vs...‘ a
H. ‘ . I n ‘ ..
A. » cw .. ..
, . I Ii u...l~ ”
«lg . -u a:
z... 5... .L ~
.. 4.

$1.. I... «J ..-ho
Q A;

v 4...

E‘

U u. . .5: u .4.
i <1. .. .

f‘w .. b. o .hrac
.1. .‘. v

.A‘ 4

”0“ .n.
Q8 a:
O 8 .n.
I
GO. ’1 0“.” . H“ . a -
V‘D . I... .i .\ u .
‘k‘q . u .n' a N
0.. A .I‘ l
C a... . .
.‘fi 0' ‘
\ t .a. .w .0
“n...” 0.3.... no 5..
.NM 8 U-.. . .
A ‘0. G I-
I I. 0.4
.9“ «\d 5 u ~
I .v (M‘ I:.,. §
.5 I. i \a
.9“ xx. ‘ h... ‘ .
h u... h”: ‘
3t.
n-fioo

.
_,
.

I.

.. : E , . ,7, ,.

 

A
4

'a
K.
#35

0469

This is to certify that the

thesis entitled

Gas Turbines for Automotive 080.

presented by

Ralph Shepherd Parka

has been aeapted towards fulfillment
of the requirements for

”.8. dogree in '03.

W

 

Major professor

Date A t 2 1

r‘. 1? w —- I

 

 

GAS TURBINES FOR AUTOMOTIVE USE

BY

Ralph Shepherd Parks
M

A l'dESIS

Submitted to the School of Graduate Studies of Michigan
State College of Agriculture and Applied Science
in partial fulfillment of the requirements
for the degree of

MAST'A‘R OF SCIENCE

Department of mechanical Engineering

1950

THHRIS

TABLE OF CONTENTS
Page

I. CALCULATIONS AND DISCUSSION FOR.AN AUTOMDTIVE GAS
TURBINE

Calculation of Compressor................... 5
Calculation of Combustion Chamber........... 14
Calculation of Nozzle....................... 1?
Calculation of Turbine Blading.............. 23
Discussion of Blading materials............. 51
Discussion of Rover Turbocar................ 35

II. FIGURESl-G eeeeeeoeeeeeeeoeeeeeeeeeeeeeeeeeeeeeeoee 37

III. BIBLIOGRAPHYOOOOOOOOOO0.00.0000...OOOOOOOOCOOOOOOOOOOO 43

*‘3
0

LIST OF SYMBOLS

Temperature rise in compressor

Compressor tip speed, f.p.s.

Slip factor

Power input factor

Axial velocity, f.p.s.

Pressure at intake of compressor, p.s.i.
Density, lbs. per cu. ft.‘

Hub diameter of compressor, inches.

Inlet eye diameter of compressor, inches.
Radial velocity, f.p.s.

Whirl velocity, f.p.s.

Acceleration due to gravity, 32.2 ft. per sec.
Volume of air, cu. ft. per sec.

Gas constant, 53.3 for air.

Joule's constant, 778 ft. lbs. per Btu.
Specific beat of air, 0.241 Btu per deg. F. per lb.
Length of blade,inches.

Mean diameter of blading, inches.

FIGURE

1.

2.

3.

4.

5.

6.

LIST OF FIGURES

Temperature-Entropy Diagram.....................
Blower and Diffuser Vanes.......................
Compressor Blade Profile........................
Cutaway of Combustion Chamber...................
Spill Type Injector.............................

Layout of Blading for Nozzles and Turbine'Wheels

PAGE

37

38

39

40

41

42

GAS TURBINES FOR AUTOMOTIVE USE

This thesis will deal with the latest advances in the component
parts of an automotive gas turbine as well as show the performance and
economy of a gas turbine propelled vehicle.

The most usable diagram for the gas turbine cycle is the tempera-
ture-entropy diagram as shown in Figure l. The inlet air is represented
at (l), and is compressed to (2) where the dotted line 1-2 is the
theoretical adiabatic compression. 1-2' is the actual compression and
. is shown with a solid line. The combustion takes place from 2-3 on a
constant pressure line and the expansion takes place from.3-4, followed
by the exhausting from.4-l at constant pressure.

It is noted that this theoretical T—S diagram.is the diagram.for a
Brayton cycle and an equivalent form of the Otto cycle efficiency applies
for the cycle efficience, that is, the efficiency is independent of the
combustion temperature; increases with the pressure ratio, but at a de-
creasing rate.1 However, in the actual cycle, the efficiency, while
increasing with higher pressure ratios, also increases with the tempera-
ture rati02T3/T1. That is, it is increased by raising the maximum.comp
bustion temperature or turbine inlet temperature and by lowering the
ambient or compressor intake temperature. It is the former factor which
above all prevented efforts in the gas-turbine field from.achieving
suceess, since it is only of recent years that materials have become
available which allowed a sufficiently high T3 for attaining a reason-

able efficiency and output. Component efficiencies are of major

 

1 D. 6. Shepherd, Introduction to the Gas Turbine (New York:

D. Van Nostrand Co., 1949) p.78.

 

-1-

importance, too, but turbine efficiencies at least have been suffi-
ciently good for the purpose for many years. Another property of
the actual efficiency is that for a fixed temperature ratio Tz/‘l‘l,
the efficiency increases with pressure ratio until a maximum.value
is reached, and then declines. The reason is that the negative work
of compression increases and reduced the net output at a faster rate
than.the reduction of fuel imput caused by the increased compression

temperature T This is obviously important, since it implies an

2.
optimum.pressure ratio for an imposed maximum.temperature.

.Also it wdll be noted that whereas in the actual Otto cycle dia-
gram.the line 1-2' veers to the left, in the Brayton cycle it veers to
the right giving a higher value for T2. The reason for this is that
the turbulence created by the compressor serves to raise the tempera-
ture.

It can also be seen from the diagram that the work done by the
compressor in compressing the air is proportional to the vertical dis-
tance between 1 and 2' and that the work done by the turbine is pro-
portional to the vertical distance between 3 and 4'. If b0 and ht
represent these two vertical distances, then the lowest possible idling
speed or the turbine will be when he equals ht. If ht is less than hc
there will not be enough work output by the turbine to turn the comp
pressor.

It has been found by experience that the best unit for automotive

use is the separate turbine type of unit, where there is a turbine for

the compressor and another turbine for driving. In.this thesis the

former will be called the GTC, (Gas Turbine Compressor), and the
latter GTPU, (Gas Turbine Power Unit), a type of notation widely used.
This is the type of propulsion unit developed by the Boeing Company
which has proved successful in operation with a Kenworth truck, and
will be the type of unit considered here.

It will be assumed also that the unit under discussion.will de-
velop 125 horsepower at full throttle conditions with the compressor
turning at 30,000 rpm and the drive turbine turning at 22,000 rpm.

In addition, there will be a connecting shroud between the GTC and

the GTPU, this shroud wdll have diffuser vanes to change the direction
of flow of the gases from the compressor turbine to the drive turbine.
For the purposes of this discussion, 1600° Fahrenheit will be used as
the operating temperature, as this temperature will penmit an engine
life comparable to a reciprocating engine of the same horsepower, us-
ing modern metallurgical advances. The unit will burn gasoline, kero-
sene or diesel fuel and an airxfuel ratio by weight of nearly 70:1 will
be used, as a large amount of excess air is used with this type of
application. This excess air is used.to cool down the hot gases to a
usable temperature. The gas turbine propelled vehicle could use a
simple two-speed transmission which would give good results on the
road and would give ample acceleration and hill climb ability.

Considering first the compressor for this unit, it is seen that
the radial compressor lends itself admirably to the automotive field
as it is cheaper to produce, stronger, can be run.at higher speeds, and
has a wider effective operating range as compared with the axial type

of compressor. The air enters the compressor casing by way of the

-3-

intake eye around the hub, is picked up by the rotating vanes of the
impeller, rapidly accelerated and discharged from.the periphery into

a diffuser. This annular chamber is provided with a number of vanes
which form a series of divergent passages, whose function is to change
velocity head to pressure head. From the diffuser the air passes to a
discharge scroll heving two outlets leading to the two burners. There
are three types of impellers used for radial compressors, (l), the
single entry web type, (2) the double-entry web type, and, (3), the
closed or shrouded type. 0f theserthe first is the easiest to manufac-
ture and approaches very closely the efficiencies of the other two
types. A pressure ratio of 4:1 is the optimum.for this type of unit
which operates without regeneration, and is what might be expected from
a single stage radial compressor. Straight blading will be used in the
compressor as it will be easier to manufacture and there will be no
complicated stresses due'to centrifugal bending moments. Backward
carved blades give a high compressor efficiency, but deliver pressure
ratios which are too low for application to gas turbines.

Perhaps the most adaptable process for the manufacture of the comp
pressor wheel would be the Thompson Products powder process which uses
light equipment, very little critical material, and limited space.

This method makes it possible to work to close tolerances with a good
economy, and gives a good finish which is desirable in this case. In
this process, powdered iron-made by reduction of crushed rolling-mill
scale - is nolded to shape in presses under approximately 150 tons.

This "green" compact is sintered at about 2000° F. in a protective at-

mosphere and later coined to the desired shape. Then the precision-

-4-

shaped part is reheated to about 20000 F. in contact with a copper
alloy in a special atmosphere furnace, the alloy being absorbed into
the porous iron by capillary action, producing an almost 100% dense
part. The corrosion resistance required in compressor blading is pro-
vided by a chromized case about 0.002 inches deep. Engine tests are
now'being run.t0 determine the suitability of electroplating a nickel-
chromium.case onto the blading instead of dipping as is now done. This
type of case gives a good corrosion resistance to the blading.

The ultimate strength of the powder part is approximately 100,000
psi, which is less than standard blade materials, but extensive engine
testing has proved the blades to be quite satisfactory. To date, more
’ than 500,000 axial-flow'compressor stator vanes have'been used in a
current production jet engine.

Aside from.the method's principal advantages of low'cost and the
use of non-critical materials, are the requirements for only 250 ton
presses, small work area, and high die life. Some 150,000 blades per
month can be made in 12,000 sq. ft. of manufacturing space. Mblding
and coining dies are currently made of conventional tool steels and
are giving a life of 75,000 to 250,000 pieces, and carboloy dies will
probably yield 250,000 to 1,000,000 blades. Excellent reproducibility
of dimensions is indicated by the long die lifeg'working tolerances
are currently : 0.003 in. to airfoil contour, ! 1/2 deg. of twist, and
i 0.010 in deviation from theoretical chord line.

Powder metal blades with the chromized finish are smooth and the

surface discontinuities are indentations between flat plateaus rather

-5-

than the usual "hills and valleys". Powder metal parts can be polished
to a high lustre by low cost polishing. It has been found by experi-
ment that a very smooth surface on the compressor wheel does not give
any added efficiency as compared with a virtually unpolished wheel so
the time spent on high polishing is wasted. However, a high polish
eases the job of searching for any cracks in.the surface of the blading.
2A.skeleton design.will be given here for the compressor for this
size turbine. The calculations are for a single-sided impeller, and
since the airzfuel ratio is 70:1, then the air mass flow should be for
a 125 horsepower unit, assuming 15% overall thermal efficiency, and
assuming 19,000 btu to be gained from burning of one 1b. of fuel,

M . 125 x 529 x 70 e 2.2 lbs/tee.
0.15 x 773 x 19,000

 

The adiabatic efficiency will be assumed to be 78%, the intake tempera-
ture 50° F., the pressure ratio 4:1, and the speed, 30,000 rpm.

(a) The temperature rise will be:

1
T . 12-1 3—3
T g 4(8 r --/j g 520 (4.0 -1) 3 324° Fe :
° 7% 5378
(b) Tip speed and diameter:
02 :- T g Jcp , where ¢ is the slip factor and O’ is

 

fie!

takingfi _ 1.04, and a' a 0.9,

the power input factor

 

U2 : 324 x 32.2 x 778 x 0.241
41.04 x 0.9
U2 : 2,090,000 U = 1450 ft./hec.

Tip dia. = 60 x 1450 x 12 . 11.1"
5 ,000

 

2 Ibid., p. 116.

(0) Eye dimensions:
Assuming an ordinary intake with eyeavanes bent for no
prewhirl and an axial velocity of 500 f.p.s.

V = 500, then 67a (Temp. rise due to air vel.) =

a
2 0
I47 '3
Temp. ratio = 520 = 1.06
520-20.9

Press. ratio : (1.06)3'5 1.23

P1 : 14.7 = 11.9 1b./'in.2
1723'

ffi_ : 11.9 x 144 = 0.064 1bs./ou.ft.
53.3 x 199.1

Flow'area

1v 0.064 x 500

 

8.

9.9 sq. in.

2 2
Area " 77- (d2 - d1 ) : 909 sq. in. 0 d2 :

Where d is the hub diameter.

1

The criterion for (12 and 6.1 are the Mach number at the eye-tip
and the general mechanical layout which might impose a minimum.for d1.

Taking the above values for d2 and d1:

-7-

d -‘n' 305 5.0 in.

U --- 458 655 ft./sec. : 77'd x rps.
tan --- 1.09 0.752 tan -_- Va
--- 47°30' 37° 0' U

Checking the eye-tip mch number,

V(ve)2 - (11)2 = (500)2 - (655)2 3 824
T1 : 499.1o R., hence velocity of sound ; yg r RT1

 

*W

)/32.2 x 1.4 x 55.5 x 499.1 g 1120

Mach. no. : 824 g 0.735
T126

This is a good average value for the Mach number at the tip, ac-
cording to Shepherd, and no prewhirl will be necessary. A larger tip
diameter than 5 inches would increase the value of U and lead to higher
Mach numbers, and a smaller diameter would increase the value of Va to
such a point that smaller diameters would also lead to higher Mach num-
bers. Thus, it seems that a 5" tip diameter is the optimum diameter for
the minimum Mach number.

(d) Diffuser dimensions

Assume eight diffuser passages and a width of easing 0.75 in.
allowing 0.5 in. between diffuser tip and impeller tip, and 1.0 in. be-
tween the diffuser tip and threat, the diameter of the diffuser tip
pitch circle is 12.1" and of the throat 14.1".

The total loss in the compressor amounts to 324 (1-0.7B) g 71.30 F.,
and this may be apportioned as 35° F. loss from inlet to diffuser tip

and the rest in the diffuser. The whirl velocity of the air at the

-8-

impeller tip is equal to Cf U 3 0.9 x 1450 g 1305 f.p.s. At the diffu-

ser throat it is 1305 x 11.1 = 1020.0 f.p.s. = 84° F. It is now neces-
. 14.1

sary to try values of the radial velocity, Vr, which will satisfy the

continuity equation.

Assume'Vr - 555 f.p.s. : 25.200 F. then

9v 84 + 25.2 a 109.2° F.

A To - 6v : 324-109.2 = 214.80 F.

Adiabatic temp. rise .-_- 214.80 F.-35 _-, 179.e° F.
Adiabatic temp. ratio a 520 - 179.8 - 1.35

520 '
Pressure ratio a (1.35)?“5 a 2.86

 

Pressure at threat a 42.0 p.s.i.a.

Density at throat . 42.0 x 144 ; 0.162 lbs./tu.ft.
53.3 x 699.8

Annulus area = 777x 14.6 x 0.75 = 0.239 sq. ft.

 

 

144
VI. = 1L ; t2.2 #_ = 565, which is close
AP 0.239 x 0162

enough agreement with the assumed 555 f.p.s.

The direction of the air at the throat, which is taken as the angle

of the diffuser passage center line, is given by tan -1'Vr g 565 3 0.433
VL- I305

and the angle is 230 30‘. The throat area is the annulus area times the
sine of the passage angle and therefore equals 0.239 x 144 x 0.398 g 13.7
sq. in.

Hence the throat width is 13.7 = 2.28 in.
8 x 0.75

This is the nominal width, and may have to be altered if surging

is encountered.

From the throat onwards the air is diffused in a passage of equiv-
alent cone angle of 11°, and introduced into the combustion chamber.
One point remains to be fixed, the number of vanes on the impeller.
The two limiting criteria are then the slip factor, (large number of
vanes), and throttling at the intake due to too large a proportion of
the eye annulus taken by vane width. In this case eighteen vanes will
be used, this number being used on another compressor which has proved
successful in the same operating range.

The next step in the process is the combustion of the fuel-air mix-
ture and a discussion of the requirements of the combustion chamber de-
sign and some of the more important calculations will be wrked out
here.

Since reaction rate is very much influenced by temperature, it
would not be desirable to mix all the air with the fuel at one step
since the and temperature is required to be of the order of 1500°--1600o F.
Little or no practical reaction would take place at this temperature
and so a primary zone mst be isolated, into which only a portion of the
total air is admitted. Maximum flame temperature indicates that the
primary airsfuel ratio should be of the order of 18:1, which is higher
than the theoretically chemically correct ratio of 15:1.

The very high primry zone temperature means that the flame tube
wall must be thoroughly cooled on the outside, and usually a thin skin
of air is admitted on the inside at the periphery, so that flames do not
impinge directly on the wall. A considerable portion of the heat trans-
mitted to the wall is by radiation, and the usual yellow-flame combus-

tion contributes more in this way than does the blue flame.

-10-

The existence of low-flame velocity requires the use of some
type of baffle in the primary zone at the point of injection, in order
to stabilize the flame by creating a zone of low air velocity. The
average air velocity entering theicombustion chamber may be from 100 to
300 f.p.s., and the baffle will allow local stagnant regions and areas
of reversed flow'which provide continuous burning zones.

The primary combustion zone should be as large as possible to allow
the maximum.time for reaction at high temperature. It is then necessary
to introduce the diluting air to reduce the temperature to that re-
quired at the turbine inlet, and this will amount to some three-quarters
of the whole, since the overall mixture is 70: 1. This mixing air is
called secondary air.

The mechanism of combustion leads to the expectation that difficulty
say be experienced in maintaining combustion efficiency over a range of
mixture strength.and that carbon formation may occur. The free carbon
is attributed to the cracking of the fuel by exposure to high tempera-
tures without sufficient oxygen.being present for'immediate combustion.
The blue flame signifies a gradual oxidation process, and can be pro-
duced in a gas turbine combustion chamber with extremely small fuel
particles with a high degree of turbulence.

The loss of pressure due to change of momentum.is<iependent only on
the temperature ratio before and after combustion.and is unavoidable as
the gases pass down the length of the chamber. It may be shown.that if
the duct diverges from entry to exit during the heating of the air the

momentum loss is decreased, and since a divergent combustion chamber is

favorable for stable burning, both factors favor its application, and
the divergent burning tube will be used in this unit.

The mixing process takes place by three methods: (a) mechanical
mixing due to baffles, orifices, etc., (used by the German Jumo engine),
(b) eddy diffusion produced by turbulence in the aerodynamic sense of
movement transverse to the bulk flow and (0) molecular diffusion accord-
ing 'to the kinetic theory of gases. The kinetic theory shows that mole-
cular diffusivity is inversely proportional to TS/z/P and therefore
that high temperature is conducive to good mixing and that high pressure
has an adverse effect.

The mixing may be effected by a baffle which causes breakaway of
the air, with the consequent formation of eddies and high general tur-
bulent motion. Flow over a sharp edge produces such an effect, and
cone baffles are often used, the interior of the core being a stabiliz-
ing region. A third method, and the one which will be used in this
particular unit, is to inject the fuel into a region of general turbu-
lence produced by the flow of air through a large number of sharp edged
orifices formed by small closely pitched holes in the burner cone. The
fuel nozzle in this case is placed at the tip of the burner cone and
the fuel is injected into the turbulent air.

The mixing of the diluting air with the hot gases to produce a
stream of uniform temperature may be accomplished by a variety of methods
and has received some attention as a special studv. Mixing by small-
scale turbulence as at the interface of two parallel streams of different

velocities is too slow a process and it is necessary to force the air

into the burning gases through holes whose axis is normal to the flow
or by ducts inclined at an angle. Holes are thersimplest method, but
may not provide enough penetration of the cold air into the hot gases,
particularly in chambers of large diameter. Penetration is improved by
placing numbers of holes in line rather than in a staggered arrangement
or*by the use of slots along the line of flow; Maximum.penetration.with
the lowest pressure loss is effected by having openings with a rounded
entrance and with the least general turbulence in the fluid streams.
While the most usual form of combustion.chamber is arranged to pass

the cold air into the flame gases, this requires that some three-
quarters of the air must be turned through a right angle, with conse-
quent loss.

Problems, therefore, arise in.maintaining stablercombustion.under
all operating conditions. There are two important limiting mixture
ratios: (1) the leanest or flame extinction limit, and (2) the richest
or ignition limit. If the fuel-air ratio is leaner than the first limit,
flame propagation ceases; if it is richer than the second, combustion
cannot be initiated. Measurements in an.experimental combustion chamber
indicate that the flame extinction limit is leaner than the ignition
limit and that the absolute pressure has no influence on either limit.
These tests also indicated that increasing the temperature improves the
ignition limit, while improving the atomization aids the flame-extinc-
tion limit. Attention must be given, therefore, in the development of
a combustion chamber to the preblem.of securing stable combustion over
the entire range of operation and quick ignition. These problems are

more difficult to solve in this liquid-fuel application than for

stationary oil-burning installations because of the high air velocities
. and the absence of radiation surfaces.

A factor bearing on the ignition problem appears to be the position
of the fuel spray cone with reference to the spark plug. The results
obtained from tests of a single burner at different air-flow with fixed
spark plug position show that ignition conditions are most favorable
'when the spray cone was slightly outside the spark plug gap; when the
spary cone was in the region of the spark plug points there was no
stable ignition.

The overall loss of total pressure is directly proportional to the

dynamic head in the chamber.

ApgK (Vzp)
23

 

The non-dimensional loss, K, is a useful criterion in evaluating
the performance of different chambers, the dynamic head being taken
at any arbitrary plane and based on the pressure and temperature of the
entry air. If the arbitrary plane is'taken as that of the greatest
cross-section, then the expression takes into account the size of the
chamber, and allows a reasonable comparison of combustion chambers of
different loading. Sufficient data are not available in published form
to givermany instances, but the Lucas design for the Rolls-Royce Derwent
engine, which is typical of aircraft applications, and will be the type
used for this particular unit, has a loss factor of around 30. This
value of K will be used in.the calculations for size of combustion
chamber and other factors in the design of the chamber. The Power Jets
design for the Whittle contra-flow type engine has a value of about 54,

and the German Jumo of about 13, all these figures being for a density

-14..

ratio of 2. The high value of the Whittle type is largely due to the
reversed flow type of design, (two 180° bends), while the low value

of the German design is achieved at the expense of combustion effi-
ciency. The above criterion is useful in comparing the pressure loss
of combustion chambers themselves, but another method allows a more
concrete value for assessing loss of output in a given engine cycle.
Here the total head loss is expressed as a percentage of the total pres-
sure at the combustion chamber entry, in this case 42.0 p.s.i., and 5%
is a representative value for aircraft engines, while 2% is a reason-
able figure for industrial applications, where space is not at such a
high premium.

It is not possible to predict pressure losses for any detailed de-
sign, and testing is the only way to obtain an accurate figure. For
preliminary design information it is logical to assume a certain per-
centage loss of pressure which is acceptable, and then to find the
combustion chanber area required to give an average value of the loss
factor by the formla given above. For large chailbers, smll scale
models tested with ai r-flow only will give an approximate value, since
K may be found by calculation.

For estimating the size of combustion chamber required for a given
duty, the pressure-loss expressions above may be used to obtain cross-
sectional area, the length being decided by taking a minimum L/D ratio
of about two and up to four if possible. The volume may then be calcu-
lated and the heat-release rate checked. The latter my reach 4.0 x 106
Btu/cu.ft./hr. for aircraft applications, falling to 1.5 x 106 for in-

dustrial units where more latitude is available.

No optiImnn length/diameter ratio for the combustion chamber has
been established, and 2:1 appears to be the minimum acceptable. It
may be convenient to have a larger number of separate chambers, or at
least a larger number of fuel injectors and primary zones if the allow-
able length of chamber is very limited. Adequate mixing for even tem-
perature becomes difficult with chambers of large cross-sectional area,
while very small sizes lead to high combustion intensities if the same
flow velocity is used.

The size of the combustion chambers will now be calculated using a

pressure loss of 3%, using a value for K 3 30, and a dmsity ratio of 2.

 

 

Then,
132-: 30 x 0.162 x 1/2 v2
64.4
v2 = 4740 v g 68.8 = ‘g
A z
1 77'0
3:1/2x2.2x2xm .A=0.19B=T“'
A A
D2 3 0.251 0 = 0.501 ft. = 6.0 1 in.

Use 6" as diameter.
Using an L/D ratio of 3, the length becomes 18".
Assuming 15% thermal efficiency, the no. of Btu/hr. at full throttle

per cu. ft. of combustion volume equals:

2545 x 125 ; 5.6 x 106

0:15 I 0.193 x 2 x TTE

 

This figure is slightly high for industrial use as compared with
the figures quoted above and perhaps on first thought it might be
assumed that the combustion volume was too small. However, it must be

remembered that an automotive engine does not operate at full throttle

-16..

except for brief periods so the actual value of combustion intensity
will be Imzch lower than the above.

The flame, once initiated, is self-sustaining and ignition presents
little trouble with light fuels. A norml type of spark plug with a
coil and breaker is sufficient to start the flame. Sometimes, when
heavy fuel is used the flame will not ignite quickly and a more volatile
fuel will have to be used first to get the flame started.

The importance of atomization will have been realized from previous
discussion, and it is the difficulty of obtaining good atomization over
a wide range of fuel flow that presents the special problem. Range of
flow is quite large on this type of installation and a range of 15/1 is
by no means unusual. In its simplest form the nozzle, which acts as
the metering element in continuous flow, is subject to the square root
law, i.e., rate of flow is proportional to the square root of the ap-
plied pressure. Hence if satisfactory, atomization is obtained for the
lowest flow required at a pressure of 7 lbs. per sq. in., and the range
required if 10 to 1, then the top pressure necessary is just over 700
lbs. per sq. in. High pressures can and are being used, but are unde-
sirable if they can be eliminated, since the fuels used are liable to
be of a non-lubricating type, and pumps to work continuously under such
conditions are difficult to construct.

The problem of selecting a suitable type of fuel injector nozzle
will be attacked next. Figure 5 shows the spill type fuel nozzle and
the action with which it produces a spray. Fuel is admitted under
pressure through the tangentially directed swirl-ports to the conical

vortex chamber in which it acquires whirl velocity, this velocity

-17-

increasing as the diameter is increased so that the fuel travels a
helical path outwards, and eventually reaches'the final orifice.
Leaving the orifice, therfuel has axial and tangential components,

and the spray takes the form of a hollow cone as it enters the combus-
tion chamber. The fuel, once away from the nozzle, however, does not
travel in a helical form,'but rather in a straight line. This conical
shape is the desired shape of the fuel, but the fuel has a tendency to
bow in towards the center line of the spray due to the low pressure at
the center of the spray. This form of fuel spray, if completely con-
verged, is called an "olive" type of spray and is not as efficient as
the hollow cone type as has been shown by experiments conducted by the
Bendix Corporation. The spray should be a complete spray in which the
spray form'begins at the minimum flow rate. The fuel should leave the
orifice Tn droplets and have a relatively thin wall thickness. The

fineness of the<iroplets increases as the pressure is further raised.

3The metering function of a plain hole is given by the following
equation:
2
W': KW'At 2 g 'W
Where 'W = lbs. of floW' t g time of fIOW'ln sec.
K = constant- 0.82 = 1 sec.
W’: wt. per cu. ft. of fuel P : injection pressure
g 51.2 lbs. per cu. ft. = 500 psi x 144
(kerosene) : 72,000 lbs. per
A : nozzle area in sq. ft. sq. ft.
Where
‘W = 125 x 550 g 0.031 lbs.
. x x ,

 

3 E. T. Vincent, The Theo and Desi n of Gas Turbines and
Jet Engines. (Few " York: iTeGi-e w-Hi 1 seal: 00., 1950) p. 531

 

-18-

an

 

d
)2 g a _._ )/64:4x1,410 =501

thus

0.051 = 2.45 x 10"6 sq. ft.
0.82 x 51.2 x 501

 

g 3.53 x 10"4 sq. in.
Diae = 0e0213 inChes

There are three main classifications of the means of atomization,
namely: (a) mechanical means, (b) use of air pressure and (0) use of
steam. The mechanical type will be used here and is the most commonly
used type of nozzle. There are three main divisions of the mechanical
type; (1) the sliding piston type; (2) mnlti-stage type; (3) spill-con-
trol type. The sliding piston type provides for variation of the swirl
port area by means of a sliding piston which moves back against a spring
under increasing pressure. The further back the piston moves, the more
port area is uncovered and the more fuel enters the swirl chamber. In
the latter stages, when the swirl ports are fully open, the discharge
orifice meters the flow; The manufacture of this type to the necessary
close tolerances does not enhance its possibilities for general use.

The multi-stage type nozzle has sets of fixed orifices which come
into use at different pressures. It is usually limited to a two-stage
type, and the‘controlling orifices may be either the swirl-ports or two
separate final orifices, with the change point operated either by

separate checkavalves or by separate fuel manifolds. This type of nozzle

gives good results, but requires careful design and watching of individ-
ual burners to avoid a sudden change of fuel flow characteristics.

The spill control type, which is the type to be used in this in-
stallation, has no moving parts and is easy to manufacture. The princi-
ple of operation is that fuel is bled off the vortex chamber at the
base, the amount being controlled by'a valve in the bleed line. In this
way a large amount of fuel may be admitted to the vortex chamber, giving
high whirl energy and turbulence, but only a portion is allowed to pass
through the final orifice. If the fuel supply to the nozzle is main-
tained at a constant pressure, the full fuel flow'goes to the injector
when the spill valve is closed. When the spill control valve is opened,
the fuel flow'is completely recirculated and by-passed to the fuel pump
and no fuel enters the combustion chamber. Thus between the two ex-
tremes of fully open and fully closed, there is an infinite range of
possible fuel deliveries to the chamber. The chief disadvantage of this
type is that a pump of relatively large capacity is required, since the
total flow of discharge plus spill may be several times that of the act-
ual output necessary. An indirect advantage is that the spill type of
nozzle is kept cool at all times'by, the volume of circulating fuel,
which is particularly advantageous when it is placed in an exposed con-
dition.

When, as in this case, two nozzles are used in parallel, it sometimes
happens that instability of flow results with two nozzles of the spill
type. This may result from.inaccurate manufacture, (e.g., eccentricity

of orifice and swirl chamber or unbalanced swirl ports), although a

-20-

very common cause is the presence of slight scores or scratches on the
final orifice, which should be polished to as fine a finish as possible.

While the most important requisite of a nozzle is good atomization
over a wide range, secondary requirements are uniform spray distribution
and optimum spray cone angle. In a symmetrical chamber the circumferen-
tial distribution should be as even as possible, and is particularly im-
portant when the air turbulence is small, as in low pressure chambers.

It may be tested by collecting the annular spray in a tray divided into
twelve or more sectors, each leading to a separate container in which
the individual amounts may be measured. A deviation above 15% from the
mean may be considered excessive. The cone angle of the spray should

be designed to work with the primary baffle of the combustion chamber,
and the variables affecting it are the ratio of swirl-port, vortex cham-
ber and final orifice dimensions. These also affect the flow character-
istics, but some independent control may be obtained by the L/d ratio of
the final orifice. large L/d reduces the cone angle, but it should never
be so small as to approach a knife-edge, since rapid wear may result.

A divergent or bell-mouthed orifice will increase the cone angle.

The turbine for the GTC unit will be considered next. Looking at
these two separate turbines as a unit in themselves, we can see that the
situation is analogous to a two stage turbine with a set of nozzle blad-
ing and one set of fixed blading so that the gas path travels first
through the nozzles, then through the first stage, then through the fixed
set of vanes and finally through the second stage or drive turbine. From

studies conducted on divergent and convergent ducts it he been determined

-21-

that the flow in convergent ducts tends to be more stable and.more
efficient, and therefore slightly better results are obtained when re-
action staging is used, other conditions being similar. The reaction
type of turbine blading utilizes the principle of "Hero's turbine" and
resembles a lawn sprinkler which rotates due to the reaction of the
water leaving'the nozzle. The impulse type, as differentiated from the
reaction type utilizes the energy transfer due to the change in direction
of the hot gas. The nozzle in.this case changes the static pressure to
velocity and is obtained by convergent passages, and the work is obtained
by change of direction of the absolute velocity. The impulse type of
blading gives maximum.blade efficiency at a wheel speed half as great as
that for’a reaction turbine, and this lower speed of course gives a much
less centrifugal load. Thus it is seen that the best'type of blading

for this type of installation.would be a combination reaction-impulse,
two stage turbine.

Some of the considerations for the design of blading and nozzles
will now be studied along with the resulting calculations and data. The
blade length is determined from the mass flow which in this case is 2.2
lbs. per second. The temperature (absolute) is 20600, and since this is
a reaction type, the efficiency will be at a maximum.over a wide range
of speed and the required horsepower can.be taken at as 100 rather than
125. This is done since the engine will be rarely used at full throttle
horsepower in automotive use. The pressure of the gas is 42 p.s.i.a. and
the exhaust back pressure will be 16.7 p.s.i.a. A.mechanical loss of 3%,

a stage efficiency of 89%, an internal efficiency of 91%, a reheat

factor of 1.02, and an axial velocity at inlet and exit from the blading
of 300 f.p.s. will be assumed.
From the tables "Thermodynamic Properties of Air", as prepared by

the U. 3. Navy:

P1 = 42 T1 = 2060 h1 = 426.2 Pr 3 406.6
'72?‘
- ° 1 -
12 _ 1582 hz _ 295.5
Ah -_- 426.2 - 295.5 = 150.7 Btu/1b. of flow

Gas leaving the Blades will be at a state determined by the stage effi-
ciency of 91%
Ah : 0.91 x 150.7 = 119.0 Btu.
Therefore h2 2 426.2 - 119.0 = 307.2 Btu. and
T2 g 16250 abs. (This figure is high-due to the method of
calculation here which is assuming straight reaction, and actually there

is an impulse factor in.the blading which will lower the exhaust temper-

ature)
Specific volume at exhaust . ‘W R T
P
= 55.5 x 1625 = 40.9 cu. ft./lb.

 

144 x'l4.7
Volume flow at exhaust : 40.9 x 2.2 g 90.1 c.f.s.
With the assumed 300 f.p.s. axial velocity the annular area required is:

Area 3 90.1 = 0.301 sq. ft.
306"

as a first trial assume that an l/Dm of 0.25 will be used, then

7H) 1 - 0.501 and Dm -_-, 0.501
0.25 x 77’

Dm .-..- 0.620 ft. = 7.42 in.

giving a blade ht. of 0.25 x 7.42 g 1.85 in., and outside dia. of 9.27
in., and a hub dia. of 5.57 in., which can be rounded off to a casing
dia. of 9 5/16 in. and a hub diameter of 5 5/8 in. including clearances,
thus giving a blade length of 1 15/16 for the second stage.

Experience shows that for the first stage blading the blade length
should be about 15% of the mean diameter or 1 1/16 in.

Thus it is seen that the second stage blade is longer than the
first stage blade which is in accordance with the reaction theory.

The next step will be to determine the number of blades for the
turbine wheels, using the aspect ratio‘. The aspect ratio is defined as,

Re. «3 blade length , and should be around 2.5 for this type of

Hide chord .
use. Thus the blade chord for the first stage blading becomes 1/2 in.,

 

and for the second stage the blade chord becomes 7/8 inch. Using a
pitch/chord ratio of 0.5, which is not considered excessive for this
unit, it is seen that for the first stage the pitch becomes 1/4 and for
the second stage the pitch becomes 7/16. The value of 0.5 for the pitch/
chord ratio is high, but in this unit a welded construction will be

used which permits this high ratio to be used safely. Thus it is seen
that the theoretically correct pitch is different for each stage of
blading. However, the pitch diameter of the second stage is larger

than the p.d. of the first stage so the calculations are as follows:

7Tx pitch dia. (lst stage) n x (7.42-0.875) g 20.50

17's: pitch dia. (2nd stage) = x 7.42 .-.- 25.5

-24-

No. blades (1st stage) a 20.50 g 82 blades
‘“0.25'
No. blades (2nd stage) = 25.5 = 55.2 use 54 blades.
0.458

In this discussion of the turbine arrangement, the term "first
stage" refers to the GTC unit and the "second stage" to the GTPU.

Further considerations of the turbine show that it is undesirable
for the blade profile to have a radius faired into a straight line,
since this will cause a large local pressure increase, and stalling will
occur. For ease of design the profile may be constructed of circular
arcs, having as feW'arcs as possible. The trailing edge should be as
sharp as possible consistent wdth stress considerations, and in no case
more than 4% of the pitch; and the leading edge may be well rounded,
since the local pressure distribution is not important.

For a given blade speed, maximum work is obtained'with the greatest
deflection, this deflection being known as "camber" in turbine work.

For a pitch/chord ratio of 0.5, which was taken in the above calculation,
the maximum camber allowable is 90°. A higher camber‘angle would pro-
duce'breakaway or'stalling, which might be avoided by increasing the
pitch/thord ratio but the welded root fixing in this unit would not
allow any increase in the pitch/chord ratio.

Data is scarce for computing the angle of incidence of these blades,
but some tests have been conducted on blading at low speed which can be
used as criteria for this case. These tests show that for a pitch/
chord ratio of 0.5 the minimum loss occurs with an angle of incidence
Of f 7°. This value will be used in the blading as the tests show'

rapidly increasing losses at more positive values.

-25-

The high mach number for turbine blading is to be avoided in the
design, but the value of unity can be reached without harm. This is
different from the case of the compressor where the high mach number
was to be avoided.

The turbine nozzles are made convergent because the convergent-
divergent form is too sensitive to off-design conditions. The controll-
ing parameter for the throat velocity is the expression M VT/P, since
the velocity at the throat is sonic. This is proved for a throat of

fixed area A;

v = M = K MT , the sonic velocity v ; Vg 3' RT
175— F' 3

therefore K172 g yga’RT, and M VT : constant.
P P

 

The nozzle blades are subjected only to gas bending stresses due to
change of momentum and to thermal stresses, but nonetheless they operate
at a very high gas temperature and are of rather thin sections, so
stresses should be calculated. The stationary set of nozzles between
the first and second stages will have to be shrouded to prevent amr
leakage loss around the tips of the nozzles.

The temperature at which the turbine blades operate is equal to the
static temperature of the hot gases plus some proportion of the dynamic
temperature due to the kinetic energy of the gases. The amount of heat
conducted through to the disc may be appreciable although the actual
radiation and convective losses may be small. Experimental results on
the Whittle turbine show that the blade assumes 85-90% of the tempera-

ture equivalent of the relative velocity, which is close to the value

-2 6-

of 84% of German results and the value of 90% predicted for a flat
plate edgewise to the stream. In a typical turbine such as the one
under consideration here, operating at 4:1 pressure ratio and Tmax of
1650° F., where Tmax is the sum of static and dynamic temperatures,
the blade temperature will be around 1450° F. at the tip and 20°-50°
lower at the root.

Some of the stresses for the turbine will now'be considered, these
stresses being very important due to the high temperatures involved.
The maximum allowable total stress is a function of the material, its
temperature and the length of life required. Turbines of the type
discussed here may be stressed as high as 15 tons/hq. in. with a rim
temperature of 10000 F. and over, this value may be taken despite the
creep limitations which are placed on the blading because the automo-
tive engine seldom runs at the wide open throttle conditions.

Blade stresses are due to the centrifugal loading and the gas load-
ing. The mass of the blades produces a centrifugal tensile stress and
a centrifugal bending moment because the centroids of successive sec-
tions do not lie on a radial line. The gas loading produces a tangen-
tial loading due to charge of momentum.and an axial loading due to
both momentum and pressure components. The trailing edge takes the
highest tensile stress, since it is the thinnest and therefore the
weakest part of the blade. This high stress may be neutralized by de-
signing the blade so that the centrifugal bending moment is opposite
in direction of the tangential bending stresses. This is accomplished

by giving the blade a slight offset from the radial direction.

-27-

Data on the fatigue strength of materials at high temperatures is
scarce, although the fatigue failure may be more important than creep
in this type of usage. New materials are constantly being developed
which wdll enable the designer to go to higher speeds and higher tempera-
tures in the turbine, but at the present the above limit of 15 tons/bq.
in. will be used.

According to Shepherd4 an approximate formula to be used in the
initial calculations is the following: 2

C.T.S. (tons/'sq.in.) a 0.048 x annulus area x r.g0§ ;

C.T.S. here is the centrifugal tensile stress, the annulus area is 0.301
sq. ft., and the rpm is 30,000 so these figures give:
(30,000)2

C.T.S. = 0.048 x 0.301 x {—1000_) g 13 tons/sq. in.
which is within the allowable limit of 15 tons/sq. in.

One of the most important parts of the stressing problem is the
fatigue failure of the'blades, since the amount of the forces is un-
known. The first step is to check the natural resonance frequencies,
and this is done by simple bowing, the use of a tuned air jet or electro-
magnetic means. Measurement of the frequencies is done by using a
pickdup of the capacitor type, amplified calibrated oscillator and
cathode-ray oscillograph.

Blades vibrate in many ways the lowest frequency being the simple
cantilever "flap", and the higher frequencies being the torsional modes,
edge modes and plate modes of complex form, together with harmonies of

all types. A further possibility is that of flutter, an oscillation of

combines flap and torsion due to aerodynamic forces. Flutter analysis

 

4 shepherd, op. cit., p. 178.

is complicated, but some knowledge is available from aircraft sources.
Flutter is most liable to occur“with long, thin blades, and it is more
likely to occur as higher performance gas turbines are developed.

The sources of vibration in.the engine must be examined after the
determinations of the main frequencies and the second or third harmonies
has been accomplished. These sources consist of the stationary rows of
blading and the blading supports, as in the stationary blades which are
being used in this unit. The critical item here is the spacing between
the nozzle tip and the rotor blades, if this clearance is too small,
excitation of the rotor blades may result. The obvious cure in this
case is to allow plenty of clearance here so that no vibration will
take place. Individual combustion chambers and the impeller vanes and
diffusers on the centrifugal compressor also are possible sources of
excitation. The frequency of the exciting force is the number of dis-
turbances per revolution times the engine speed, and a diagram may be
made plotting the natural frequencies versus the engine r.p.m. Thus
the possible points of interference may be seen.

The experimental determination of blade frequencies is usually
done at room.temperature and with no loading on the blade, hence a
correction is necessary to find the vibrational modes under running
conditions. A knowledge of Young's Modulus, E, both at room tempera-
ture and at running conditions is necessary because the frequency of
any simple vibration is proportional to )ril Centrifugal stress intro-
duces another correction factor and the formula for this correction is:

F2 3 F02 - KHZ , where F is the frequency under running

conditions, F0 is the static frequency, N is the speed in r.p.s., and

K a constant. K can be calculated by reference to the works of
Timoshenko, and its value depends chiefly on the ratio of blade length
to disc diameter.

A knowledge of damping will enable the designer to have a certain
amount of control over the amplitude of vibration present in the blading.
The blade is subject to three kinds of damping: (1) internal damping
dependent on the material, (2) root damping due to the method of fixing,
and (3) aerodynamic damping caused by air flow. Material damping is
only a small portion of the total, so it is unwise to sacrifice any of
the material properties in order to improve it. Root damping will be
very small in this case as the welded fixing is very tight. Aerodynamic
damping is difficult to estimate, but probably accounts for a major
part of the total damping. It is due to the shedding of vortices at
the blade edges, and can be calculated, but is very sensitive to vary-
ing flow conditions. If vibration is encountered, one method of con-
trolling it is to use partial admission, i.e., the use of only a sector
of the nozzle annulus. Another method is by redesigning the airfoil
section of the blade. The latter method sometimes changes the aerody-
namic damping to a marked degree. Due to the lack of published informa-
tion on this problem, empirical means of correction have been largely
employed by engineers in the aircraft gas turbine field when blade
failure resulted from.vibration.

The shaft sizes for the GTC unit and the GTPU can be calculated
using the natural period of vibration of the shaft with the attached

weights, the object being to secure smooth operation free of vibration.

However, an initial assumption may be made based on experience which

shows that the shaft diameter should be about 0.2 times the mean dia-

meter of the blading. Thus it is seen that the shaft diameter will be:
0.2 x 7.42 = 1.484, use 1 1/2 " shafting.

Much research has been done on materials for gas turbine engines
to increase their utilization, this utilization being accomplished in
two ways. The first would extend the lives of component parts, and the
second would permit the use of higher operating temperatures. Operating
temperatures in the neighborhood of 3500° F. are desired.

In the NACA tests on.turbine blades, four turbine disc of Timken
16-25-6 steel were assembled with cast Vitallium.b1ades, and the wheels
subjected to rotation at high speeds in a cyclic fashion. The wheels
were run until a blade failure occurred, and then hardness, micro-
structure, and grain size tests were conducted on the failed and un-
failed blades. Failures were divided into two groups, primary and sec-
ondary. The primary failures were those produced by environmental con-
ditions and the secondary failures were produced by impact of fragments
from primary failures. The fracture face of a primary failure has three
distinct zones: (1) a zone which has the smooth appearance and concentric
rings characteristic of fatigue failure; (2) a zone having a fibrous ap-
pearance suggestive of fractured ductile material; (3) a transitional
zone composed of both zones 1 and 2. Fractures of the primary class were
transcrystalline across the entire blade, and the failures occurred pri-
marily in the center third of the blade length. During operation this

portion of the blade is under the most stress.

-31..

These observations showed that the blade failure was a case of
fatigue which progressed until the centrifugal forces on the blade ex-
ceded the ultimate strength and the blade failed in tension. Thus the
important criteria in selecting a turbine blade are: fatigue strength,
rupture strength, and impact strength. Creep should also be considered
so that the'blade will not elongate sufficiently to rub the inside of
the casing.

A study was made on thermal shock and consisted of heating a thin
symmetrical edge of a specimen of the various alloys to 17500 F. for
two hours and then quenching in water at 45° F. The alloys used'were
3-816, S-590,'Vitallium, 422—19, K-40, and Stellite 6. This test showed
that alloy 8-816 had the greatest resistance to thermal shock and
Stellite 6 the least resistance. The criterion which was used in the
test was the presence of a crack which extended across the face of the
edge. Thus it is seen that the use of 3-816 would prolong the blade
life and its use in.the unit under consideration here would be advantag-
sous.

Referring again to the tests on the Vitallium blades, it was ob-
served that the range on.the range of primary failures was from 3.9
hours to 72 hours, a very large range. If variability could be elimi-
nated, the goal of 72 hours might be achieved, but this "scatter” of
results is a property not only of Vitallium, but is known to exist for
other alloys both forged and cast.

One treatment which was found to be beneficial to Vitallium is that

of aging, which in the tests consisted of aging for 48 hours at 1500° F.

Tirenty of the blades were aged and thi rty—three were used in the as-cast
condition, and the unit was operated at 22,500 rpm and 1650° F. until
all the blades had failed. This test showed that the aged samples had a
15.2 hour deviation as against the unaged samples' 20.7 hour deviation.
The time for the first failure was 4.3 hours for the unaged and 35.8
hours for the aged sample. The mean time of failure for the unaged
sample was 55.2 hours and for the aged, 63.6 hours. Thus it was conclu-
sively proven that aging has a beneficial effect on blading.

In actual use, the automotive turbine would Operate most of the time
at a temperature less than that necessary for wide open throttle condi-
tions, much of the operation is starting and stopping. For this reason
the alloy 8-816 was selected as the steel for the automotive turbine be-
ing considered here, inasmuch as it has the highest resistance to therml
shock.

In considering the essential elements of control, it is seen that
the fuel pump and throttle are the bare essentials as the engine power
is proportional to the fuel flow, the mixture strength being fixed
only by the intrisic characteristics of the engine. Since the amount of
fuel required by the engine is a function of the rpm and inlet air pres-
sure and temperamre, a simple throttle controlling the bypass valve on
the "spill" injector is sufficient for automotive use. However, an over-
speed governor should be incorporated into the mechanism, this governor
would be of the ordinary centrifugal type which would open a valve in the
bypass line for the injector, by-passing all the fuel through the fuel

pump. Another control which would be desirable would be a temperature

control which would open the valve in.the bypass line when the tempera-
ture reached an excessive value.

The basic element of the fuel system, the fuel pump, would be the
positive displacement gear type, delivering about 500 p.s.i. Since this
gear type pump would present starting loads which would be rather high,
it would be advisable to have the valve held open on the bypass line
until the engine was being turned at idling speed by the starter. Then,
‘with'the engine turning at idling speed, the bypass valve would be suffi-
ciently to allow fuel to be sprayed into the combustion chamber where it
would be ignited.

Bearings and lubrication systems are matters of detailed mechanical
design and will not be dealt with to a great extent here. Ball and
roller‘bearings are the most widely used in the aircraft field and would
normally be used in this country, although in Britain plain bearings are
used in this type of engine. Since there is considerable end-thrust
present in the unit, it would be advisable to employ tapered roller bear-
ings.

Lubrication is accomplished by an oil spray in the aircraft turbine,
but for the automotive unit, solid lubrication is used successfully.

The bearings are not sensitive to the amount of oil which they receive,
and the simplest lubrication.arrangements are the best in this case.

There will be an integral gear reduction.used in this engine, the
gear ratio being of the order of 25:1. The planetary type of gearing
lends itself well to the unit and will be used here, since it can accomp-
lish a large gear reduction in a small space. The gearing will demand

good design and close tolerances to avoid noise.

_'ZA_

The gas turbine has many advantages when used in automotive work,
and one of the most important advantages is the excellent torque char-
acteristics which it exhibits. The torque at zero speed is the maximum
obtainable, and thus the use of a clutch is unnecessary. Although it
would be possible to use this unit without a transmission, one gear
change would give better acceleration and thus a two speed transmission
'would be advisable. Also, a reverse gear would be necessary. It is
obvious that the transmission problem is much simplified using a gas
turbine.

Another advantage of the gas turbine is that it is an.airbcooled
unit, there is no need for a radiator, anti-freeze, and all the plumbing
difficulties that go with the water cooling system. A.further advantage
is the simpler construction of the gas turbine. Instead of converting
reciprocating motion into rotary motion through a piston, piston pin,
connecting rod, etc., the turbine transmits rotary motion directly
through a gear reduction.

G. Goeffrey Smith,5 who drove the British Rover Gas turbine powered
auto,.says, "it left me with pleasurable astonishment at the<degree of
excellence and entire ease of handling this unique vehicle." The
present Rover mgine develops nearly 200 hp., idles at 7000 rpm., and
turns up about 40,000 rpm at full speed.

Starting is prompt and there is no warmeup period required. In the
Rover car there is one speed ahead, so the only controls are the accel-

erator and the brake. The car cannot be jerked forward on starting as

 

5 c. Geoffrey Smith, "Rover's Turbocar", as: Journal, 58:36,

the gas turbine acts like a fluid torque converter and the forward
movement is very smooth. The acceleration of the Rover is comparable
to that in a conventional auto, so no difficulties in heavy traffic
would be encountered.

On one run around the track the Rover attained a speed of 89 mph,
but mr. Smith feels that the car was not being "pushed", and that he
could have attained a speed of 100 mph. .A test was made on the accel-
eration of the Rover and it was found to attain.a speed of 60 mph in
exactly 14 seconds. The acceleration test was made by holding down on
the brake, bringing the turbine up to full speed with the foot throttle,
and then releasing the brake.

Some of the disadvantages of the Rover car included noise, high
fuel consumption, and a perceptible odor when slowdng down. Of these,
the high fuel consumption is the most critical disadvantage, and the
one which will demand the most effort to improve. The Rover car gets
only seven miles per gallon, and many advances will have to be made in
the field of fuel consumption of'the gas turbine'before it will become
commonly used.

However,'metalurgica1 strides are being made in.the direction of
higher temperature alloys for gas turbines. With the use of these higher
temperatures and continued improvement in the component efficiencies of
the unit, there is every reason to suspect that the gas turbine may gain

a place in the automotive field.

-36-

£70“. _/

-37-

7". .- r ”/71?- .
/ .‘I/vl/r"f'—7 ” _/ l /i {4’ 4‘ i"...

\u

i
‘\
“IA,A" "1
( .‘_(

 

/ "3 €/ -".‘/')..' r‘
. _- ' ,9 _
’/ /.' 7./_1/__{:_,
w . v.’
.. u v . .
_\‘ I
j!
l
, _, W M MM.
-" [7/ ‘) p.¥ . [7,", "‘
~‘ ~ ic‘l;¢". I — ’
. . ‘ fl”, 1/ J / L. L.
z .1 - .
' m." -‘ "’ L- --
’r [I [1" ‘3‘“q.
/ 3/ f
.'./' ‘
//l~ , , .
I/ f f/
. 1 ,’
. A

\\‘
A .\\r
/ ‘
r. - s \\
4 \
I! ‘ / -' ‘ “e\» \
l
I . '
l/ 1‘ t i
C
. \

I _ , __ _ .A__ __- ,3 _. _,_o._. -

-- 3—— .-~—»-—. I - :1

‘s
.
l. r
. _, \
~ r
,-" ~ \
’ A
. S, J
up, . _-/' .\ ; r
i 'f -' I /./’ ‘
,/, ' \ 3/ \\ ’. .
f ' ‘
/ I ' l
I .l 1' / ~
- I I \ “I
.A’ I ‘ v ‘
\ , ./' N 4 ‘\
\ ‘ v . ‘ \ \ V
/ t . . _‘ ~
\ ‘ ‘i \ e
‘1; I q \ ‘\
A \
x' «x
, , ‘\
,l I 1‘
\,._—-‘
\

 

 

 

5-

 

\
.. w.--‘~- .- ...., .b—a-H-n-uo—n—‘.

!
.—
I

 

.k“.-- ”’- ..—-ew a“. w—. .“~————-—J

L._._____.-___e - .. .--

AL...

 

 

_.

 

 

7”" /?/4 r 7.

-‘I'

/-?.x;:-:,»:r"/g g.“

-...-,-———-—.

 

272% £11. "5.5 g? 3725i

‘~0é5’/'[/? 634.) #21 '12-?

 

 

“:I

I
I

’2‘.

-... -. r

 

--..-.~ ~-——.—-—

I I

1

I

1 9 1”” 5&29?L/7’/ A/A’ m 1:: .,
I

.. 7737'. 7‘1:
n
3
0
fi
h

3
’D
r
C
”I

I
I
'I
II
I
I
I
I
I
I
I
i
I

 

far-:7— .“ -'
1"
K
J
C
J

. .\9-

12
I 147/ 27f”?

' -.L:A.
h.._ _. ‘2‘-

 

 

.“ ,1
:‘ I"! £48719: [32.7” 7
MW; / 444/? wig -« " g
I,
1.
I

 

 

g mew pf e 19 new f/c I, G‘z/fi’fl/nmj
[1.6.17

 

 

 

 

 

 

a»
m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I J

74V?" ‘" 1. II I 7‘3
Ilzlz‘T-‘r : N. r

.24.;45 1 .II \ fl?

 

 

 

 

 

I, ‘n ‘a-

LIT/(71': A Z- . «kg. //y’£/::/ . f ‘77.”?

14:9;

 

-41-

 

 

 

/57' (“7’1”” ‘7' 2‘
1' v

I I" ‘ .- .
/ ,‘r ”IN. A. .-.. ..._ ,_ ..

 

 

—_ ._.___. - _._. - K _.__.
.
/“' ’3‘ f :2 ""/, .'...‘ s."
L/\ "I a 4 I a u ‘ u ‘1 A4 .
‘ ‘ '
’7 ,I\'{. (5 ‘7" ' I q ' I
: ‘ x ‘ C J/I /‘V0a/JZA..;£'
' 3 .. ........ ._.__..._.....-,.... _...
i x Y
i g 1
‘ 1er :' 7‘ r':.-., v
' L : ' M", £51" 1’ L " ad
I I . ' T _..__-__...___.- __...____--,_. 4—H
l

2;
/

 

 

 

 

V ‘ \o ' '
. A ’ ‘ " -_, ~\ “ n
' V in

't ,n a -‘;/£-*£‘1
IV - ~ —-
kl-r ..( t . l _

m.-m—-...—

 

. d-
. / ('Iv', (5

BIBLIOGRAPHY

A. BOOKS

 

Shepherd, D. G., £5 Introduction £2_the Gas Turbine. NeW'York, Van
Nostrand Co., Inc. 1949.

 

 

Vincent, E. T., The Theo and Design 2: Gas Turbines and Jet Engines.
New York, MCGPaWbHi Book Co., 1950.

 

 

Zucrow, M. J., Principles 2: Jet Progulsion and Gas Turbines. ‘NeW'York,
John'Wlley & Sons, Inc. I945.

 

 

Timoshenko, 8., Vibration Problems in Engineering. NeW'York,'Van Nostrand
Co., 1937.

 

B. PERIODICAL ARTICLES

 

Colwell, A. T., Bartlett, K. M., Cummings, R. E., "Seven'ways to Produce
Turbine Blades,“ SAE Journal, 57:48, June, 1950.

Smith, G. Geoffrey, "Rover's Turbocar," SAE Journal, 58336, July, 1950

A?

I! 5”!“ ”CC n:%“..~éa v
'1 a.“ -.' ' :3 :r' u, ‘~-.’ 2 ,4 .

Feb 16"§’
M“ 5 '“

 

1”" “N15 ‘ -
emu“. m 3;.

 

 

9
8
6
7
2
4
1
3
0
3
9
2
1
3

 

Ill
.llll
|
|
ll
l
|
I
I
I'll
I
ll
III
II
I'll
'Il
ll
IIII
III
III
III
|
l
I
I'll
II
I
I
II:
III
I
l
I
|