycle and if
other things are equal one should be the best for application to
engines. But in this comparison we should consider not only
which cycle transforms the largest amount of the heat energy
supplied to it into work but also through just what range of pres-
sures, volumes and temperatures these cycles must operate to
produce the work This comparison will be purely mathematical
and will, when completed, enable us to select the best cycle or best
two cycles as the case may be, i. e., that one or those two cycles
i
2 ... 9 . t ..Ttq:..yEAT ENGINE PROBLEM.
that promise the best returns for the labor spent on designing
mechanism to execute the cyclic changes.
Having made the mathematical selection of the cycles best
adapted to our purpose we are called upon to consider how to
heat or cool to cause expansion or contraction with the means at
our command and at the rate required. This second part involves
all questions of possibility or practicability of doing what seemed
mathematically to be desirable.
To place each of the cycles in proper relation each with the other
and to show the physical possibility of executing those promising
good returns as power generators is the general problem. More
particularly the question resolves itself into a search for an effec-
tive competitor of the Otto cycle engine which now is the only
good heat engine of the perfect gas sort.
As the work progressed beyond the mathematical analytic stage
there appeared a cycle which promised good returns for any labor
expended on its development but which has been comparatively
neglected. The latter part of the work is taken up with a study
of physical and engineering problems entering into the execution
of this theoretically desirable cycle in engines and includes the
determination of many of the physical constants necessary for
computation of designs. In this part also there is set down all
the difficulties to be encountered and both the solutions obtained
and the need of solutions for those questions still open are noted.
RESUME OF WORK AND RESULTS.
The work was taken up in detail as follows, and each section
brought to a successful conclusion except where otherwise stated :
PART I.
NEW CLASSIFICATION OF CYCLES AND DIAGRAMS OF SAME IN
P. V. & 6$ COORDINATES.
Cycle I Isometric heating; adiabatic expansion; isopiestic cooling.
Cycle IA Isometric heating ; adiabatic expansion ; isometric cool-
ing ; isopiestic cooling.
Cycle IB Isometric heating ; adiabatic expansion ; isothermal cool-
ing ; isopiestic cooling.
Cycle 1C Isometric heating ; adiabatic expansion ; isothermal
cooling.
Cycle II Adiabatic compression ; isometric heating ; adiabatic
expansion ; isopiestic cooling.
Cycle IIA 2 Adiabatic compression ; isometric heating ; adiabatic
expansion ; isometric cooling.
Cycle UA l Adiabatic compression ; isometric heating ; adiabatic
expansion ; isometric cooling ; isopiestic cooling.
Cycle IIB Adiabatic compression; isometric heating ; adiabatic
expansion ; isothermal cooling ; isopiestic cooling.
Cycle IIC Adiabatic compression ; isometric heating; adiabatic
expansion ; isothermal cooling.
Cycle III Adiabatic compression ; isopiestic heating ; adiabatic
expansion ; isopiestic cooling.
Cycle IIIA Adiabatic compression ; isopiestic heating ; adiabatic
expansion ; isometric cooling ; isopiestic cooling.
Cycle IIIB Adiabatic compression ; isopiestic heating ; adiabatic
^expansion; isothermal cooling; isopiestic cooling.
Cycle 1 1 1C Adiabatic compression ; isopiestic heating ; adiabatic
expansion ; isothermal cooling.
Cycle IV Adiabatic compression ; isothermal heating ; adiabatic
expansion ; isopiestic cooling.
Cycle IVA Adiabatic compression ; isothermal heating ; adiabatic
expansion ; isometric cooling ; isopiestic cooling.
Cycle IVB Adiabatic compression ; isothermal heating ; adiabatic
expansion ; isothermal cooling ; isopiestic cooling.
3
4 THE HEAT ENGINE PROBLEM.
Cycle IVC Adiabatic compression ; isothermal heating ; adiabatic
expansion ; isothermal cooling.
Cycle V Adiabatic compression ; any law of heating ; adiabatic
expansion ; isopiestic cooling.
VB > Similar meanings to preceding cases.
vcj
Cycle VI Atmospheric heating ; isometric cooling ; isothermal
cooling.
Cycle VII Atmospheric heating ; adiabatic expansion ; isopiestic
cooling ; adiabatic compression.
Cycle VIII Atmospheric heating ; adiabatic expansion ; isothermal
cooling.
Cycle IX Atmospheric heating ; adiabatic expansion ; isometric
cooling ; adiabatic compression.
Cycle X Atmospheric heating ; adiabatic expansion ; any law
of cooling ; adiabatic compression.
FOR EACH CYCLE A FORMULA is DERIVED EXPRESSING EACH
OF THE FOLLOWING VARIABLES AS A FUNCTION
OF THE HEAT SUPPLIED H v
(p y v, T) for every point of the diagram.
7/ 2 or the heat discharged as unavailable.
W H l H 2 or the amount of energy transformed into work.
TT
EI~ or the efficiency, the fraction of energy supplied
"\
that becomes transformed to work.
R^ or entropy range.
(TT , TT \
l - 2 - ] or mean -effective temperature.
K(f )
R v or volume range.
M.E.P. = ~ r or mean-effective pressure.
R ,
R n or pressure range.
M.E.V. -=r-0r mean effective volume.
R P
R ( or temperature range.
CYCLES COMPARED.
The formulae derived are here collected and curves drawn in
two coordinates. One coordinate is in every case //, the heat sup-
RSUM. 5
plied and the other coordinate the variable under consideration.
This gives one curve in every set for each cycle and as many sets
as there are variables. Some of these are less important than
others and the former are omitted and a set presented for each of
the following.
Curves of Temperature after Heating.
Curves of Pressure " "
Curves of Volume " "
Curves of Temperature after Expansion.
Curves of Pressure " "
Curves of Volumes " "
Curves of Heat Discharged or Abstracted.
Curves of Efficiency.
Curves of Mean Effective Pressure.
Curves of Mean Effective Volume.
Curves of Mean Effective Temperature.
Comparison and interpretation of curves.
Selection of a cycle to be applied to engines, the selection based
on theoretic grounds alone.
PART II.
THE EXECUTION OF THE CYCLES BY MECHANISMS.
All heat to be derived from a fire and may be imparted to the
gas in three ways : (a) Through walls, (fr) by introduction of
hot body, (c) internal combustion.
External heating condemned.
Introduction of hot masses impracticable.
Internal combustion advocated.
Internal heating by coal, oil, gas.
Explosive internal combustion.
Non-explosive internal combustion.
Explosive internal combustion in engines discussed.
Explosive engines pretty well known and now receiving much
attention hence this question left for study of less well known types.
Other types of internal combustion engine considered alone and
in relation to others.
Two typical classes of these non-explosive engines stand far in
front of others from every point of view, the Brayton and the Diesel.
Review of cyclic analysis so far as it refers to the three typical
cases of the practicable cycles.
6 THE HEAT ENGINE PROBLEM.
Left for further study Diesel, Otto, Brayton and their variations.
Diesel an imperfect Carnot and from analysis may be neglected
in comparison with the Brayton for power generation.
This leaves as the cycle worthy of application but little known
and not at all recognized, Brayton's with its variations.
Special problems introduced by the internal-combustion method
of heating, (a) What specific heat is to be used in calculating
rise of temperature during a chemical change, that of the constitu-
ents, that of the products, or something different from both, (b)
Volume change due to chemical action, (c) Is the heat of com-
bustion of a mass of fuel constant or does this depend on conditions,
and if so determine them.
Heat suppression in combustion as evidenced by the discrepancy
i) i)
between observed and and theoretic values of the same for
A "i
the isometric and isopiestic combustion respectively.
Effective specific heat versus effective calorific power.
A variation of calorific value with the method of combustion
would give Otto or Brayton the preference in efficiency as the case
might be.
The non-explosive, internal combustion engine has three ele-
ments, fuel and air supply, fire-box, expansion parts.
Fuel and air supply require no study as pumps and compressors
are well-known machines.
Utilization of hot expanding gases in cylinders and turbines has
been done and requires only enough study to reduce to good
practice ; there is nothing of the impossible.
The combustion phase is where the trouble has been, no entirely
successful fire-box has yet been proposed that will meet all require-
ments though some there are that work very well under specified
conditions.
The engines of this class would have great elasticity of action
in speed, power and direction of motion, they would be able to pull
up to an overload and they can be constructed to burn any fuel.
Coal-burning engines build and proposed, typical cuts.
Oil-burning engines built and proposed, typical cuts.
Gas-burning engines built and proposed, typical cuts.
Details of Construction Compared. Cylinders, igniters, gover-
nors, preheaters, regenerators, fuel feeds, gas burners, oil burners,
mixers, proportioners, use of water, position of fire.
RESUME. 7
All cycles possible with the non-explosive internal combustion
engine.
The engine built by George Bray ton, its abandonment and
eclipse by the Otto machine.
Only non-explosive internal-combustion engine working to-day
is that of Diesel.
The Diesel engine in practice working not under the modified
Carnot cycle but under the Brayton cycle. It is then rather a
modified Brayton.
The cause of failure in other attempts at application of Brayton
or modified Brayton cycles invariably traceable to the fire-box or
method of combustion, therefore investigation of methods of burn-
ing oil and gas necessary.
Methods of gas combustion classified.
Combustion of gases and mixtures requiring an atmosphere and
producing a volume of flame.
Combustion of explosive mixtures by self propagation : (a) at
rest and (b) in motion.
Requirements of proper method for the combustion of explosive
mixtures in motion.
Experiments made in search for means to fulfill the require-
ments.
New method of combustion of explosive mixtures in motion a
close approach to ideal.
The explosive gas fire.
Operation of the internal combustion engine by intermittent com-
bustion in which the mixture leaves after passing the inlet valve to
expansion space.
Operation of engines with continuous combustion, the fire burn-
ing steadily and the inlet valves controlling the burnt hot gases.
PART III.
LIQUID FUEL COMBUSTION.
Oil combustion, a series of physical actions involving a knowl-
edge of gas combustion and to be studied in the light of that
knowledge.
Different oil systems differ, (a) in the method of producing the
vapor or oil -gas, and (ti) in the methods of causing a meeting of
this vapor with the air.
Oil combustion classified.
8 THE HEAT ENGINE PROBLEM.
Historical review of different classes by studying characteristic
systems.
Requirements for enclosed pressure system to be used in the
internal-combustion engine.
Report of series of experiments having for their aim the dis-
covery of a suitable system.
The " explosive oil fire," as developed, proves satisfactory and
suitable.
Some experiments and notes to test the availability of the " ex-
plosive oil fire " for other uses.
PART IV.
PHYSICAL PROPERTIES OF EXPLOSIVE MIXTURES.
Historical sketch reviewing present knowledge of the properties
of explosive mixtures.
No data sufficient for computation of many of the quantities
needed in the application of explosive mixtures to engineering work.
Object of this part not only to discover if possible a properly
simple and accurate means for obtaining such data, but also to use
the apparatus in the making of such observations as time might
permit.
Apparatus designed and used for
fg as >
i Measuring^ air,
t neutral products of combustion.
2 Mixing, compressing and storing the mixture.
3 Producing products of combustion by method available for
collection and storage.
4 Measuring the heat of combustion of explosive mixtures
directly by burning at constant volume.
5 The same as 4 but by burning at constant pressure.
6 Measuring pressures due to constant volume combustion
directly.
7 Measuring volume's increases due to constant pressure com-
bustion directly.
PART V.
CONCLUSION.
Review of work done, results attained, and statement of what
remains to be done.
A METHOD OF CYCLIC ANALYSIS OF HEAT
ENGINES.
HEAT ENGINE CYCLES ANALYZED.
Prime movers are useful when they produce motion in required
directions against resistances. Nearly all our machines which in
general constitute the resistance to prime movers are designed to
be operated through an applied forceful rotary motion ; therefore
the prune movers that are to be of most service to us in our ordi-
nary working operations must develop forced rotary motion. By
far th'e largest number of these rotary motion prime movers come
under the head of Heat Engines. These heat engines may be
divided into two classes :
(a) Those that do work by utilizing the expansion of a sub-
stance when changing from the liquid to the gaseous state.
(ft) Those that do work by utilizing the expansion of a perfect
gas, this expansion being caused in some mysterious way by ab-
sorption of heat.
The engines of class a usually consist of two parts, a part for
the production of the vapor, and a part for the utilization of this
vapor, converting an increase of volume into a forced rotary motion,
in ordinary language, of a boiler and an engine proper. The
amount of work that can be done with a given amount of heat by
a prime mover of this class, is definitely known within certain
limits, when we know how much liquid can be converted into
vapor by this heat, and the relative specific volumes of the liquid
and resulting vapor. It therefore depends chiefly on the liquid
chosen, and, of course, on the mechanical efficiency of the system
as a converter, or, as we may say, on the design of the machine.
The cycle of operations is : (I.) Add heat to liquid and pro-
duce vapor. (II.) Allow vapor to expand to as low a pressure as
possible, and then discharge it either as vapor or as a reconverted
liquid.
This cycle is unchangeable except in incidental details. On the
contrary, however, when we employ a perfect gas to which to
apply our heat, and whose expansion gives us our work, we may
have a large range of different cycles or series of operations that
9
10 THE HEAT ENGINE PROBLEM.
may be performed on or by the gas in question. The amount of
work done by our expanding gas due to the initial application of a
given amount of heat will depend on the manner of heating, method
of expansion, ultimate disposition of the gas, and, of course, on the
mechanical efficiency of the machine for performing the operations
desired, and will depend not at all on the gas chosen. In short,
the varying amounts of work that may be done will depend solely
on the cycle itself. It is therefore evident that there is consider-
able importance in knowing just how the cycle can effect this
change of ultimate useful work for given heat supplied.
In the actual application of any cyclic principles we find various
other questions beside the ultimate useful work that demand at-
tention and study. For example, one cycle requires a larger vol-
ume of gas to do same work as another ; a larger engine is therefore
necessary ; some cycles operate under higher temperatures than
others ; some through wider ranges of temperature and pressure.
Many other questions might be cited, but enough are given to show
that it is necessary that we study the cycles as such, and obtain a
statement of every question in terms of the cycle, before we begin
the consideration of the mechanical difficulties involved in its car-
rying out.
It is possible to cause a similar mass of perfect gas to pass
through each of the cycles, and obtain an equation for every vari-
able entering into the cycle in terms of the initial conditions and
the quantity of heat supplied. For example, we can write
For cycle I. Efficiency = =/ I (H l C')
For cycle II. E=f n (H l C lt )
For cycle III. E=f m (H'"}
For cycle;., E=f n (H,C") '
where H^ is heat supplied, and C a constant.
We thus get a series of curves of efficiency, one for each cycle,
in terms of the same variable, and get exact relations of the cycles
regarding efficiency at a glance. Instead of efficiency we might
have chosen the final volumes or the maximum temperatures.
It is the object of this part to consider the various cycles as
above outlined and cause one pound of air to pass through each
of the cycles under ideal conditions, and to. determine every cyclic
variable in terms of H v and arbitrary initial conditions. To pass
from ideal conditions to practical ones we need only apply a cor-
rection factor.
CYCLIC ANALYSIS OF HEAT ENGINES.
1 1
In what follows we shall not consider how the heat is applied
or abstracted, the mechanisms involved, nor the practicability of
the process.
The following cycles will be considered :
CYCLE. I e
B
FIG. i.
FIG. 2.
Let Fig. i be a P.V. and Tig. 2 be a O(p diagram for the cycle.
Then we have :
From B to C. Addition of heat isometric'ally from atmospheric
pressure.
From C to D. Adiabatic expansion to atmospheric pressure.
From D to B. Cooling at atmospheric pressure.
P
CYCLE I -A.
e.
B
E.
FIG. 3. FIG. 4.
We have :
From B to C. Addition of heat isometrically from atmospheric
pressure.
From C and D. Adiabatic expansion to point above atmos-
pheric pressure.
12
THE HEAT ENGINE PROBLEM.
From D and E. Cooling isometrically to atmospheric pressure.
From E to B. Cooling at atmospheric pressure.
CYCLE IB.
P>
FIG. 5. FIG. 6.
We have :
From B to C. Addition of heat isometrically from atmospheric
pressure.
From C to D. Adiabatic expansion to below atmospheric pres-
sure.
From D to E. Cooling isothermally to atmospheric pressure.
From E to B. Cooling at atmospheric pressure.
CYCLE 1C
C. '
B
I).
V
FIG. 7. FIG. 8.
We have :
From B to C. Addition of heat isothermally from atmospheric
pressure.
From C to D. Adiabatic expansion to pressure below atmos-
phere such that we get,
CYCLIC ANALYSIS OF HEAT ENGINES.
From D to B. Cooling isothermal ly to original volume and at
mospheric pressure.
p
CYCLE
A.
FIG. 9. FIG. 10.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C. Addition of heat isometrically.
From C to D. Adiabatic expansion to atmospheric pressure.
From D to A. Cooling at atmospheric pressure.
CYCLE HA. @
V
FIG. ii. FIG. 12.
We have :
From A to B. Adiabatic compression from atmospheric pres-
sure.
From B to C. Addition of heat isometrically.
From C to D. Adiabatic expansion to pressure above atmos-
phere.
From D to E. Cooling isometrically to atmosphere.
From E to A, Cooling at atmospheric pressure.
H
P
THE HEAT ENGINE PROBLEM.
CYCLE JPB. e
B
A. E.
C.
E.
D
V
FIG. 13.
FIG. 14.
We have :
From A to B. Adiabatic compression from atmospheric pres-
sure.
From B to C. Addition of heat isometrically.
From C to D. Adiabatic expansion to pressure below atmos-
phere.
From D to E. Cooling isothermally to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
CYCLE 1IC
B
A.
D.
FIG. 15. FIG. 1 6.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C. Addition of heat isometrically.
From C to D. Adiabatic expansion to pressure below atmos-
phere such that we get,
From D to A. Cooling isothermally to original volume and -at-
mospheric pressure.
CYCLIC ANALYSIS OF HEAT ENGINES.
CYCLE 111 @
FIG. 17.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C. Addition of heat isopiestically.
From C to D. Adiabatic expansion to atmospheric pressure.
From D to A. Cooling at atmospheric pressure.
CYCLE JIIA.
FIG. 19.
FIG. 20.
We have :
From A to B. Adiabatic compression from atmospheric pres-
sure.
From B to C. Addition of heat isopiestically.
From C to D. Adiabatic expansion to pressure above atmos-
phere.
From D to E. Cooling isometrically to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
i6
THE HEAT ENGINE PROBLEM.
CYCLE 1HB
' A
FIG. 21.
FIG. 22.
We have :
From A to B. Adiobatic compression from atmospheric pressure.
From B to C. Addition of heat isopiestically.
From C to D. Adiabatic expansion to pressure below atmos-
phere.
From D to E. Cooling isothermally to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
CYCLE me.
A.
V
I),
FIG. 24.
We have :
From A to B. Adiabatic compression from atmospheric pres-
sure.
From B to C. Addition of heat isopiestically.
From C to D. Adiabatic expansion to pressure below atmos-
phere such that we get,
From D to A. Cooling isothermally to original volume and at-
mospheric pressure.
CYCLIC ANALYSIS OF HEAT ENGINES.
P B. 8.
B
FIG. 25. FIG. 26.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C. Addition of heat isothermally.
From C to D. Adiabatic expansion to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
P IB. CYCLE IV A. 6-
B
E.
FIG. 27. FIG. 28.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C. Addition of heat isothermally.
From C to D. Adiabatic expansion to pressure above atmos-
phere.
From D to E. Cooling isometrically to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
THE HEAT ENGINE PROBLEM.
CYCLE IVB
B
C.
FIG. 29. FIG. 30.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C, Addition of heat isothermally.
From C to D. Adiabatic expansion to pressure below atmosphere.
From D to E. Cooling isothermally to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
P B CYCLE IVQ
B
A.
D,
FIG. 31.
FIG. 32.
We have :
From A to B. Adiabatic compression from atmospheric pressure.
From B to C. Addition of heat isothermally.
From C to D. Adiabatic expansion to pressure below atmos-
phere such that we get,
From D to A. Cooling isothermally to original volume and
atmospheric pressure.
CYCLIC ANALYSIS OF HEAT ENGINES.
CYCLE V.
FIG. 33.
FIG. 34.
We have :
From A to B. Adiabatic compression from atmospheric pres-
sure.
From B to C. Addition of heat at variable pvd.
From C to D. Adiabatic expansion to atmospheric pressure.
From D to A. Cooling at atmospheric pressure.
Cycles V., A, B and 7 may have the same modification on Cycle
V. as II. A, B and C have on III., for example.
THE ATMOSPHERIC OR VACUUM CYCLES.
Here all the cyclic operations take place at or below atmos-
pheric pressure.
CYCLE W.
P
0.
A
B.
B,
V
FIG. 35.
FIG. 36.
We have :
From A to B. Addition of heat at atmospheric pressure.
From B to C. Cooling isometrically.
From C to A. Adiabatic compression.
20
THE HEAT ENGINE PROBLEM.
CYCLE W.
FIG. 37. FIG. 38.
We have :
From A to B. Addition of heat at atmospheric pressure.
From B to C. Adiabatic expansion.
From C to D. Cooling isopiestically.
From D to A. Adiabatic compression.
CYCLE VIII.
FIG. 39.
FIG. 40.
We have :
From A to B. Addition of heat at atmospheric pressure.
From B to C. Adiabatic compression to such a pressure, that
we get,
From C to D. Isothermal compression to original state.
CYCLE IX.
V
FIG. 42.
FIG. 41.
We have :
From A to B. Addition of heat at atmospheric pressure.
From B to C. Adiabatic expansion.
From C to D. Cooling isometrically.
From D to A. Compression adiabatically.
CYCLIC ANALYSIS OF HEAT ENGINES. 21
We might have many modifications of these but as a discussion
of the type throws sufficient light on the variations considering
the importance of the cycles, these modifications will not be dis-
cussed.
CYCLE I.
Fig. i. Fig. 2.
Let HI be the heat added from B to C.
Let C v be the specific heat of gas at constant volume, and here
assumed constant for simplification. It is probably a variable, but
so assuming it gives unmanageable formulae. A correction may
afterward be applied, if desired. C v = heat to raise one pound
gas i F. at constant volume.
Let v b be the volume of the gases at point B of the diagram,
i. e. t before heating and expressed in cubic feet.
Let p b be the corresponding pressure in pounds per square foot.
Let T b be the corresponding temperature in absolute degrees
Fahrenheit.
Then will the increase in temperature be given by'
T _ H \
<~ b ~^ v
or
Since volume is constant from B to C,
whence
A-Af-
y />
From (i)
'-'*i-3L
?;~ h c-X
Since this quantity
C.T;
will enter into many of our equations, let us denote it by
whence
22 THE HEAT ENGINE PROBLEM.
P.-PJ- (2)
The adiabatic relation
p/>if
gives
But p d p b by hypothesis, hence
-^)^ (3)
Another adiabatic relation gives
whence
remembering p d = p b and substituting the value of T e
Let H 2 be the heat discharged. Then
Where C p specific heat at constant pressure and assumed con-
stant. Hence substituting
- 1). (s)
The work done in heat units will be
W-H^-H, i (6)
~H l -C f TJ < X^-iJ. (7)
And in foot pounds
This work of expansion could have been obtained by tempera-
tures and by integration as well.
We have
CYCLIC ANALYSIS OF HEAT ENGINES. 23
But
..IT- CT f -CT b -C,T d+
We know also
and
in heat units. This second term is the area of the rectangle be-
tween { * = and \ V == V * and lying below atmos-
( / = atmosphere ( v = v d
phere is not available for work.
By integration W j^ pdv = area between expansion curve
and axis of volumes. The expansion is adiabatic.
i r i r
// ^ ^
. yy = tf _ -_C
i -r
Since
c, c,
= JC v (T c - T d ) in foot pounds.
Subtracting the rectangle p b (v d v b ) we get
24 THE HEAT ENGINE PROBLEM.
in foot pounds, or in heat units
as before.
Before going farther let us apply a test to each of the states B,
C, D from the law of perfect gases
* b b _ bb __ r,
= " T b -
T l ' T
r,
hence these are identities, as they should be.
Denote the volume swept through or volume range by R v .
Then will
i_
R* = v d -v b = v d -v c = v b \Xi - i] (8)
Whence mean effective pressure
Efficiency
The entropy range is given by
Mean effective temperature
The temperature range
The pressure range
CYCLIC ANALYSIS OF HEAT ENGINES. 25
Whence we may write a mean effective volume
MEV 7
- - J
- i). (16)
These results are here tabulated for reference and comparison
with what follows :
We might take the formulae* derived for mean effective tempera-
ture, but as these were the results of a comparison of cycles, none
of which -ran below atmospheric pressure, it would be better to
take another standard here. Let us take arbitrarily as the mean
effective temperature one half the sum of the mean temperature
of heat addition and the mean temperature of heat abstraction.
CYCLE I. X= i + - L
Formula Reduced to Initial
Symbol. Formula as Derived. Conditions.
' 6 Arbitrary p h
B
*T* x b b * b b
b " ~R ' ~R~
T
p p.-^ p.X
t e to i * b
* b
T
D\
Pa Pb
a
T < MA' r " Xy
H, C f (T t -T b ) CT^- i)
W J(H,-H^ .J{ff, - C p
* School of Mines Quarterly, XXI., 4, 1900.
26 THE HEAT ENGINE PROBLEM.
Formula Reduced to Initial
Symbol. Formula as Derived. Conditions.
E
'-'
*. *-'* ^ Y -I)
M.E.P.. ./
*, A -A A(^-
FT
M.E.V.
R * k&r ^ lo g^
i
7;
r ................ T c -T b .............. T b (X-i).
CYCLE I. A.
FIG. 3. FIG. 4.
We have as in Cycle I. for point C.
v c = v b (i)
(3)
Assume
Then from the adiabatic relation
v
or
Also
" v *> (s)
CYCLIC ANALYSIS OF HEAT ENGINES. 2/
Substituting, values of p c and T c in (4) and (5) we get
**(0' (6)
T <- T *\-&)* - T ^\jr- (7)
If we write
j-"'
A
then JL_
7 (8)
(9)
ff
l_
v e = v d = v b (Xn)y, (10)
A -A.
. (u)
6
Let us apply the perfect gas law to the points B, C, D and E.
A? h r>
-r = A,
/^c c fb b r>
^r~ = ~T~y = R >
J. J. i^A.
c o
l r,
T
*
Heat is abstracted in two parts, the first at constant volume
from D to E and the second at constant atmospheric pressure
from E to B.
28 THE HEAT ENGINE PROBLEM.
Hence
The work done in foot pounds is
(13)
(4)
('5)
The mean effective pressure
W
M.E.P. = --. (16)
But
,M.E.P,/|- -j- K(i8)
r
C 1 / i \ 11
I JLJ T (Xn\i I - I I -4- ^* 7" ff^Y??^ il I
II b \ n j P b j.
(2(
As before, the entropy range is
R^=C v \ogX. (21)
CYCLIC ANALYSIS OF HEAT ENGINES.
29
Taking the mean effective temperature as the mean of the aver-
age heating temperature and the average cooling temperature,
C.\og.X
The temperature range is
The pressure range is
Whence
(23)
(24)
(25)
/
'.)7 (I-:)-
' \n 1
(26)
Tabulating these results we get for
CYCLE I. A.
Symbol. Formula as first derived. Formula reduced to initial condition.
p b ---- Arbitrary ---- . ................. p b
B<
C
T
A
R
R
THE HEAT ENGINE PROBLEM.
Symbol. Formula as first derived. Formula reduced to initial condition.
i
D\
E\
A
v.
E
R .
M - KP R-
H - 2
* T
,- -r 1 ~
ff,..C.(T i -T)+ C(T-T t )~C.T t (Xn)-> \--i\-
MET
A -A
)7g- i) -C f T&Xnji- i]
CYCLIC ANALYSIS OF HEAT ENGINES. 31
CYCLE I. B.
FIG. 5. FIG. 6.
As the operations up to the point C, i. e., after addition of heat
are the same as in cycle I., we may assume these results :
V e= V l (0
A = A* (2)
T.= T b X. (3)
Choose p d so that
A>A>o- (4)
Expansion CD gives
Also
From the isothermal relation along DE,
A = A b y hypothesis (8)
or
Applying the perfect gas law to the various points
T T,X ~ ' T,
cb b
32 THE HEAT ENGINE PROBLEM.
i) v b b n i)^v
fe e ___ fb b __ r>
-jT = - -~- - ^-.
' T (XnT b
^ n
Heat is abstracted in two parts, first a part isothermally and
second a part at atmospheric pressure. The part abstracted
isothermally is extremely difficult to calculate without the aid of
the 60 diagram and its relations.
The entropy range along BC has been found to be
Now it is evidently the same so far as entropy range is con-
cerned whether we cool at constant pressure from E to B or heat
isopiestically from B to E, thus
T f
Hence the entropy range for the isothermal operation will be
given by
(.3)
This latter isothermal change taking place at temperature T e =
T d the heat of cooling will be given by
Hence the total heat abstracted is
= C p (T c - T b ) + T, \ C, \og c X- C p log 5] (15)
- * 'h J
But
_
C f l S, "~ = C r log e X V + C f log, i
= C,\ 0ge X+(C,-C p )l g.
,,.
CYCLIC ANALYSIS OF HEAT ENGINES.
33
Since
And
Hence
The work in foot-pounds is
-*. (16)
(17)
-.= I -
(18)
M.E.P. =/
.-. M.E.V. =/]
(20)
(21)
(22)
The mean effective temperature being the mean of the heating
and cooling means is given by
M.E.T. =
where R^ is same as in previous cycle.
34 THE HEAT ENGINE PROBLEM.
log
(23)
Tabulating :
*
CYCLE I. B.
Symbol. Formula as first derived. Formula reduced to initial conditions.
p b Atmosphere Atmospheric p b
v b Arbitrary v b
c
" R ' R
A A? A*
T t (Xn)
A
T,
CYCLIC ANALYSIS OF HEAT ENGINES.
3S
C,(T.-T t ) + 7-Jc.log.A-- C,log.
w.
I
- .) .............. ^[(-
M ' E - P
-A
- A
M.E.V.
log, 5
T,
M.E.T.
36 THE HEAT ENGINE PROBLEM.
CYCLE I. C.
FIG. 7. FIG. 8.
Assume all results to point C from Cycle I.
A -A* (0
"< = ^ (2)
T.~T,X. (3)
From the adiabatic CD.
This adiabatic must meet the isothermal from B in point Z>,
hence
'I ;' ^-^^- - ; - (5)
Equate (4) and (5)
(6)
This is the pressure at which the isothermal through B will
meet the adiabatic through C. Its corresponding volume is
*< = , - = ^^ (7)
T*-1v (8)
The heat abstracted by the isothermal cooling is found as before
from 60 relation.
= c\o Tc =C\o X.
6 * i
CYCLIC ANALYSIS OF HEAT ENGINES. 37
Hence
log e X. (n)
The work done in foot-pounds is
W=J(H, = /Q =/(//, - 7;C.log.Jr> (12)
The efficiency is
The volume range is
i
Hence
The entropy range was found R^ = C v log c X hence
M-[- -p< / *** i" * V
.E.T. =
The pressure range is
Xr*
And
Tabulate
CYCLE I. C.
Formula reduced to initial
Symbol. Formula as first derived. conditions.
p b Atmospheric Atmospheric/^
v b Arbitrary v b
Tb J? ~'R
THE HEAT ENGINE PROBLEM.
P'ormula reduced to initial
Sym
c-
D
bol.
Formula as first derived. conditions.
T b
f c
"V
f b y f b
c
. .V . . V
T
T 1 I T _i_ 1 l TV
4)
*V h C,TJ-
I A j A
Pd' ' '
1)
1 1
T,.
b b
....Wog.y' V.log^
E.
R
H
M.E.P ............ ....- ...... ...... /
M.E.V.
>. -
M.E.T.
.'. T- T I: : .'.... .T,(X-
CYCLIC ANALYSIS OF HEAT ENGINES.
39
CYCLE II.*
Let Fig. 9 be the P. V. and Fig. 10 the 0
(12)
(H)
(i5)
CYCLE II.
Symbol. Formula as First Derived. Formula Reduced.
. .............. -' fr arbitraf y)
T M
A AV
42 THE HEAT ENGINE PROBLEM.
Symbol. Formula as First Derived. Formula Reduced.
Pi Pa A
H, -C r (T d -T a ) <
w. ff t -ff t HI- > 7 i< jrv -
w H.-CTJ.XI-I)
M - E - P JR /-^-tr
M - KT
.log. .............. C,\ogX
2 b
-0
R. .......... ..-.... -A -A ......... .-.- -p a (r x -
7;- r
CYCLIC ANALYSIS OF HEAT ENGINES.
43
CYCLE II. A.
Let Fig. ii be the P. V. and Fig. 12 the 6(P diagram for the
cycle.
p ,C- ' CYCLE 11A.
V
FIG. n.
FIG. 12.
Then we have, since the compression is as in Cycle II.,
-
A=Ar v .
Also for C the heat addition being as before
(0
(2)
(3)
(4)
ill . (s)
~ 1 X. (6)
The point D lies arbitrarily between C and the atmospheric line
on the adiabatic
(V
^r
-. (7)
From this point we will consider two. cases: i, the general
where v d is greater than v a , and 2, a particular case where
44
THE HEAT ENGINE PROBLEM.
Then we have
Y-l
v V*- 1
P.=Pa
(8')
(9)
T f TV
j == * *'*
(9')
(10)
V .' = ""a
(10')
(n)
A'=A
(ii')
'A
*>
(12)
(12')
Apply the perfect gas law
AT'
R,
R.
<-R,
p'v' v p
* < __ a^o- . J
CYCLIC ANALYSIS OF HEAT ENGINES. 45
Heat is abstracted as follows :
The work is given by
(H)
-V- 1?-]-
Volume range is
j
R, = v d -v b = v d -^ (is)
w w
M.E.P. =/__=/--- (I6)
(13')
C.TJX- i].
M.E.P.=/
Entropy range is the same for both cases,
4 6
THE HEAT ENGINE PROBLEM.
Mean of mean temperatures of heat addition and abstraction,
M.E.T.'
MET -
-'
2 C,\Og,X
(18)
(i 8')
Pressure range is same for both cases
Mean effective volume
W
(19)
^ (20)
Temperature range is also the same for both
w
(21)
CYCLE II. B.
Let Fig. 13 and Fig. 14 be the P.V. and 0 diagram respec-
tively of the cycle.
CYCLE ITB.
B
c.
E.
FIG. 13.
FIG. 14.
CYCLIC ANALYSIS OF HEAT ENGINES. 47.
Assume same results as before up to the point c. Take p d
something less than atmosphere, i. e.,
0- (0
Then
And
Through D and a point E whose volume is greater than the orig-
inal we have an isothermal
(4)
* '"'/."'M A r/,
Hence
A = A- (6)
Apply the perfect gas law to the points
p v
a n f?
i i
A
T
AA_ o
r " ~
48 THE HEAT ENGINE PROBLEM.
During the isothermal compression heat must be abstracted, the
amount can best be calculated by Oy coordinates. Call this
amount w, then,
m =
But
f!t- f.= Cfe- f) ~ (f. -
and
"
Besides this amount m we must abstract a quantity ^(7^ 7 a )
isopiestically, whence
, = CJtT. - T.) + T, j C. log,X+ C r log. \X^ ( g)*T ] J
(7)
-//,-// (8)
- l ~W t - (9)
The volume range is
(10)
(II)
CYCLIC ANALYSIS OF HEAT ENGINES.
w
49
(15)
CYCLE II. C.
i
Let Fig. 15 be the P.V. and Fig. 16 the 60 diagrams of the
cycle.
CYCLE JIG
FIG. 15.
FIG. 16.
All values for the compression and heat addition found in Cycle
II. may here be assumed. The point D lies at the intersection of
two curves, one an adiabatic through C, the other an isothermal
through A and the relations can be written. From the adiabatic
relation
From the isothermal relation
Equating we get
50 THE HEAT ENGINE PROBLEM.
Pa
This is the pressure at which the intersection will take place.
By substitution we get
^-f.AFi -." " (2)
; T d -T, . '- \ (3)
Applying the perfect gas law to D
-K.
All the heat is abstracted at constant temperature during the
compression D to A. The entropy range is evidently the same as
for heat addition and this is
R d P d >Pa (0
and
V *> V a- ( 2 )
This latter (2) will not necessarily follow from (i) but where it
does not hold the cycle is decidedly imperfect and this case is here
neglected, i. e., the case where the isometric DE cuts the adiabatic
AB.
We have then
P.- P. (5)
Apply the perfect gas law to D arid E.
AA
r
This verifies the formulae.
Heat is abstracted in two parts and the amount is
(8)
CYCLIC ANALYSIS OF HEAT ENGINES.
57
-'}
The work done is
and efficiency
- H
H n
W
W
:.E.P. = /^- =
Rtp C p \og e Fas before for III.
R P = P- Pa = P(r y - i) as in III.
M F V / /
- J R r - J p.(r-ir
As before III. the temperature range
CYCLE III. B.
Figs. 21 and 22 are its diagrams
P CYCLE MR a
B
A. E.
(10)
(II)
(12)
(13)
(14)
(15)
(16)
(18)
FIG. 21
FIG. 22.
58 THE HEAT ENGINE PROBLEM.
All results of III. A up to period D may be assumed except
that p d which was there arbitrary was assumed greater than p a is
here less than p a , i. e.,
p c >p d >o. (i)
We had
and
/AiY ~~ l . (3)
' \P.
Through E and D there must pass an isothermal and
l 'e V d p " C 'd p ~ V a \ p ] p
re -fa -f d *a
. . (;)
Applying the perfect gas law to E
Heat abstracted I isothermally a quantity ?.
2 isopiestically " n.
n=C p (T e -T a )
But
(4)
(5)
CYCLIC ANALYSIS OF HEAT ENGINES.
59
= H, - H v
(10)
M.E.R=/^=y
w
= y />-/;
r = T a (-{*- 1 Y- i) as before III.
CYCLE III. C.
Let Figs. 23 and 24 be its diagrams.
CYCLE inc.
A.
FIG. 23-
V
(13)
C4)
(15)
(17)
G
D.
FIG. 24.
6o THE HEAT ENGINE PROBLEM.
We will assume all results to C already derived. The point D
is determined by the intersection of the adiabatic through C with
the isothermal through A. From the adiabatic relation
From the isothermal relation
y y -i
By substitution
r i Pi
- : , .", -*' (0
: : - (2)
" ; (3)
log, Y (4)
(6)
FT-'
(8)
CYCLIC ANALYSIS OF HEAT ENGINES. 6 1
M.E.T. = i (3+3) = * (3r f^r\
* \ Ry J 2 \ (7 log F /
(10)
+
= 7:-r = 7;[ r v-iF-i]asinIIL (n)
CYCLE III. C.
Symbol.
Formula as First Derived. Formula Reduced.
ft 1 _?Ll v, r y
\V 1
V a V a
b
T .
f r
.7^ f5A ; Trv- 1
P
a \ ?; / ft
re
-* ft / rt/
^ v F
c *
T . .
6 ^> r
r (i + 1 \ . T r^- 1 F
*\ i T t l
y_
i) T^V !
v .
' 'fa 1
d
T } .
d
T
E
e 'p P fc>e
2 \ ^?t: / ^ y IQCT F a /
62
# r
M.E.P..
R ,
M.E.V.
THE HEAT ENGINE PROBLEM.
..*,-*.. ... (Yri-L\
a \ r ;
H.-T^rio
J-i
R
J
A-A p\r-
y y -i
T -T
CYCLE IV.
Figs. 25 and 26 are its diagrams.
p B. CYCLE IV
B
FIG. 25.
FIG. 26.
We may assume the results already obtained for the compres-
sion but beyond that new conditions arise. By isothermal heating
the curve approaches the atmospheric line and there will be a cer-
tain quantity of heat that will bring the isothermal down to the
atmospheric line leaving a subsequent adiabatic expansion an im-
possibility. This quantity of course depends on the location of B,
i. e., the amount of previous compression. The higher the pre-
CYCLIC ANALYSIS OF HEAT ENGINES. 63
vious compression the more heat may we add isothermally before
reaching atmospheric pressure.
The quantity of heat which will make adiabatic expansion im-
possible and stop the isothermal on the atmospheric line can best
be determined from 60 relations. Denote this quantity by Q.
FIG. 45.
On the 80 diagram Fig. 45 the point 3 lies at the intersection
of the isothermal 23 drawn at temperature 02 the compression
temperature and the isopiestic 13 drawn from atmospheric tem-
perature Oi to the intersection 3. In each case the entropy range
C
.-. <2 =
Apply now to the Cycle IV.
* i
T,
iog eT y l . (0
This is the amount of heat that will bring C down to atmosphere
with no adiabatic expansion. In order that the cycle may exist
according to the hypothetical definition we must add less heat
than this quantity Q. Hence we have the equation of condition
for the existence of the cycle
(2)
or
46 THE HEAT ENGINE PROBLEM.
A similar method can be used to find the amount of expansion
or resulting pressure and volume after addition of H iy BTU of
heat.
Draw on both diagrams the isopiestic through the termination
C of the isothermal and cutting the adiabatic AB at point c'.
Then
T T
Y' t = C P lo S' T,
T r y ~ l
But
And the amount of heat necessary for this isothermal expansion
from B to C
But
r - i - ' C - C
c
and
CYCLIC ANALYSIS OF HEAT ENGINES. 65
Put
n,
Then will
That is to say if we start at state B and add a quantity of heat
H v isothermally the resulting pressure is
A P T y
L r b __ * at ., v
C x^5 /? \ 3 )
Since
Now
- Q
117 1 1 TTT
We had III.,
Hence
Whence
(4)
(5)
(6)
(7)
66 THE HEAT ENGINE PROBLEM.
Similarly
Apply the perfect gas law
T.
R
f-fzr-R.
Verifying the formulae,
E
l
H
-' +
M.E.P. -/-/- -^
(8)
(9)
(10)
(12)
ap^J ; 03)
(H)
(15)
CYCLIC ANALYSIS OF HEAT ENGINES. 67
MEV T W jH.-C,W*-*) , }
- J R- J p a (r-i}
-i). (18)
CYCLE IV.
Symbol. Formula as Derived. Formula Reduced.
A - M)* W
r
T,, T^ Tjr
T
Equation of condition. . . H^ < T b \og e ^ . . . H l < Tj*~ l
A Pj y
T, T,
p p p
id fa fa
7, :
H, C p (T,-T a ) C p T.(e r - 1 - i)
i-g i - C,T.(<'- 1 ~
r> "1
^* 7; r- l f
68
THE HEAT ENGINE PROBLEM.
M.E.T r-vl-inf
V
J r-p j j x-- /y~i / Y 1 \ -,
"T^l" JL ^~
(l V^
W r//
M - E - P JR, J \- v{e ^_
r
w
M.E.V.. .../
CYCLE IV. A.
Figs. 27 and 28 are its diagrams.
CYCLE IVA, 8.
B
FIG. 27.
E.
FIG. 28.
We may assume the results of IV. up to point C. The point
D lies somewhere on the adiabatic between C and atmosphere and
is subject to the conditions
P.>P>P. (0
CYCLIC ANALYSIS OF HEAT ENGINES. 69
*.>* (2)
Then
r ~': ' . (3)
Similarly
(4)
(5)
Apply the perfect gas law to D and E.
T ' JL
The heat abstracted is
(7)
THE HEAT ENGINE PROBLEM.
M.E.P.=/
W
(8)
(9)
(10)
M.E.V. = /
w
CYCLE IV. B.
Let Figs. 29 and 30 be its diagrams.
B. CYCLE IVB
B
FIG. 29.
FIG. 30.
(12)
(H)
c
D.
FIG. 29. FIG. 30.
The operations up to C are as in IV. and we may assume those
results.
CYCLIC ANALYSIS OF HEAT ENGINES. /I
The point D is subject to the condition
Pa
V a - (2)
Then
V = V I " I V *'~~ 1 f 7 ^
a \ P j ^ '
and
^-^"r-i /\
^ (4)
Following the methods already adopted we can write
ffi-Tfa-ri+Ctf-Tj.
But
()'
'
r r
T; 711-
(y
Equation of condition H l > o
A
W h
E.. .1
b p c ' r
T.. .71. .7>>
A
.z;^
r
CYCLIC ANALYSIS OF HEAT ENGINES.
75
1
M.E.T * '^"^) f (l +
f
M.E.P J~ JHA
- - , ||t' I
-# . . ^ A . -P \r y F I
p -fd ft -fa \' g% I
M.E.V /-^r
-p
^r T.-T,
CYCLE V.
Let Figs. 33 and 34 be the diagrams of the cycles.
CYCLE f.
P
FIG. 33.
FIG. 34.
If we add heat at increasing /, v, and T the curves of states will
lie somewhere between the isometric and isopiestic on both dia-
grams and the cycle is somewhere between III. and II. If the
7 6
THE HEAT ENGINE PROBLEM.
heat addition took place at decreasing p, increasing v and y the
curve of states would lie between the isopiestic and the isothermal
and the cycle lie between III. and IV. We cannot, however, cal-
culate the appropriate set of formulae without knowing the law of
variation of states. The number of ways of variation is infinite,
and while any one might be assumed, nothing could be gained by
the calculation unless the law of variation chosen were preemi-
nently simple or maintains in practice. Whatever it may be, how-
ever, the previous discussion will enable us to class it pretty well
without entering much into details.
CYCLE VI.
Let Figs. 35 and 36 be the diagrams of the cycle.
CYCLE VI.
0.
P
V
FIG. 35.
Heat being added isopiestically
71-
FIG. 36.
H \
T> / H\
= T " ( l + ~CT
* { ^^ J - .
(0
The point C lies on the adiabatic through A, hence
(3)
(4)
T =
x
(5)
CYCLIC ANALYSIS OF HEAT ENGINES.
The perfect gas law is seen by inspection to apply
- n = c.
77
(6)
(7)
.\RA>o (4)
(5)
(p \v-*
)v- *
But
Hence
^=?- ( 8 >
Similarly
T a T b
And
(10)
C p \og e x as in VI. (13)
CYCLIC ANALYSIS OF HEAT ENGINES.
79
w
W
CYCLE VIII.
Figs. 39 and 40 are its diagrams.
CYCLE VIII.
0.
We have for
FIG. 39.
(15)
(16)
(18)
(19)
FIG. 40.
(0
(3)
The isothermal through A intersects the adiabatic through B to
determine C.
From the adiabatic
From the isothermal
(
But
8o
THE HEAT ENGINE PROBLEM,
v 1*1 v
JL
By substitution
(4)
, = v ,-v=v*( xy - i -
(6)
(7)
(8)
(9)
(10)
CYCLE IX.
Let Figs. 41 and 42 be its diagrams.
CYCLE IX-
p
(12)
V
FIG. 41. FIG. 42.
Up to the point C the results of VII. may be assumed.
CYCLIC ANALYSIS OF HEAT ENGINES. 8 1
The point D lies on an adiabatic through A and is subject to the
conditions
v,= v (i)
d c \ /
(3)
5 (5)
(6)
-H t -H,-CT* ? r i-~ (7)
(8)
^ = 6^ log e x as before (9)
M.E.T. = i(^) (10)
W
:.E.P.=/- -rr ()
^=A-A = A-^ 7 (13)
(^
M.E.V.-/- ~ (14)
L f C
R T =T b -T n = T a (x- i). (15)
82 THE HEAT ENGINE PROBLEM.
CYCLE X.
In this cycle as in the last four heat is added at atmospheric
pressure, then follows adiabatic expansion after which heat is ab-
stracted according to some law as yet undefined. Adiabatic com-
pression completes the cycle. As the law of abstraction of heat
is as yet undefined we cannot, of course, derive formulae for the
cycle and will leave its discussion as we did Cycle V.
We might have derived formulae for the imperfect carrying out
of cycles VI., VII., VIII. and IX. but they are of such slight im-
portance in practice that it did not seem desirable.
Besides the twenty-two cycles considered there may be others
due to the combination or differentiation of these typical ones, but
the object of this paper will be best accomplished by a study of
types, the non-typical or synthetic cycles having been omitted.
The method of study here set forth being of universal application
to all possible cycles will furnish means of reaching a clear under-
standing of any of the unconsidered cycles should need arise.
COMPARISON OF CYCLES.
Of the many cycles considered we will choose for comparison
only those that might be called the perfect cycles because accurately
defined and these are Cycles I., I. C, II., II. A 2 , II. C, III., III. C, IV.,
IV. C. The atmospheric cycles are of comparatively little impor-
tance and will be neglected in the comparison. We will take up
each variable separately and study its value in the different cases
by formula and by calculated examples expressed in curves which
are then the graphical formulae. The curves given are approxi-
mately correct and as the same approximation will probably main-
tain for all the cases the curves will serve as well for comparison as
if absolutely exact. Two cases of each are given, one with com-
pression 2 : i and one with 10 : I (volume ratios). Call the at-
mospheric values p a , v a , T a .
TEMPERATURES AFTER ADDITION OF H^ B. T. U.
Cycle.
I.,I.C T c =TX=T a i + r CO
CYCLIC ANALYSIS OF HEAT ENGINES.
3
ii., ii. A., ii. c
(2)
-T2T2
I2C2.
JSZTQI.
300 400 900 . . (43)
II. A v d = *', (44)
II. C v tl = s^ = *. ( i 4- ^ )^, (45)
in. f.,-.y; (46)
III. C v t = Yri -v.i + -" ^, (47)
IV. i., = t^'-i = fcyVl, (48)
IV. C v d = v a e z ' = v a e (C p- c ^ T . (49)
These curves will admit of considerable discussion, but the curves
of Fig. 5 i show at a glance all we wish to know in general.
HEAT DISCHARGED OR ABSTRACTED.
Cycle.
I- H 2 = C p T a (Xy - i), (50)
/ H^\
>e \ C***''
II. H 2 = CT n (X-i - i), (52)
II- A H 2 = CT a (X- i) = ^, (53)
II- C H 2 =C v T a \og e (i-^), (54)
HI- H^CT a (Y-i}^^ (55)
III. C H 2 = C p T a log e ( i + ^ ), (56)
* ^p 1 * 1
IV // C T (e Y ~ l T "\ C T (e^ rb 1} ( c 7")
1 v J ' 7 2 u ^\^ 1 /' ^o 2 a\ e L J> \5r)
IV. C
CYCLIC ANALYSIS OF HEAT ENGINES.
(58)
Equations (33), (55) and (58) are identical, that is, these three
cycles will discharge the same amount of heat and have the same
efficiency ; moreover this efficiency will be independent of every-
thing but the compression. These three cycles have, further,
a common property not seen by the formula, but from their defi-
nitions each receives and discharges all its heat according to the
same law.
- Cycle II. A receives all heat at constant volume and discharges
all at constant volume.
Cycle III. receives all heat at constant pressure and discharges
all at constant pressure.
Cycle IV. C receives all heat at constant temperature and dis-
charges all at constant temperature.
A consideration of the above would seem to warrant the prop-
osition :
When all the heat is discharged according to the same law
under which it was received then the cycle will have an efficiency
independent of everything but the previous compression and will
be given by
We may remark here that as IV. C is the Carnot Cycle we can
state that Cycles II. A and III. have the same efficiency as the
92 THE HEAT ENGINE PROBLEM.
Carnot Cycle with same previous compression. This is an im-
portant supplementary to the old theorem that the Carnot Cycle
has the highest efficiency for its temperature range.
The relation between the other values of H 2 are best shown
by the curves of Fig. 52 by implication. The quantities: Pres-
sure range, Volume range, Temperature range, do not need sepa-
a >*
A 2
100 300
rate sets of curves as we can get a fair idea of the values from an
inspection of the previous curves. If, however, any case seems to
call for an exact solution it can be obtained by a simp'e substitu-
tion in the formulae already given.
CYCLIC ANALYSIS OF HEAT ENGINES.
93
Mean effective pressure, volume and temperature, however, are
important values and not easily located relatively from the formulae.
Figs. 53, 54 and 55 show these curves as calculated for two cases
of compression. It may be here remarked that in the case of
Cycle IV. when the compression is 2:1 only about 44 B.T.U. can
be added to i Ib. air and with a compression of 10:1 about 282
B.T.U., this is why the curves end abruptly at these values of H r
OIL2.
C-
3
r
3HQ2
3HCI
A thorough discussion of the equations derived while important
and leading no doubt to many new and useful results would be
very long and would extend beyond the limits set for this paper
which had for its object rather the exposition of the method of
procedure than a thorough application of that method.
Besides the complete discussion referred to there is another im-
portant point of view to be taken of these formulae that of inter-
94
THE HEAT ENGINE PROBLEM.
pretation with respect to operating engines ; this is also reserved for
later treatment.
irz
The curves of the important cyclic variables as functions of the
heat supplied admit in their interpretation of the statement of
many important new propositions. Some of these are quite gen-
eral, while others are more specific. A few of the most obvious
will be noted.
The cycle consists of a series of operation or pressure volume
temperature changes resulting in a return to the original state of
pressure volume and temperature.
GENERAL PROPOSITIONS.
I. The P.V.T. at any point of a cycle depends on : (a) The cycle
itself qualitatively considered, i. e. y the nature and order of succession
of the processes or phases already completed ; () the extent or
intensity of each phase of the cycle qualitatively ; (c) the amount
of heat, ff, added before reaching the point considered. For exam-
ple, the temperature at the end of combustion will be different for
different cycles, will vary with the compression before heating, the
law of compression and the amount of heat added.
CYCLIC ANALYSIS OF HEAT ENGINES. 95
2. The part of the total heat transformed into work is a function
of the cycle, and will vary with the order, nature and extent of the
cyclic phases, except when all the heat is added and all abstracted
according to the same law.
3. When the laws of heating and of cooling are identical then
the part of the total heat supplied that becomes transformed into
work is constant for the same previous compression, and this re-
sulting efficiency is a function of the previous compression only
when these other two phases, compression and expansion, complet-
ing the cycle, have likewise the same law.
4. The range of changes in pressure volume and temperature is
different for different cycles, and in any one cycle will depend on
the amount of heat added.
5. While the variations noted do in general hold, yet in the dif-
ferent cycles each variable may be a different function of H lt so
that two or more curves may intersect, and for that particular value
of HI the variable will have the same value in two or more different
cycles simultaneously.
EFFICIENCIES (Fig. 52).
By inspection and plotting of formulae Nos. 40-58, page 90, for
the values of heat necessarily discharged we may draw some con-
clusions concerning efficiencies of the transformation process for
various cycles. Denoting as before the ratio of heat energy trans-
formed to that supplied by E = efficiency, it will be possible to
draw the following comparisons :
6. For Cycles II. A 2 , III., IV. C the efficiency is a function of
the adiabatic compression only and the same function for each.
It is independent of the amount of heat supplied, i. e., is not a
function of H.
7. For all cycles the efficiency increases with the compression,
but not according to the same law.
8. For Cycles IV., IV. A, IV. B, IV. C the efficiency decreases
with increase of heat added to the same mass of gas.
9. For all other cycles except II. A 2 , III., IV. C the efficiency
increases with H lt but according to different laws, so that the dis-
tance between efficiency curves will vary.
10. For these cases a change in H l will produce more effect
when HI is small than when it is large.
11. After heat has been added the efficiency will vary with the
96 THE HEAT ENGINE PROBLEM.
degree of expansion. Cycle II.. therefore, will have an efficiency
always higher than II. A and lower than II. B or II. C.
12. Cycles in which an adiabatic compression precedes heating,
will always have a higher efficiency than those lacking this com-
pression, other things being equal.
I 3. For the same initial conditions and same heat added it H^
is large enough Cycle II. C will always have the highest efficiency
(II. A )
and then come in order III. C; I. C; II. \ III. > always re-
I IV. C j
membering that IV., IV. A, IV. B, IV. C cannot exist if // a be
large.
14. The difference in efficiency between the curtailed expansion
of Cycle II. A 2 and that of II. increases with the amount of heat,
the difference being small when H^ is small and greater as H v in-
creases, the greatest possible being about 12 per cent.
15. Expanding Cycle II. to original temperature making Cycle
II. C, may increase the efficiency from 5 to 15 per cent, approxi-
mately for possible values of // r
1 6. Cycle III. may add by expansion to original temperature as
much as 25 per cent, to the efficiency for possible values of H^
17. Cycles IV., IV. A, IV. B have an efficiency decreasing with
increase of H provided H remain small ; when H passes a certain
limit the cycle ceases to be possible.
1 8. A change in the volume ratio of compression from \ to y 1 ^
will increase the efficiency of the cycles as follows for possible
values of H^.
Cycle II. . . . 30-20 per cent, approximately, depending on H lt
" II- A, )
" III. r 35 P er cent - approximately, depending on H 9
IV. C j
" II. C . . 40-5 per cent, approximately, depending on H v
TEMPERATURES (Figs. 46, 49, 55).
I. For the same previous compression the temperature resulting
in each cycle from heat addition and which is the maximum for
the cycle, will be different. That is, the addition of the same
amount of heat will result in a different temperature for each group
of cycles and the lines of Fig. 46 show that no two can be iden-
tical except I. and III., which cross.
CYCLIC ANALYSIS OF HEAT ENGINES. 97
2. Gases passing through Cycle I. may, on addition of a certain
amount of heat, H r have a temperature equal to what the same
gas would have passing through C>cle III. However, for more
heat added the temperature for I. will become higher than that for
III. while for less heat added III. will be higher.
3. Increase of compression before heating changes the temper-
ature after heating by only so much numerically as the varied
compression has resulted in changing the temperature before heat-
ing begins.
4. The temperature increase due to heating is proportional to
the amount of heat added // lf and the constant of proportionality
involves the reciprocal of the specific heat for the process and the
weight of the gas present.
5. After the gas has expanded to the greatest volume possible
in the cycle, no two cycles will leave the gas with the same tem-
perature except in a few special cases.
6. Cycle I. C, II. C, III. C, IV. C by definition have the same
temperatures at the end of expansion, and this is moreover con-
stant no matter what H may be and is equal to the initial temper-
ature of the cycle.
7. There will be a value of H l for a limited range of compres-
sions for which cycle III. may give to the gas the same final expan-
sion temperature as Cycle I.
8. Similarly II. for one compression may coincide in final tem-
perature with II. A for some other compression.
9. The temperature after expansion for cycle II. A, will always
be higher than for III. and III. higher than for II.
10. In round numbers II. A may be 25 per cent, higher than III.
and may even be 100 per cent, higher than II. for the same com-
pression for possible values of ff r
11. With variation of compression the temperature at the termi-
nation of expansion will vary, always becoming lower but the
extent of the lowering will depend on how much heat was added
before expansion and in case II. A and III. is exactly proportional
to H r
12. A change of compression J to y 1 ^ may change the tempera-
ture at the end of expansion in the case of cycle II. A and III. as
much as 80 per cent, for possible values of ff r
1 3. Mean effective temperature, Fig. 55, are different for different
cycles and for different compressions in the same cycle.
9 THE HEAT ENGINE PROBLEM.
14. Cycle IV. C is the only cycle with constant mean effective
temperature.
15. Mean effective temperature of all other cycles increase with
H,
1 6. For large values of H^ the order of magnitude of mean
effective temperatures will be: Lowest, IV. C, III. C, I. C, II. C,
III., I., II., highest, II. A.
17. For lower values of H l this order may be somewhat changed
and there will be points at which two different cycles will have
simultaneous values of M.E.T. and // r
PRESSURES (Figs. 47, 50, 53).
1. The pressures resulting from heat addition are different for
cycles with different numerals, but the same in any one group.
Thus, II., II. A, II. B, II. C or Group II. will all have the same
pressures, whereas those of Group II. will differ from those of
Groups III. and IV.
2. Lines representing pressures or functions of the heat supplied,
H r will cross as these functions are different for different groups,
and it will hence be possible for the different groups of cycles to
have the same pressures for certain values of H^.
3. Groups I. and II. have pressures after heating that increase
with H^ while in Group III. the pressure is constant and in IV.
decreasing with increase of // r
4. For same compressions Group II. will always have the high-
est pressure after heating, and III., IV. and I. come in the order
named for moderate H l% while for large H l IV. cannot exist.
5. Increase of compression will change the pressure after heat-
ing in Group III. only so much as results from the changed com-
pression before heating. In Groups II. and I. the change is such
as to keep the pressure ratio before and after heating constant ;
so that for a given change in H l the resulting pressure change in
II. will be greatest for high compressions, less for moderate com-
pressions and least for no compression, i. e., for Group I.
6. After expansion by definition the pressures of I., II., III. and
IV. are all atmospheric and equal.
7. The pressure which II. A 2 will reach when the gas has ex-
panded to original volume increases with H l and is such that the
ratio of this pressure to atmospheric is the same as the ratio of
pressure after heating to that before.
CYCLIC ANALYSIS OF HEAT ENGINES. 99
8. Cycles with letter Call go below atmosphere in expanding to
such a pressure as will bring the temperature down to that origi-
nally existing in the gas. These resulting pressures after expan-
sion are different for each cycle, but the lines representing them as
functions of H^ may intersect.
9. The lines for IV. C may cross others, but I. C, II. C, III. C
cannot intersect and these will always be in the order of magni-
tude II. C, III. C, I. C and all asymptotic to axis of //, so that the
terminal pressure can never be zero.
10. An increase of compression will cause an increase in final
pressure for same // r
11. Mean effective pressure expressed as a function of H l will
give for every cycle and every different compression a different
M.E.P. curve, but as before these may intersect.
12. For all cycles except those ending with isothermal return
to the original state, the M.E.P. increases with H l but for those
bearing the letter C the M.E.P. decreases and for no cycle is it
constant.
13. For the same previous compression the cycles have M.E.P.
of about the following order of magnitude when H l is large
enough.
Greatest M.E.P., II. A 2 , 200; II., 40; I., 25; III., 15; II. C,
1.5; I. C, 0.3; III. C, 0.2. When H^ is small IV. will probably
come between III. and II. C,
14. A change in compression from \ to T x (vols.) may cause a
change in II. A 2 of 35 per cent., II of 100 per cent., III. of 300
per cent, for the same possible values of H r
15. The effect of changed compression before heating is the
more marked on M.E.P. resulting when M.E.P. is lowest and the
extent of the increase is greater with H^
VOLUMES (Figs. 48, 51, 54).
1. The volumes after heating are the same for cycles of the
same group and for all groups increase with H v except in Groups
I. and II. where by definition they are constant and equal to the
volumes existing before heating.
2. In Group III. the volumes after heating are proportional to
HI with the same constant of proportionality for the same com-
pression. Increase of compression decreases this constant of pro-
portionality.
100 THE HEAT ENGINE PROBLEM.
3. In Group IV. the volumes increase rapidly with H v but are
not proportional to H^ so long as H^ is small ; with large //, Group
IV. cannot exist.
4. Lines of volumes after heating represented as functions 01
H l may cross in some cases. II., IV. and III. may cross I., i.e., the
compression cycles may cross the non-compression ones. But for
the same compression II., III. and IV. can never have the same
volumes after heating. Lines of III. and IV. for high compres-
sion may cross II. for a lower compression but cannot cross each
other.
5. For possible values of H l the volumes after heating for the
different groups may have the following order of magnitude if
//j is large enough : Group III., 55.00; group I., 12.38; group II.,
6.00.
6. After expansion is completed the volume occupied by the gas
in the different cycles will vary through very wide limits, increasing
with H^.
7. The volume occupied by Cycle III. will be such as to keep
the ratio between this final volume and the volume before com-
pression the same as the ratio ot volume after heating to that be-
fore and the final volume is proportional to H v The constant of
proportionality is decreased by compression increase.
8. The final volume of Cycle II. A is least and equal to that
existing before compression.
9. When H^ is large enough there may be a value for which
the final volume may exist in the following order of magnitude :
III. C lf 7,000.00; I. C, 4,200.00; II. C, 2,300.00; III , 75.00; I.,
65.00; II., 51.00, II. A 2 , 12.38. A change of compression by
which the volume after compression is one fifth that for the pre-
vious case may change this list to the following : III. C., 1,000.00;
I. C, 4,200.00; II. C, 500.00; III., 40.00; I., 65.00; II., 34.00;
II. A 2 , 12.38.
10. The mean effective volumes increase with H v for all cycles
except II. A 2 in which this variable is constant.
11. For cycle III. the M.E.V. is proportional to H^ and increase
of compression increases the constant of proportionality.
From the data here set down the selection of a cycle on purely
ideal grounds can be made with a full knowledge of all the con-
ditions surrounding the selection ; that is knowing what results are
CYCLIC ANALYSIS OF HEAT ENGINES. IOI
desired the cycle that theoretically, ideally or mathematically con-
sidered gives the results can be found and in addition it is easy to
see what accompanying circumstances are inevitable. If that
cycle that transforms the greatest amount of heat into work ideally
is wanted it is readily seen that II. C with as high compression as
possible must be selected, but it is also evident that a very large
volume range must be submitted to. If that cycle with the lowest
temperature range is wanted then any of Group IV. must be taken.
If a cycle is desired that will convert of any amount of heat the
same proportion into work then any one of II. A, III. or IV. C,
but of these one has the lowest pressure range, another the lowest
temperature range and the last the lowest volume range.
Examples could be multiplied almost indefinitely, but enough
has been said to make clear the purpose of and justify this laborious
analysis, for the results desired can be set down at once for cycles
considered, and, moreover, for any cycle not considered it is evi-
dent that similar treatment will place it at once in comparison with
all these presented.
NOTE. This paper is sent to you that you may examine it in advance of the
meeting, and prepare any discussion of it which you may wish to present.
" Tt is issued to the membership in confidence, and with the distinct understand-
ing that it is not to be given to the press or to the public until after it has been
presented at the meeting.
The Society as a body is not responsible for the statements of fact or opinion
advanced in papers or discussion. (Art. 44 of its Rules.)
As there will be no adequate supply of extra copies, and papers are liable to
be read by abstract only, preserve this copy for your use, and
BRING THIS COPY WITH YOU TO THE MEETING.
(Subject to Revision.)
No. 926.*
THE HEAT-ENGINE PROBLEM. \
BY CHARLES E. LUCRE, NEW YORK.
(Non-Member.)
* AND PRESENTED BY R. H. FERNALD.
(Associate Member.)
1. A MATHEMATICAL analysis of the different cycles of variation
of state through which a mass of gas may pass can give no more
than a provisional idea of the value of those cycles for convert-
ing the energy of heat into useful power. Such an analysis
must presuppose certain ideal conditions that may or may not
be possible in practice, and though mathematically we may find
that one cycle should convert more of the heat supplied into work
than any other, there may be difficulties in the way of practically
getting this result. It may happen, for example, that a very
complicated large or heavy machine is necessary, or that the
* To be presented atlhe New York meeting (December, 1901) of the American
Society of Mechanical Engineers, and forming part of Volume XXIII. of the
Transactions.
f For further discussion on the same topic consult the Transactions as
follows :
No. 843, vol xxi., p. 396 : "An Efficiency Test of a One Hundred and Twenty-
five Horse-power Gas Engine." C. H. Robertson.
No. 861, vol. xxi., p. 961 : "The Gas-engine Hot Tube as an Ignition-timing
Device." Win. T. Magruder.
No. 875, vol. xxii., p. 152 : '" Efficiency of a Gas Engine as Modified by Point of
Ignition." C. V. Kerr.
No. 879, vol. xxii., p. 312: "A New Principle in Gas-engine Design," C. E.
Sargent.
No. 895, vol. xxii., p. 612 : " Efficiency Tests of a One Hundred and Twenty-five
Horse-power Gas Engine." C. H. Robertson.
2 THE HEAT-ENGINE PROBLEM.
required changes of state in the gas cannot be carried out at
all, or, perhaps, not fast enough to be useful in a prime mover.
In the general study, then, of the heat-engine problem, we
must add to the analytic cyclic discussion a careful considera-
tion of a number of practical questions, the results of which,
when allied with the mathematical analysis, will permit of a
logical selection of the proper cycle to which we should devote
our executive energies ; their goal is the production of that
prime mover whose source of energy shall be heat, whose
medium of transformation of this heat into work a perfect gas,
and which shall call for the simplest machine, giving the greatest
power in the smallest space with the least metal and under the
most favorable circumstances.
2. Every cycle available for transforming heat energy into
mechanical energy by the moving of a part against a resistance,
must include as one of its phases the heating of the gas in some
particular way peculiar to that cycle. This giving of heat
energy to the transforming gas presupposes a source of heat
which in practice must be a fire. The heat of a fire may be
imparted to a mass of gas in three ways :
I. The fire may be placed on one side of a wall through which
the heat must pass to the mass of gas on the other side ; this
may be termed external heating.
II. The fire may be caused to heat a solid mass, which is
afterward shut off from the fire and brought into contact with
the mass of gas ; this is a combination of external and internal
heating.
III. The fire may be enclosed and maintained by the mass of
gas itself ; in this case the gas must be, at least in part, air which
will furnish oxygen for this internal combustion.
3. Any system which depends on the heating of the gas by
contact with solid matter at a high temperature, must necessarily
be slow in operation and involve large masses of gas. For the
transfer of heat, the source must be hotter than the receiving
mass, and a difference of temperature, for a given rate of transfer
sufficiently high to be of practical value, must be greater than
the medium of transfer can stand without injury. Consider
how hot the walls of a chamber would have to be to heat a mass
of gas as rapidly as is done in the gas engine, and the point
made above will be clear. Nevertheless, engines with this kind
of heating have been built, but, admirable as some of them have
THE HEAT-ENGINE PROBLEM. 6
been in conception, they bave proved failures as prime movers
in competition with others because of the points noted. The
engines of Ericsson, Eankine, and the Stirlings are all in-
cluded in this class, with results that are well known. Erics-
son's large engine of 300 horse-power showed a mean effective
pressure of about 2 pounds per square inch with a piston area
of 600 square feet. The only machine now working with this
external heating is the one known as Eider-Ericsson, used in
small sizes only for the slow pumping of water.
4. Nothing that this system can do will compare with what
may be derived from the use of the internal-combustion method
of heating. This internal-combustion heating of a mass of gas
will permit of a heating as rapid as we choose, and to any
temperature up to a certain maximum. If all the air sup-
plied has its oxygen converted with the fuel to CO 2 , H^O, etc.,
there being no excess of either oxygen or fuel, then the mass of
gas which, it is true, has changed in chemical composition, but
not materially in physical properties, has received the maxi-
mum amount of heat obtainable from the combustion of the
fuel used. If only a part of the air support combustion and the
products be diluted with unused air or by steam, etc., then any
desired temperature between the original temperature of the
gases and the maximum may be obtained. The problem of
heating gases by an internally maintained fire is difficult, com-
pared with the other method of external heating, and this may
account for its later application. We might say in brief that
externally heating a gas is thermally bad but easily done,
internally heating the gas, thermally good but not so easy
to do.
5. Heating working gases by internal combustion has been
done with coal, oil, and gas. The methods used might be tabu-
lated briefly.
I. With coal :
(a) Air is passed through a coal fire with or without a grate.
Cayley, Shaw, and Genty.
(6) A coal fire is moved through an enclosed mass of air.
Lord.
II. With liquid fuel not previously vaporized :
(a) The enclosed air acts as a quiet atmosphere supporting
the combustion of a jet of oil flame. Diesel.
(b) The air is caused to move past a burner, and in passing
THE HEAT-ENGINE PROBLEM.
supports combustion, the heated products passing on. Wilcox,
Bray ton, Nordberg, and Shad all.
(c) Oil is thrown into a hot chamber, there vaporized, and
brought into contact with the air, the proportions being so
maintained as to make the resulting gaseous mixture explosive.
Combustion is of the self- propagated sort. Hornsby, Mietz &
Weiss, and Capitaine.
III. With gas or previously vaporized oil :
(a) An enclosed air atmosphere supports a quiet jet of gas
flame. Diesel and Gibbs.
(b) Air in motion passes a fixed gas flame as in most atmos-
pheric engines. Wilcox, Weiss, and Otto atmospheric.
(c) Air mixed with gas in explosive proportions is caused
to pass a point where the combustion is localized. Brayton,
Schmid, Beckfeld, and Reeve.
(d) Air mixed with gas in explosive proportions is enclosed
in a chamber, and while at rest burned by self-propagation,
after inflammation was provoked by a local ignition. Otto,
Priestman, Nash, Westinghouse, and in fact nearly all existing
internal-combustion engines.
The above classification leads directly to the broad division
of internal-combustion engines into two great classes, the explo-
sive and non-explosive. The term " explosive " we shall apply to
all those engines in which a mass of gaseous mixture at rest is
ignited at one point, and the whole burned by self-propagation.
The other term, " non-explosive," we shall apply to those engines
in which the gases are in motion and in that motion pass a
point where combustion is localized, and are there heated in the
passing. To complete our terminology, we add the expressions
" intermittent non-explosive," to those machines in which the
combustion is periodically interrupted at the cylinder end as
in Diesel's, and " continuous non-explosive " to those in which
the combustion is maintained in a chamber, and the hot gases
used as required, as in Reeves, Schmid, and Beckfeld.
We have, then, explosive engines ; non-explosive engines with
intermittent combustion, or continuous combustion as the dif-
ferent kinds of internal-combustion engines.
6. The explosive engine as developed and perfected, chiefly
by Dr. Otto, holds the field to-day, and its very general use
has brought out its merits and demerits. It has been, and is
to-day, the subject of many researches and experiments, all
THE HEAT-ENGINE PROBLEM. 5
tending to perfect it by the discovery of its faults. All this has
resulted in its present position, which might be summed up as
follows :
It is extremely simple in construction, having comparatively
few working parts. .
The thermal changes of heating and expansion are all per-
formed in the same place, on a quiet mass of gas, and nothing
but the gas is heated.
The best engines those of rational design do not differ much
in construction and results, and this brings out an important
point that in the handling of a mass of gas to be exploded, we
accept a certain inflexibility from which we cannot escape.
7. As a machine, it cannot compare with the steam engine.
It is not easy to start, and cannot be worked at widely variable
speeds ; its governing is bad, the speed varying at different points
of the stroke, but adding up to a fairly constant total number of
revolutions per minute ; it has no margin of power and carries
an ignition system that once deranged stops the machine ; it is
non-reversible ; it has a low mean effective pressure for high-
pressure range, hence is heavy ; it can use only one kind of
fuel, and that gas, and whether this be produced from oil or
coal, it must, nevertheless, be produced outside the natural gas
regions. It cannot, and never will be, able to use crude, unre-
fined oils directly ; it operates under only one fixed cycle.
8. About as much can be said for the explosive type as against
it. It has occupied nearly all workers in the internal-combus-
tion field for the past thirty years, and the success attained
continues to draw to the problem large numbers of men and
an immense amount of capital, and these, working together,
must do much for this type in the future. But while this good
work goes on, there is no reason why the other types of internal-
combustion engines should not receive their share of attention.
Some have been built and many proposed; some were successful
and some failures ; but a careful study of what has been done
successfully and the cause of failure of the unsuccessful en-
gines would, if no more, show clearly the possibilities of this type.
If the difficulties are clearly set forth, the solution will be the
easier, and if in the study of the difficulties the solution appears,
so much the better.
Of the successful engines of the non-explosive type, there
may be mentioned two that easily head the list, the old Bray-
THE HEAT-ENGINE PROBLEM.
ton and the modern Diesel, and the results obtained from these
machines are certainly encouraging. However, before entering
into a discussion of the various non-explosive machines, it would
be well to make sure of our theoretical grounds.
9. The different cycles of operation that might be performed
on a mass of gas are infinite/ but there is a limited number which
are striking and simple. These are given below.
Let Fig. 1 be a pressure-volume diagram for the Cycle I.
Then we have :
From B to (7. Addition of heat isometrically from atmos-
pheric pressure.
From C to D. Adiabatic expansion to atmospheric pressure.
From D to B. Cooling at atmospheric pressure.
P CYCLE. I
D.
v
FIG. 1.
In Fig. 2 we have for Cycle IA.:
From B to (7. Addition of heat isometrically from atmos-
pheric pressure.
CYCLE I A.
FIG. 2.
THE HEAT-ENGINE PBOBLEM.
From C to D. Adiabatic expansion to a point above atmos-
pheric pressure.
From D to E. Cooling isometrically to atmospheric pressure.
From E to E. Cooling at atmospheric pressure.
In Fig. 3 we have for Cycle IB.:
From B to C. Addition of heat isometrically from atmos-
pheric pressure.
From C to D. Adiabatic expansion to below atmospheric
pressure.
From D to E.
From E to B.
Cooling isothermally to atmospheric pressure.
Cooling at atmospheric pressure.
CYCLE I.B.
V.
FIG. 3.
In Fig. 4 we have for Cycle 1C.:
From B to C. Addition of heat isothermally from atmos-
pheric pressure.
CYCLE 1C
V
FIG. 4
8
THE HEAT-ENGINE PROBLEM.
From C to D. Adiabatic expansion to a pressure below
atmospheric such that we get,
From D to B. Cooling isothermally to the original volume
and atmospheric pressure.
In Fig. 5 we have for Cycle II :
From A to B. Adiabatic compression from atmospheric
pressure.
From B to C. Addition of heat isometrically.
From C to D. Adiabatic expansion to atmospheric pressure.
From D to A. Cooling at atmospheric pressure.
CYCLE II
FIG. 5.
In Fig. 6 we have for Cycle TLA.:
From A to B. Adiabatic compression from atmospheric
pressure.
From B to C. Addition of heat isometrically.
CYCLE HA.
FIG. 6.
THE HEAT-ENGINE PROBLEM.
From C to D. Adiabatic expansion to a pressure above
atmospheric.
From D to E. Cooling isometrically to atmospheric pressure.
From ^to A. Cooling at atmospheric pressure.
In Fig. 7 we have for Cycle IIB.:
From A to B. Adiabatic compression from atmospheric
pressure.
Addition of heat isometrically.
Adiabatic expansion to pressure below atmos-
From B to C.
From C to D.
pheric.
From D to E.
From Eio A.
Cooling isothermally to atmospheric pressure.
Cooling at atmospheric pressure.
CYCLE KB.
A.
V
FIG. 7.
In Fig. 8 we have for Cycle 11(7.:
From A to B. Adiabatic compression from atmospheric pres-
sure.
CYCLE JIG
V
FIG. 8.
10
THE HEAT-ENGINE PROBLEM.
From B to C. Addition of heat isometrically.
From C to D. Adiabatic expansion to a pressure below
atmospheric such that we get,
From D to A. Cooling isothermally to the original volume
and atmospheric pressure.
In Fig. 9 we have for Cycle III :
From A to B. Adiabatic compression from atmospheric
pressure.
From B to 0. Addition of heat isopiestically.
From G to D. Adiabatic expansion to atmospheric pressure.
From D to A. Cooling at atmospheric pressure.
CYCLE 111
FIG. 9.
In Fig. 10 we have for Cycle
From A to B. Adiabatic compression from atmospheric
pressure.
From B to C. Addition of heat isopiestically.
From G to D. Adiabatic expansion to a pressure above
atmospheric.
P CYCLEIA.
FIG. 10.
THE HEAT-ENGINE PROBLEM.
11
From D to E. Cooling iso metrically to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
In Fig. 11 we have for Cycle Illi?.:
From A to B. Adiabatic compression from atmospheric
pressure.
From B to C. Addition of heat isopiestically.
From C to D. Adiabatic expansion to a pressure below
atmospheric.
From D to E. Cooling isothermally to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
CYCLE fflli
FIG. 11.
In Fig. 12 we have for Cycle III6 Y .:
From A to B. Adiabatic compression from atmospheric
pressure.
From B to C. Addition of heat isopiestically.
From C to D. Adiabatic expansion to a pressure below
atmospheric such that we get,
p
CYCLE 111C.
A.
V -
FIG. 12.
12
THE HEAT-ENGINE PROBLEM.
From D to A. Cooling isothermally to the original volume
and atmospheric pressure.
In Fig. 13 we have for Cycle IV.:
From A to B. Adiabatic compression from atmospheric
pressure.
Addition of heat isothermally.
Adiabatic expansion to atmospheric pressure.
Cooling at atmospheric pressure.
CYCLE IV.
From B to C.
From C to D.
From Eto A.
FIG. 13.
In Fig. 14 we have for Cycle IV A.:
From A to B. Adiabatic compression from atmospheric
pressure.
a CYCLE IVA.
FIG. 14.
THE HEAT-ENGINE PROBLEM.
13
From B to G. Addition of heat isothermally.
From G to D. Adiabatic expansion to a pressure above
atmospheric.
From I) to E. Cooling isometrically to atmospheric pressure.
From E io A. Cooling at atmospheric pressure.
In Fig. 15 we have for Cycle IV B.:
From A to B. Adiabatic compression from atmospheric
pressure.
From B to G. Addition of heat isothermally.
From G to D. Adiabatic expansion to a pressure below
atmospheric.
From D to E. Cooling isothermally to atmospheric pressure.
From E to A. Cooling at atmospheric pressure.
P iB
CYCLE IV B
A.
FIG. 15.
In Fig. 16 we have for Cycle IV C.:
From A to B. Adiabatic compression from atmospheric
pressure.
From B to C. Addition of heat isothermally.
From G to D. Adiabatic expansion to a pressure below
atmospheric such that we get,
From D to A. Cooling isothermally to the original volume
and atmospheric pressure.
Besides these there are various atmospheric cycles some-
times called vacuum cycles in which the first step is the heat-
ing of the entering charge at atmospheric pressure. Because
of their slight importance they are here omitted.
THE HEAT-ENGINE PROBLEM.
CYCLE JVC.
A.
FIG. 16.
10. A very careful mathematical analysis* of all these cycles
leads to these conclusions :
(A) Cycle I. and its variations, by reason of its poor showing
in efficiency and mean effective pressure as compared with the
previous-compression Cycle II., must be set aside.
(B) The atmospheric cycles, by reason of their low mean
effective pressure and consequent large volume range, are use-
less for power purposes as compared with the other cycles.
(C) This leaves as the only cycles worthy of application, II.,
III., IV., and their variations.
(D) Of the last mentioned, there are three which are peculiar,
and these are : Cycle 11,4., Otto, heating and cooling the gas
at constant volume ; Cycle III., Brayton, heating and cooling
the gas at constant pressure, and Cycle IV (7., Carnot, heating
and cooling the gas at constant temperature.
All these have the same efficiency for the same compression,
and should, consequently, with the same heat supplied, do the
same work.
The efficiency of each is given by
where V a is the volume before compression,
F 6 " " after i
;/ " ratio of specific heats, and for air, y ~ 1.406.
* Columbia School of Mines Quarterly, Nos. 1, 2, 3, vol. xxii., 1901.
THE HEAT-ENGINE PROBLEM. 15
(E) The other cyclos, II., IIB. and C.; Ill A. and B.; IV.,
IV A. and B., can easily be given their proper comparative
position by remembering that each is a more or less complete
expansion of one of the above three. For example, if in the
Otto the expansion were carried to atmospheric pressure, the
efficiency would be greater than for the Otto. Similarly with
the Carnot, if the expansion were stopped at atmospheric pres-
sure as was first suggested by Diesel, the resulting Cycle IV.
would have an efficiency less than the Carnot, and hence less
than either the Otto or Brayton cycles.
(F) The other variables entering have the values tabulated
for each of the cycles adopted for comparison.
( Same mass of gas,
Given the < Same heat supplied after,
' Same compression.
There will result for
Cycle II A., Otto }
" III. Brayton > Same work done, and lience same efficiency.
" IV C. Carnot )
And, further,
Lowest. Intermediate. Highest.
Maximum temperature Carnot Brayton Otto
Pressure range Brayton Carnot Otto
Volume range Otto Brayton Carnot
Temperature range Carnot Brayton Otto
Mean effective pressure Carnot Brayton Otto
Pressure range
;rj ~ T^ - Brayton Carnot Otto
Mean effective pressure
Mean effective temperature Carnot ayton Otto
The relation of the Diesel to the Otto an Brayton is easily
seen, if we remember it as an imperfect Car..ot.
11. Some of these variables should be a maximum and some
a minimum. For the maximum teii*~ crature the Carnot holds
first place, but its impracticability yields the place to Brayton.
Neither pressure range nor mean effective pressure is wanted
by itself, but only the ratio between them, for it is to this ratio
that the weight of the engine must be approximately propor-
tional ; here Brayton holds the most favorable place.
Volume range should be low, and here first place is held by
the Otto. The mean effective temperature should be low, and
the Brayton is exceeded only by the Carnot.
The low mean effective pressure of the Carnot, and all other
16 THE HEAT-ENGINE PKOBLEM.
isothermal combustion cycles, is sufficient warrant for cutting
them out of consideration in comparison with the Cycles II.,
III., and their variations.
We have thus arrived at the conclusion that, theoretically,
the last-named cycles only are worthy of further consideration.
Of these the Brayton, III., holds a most favorable position,
being surpassed by the Otto only in the position of volume
range.
12. In the above, the hypothesis that heat could be added to
the gas has been assumed, and no account taken of the means
of so doing, but this point needs consideration. If heat be
added through walls from a source of known supply, of which
we can control and use as much or as little as we please,
there will be no alteration in the formulae or results of the
analytical comparison; but the internal- combustion method of
heating presents some new questions for solution. First, the air
and fuel become carbonic acid, steam, etc., and as to what value
of the specific heat should be used, who can say ? Second, the
chemical change is accompanied by an intrinsic volume change.
Third, there may be reasons why the fuel should give out more
heat when burned in one way than when burned in another.
13. The only ways of heating by internal combustion that are
worth anything for power are the constant volume and constant-
pressure methods. On theoretical grounds, we have no reason
for saying that, for any particular system of combustion, more
heat can be developed one way than the other. The evidence
that heat has been added to a mass of gas in an engine is, for
the two cases : (A) an increase of pressure, and (B) an increase
of volume. This pressure increase on the one hand and vol-
ume increase on the other we can readily observe by indicators,
and the results of these observations on a large number of
indicator cards show that the increase is not what it should
be if all the calorific value of the fuel had developed.
In short, there is in practice abundant evidence of heat supres-
sion, and whether this be due to radiation, conduction, dissocia-
tion or an increase of specific heat, or to an actual non-production
of heat is unknown. What we do know and can assert is that the
effects on pressure and volume are such as they would be if
only a part of the heat supposed to be generated had appeared.
The result might be worked up to give a new value to the heat-
ing power of the fuel, to be called its effective calorific value, or a
THE HEAT-ENGINE PROBLEM.
new value given to the specific heat, to be called the effective
specific heat of the process.
14. For constant-volume combustion we have, for H } , the
British thermal units per pound of mixture,
where p l = pressure before compression.
T l temperature before combustion.
p 2 =. pressure after combustion.
T. 2 = temperature after combustion.
C v = specific heat at constant volume.
This ratio in the general run of gas engines will average
about 3.5, In some cases it may reach 4, but I know of no case
where it has reached 5. Some values are given below:
J92
Engine. p l Remarks.
Westingliouse 3 On Gas
Otto 4.5 N.Y.Gas
Hornsby Ackroyd 3.5 Kerosene
Nash 4 ....N.Y.Gas
Clerk 4 Glasgow Gas
Crossley 3 Dowson Gas
f Priestman 3.5 Kerosene
Crossley oil 3.5 .* Kerosene
A general statement, very nearly true, would give these pres-
sure and temperature ratios about 50 per cent, of what the usual
values of II { and C v would produce. These figures, while not
strictly true for any one case, give a fair average value.
15. The other system of combustion that at constant pres-
sure may be observed in the same way. The only indicator
card I have of this type of engine was taken from a Brayton oil
engine with its smoky fire. The volume ratio, in this case, is
quite well given by the relative lengths of the delivery line of
the compressor and the admission line of the power cylinder,
and is given by
? = 3-2.
Let us see how this compares with the pressure ratios given.
Theoretically,
* = -?" -i + L?L
91 ~ V ~ f^ '7 T '
18 THE HEAT-ENGINE PROBLEM.
where C p is the specific heat at constant pressure and the
other symbols are as heretofore ; combining this with the simi-
lar one for the other type we get
or
'
V
= 1 + y ~ - 1 Take y = 1.4 ; and
By substitution, when
l?2 =l, we get
'2 9 * 9 A
= &. as,
V, Pi
^- = 3, " " ^- = 3.8.
Vl Pi
-^- = 3.2, " " ^- = 444.
16. This shows thai when a Brayton engine gives a volume
ratio in combustion of 3.2, there is evidence of as much heat as
would cause a pressure ratio of 4.44 in an explosion engine ;
hence it would seem that, for the combustion process alone, the
Brayton engine, even with its poor fire, was giving evidence of
as much heat as the very best explosion engine, and more than
can most of them. This point is very striking, and, in order to
verify or disprove it, a large mass of data is necessary, which
can be collected only after considerable time.
The above point bears strongly on the formulae of cyclic com-
parison. The analysis showed that the Otto and Brayton cycles
must have the same efficiency for the same heat added ; but, if
one, by reason of its system of combustion, can take from the
fuel more heat than the other, then that one must have the
higher efficiency.
17. All non-explosive internal-combustion engines, except the
atmospheric types, must provide for three stages : first, the sup-
plying of working gases, which are derived from air and fuel,
hence, we need an air and a fuel supply ; second, the causing of
THE HEAT-ENGINE PROBLEM. 19
the combination of the fuel and air in combustion to raise their
temperature, and thereby vary either pressure or volume of the
gas, as we desire ; third, the utilization of the hot gases thus
produced to actuate a mechanism by the action of expanding
gas on a moving part.
I. The air and fuel supply may be accomplished in any of the
ways known to and accepted by engineers ; the results cannot vary
much with changes in design of this part, since compressors and
pumps are well-known machines.
II. The burning of the fuel in the air supplied offers what is
probably the most difficult problem for solution. Its difficulty
is attested by the variety of the means proposed and the indif-
ferent success of those that have been tried. When solid fuel
was used, as in Cayley's engine, no means, without great compli-
cation of parts, were found adequate to cope with the smoke, dust,
and distillation of gas from the coal. With liquid fuel, Bray-
ton was troubled with soot, and those burners which have
burned clean required a large excess of air. With gas, Brayton
also failed, and he was not alone, as no adequate system of burn-
ing gas, when enclosed and under pressure, had then been pro-
posed, the trouble being not so much in getting a burner to work
under specified conditions, as to get one that would work under
wide and sudden variations of feed and pressure.
III. The utilization of the hot gases has been successfully
tried in cylinders, and rotary machines have been proposed, in-
cluding the turbine ; though none have appeared on the market,
ifc is inconceivable that there can be any serious trouble to
be anticipated in such utilization. The reason of the general
failure of the machines proposed is probably the difficulty
noted above, for gases at high temperature are used every day
in exploding engines with the greatest ease. All of these non-
explosive engines may work under any one of several cycles,
depending on the relation between the last two processes
the amount of heating compared with the amount of expansion
permitted. Here is an important point, for by a simple control
of the above relations, by passing air around the fire and vary-
ing the cut-off to the power cylinder, we can vary the cycle,
hence the work output ; thus an engine equipped with means to
do this would be able to work at all loads equally well, and be
able to pull up to a temporary overload, just as do steam engines.
This great elasticity of action is beyond comparison with the
20 THE HEAT-ENGINE PROBLEM.
rigidity of the explosive engine. Moreover, the question of
available fuel again comes up ; anything that will burn may be
used, and with it a working elasticity obtained - two desirable
results.
Before we examine the details entering into engine construc-
tion, let us look at some of the complete machines that have
been proposed for carrying out the system, dividing our study
according to the fuel used, taking up, first, coal-burning ; second,
oil-burning; and third, gas-burning engines.
Coal-bu r n ing En gin es .
18. The first working engine of this class was Cay ley's furnace-
gas engine. The air was forced into the fire-box, where a coal
fire was maintained, and the hot gases used in a cylinder. This
engine worked on what has been called the Brayton cycle.
Rankine says of it : " The cylinder, piston, and valves were found
to be so rapidly destroyed by the intense heat and dust from the
fuel that no attempt was made to bring it into use."
In the United States, Philander Shaw proposed the engine of
Figs. 17 and 18 in 1861. Air from a pump cylinder passes more
or less through a coal fire, becoming heated in its conversion, and
finally with increased volume working in a larger power cylin-
der. The fire-box is provided with a grate, c, and is lined with
brick e/(Fig. 17), fuel is fed through //(Fig. 18), and is moved
by a piston head g. The furnace has openings, a\ below the grate,
and others, a 2 , above the fuel. Two single-acting cylinders are con-
nected at 180 degrees to one shaft. Each cylinder is in two parts :
the upper, A, is finished to a fit with piston D the lower part,
J5, is left rough, as the lower part, F, of the piston does not touch
it. The top of the piston has a trunk, C, which acts as an air-
pump cylinder with the main cylinder casing. The motion of the
air is indicated by the arrows, and the proportion that enters the
fire-box above or below the fuel is controlled by a valve r (Fig.
17). The heated air and gaseous products of combustion pass
into the cylinder through valve p, and the exhaust passes out
through a heater for incoming compressed air. A little flange,
#, is placed to catch dust and a groove, v, for oil.
19. All parts where radiation is likely to occur are jacketed
by the incoming air. The working part of the piston fits loosely,
and at a point just above the highest position of its bottom is an
THE HEAT-ENGINE PROBLEM.
22
THE HEAT-ENGINE PROBLEM.
annular space kept filled with cool air to prevent overheating of
the working faces. Governing may be made two-fold : first, the
amount of heat added fco the air can be regulated by sending
more or less through the fire ; and, second, the power may be
directly controlled by the main cylinder cut-off. The air receives
heat really from one source, though in two places the one source
being the fire, and the two places the exhaust-warmed pre-
heater, or regenerator, and the fire-box. A hand pump is pro-
posed to raise the internal pressure for starting.
20. The engine proposed by Henry Messer in 1863 is shown
in Figs. 19-21. The air pump is double-acting, and the power
FIG. 19.
JLJL
FIG. 80.
THE HEAT-ENGINE PROBLEM.
23
cylinder single-acting, hence there are two stages : 1. On the
down-stroke air is compressed and heated at decreasing
volume ; 2. On the up stroke the air, previously compressed
and heated, is sent through the fire and thence to the working
cylinder. When the up-stroke begins, the high-pressure gases
in the reservoir begin to expand at the same time that the air
in the pump begins to increase in pressure, and, finally, when
the increasing pressure in the pump equals the decreasing pres-
sure in the receiver the second cylinderful of air becomes avail-
able. In its passage the air may take up enough heat to maintain
p constant, T constant, or neither ; it is impossible to say, hence
FIG. 22.
the cycle is indeterminate. Later Messer suggests some changes
in valve construction to prevent overheating, also in governing
by throttling between the fire and air receiver.
21. Cyrus W. Baldwin proposed the engine of Fig. 22 in 1865.
The engine cylinder has between its upper cool part and lower
hot part an //-shaped water passage f in accordance with his
belief that more trouble with overheated working faces will be
caused by heat conducted through the metal cylinder walls than
by contact with the hot gases. He provides for distilling gases
from the coal fire by supplying an auxiliary fire beyond the main
furnace. He says : " Part of the fuel in the large furnace is
changed by the heat therein to volatile gases, which do not burn
when generated, but will burn if, while they are hot, they are
24 THE HEAT-ENGINE PROBLEM.
brought into flame. To supply such a flame through which all
the products of combustion from the large furnace must pass,
a small furnace is supplied, and the results which follow its
application are found in practice to be highly beneficial."
22. In the engine of L A. L. Soclerstrom, 1869, a radical
change of arrangement is proposed, as shown in Fig. 23. The
FIG. 23.
upper part of the cylinder acts as the compressor and sends air
down and around the casing, the air passing the exhaust reheat-
ing coil B. After this reheating by contact with the exhaust
coil the air is sent out through the opening yx on top of the fire
at a, the top combustion helping to prevent coking and distilla-
tion, but adding trouble from dust and ash. Governing is effected
by a split current and variable cut-off.
THE HEAT-ENGINE PROBLEM.
25
23. Thomas M. Fell in 1880 proposed a very interesting
though complicated machine (Fig. 24). An air compressor, C\
sends air through the pipe M l to the fire at A. Exhaust gases
from the power cylinder A 1 are thrown first around the tubes
D\ and then around the water- cooled tubes E l , and are returned
by the pump J3 l , through the tubes D\ and back to the hot
chamber T through the valve IP. Accumulation of gases in
the closed system is prevented by the blow-off G l , arranged to
maintain a constant pressure on the high-pressure side of the
system. An attempt is made to keep a two-par 1; system, one
FJG. 24.
high-pressure heating, and the other low-pressure cooling. This
system would probably give the results of Cycle IIIC., OL' one
somewhat similar.
24. Hiram S. Maxim in 18S4 proposed a machine as shown in
Fig. 25. Air is drawn by the right-hand side of the piston from
the space A', which in turn gets its supply from the atmosphere
by a throttling slide T. This space A 1 has a diaphragm E so
arranged that a partial vacuum will cause the slide valve N to
by-pass air around the fire. In its normal operation, or when
running slowly, most of the air passes through the fire and is
heated. At the time the air is passing into the receiving cham-
ber no air is entering the working side of the cylinder. Hence
26
THE HEAT-ENGINE PROBLEM.
the heating of the air during this stage must take place at de-
creasing volume and increasing pressure. When this is finished,
the valve V opens and gases enter on the impelling side of the
piston. The tendency now is to decrease pressure, but air in
the space G l will pass the fire, tending to uphold pressure ; hence
there is a second heating at uncertain pressure, giving an inde-
terminate cycle.
25. Lucien Genty in 1893 proposed the engine of Figs. 26-28.
This engine has the usual single-acting cylinder with elongated
piston. Air is drawn from a chamber under the floor to obviate
the noise of suction. The air thus received by the water-jacketed
compressor 10 is sent through the valves 17 into the hollow
bed 4, and thence past the ribbed-coil preheater, here to be
warmed by the exhaust. The preheater is arranged to take up
expansion, and exhaust gases are prevented from touching the
metal by brick lining. The partly heated air enters admission
valve 36 by pipe 35. Yalve 36 is double balanced, and held to
its seat by a helical spring, and is actuated by a cam on the
horizontal shaft.
THE HEAT-ENGINE PROBLEM.
27
2G. The valve gear is arranged to govern by varying the open-
ing of valve 36. A weighted piston controller regulates the pro-
portion of air admitted to the two passages 41 and 42 after pass-
ing the admission valve 36. Thus the method of governing is
two-fold ; 1, by a variable cut-off to the power cylinder and com-
bustion chamber, and 2, by means of keeping the pressure of
the air admitted initially constant. Passage 42 leads above the
fire and 41 below.
FIG. 26.
The combustion chamber is a cast-iron funnel lined with
brick, enlarged at the bottom ; the coal is thus burned without a
grate. The lower and hottest part has a water-jacket, 49. Means
for charging and cleaning are provided, and a ball-and-socket
swivel, 57, carries a stirrer, 58, which is removable. After pass-
ing the admission valve and reaching the fire, the air is first
heated and afterwards expanded in the presence of the fire under
the piston. The motor cylinder is of two parts ; the lower, 72, is
lined with fire clay, the upper, 74, is bored true and water-jack-
eted. The lower part of the piston has air-cooling ribs, 76, and
the rod is rigidly connected to the yoke 78, leaving no lubri-
cated joint within the piston. The piston fits loosely in parts
28
THE HEAT-ENGINE PROBLEM.
75 and 78, and tight in part 74. Besides having the water-
jacket on this part it may be further protected from heat and
dust by a groove, feeding air down to the combustion chamber ;
to do this air is provided by a small pump on the piston at 89.
A flexible pipe provides water to cool the piston interior.
27. After being heated and expanded the gases are discharged
through the self- cleaning valve, 110. This valve is hollow and
water-cooled, as is also its casing, 116. Compressed air seats
the valve, and mechanism raises it. Injury from sparks, etc., is
FIG. 28.
prevented by drawing the valve up in its sleeve casing. To pro-
tect the seat it is made narrow and cleaned by escaping air.
Exhaust gases are discharged through the regenerator, or pre-
heater. This machine, though very complicated, presents many
suggestions that may be of value.
It will be observed that the combustion of coal calls for a
great complexity of parts and functions, and this must be so.
We have, as one of the greatest troubles, the impossibility of
regulating the heating power with time of a coal fire, and there
is the inevitable dust and ash. This makes the use of oil and
THE HEAT-ENGINE PROBLEM.
29
gas with the necessarily simplified apparatus almost mandatory
in internal-combustion engines, especially those of moderate size.
While here we should have none of the coal troubles, we will
meet others incidental to the feeding of fuel as it is required,
and only for the instant that it is required.
Oil-burning Engines.
28. In 1865 Stephen Wilcox, Jr., proposed an oil burning
engine, shown in Fig. 29. He said, " The extraordinary devel-
FIG. 29.
poment of what is known as petroleum oil and the several prod-
ucts obtained therefrom, makes it practicable to produce and
work very small engines on this plan."
A, the working cylinder, and a 1 , the pump, differ in no part
from what we have seen in the coal-burning engines. The
means for feeding and burning the oil are, however, new and of
especial interest. Fuel is fed from an elevated tank 1 through
pipe 2 ; the interior pressure of the furnace is balanced on the
surface of the oil by pipe 3. A stop-cock, 4, is provided, which
is arranged to shut off oil by a piston and links, when the
furnace pressure exceeds what is desired ; this serves to govern
30
THE HEAT-ENGINE PROBLEM.
the machine. Oil flows into a vaporizing reservoir, 12 (Fig.
33), having Ings, 13, to conduct heat from above. The upper
part of this reservoir is cylindrical and provided with holes,
10, and fitted with a loose part, 11, with inclined top and a hole
in the centre matching the movable pin, 14, actuated by the
governor. This pin and hole act to close the vapor outlet. The
special construction of the burner is intended to give a constant
velocity of efflux to the vapor.
This combustion may be classed as the burning of a jefc of
vapor in an atmosphere of air, the air about the flame being
kept fresh only by convection.
Governing is effected by a double means ; the fly-balls act to
shut off vapor and increase of pressure cuts off the fuel.
29. A. H. De Villeneure in 1872 proposed an engine in which
lie provided a combustion chamber, having a platinum rose, p
(Fig. 30), on which impinge jets of oil vapor from b and air from
m provided by pumps. He thus expects to obtain a combustion
to heat the mixture.
George B. Brayton in 1872 proposed and built an engine that
was very complete and fairly successful. Fig. 31 is a general
view and Fig. 32 his oil burner.
Air is compressed in the single-acting pump, which has a vol-
ume one-half that of the power cylinder. The compressed air
passes from the constant-pressure receiver through pipe 1) and
over the absorbent material e, through which the fuel is fed by
a pump. Here it takes up vapor and the mixture passes the
wire-gauge grating and into the cylinder, where it burns.
THE HEAT-ENGINE PROBLEM.
31
Means are provided to prevent entirely shutting off the air from
the power cylinder, and thus there is kept constantly burning a
small flame which increases for the power stroke. Governing
is effected by a variable cut-off to the power cylinder.
The power cylinder is water-jacketed, and no trouble is ex-
perienced through overheating. A safety valve is placed on
the reservoir.
32
THE HEAT-ENGINE PROBLEM.
30. Joseph Hirsch, in 1874, proposed the engine shown in
Fig. 33. An air pump, L, supplies air to a cylinder, //, to take
the place of converted air which is periodically expelled ; J is a
regenerator connecting the two cylinders, A and /// N is a
water chamber for cooling gases. Fuel is injected at I>' 2 . When
the piston C moves down, air from G is sent over through the
regenerator, partly heated here and then further heated by
fuel at B\ On the up-stroke the products of combustion pass
over the regenerator to //, being thus doubly cooled, first by
the regenerator and second by the injected water. Part of the
products of combustion escape at k and are replaced by fresh
air. No means for ensuring the combustion of fuel in the
atmosphere of air and products of combustion are provided.
FIG. 33.
31. Stephen Wilcox, in 1885, proposed the engine of Figs. 34
and 35. The power cylinder, A, is double-acting and tande re-
connected to an air compressor, B. Plates of non-conducting
material are applied to piston and cylinder heads to keep the
gases admitted as hot as possible. Cylinder walls, where lubri-
cation is necessary, are water-jacketed, and water-cooled rocking
valves of the Corliss type are provided ; air is sent to the exhaust
preheater C from the pump before passing to the power cylinder,
and is, therefore, first warmed and later heated by combustion,
in its passage over the gauze grating, 8, where it meets the
liquid fuel. The valve mechanism permits varying the cut-off
and reversing.
A reciprocating tube carrying a lamp works in each port r
and serves as an igniter in connection with an exterior relighter.
THE HEAT-ENGINE PROBLEM.
33
The receiver is first charged with air by a hand-pump, and a
little fuel sent to the burner by hand. Valve 24 in the escape
pipe 22 is closed and a torch applied to the burner through w-.
The engine may now be started by opening the main stop-valve,
28, and automatic action begins.
32. Kudolph Diesel, in 1892, proposed the oil-burning motor
shown in Fig. 36. A single-acting cylinder C carries the plunger
FIG. 34. | T
P, air-valve F, and fuel-valve D. High compression of the air
is followed by fuel injection and later by expansion. The tem-
perature developed by the compression must be sufficient to
ignite any fuel thrown into the air. Later engines vary some-
what in detail, but the principle of operation is the same.
Gaseous and powdered solid fuel can also be used.
It is obvious that the quantity of fuel injected per stroke will
determine the cycle. If only sufficient isjfadmitted to keep T
3
34 THE HEAT-ENGINE PEOBLEM.
constant, we have Cycle IV., or some of its variations ; if enough
heat 'results to keep p constant during combustion, we have
one of the Cycles III.
33. B. V. Nordberg and C. E. Shadall proposed in 1895, not
a complete engine, but a system of operating engines by internal
combustion ; the apparatus is shown in Fig. 37. The products
of combustion are to be used in any way deemed advisable.
FIG. 35.
A source of airjsupply and means for using the products are
assumed. Oil is fed from a tank B by water displacement, the
same water-jacketing the combustion chamber C. This com-
bustion-chamber jacket may add steam to the products at c^
The air current from the compressor acts (1) on the surface of
the water A ; (2) at the oil atomizer E; (3) at the lamp /; and
(4) with reduced pressure atjhe_opening c in the burner. The
THE HEAT-ENGINE PROBLEM.
35
atomizer is fed with oil through the pipe b, and the float 7>, 3 pre-
vents, by proper specific gravity, an overflow of water. The oil
passes up to the nozzle, where it meets an air current and is
there sprayed into the chamber C 1 ; the spray meeting a flame
jet from the lamp at i in an atmosphere of air provided through
c, is enabled to burn. Increased pressure cuts off the oil supply
Oil
FIG. 36.
Gas-burning Engines.
31 The operation of engines of the class we>re considering
by a gas combustion offers what would seem tcvbe the simplest
solution, but in reality the difficulty is much greater than might
be supposed. Of course, no trouble will be ^experienced with
dust, vaporizing of oil, or soot from imperfect oil combustion,
but there appears the difficulty of finding a burner which will
36
THE HEAT-ENGINE PROBLEM.
THE HEAT-ENGINE PROBLEM.
37
completely consume the gas without excess of oxygen under the
widest possible range of conditions. The engines proposed
differ chiefly in the means proposed for accomplishing this gas
combustion.
Stephen Wilcox, in 1865, proposed the machine of Fig. 38.
Air and gas are supplied through pipes K and J from feed-
pumps G and If to burner j, and the mixture is ignitejd at R.
When the gas is to be supplied by vaporizing oil, the exhaust-
heated vaporizer N is introduced. A is the power cylinder, B
a changing cylinder, and F a regenerator. When b descends, the
valve M allows cold air to fill the top of B ; this air, on the up-
Fio. 38.
stroke, moves through the regenerator and burner, which at
this time delivers its mixture ; this double heating raises the
pressure in the system, and the piston rises. Then both de-
scend and exhaust some gases into P. The engine is thus oper- '
ated not solely through the heating of products of combustion,
but by the heating of a mixture of these gases with pure air.
35. Albert Schmid and J. C. Beckfeld proposed, in 1889, the
system of obtaining hot gases through internal combustion of
air and gas by the apparatus of Fig. 39. Gas and air are sup-
plied by pumps to tanks A and G ; they are mingled in an
injector chamber $ 2 ; thence passing through the perforations t,
they are burned at 8. To maintain a difference of pressure in
38
THE HEAT-ENGINE PROBLEM.
the tanks and combustion chamber a relief valve 0, controlled
by a diaphragm, is provided. To dilute the products of com-
bustion and reduce their temperature, a pipe L conducts fresh
air to the mixer S l . Electric ignition is suggested.
Later, another arrangement, shown in Fig. 40, was suggested.
Here a long perforated brick is inserted to aid combustion
and act^s a reigniter. A receiving reservoir is added, to which
the blow-off is attached. An igniting plug V of coke or carbon
is also added.
In Fig. 41 is shown an addition of a steam boiler with an ex-
FlG
haust gas feed^water heater. The boiler is to be used in start-
ing with an ordinary external combustion fire-box, and, later,
enclosed, mixing the steam and products of combustion.
36. Ilerman Schumm, in 1895, suggested the engine of Figs.
42 and 43, which offers some novelty. A is the engine cylinder,
B the piston, C an inlet for combustible mixture, D an inlet for
pure air, and E the*, exhaust valve. An electric igniter, i, is
provided, and a gauze diaphragm, g, prevents back flash. Air
is compressed by the pump and stored in 6r, gas similarly com-
pressed by H. Air is admitted through J) until the piston has
moved out a short distance and then cut off; at the same time
THE HEAT-ENGINE PROBLEM.
39
the air and gas mixture is admitted through C and ignited, the
combustion operating on the gauze until the mixture F ^is in turn
shut off, when adiabatic expansion begins.
37. Sydney A. Eeeve, in 1897, proposed the apparatus of Fig.
FIG. 40.
44 to obtain by internal combustion working gases to be
used in an engine. He lays stress on two points : one, re-
lating to the combustion, that the proportions of fuel to air
shall not materially vary ; and the other, the reduction of the
FIG. 41.
temperature before sending the products of combustion to the
engine.
Both air and fuel are to be supplied by separate pumps,
and the proportions regulated by maintaining the pressures in
40
THE HEAT-ENGINE PROBLEM.
the two receivers, C and C l , equal by water seal and float valve,
and by passing these gases of equal pressure through a double-
ported valve D of proper areas. The pressure in C is controlled
by that in C 1 , and that in C kept above that in the combustion
chamber by the loaded check G. This also permits mixing
fresh air with the "products of combustion if more than is
wanted for combustion is available.
The products of combustion pass through water so supplied as
to keep a given quantity always on hand and at the boiling point
FIG. 42.
corresponding to the pressure, so that the hot gases will, in part-
ing with their heat, evaporate water and be themselves cooled to
the temperature of saturated steam at their pressure. Of course,
any feed-water supplied must be heated before evaporation, but
this only has the effect of decreasing the rate of evaporation
without stopping it.
Another device proposed for equalizing pressures in the air
and gas receivers is to let fuel pass through a flexible-walled
bag suspended in the air tank.
38. The cooler and burner might also be arranged as shown
THE HEAT-ENGINE PROBLEM.
in Fig. 45. The regulator valve has a spring-balanced dia-
phragm, 4, which actuates a similar plunger, 6, by rod 5, the
plunger moving in a perforated sleeve. Gas enters through tube
12 in the centre and air through 10, the mixture passing an
igniter at b. A liquid seal is provided here for maintaining a
decrease of pressure between supply and discharge valves.
39. Lucius T. Gibbs, in 1897, proposed a system (Fig. 46), in
which the motive power shall be air admitted to the power
cylinder from a source under pressure, and when the pressure
after cut-off has become so far reduced as to reach that of a
FIG. 44.
stored mass of gas maintained lower than the air supply, the
gas will enter and be ignited, thus tending to keep up the pres.
sure during the expansion. Thus the adiabatic will be raised
to perhaps an isothermal or higher.
40. If we would trace any line of progress through these
machines, we could not make our division according to the fuel
used, as at times engines burning all kinds of fuels have been
suggested simultaneously. To be sure, before the possibilities
of petroleum were known, the principal fuel was coal, and,
naturally, in early engines coal fires predominate ; though it
would seem that they would gladly be dropped for oil and gas,
yet they were not, and continue to appear from time to time
42
THE HEAT-ENGINE PROBLEM.
There is a division, however, in the stages of progress that is
significant, as showing how strong is the influence of the known
and tried on the proposals of apparent novelty. The old so-
called hot-air engines of the Ericsson type had a' mass of air
FIG. 45.
enclosed, and means were provided for heating and cooling the
same mass without exhausting, the heat, of course, being sup-
plied through walls from an outside fire. So we find the earlier
internal-combustion engines working on a system only slightly
different from the above. There is a mass of gases enclosed,
THE HEAT-ENGINE PROBLEM. 43
and means for transferring them from a hot part to a cold part,
and so varying their internal pressure, but the hot part is here
provided for not by a hot plate, but by a short flame injection,
or passage over a fire. The operation was to depend chiefly on
the alternation of hot and cold in the same mass.
41. Later this system was developed by injecting more and
more fuel, or by causing the mass to pass entirely through the
fire, necessitating more fresh air for the next time and calling
for an exhaust ; finally we have a regular admission and exhaust,
the gases passing continuously in the same direction, and no
alternation of heating and cooling being attempted in the sys-
tem. Any cooling that is to to take place must occur 'outside
FIG. 46.
the machine in the atmosphere, and the resulting contraction
of the gases forms no part of the working cycle.
42. The engines presented represent by no means all of those
suggested, but are selected from a very large number to show
the principal ideas advanced in the past. By studying them we
can reach an understanding of what points may be accepted as
solved, and what are still open for discussion and research.
It will be observed that these non-explosive internal-combus-
tion engines may be divided into (I.) intermittent internal com-
bustion, as in Genty's, Brayton's, Wilcox's, Diesel's, etc., in
which the air and fuel pass the power- cylinder valves before
combustion ; and (II.) continuous internal combustion, as sug-
gested by Schmid and Beckfeld, Reeve, and Nordberg and
Shadall, in which the hot gases are continuously produced, and
afterward utilized by passing through valves to the power
cylinder, or by expanding through a turbine without valves.
44 THE HEAT-ENGINE PROBLEM.
43. Different ways of doing the same thing by varying details
have been advanced, and it may be well to bring these together
in some cases for comparison.
Cylinders :
Both single and double acting, jacketed and unjacketed are
found. Air and power cylinders may be independent, or the
two operations of compression and expansion performed in
opposite ends of the same cylinder. When this is done, the
cylinder may be of the same diameter throughout, or reduced
at one end. Independent cylinders have been connected by
beam, separate cranks on the same shaft, and in tandem. In
most of the intermittent-combustion type the piston fits a part
of the cylinder only, and is loose in the hot end ; in this case,
heat is prevented from reaching the working faces by prolong-
ing the piston, and by blowing cold air or steam down around
the loose part toward the hot end.
44. Igniters :
With engines using oil or gas, where combustion may be
entirely interrupted for a time, some form of relighter must be
provided. This may be
I. A plate heated by an external jet from outside. Nordberg.
II. A platinum rose, with mixture impinging. Yilleneure.
III. High temperature of compression. Diesel.
IV. A plate or wire electrically heated. Wilcox.
Y. Introduction of incandescent solid. Schmid and Beckfeld.
VI. Electrodes with spark gap. Babcock.
VII. Introduction of auxiliary flame continuously. Nordberg
and Shadall.
VIII. Introduction of auxiliary flame intermittently. Wilcox.
IX. Constant-burning flame in cylinder. Brayton.
45. Use of water :
I. In compression to cool air.
II. To be evaporated by products of combustion by contact
and adding steam produced.
III. To cool expanded gases and produce pa.rtial vacuum.
IV. To cool hot parts and be discharged.
V. To cool hot parts and add steam.
VI. In an internally fired steam boiler, which may act as a
starter.
VII. As a jacket to hot parts, the steam used in a separate
cylinder.
THE HEAT-ENGINE PROBLEM. 45
VIII. As an annular piston cooler to prevent contact of hot
gases with working face of cylinder.
46. Methods of governing ;
I. Throttling air intake.
II. Throttling passage between the air and power cylinders.
III. Varying cut-off at power cylinder inlet.
IV. Varying combustion chamber pressure by fuel cut-off.
V. Varying internal pressure, by splitting the air from the
compressor and sending more or less of it through or around
the fire.
. VI. Varying internal pressure by blow-off.
VII. Combination of varying initial pressure to power cyl-
inder, and varying cut-off to same.
VIII. By throttling the exhaust.
47. Preheating and regeneration :
A heating of the air after compression, and before reaching
the fire, will insure a good combustion, and if this is done by
exhausted heat, it aids the economy of the engine.
I. Causing the air approaching the fire to first surround the
fire-box as a jacket.
II. Causing exhaust and compressed air to pass on opposite
sides of a plate.
III. Alternately sending fresh air and exhaust through the
same chamber, which must then be present in duplicate.
IV. Combination of exhaust regenerator and fresh air fire-box
jacket.
48. Fuel feed :
Coal may be fed by any automatic stoker, and from above,
below, or the side ; we will omit these.
I. Oil by gravity, cut off by internal pressure.
II. Oil by pump, cut off by by-pass.
III. Oil by water displacement, cut off by internal pressure.
IV. Oil by injector suction.
V. Gas by pump.
49. Gas and oil burners :
I. Air and gas fed from opposite openings, jets impinging.
Wilcox, Reeve.
II. Air and gas fed through separate openings, the jets mix-
ing by impact with platinum rose. Villeneure.
III. Air and gas mixed, fed through simple opening. Bab-
cock.
46 THE HEAT-ENGINE PROBLEM.
IV. Air and gas mixed, fed through gauze screen. Brayton,
Schumm.
Y. Air and gas mixed by injector at burner, passing through
openings into larger chamber, and products through openings
similar to first, but larger in area. Schrnid and Beckfeld.
VI. Volatile oil vaporized by heat of burner burning in
atmosphere of air. "Wilcox.
VII. Volatile oil vaporized by falling on hot plate in atmos-
phere of air. Schmid.
VIII. All kinds of oil sprayed into atmosphere of air.
Nordberg and Shadall.
IX. Volatile oil dropped on grating, where it meets air, the
mixture burning on one side. Brayton, Wilcox.
X. Hot-air atmosphere for an injection of fuel. Diesel.
50. Mixers and proportioners :
I. Mix at burner, proportions maintained only by pump of
proper size.
II. Pass air through volatile oil kept at fixed temperature by
exhaust.
III. Mix at burner, proportions maintained by pumps of
proper size, aided by a device to feed both pumps at constant
pressure.
IV. Proportions maintained by (a) movable partition between
air and gas receivers, to keep pressures equal, and (b) proper
double valve at the discharge.
V. Mix at injector, with no special device, for proportioning.
Position of fire :
I. Directly under piston.
II. At side of cylinder; each end, when double acting.
III. In separate, continuously operated, highly heated cham-
ber.
Cycles :
Nearly every cycle of operations can be followed. Engines
that take air in one side of the cylinder and send it around to
the other side, and those with two pistons moving in different
directions, causing a change of position of the air only, will give
a constant-volume combustion, and whether the engine follows,
Cycle II., HA., B, or C will depend on the amount of expansion
permitted. Engines feeding the cylinder with a flame may have
a constant-temperature, or constant -pressure combustion, de-
pending on the fuel used ; hence III. or IV., with all their varia-
THE HEAT-ENGINE PROBLEM. 47
tions, are possible. Those engines that maintain a constant
pressure in the combustion chamber will follow III., or some of
its variations which one, will be determined largely by the cut-
off. A turbine system would follow III. most nearly.
51. Of the engines considered, and the much larger number
not mentioned, not one, except the Diesel, is on the market to-
day. This, with some, would be a sufficient argument to con-
demn the whole system, but a little study will show that the
trouble is nearly always in the same place, and a little persever-
ance is all that is necessary to remedy the defect. Much more
difficult problems than this have been solved successfully, but
there was first necessary a recognition of the trouble and a good
reason for spending time and money to overcome it.
The compressing of air is no new problem, and the using of
hot gases is done every day in thousands of horse-power of ex-
plosion engines, so that these two parts of the engine may be
considered solved, leaving, as the only doubtful essential, the
fire, and here is the seat of most of the trouble.
52. The old engines, like Cayley's, using coal, were found to
cut with ash, etc.; but, nevertheless, when everything was right,
they ran and gave a good account of themselves. In the natural
order of things, Brayton appeared with a gas machine. The air
and gas were mixed in proper and explosive proportions, at the
compression intake, and sent through a wire gauze grating to be
burned in the cylinder. Clerk says of it : " The engine worked
well and smoothly ; the action of the flame in the cylinder could
not be distinguished from that of steam, it was much within
control and produced diagrams similar to steam." The flame
grating of gauze was the weak part, as an accidental piercing or
overheating caused an immediate back flash and stopped the
engine. Brayton could not stop this, so he tried a volatile oil
with compressed pure air, but his burner was very crude and
resulted in a goodly soot deposit. The case seemed hopeless,
and doubly so when Otto appeared in the field with his success-
ful engine, so Brayton gave up.
Here, however, was a working engine giving good results, both
in economy and regulation, needing only a good burner to keep
it going.
53. The immediate success and attractive principle of the
simple one-cylinder Otto has held the attention of nearly all
from then until to-day. One man there was who, not only did
48 THE HEAT-ENGINE PROBLEM.
turn aside, but, having turned, persevered, and he was rewarded
by success that was Diesel. He prepared an elaborate plan
to imitate as nearly as possible the Carnot Cycle, with its iso-
thermal combustion in a cylinder certainly a striking novelty.
But he did not follow it, as the low mean effective pressure of all
the Cycles IV., which he attempted to follow, necessitated immense
machines for the power produced ; what he did do was to repro-
duce the Brayton engine with another burner and igniter. His
hot compressed air did what Brayton could not do, but in every-
thing else he was strictly Brayton with his Cycle III. of opera-
tions, which lie ultimately followed. This is one solution, then,
but not necessarily the best, as Diesel needs a very high
compression to run, and, while this is the reason for his high
efficiency, it makes a heavy machine. A little lower efficiency,
with less weight, would be very acceptable, but this would pre-
clude the Diesel burner.
54. A detailed study of the^combustion of gas and oil should
certainly lead to a still further opening up of this promising,
though neglected, field of engineering. Produce a good fire and
you must inevitably produce an improved Brayton engine, and
this, in view of what has been said, is certainly a very desirable
end.
While combustion, as a purely chemical process, has formed
the subject of numerous papers and researches, leading to most
interesting theoretical results of great value to physical chemists,
not so many have resulted in the discovery of a new or useful
mode of burning fuels.
55. Investigation long ago showed that the oils undergo a
vaporization before combustion, and that the oil flame is really
an oil vapor, or gas flame, so that a knowledge of the laws of
combustion of gases will give us those of oil combustion, with
the exception of the means of previously vaporizing the oil. In
fact, the different methods of burning oil now in use vary chiefly
in this second respect, the means provided for gasifying the oil.
It would, therefore, be well to look at the question of gas com-
bustion first.
Gases burn by combining chemically atjhigh temperature with
oxygen,*and the study of their combustion may be most readily
divided into classes whose characteristics are the ways in which
this coming together of the gas and its oxygen^are provided for.
56. This leads to the division :
THE HEAT-ENGINE PKOBLEM. 49
Class I. Gas issuing from an orifice into a supporting at-
mosphere and where all the oxygen for combustion is derived
from that atmosphere.
Class II. Gas mixed with oxygen insufficient in quantity for
its combustion or for the formation of an explosive mixture,
issuing into a supporting medium from which all necessary
additional oxygen is derived.
Class III. Gas mixed with oxygen in quantities insufficient
for complete combustion, but sufficient for the formation of an
explosive mixture, issuing from an orifice into a supporting at-
mosphere, from which all necessary additional oxygen is to be
derived.
Class IV. Gas mixed with oxygen in just sufficient quantities
for combustion, issuing from an orifice into any sort of atmos-
phere. We shall call a mixture of this sort a "chemical"
mixture.
Class V. Gas mixed with oxygen in such quantities as to
form an explosive mixture, but with insufficient oxygen for com-
plete combustion, burned in a mass by a single explosion.
Class VI. Gas mixed with oxygen in chemical proportions,
burned by a single explosion in mass.
57. The first class of combustion is very imperfect, conse-
quently only low temperatures result, while large excesses of
oxygen over what is chemically necessary are required. It is
to this very imperfection that we owe the efficiency of our ordi-
nary gas jet as a source of light. The unequal distances trav-
elled by molecules of gas before reaching the place where they
can find and combine with the necessary oxygen, gives the flame
a volume; i.e., a certain portion of space is filled with the flame.
In the study of combustion, as the origin of heat, this class is of
no importance. Mixing the oxygen with the gas, previously to
heating for ignition, as in Class II., is a direct aid to nature,
eliminating the hunting process of Class I., or, at any rate, re-
ducing it, and making necessary only the heating to the igni-
tion temperature to cause combustion. This is shown in the
immediate shortening of the flame over that of the previous
class, and its loss of luminosity, while still retaining the vol-
ume character of the flame. It is the principle of the Bunsen
burner, and the large class that follow it for use in furnaces,
heaters, cooking stoves, and heating water in steam carriages.
58. In most of these the mixture of air with the fuel is made
50 THE HEAT-ENGINE PROBLEM.
by causing the jet of gas to impinge on a mass of air, some of
which is carried along with the air under the double influence
of gas friction and the heated top of the burner, whence the
mixture issues.
Some other systems, of which the American Gas Furnace
System is one, effect the mixture in closed chambers before exit
at the burner.
Combustion of Class II. is characterized by the fact that
there is an actual volume of flame ; the flame is hotter than in
Class I., which means that for a given flame volume, either more
gas is burned, or the products of combustion are less diluted ;
the flame is less luminous and not of uniform color throughout
its volume.
An infinite variety of details of arrangement in the exit and
mixing of the air and gas may be devised with varying results
for special cases, but we may say of all of them that though the
combustion be very perfect and the amount of heat generated
large, yet there is always a " flame volume," indicating a strug-
gle, as it were, on the part of the gas and air in their final com-
bustion. The combustion, though approaching perfection in
many cases, is rendered so only by the use of a large excess
of the oxygen chemically required giving oxidizing products of
combustion.
59. It is only when we previously mix the gas and air com-
pletely and uniformly in the proper chemical proportions that
we can get non-reducing, non-oxidizing products of combustion,
and, since none of the heat goes to warm excesses of oxygen or
of fuel, the temperature of these products must be the highest
possible. Combustion of this sort is flameless, or, rather, what
flame there is is without volume, having only length and
breadth without thickness, and is, in fact, a surface.
Such combustion is governed by laws quite different from
those under which the classes already noted operate, and it is
to the combustion of chemical, and other explosive mixtures,
that tliis section is mainly devoted.
60. Let us consider first the class mentioned as Class VI., in
which a mass of chemical mixture i.e., gas and its needed
oxygen is confined in a chamber. If inflammation be provoked
at any point of the mass, it will, by self-propagation, finally and
successively inflame the whole mass. This is the first and
fundamental principle of this sorfc of combustion. The investi-
THE HEAT-ENGINE PROBLEM. 51
gation of this propagation of inflammation by such men as Davy,
Bunsen, Mallard and LeChatelier, Berthelot and others, has
shown that :
(a) In any mixture, the rate of propagation is constant for a
given temperature before inflammation.
(b) The rate of propagation for such mixtures varies with
different combustibles, being, for example, very fast for hydro-
gen and slow for marsh gas.
(c) The rate of propagation increases with the temperature of
the mixture before inflammation.
-(d) The combustion is visible by reason of a flame-cap, or
deep blue film of flame, which travels through the mass, and
which, at any instant, completely separates all the burned from
the unburned mixture.
This uniformity of velocity of inflammation would indicate
that in a mass where inflammation had started at a point, the
flame-cap, or surface of combustion, exists at any instant on the
surface of a sphere whose radius is proportional to the time
elapsed.
61. All this has been assumed to take place in a large mass of
gas. If, however, the enclosing vessel be given special forms,
certain other characteristics are brought out. One which is of
interest to us is the fact that, when the enclosing vessel is a cyl-
inder, or prism, in which the combustion surface travels with its
centre on the axis, the velocity becomes affected by reduction of
cross-section and that there will always exist for every such mix-
ture an area of cross-section so small that the self-propagation
ceases. This has been explained by saying that the walls car-
ried off heat so fast that the small flame-cap could not generate
heat enough to keep itself above the temperature of ignition.
Davy secured the same effect by using his screen of wire gauze,
which, if interposed in the path of the combustion surface, in-
stantly cooled the same sufficiently to prevent the ignition of
the mixture on the other side, provided, of course, the tempera-
ture of the gauze itself is sufficiently low.
62. When a neutral diluent gas, such as N or CO 2 , is added
to a chemical mixture arranged for the above-discussed com- .
bustion, its effect is to reduce the rate of propagation, though
not in conformity with any law yet discovered. Of course, there
will be a point when so much of the neutral gas is present
that combustion is impossible, but no reliable data are at hand
52 THE HEAT-ENGINE PROBLEM.
on this point, as the same conditions often give widely varying
results.
While large quantities of a neutral gas may be added, with-
out affecting the combustion except to decrease the rate of prop-
agation, a dilution by a comparatively slight amount ( f oxygen
will prevent it altogether. An excess of ga , it has been found,
will act within certain limits like the presence of a neutral gas.
By far larger amounts of fuel than of oxygen may be present in
excess without arresting combustion.
63. This brings us to class V., where explosive mixtures are
burned in mass, the mixtures having excess of fuel. The com-
bustion is possible within quite wide limits, with no other effect
than varying the rate of propagation. In fact, we see a great
deal of it to-day in our gas engines. While, of course, we should,
in these engines, invariably use the proper chemical mixture, they
are seldom, if ever, constructed to maintain this properly, and, as
a slight excess of oxygen will completely prevent inflammation,
the error is always made on the other side ; sooty exhausts bear
testimony to this. The gas engine also gives evidence of the fact
that neutral gases decrease the rate of propagation, for in some
two-cycle engines which I have lately examined I find it im-
possible to get a vertical combustion line on the indicator dia-
gram with a fixed ignition, except at very slow speeds about 60
revolutions per minute. This is due entirely to the presence
of exhaust gases in excessive quantities as diluents to the charge.
Some of the principles above noted as belonging to masses of
mixture at rest will make clearer the nature of the problem of
combustion of the same mixtures when in motion issuing from
an orifice. x
64. The desirability of being able to burn an explosive mix-
ture continuously and non-explosively under commercial, rather
than laboratory, conditions having long been obvious, a series of
experiments was undertaken at Columbia University with this
end in view. Many experiments were made and various results
obtained, but as a full account would take too much space and
avail little, only a few characteristic experiments will be noted
as leading up to the result. Consider a mass of explosive mix-
ture passing through a non-conducting tube with a uniform ve-
locity v. Then, if inflammation be started at some point, the sur-
face of combustion may remain at rest or move with or against
the current. J>enate .the rate of propagation by r. Then, when
THE HEAT-ENGINE PEOBLEM. 53
v >' r the surface of combustion will move with the current, and
if the tube has an end, the flame will " blow off" and combustion
cease ; if v r the surface of combustion will remain at rest,
other influences being inoperative ; if v < r, the surface of com-
bustion will move back toward the source or " back flash."
Of course, a small tube of heat-conducting material will exert
considerable cooling effect, but for the present we will not con-
sider such tubes.
In a practicable system of burning an explosive mixture con-
tinuously, we may state the following as desiderata and later see
ht>w they can be secured.
I. "Back flashing ' must be prevented.
II. "Blow off" must be prevented.
III. Combustion surface must be localized.
IV. It must remain localized for wide ranges of feed or veloc-
ity of flow of the mixture.
V. The localization must be unaffected by changes of tem-
perature.
VI. Large or small quantities must be burned without affect-
ing the above, and the transition from very small quantities to
very large, or vice versa, however sudden, should be easy.
65. The first requirement might be accomplished in three
ways :
(ft) By using a long tube of some conducting material and
so small in diameter as to prevent the passage of the flame-cap
under any circumstances.
(b) By using wire gauze screens.
(a) By causing the mixture to flow at some point with a
velocity always greater than the rate of propagation.
The first (a) is impracticable, as it permits of only small quan-
tities being burned ; the second (b) will not work when the wire
gauze gets hot ; this leaves (), which is practicable, as a valve in
a pipe will answer for the necessary contraction and consequent
increase of velocity. Hence we must put down as the first re-
quirement in our desired method of combustion the following.
At some point before the combustion surface is reached the ve-
locity of feed must be such that v > r.
66. Requirement II. might be accomplished in three ways : -
(a) By so reducing the velocity after passing the high-speed
point that we have at some surface v = r.
(b) By suddenly increasing the temperature of the mixture so
54 THE HEAT-ENGINE PltOBLEM.
as to increase the rate of propagation while v remains con-
stant ; or,
(c) By both reducing v, by spreading the current, and in-
creasing r by heating.
All of these ways are practicable ; but, as a reduction of
velocity alone, or a sufficient heating alone would not produce
the desired results so well as both operating together, there was
introduced as the second requirement in our desired method,
the following. After passing the point where v > r, the ve-
locity of the mixture should be so reduced and its temperature
increased as to make v l = r 1 .
67. With these conditions in mind, let us consider an experi-
ment. Let the mixture issue from an orifice into the air. By
properly regulating the velocity of exit, the flame-cap can be
maintained at the orifice the only device with which I suc-
ceeded in this experiment was by causing water to drip into the
supply chamber ; the position of the flame-cap is so extremely
sensitive to changes of flow that all other methods which were
tried for obtaining a constant velocity of exit, variable at will,
failed increase the flow slightly, and the flame-cap will lift off.
This may be done until the flame-cap is as much as 2 inches
(with illuminating gas and air) from the orifice before extinction
takes place. It would seem that the impinging of the jet on the
atmosphere should spread it and so reduce its velocity, but no
appreciable increase of diameter could be observed. When the
cap is close to the orifice, it is of a deep blue color, uniform in
shade over the disk, and the edges are sharply defined ; whereas,
as it lifts off some distance, it becomes indistinct and unsteady
at the edges until, at the moment of extinction, it fades into
nothing. When the cap is away from the orifice, while there
is no visible connection with the source of supply, there really
exists a column of mixture extending from the orifice to the cap
and passing through the atmosphere. Naturally, at the surface
of this column, diffusion will take place, and the longer the
column, the greater will be this diffusion effect, thus affecting
the composition of the advancing column of mixture and caus-
ing partial loss of gas. This is the real cause of extinction.
68. From these experiments we can draw the conclusion that
the current cannot be sufficiently reduced in velocity by issuing
into an atmosphere of lower pressure to prevent " blow-off "
before diffusion with the atmosphere so alters the character of
THE HEAT-ENGINE PROBLEM.
55
the mixture as to cause extinction before reaching the surface
of combustion. This calls for a new condition besides those
noted in the requirements for combustion. The reduction of
velocity of the mixture, after passing the place where v > r,
must be accomplished in such a way as to prevent diffusion
with any other gas.
69. To prevent this diffusion, there naturally suggests itself
the expedient of surrounding the issuing jet with a shield of
larger diameter, to still permit of the desired expansion. This
is shown in Fig. 47, and is essentially the same as proposed by
Ladd, Schmid, Beckfeld, and others. The mixture must issue
from orifice a with a velocity v a > r ; this will prevent " back
flash." If the distance from a to b is long enough to allow the
gas to spread and reduce velocity, " blow-off " will not occur
until v b > r, and within these limits the flame-cap should re-
FIG. 47.
main within the shield. A trial shows that when (Dia) b is but
slightly larger than (Dia) a> the feed velocity may be varied in
about the proportions noted, but this means that we are confined
within very narrow working limits. The flame-cap seems to
lose its flat, volumeless character for some reason not at first
clear. When (Dia) b is much larger, say four or five times
(Dia) a i a slow increase of feed velocity above r reveals the ad-
vancing flame-cap just as if the shield were not there. Later, a
slight spreading is noted, and then the flame actually begins to
show volume, as if there were no longer an explosive mixture
present ; this heats up the shield. A little consideration will
show this to be due to the diffusion of the advancing and
slightly spreading column with the products of combustion
within the shield, and the high temperature of the shield helps
to maintain a combustion of what is now a diluted explosive
mixture beyond a point where that combustion would be pos-
sible if cold. An increase of velocity will cause extinction by
"blow-off."
56 THE HEAT-ENGINE PROBLEM.
70. Here the results are somewhat better than in the last
case without the shield. The principles operating, with the
results are : back flash prevented by sufficiently great initial
velocity at a ; a spreading to reduce velocity, but very slight
and insufficient, as proved by the narrow working limits ; diffu-
sion is not prevented ; gas is partly heated before burning by
the shield, which helps to continue combustion. If the advanc-
ing column did increase in cross-section and decrease in velocity
while advancing, successive possible, positions of the flame- cap
would be as shown at 1, 2, 3, 4, etc., of Fig. 47.
It is obvious that at any point between a and 7, such as 4,
the cap is surrounded by products of combustion, and the
advancing column of mixture is passing through an atmosphere
chiefly composed of the same, resulting in disastrous diffusion.
FIG. 48.
This at once suggests giving the shielding envelope the form of
a cone, supposing the orifice circular, so that the flame-cap at
any instant may entirely fill up the space between the walls.
71. Apparatus with this end in view was tried and gave some
interesting results. Fig. 48 shows a cone of 45 degrees angle,
with a ^-inch orifice such as was used. The velocity of feed was
so adjusted as to cause the flame-cap to advance slowly from a,
with the expectation stated above. The flame-caps at successive
positions took the forms shown at the lines 1, 2, 3, 4, 5, 6, etc.,
and finally "blow-off" occurred. Since the only place where the
combustion surface can remain at rest is on a surface where
v = r, and since, secondly, r is here constant, the curves indi-
cating the intersection of the combustion surfaces by meridian
planes, give us graphical values of the velocity of the advan-
cing column of mixture. It is seen that the expected spreading
did not take place, and that at any circular cross-section of the
THE HEAT-ENGINE PROBLEM. 57
cone, the velocity was greatest at the centre, decreasing toward
the edges.
The curves 1, 2, 3, etc., are really cross-sections of successive
constant velocity surfaces in the advancing column, and the
surface of combustion will lie on that surface of constant-trans-
mission velocity where v = r.
72. A constant-velocity surface may be defined as a surface at
every point of which the moving particles of gas have equal in-
stantaneous velocities. If these successive surfaces had remained
flat or nearly so, the proper sort of spreading of current and
uniform decrease of velocity would be indicated. This gives us
an accurate definition of how we want our velocity reduced after
passing the point where v > r. The velocity of the advancing
mixture must be reduced without diffusion, so as to keep the
surfaces of constant velocity of such form that adjacent points
on any one will be at approximately the same distance from the
point where spreading begins. Reducing the angle of the cone,
while helping matters considerably, reduces the range of feed
velocities within impracticable limits.
73. Many ways of bringing about the above were tried, but
only one seemed preeminently good both by reason of its
simplicity and effectiveness, for it fulfils almost perfectly the
requirements proposed for our desired method ; this is, to fill
the cone with fragments of refractory material such as pottery,
broken crucibles, bits of magnesite, or any other rock that will
stand the high temperature without fusing. In cones of 60
degrees, and with a ^-inch orifice, I have found pieces about
| inch diameter to answer well.
These separate pieces of solid matter interpose many reflect-
ing surfaces without materially hindering the advance of the
mixture, and cause it to spread in the way desired, keeping the
surface of combustion spherical and preventing diffusion. A
variation of velocity causes the spherical surface of combustion
to vary only in diameter, and the limits of feed are determined
only by the size of the cone.
74. A cone of given altitude will give the greatest range of
variation of diameter of cross-section when its angle is 180 de-
grees. This is a plane surface which, with the orifice and broken
rock, should appear as in Fig. 49. Here the surface of com-
bustion is approximately a semi-sphere. Trial shows that tbis
arrangement works perfectly, and the limits of feed are deter-
58
THE HEAT-ENGINE PROBLEM.
mined only by the size of the pile of rock surrounding the open-
ing. A cone of 360 degrees, or no cone at all, suggests the
surrounding of the nozzle by broken rock without any enclosing
walls (Fig. 50). This arrangement also works remarkably well.
FIG. 49.
The surface of combustion is here approximately a sphere,
giving the greatest possible increase in area of the surface of
combustion for the distance travelled from the nozzle.
If d denote the distance from the point where spreading
begins to the surface of combustion and S the area of the sur-
face, we have :
For a cone,
7id~ tan 2 of.
For no walls (Fig. 50), S l =
FIG. 50.
75. Not only is the greatest possible range of action by veloc-
ity reduction thus obtained, enabling the greatest possible
amount of mixture to be burned in a given volume, but this
THE HEAT-ENGINE PROBLEM. 59
amount is further augmented by reason of the increase of the
rate of propagation caused by the passage of the mixture
between the hot fragments. Hence both principles operate
simultaneously toward the desired end.
We have thus arrived at a method of continuously burning
explosive mixtures of all sorts, whether in the chemical propor-
tion or not, as classified in IY. and V.
76. The method fulfils all the conditions set down as neces-
sary, and may be stated as follows :
I. Cause the mixture to pass a point where its velocity of
transmission shall be always greater than the rate of propaga-
tion of inflammation through the mixture. This may be done
by a valve in the feed-pipe.
II. So spread the current of mixture after it passes this point
of high velocity that surfaces of constant-transmission velocity
shall be of such form as to keep adjacent points on any one at
approximately the same distance from the point where spread-
ing begins. The whole spreading must take place so that the
advancing unburned mixture cannot diffuse with any other gas.
This can be accomplished by surrounding the orifice with
solid fragments, introducing numerous reflecting surfaces which
accomplish the spreading; also, by the passage through the
interstices between this solid matter, the mixture is heated and
the rate of propagation increased, making possible the burning
of more mixture in unit volume.
77. When a chemical proportion is maintained in the mixture,
all the combustion takes place on the combustion surface, giving
absolutely neutral products of combustion ; but when an excess
of gas is present within certain limits, all gas that can find
oxygen burns explosively between the solids, while the excess
acts merely as a neutral diluent to be burned when it meets an
oxygen atmosphere later on. By properly placing the oxygen
atmosphere to burn the excess gas, we can get the hot products
either reducing or oxidizing reducing after leaving the explosive-
combustion surface and before meeting the excess of oxygen in
the atmosphere, oxidizing after that meeting.
It might be here noted that the principle well known in explo-
sive combustion at constant volume, and constantly operating
in the gas engine, that " to a chemical mixture of air and gas
there may be added large quantities of gas without altering
the explosive properties of the mixture," is, by these experi-
CO THE HEAT-ENGINE PROBLFM.
ments, extended. It appears that in explosive combustion at
constant pressure, or, as I have called it, " continuous combus-
tion of explosive mixtures," the same principle applies, and,
though no real proportion measurements have yet been made,
it seems to a wider degree. That is to say, that in the method
here described, mixtures of air and gas, with gas in excess of
the amount the air present can support, will burn explosively.
The excess gas present acts merely as a neutral diluent, such
as the nitrogen of the air. It is a fact also that, as the solid
fragments heat up, the excess may be greater than when they
are cold.
78. Another interesting thing noted in these experiments is
that an explosive fire will sometimes emit a musical note ; it
may be that this is always true and that its absence at any time
is due to lack of the proper resonator. This would seem to
indicate that what to the eye appears as continuous combustion,
is only approaching the limit, which is continuity, and that in
reality single explosions in rapid and reyular succession are
taking place. It would be interesting to determine whether
the temperature or kind of gas has any influence on this note.
79. The perfection of the gas combustion above discussed and
the simplicity of the apparatus make the method highly satis-
factory, and the solution of the difficult problem of explosive-
gas combustion lends encouragement to the even more diffi-
cult case of oil combustion. The experiments with oil, though
not yet complete, promise to give equally satisfactory results ;
in fact, it is almost certain that they will. However, the oil
system has so far been tried in only a few cases, and it is not
wise to announce the complete success of the system until all
possible conditions have been met.
80. It was shown that there were only two classes of com-
bustion worthy of consideration for use in internal-combustion
engines, and only two cycles that promised returns commen-
surate with the labor and time that might be expended in their
development the Otto and the Brayton. The Otto is simple
to carry out in practice, and is now, to all intents and purposes,
fully developed, while the Brayton has hitherto failed, chiefly
because of the difficulty of handling explosive mixtures in the
dsireed way. This difficulty now removed, puts the Brayton
cycle on a different basis, making the system quite as feasible
as the Otto, and, in most respects, promising better results. Not
THE HEAT-ENGINE PROBLEM. 61
only this, but the fact that the oil combustion will almost cer-
tainly be put within as easy reach, adds another point in favor
of the Brayton cycle, in the carrying out of which any sort of
oil may be used, whereas the Otto is here barred.
It is not necessary to enumerate here the comparative merits
of the two systems, for that can be easily judged by what has
already been stated.
81. There is one point, however, that should receive notice,
that is, should we operate Brayton cycles with intermittent or
continuous combustion ? With intermittent combustion the fire
burns within the cylinder, and as nothing, but fuel and air pass
the inlet valves, they can be the more easily kept cool ; while,
on the other hand, the placing of the burner beyond the valve
presents two undesirable features : first, the clearance must be
unusually large ; and second, the intermittent feed and cut-off
of air and fuel at just the right time, without alteration of pro-
portion in a fraction of a second, introduces a condition very
difficult to meet. Continuous combustion within a fire-box is
easier to handle, there being no alterations of feed and the
clearance may be as small as we please, whereas we have as
undesirable the feeding of hot gases past the inlet valves.
Which of these alternatives will prove the better for use, in
the system of engines under treatment, can be decided only by
actual construction, but as either will work, there is no great
risk involved in building.
The paper on liquid fuel combustion which follows belongs prop-
erly in the body of the previous paper and should follow other
matter of paragraph 79 page 60,* but at the time this was written
the work on oil had not yet been completed. The slight lack of
continuity also which is apparent as well as some repetition is due
to the way in which it was found necessary to present this some-
what extended work, i. e. the preparation of separate papers dealing
each with one phase of the work but each sufficiently complete
within itself to make good reading and hold the interest to but one
topic. With this in mind it is believed that the connection and
interrelation existing will not seem too strained.
*A. S. M. E., Dec., 1901.
g This paper is sent to you that you may examine it in advance of the
meeting, and prepare any discussion of it which you may wish to present.
It is issued to the membership in confidence, and with the distinct understand-
ing that it is not to be given to the press or to the public until after it has been
presented at the meeting.
The Society as a body is not responsible for the statements of fact or opinion
advanced in papers or discussion. (Art. 44 of its Rules.)
As there will be no adequate supply of extra copies, and papers are liable to
be read by abstract only, preserve this copy for your use, and
BRING THIS COPY WITH YOU TO THE MEETING.
(Subject to Revision.)
Hio. 0934.*
LIQUID FUEL COMBUSTION.
BY CHARLES E. LUCRE, NEW YORK.
(Non-Member.)
PRESENTED BY R. H. FERNALD, NEW YORK.
(Associate Member of the Society.)
1. OIL combustion, considered as a rather complicated series of
physical actions, has never received the attention due to its impor-
tance. There have appeared from time to time men who, taking
up the corresponding question for gases, gave to the world a series
of researches which leave but little to be desired, and the very
perfection and elasticity of our methods of burning gases brings
into stronger relief the narrow limits of present practice in oil
combustion. Before we can hope to design special and proper
furnaces this problem must be attacked from this standpoint, and
the physical operations will when brought together and classified
give us the principles of oil combustion. A detailed and minute
treatment of this question would call for a lifetime of study, but
some of the principles are more prominent and appear more
evident than the others; a few of these have appeared in the
course of some experiments undertaken for an object noted
later.
2. The analytical treatment of the combustion of gases greatly
simplifies the problem of oil combustion. By classifying the gas-
burning methods according to the mode of bringing the air and
gas together, it was found that there were, broadly, two great
* To be presented at the Boston meeting (May, 1902) of the American Society of
Mechanical Engineers, and forming part of Volume XXIII. of the Transactions.
2 LIQUID FUEL COMBUSTION.
divisions of all systems, those in which a supporting atmosphere
was necessary, and those in which, because of the self-propagation or
explosive property of the burning mass, no supporting atmosphere
was necessary. Moreover a distinctly different set of laws of
physical action holds in each case. The laws of combustion for
explosive mixtures with their volumeless flames are radically
different from those for all other mixtures the combustion of
which calls for a supporting atmosphere, giving rise to a volume
of flame due to the meeting of the fuel and oxygen at different
points, and at each springing into flame when juncture is effected.
The volumeless flame of the true explosive fire depends for its
localization and maintenance on the relation between the rate of
propagation of inflammation in the mixtures and the velocity of
translation, together with the extent of freedom from diffusion
of the fresh mixture with the products while approaching the
combustion surface.
3. For .a comprehension of the different cases of oil burning,
we must add to our knowledge of gas combustion something on
the vaporization of the oils, since it is conceded that oil will not
burn as such a distillation or vaporization preceding the actual
combination with oxygen. So that different oil systems will differ
chiefly in the method of producing the oil vapor, and in the
method of causing a meeting of this vapor with the air. Any
two systems which agree in these two points must be brought
together as coming under one class, but perhaps differing in details
which may or may not be essential to good results.
4. Of all the different systems proposed we can, according to
the above, note only three different general classes :
I. The burning of oil in an atmosphere without previous treat-
ment by air or heat. This class burns the oil (a) from a surface
either that of the liquid mass or that of films artificially produced
by sand, stones, fibrous or metal wicks. The vapor burns imme-
diately as formed, and hence there can be no mixing with air, the
flame existing merely in an atmosphere which may be often
renewed or not, i.e., depend on blowers or mere convection. The
fires resulting from this class are grouped for action and effects
with the first kind of gas combustion, a jet of gas issuing unmixed
into an atmosphere of air.
There may also be included (b) those retort or (c) spray vapor
producers which deliver oil gas, as just noted.
Oil burning by methods coming under this class must be subject
LIQUID FUEL COMBUSTION.
to the same merits and defects as the gas combustion noted, how-
ever diverse, complicated, or ingenious the details of operation or
construction may be.
II. Under this class will be grouped all oil fires capable of
producing what is known in gas combustion as the "Bunsen
effects." That is to say, the oil is vaporized in such a way as to
permit the mixture with it of a certain amount of air before it
reaches the existing flame, and having reached the existing flame
requires an oxygen atmosphere to support combustion.
Any system producing vapor which can be handled as can a
gas may also be included.
III. The third class of oil combustion will include all those
methods in which the oil is vaporized and mixed with air in such
proportions and in such manner that there will result an explosive
mixture of oil vapor and air. Such oil fires will be subject to the
laws of combustion of explosive mixtures. The vapor may be
produced in retorts by boiling a mass of liquid, or by spraying oil
on hot surfaces and then conducting it to a point where it may
mix with air, or the oil may vaporize by contact with or approach
to a hot surface in the presence of the air.
5. The most natural and earliest practical method of oil burn-
ing was that of simply lighting the surface of a mass resting in
a pan. The amount of heat that could be developed depending
on the surface of oil exposed led to a spreading of the oil over
plates and running over numerous grooves and in the formation
of cascades, etc. The high flash point of some oils, i.e., the high
temperature at which they give off combustible vapor and the
presence of the liquid mass made it impossible to burn them in
this way, and hence was brought about one of the first improve-
ments in oil combustion. The wick system results from a desire
to produce more vapor, and this from oils of high flash point ; by
it oil is caused to spread out over a large surface in as thin a film
as possible, and is then subjected to heat. Being in a thin film it
is easily vaporized because of the small quantity at any point and
the ease with which the substance supporting the film can be
heated and kept hot, the vapor thus produced burns as it appears
in an air atmosphere.
When the film bearer has a4ow specific head the vaporization is
the more rapid at the surface but slower beyond ; with a metal
film bearer the conduction of heat beyond the surface causes a
vaporization at more points and insures more complete vaporiza-
LIQUID FUEL COMBUSTION.
tion by a superheating at the surface ; the superheating may even
decompose the vapor.
6. These wick burners are easily illustrated by a pile of sand,
fragments of brick or fibrous material in a pan of oil ; wire net
may also be substituted for the fibrous or other material.
FIG, 1.
For these burners to work at all the surface, at least,, must be
hot, and when acting in an atmosphere the latent heat of evapora-
tion of the oil will tend to keep the temperature down y resulting
in steady conditions. There will always be a limit of rapidity in
such combustion, since a constant state will be reached for the
FIG
wick temperature and rate of eraporation, and, consequently, for
the combustion, making regulation difficult.
Tr Such systems must necessarily be slow heat producers ; how-
ever perfect the combustion and disadvantages of the first class of
gas combustion, we must add a few more characteristic of the
methods of evaporation.
Fig, 1 shows the simple pan-wick system of Weeks, and Fig. 2
LIQUID FUEL COMBUSTION.
the cascade of Verstract, which is combined with a wick at the
bottom to burn what escapes from the cascade. This is- selected
for illustration because it is also an example of an attempt to pro-
duce Bunsen effects in the mixing of air with the vapor. It does
not succeed in this, however, because when the liquid surface is
present the flame will locate there, and the air blown through
the slits R on the falling oil sheet can only have the effect of an
often renewed atmosphere, i.e., a wind ; no mixing of vapor and
air is possible.
8. However, the surface of vaporization is increased, hence
more vapor is produced, and, in addition, the air blown through
helps to accelerate the combustion ; but in spite of this the action
is precisely the same as before, a flame of oil vapor burning in an
f
* If c \r Ir
FIG. 3.
atmosphere of air. The improvement then is not one of class but
of detail.
Improvements of the same sort on the wick method, aiming to
lift the wick from Class I. to Class II., and get Bunsen effects re-
sult in the same way. Air blown through the wick chills it and
retards vaporization, in addition to slightly lifting the hot part of
the surface-flame farther from the vaporization surface, which
should be kept hot. An illustration of this, Fig. 3, is the method
of Hubbard. A mat is provided with a pipe system to deliver oil
at numerous outlets in the mass, with the intention of saturating
the mass. Then air is blown through the mat. The intention is
to vaporize the oil in the mat by the heat from above, and the
vapor, mixing with the air passing through, is to ignite on the
top. It will be readily seen that as each outlet is a source of oil
feed to the mat, we have a large number of wicks grouped with
air blowing on them.
6 LIQUID FUEL COMBUSTION.
9. Supposing a vaporization to take place immediately on
issuing, as is expected, and which fact, of course, depends largely
on the fuel used, we will have each nozzle a source of gas, and
there will result a number of gas jets blowing into the mat. The
ascending air current will lift the gas jet, and there will result
practically a cone of gas with apex at the orifice, 'irrounded by
air. At the edges there would be some diffusion, and beyond the
cone a moving atmosphere of air. If the mat were thick enough
and the air current not too swift, there might result an approach
to a Bunsen flame in an atmosphere of air within the mat. If
the mat were not thick enough, and the air current moved too
fast, there would be at the surface a Bunsen effect. In no case
could there be an explosive effect of Class III., because of the rela-
tion of air and oil vapor supply, the one surrounding the other,
making at every point a constantly changing proportion. Were
the air and oil vapor discharged into the mat through the same
orifice the effect would be quite different, as will be seen
later.
10. Air blown on the top of a wick makes the flame a little
more vigorous only because it renews the atmosphere instead of
depending on convectk>n, but the process would not change the
combustion otherwise. The two systems of surface evaporation
from the liquid mass and from the wick film both demand, in
order that the action may be continuous and non-clogging, an
easily vaporizable oil that will not distil into parts of different
vaporization temperatures. With such an oil obtainable, of
course the next obvious step is to simply boil it in a retort, pro-
ducing vapor which can be used exactly as gas and by all the
means known for gases. However, an additional precaution must
be taken, that of preventing decomposition of the vapor produced
by undue heating before burning.
This would be a great advance over the methods noted before,
given only the proper fuel, and we can produce any sort of fire
from the illuminating flame, through Bunsen and blow-pipe effects
to the more recent explosive fires with their high temperature and
rate of consumption, and with each obtain perfect regulation.
11. To vaporize oil in retorts requires that
(a) The oil be not subject to fractional distillation, but that all
of it must vaporize at the same temperature for any given
pressure.
(b) The temperature shall not rise above this vaporization tern-
LIQUID FUEL COMBUSTION. 7
perature or decomposition of vapor will result with deposit of car-
bon to choke the passages.
(G) The vapor once produced must be prevented not only from
superheating before reaching the fire, but also from condensing.
These conditions are exceedingly difficult to get, and no oil
cheap enough to be used for fuel in competition with coal is avail-
able for the system which is otherwise very attractive in its sim-
plicity and range of possible effects.
12. These oil vapor producers may be operated by (a) the
boiling of a mass of oil, (b) the vaporization of a spray, stream, or
drops by contact with hot parts, and (c) by the carburated air
method. The first two, so far as their resultant effects go may be
grouped together, but the third has been found of value in many
cases where the selection of the required fuel is not prohibited.
Air is brought in contact with liquid surfaces, and passing off
carries some vapor with it. We have here a mixture of air and
vapor burnt in the atmosphere of air, or we may go farther and
form the explosive gas requiring no atmosphere to burn. In just
what proportions of air and vapor the mixture will be delivered
from the carburettor depends on the temperature of the air, the
intimacy and length of time of contact with the liquid, and the
temperature at the evaporation surface. Of course, the tempera-
tures of the carburettor will tend through evaporation to become
continually lower, and this must be guarded against.
13. These oil vapor systems differ but little from the pure gas
systems, and whatever can be done with gas can be done with
these vapors, giving fires of classes I., II., and III., provided the
proper fuel is available, and, if the necessity for the fire is so urgent
that cost is not the most important consideration, they may be
very useful.
We have not yet noted, however, any system which will enable
us to burn heavy oils, or those which fractionally distil, leaving a
residue and of these petroleums and some of the petroleum
products form the largest and cheapest source of liquid fuel sup-
ply.
With the spray or atomizing system we have something
radically different from these so far considered, inasmuch as any
kind of oil may be used and a good fire obtained with each. Here
the oil passes through a small opening, where it meets air issuing
at a high velocity and is by it thrown into the firebox as spray.
The firebox being filled with flame and lined with brick also quiet
8 LIQUID FUEL COMBUSTION.
hot, each particle of oil is vaporized in the presence of air, and the
products of combustion of previously consumed oil particles.
14. The temperature of the fire resulting is extremely high,
and this led to the use of steam for the spraying agent, the
injecting nozzle having other openings through which air passes
under the influence of the chimney draught and partial vacuum
produced by the jet.
The action here is probably more nearly explosive than any-
thing else. It was noted that the rate of propagation of com-
bustion in explosive mixtures is very much increased by high
temperatures. When an explosive mixture is forced into an
enlarged chamber previously made very hot, blow-off is prevented
for quite a range by this increase in the rate of propagation. The
oils commonly used in the spray have a very high temperature
of vaporization, and it is more than probable that, moving with
the air in the hot parts of the firebox, at the high temperature
of the mixture when vaporization takes place, the rate of prop-
agation becomes so high that blow-off does not occur. How-
ever the action is not the best even though explosive, for there
is a large admixture of products with the jet, particularly at the
edge and at reflecting surfaces.
15. The entering jet, approximately conical in form, is com-
posed of a large number of oil particles, each surrounded by some
air and some steam. As the jet approaches the hot section, the
oil springs into gas and the gas with the air into flame, the steam
is inactive until very high temperatures are reached, when it
decomposes and acts as a cooler. The vaporization of the oil is
accomplished either in space while the oil particle is in motion
surrounded by air, or by contact with some of the solid surfaces
of which a good many are provided in the form of arches,
bridges, baffles, etc. All that can be seen is an orange glow and
the course of the jet is invisible, except near the nozzle.
The system requires that the spray be formed, and for this air
or steam under sufficient pressure must be provided, numbers of
baffles and bridges to break the current after it has entered, in
order to scatter the jet and distribute the heat ; a firebox of
sufficient capacity to allow the formation and vaporization of
spray and its final combustion ; small openings at the nozzle.
16. Many auxiliaries to the spray have been used, but of these
the most notable is that of Kermode, who sprays with heated air
directed upon a bed of bricks or asbestos placed on the fire area.
LIQUID FUEL COMBUSTION. 9
He provided this loose fire-bar covering simply to cover the bars
easily but noted that by their presence the action was improved,
for a reason which will be seen later.
While most of these spray systems depend on a pressure drop
of air or steam to atomize the oil and these seem to have given
the best satisfaction, yet there are some others which work on
the few ounces pressure of a fan. The oil is conducted to sharp
points by capillary action and blown from them by the blast ;
tests of these generally show higher oil consumption than the
compressed-air system, probably because of the lower tempera-
ture, resulting from more air and slower burning.
17. The subject of oil combustion in general is very interesting
to the laboratory experimenter, and as a system was desired
which would burn enclosed under pressure for use in the internal
combustion-engine, a series of experiments was undertaken at
Columbia University to find, if possible, a method which was
adapted to these conditions.
With the knowledge of what had been done with oil fires in
other applications as a guide, the first series had for its object
the determination of the principles that should govern enclosed
pressure fire systems. These principles once determined, it was
hoped that the desired method would appear. Some of the
experiments made together with the deductions therefrom are
here briefly presented for a record, as they may be of value to
other workers in the field. Engines which work by passing air
through a fire and thus expanding the volume at constant
pressure, impose on the fire some conditions not easy to satisfy.
18. Air must be compressed into the firebox, and at each
delivery of the compressor there will be a pressure increase on
the fire; similarly fit each admission to the engine cylinder there
will be a pressure drop, and while we may call the system a
constant pressure combustion system, this cannot be strictly true.
What is constant is the mean pressure, and even this may vary
after a limited time, for variation of admission and cut-off will
change it. So that a fire to work successfully in this apparatus
must be unaffected by pressure variation whatever may be its
extent or suddenness.
One of the greatest advantages that may be derived from this
type of engine over the explosive, for example, is the possibility
of employing the cheap and safe heavy oils. But to realize this
advantage we must add to our conditions one imposing the require-
1*
10
LIQUID FUEL COMBUSTION.
ment that heavy oils shall be burnt. And finally, the products of
combustion must be delivered at a constant temperature, and that
as high as possible. Moreover, this maximum must be known to
the designer who proportions his cylinders and mechanism for
some particular volume expansion dependant on this maximum.
19. The most radical difference between these conditions and
those imposed on an ordinary fire is that of burning, enclosed under
a pressure which may vary widely and suddenly ; so it was
decided to first try to obtain a fire which would do this regardless
of the fuel used or the delivery temperature, and this being
attained to experiment with the other conditions by making
appropriate modifications.
20. One burner which seemed to give a good steady Bunsen
effect in ordinary use was that of the gasolene or kerosene
soldering, torch, or cook stove. This seemed so simple and un-
likely to be affected by pressures that the principle envolved was
that first tried, oil fed through a self-vaporizing apparatus and
escaping as gas.
Kerosene was fed through a coil of small brass tubing as shown
in Fig. 4, the oil flowing from the top toward the bottom burning
at B and playing on the coil. It was expected by a long coil to
obtain a perfect vaporization. This device was found exceedingly
irregular in action, no matter how carefully the feed was adjusted,
the vapor delivery was never steady, varying from a long flame to
total extinction. Matters were somewhat improved by enclosing
the coil in a shell insuring a uniform heating throughout. After
working for some time in this way the operation stopped, and
the tube was found full of solid carbon at the lower part, showing
a decomposition of vapor in that part. Gasolene, alcohol, etc.,
could be used, but not petroleum and heavy oils.
LIQUID FUEL COMBUSTION.
11
21. There were two faults prominent in this arrangement : first,
the down-feed through a variously heated coil, gave rise to un-
even vapor generation ; second, the passage of the formed vapor
through the heated part where decomposition could occur.
In the next burner it was intended to do away with both of
these faults, the first to be eliminated by having a large mass of
liquid boiling, and delivering vapor in such a way as to avoid
superheating and so eliminate the second fault.
The apparatus of Fig. 5 was made. Oil enters at A, is dropped
to X, where it boils in the chamber, being heated from below ;
the vapor generated passes around BC, feeding the flame. B is a
valve which permits shutting off vapor delivery, and by the
increase of pressure also the feed which was under constant head.
FIG. 5.
Any rise of oil level was prevented by the overflow D. Air enter-
ing at the bottom with the vapor a very good Bunsen effect
could be produced when burning free and with naphtha as fuel.
When enclosed, however, and with pressure put upon the chamber
the flame became very irregular, and any quick change of pressure
always resulted in extinction. With kerosene there was considera-
ble vapor condensation in the drip. Various modifications of
these vapor generating pressure oil burners were tried, but all were
unsatisfactory for enclosed pressure use. The boiling oil generates
within its chamber varying pressure depending on the rate of boil-
ing, and rate of efflux of the vapor. The rate of boiling or vapor
generation, if the flame below is kept constant, will depend on the
pressure on the surface of the liquid.
22. This, in turn, depends upon the pressure on the flame and
the size of opening in the vapor pipe. We thus have a number
of conditions surrounding the vapor supply, from which the air
12 LIQUID FUEL COMBUSTION.
supply is free, but the air supply has varying conditions of its
own, and as these double conditions are never, as it were, in phase,
the result is failure. The only way in which we can keep the
proportions of air and vapor right is by observing the flame, and
this is, of course, out of the question when it is enclosed. When
conditions can be kept right, a very good fire can be made with
this burner, using kerosene, gasolene, naphtha, alcohol, etc.
Some other experiments leading from this brought out the fact
that much better results could be obtained if the boiling is con-
fined to the surface of the liquid rather than allowed to exist
throughout the whole mass. To get this result a pipe was bent,
as shown, Fig. 6, and oil fed from below to the enclosed length,
which becomes hot on top from the oil vapor jet B. With a con-
stant head, a flame could be kept lighted under pressure, and
enclosed up to the feed-head. A sudden change, however, created
trouble. The vapor delivery depends on the difference in pressure
between the chamber and the feed-head, and the flame grows
smaller, allowing the hot plate to cool when it should be getting
hotter. The proportions could not be maintained at all constant
under variable pressure, though the burner would work all right
when proper adjustments could be made.
23. With a feed varying with the chamber pressure a slight
improvement resulted, though even then the result was not satis-
factory. There was carbonization at the orifice with kerosene,
and the apparatus would not work at all with heavy oil. Sudden
pressure changes invariably caused extinction. The amount
which can be fed economically can be varied but little, and not so
to keep any constancy of proportions with the air.
To maintain some such constancy of proportion was necessary,
because the ultimate object of the oil fire was to heat the air, and
different quantities of oil burnt in the same air will give different
temperatures ; and if the proportion cannot be predicted certainly
LIQUID FUEL COMBUSTION.
13
the final temperatures cannot, and the fire is useless for the pur-
pose in hand. With the purpose of keeping some sort of ratio
between fuel and air, an air suction oil lift was tried, Fig. 7.
24. It was not intended that the complicated action of the com-
mon atomizing spray should result, but only that the air should
lift oil in quantity somewhat in proportion to its own quantity.
This oil is blown with some air through a flattened attenuated
opening A, where it is spread out without changing its velocity,
FIG. 7.
and then brought in contact with an externally heated plate, B, to
be vaporized, the action being similar in effect to that of the
carburettors used in Priestman oil engines. It was found that
there was one rate of air feed at which just the right amount of
oil would lift, a variation either way changing the action materi-
ally. When enclosed the slightest change of pressure results in
bad action, sooting, flooding; and extinction. A number of similar
injector oil lifts were made, and the conclusion reached was that
none could be depended upon to produce the action desired. To
further test the principle of carrying oil by the moving current
FIG. 8.
of air either as mist or vapor, the arrangement of Fig. 8 was tried.
Here an irregular mass of wire net fills the chamber J, which is
about one-half full of oil. The wire threads draw up by capillary
action the oil from the surface, spreading all over the wires and
some of the spaces between, making conditions very favorable for
the air to take up either mist or vapor as the case may be. The
issuing jet is reflected back upon itself and heats the nozzle, insur-
ing that any mist shall become vapor.
14 LIQUID FUEL COMBUSTION.
25. The opening need not be small. It worked very well for kero-
sene and better for gasolene, and much better for both when
heated air was supplied. This fact in addition to that of liquid
collecting in the cone, J$, seems to indicate that a mist rather than
vapor was the result of the air passage over the net. This fact is
further proved by the working under kerosene without dropping
of temperature such as would occur with evaporation. With a
steady set of conditions this apparatus worked well as was noted,
but like the injector devices, no great variation of the fire could be
made. It was tried with petroleum, and the result showed a col-
lection of residues in A, only the lighter volatile parts coming off.
A little carbonization appeared at the nozzle.
26. All these devices depending on the vaporization of oil at
some point have given great trouble from regulation when en-
FIG. 9.
closed, and none was found satisfactory for variation of combustion
pressure. It is probable that one could be designed, but it would
necessarily be complicated. With the ones tried the temperature
of the products could in no way be kept constant, and while most
required large and variable excess of air, a few were found which
could be operated by little ; but the high temperature resulting
invariably produced vapor decomposition. They required, more-
over, special oils, but even this might be endured if a steady
fire with always the same temperature delivery could be made to
work under variable pressures ; but these results could not be
obtained.
27. The vaporization systems were now abandoned in view of
these difficulties for the attractive simplicity of wick combina.
tions which, while perhaps not offering the greatest perfection of
combustion, yet would not go out when put under pressure,
Fig. 9 was tried, with a bottom oil feed to the wick, and air sup-
LIQUID FUEL COMBUSTION. 15
plied above. It was found that the wick at the bottom of the
chamber was not affected by pressure, and burned steadily when
enclosed, but a steady discharge was necessary, for when the dis-
charge was interrupted the flame was extinguished. It seems as if
the products must be conducted away at once, and this is probably
because, with the wick, the vapor generation will go on some time
after the oxygen supply fails. It also seemed advisable to have
the air current and flame tend towards the same opening and not
oppose. Opposition produces a violent flame and irregular action
which may cease entirely at any time.
28. To improve the means of renewing the atmosphere of this
fire the burner of Fig. 10 was made. A is an asbestos mat sup-
plied with oil from B. C is the air-supply pipe ending in the
funnel Z>. If the oil be lighted at A, and time allowed for the
whole to heat up, the burner can be enclosed and pressure applied
FIG. 10.
through the air-supply C without causing extinction. The pres-
sure in the combustion-chamber has absolutely no effect on oil
flame thus produced.
29. The products of combustion thus produced were piped to
a small Shipman steam-engine to observe the effect of the impulse
of engine admission on the action of the fire. Good results for
any pressure were obtained with only one drawback. If the velo-
city of the air over the flame be too high, the flame will go out.
With gas or oil previously vaporized a surplus as well as a deficiency
of air will cause extinction ; here, any surplus will have no effect,
provided only that it move slowly enough. A most important
result was here attained, viz., the flame could be kept going under
working conditions.
For a more perfect and rapid combustion of oil by the wick
method, it seemed desirable to keep the flame hot even beyond its
visible part, and everywhere supplied with fresh air. This could
be done either by drawing the flame out to a thin sheet, or by
shooting across it warm air currents, as in the blow-pipe. Ac-
16
LIQUID FUEL COMBUSTION.
cordingly, the apparatus of Figs. 11 and 12 were constructed.
In A is asbestos, on top of which the oil rests, and through which
oil trickles to the part below enclosed between pipe, (7, and sur-
rounding pipe, B. The flame having been started at /, air was
turned on through (7; the flame was conical, with a well defined
FIG. 11.
blue interior, and was blue even at the tip. This method of feed
might be duplicated by dropping oil In varying quantities on
the loosely-packed wick if a variable combustion is desired. This
burner can be enclosed and put under pressure, and, moreover, un-
like the last, is not sensitive to change of velocity of air through
it.
30. While not all the conditions desired are here met, many
FIG. 12.
that are most important are fulfilled. The burner will work, en-
closed, fairly steadily, and is not affected by pressure changes, but
it always requires a large excess of air, and, therefore, delivers
products whose temperature, though fairly constant, is yet not the
maximum. Fig. 12 was made to be an advance on this type, by
introducing a hot air blow-pipe effect. The flame here, instead of
LIQUID FUEL COMBUSTION. 17
having a blue center, has a deep yellow core forced by the air
currents into blue at the edges. The center is the gas generator,
which gas is completely burned at the edges by air from the
heated lips of the tangent tubes. External heating to redness
will cause internal ignition, and wicks placed in the path of the
products seem to improve the action, both in completeness of com-
bustion and as re-lighting after extinction. This burner could be
used with satisfaction in every point, except that it used such
large quantities of air and delivered products of comparatively
low temperature.
31. The tendency of the preceding experiments is evident,
always away from special vaporizers to arrangements with autom-
atic regulation, the vaporization taking place in the firebox, in the
presence of air if possible, and so preventing not only decomposi-
tion and carbonization, but also condensation. It seems rather
odd that in general the means which worked best under the diffi-
cult conditions imposed were in general the simplest. For the
use of all oils, including those of low and those of high boiling
points, probably the following conditions would, 'if they could be
fulfilled, produce a very good oil fire :
The oil to be introduced as liquid with the air and brought im-
mediately, still with the air, to the hottest part of the fire, with
means added to prevent the mixture of the vapor thus produced
and its air from mixing with products of combustion of matter
already burnt. At the time these conditions were formulated, it
seemed impossible that any apparatus could be constructed which
would permit of such action ; but in fact, such an apparatus was
found, and worked so well as to entirely justify all the labor of
classification and minute experimental observation, which made it
possible to predict what conditions should produce a good method,
even when the means seemed impossible to find.
32. This resultant method was not the outcome of this series of
experiments alone, but rather of the combined oil and gas exper-
iments, some of which have been previously reported. Just about
the time the experiments above described were completed, and the
probably necessary conditions for the good oil fire formulated, the
explosive gas-fire described in the author's paper, No. 923, Vol.
XXIII., p. 292, was discovered.
By the use of the explosive mixture, a fire can be made in a
closed chamber, requiring no atmosphere beyond the mixture fed,
and such a fire will deliver hot gasses at a constant unvarying
18
LIQUID FUEL COMBUSTION.
temperature, no matter what the quantity burned ; this tem-
perature is the maximum possible, as no excess of air is heated ;
and, finally, this very excellent fire calls for no special apparatus,
requiring simply an opening through which the feed must be
made at a rate exceeding the rate of propagation, which opening
is surrounded by a pile of broken rock. The function of this
broken rock is to decrease the velocity of translation by increasing
the area of cross-section of the advancing stream, and to increase
the rate of propagation of inflammation by heating until these
FIG. 13.
two rates become equal, allowing the combustion to localize
within the fragments.
33. This suggested a revival of the experiments on oil along the
lines laid down as follows : It was desired to vaporize the oil and
produce with the vapor an explosive mixture which, in return,
was to burn under pressure as desired in one of the explosive
burners. The apparatus in Fig. 13 was constructed to do this ;
gasolene is held in a chamber kept at about 60 degrees Fahr., and
bubbles through from A, the carburetted air was rendered ex-
plosive by the manipulation of the by-pass B, admitting pure air
above the liquid. The resulting explosive mixture was burnt in
the explosive-burner, C.
34. This arrangement fulfilled the requirements exactly so far
as this particular fuel was concerned, giving a fire under pressure
LIQUID FUEL COMBUSTION.
19
not affected by any changes in pressure however sudden, and
delivering at all feeds hot products of exactly the same tem-
perature.
This burner was also piped to the steam-engine and a second
by-pass, />, permitted, keeping the temperature of the air entering
the engine under perfect control. Wide variations of speed and
pressure had no effect, neither had the pulsation due to engine
admission and cut-off.
35. Here, then, was a very encouraging result, but, unfortun-
FIG. 14.
ately, only, the fuels easily vaporized, such as naphtha, gasolene,
benzine, alcohol, etc., were available.
The next step was an attempt to utilize kerosene in a somewhat
similar way. To this end the apparatus of Fig. 14 was set up.
Here oil is fed to chamber A and kept at variable level ; air is
admitted at B and passing C throws a spray into D where it is
vaporized by the heat of the fire ; the end D being covered with
the rock an explosive fire resulted, the correct proportion of air
to vapor being maintained by varying the air supply and oil fuel.
Thus, while it worked under some circumstances and gave a
very satisfactory fire, showed the same trouble that was experi-
enced with sprays in the other series of experiments and was
20 LIQUID FUEL COMBUSTION.
abandoned for the device there found to be more satisfactory,
i. e., a surface boiling of the liquid. Fig. 15 shows the device
constructed for this purpose.
36. Air was admitted at A, and with it, at first, some gas, mak-
ing an explosive fire at B. The oil was fed from below to the
cone under the plate, -C, heated from above, Yarying level
exposed more or less surface to be heated and varied the distance
from the fire plate. The regulation in practice was not as good
as one might expect from the device. The vaporization by
approach of the oil to the hot parts suggested the next step
which is so obvious that it seems as if it should have been tried
first. This was to simply feed the oil and air through the same
pipe to a pile of rock where the explosive fire is maintained, with
the expectation that the hot pipe will do the vaporizing. The
oil is fed through cock A, Fig. 10, and the air through B, both
reach the fire through the same pipe (?, and burn explosively in the
FIG. 16.
mass of rock. This was eminently satisfactory, and showed an
action which was very fine, if unexpected.
37. When the feed is slow the pipe C becomes hot and then
does undoubtedly act as a vaporizer, but when the feeds are in-
creased the fire is forced away from the nozzle, as in the case of
gases, and the pipe C remains almost cold, no matter how hot the
fire in the rocks, but the perfection of the action is maintained
LIQUID FUEL COMBUSTION. 21
and it is found that not the pipe, C, but the hot rocks themselves
act as the vaporizer. The air and oil impinge together on the
hot mass, spreading out in constant velocity surfaces; the com-
bustion takes place on that surface where the velocity is equal
to the rate of propagation and in the passage the oil automati-
cally vaporizes by contact with the same rocks which make the
explosive fire possible, and all this happens without diffusion with
the products of previous combustion. Thus the function of the
rocks becomes complicated ; first, starting with gas the explosive
fire is made possible by their presence, and the result is the heat-
ing of the entire mass from top to bottom, the mass thus heated
is a perfect vaporizer for the oil, which, fed with its air makes an
explosive mixture and maintains the temperature of the rocks,
the whole interrelated series of actions and reactions producing
what I have named the " Explosive Oil Fire."
38. Were the proportions not explosive the interior of the mass
would chill and the vaporization would stop. It is a very strik-
ing experiment to withdraw the nozzle from the intensely glowing
mass of rock, of a properly working fire, and note the oil drip,
drop by drop, giving off each time a dull red flash and a cloud of
smoke, while the whole rock mass cools down ; a re-insertion of
the nozzle causes at once a resumption of the intense rapid high
temperature combustion. And, secondly, by a simple change of
proportion observe an instant cessation of the action, producing
first smoke and then total extinction.
This method of burning the oil is perfectly adapted to the pur-
pose for which it is designed, i.e., the combustion of any oil in a
closed pressure-chamber, as already described, and the action
leaves nothing to be desired.
39. Naturally the next question would be to determine the
action with residue oils of petroleum. It need only be remarked
here that with every oil tried the action was the same ; and three
fires side by side, burning respectively kerosene, cylinder oil, and
linseed oil, showed no difference in action. The so-called residue
oils leaves no residue this way. The experiment of feeding the
several oils successively through the same fire without interrup-
tion resulted in no apparent change of action. We can now see
how the action of the brick that Kermide placed on his grates
improved the action, which would have been still further im-
proved if the spray had been prevented from diffusing with pro-
ducts before reaching the brick. Moreover, by this simple change
22 LIQUID FUEL COMBUSTION.
the spraying process would be rendered unnecessary. Also in the
case of the saturated mat referred to in the earlier part of the
paper, it will now be readily seen how the feeding of both oil and
air through the same opening, instead of as designed, would have
completely changed the action.
Originally firebrick fragments were used, but the fire was in-
tensely hot, and fused such fragments together, even fluxing
them and causing a flow. Later other rocks were tried, and mag-
nesite was found not to fuse ; dolomite, also infusible, crumbles
slightly.
It was noted in the experiments on gas that considerable gas
might be added to a mixture in excess of that required for
chemical proportions without injuring the explosive properties of
the resulting mixture, which fact was of value in producing sur-
face flames above the explosive fires ; of course, there will be a
point where the explosive property will be lost and extinction
ensue.
40. When an excess of oil was tried the explosive fire between
the rock fragments, which act as the automatic non-diffusing
vaporizer, was maintained by the lower and explosive part of the
fire, while the excess of oil passed on to be burned above. It
was found possible to varf the oil 100 per cent, without stopping
the explosive action below, the effect being merely a variation in
the length of the surface flame. This variation at will in the
character of the surface flame is of no importance in the problem
which was originally set, i.e., the production of a fire for an in-
ternal combustion engine working by the increase of volume at
constant pressure. It is, however, of the utmost importance in
metallurgical and steam-boiler furnaces, and a few experiments
other than those originally set were made on these applica-
tions.
41. Fig. 17 shows a series of burners which were used experi-
mentally with success on both open and closed fires, showing the
great simplicity that here meets with success. The one at the
lower left-hand corner, shows an air chamber of 2-inch pipe
through which the oil pipe is led, the air and oil passing down-
ward at an incline of about 30 degrees to the rock bed. Propor-
tions of mixture are maintained by external valves ; the outlets
may in this type be easily duplicated.
The one passing the stool shows a down bending quarter pipe
fed with air and oil, and provided with a starting gas cock. About
LIQUID FUEL COMBUSTION.
23
one square foot of rock several inches thick can be kept in a glow
with this.
42. To show that the theory of the formation of combustion sur-
faces holds with this oil tire as with the gas, a pile of broken brick
was arranged to run about fifteen minutes, covered with clay ; the
result is shown on the top of the chair, a round cavity was fused
out and a center lump left, showing what would be expected with
FIG. 17.
this down bending nozzle from the theory, viz., an annulas form
of combustion surface.
The plate on the floor at the right is tapped to receive from be-
low the nozzle lying on the top containing a center vertical oil
feed surrounded by the air feed. This does not work well on a
flat plate as some oil collects in a circle on the plate, where it meets
little air, and is moreover chilled by the plate, a conical brick
bottom works better.
LIQUID FUEL COMBUSTION.
Two 6-inch nipples arranged for closed fire-pots are shown,
the lower one provided with oil, air, and gas inlets delivering to
a pipe which merely enters the wall of the fire chamber. This
works very well; after heating by a properly proportioned mix-
ture, the whole becomes dazzingly hot, and a blue to orange surface
FIG. 18.
flame several feet high can be obtained without disturbing the ac-
tion of the lower fire. While the whole firepot is white hot and
would melt in time, the horizontal feed pipe is always cool enough
to be borne by the hand. The very high temperatures that can be
produced may be easily estimated when it is stated that this
burner can consume a gallon of crude petroleum in about ten min-
utes, and in so doing uses no excess of air.
LIQUID FUEL COMBUSTION.
43. The upper nipple shown is connected for use with gas, and
is provided with a 1-inch clay lining. It is fed from below
with a mixture of air and gas from the motor-driven 6-inch
positive blower. This was used for melting crucibles of tin,
aluminum, lead, copper, etc., in the calibration of a Le Chatelier
pyrometer for some experimental determinations of the passage of
heat through metal from a hot gas to a cold.
44. In the center of the shelf is shown a 2-inch cross bottom
fed by air and gas direct from the mains and used for heating
soldering irons. Kerosene has also been used in this apparatus for
the same purpose.
The first application of this method of combustion to a steam
boiler is illustrated in Fig. 18. The oil tank in the rear has a
delivery pipe starting at the bottom of the tank, and air from the
FIG.
main is piped to the oil surface and to the burner through the
hose ; a slight throttling at the boiler will put enough pressure in
the oil to lift it to the burner, where it passes the valve seen in the
right front. A half-inch pipe leads to the center of the firebox
and then turns down by an elbow ; the grate is covered with clay
and the firepot filled with broken rock A gas connection is
shown for initial heating and a steam pipe passes from the dome
vertically downward in front. It was for observing the action of
steam in the fire that this particular apparatus was^set up. It was
hoped that by the decomposition of the steam in the fire the
excessively high temperatures would be avoided and the use of
special rock of high fusing point rendered unnecessary. If the
fire be started and brought to a steady glow and steam be then
admitted, there will at once appear an almost invisible surface
flame showing the action desired, a decomposition of steam in the
hottest parts and a recombination, more or less complete, beyond
at the surface. The steam is thus a sort of heat distributor, and
26 LIQUID FUEL COMBUSTION.
in this^way it was found feasible to use common fire-brick; an
occasional sticking together is easily broken up at the end of a run
by a bar, and everything made as- good as new.
45. It was also thought worth while to try what could be done in
producing reverberatory action, similar to that of coal fires. To
this end, this apparatus, Fig. 19, was made of brick and clay.
With a 5-inch fire at A, and an air inlet at B and (7, a good hot-
colorless flame 2 feet long could be produced, heating the chamber
D to an even glow with an atmosphere reducing or oxdyzing as
desired.
It should be noted that all of the explosive burners described
will work under any air pressure whatever, a variation merely
altering the distance of the combustion surface from the outlet,
but for burning a given amount of oil a larger air pipe must be
used with low pressure air feeds than with high.
46. In conclusion, it may be said of the method resulting from
this experimental research that it seems to be in every way
satisfactory for the purpose for which it was derived, and may be
of use in other applications. It has no small openings for oil, no
possibility of carbonizing; will burn any oil with air at any pressure,
provided only that enough air be supplied, and is subject to an
almost unlimited variation of form ; it will deliver gases at a con-
stant and maximum temperature, which may be lowered to any-
thing desired by air dilution ; is capable of burning more oil in less
volume than any of the other forms tried, and this with the least
possible amount of air. It must be stated as a drawback that
without the use of steam it calls for the use of selected rock to
prevent fusion.
PHYSICAL PROPERTIES OF EXPLOSIVE MIXTURES.
Power generation, involving as one of its phases the internal
combustion method of heating a gas, demands a knowledge of the
properties of explosive mixtures not only qualitatively, but quanti-
tatively as well. For, internal combustion presupposes the fuel
and requisite air in proper or otherwise known proportionate
amount introduced into the closed system, and investigation has
shown that to obtain the best results in heated products of combus-
tion these two elements the fuel and air should be mixed before
combustion, producing thereby an explosive mixture. Moreover,
it has developed that, no matter whether the fuel be liquid or gas,
the explosive combustion of the resulting explosive mixture is not
only the best from the point of view of physics, but also from that
of simplicity and practicability, that is, it is not only the best
way, but it is the simplest and easiest to carry out.
Researches by very eminent scientists on this subject have
shown : ( i ) That explosive mixtures have properties not possessed
by other mixtures, (2) they have pretty well developed the nature
of these special qualitative properties and (3) they have measured
the extent and intensity of many of these physical reactions of these
mixture. But in spite of the information developed by these men
the fact remains that to-day, when we are so extensively using
explosive mixtures in our exploding gas-engines and contemplating
their utilization in other ways, there does not exist data sufficient
for the calculations of many of the quantities needed, nor is there
obtainable apparatus sufficiently reliable and practicable to enable
designing engineers to obtain the data needed for their work. It
has been the aim, then, of this part of the work not only to work
out if possible a properly simple and accurate means for obtaining
such data, but also to use the apparatus in the making of such
observations as time might permit. Before entering into the work
forming the subject matter of this chapter it seems advisable to
first look at the work of the scientists referred to and to note their
results.
Bunsen, in the course of his work on gas analysis considered:
2 THE HEAT ENGINE PROBLEM.
(i) The heat of combustion of a gas; (2) the temperature of
combustion; (3) " the explosive force of gases "; (4) temperature
of ignition of gases, and (5) limit of inflammability of mixtures,
as influenced by dilution. All of these were undertaken chiefly
in reference to one phase of his system of analysis, i. e. } the deter-
mination of combustible gases in a heterogeneous mixture under
analysis.
The heat of combustion he calculates from that of elements
determined by analysis, using the elemental values of Favre and
Silberman. The temperature of combustion and " explosive
force," or pressure after constant volume combustion, are calcu-
lated from this last by assumption of a constant value of Cv.
Under temperature of ignition he simply noted that a gas which,
by reason of dilution, became uninflammable regains its combusti-
bility if prevented from expanding freely during ignition or when
its temperature is raised. Under limit of inflammability it was
observed that inflammable mixtures might be rendered uninflam-
mable by dilution, and that the point of difference is sharply
marked. His table is :
i Volume of detonating gas with{|;| CO. is
i Volume of detonating gas with^ H i.
He also determined some rates of propagation of explosive
mixtures with the pressure tank and orifice method ; results later
shown to be quite erroneous.
Following him, whatever was done by other individual investi-
gation was overshadowed by the work of Berthelot and Vielle
ariti Mallard and Le Chatelier. The only extensive work under-
taken with the sole object of studying the properties of explosive
mixtures was that of Mallard and Le Chatelier, published in var-
ious papers, and collected and republished in the collected works
of the " Commission de Orison," 1883, with the title " Recherches
Experimentales et Theoretiques sur la Combustion des Melanges
Gazeux Explosifs par MM. Mallard et Le Chatelier Ingenieurs
au Corps des Mines."
They say in the introduction : " We have not limited our work
to mixtures formed by air and fire damp ; we have extended it to
CYCLIC ANALYSIS OF HEAT ENGINES. 3
the principal combustible mixtures. We believed that we should
thus be able to profit by apparatus often costly, set up by us,
and by the experience gained in its manipulation to furnish to
science some new facts on questions still but little known " (1883).
This work and notes on that of previous investigations, which
is reported in some three hundred pages and several plates of cuts,
was divided into three parts :
i. Conditions necessary for starting active combustion and the
temperature of inflammation.
2. The rate with which inflammation, once started, will propa-
gate itself through the gaseous mass, and in general the circum-
stances characterizing that propagation.
3. The pressure produced in a closed vessel after the combus-
tion of the gaseous mixture enclosed in it, from which can be
deduced: (a) Law of cooling of hot gases in cool walls (b)
temperature produced by combustion; (c) nature of variation of
specific heat at high temperature.
The subject of temperature of inflammation is treated as
follows :
Historical. The work of Davy, who observed that, at times,
when a metal bar while hot might not inflame mixture a flame will.
He arranged some gases in order of inflammability : Marsh gas,
ethylene, carbonic oxide, hydrogen and phosphoretted hydrogen.
To the last, PhH 3 , he gave 116 C.
Davy also noted that slow combustion unaccompanied by heat
and light always took place in mixtures.
After him, Bunsen, who worked on questionable theoretic
grounds, gave these figures :
i Volume (H + O) + 2.85 Volumes of CO 2 1790 C.
+ 3-65 Volumes " H 2116 C.
+ 10 Volumes " 857 C.
Nothing more was found, probably due to difficulties.
Method of Experimenting. After considering several methods
all are rejected as inaccurate or impracticable except the one
adopted. Mixture is admitted rapidly into a chamber previously
heated to a known temperature which may be empty or filled.
Both methods were used. It is then observed whether the gas
ignites or not, and two limiting temperatures can be determined
4 THE HEAT ENGINE PROBLEM.
between which the temperature of ignition must lie. It was
found very slow work and difficult to avoid both accidental
and systematic errors; however, results were tabulated for mix-
tures of H and air in all proportions and diluted with CO 2 and O.
Similarly for CO and C 2 H 4 , fire damp. The limits in the three
cases are for all mixtures :
H> 5I7-595 , mixed with air, O and CO,.
CO, 630-725, mixed with air, O and CO 2 .
CHj, 640-760, mixed with air and O.
Experiments on slow combustion show a discontinuity between
it and that accompanied by light and heat changes.
The whole is summarized as follows :
The temperature of inflammation can be fixed at
555 for explosive mixtures of H and O.
655 " " CO and O.
656 " " GH 4 and O.
The addition to explosive gas of even a considerable volume of
inert gas modifies little or not at all the temperature of inflam-
mation.
However, with mixtures of CO and O the addition of notable
quantity of CO 2 seems to elevate that temperature to a sensible
degree. One volume of CO 2 added to explosive mixtures CO + O
raises the temperature from 655 to 700.
For mixtures in which H and O are the elements the com-
bustion takes place as soon as the temperature of inflammation is
reached. It is entirely otherwise for marsh gas, which we may
liken to fire damp. The mixtures formed by this gas with air or
oxygen do not burn except after having been brought to and kept
ten seconds perhaps at a temperature equal to or superior to that
of inflammation. The retard of inflammation increases with dif-
ference of temperature of gas and that of inflammation and with
the increase of inert gas. This latter reason explains why, accord-
ing to Davy, a bar of red-hot iron, though above 650, will not
ignite a mixture of fire damp. By opposing circulation one may
easily provoke inflammation because when it circulates freely the
gas does not remain long enough exposed to the temperature of
inflammation.
CYCLIC ANALYSIS OF HEAT ENGINES. 5
RATE OF PROPAGATION.
Davy is recognized as the first to study the question. Without
measuring exactly he knew it took less than one second for the
flame to travel through the best mixture of air and fire damp one
foot long, and also came to the conclusion that small diameter
tubes and metal gauzes will prevent passage of flame of the
majority of mixtures. * ,
Bunsen is noted as having found, by his orifice and tank method,
the figure of 35 m. per second for H and O.
MM. Schloesing and De Mondesir did some unpublished work
for gas-engines, observing in glass tubes the progress of the flame.
They used mixture of CO where r is so slow as to be easily
followed by the eye. They noted that for these slow mixtures
an agitator such as that of the jet of gas into a quiet mass makes
r very great and that combustion itself causes many agitations,
so that the values observed may vary widely and be always dif-
ferent from normal. The agitations are due to : ( i ) Difference of
density between burnt and unburnt gases; (2) dilatation of burn-
ing part; (3) vibratory actions of several kinds due to compres-
sibility of gas when subject to impulses.
M. Fonesca experimented with mixtures of O and various gases
that burn with it. A stream of mixtures is given a high velocity
till the flame cap rests some distance from the orifice ; the velocity
is then reduced till contact occurs and some figures deduced.
H + O 35 m. per sec.
CO + O 1.40
C 2 H* + 80 2.10
PhH 3 + 80 9.20
M. Gouy tried to deduce r from the angle of the luminous cone
in Bunsen flames. Berthelot and Vieille worked on the subject
and found the rate of propagation abnormal and extremely high,
for certain cases moving several thousand meters per second
very superior to sound. They called this mode of propagation the
explosive wave and recognized that the wave itself must travel
with the velocity of sound.
With this experience to guide them Mallard and Le Chatelkr
began work. Though the method of orifice was recognized as
introducing many errors and as more or less dangerous, it was
THE HEAT ENGINE PROBLEM.
employed for mixture where r did not exceed one meter per
second. The orifice was o.oi m. in diameter. The second method
was that of a tube closed at one end and open at the other with
ignition at the open end.
Time was measured by automatic machines electric, pneu-
matic and photographic. The electric depended on a passage of
a spark through the flame when gap was too large for passage
through cold gas. The pneumatic depends on the explosion of
gas in chambers connected with the tube and ignited by the pas-
sage of the flame. The photographic consisted of a moving plate
receiving the action of the flame in a glass tube giving a curve
whose abscissae are distances in tube and ordinates time. All
these methods called for a delicate and expensive apparatus and
the results obtained are not likely to be soon duplicated.
It was found that various influences acted to change the rate
of propagation and figures are given for each.
1. The material composing the tube
[CO + O] in .01 m. tubes
f For gflass . 2.20 "1
-{ y m. per second.
I For lead . . . 2.35 J
2. Diameter of tubes containing the mixture. One limit is
r=3.oo m. per sec.
D = .003 m.
3. The temperature of the mixture H and air with 30 per
cent. H.
15 3.28 m. per sec.
100 v
4.35 m. per sec.
4. Nature and proportions of mixtures in tubes .01 m. in
diameter :
H per 100.
Velocity.
C 2 H 4 per zoo.
Velocity.
11.65 (C a H 4 )+jr.
Velocity.
6
00
5-6
0.00
-5.N
.42
IO
.60
6.0
03
i.oN
30
20
J -95
JO.O
.42
1.4 N
.19
30
40
3-30
4-37
12.0
14.0
.61
.36
. 5 C0 2
i. oo CO 2
$
50
3-45
16.0
.10
60
2.30
16.2
0.00
70
1. 10
80
o.co
CYCLIC ANALYSIS OF HEAT ENGINES. 7
Illuminating gas and air :
Gas per 100. Velocity.
10 .44 .48
12 .68 .84
15.0 1.02 I.OS
17.5 1.16 i.2i
20.0 .88 0.98
H and O give rates from 40 480 m. All the above concerns
only the uniform movement, but this is always followed by the
vibratory movement and later by the explosive ware if the mass
be large enough and sufficiently extended.
A summary of the work on propagation of inflammation brings
out the following facts :
There are two modes of propagation: (i) Normal, that by
conductivity, and (2) that which takes place by the transmission
of a pressure sufficiently high in the propagation by explosive
wave. These correspond to deflagration and explosion of dyna-
mite, etc. Each has a fixed velocity for a given mixture at a
given pressure.
R due to normal propagation never exceeds 20 m. per sec.
For H and air the maximum is 4.30 m. per sec. for a 40 per cent.
H, i. e. } an excess (30 per cent.).
For C 2 H 4 and air the maximum is 0.62 m. per sec. for a 12.2
per cent., i. e., an excess (9.4 per cent.).
For illuminating gas and air the maximum is 1.25 m. per sec.
for a 17.0 per cent., i. e., an excess (15 per cent.).
For CO and O and air the maximum is 2.00 m. per sec. always.
R increases with I and when tube is large is independent of
diameter, but a tube small enough may cause extinction.
Agitation increases R. Combustion in tube with slow R sets
up oscillation which may cause extinction.
When for any reason of vibration or explosion of burnt gas
the pressure transmitted to a layer next is equal to that which
would elevate it to the temperature of inflammation, the com-
bustion propagates with the same velocity as the compressive
wave resulting in the explosive wave.
TEMPERATURE OF COMBUSTION.
Dulong, Favre and Silberman, Thomsen and Berthelot all
worked on Q, from which the temperature of combustion was to
8 THE HEAT ENGINE PROBLEM.
be calculated with a known value of Cv the specific heat. But
Saint Claire Deville showed that dissociation could take place.
He tried dropping hot metal from the flame to water. Crova and
Rosetti used optical methods on flames. Vieille used spherical
bombs and noted displacement of piston to get maximum pres-
sures.
Bunsen, among others, tried to compute the temperature of
combustion from observed values of pressures resulting from
explosion. Some of the pressure ratios determined by him
follow :
Gas Added to i Volume of
Explosive Mixture. For Mixtures of H.
O 9-97
o 975
I.26N 7.49
6
Gas Added to i Volume of p
Explosive Mixture. For Mixtures of CO.
o .............................. . ........ 10.78
o ....................................... 10.19
.io8O ................................... 9.05
.686CO .................................. 8.89
.8550 ................................... 8.44
i.o86O .................... ........ ....... 7-86
I.2S6N ................................... 7-73
I.256N ................................... 7-35
i.7iO .................................... 6.67
2.i6O .................................... 5-83
4-79
RESUME OF TEMPERATURE OF COMBUSTION
Pressures developed are higher than the static due to the heat
developed. Before communicating itself to the whole mass the
increase of pressure concentrates itself on the layer in contact
and the effect is greater the greater the rate of propagation.
Permanent gases cool according to
dd
^dt= ae + be
when e is (temperature of gas) (temperature of walls) ;
a is independent of pressure.
b is inversely proportional to density.
CYCLIC ANALYSIS OF HEAT ENGINES. 9
If the gas can condense the fall in pressure is given by
dw
dt " ~~ ^''
^v is the variable pressure.
p is tension of the vapor at temperature of the walls.
When gas is partially condensable we have
I dw
where w is the variable pressure and p is pressure of whole mix-
ture at temperature of walls.
Dissociation. Information from cooling curves. Dissociated
CO 2 will recombine when mean temperature of gas reaches 1800.
Gases mixed with it have apparently no effect. No dissociation
of H 2 O noticed.
Temperature of Combustion at Constant Volume. Calculated
from pressures when dissociation = o, i. e., from pressure ratios.
Specific heats are found by the method noted to follow closely
these formulae
.
2000
,, = 4.74
For H 2 O C v = 5.61 -f 3.28 Tio-
Perfect gases C v = 4.8 -f .0006 T
6.313.6
5.6 12.2
4.8 6.0
Temperatures of combustion at constant pressure are calculated
by making change in the value for specific heat.
After these facts were obtained there has appeared periodically
attempts at development of special points. Dugald Clerk obtained
some pressures due to explosion, and more recently others, includ-
ing the Massachusetts Institute of Technology, have given
some values of pressure ratios for mixtures of fuel and air, but
the results do not always agree and seldom cover working
conditions.
After all the work is looked over and the labor and expense
attached to the results realized it seems rather a pity that we have
nothing of any immediate value for the designer of gas-engines or
the user of explosive mixtures in other fields. For example, it
10 THE HEAT ENGINE PROBLEM.
is absolutely impossible to calculate the maximum pressure that
may result in a cylinder of a gas-engine, even when the composition
of the gas is known, or secondly to determine the change in vol-
ume due to the combustion of a mixture at constant pressure.
Of course a calculation can be made, but it will be far from that
realized by actual trial and the reason can no doubt be found
in the great complexity of the process involving many unknown
influences.
As another illustration of the unavailable form of much of
the present information and apparatus there may be cited the
case of determining data for the mean effective pressure that
must be counted on in designing an exploding gas-engine.
Assuming the compression and expansion lines as constant
curves the mean effective pressure of an Otto cycle card will
depend on the compression, i. e. } cylinder clearance, and on the
length of the explosion line, i. e., on kind of fuel and composition
of mixture. From the clearance can be computed the amount
of burnt or partly burnt gases that will be mixed with a fresh
charge, and the resulting complex mixture will have a certain
pressure range for its explosion and this, moreover, for that mix-
ture must be constant. But with the information at hand this
question fundamental to engineers designing gas-engines cannot
be computed.
The questions set down for clarification are these :
1. Pressures resulting from constant volume combustion of a
mixture of gas with air in all explosive proportions to determine
(a) best mixtures and compare with chemical determinations of
the same, and (b) the maximum pressure for each mixture with
variation due to change of composition.
2. Volumes resulting from constant pressure combustion of
mixtures.
3. Heat of combustion of these mixtures burnt at constant
pressure.
4. Heat of combustion for constant volume combustion.
5. The effect of dilution by products of combustion on all of
these quantities.
These questions called for the design of apparatus :
1. For measuring air, gas and neutral products of combustion.
2. For mixing, compressing and storing mixtures.
CYCLIC ANALYSIS OF HEAT ENGINES. II
3. For producing products of combustion by methods available
for determinations on an engineering scale.
4. A constant pressure combustion calorimeter for explosive
mixtures.
5. Same for constant volume.
6. A chamber for determining pressures due to explosion.
7. A chamber for determining volume increase due to constant
pressure combustion.
With the apparatus at hand there are few of the questions vital
to engineers entering into the thermal properties of explosive
gaseous mixtures that cannot be determined in a way immediately
available, i. e. } without computation, for use in design. It must
be noted that, while many of the actions and processes are complex
and a pure scientist would be bound to analyze them and assign
to each element a value, the engineer is more concerned with
resultant effects than elemental ones, and is, moreover, saved the
possibility of multiplied error in computing the resultant from
the elemental if the resultant can be measured directly with suffi-
cient known conditions to insure constancy and serve as a specifica-
tion for the process.
SOME NEW WORK ON PROPERTIES OF EXPLOSIVE
MIXTURES.
THE APPARATUS.
All work in the experimental study of heat is, as is well known,
very difficult, calling for most careful observations with apparatus
sometimes impossible to construct with sufficient accuracy, and
always expensive even to a slight degree of accuracy. The
study of the characteristics of explosive mixtures is no exception
to the rule, partaking of the general difficulty of all heat work,
that of isolation of phenomena of observation, and the prevention
of the manifestation of more than one at a. time. In most cases
each experimenter has designed and constructed apparatus of his
own, and in no case, it seems, has any one used instruments first
employed by a predecessor. It is probably due to this that the
results of different observers do not always agree. In none of
the researches does there seem to have been adopted a sufficiently
direct and simple means for obtaining ALL the results ; having an
instrument to measure one constant the observer has rested con-
tent with COMPUTING others equally important, which were mathe-
matically related, though these computed values might also have
been observed directly. As actual trial has shown that the results
obtained by this process of computation of constants from ob-
served values of some other related constant is not always reliable,
introducing, as it generally 'does, a multiplied error if not involving
unproved assumptions of interrelation, it seemed desirable, (i),
that each constant needed be measured directly under well-defined
conditions, and (2), that the apparatus for the measurements be
made simple enough for duplication by others whose observations
could act as a check on the results. In this reduction to simplicity
it is essential that the apparatus be so constructed as to permit of
measurements on the mixtures under as nearly the same condi-
tions as found in engineering practice as was possible. Such a
set of apparatus once set up in a laboratory can be used to rapidly
determine what is wanted from time to time as occasion might
demand, and the results could easily be checked up by other ob-
12
CYCLIC ANALYSIS OF HEAT ENGINES. 13
servers on the same or duplicated apparatus. The actual construc-
tion should involve little or no machine work, and the instruments
of observation must be those in common use in the general engi-
neering laboratory.
The results of the work on continuous explosive combustion of
gaseous mixtures furnished a ready means for obtaining burnt
gases for dilution of mixtures of otherwise known composition
of air and gas. The action and appearance of the fire itself fur-
nishes a very good and simple means for determining the propor-
FJG. i.
tions being fed, and hence the character of the products of com-
bustion, whether oxidizing reducing or neutral, within limits pretty
close ; much closer than we can attain to constancy in gas-engine
work.
For a constant pressure combustion-calorimeter this same
method also filled requirements. For, as has been shown, a mix-
ture of air, gas and neutral in any proportions that can be exploded
may be burnt in a closed chamber and under pressure, so there
14 THE HEAT ENGINE PROBLEM.
only remains as a necessity, the provision for measuring gas burnt
and heat developed by submerging the whole apparatus in water.
The details of this calorimeter as adopted and used will be noted
later.
Gas measuring, mixing and storing was accomplished by water
displacement in thin metal tanks ; the quantities read by the levels
in a water glass before and after more water had displaced the gas
required. All connections and communications between the parts
were made of combinations of pipe fittings and rubber hose. Four
tanks were provided. (Fig. I, A,B,C,D), each fitted with water
inlet and discharge communicating with water main and sewer,
and each fitted also with gas inlet communicating with the appro-
priate source and discharges connecting A, B, and C with D and
that of D with the apparatus in which mixtures were to be used.
The first three tanks then were measuring tanks, A, for illuminat-
ing gas, B, for air, and C for products of combustion prepared pre-
viously, while the fourth, D, is the mixing and storage tank and is
the only one in which the pressure was allowed
to exceed atmosphere. Two glass tubes, E
and F, were provided to each of the measuring
tanks and extending the entire length. One
of these tubes, F, was connected by both ends
to the interior, while the other, E, was con-
nected at the bottom, only the top being open
to atmosphere, Fig. 2. The doubly connected
tube then serves as an ordinary water column
showing the water level, while the other
gives an indication of the interior pres-
FlG 2 sure. When the level of the open column is
equal to that in the closed column the pressure
on the interior is atmosphere ; if the level of the open column rises
above that of the closed column then the interior pressure exceeds
atmosphere by an amount exactly equal to that due to difference
of level.
To obtain a mixture of any composition the method of operation
was as follows: All tanks were filled with water till they over-
flowed; then, one at a time, the water was allowed to flow out
of the measuring tanks A, B, C and illuminating gas, air and
products of combustion, flowing in each to its own tank, filled the
CYCLIC ANALYSIS OF HEAT ENGINES. 15
space left by the receding water. Manipulation of the gas dis-
charge and water inlet valves, guided by the relative position of
the water levels in the two columns, enabled the operator to keep
the pressure on the inside equal to atmosphere, and relieving the
walls of any stress as well as preventing compression or rarefica-
tion of the gas. When each tank is filled with its respective gas
the amounts of each desired in the mixture are laid off on the
water glass above the present water level and, as these have all
the same diameter, the height on the columns will be a measure
of the quantity of the gas. Water is then turned out of the storage
tank D and a partial vacuum created therein. With one hand on
the gas cock x and the other on the water cock y f the amount of
gas wanted is caused to flow from tank A, where it was measured,
to the tank D, where it is wanted, the transfer being made without
allowing any change of pressure in tank A by simply regulating
the relative openings of the valves x and y and watching the two
water columns. After the gas wanted is transferred air is simi-
larly measured and transferred, and later, if desired, products of
combustion, until finally the tank D contains all the constituents
of the desired mixture. All openings to D are then closed and
water from the main G admitted through the three-way cock H,
until the full pressure (in this case 60 Ibs. gauge) has compressed
the mixture. This method of feeding and afterward compressing,
results in a perfectly uniform mixture, as was proved by com-
paring effects derived from burning parts first drawn off with the
last that remained.
We have now in tank D a perfectly uniform mixture of known
composition, compressed to 60 Ibs. and available for whatever
tests or experiments we may desire to make, by the simple opening
of the valve /. All this measuring, transferring and compressing
the constituents of the mixture takes about five minutes from the
beginning up to the time the mixture is ready for use. When the
operator is working with one mixture this time may be lessened if
an assistant is at hand to recharge measuring tanks and get the
quantities desired measured off for the next mixture desired.
NEUTRAL GAS GENERATOR.
To generate neutral products of combustion a positive blower
A, Fig. 3, driven by a Crocker- Wheeler motor B, fed an ex-
i6
THE HEAT ENGINE PROBLEM.
plosive mixture of air and gas to the two-inch tee C. The propor-
tions were obtained as desired by air-cock D and gas-cock E on
the blower suction and the mixture thus obtained burned within the
mass of broken magnesite in the tee C. Ignition was effected
through the opening G, and when the proportions were found
correct by observing the fire this opening was closed, thus sending
the products of combustion over through pipe / to bell K under
FIG. 3.
water in tank L. When wanted the products of combustion could
be drawn off from K through pipe M under a constant pressure,
being that due to the height of water above the bell. More gases
than were wanted were continuously generated, the surplus always
bubbling off from bell K; this was to insure getting fresh gases
delivered at constant and small head. The water around the
pipes served to condense and catch any steam in the products.
CONSTANT PRESSURE COMBUSTION GAS CALORIMETER.
The constant pressure calorimeter consisted of a one-half-inch
tee F, Fig. 4, nearly full of broken magnesite and fitted with a
jump-spark plug G operated by a vibratory primary circuit breaker
induction coil. Explosive mixtures of the previously determined
composition and of known amount were fed to the bottom of the
combustion chamber F through a one-eighth-inch copper tube E.
As it was necessary to discharge the whole of the measured
quantity of mixture from the pressure tank D f and necessary
secondly that no water should follow, the trap B was introduced.
CYCLIC ANALYSIS OF HEAT ENGINES. I/
It is simply a glass bottle with two glass tubes, one a feed and the
other discharge, fitted to a perforated cork. The discharge tube
was connected with the calorimeter by the light rubber tube D.
Hot gases from F were discharged through H to the bottom of
the water tank and thence passed up through square coil of one-
FIG. 4.
eighth-inch iron pipe. The end of this coil was supplied with
rubber and glass tubes, so that the products of combustion could
be directed to the atmosphere or to any point of the water to act
as a stirrer when so desired.
CONSTANT VOLUME COMBUSTION PRESSURE RATIO CHAMBER.
The explosion chamber for determining the pressures due to
constant volume combustion consisted primarily of tee A, Fig. 5,
nipples B and caps C, C . A Crosby gas-engine indicator was
attached to the tee as shown. The spark plug D was carried in
one branch of the three-way cock on top and on other branch
was connected to mixture feed and water overflow openings E.
The apparatus was first filled with water through valve F until it
overflowed through valve G, the mixture feeding valve H and
18
THE HEAT ENGINE PROBLEM.
water discharge valve / being closed ; the spark points meantime
being protected from the water by the three-way cock. When
entirely full of water valves G and F are closed and H and /
opened, the mixture from tank D thus expelling the water; the
three-way cock is then thrown to permit contact of points with
FIG. 5.
mixture, which is allowed to blow through freely to fill chamber
at atmospheric pressure. Then all openings are closed and the
mixture fired, the pressure rise being shown by the length of line
drawn on the indicator drum to the proper scale. To expel the
burnt gases water is admitted as before and the new charge sub-
sequently used to drive the water out.
CONSTANT PRESSURE COMBUSTION VOLUME RATIO APPARATUS.
The quantity of gas that could be stored in the tank D is so
small and the time to attain maximum effect in a heating chamber
CYCLIC ANALYSIS OF HEAT ENGINES.
so long, that with these tanks the combustion chamber could not
become heated sufficiently to make a measurement of maximum
volume increase. The apparatus of Fig. 6 was constructed with
this end in view. It depends for its action on the principles of
gas flow through an orifice. The rate of flow of a gas through
an orifice is proportional to the form of orifice and to the pressure
drop through the orifice. Now if the gas be caused to pass
through a hole in a plate before combustion, and later, after com-
bustion, pass through a similar hole in a similar plate, the constant
.due to the form of orifice would be eliminated in comparing
velocities through the two holes. Secondly, when the fall in
pressure through each hole is the same the velocity of flow through
each plate will be equal, and the volume passing will be propor-
FIG. 6.
tional to the area of the orifice only if the pressures used be small
enough to make correction for compression vanishingly small.
Gas and air are mixed in any proportion desired at the compressor
intake and delivered, mixed, to the chamber AB, from which the
mixture will pass to the upper chamber C through a hole in the
plate secured between the flanges. In chamber C there is placed
a cone of brick to keep lower plate cool, and in the cone broken
rock to permit of the combustion of the explosive mixture. The
top plate between the flange D is provided with asbestos sheets
to keep the hot gases from chilling just before issuing.
At times both the brick cone for the lower and the asbestos sheet
protection for the upper plates were removed for the taking of
20
THE HEAT ENGINE PROBLEM.
observations, while at another time a one-inch lining of fire clay
was supplied to prevent radiation. Mercury manometers to both
chambers indicate the interior pressures, and hence the drop in
pressure through each plate.
RESULTS OBTAINED WITH APPARATUS. FLAMES IN ATMOSPHERE
OF DIFFERENT AIR-GAS MIXTURES.
Before proceeding to the effects of the combustion of different
mixtures it is necessary to first determine the limits of combusti-
bility, and in so doing opportunity was afforded to observe the
characteristics of the flames of different mix-
tures.
The mixture from the compression storage
tank D was led to the apparatus, Fig. 7. This
consisted of a one-quarter-inch tee with a
valve A and manometer B, the flame locating
at C. Mixtures were ignited and the flow
regulated to determine the maximum length
and character of the flame and the pressure
at which the flame would blow off.
Appearance and blozv-off pressures of mix-
tures of air and gas burnt at opening of one-
quarter-inch pipe in atmosphere of air:
Mixture, {Air .
FIG. 7
Blow-off pressure, 1.25" H 2 O
With a length of one-half inch a clear blue flame results; an
increase to three inches in length develops a green core and faint
spots of yellow appear. A still further increase of the pressure
causes the core to become less distinct and the end of the flame
becomes wavy and oscillatory. A roaring noise develops also.
Just before blow-off the flame becomes violet and green near
nozzle ; the end is quite wide and spreading. Blow-off occurred
at 1.25 inches of water pressure with a length of fourteen inches
after some trembling at the nozzle.
Mixture, j ^ ir ' ' ' 2 \ Blow-off pressure, .8" H 2 O.
(^ vJclS JL )
Flame all blue and remained blue as length was increased with
CYCLIC ANALYSIS OF HEAT ENGINES. 21
very much less spreading at the ends. There appeared no core.
A length of twelve inches was the maximum at a water pressure
of eight-tenths of an inch.
Mixture, { ^ ir ' ' ' 3 } Blow-off pressure, .5" H 2 O.
(. vjraS . I )
, j ^ ir ' * ' 5 } Blow-off pressure, .12" H 2 O.
(, vjas . i )
First appearance of the flame cap, which was- of yellowish color
surrounded by light blue and could be extended to a maximum
length of six inches with a blow-off pressure of one-half inch
water. The cap, however, instead of being smooth, had serrated
edges.
Mixture, j ^ ir * ' ' 4 1 Blow-off pressure, .3" H 2 O.
v. vjas . . . i )
The flame cap is now more distinct, with a filmy halo surround-
ing it, color blue-green. The flame on extension becomes sharply
pointed at the end with little vibration. At the maximum length
of four inches and blow-off pressure of three-tenths of an inch
the flame became very pale.
Mixture,
The flame was very similar in appearance to the last with the
exception of being more blue than green, the tip was very pale
and sharp-pointed at the maximum length of three and one-half
inches.
Mixture, |g ' ' ' ^ j Blow-off pressure, .08" H 2 O.
Flame was somewhat shorter and somewhat more filmy or
cloudy in character, the maximum length, of a little more than
two inches was reached with a water pressure of about eight
one-hundredths of an inch.
Mixtures containing more air than the last could not be burnt.
The limits of explosive combustibility differ from the preceding
limits for flames. When admitted to the explosion chamber,
isolated and exploded in bulk the limits of combustibility were :
Air . . 3 ) ^ f Air ... 7.
Gas . . . i j Upt iGas . . . i.
But with the mixtures 6.5/1 up to 7/1 the ignition was very
uncertain, occurring only after long passage of the spark and
22 THE HEAT ENGINE PROBLEM.
often failing entirely. This is an unfailing characteristic of
extremely dilute mixtures, i. e., mixtures containing a large per-
centage of neutral or excess gases.
Constant Pressure Combustion Calorimeter.
Before stating the results of this calorimeter on an unanalyzed
water gas it will be well to look at some characteristics of a water
of typical composition. It is intended that this calorimeter be
used by men unskilled in gas analysis and in places where such
an apparatus is unavailable. The results that theoretically should
accrue from this typical water gas will be compressed until what
was actually observed on the action of the gas used.
Stillman gives as an ordinary water gas the following mixture:
C0 2 3-8
C 2 H, 14.6
CO 28.0
H 35-6
CH 4 16.7
N 1.3
Total 100.0
of this we have
NEUTRAL.
C0 2 3-8
N ^3
Total 5-i
This gas, moreover, will heat yield 691.59 B. T. U. per cubic
foot products condensed, and will call for in its combustion 5.21
parts of air per one part of gas.
A chemical mixture then would have these characteristics :
Air ' 5- 21 volumes
Gas LOO
Total 6.21
of which we have
Neutral {Neutral in gas .051
I Nitrogen in air 4120
Total neutral 4.17 in 6.21 parts or 67 per cent, neutral.
Let us then tabulate various mixtures and note some of their
characteristics.
CYCLIC ANALYSIS OF HEAT ENGINES.
Gas.
i
i
i
i
i
i
i
I
I
Air.
3.o
3-5
4-0
4-5
5-o
5-5
6.0
6.5
7-o
Inactive Air, i. e., Excess.
.29
.79
1.29
1.79
Active Air.
3-o
3-5
4.0
4-5
5-o
5.21
5.21
5.*i
5-21
Inactive Gas, i. e., Excess.
.427
.328 .232
135
.050
Active Gas.
Neutral in Active Air.
Neutral in Active Gas.
573
2.372
.029
.672
2.768
034
.768
3-163
039
.865
3-559
044
950
3-954
.048
I.OOO
4.120
.051
I.OOO
4.120
.051
I.OOO
4.120
.051
I.OOO
4.120
.051
Total inactive or excess.
Per cent, inactive or excess.
2.828
.701
3- '3! 3-434
.696 .687
3-735
.680
4.052
675
4.461
.686
4.96!
.704
5.46!
.729
5-961
.746
It should be noted how very slightly the increase in percentage
of dilution increases with the excesses of air and gas ; though the
proportions may vary over 100 per cent., the dilution varies through
but little more than 5 per cent. This is very striking, as will be
noted again when the results of increasing dilution by neutral
additions is taken up. There can be little doubt that the limits
of combustibility is intimately associated with the per cent, of
neutral or inactive gases present.
Next let us examine the calorific values of some of these mix-
tures, i. e.j the amount of heat that one cubic foot of gas can
deliver when burnt explosively in mixtures within the limits of
explosive combustion. The heat developed by a cubic foot of the
gas in question is 691.59 B. T. U. when completely burnt, *. e. f
in a chemical mixture, or in a mixture in which air is in excess,
within of course the limits of combustibility.
Gas.
i
i
i
i
i
i
I
I
I
Air.
3-o
3-5
4.0
4-5
5-o
5-5
6.0
6.5
7-o
Gas Burnt, i. e., Gas that could
find Air Enough to Burn it.
B. T. U. Available.
573
396.3
.672
464.8
.768
53L2
.865
598.2
950
657.0
I.OOO
691.6
I.OOO
691.6
I.OOO
691.6
I.OOO
691.6
These results are shown graphically on the curve A of Fig. 9.
Curve B shows the results of observations on the water gas used
in the calorimeter and of unknown composition.
The calorimeter can was filled with 16.5 Ibs. water for each run
and the water equivalent determined by experiment to be 2.1 Ibs.
of water ; this is for both can and coil. Radiation was assumed
zero, as the temperature of the room was in every case between
the initial water temperature and the final.
2 4
THE HEAT ENGINE PROBLEM.
Four inches altitude of gas was used every time from the eigh-
teen-inch tank mixed each time with varying quantities of air.
The results are as follows reduced to cubic feet of gas :
B. T. U. PER CUBIC FOOT OF GAS WHEN MIXED WITH AIR.
Gas.
700
Air.
3-0
3.5
4.0
4-5
5-0
5-5
6.0
6-5
B. T. U. per cu. ft. Gas.
27S.I
347.82
401.57
471.00
541.70
6l6.59
600.78
powers of
bo
7oo
400
\
,00
ir |3er One
FIG. 9.
This is one set of readings, and the curve is a very good one.
other sets of readings were taken, and if the results had been re-
duced to a mean the curve would have been perfect. The one set
given were obtained in a space of about two hours. The values
for mixtures 6.5/1 and 7/1 were very erratic but always below
the maximum, and this result is extremely important, viz., that
very dilute mixtures after combustion has been started may cease
to burn before combustion has become complete.
CYCLIC ANALYSIS OF HEAT ENGINES. 25
It may be well to note a few peculiarities of this calorimeter
before leaving the subject. When starting a continuous stream of
sparks is provoked between the points and the mixture then turned
on. As soon as the flame cap has settled and combustion is well
started the spark may be turned off and attention turned to watch-
ing the thermometer and directing the stirring of the water. The
feed may be depended upon to take care of itself. The process of
starting, however, will not be successful unless good judgment
in regard to certain points is exercised; however, a few trials
are sufficient to show up these difficulties, and the means for avoid-
ing them.
If the entering stream be so small in quantity as to give the gas
too small a velocity through the feed pipe, then when the first part
of the mixture reaches the spark there will be back-flashing, which
may result in a succession of explosions. These successive ex-
plosions will have a period depending on the velocity of feed;
they become more frequent as the velocity increases, until finally
the velocity becomes high enough to force the flame beyond the
feed pipe, when it will lodge in the rocks and stay there. It may
even happen that the back-flash will extend to the trap, but no
harm will be done except blowing off the rubber tube and causing
a loss of the charge. At first this back-flashing was very trouble-
some and necessitated investigation. When the combustion cham-
ber and coil were removed from the water no back-flash occurred
even with a very slow feed, but a reinsertion caused the trouble
to reappear. This was attributed to the intermittent back pres-
sure effect of the bubbling of discharging gas through the water;
it was then the flexible end to the coil was attached so that the
discharge could be directed above during and below after starting,
when it would do no harm. Since it was now shown that back
pressure could have an appreciable effect on the action, and back-
flash still occurring occasionally, the discharge gas coil of one-
eighth inch copper tube shown in Fig. I was removed and the
square coil of larger iron pipe substituted to reduce back pressure.
The back pressure was thus made constant and less than originally,
and the charge could be ignited at once without any failure with
a constancy that 'was very gratifying.
26 THE HEAT ENGINE PROBLEM.
Volume-ratios During Constant Pressure Combustion.
When the fire is enclosed and insulated the immediate effect
of constant pressure combustion is to increase the volume of the
gases ( neglecting the small changes due to chemical regrouping
of molecules). Knowing the amount of heat developed by the
combustion and the specific heat of the gases, the volume increase
should be easy to calculate. But in such a calculation no account
can be taken of the large number of other influences, among them
radiation, conduction, dissociation, etc., involving loss of heat
to other phenomena that may be present and no assurance can
be held out that the specific heat during the process is constant
or equal to that of the products of combustion. A computation,
therefore, by this the only method is of no value in engineering
work, and the only way to obtain a result of real worth is to
measure the increase directly under specified conditions. The
method used has already been noted. The firebox consisted of a
piece of six-inch standard steam pipe twelve inches long and the
plates containing the orifices were of black iron one thirty-second
of an inch in thickness. Unprotected, i. e., with plates bare and
pipe unlined, it was found for a pressure drop of four inches of
Hg through each orifice that the maximum effect was vjv =1.50
for best mixture air and gas. Protecting the interior of the com-
bustion chamber by one inch of fire clay and sand on the inside,
the lower plate carrying a clay cone three inches high, and the
upper plate protected externally by a quarter of an inch of asbestos
sheets, the maximum effect was found to be vjv = 4.20 for
best mixture air and gas. Comparing the heating value of the
gas used with that of say kerosene oil, and making an estimate of
losses from above values it is probable that with the best mixture
of kerosene and air that a value z/ 2 A'i = 6.00 might be expected,
but it must be remembered that this is only an estimate and of but
little value compared with the last two figures of actual ob-
servations.
The apparatus used here was so certain in operation and constant
in results that the readings could be obtained in a very short time
and, when obtained, relied upon. The method of operation was
as follows: The fire was started in the chamber with top plate
removed; once burning steadily this plate was bolted down to
CYCLIC ANALYSIS OF HEAT 'ENGINES. 2J
the flange and the whole allowed to heat up. The top orifice was
originally the same size as the lower and as the fire pot heated
up was continually reamed out to keep the manometer readings
at the ratio of 2 :i, i. e., so that the drop in pressure was the same
through both plates. When continued heating showed no in-
crease in size of the top hole possible, then manipulation of the
mixture was resorted to to cause, if possible, a rise in pressure in
the combustion chamber and so permit a further enlargement of
the upper opening to bring the pressure again to one half that
existing in the lower chamber. Thus the maximum effect was
obtained. It must be remembered that the readings included the
friction effect of passing through the mass of broken rock forming
the burner proper.
Pressure-Ratios for Constant Volume Combustion.
As is the case with volume ratios in constant pressure combustion
it is impossible to calculate the values desired from calorific
value and specific heat. Many determinations of the presence of
ratios for various substances have been made but none for a wide
range of mixtures including as one of the constituents neutral
products of combustion.
Each mixture of air to gas within the range of combustibility
was fired and then to each was added in turn successively increas-
PlG. XIII
FIGS, n, 12, and 13.
ing amounts of neutral gases obtained by burning an explosive
mixture as described.
It appeared that the resulting pressures were intimately con-
nected with the percentage of dilution of neutral or excess gases,
28
THE HEAT ENGINE PROBLEM.
and as the gas used has already exhibited some agreement with
what is possible with the typical water gas chosen in comparison it
will be well to work out a table of percentage of dilution of dif-
ferent mixtures and these figures will be placed on the curves of
Figs. 11-16. The agreement and evident existence of a law is
apparent.
. WATER GAS OF NOTED COMPOSITION.
Mixture
f Air, 3 1
' I Gas, I }
diluted.
Gas.
Air.
Added Neutral.
Primary Neutral.
Total Neutral.
Per cent. Neutral.
I
I
* I
I
3
3
3
3
O
I
2
3
2.8 3
2.8 3
3.83
4.83
5-83
70.1
76.0
80.5
83.0
Mixture,
f Air,
(Gas,
diluted.
Gas.
Air.
Added Neutral.
Primary Neutral.
Total Neutral.
Per cent. Neutral.
4
3-43
3-43
68.7
4
I
4-43
74.0
4
2
5-43
77 .6
4
3
6.43
80.6
4
4
7-43
82.8
f Air,
(Gas,
diluted.
Gas.
Air.
Added Neutral.
Primary Neutral.
Total Neutral.
Per cent. Neutral.
5
O
4-05
4-05
67.5
5
I
5-05
72.1
5
2
6.05
75-7
5
3
7-05
78.4
5
4
8.05
80.5
Mixture
f Air, 6 1
' I Gas, i }
diluted.
Air.
Gas.
Added Neutral.
Primary Neutral.
Total Neutral.
Per cent. Neutral.
6
6
6
6
6
6
I
2
3
4
5
4.96
<
4-9 6
5-96
6.96
7.96
8.96
9.96
70.4
74-4
77-3
79.6
81.3
84.0
CYCLIC ANALYSIS OF HEAT ENGINES .
2 9
Mixture, { ^ 7 1 diluted
{ Gas, I j
Air
Gas.
Added Neutral.
Primary Neutral.
Total Neutral.
Per cent. Neutral.
7
5.96
5.96
74.6
7
I
6.96
77-3
7
2
7.96
7#6
7
3
u
8.96
8i-3
7
4
9.96
83.0
The curves of Figs. 12-16 show the pressures given by the
indicator for each mixture and are the mean values from a large
number of lines drawn by the indicator. These curves are com-
bined in Fig. 17, which is, therefore, a curve of pressures for all
FIG. xiv
FIG. xv
FIGS. 14 and 15.
mixtures diluted or not within the range of explosive combusti-
bility. The numbers on the curves show the percentage of dilu-
tion of the typical water gas. The results are most remarkable
30 THE HEAT ENGINE PROBLEM.
and can be accounted for only by assuming that the presence of a
large amount of dilution hinders combustion. The limits at which
combustion ceases to be possible on too great a dilution are here
indicated, whether that dilution be due to excess gas, excess air
or neutral gases. It will also be observed that the character of the
diluenl has an appreciable effect, but that when the dilution is least
the pressure is greatest, about 60 pounds above atmosphere or a
ratio of 5 ; and the presence of a constant per cent, of neutral will
make combustion impossible no matter what the mixture of air
and gas. The greatest neutral dilution gives the least pressure
about 15 pounds above atmosphere, or a ratio of about 2. These
results give a reason for the decreased pressure in exploding gas-
engines in which the mixture is always diluted by burnt products
.to an extent of 20-40 per cent, of the volume of neutral addition
to the gas mixture which may already have neutral gas present
to the extent of 65-70 per cent.
Neutral additions to the gases sent to the calorimeter and to
the other apparatus showed, besides a corresponding and proper
heat value for the resulting mixture, a decreased rate of propaga-
tion accompanied by a difficulty in ignition and constant tendency
to incomplete combustion, i. e. } tendency to cease burning after
inflammation had been started and before the mass had been-
entirely burnt.
CONCLUSION.
The next step in this work, now that the mathematical dis-
cussion and determination of the physical action and constants is
finished, is naturally to apply the results to an operating machine.
It seems that the best arrangement would result in the combina-
tion of a compressor, a fire chamber and gas-expansion turbine,,
and it is the construction and test of such a combination that
will form the subject of the next work. However, as this point
marks a natural division of the subject and as sufficient has been
developed to more than fill the requirements of a doctor's disserta-
tion, the remainder of the work will be left till later, though it is
with sincere regret that this investigation, already so fascinating
and so fruitful, is even temporarily laid aside.
THIS BOOK IS DUE ON THE LAST DATE
STAMPED BELOW
AN INITIAL FINE OF 25 CENTS
WILL BE ASSESSED FOR FAILURE TO RETURN
THIS BOOK ON THE DATE DUE. THE PENALTY
WILL INCREASE TO SO CENTS ON THE FOURTH
DAY AND TO $1.OO ON THE SEVENTH DAY
OVERDUE.
LD 21-100m-7,'33