Calhoun: The NPS Institutional Archive

Reports and Technical Reports All Technical Reports Collection

1964-04

Radial-Inflow Turbines

Vavra, M. H.
Monterey, California. Naval Postgraduate School

http://hdl.handle.net/10945/31854
 64T-4
TN NO.

UNITED STATES
NAVAL POSTGRADUATE SCHOOL

DEPARTMENT OF AERONAUTICS

TECHNICAL NOTE
NO. 64T-4

RADIAL- INFLOW TURBINES

by

« H.^VAVRA
II
T^ Tn.-\.\/ a
 TN NO. 64T-4

UNITED STATES NAVAL POSTGRADUATE SCHOOL

DEPARTMENT OF AERONAUTICS
PROPULSION LABORATORIES
TECHNICAL NOTE
NO. 64T-4
 TABLE OF CONTENTS

Page

INTRODUCTION 1

DIMENSIONLESS PERFORMANCE PARAMETERS 6

SPECIFIC SPEED AND SPECIFIC DIAMETER 9

CALCULATED PERFORMANCE VALUES 13

CONCLUSIONS 16

BIBLIOGRAPHY 17

Fig. 1 Specific Speed Characteristics of Turbines 18

Fig. 2 Influence of Specific Speed on Efficiency of Radial 19

Turbines

Fig. 3 Calculated N -D Diagram for Single-Stage Radial 20

Turbines

Fig. 4 Radial-Inflow Turbines 21

a) Dimensions
b) Velocity Triangles

Fig. 5 Expansion Process in Radial Turbine represented in 22

Entropy Diagram

Fig. 6 Outer and Inner Radii of Discharge Annulus of Radial 23

Turbine for Different Ratios Q. and R /R,
A I 1

Fig. 7 Calculated Parameters from Fig. 3 24

Fig. 8 Calculated Stage Parameters (T|. , k. , 39 ) 25

X IS £-

Fig. 9 Calculated Stage Parameters (r*) 26

Fig. 10 Assumed Change of Velocity Coefficient Y of Rotor with 27

Relative Discharge Angle 3_

Fig. 11 Calculated Stage Parameters (71., k. , r*) for Y 28

* 1S
of Fig. 10

ii
 INTRODUCTION

During the past year the writer has been associated with three pro-

jects where radial turbines were used instead of axial turbines. In a 1

MW Bray ton-cycle space power plant, operating with a mixture of Helium

and Xenon to obtain reasonably high molecular weights of the operating

fluid, a radial-inflow turbine of 7 in. diameter was found to be the best

choice to generate abt. 6000 KW at 170 psia and 1500 F inlet conditions.

Its speed would be 38,000 rpm and an efficiency of about 897» should be

obtained. This nuclear gas turbine would have a centrifugal compressor,

and with a radial turbine it is possible to completely balance the axial

thrust of the rotor. This condition is necessary since it is desirable

to use gas bearings with a limited thrust capability, to avoid a lub-

oil system which could contaminate the reactor and the heat transfer sur=

faces. Besides that, the simplicity of the turbine itself, its inlet

and discharge manifolds, and the ease with which they can be arranged in

the loop were additional considerations which lead to the choice of a

radial turbine.

In a project for a neclear gas turbine of 1.5MW for land use, that op-

erates with air in a closed system at high pressures, a radial turbine was

preferred to an axial one for similar reasons, A wheel of about 8 in, dia-

meter at 40,000 rpm can generate 5300KW at inlet conditions of 1100 psia and

1400 F at a pressure ratio of about 2. Efficiencies of 887o are expected.

With a radial turbine it is possible to obtain a favorable arrangement of

the inlet casing which is then stressed only by a small pressure difference

Thrust balancing was of great importance at the high operating pressure

although oil-lubricated outboard bearings will be used. As planned, the

unit could have a simple tapered - land thrust bearing to take up the

resulting axial force of about 2000 pounds. Ruggedness, simplicity, and

cost were other factors that made the radial turbine more advantageous

than an axial type.

A very special application of a radial turbine was found for a turbo-

pump unit of a rocket motor with extreme pressures. Because of the small

volume flow rate in the turbine the leakage losses in an axial turbine

would have become excessive without special sealing arrangements at the

blade tips, which were thought undesirable because of the distortions of

the casing at the high pressures. Complete thrust balancing is necessary

and the radial turbine in question is symmetrial with dual discharge be=

cause of this requirement.

Radial turbines have been in use for many years, especially as expand-

ers for air-conditioning systems in aircraft, for cryogenic purposes, and

small gas turbines. At present the tendency toward smaller and smaller

turbomachines has given a great boost to radial machines since they can

cover a field where the axial turbine cannot operate at good efficiency

because of its small blade heights. Interesting applications and designs

are shown in Reference LI J . While, until recently, it was not possible

to fabricate small radial type rotors from forgings, such designs have

now become a reality with modern production methods, especially by means

of electric discharge machining. Excellent tolerances can be maintained

for very thin blades since no force is applied to the rotor blades during

machining.

The life support system of the Mercury capsule has two compressor

wheels of 2 in. diameter running at 40,000 rpm. Efforts are now made to

build a self contained life support system for extravehicular space explor-

*References - refer to bibliography at end of paper.

ations which would use a turbo-compressor set with a radial turbine of

4.4 in. diameter, operating at 80,000 rpm, and driving a centrifugal com-

pressor with a wheel of 1.32 in. diameter L2J. It is of interest to note

that the waste heat of the astronaut and his C0 ? production will be used

as the sole heat source of the little turbopump and refrigeration unit.

Recently, the literature dealing with radial turbines has increased

greatly. Concepts that are somewhat foreign to the designer of axial tur-

bines are being applied, such as specific speed and specific diameter,

which have been used in hydraulic turbine designs for many years. Actually,

radial hydraulic turbines, so died Francis turbines, have been built since

1900, and the designers of radial turbines for compressible fluids now try

to benefit from the experiences made in the hydraulic field.

However, some of the recommendations and design criteria one finds

in the literature are confusing and contradictory. From Fig.l, for in-

stance, it seems that the only requirement for high efficiency is a high

specific speed N . (This value and others will be defined later in the

paper). However, Ref.5 indicates also that the value of \]Jc for best ef-

ficiency should be between 0.69 and 0.725. The corresponding head coeffi-
2
cients k. (C
v /U-.) , which the writer prefers to use, are therefore
is o 1

between 2.1 and 1.9. Fig. 2, which has been adapted from Ref.6, seems to

convey the same recommendations, but it contains one curve from another

source (Hydro-Aire, Inc.) that seems to indicate that there exists an

optimum value of the specific speed, so that the efficiency tends to de-

crease if this value of N is exceeded. In Ref.6 it is shown that the value

s

of IL /C at which the best efficiencies were measured is 0.65, correspond-

ing
° to
k, 2. 37.
is
 .

Ref.4 introduces the so-called specific diameter D , and in Fig. 3 it

is shown that a relation exists between D and N for optimum

r efficien-
s s

cies. Further, the highest efficiency occurs at values of U, /C between

0.6 and 0.7, or at values of k. between 2.78 and 2.04.

is

The results of Figs . 1 and 2 differ from those of Fig. 3. The first-

mentioned figures show the efficiency TL of a turbine if it is referred

to the isentropic enthalpy drop from the total inlet pressure to the total

discharge pressure, whereas in Fig. 3 are given the efficiencies T) refer-

red to the enthalpy drop from the total inlet to the static discharge

pressure. If TL is used as a criterion, two turbines with the same effi-

ciency Tl can generate different powers for the same inlet conditions and

the same static discharge pressure at the wheel exit, depending on the

velocity of the flow at the turbine discharge. It is possible of course,

to recuperate part of the kinetic discharge energy by means of a diffusor,

similar to the use of a draft tube in hydraulic turbines, but this effect

should not be considered if different turbines are compared which must

produce the same power. It is necessary, at least, to compare not only

the "total-to-total efficiencies TL but also the "total-to static" effi-

1

ciencies T).

In order to clarify the situation the writer has undertaken theoreti-

cal studies to define more clearly the parameters that bear on the oper-

ating performance of radial turbines. The purpose of the study was to

establish the range of applicability and the efficiency of such turbines.

In particular, the writer does not like to operate with parameters that

are semi-dimensionless , which require changes and conversion factors if

different systems of units are used. Not only is the use of such para-
meters awkward but in most cases they do not show up the actual physical

nature of the phenomenon they describe.

The study is limited to radial turbines where the blades extend from

the outer diameter of the rotor to the discharge annulus, and whose meri-

dional channel is such that nearly uniform discharge velocity can be ex-

pected. In particular, the rotor blades are supposed to be in meridional

planes at the inlet to avoid bending stresses in the blades. Rotors with

such "radial" blades can operate at higher peripheral speeds than others

for given physical properties of the disk material. Furthermore, since

the use of multi stage radial turbines is associated with considerable

design difficulties, the turbines treated in this paper will be single-

stage units with whirl-free flows at the discharge to minimize the so-

called leaving loss, which equals the kinetic energy of the absolute

flow at the turbine discharge.

The investigation has been carried out because of the research work

in radial turbines that is now under way, or is planned, at the new Pro-

pulsion Laboratories of the United States Naval Postgraduate School. It

is hoped that these efforts will create a better understanding of the

flow phenomena in such machines, and establish more clearly when and

where a radial turbine should be used. The study is preliminary in

nature. Three dimensional flow conditions are not treated, and it does

not touch on the effects of Mach and Reynolds numbers, although the lat-

ter have a pronounced influence on the performance of turbines, both at

very low and very high values.

 DIMENSIONLESS PERFORMANCE PARAMETERS

The analysis of the radial turbine of Fig. 4 will be carried out with

the simplified one-dimensional approach described in chapter 15 of Ref.7.

The conditions along a mean stream surface from station 1 to station 2

are considered to be representative of the stage, and R„ is the mean

square radius at the discharge. For the assumed uniform velocity distri-

bution at the rotor discharge, equal flow rates will pass between the

radii R~ and R„ , as do between R« and R~ . • The expansion process in the

turbine is shown in Fig. 5.

For assumed radial inlet of the relative velocity W, , and axial di-

rection of the absolute discharge velocity V~ , there are from Eqs.15 (30)

and 15 (32) of Ref.7

» - ji = cpsin^ /T—5» - -£- (1)

where, in addition to the symbols of Fig. 4,

cp velocity coefficient of nozzles

¥ velocity coefficient of rotor blades

r* theoretical degree of reaction of the

stage (see Fig. 5)

X function, given by Eq.15 (27) of Ref 7

C theoretical velocity of isentropic expansion

from P to P„.
o 2

V m velocity ratio of stage

Neglecting the reheat factor f in Eq.15 (27) of Ref.l, and assuming

complete recovery of the relative velocity W, in the rotor (^p 13 1) , there

is with Eq.l, and the symbols of Fig. 4,

pg - C =
2
<p I si„\ [ 2 + (R
2
/ Rl )
2 Qi - I)] cos\
(
^^ - lj(2)

This relation shows that the degree of reaction r* of a radial tur-

bine with radial blades at the outer diameter, and axial absolute velocity

at the discharge, cannot be chosen arbitrarily for chosen radius ratios,

and ratios V „/V of the through-flow velocities. .

m2 ml

From Eq.2

r* - -JL_
l + c

The following performance parameters of the stage are introduced:

2
k - Ah / (u\ /2 g J) (Head Coefficient) (3)
IS IS J.

2
k - Ah / (D. /2 g J) (Work Coefficient) (4)
w l
Ah
TL =
2vT (Total-to Static Efficiency) (5)
is
Ah
TL - (Total-to-Total Efficiency) (6)
J"?
is

The head coefficient k . represents the isentropic enthalpy differ-

ence Ah. of Fig. 5, from the total inlet pressure P to the static dis~
Is .
o

charge pressure p- that can be handled by the stage } as a multiple of

the "kinetic energy" of the peripheral speed U, at the outer radius R 1

of the rotor. This head coefficient is related to the velocity ratio u

of the stage by

N2 + C
. JL ('f£ 1 ,.
* V
)

(7)
is ,
V
,2 n
u
l
J r
2 sin 2 o
cp
.

The work coefficient k expresses the useful mechanical energy Ah of

2
Fig.5, generated by the rotor, in terms of U, /2 g J. From Eq.15 (28)

of Ref. 7
 2 cp2 sin Q'
1

k. 1 + C (8)
is

and, from Eq.15 (19), if T] is replaced by T) .

\ \
\ 1 - < V /V
2
cot
2
cr. 1 < V /U
2
m 2 m l) m2 l)
k. k.
is is

2
(9)
2
k. - (V
N /U,)
1'
is m2

Other useful relations are:

— m2
1
=
T~
m2
ml
cot «1
,^
<
n\
10 >

and
-i /V R i

The nozzle discharge angle a cannot exceed about 80 to 84 . The

writer has carried out tests on a turbine inlet scroll with guide vanes,

where average flow angles ot of about 83 were measured The losses

from the scroll inlet to the rotor inlet were about 67„ of the theoretical

velocity head from inlet to nozzle discharge, corresponding to a value of

Cp of about 0.97. For the calculations presented later a velocity coef=

ficient cp = 0.96 will be used. The velocity coefficients Y of the rotor

blades are not known with great accuracy, in fact, this investigation

was undertaken to establish more precisely the values of Y for rotors

with different geometry.

The radius ratios R„/R, cannot vary greatly since the outer radius

R- at the discharge should not be larger than about 757„ to 807o of R, to

2o 1
obtain reasonable wheel shapes. Fig. 6 shows that for values R„/R, smal-

ler than 0.4 the available annulus area at the discharge becomes too

small, and for R ? /R 1 larger than 0.6 the radius R„ becomes excessive.

Hence R„/R 1
can vary from about 0.4 to 0.6.

One major difficulty with the design of radial turbine wheels is

produced by the fact that the volume flow rate at the discharge is larger

than that at the inlet, and that this flow rate must pass through the
2
outlet annulus which is a fraction only of the projected area tt R of

the rotor. Hence in many cases it is necessary to use through- flow

velocities V „ that are larger than V , at the rotor inlet. In general,

mZ ml
therefore, the three parameters
12R»/R
&, , n
1
, and V
mz
/V ,
ml
m V /W,
2 1
can be

varied to satisfy particular operating conditions. However, centain com-

binations of these quantities may produce angles f3~ that are larger than

those which can be realized in actual rotors. It must be borne in mind

that the velocity triangles of Fig. 4(b) represent the actual average ve-

locities and flow angles at the different stations. Hence the actual

blade angle (3_ at the rotor discharge, that produces a flow angle |3

must be larger than 3„ because of the thickness of the blades, and be-

cause of an effect similar to that which is known as "slip" in centri-

fugal compressors. For these reasons the flow angle (3 cannot exceed

about 65° to 70°.

SPECIFIC SPEED AND SPECIFIC DIAMETER

These concepts are obtained from similarity considerations, in parti

cular from Buckingham's tt - Theorem l8J if changes in mass density are

ignored. Specific speed N and specific diameter D are regularly used

s s
for hydraulic pumps and turbines. For compressible fluids it is general-

ly accepted to apply these quantities to the station where the lowest

mass density of the fluid occurs For radial turbines then

si!, 4

where: N rotative speed of impeller

Q~ " volume flow rate at rotor discharge

H = isentropic head

Further ,

D H *
D 5 (13)
s 2
\
where D. is the outer diameter of the rotor. Evidently the values obtain-

ed from Eqs.12 and 13 for particular designs depend on the units that are

used to express the different quantities. Even in the English system of

units a number of conversion factors must be used to correlate different

values.

The values of N and D in Figs.l, 2, and 3 are calculated with the

s s

usually applied units of

N in rpm
3
Q2 in ft /sec

H in (ft-lb)/lbm

D, in ft
0// 1

Then N and D have the dimensions (ft

v /sec ) min , and
s s '

2 -i
(ft/sec ) , respectively. Sometimes a so-called dimensionless specific

speed N , and a specific diameter D . are used, where N is introduced in

si si
3
revolutions per second, Q_ in ft /sec, D. in feet, and H in ft-lb/slug

10
 —

2 2
or ft /sec . Then

N - N (810.55)
s sl

and

D - D (0.420)
s al

Adding to the confusion is the situation that some sources intro-

duce the speed N in radians per second. The resulting specific speed

N is then related to N by
s2 s

N - N (129.01)
s s2

It can be shown that turbines with similar geometry and equal spe-

cific speed and specific diameter must have the same efficiency if Mach

number and Reynolds number effects are ignored. Furthermore it is evident

that N and D represent the speed and the diameter, respectively, of a

9 S

turbomachine that discharges a unit of volume flow rate at a unit of head.

It is possible however to relate these concepts directly to the per-

formance rparameters that were established earlier. For N and D of

s s

Figs.l, 2, and with

^
3,

N -
1

2
Q2 A V R V 14 >
A " l
" ' C <
2 m2 m2

H Ah J k. ~
is is 2 g

there are
%
V /U
N 384.55
«A m2 l>
(15)
s 3/4
k.
is

k *
ls
and D 0.3984 (16)
s
V /U %
«A m2 l>

11
Further
4 2
k. - 2.3472 (10 ) (N D )" (17)
s g s

and
C
A
V /U
m2 l
= 6 7624
"
<
10
" 6
>
k
is
^ N
s
2
< 18 >

The factor C of Eq.14 is shown in Fig. 6. It represents the annulus

area at the discharge as a fraction of the area of a circle with the

wheel radius R. . Hence, the quantity C. establishes the ratios R„ /R.

and R /R of a turbine rotor with a given radius ratio R~/R- , operating

at a certain value of V /U, . The value £. has not appeared in the pre-
m^ 1 A
ceding calculations, nor is there a way to establish it by means of other,

more detailed flow calculations without knowing more about the interde-

pendence of wheel geometry and flow losses. The principal advantage of

introducing the specific speed concept lies exactly in the fact that £

can be obtained from it. However instead of making a detour via the spe-

cific speed it would be just as easy and useful to establish the dimen-

sionless quantity Q v o/U, as a function of, say, the head coefficient

k. , since Eq.17 shows that giving a relation between N and D , as in

IS s s

Fig. 3, is tantamount to specifying the magnitude of k. .

XS

These investigations also show that it is not sufficient to state

simply that a turbomachine must have a certain specific speed for best

efficiency, without implying or specifying that it must also operate at

a definite head coefficient.

Although Fig. 3 presents calculated data only, a curve was drawn

through the points that give the optimum efficiency for chosen values of

N . For the corresponding values of D and T). it is then possible to

S S X

12
draw the curves of Fig. 7, which in turn can be used to establish the re-

lation between C. V /U. and k. with the efficiency

J T). as parameter.
A m2 1 is 1

Fig. 7 indicates that a value of C V /U, = 0.11 produces the highest

2

efficiency with a k.
lb
2.35, corresponding to N =68 and D = 1.47.
S S

If the curves of Figs. 3 and 7 were correct they would definitely be at

variance with the presently used criterion that a radial turbine must

have an N of at least 80 to reach optimum

r efficiencies,
s

CALCULATED PERFORMANCE VALUES

Eqs.2 to 11 were used to establish the performance of radial-inflow

turbines for different values of

12
ot. , R-/R,
1
, and V „/V ..
m2 ml
For the data

presented in Figs. 8 and 9, a constant value V m 0.9 was used for the losses

in the rotor, and cp was taken to be 0.96. Because the ratio V /\J, ap-
mzn 1

pears in Eqs.15 and 16 the results of the calculations were plotted as

functions of this quantity. Extremely interesting is the fact that

V has the strongest influence on the performance of the turbine.

2 /Ui

Although not clearly seen in Fig. 8, the curves for o>. 75 and 70 over-

lap almost completely with the curves for ex = 80 . The influence of

the radius ratio is small also and the curves indicate that, the smaller

the value of the ratio V /U, is, the higher the internal efficiency T).
mz 1 1

will be. However, because of the limitations imposed by too large flow

angles P_, the minimum value of V ^/U. cannot be smaller than about 0.25,

depending on R /R,
2

Fig. 8 further shows that the total efficiency TL is not affected

greatly by V ,/U.. Hence, the total efficiencies of 90% to 93% quoted

mz 1
2
in Ref.5 are not surprising, all the more as the quantity (V /U.) is
mz 1

equal to the leaving loss coefficient k^ which is defined by Eq.15 (17)

of Ref.7.

13
 Experience with axial-flow turbines shows that the rotor losses, or

the velocity coefficients Y, are primarily depending on the deflection

of the flow in the rotor blades (see Fig. 15(5), p. 434, Ref.7). A similar

inter-dependence between ¥ and f3„ seems likely in radial turbines. Since

the flow is turned by 90 even though the angle 3„ is zero, it can be

argued that the flow deflection in a radial turbine is 90 + 3 . Based

on this premise the curve for ¥ of Fig. 10 has been established, and the

corresponding results are plotted in Fig. 11. It is seen that the curve

7| vs. V o/U-, now shows a decrease in efficiency for values of V o/U,

below about 0.15. The writer believes however that Y will decrease more

with increasing angles 3_ than given by Fig. 10, and there is reason to

assume that the best efficiency will occur at values of V o/^i between

0.3 and 0.35. The optimum value of V 2

/u i is affected also by the disk

friction losses of the impeller. The frictional moment M f is obtained

from
2 3
Mf - C
f
P U
x Rl

where, according to von Karman,

" 1/5
Cf constant -Re

The mass density p.. is that at the rotor inlet. Ignoring Reynolds

number effects, or taking C f as a constant, the power P f absorbed by disk

friction is
2 2
P M oo - C Pj_ Uj_ R
f f x

The power P generated by the rotor is

P » w Ah, T\.
w is i

Hence the efficiency 7| of the turbine, by taking account of disk

14
 2 3
C P U R
x x 1
friction, is T| * TL 7-7-
o i w Ah
is

Now, with the mass density P_ at the wheel discharge

W = P
2
A
2
2
V
m2
= P
2
C
A * ^ V
m2

and, with Ah. - k. U, /2 g

°
J,
is is 1

P /P
1 2

where K is a constant if Reynolds number influences are ignored. It is

evident from Eq. 19 that the optimum value of the efficiency T] will be

shifted toward higher values of V o/U, because of disk friction.

Now, if the results from Fig. 7 were considered to be correct, hence

if the optimum ratio C. V /U-, would have to be equal to about 0.11, a

A mZ 1

value of C = (0.11)/(0.3) = 0.37 should give the best turbine for

V n /U, 0.3. This value of CA then establishes the radii R„ and R„

m2 1 A 2o 2i

in relation to R
1
for given ratios R_/R
1
. For R„/R 1 = 0.5, for instance,

there are from Fig. 6:

R /R = 0.25 R R = °* 66
2o 1
;
2 i^ l

For given operating conditions and flow rates of the turbine the

value of C. not only gives these radius ratios but also determines the

actual diameter of the wheel and, because of k. , the rpm

r of the turbine.
is

To establish the blade width b, at the rotor inlet it is necessary

to know the relation between the mass densities P. and p . Modifying

Eq. 15(40) of Ref.7 for the present application

2
.
2 +
<WV Y Y/(Y-1)
p
l L
^ Kis - 1

(20)
P 2 <]
2 1 - cp (1 - r*) Y

15
where
2
Y - k (Ul /a o ) (21)
*-f± ig

The quantity a in Eq.21 is the velocity of sound of the fluid at the

total inlet temperature T . With this information the geometry of the

rotor is completely determined.

CONCLUSIONS

The study shows that the efficiency of an inward- flow radial turbine

with radial blades at the inlet of the rotor, and whirl- free discharge,

is primarily depending of the ratios V /U, and Q. V /U, . If the inter-

mz l A mz 1
dependence between Q. V /U, , k. , and T|. were known it would be possible
A roZ 1 is l /

to establish the dimensions of designs with optimum performance. Whereas

the quantity V o/U, seems to be a measure for the effect of the flow de-

flection on the losses, the ratio C V „/U, seems to include also the
A m2 1
influence of the geometry of the flow path,,

Various approximate and tentative limits of the design parameters

are indicated in the paper, but it is hoped that research at the Propul-

sion Laboratories of the U.S. Naval Postgraduate School will eventually

produce more accurate design criteria which are necessary to enhance the

state of the art in this promising field of turbomachines.

16
 BIBLIOGRAPHY

1. W. T.von der Nuell, "The Challenge of Miniaturizing Turbomachinery"

Interavia, Vol.17, no. 8, 1962, p. 1022.

2. F.E. Maddocks; R. V. Hoffman; N. Van Le, "Problems and Advances in

Small-Size Turbomachinery", Symposium on Small Turbomachinery,
Ninth Annual Gas Turbine Conference, Houston, Texas, March 1964
(Abstracts prepared by AiResearch Mfg. Co. Los Angeles, California.).
,

3. W. T.von der Nuell, "Single-Stage Radial Turbines for Gaseous Sub-

stances with High Rotative Speed and Low Specific Speed",
Trans. ASME, Vol.74, no. 4, 1952, p. 499.

4. 0. E.Balje, "A Study of Design Criteria and Matching of Turbo-

machines", Parts A and B, Journal of Engineering for Power, Trans.
ASME, Series A, Vol.84, no.l, 1962, p. 83 and p. 103.

5. Ho JoWood, "Current Technology of Radial-Inflow Turbines for Compress-

ible Fluids", Journal of Engineering for Power, Trans. ASME, Series A,
Vol.85, no.l, 1963, p. 72.

6. W. K. Jekat, "A Radial-Flow Cryogenic Expander Turbine", ASME Paper

64 - GTP - 19, March 1964.

7. M, H. Vavra, "Aerothermodynamics and Flow in Turbomachines" , Wiley

& Sons, Inc., New York, 1960.

8. D.G. Shepherd, "Principles of Turbomachinery", The MacMillan Co.,

New York, 1956.

17
TN 6AT-4

PLOTTED POINTS

190*IF.R., D,«O.Z9'
ij/5 »ao (air)
/
Z AXIAL, D,«a42
P/*^-6.0 LAIR)
A FRANCIS (WATER)
4.90TLF.R.
g/^J«-£.3
(GAS) OfOSlS'.hO
A AXIAL, CL •1.16'
.(AIR) ?
6. AXIAL
*MR)
7 KAPLAN (WATER)
ft9O-|.F.R. > D .0.6*S'
1

^•6.0 (AIR) <20

ft ASME PAPER
No.Sl-GTP-io

IOOO

Fig. 1 Specific Speed Characteristics of Turbines

( Adapted cum Fig* 12 of Reference 5 )

18
 —

TN 64T-4

.4
— -<>>/d.
-rJ r
DATA
?t
9 Jbl -few r^rJr r

r Vq,
.6

.7
>^
7^\
f 7^
RADIAL
_ TUR3IN£_
HYDftO-MRE
IMC
.6

.8

.4

WHEEL SHAPES
.3
FOR II •
Q..

o t
10 20 30 40 SO 60 70 60 90 100

NS"
Q>.70
H f^O.78
J

Re. 2 Influence of Specific Speed on Efficiency

of Radial Turbines

( Adapted from Fig. 9 of Reference 6 )

19
TN 64T-4

if. de--t£» irncii't*

b DELATED TO STATIC m-
M*.'«T rRlSJUHE AUO
TOTAL IHLCT P(Uk»v»»E.

/»,-o*

too »oo»

N,

Fig. 3 Calculated N' -D^ Dic^ron for Sinsle-Stage Radial

r

iuvoir.c:

( Adapted from Fig. 16 of Reference 4 )

20
TN 64T-4-

Fig. 4 Radial - Inflow Turbines

(a) Dimensions
(b) Velocity Triangles at Inlet and Discharge of
Rotor

Fig. 4 (s)

At station (I)

V
V rv»l * W
wl

Fig. 4(b)

At itntlon (2) V„». V

? j
A^

u.

21
TN 64T-A

Fig, 5 Expansion Process in Radial Turbine represented

in Entropy Diagram

A - 2 g J Cp for temperature differences

A - 2 g J for enthalpy differences

h - static enthalpy T - static temperature

H - total enthalpy T - total temperature
H_- relative total enthalpy s - entropy
p - static pressure V - absolute velocity
P - total pressure W - relative velocity
P - total relative pressure U - peripheral velocity
R

22
 U.S.WAVAL POSTGRADUATE SCHooL PROPULSION] LABORATORY M.K. VAVRA'
'RADIAL- FLOW TOR BikJES" TNi 64-T-4 FIG. 6
i

i
-

OUTER AMD IN/NEJ? RADII OF;J>!SCHARSE ANJWULU.S

.

. OP RATMAL TURBINE FOR DIP FERE NT RATIOS ¥bA

iT
- AMD R a /R, . _.
'
'

LL*4i
•

•
: i
'.
.

!
i

H . :

.
!

±m it;
•

i. .
?

___- u m . , •

1
!

._

'

: ffl
\ ;• N
. —.' ::: hL

. — .'..
• >

ii_.4_ — ,
. . i
Mm.mm R_*£-
-
; .

R ^
el V , : Li
K
'

iffi J...;

,- •»»..» —v -/ — I ...»4- i i .. J ~_.- :.._.

|

:-r.'

-t.Po ^ | i

fa :

| - -
I
:

! ,
-... ffl
. :
!fi i'4ti i
_^— — n
.4,,. t.U.l

'
:;:
_—..... -
-_j _
: - ;

ttyFtrffl
I
'

I tttffl
4+HT+-:
'
i
:
r.iii _
I

\
•
i
|

-' - -
H
!
:
.h': ;

"

If!
*2o/^ tirrBEe;

Br
I t'H
m
'
! ,-

-! '
**** ^\
^ *J/i?;.o.;
tJHE tflj

. i
/
>4- i 1
-

V*.*' -*
mm
-

I
•• :"

P pit
ol —»*-
^
I

- .I I...
:

. .

- o.i \
.

r o \
O-S O-fc

L i J r - i 1 ! - . —
23
J- ; -

• '2?
In
1 —
24 fe. *//*/6*
 U-».NftVAL POST&KA.D'.'PO'E SC. HO O L PROfm.S'ON LABORATORY M.H.VAVRA
Radial -flow turbines" tn. 6474 Fl<3. 8
r »oo i

IftSSflS ^1
I

2
*n ;-— —rr-ui 7- .^ \^ i,i , ..

/t !* n £HSS?b.
*<0"*» >K,
•/• ill
£ E
lu _

•
I .. jH .

Is ? h
»- o Htffi-
.
I
-«, .To'
O O ;
.
i

.
. —
. ...
J_-..

o p
x a
.-.,

*
h b±j
I
:
<V 7.
t [_i,i

-
.
I

EH !

.
tlxinxi

h6o r
o o&
; i
-
|

.4.0-—
J

•
:.
- :

ra 1.1

j
;

msi
:

•
•
.
•

1 " -* — — ——
I -
*=^-r-

v h- L . . ..
!

— ——hf~~H
t_i ,

—
m
K . t-j -
ttsra ;

2o —
ffi
I
t-i

0.2 0.4 o.fe o.8

V2C Vrrt2 VEtoc.TY at Rotor discharge if

« PERIPHERAL SPEED AX 1MAX.DIAME?ER
.„ of'rotOR 1
'

.
i

. N
|

:
.

15<
,,

W
•

c,; ~ -
o
•

2 c.c-. o ,
1.0 1
I
I
-i I
•:! .
!
; i i,i ; ; 'i., i

25
 U.S. NAVAL POSTGRADUATE PROPULSION LABORATORY
SCHOOL M. H. VAVRA
./.RADIAL- PLOW TuRBiiMEs"XN.64T-4- FIG. 9

•rr+

cm V 2 . v m2 « VELOCltX". .AT
' ROTOR DISCHARGE h

SPEED AT. MAX. DIAMETER

' .

tlt-l
-l-T-
..PERIPHERAL i
_. Rotor
:

vi!). i 1 1 iii, 1 1 ..I' iii i i ii —

26
At. +M/*4
*//«?/&«
U.S. naval postgraduate school PROPULSION/ LABORATORY M. M. VAVRA
RADIAL- FLOW ToR^kiEs", TN . G4-T-4- Ft<*. 1 I

-CONDITIONS FOR CONSTANT LOSS |U RoToR YeO-9

————
.

j
>
Conditions For losses according to ri&. to

...

28
U-89.353

U 89353
DUDLEY KNOX UBRARY-
RESEARCH REPORTS
U 89353
5 6853 01058281