University of Central Florida

Retrospective Theses and Dissertations Masters Thesis (Open Access)

Convective Heat Transfer from a Cylinder Rotating

in Air
Spring 1983

O. Kemal Pasamehmetoglu
University of Central Florida

Find similar works at: http://stars.library.ucf.edu/rtd

University of Central Florida Libraries http://library.ucf.edu

Part of the Engineering Commons

STARS Citation
Pasamehmetoglu, O. Kemal, "Convective Heat Transfer from a Cylinder Rotating in Air" (1983). Retrospective Theses and Dissertations.
Paper 707.

This Masters Thesis (Open Access) is brought to you for free and open access by STARS. It has been accepted for inclusion in Retrospective Theses and
Dissertations by an authorized administrator of STARS. For more information, please contact lee.dotson@ucf.edu.
 CONVECTIVE HEAT TRANSFER FROM
A CYLINDER ROTATING IN AIR

BY

O. KEMAL PASAMEHMETOGLU
B.S., Bogazici University, 1981

THESIS

Submitted in partial fulfillment of the requirements

for the degree of Master of Science in Engineering
in the Graduate Studies Program of the College of Engineering
University of Central Florida
Orlando, Florida

Spring Term
1983
 ABSTRACT

This study had a two-fold purpose. The initial emphasis

was placed upon the analysis of heat transfer to ambient air from

a rotating cylinder. Three distinct heat transfer regimes can

be identified. For low rotational speeds corresponding to a Rey-

nolds number less than the critical value for initiation of

turbulence, the flow is laminar and the rotation has no effect

on the average heat transfer coefficient. In the transition re-

gion, the heat transfer coefficient depends upon both natural con-

vection and rotational effects. For higher rotational velocities,

the flow is fully turbulent and rotational effects dominate.

Previous analytical and experimental studies have been con-

ducted for all three regions. These studies are summarized in

this thesis and it is seen that there are gaps and limitations in

the existing state of knowledge. Therefore, further study is re-

quired especially for high rotational speeds.

In the second phase of study, an experimental program was

designed to determine heat transfer coefficients at various rota-

tional speeds and heat transfer rates. The rotating cylinder is

designed in such a way that it can be inclined by 30, 45, 60 or

90 with the horizontal. High rotational speeds are possible

yielding Reynolds numbers of 2 x 10

5 The maximum power is

.approximately 400 W.
 ACKNOWLEDGEMENTS

The author wishes to acknowledge the financial support received

from the University of Central Florida for the completion of his

graduate study.

Mr. Pasamehmetoglu wishes to thank the faculty, the graduate

students and the personnel of the Department. He also wants to

acknowledge the help of Dr. P. Bishop, Mr. A. Minardi and Dr. E.R.

Hosler who served on his committee. But especially, appreciation

is extended to Dr. P. Bishop for her assistance, her lasting inter-

est, and valuable suggest ions during this study.

The author is also indebted to Mrs. Sharon Darling for her per-

fect typing job.

iii
 TABLE OF CONTENTS

LIST OF FIGURES . . v

LIST OF TABLES . vii

LIST OF SYMBOLS. viii

Chapter

I. INTRODUCTION 1

II. CONSERVATION EQUATIONS 3

Laminar Flow . 3
Turbulent Flow . 7

III. THEORY AND ANALYSIS. 9

Critical Reynolds Number 9

Slow Rotational Speeds Below Critical Reynolds
Number . . . . . . . . . . . . . 13
Transition Region. . . 18
High Rotational Speeds 21
Horizontal Plate Analogy . 22
Reynolds Analogy . . . 23
Temperature Profile. 24

IV. THE EXPERIMENTAL SET-UP . . 31

Cylinder 31
Supports . 33
Rotation . 33
Temperature Measurement. 35
Heat Input . . . . . . . . . . 40

V. EXPERIMENTAL PROCEDURE . 41

VI. CONCLUSION . . .... . . .. . . . . 47

APPENDIX . 51

REFERENCES 70

ADDITIONAL BIBLIOGRAPHY . . 72

iv
 LIST OF FIGURES

1. Approximate Nlln vs. R~ graph . . . . . . . . . . . 10

2. Ascending and descending side of velocity profiles for

rotation and natural convection. . . . . . 12

3. Temperature field for a =0 .1122, 8 = 1. 0 (cylinder surface),

0.8, 0.6, 0.4, 0.2, 0.1. . . . . . . . . . . . . . . 16

4. Nusselt number variation around the periphery and the

average Nusselt number for a =0.1122 . . . . 16

5. Temperature field for a =0.7407, 8 1.0, 0.8, 0.6, 0.4,

0.2, 0.1 . . . . . . . . . . . . . . . . . . . . . 17

6. Nusselt number variation around the periphery and the

average .Nusselt number for a= 0.7407. . . 17

7. Transition region from laminar to turbulent. . 20

8. NuD vs. ReD graph showing the results of the horizontal

plate analogy and the Reynolds analogy . . . . . . . 25

9. Dimensionless velocity distribution . . . 28

10. Dimensionless temperature distribution 29

11. Experimental set-up showing the various equipment. 32

12. Inclination angles . . 34

13. Thermocouple locations along the cylinder. . 37

14. Graph of temperature vs. emf produced for various thermo-

couple wires . . . . . . . . . . . . . . . . . 38

15. Thermocouple circuit 39

16. Section of the left half of the cylinder . 52

17. Section of the right half of the cylinder . . 53

18. Aluminum core. . . . . . . . . 54

v
 LIST OF FIGURES (Continued)

19. Copper tube . . . . 55

20. Left end insulation cap . . 56

21. Right end insulation cap. 57

22. Section of the left shaft (with accessories). 58

23. Section of the right shaft (with the accessories) 59

24. Front view of the bearing support . 60

25. Section of the bearing support . . 61

26. Bearing casing . . 62

27. Left shaft flange . . 63

28. Right shaft flange . . 64

29. Thermocouple slip-rings and insulation rings. . . . . .. . 65

30. The insulation envelope for the slip-rings. . . . .. 66

31. Brush arm and brush holder. . . . . . . . . . 67

32. Insulation envelope and copper slip-rings 68

33. Brush arm and brush holder. . . . . . . . . . . . 69

vi
 LIST OF TABLES

1. Some Experimental and Theoretical Values for ReCRIT. 11

2. NuD for Various ReD Using Equations. . . . ... 26

3. Expected Surface Temperatures (Within 5 C Accuracy). 44

4. Some Expected NuD and ReCRIT Values . . . . . . . . . 45

vii
 LIST OF SYMBOLS

Symbol Description

A Surface area of the cylinder

B Proportionality constant

cp Specific heat at constant pressure

D Diameter of the cylinder

d Damping factor

E Voltage

g Gravitational acceleration

H Convective heat transfer coefficient

I Intensity

F Friction factor

K von Karman constant

k Thermal conductivity

1 Mixing length

1 Apparent mixing length

c
p Pressure

q
. Heat flow rate

R Radius of the cylinder

r Radial coordinate

T Temperature

~T Temperature difference = T - T00

w
viii
 LIST OF SYMBOLS (Continued)

u Tangential velocity

v Velocity

y Distance from the wall

u Friction velocity
T
a Thermal diffusivity

a Proportionality constant (equation 19)

s Volume coefficient of expansion

y Plume shift angle

Surf ace emissivity

Eddy viscosity

Boundary layer thickness

K Surf ace curvature

Dynamic viscosity

Kinematic viscosity

p Density

w Angular velocity

a Stephan-Boltzman constant

T Shear stress

Tangential coordinate

ix
 LIST OF SYMBOLS (Continued)

DIMENSIONLESS GROUPS

Symbol Description
HL
Nu Nusselt number =
k

3
gf36TL
Gr Grashof number
2
\)

vL
Re Reynolds number = \)

72u d
dr (ur)
Ri Richardson number =
[ r ddr (u)]2
r

\)
Pr Prandtl number = -
a

8 Temperature Difference 8 = (T-T ex) I (T -T )

w 00
a Rotation parameter= Re/~ (equation 17)
1/4
Small parameter = 1/ (Gr) (equations 4 and 6)

u+ Dimensionless velocity = u/u

T

r+ Dimensionless radial coordinate = ru /v

T

T - T
T+ Dimensionless temperature w

y+- Dimensionless distance from wall = (r, - R)uT /v

h Dimensionless radial coordinate = 1 + y/10.R

x
 LIST OF SYMBOLS (Continued)

SUBSCRIPTS

r Radial component

Tangential component

c Natural convection effect

w At the surf ace of the cylinder

00 Ambient conditions

w Rotation effect

D Characteristic length taken as diameter

b At the edge of laminar sublayer

CRIT Critical

RAD Radiant

I Input

SUPERSCRIPTS

Time average

Instantaneous fluctuations

xi
 CHAPTER I

INTRODUCTION

In the field of fluid mechanics and heat transfer, the problem

of rotating bodies is not only of theoretical interest, but is also

of great practical importance. Until about the 1960's, the practi-

cal application of heat transfer in rotating systems was confined

to the cooling of rotating machinery, such as electrical motors and

turbines. After the 1960's, the problem has received more emphasis

because of the further application of rotating systems, and the new

technology has required more sophisticated solutions. For example,

Hickman and others, cited by Kreith fl], have shown the potential

of rotating condensers for sea-water distillation and spacecraft

powerplants in a zero gravity environment. Ostrach and Braun [2]

envisioned the possibility of cooling the nose cones of space vehi-

cles during re-entry by a fluid set into rotating motion by spin-

ning the nose. Fendell [3] has outlined the problems caused by

heat transfer to rotating cryogenic fuel storage tanks in orbit.

Rotating heat exchangers operating on a variety of different prin-

ciples have been introduced in the chemical and automotive [4]

industries.

In this study, the heat transfer characteristics of rotating

cylinders are investigated. The first step was to present the prob-

lem of the convective heat transfer from a rotating horizontal

 2

cylinder. This first step was intended to be more than a mere

literature survey by indicating the gaps and limitations in the

existing state of knowledge. As the second step, an experimental

apparatus is designed for the study of horizontal cylinders. This

experimental set up will be used to determine the numerical con-

stants needed for further analytical study, as explained in Chapter

VI. The design is extended for inclined cylinders with different

angles of inclinations.
 CHAPTER II

CONSERVATION EQUATIONS

In this chapter, the basic conservation of mass, momentum and

energy equations are given in cylindrical coordinates for free con-

vection around rotating bodies. All the equations are given for

two-dimensional steady flow with constant properties where velocity

and temperature are a function. of radius and angle .

Laminar Flow

The general conservation equations are given below as:

Conservation of mass:

(1)

Navier-Stokes equations:
av vrfl av v2
r component: p (v r-ar
r + ~
racf)
r _jp_)
r
(2a)

component :

(2b)

3
 4

Energy equation:

(Laminar flow is observed only for slow rotational speeds, there-

fore viscous heat dissipation is neglected)

(3)

For an unheated cylinder rotating in air, equation 1 reduces

to:

(4)

Equation 4 shows that V is a function of the radial coordinate,

r, only. Therefore, the Navier-Stokes equations reduce to:

v 2
l= 3P
-p (Sa)
r ar

(Sb)

Solving equation Sb with the following boundary conditions:

at r = R

when r -+ 00

The following velocity profile is obtained:

2
V = wR (6)
r

Substituting this velocity profile into the energy equation, equa-

tion 3, conduction is the only mode of heat transfer through the

circular streamlines of air.

 5

But, in the case of a heated cylinder, the flow is laminar for

very low rotational speeds where natural convection effects are pre-

sent. Therefore, the velocity and pressure terms must include the

contribution of the natural convection. Defining the velocity

components and pressure as the sum of convective and rotational

effects:
vr = vr,c + vr,w

vcp,c + v, w

p p + p
c w

The subscripts c and w stand for natural convection and ro-

tation, respectively. By substituting the rotational terms already

obtained, which are:

v r,w = 0

v cp ,w WR2 /r
()P
w
ar = - p

()P
w 0
acp

the following mass, momentum and energy conservation equations are

obtained:

0 (7)
 6

av v 2 av
P[V rz~ + (~+WR) r,c 2 2
( Vcp ,c + WR Ir) Ir] =
r,c ar r 2 acp
r
(Sa)

W2R4 ap 'iv
-P - - - __ a
c + {- i a
[- - (rV )] + .L r,c 2 av <P
r
3 dr ar r ar r,c
r
2 acp2 r acp'c} +pg sin

av rh wR2 v . 2 av v 2
p [V ( ~, c _ - -) + (~ + wR ) , c + ~ (V + WR ) ]
r,c ar 2 r 2 acp r ,c r
r r

(Sb)

1 ap c a a2v av
- - - +{- [1 a (rV~ )] + L <P,c + L r,c} + pg cos
r acp cir rar ~,c r2 acp2 r2 acp

v 2
V aT + (~ + WR ) ciT (9)
r, c cir r 2 a(j) =
r

Assuming an order of magnitude of 1 for VA- and racp, and an order

~,c

of magnitude of o(where 8 << 1) for V and ar, it is seen that equa-

r
tion Ba is dominated by equation 8b. After the order of magnitude

analysis, the momentum and the energy equations reduce to:

av 2 v 2 av aP
( ,c _ wR ) + (~+WR) <P, c] - c
-1 --.---
pfvr,c ar 2 r 2 acp r acp
r r
(10)
a 1 a
+ {-f-~(rV )]} + pg cos
ar r 8r ,c
 7

v 2
V 3T + (~ + wR ) aT (11)
r' c ar . r 2 3 -
r

These three equations, equations 7, 10 and 11, have to be solved

simultaneously to obtain the exact temperature and velocity distri-

butions around the slowly rotating cylinder. Because of their com-

plexity, they require numerical solution. A numerical solution tech-

nique for similar equations is given in the article by Goettler and

Fillo [5]. The results of this study are summarized in Chapter III.

Turbulent Flow

When the rotational speed is high enough to generate fully tur-

bulent flow, the effect of the natural convection is negligible.

The velocity, p~essure and temperature expressions can be written

as follows:

Vr = Vr + V' r

p = p + P'

T = T + T'

where denotes time average values and ' denotes instantaneous

fluctuations. When these expressions are substituted in the lami-

nar flow equations, after an order of magnitude analysis similar

 8

to the laminar flow case, and after taking the time average of each

term, the following equations are obtained:

0 (12)

But Vr = 0, therefore V is a function of + only. Then, the Navier-

Stokes equation in the -direction becomes:

vl_ [_!_ l_ (rV!h)] - a (V' V' ) = 0 (13)

dr r ar ~ ~ r

and the energy equation becomes:

a
a; (V'
r
T') (14)

These equations for the turbulent boundary layer are solved by

Teng and Grief [6]. The results are summarized in Chapter III.

The equations for the transition region are not included for

two reasons. First, these equations are very complex, and the

order of magnitude analysis is difficult to apply. Second, no

attempt for such a solution has been found in the literature. This

region is briefly discussed in Chapter VI.

 CHAPTER III

THEORY AND ANALYSIS

The heat transfer characteristics of a rotating horizonta l

cylinder can be analyzed for three different heat transfer regimes .

These regions can be observed by looking at a plot of Nu vs. Re

obtained in earlier experimental studies (Figure 1).

For these different regions, some experimental and theoreti-

cal work has been done. By examining these studies, a complete

overview can be made for the problem. In the rest of the chapter,

these three regions will be considered more closely.

Critical Reynolds Number

The Reynolds number is expected to reach the critical value

when the tangential velocity of the surface of the cylinder becomes

approximately equal to the velocity of the vertical free convec-

tion flow at the side of a stationary cylinder.

Although there is an exact solution for the natural convec-

tion flow around a cylinder, an approximate value is used by An -

derson and Saunders [7]. They used the well-known solution for

free convection with a vertical plate, assuming that the velocity

at the side of the horizontal cylinder is approximately the same as

the velocity at the top of a vertical plate whose height equals

the diameter of the horizontal cylinder.

9
 10

300

I
YB
I
A I
I
I
2 3 Log ReD 5
Re 10
CRIT 2
Re
CRIT 1

Region A: Laminar region

Region B: Transition region

Region C: Turbulent region

Figure 1. Approximate NuD vs. ReD graph.

 11

The same authors, considering that the boundary layer is thin,

took the maximum velocity in the boundary layer and compared it to

the surface velocity of the cylinder. This leads to a functional

form for the critical Reynolds number, although the numerical value

of the constant is uncertain. These velocity profiles mentioned

above are shown in Figures 2a and 2b.

Table 1 shows the experimental and theoretical values of the

critical Reynolds number.

TABLE 1

SOME EXPERIMENTAL AND THEORETICAL VALUES FOR ReCRIT

D (inches) 1.0 1.8 3.9 1.8*

Experimental 320 780 2100 3000

Theoretical 240 625 1800 2540

*Pressure 4 atm.

It can be observed that the theoretical values are consistent-

ly lower than the experimental values. The theoretical values

are calculated based on the equation derived by Anderson and Saun-

ders [7], by following the assumption described above. The equa-

tion is:

ReCRIT = 0.93 /Grj/P-i (15a)

This reduces to:

 12

Figure 2a. Ascending side.

Velocity profiles for rotation and natural convection

------ Natural convection

Rotation

Figure 2b. Descending side.

 13

ReCRIT = 1.09 /Gr (15b)

for Pr= 0.73 which is approximately the case for air.

There is an uncertainty in the numerical constant of equation

15b due to the assumptions. But there is also an uncertainty in

reading the experimental curves. Therefore, the result of having

the theoretical values consistently lower than the experimental

ones is not significant since we have a certain margin of error in

both values.

Therefore, the critical Reynolds number region requires a more

detailed analysis in a future study.

Slow Rotational Speeds Below Critical Reynolds Number

The laminar region shown in Figure 1 is characterized by a

constant average Nusselt number, up to the critical Reynolds num-

ber. By examining Figures 2a and 2b, it is seen that the velocity

on the ascending side (a) due to the combined effect of natural con-

vection and rotation is higher than the velocity on the descending

side (b) of the cylinder. Therefore, a higher heat transfer coef-

ficient is expected on the ascending side than the descending

side. Although the experimental data show a constant average Nus-

selt number, the local Nusselt number is a function of and

changes around the cylinder.

In this region, the average heat transfer coefficient

can be calculated from a heat transfer correlation for a

stationary cylinder subject to natural convection. It is:

 14

~
4
N uD = 0 4 7 ( G rD P r) (16)

In slow rotational speeds, the flow is observed to be laminar,

except for the plume region. The plume region is defined as the

region where the flow separates from the cylinder surface, and it

can be seen in Figures 3 and 5. A detailed analysis of this lami-

nar flow is made by Goettler and Fillo [SJ.

In this study, laminar convective heat transfer from a heated,

slowly rotating horizontal cylinder to ambient air is examined

analytically and experimentally. For low rotational speeds, below

the critical speed for the initiation of turbulence, laminar boun-

dary layer equations, equations 7, Sa, 8b and 9, are deduced on

the basis of an order of magnitude argument. A parameter based on

the rotational Reynolds number and Grashof number is established

characterizing the influence of rotation upon the free convective

motion. This rotational parameter, o, is defined as follows:

a = ReR I /GrR (17)

In addition to rotation, the curvature effects are also taken into

account. For this study, in the calculation of dimensionless

parameters, the radius of the cylinder is taken as the character-

istic length.

The boundary layer equations are solved numerically and the

solutions are restricted to the following cases:

 15

(a) a ~ o where
rotational effects are small and curvature
effects are retained, and

(b) a - 1 where rotational effects are important and curva-

ture effects are neglected.

o is the boundary layer thickness and is defined as:

Y.4
o= 1 I (Gr R ) ( 18)

The experimental values are obtained for a= 0.1122 and a=

0.7407.

In Figures 3 through 6, which are obtained from Goettler and

Fillo [5], the experimental results are compared to the numerical

solutions for these two a values.

The nomenclature used in these figures is as follows:

Y.4
s =small parameter= l/(Gr)

K surface curvature = 1

<P = angular coordinate

e temperature difference= (T-T00 )/(Tw -T00 )

h = radial coordinate = 1 + y/lOR

The following notation is used in all figures:

experimental results:

perturbation (a-o) solution:

first order (a~l) solution:

The figures show that the perturbation solution is in better

agreement with the experimental results, compared to the first

order solution.
 16

\ 0

\. ~
'' . ,,'
\'

Figure 3. Temperature field for a =O .1122, 8 = 1. 0 (cylinder

surface), 0.8, 0.6, 0.4, 0.2, 0.1 [5].

90 180 . -90
O'--.....a..~.____,,_~_._--.___._~.___.~...._--.__-...__.

Figure 4. Nusselt number variation around the periphery and

the average Nussel t number for a= 0 .1122 [ 5] .
 17

\
\
\ .
\
\
\
h

Figure 5. Temperature field for a =0.7407, 8 = 1.0, 0.8, 0.6,

0.4, 0.2, 0.1 [5].

ac
" .2
.1

Figure 6. Nusselt number variation around the periphery and

the average Nusselt number for a= 0 J407 [5].
 18

An important result of this study is that it confirms

the proportionality of the plume shift, y, to rotation parameter,

0:

y = (JjJ ( 19 )

The value of a determined experimentally is 1.16 and it is h igh er

than the value (a = 0.6735) calculated when o-cS, when the curvat u re

effects are important. a calculated for 0-1, when the curvature

effects are neglected, deviates from experimental observation in

that it is not linearly proportional to the rotation parameter, a.

However, the values correspond more closely to the experimental

data than those where a = 0.6735, whenever a > 0.1.

Transition Region

This corresponds to a region B in Figure 1. In this region,

turbulence starts and the effect of both natural convection and

rotation are present.

Examining the experimental Nu vs. Re curve of Figure 1,

D D
just after the critical Reynolds number, for low Grashof number,

a sudden decrease, although small, in Nusselt number is observed.

This may be due to flow separation phenomena which are also ob-

served in forced convection situations, such as in the case of flow

over a circular cylinder.

For this region, no particular studies exist, neither experi-

mental nor analytical. An effort to explain the transition re-

gion, although it is far from being a detailed analysis, is made

by Richardson [8].
 19

Anderson and Saunders [7] expressed the critical Reynolds num-

ber by an equation of the form:

Grl/2

Also, from equation 16, for free convection in laminar flow the

Nusselt number is found to be proportional to the one-fourth power

of the Grashof number:

Nu ..., G l/ 4
D rD

Therefore,
2
Re ..., Nu (20)
CRIT

Later in this chapter for high rotational speeds where the flow is

completely turbulent, a relation for the N uD in terms of R~ is

given from the study of Anderson and Saunders [7].

Nu = 0.10 R eD2/3 (21)

D

For these two equations, equations 20 and 21, the intersection

might be expected to occur under conditions where there is no tran-

sition region. This would probably occur at the Grashof number

which corresponds to transition in natural convection for the sta-

tionary cylinder. The corresponding Nusselt number is about 47.

The line for the critical Reynolds number is made to intersect

the line for the fully developed turbulent flow at this point, and

the triangular transition region is obtained as shown in Figure 7.

It is observed that the experimental data fall in this region,

and therefore, agree with the explanation.

 20

47 - -------- --- ------------ -- ----------------- ---

10

0
2 3
3xl0 3xl0 Re
D

Figure 7. Transition region from laminar to turbulent.

 21

In the transition region, hoth the rotational and the natural

convection effects are present. Therefore, the NuD is a function

of both the Grn and the Ren Although there is no analytical

expression available, two empirical relationships are found in the

literature. The one derived by nropkin and Carmi [9] is:

(22)

It is in very good agreement with the relationship correlated by

Etemad flO] as:

2 0.35
Nun= 0.11 [(0.5 Ren + Grn) x Pr] (23)

High Rotational Speeds

This region is shown as region C in Figure 1. In this region,

the flow is fully developed turbulent and the Nusselt number is

dependent on the Reynolds number only; this means that the rota-

tional effects are dominant and the natural convection effects can

be neglected.

There are two empirical equations available in this region.

One, derived by nropkin and Carmi [9], is given as:

(24)

The equation derived by the same authors for the transition

region, equation 22, reduces to this equation when the Grn term,

accounting for natural convection effects, is neglected.

 22
The second empirical equation correlated by Etemad [10] as:

0.7
N~ = 0 076 (ReD) (25)

reduces to equation 23 when the term GrD is neglected.

There are also three analytical studies that have offered ex-

planations of this region. The first two are based on analogies,

namely the horizontal plate analogy [7] and the Reynolds analogy

[11], and they directly obtain a relation for NuD in terms of R~.

This third one is an analytical solution for the temperature dis-

tribution where comparison is made with experimental results.

Horizontal Plate Analogy

For high rotational speeds, an analogy has been established

between the free convection from a heated horizontal plate facing

upwards and the turbulent transport from a rotating cylinder.

For a heated plate facing upward, the heat transfer relation

is given by [12]:
3 2 c 1/3 (26)
HkL = O.l 4 [L p gplT (_p_)]
2 k

To obtain the heat transfer from a rotating horizontal cylinder,

the term (gSlT), which represents the buoyancy acceleration on a

2
fluid particle, is replaced by (u /r), which represents the centri-

fugal acceleration. Assuming a thin boundary layer, r is taken as

Rand u as u , the surface velocity, since 6T is also measured

w
from the surface.

Therefore:
 23

2
HL L3 2 u c 1/3
k= 0.14 I-p-~
2 R (+)]

c
Rearranging, taking P;t; ~= 0.73 for air, and putting u WR,
k w
the equation becomes:

HD 2/3
0 lo R eD (27)
k

Reynolds Analogy

The Reynolds analogy is applied to a cylinder rotating in

air by Mayhew [11]. The details of the analogy can be fotmd in

his work. A brief summary and the results of the analogy are given

here.

The procedure adopted by Prandtl and Taylor to extend the

validity of the Reynolds analogy to fluids of Prandtl numbers be-

tween 0.5 and 2.0 is followed to obtain the following dimension-

less relation:

F
Nu 2 1 + Ub (Pr - 1)
u 00

For the case of the cylinder, the velocity at the edge of the larni-

nar sublayer relative to the cylinder surface is replaced for ~

and the free stream velocity relative to the cylinder surface is

replaced for u 00 The ratio ~/u 00 is found from the tmiversal

velocity profile. In this process, for air (Pr...::: 0.7), the exis-

tence of the buffer layer is neglected and the boundary layer is

assumed to c~nsist of a laminar sublayer and a turbulent layer.

 24

Finally, the following relation is found for the N u:o=

0.35 F ReD
(28)
1 - 2.47 IF
where F, the friction factor, for a rotating cylinder can be cal-

culated from:

1 ReD IF
- 0.6 + 4.07 lo~
IF 0 2
(29)

according to Theodorsen and Regier [13].

Figure 8 shows a plot of equations 27 and 28 and the heat

transfer data obtained by Anderson and Saunders [7]. NuD and F

values are calculated from equations 24, 25, 27, 28, and 29 and

are listed in Table 2. This table shows that, although for lower

ReD values the horizontal plate analogy seems to be in agreement

with the empirical results, at higher ReD values the Reynolds an-

alogy approximates quite well the empirical relations.

Temperature Profile

Teng and Grief [6] have obtained the velocity and temperature

distributions by solving the conservation equations for turbulent

flow given in Chapter I. For the solution, the apparent mixing

length, 1 , used is of the form:

c
1 = 1 (1 - BR.) (30)
c 1

where 1 is the mixing length without rotation. 1 takes care of

c
rotation, and therefore, of the centrifugal effects. The quantity
 3
10 /
I I /
F /
0. 35 Re /
Nu = D (28) /
- /
D 1 - 2.47 F /
/
/
2/3 /
- - - Nu = 0 .10 Re (27)
D D

10 2L_ ~Experimental Data [ 7)

Nun

Transition
10 N
\Jl

1
2 3 104 5 6
10 10 10 10

Ren

Figure 8. Nu vs. Re graph showing the results of the horizontal plate analogy and
the Reynolds analo~y [l]. D
 TABLE 2

Nu FOR VARIOUS Re USING EQUATIONS

n n

Horizontal
Reynolds
F Empirical Empirical Plate
Re Analogy
n Equation Analogy
29 Nun Nun
Nun Nun
Equation 24 Equation 25 Equation 27 Equation 28

300 0.0355 3.86 4.12 4.48 6.97

800 0.0227 7.86 8.18 I

8.61 10.l

2,000 0.01584 14. 93 15. 51~ I 15.9 16.1

I

N
0\
8,000 0.01002 39. 4 41.02 40 37.3

20,000 73.7
0.00772 74. 82 77 .90 I
69.0
I

80,000 186
0.00542 197.46 205.6 187 .o

200,000 342
0.00446 375.01 390.42 374 .0
I 861
800,000 0.00318 989.66 ! 1030. 33 I 1030.0
 27

B is an empirical constant, and it is set equal to 5 in previous

experimental studies. But the authors note that even better agree-

ment is obtained for B = 7 recommended by Bradshaw [6]. The mixing

length without rotation is chosen to be of the form:

1 =Ky [l - exp (-d)] (31)

Therefore, the following relation for the eddy diffusivity is ob-

tained:

(32)
2
- K (Y+)
2
[l - exp (-d)J
2 2
(1 - BR.) r + ~ (u+)
l dr+ r+

where the damping factor d is given by:

3/2 (33)
d = y+ (1-)
26 l
w

and K, the van Karman constant, is experimentally set equal to 0.42.

By using equation 32 in the conservation of momentum and energy

equations, the velocity and temperature profiles are obtained and

are shown graphically in Figures 9 and 10. For the temperature

profile, a semi-empirical solution is also obtained by using the

empirical result for the eddy diffusivity, directly obtained from

the velocity data, instead of equation 20. It is observed in

Figure 10 that this result is in good agreement with the experimen-

tal results for high Reynolds numbers, but fails at low Reynolds

numbers. In this figure, the experimental data of Kasagi and

Hirata [14] is used to compare the theoretical results.

 28

30 --------------....----.---r---r---..,----,..--.--...,...---r-----r--,r--~-1

65,000

------B
------B
0

25 5

u + - u+
w

20
u
w
+

.... .
50,000

_...
...-

25,000
15

experimental
data [14]

t 1 10 y+

Figure 9. Dimensionless velocity distribution [6].

 29

Re 72 ,000
B = 0 / D
B = 5
/
/

- - s from /
/
/

Ml ocity
ve .
25
T00+ data

T+ 48,ooo.


20
/
/
/
/

....
/

0115 .. 24 ,00~'

...........
//
/
/
/"

. .... ;J. /
/ ~

0110
"./

.. .........
/ /
//
Y"
/
//
/

/
/ 16,000
/
/,,


experimental
data [14]

t 0
1 10 10
2 10
3 10
4

Figure 10. Dimensionless temperature distribution [6].

 30

This brief description for the analysis of a rotating hori-

zontal cylinder shows that there are still matters to be clarified,

or to be improved both experimentally and analytically. A future

study should concentrate on these points. For this purpose, an

experimental set-up is designed which can provide accurate data

in a wide range of Reynolds numbers and Grashof numbers. The set-

up will also allow study of an inclined cylinder at slow rotation-

al speeds, an area of study missing in the literature.

 CHAPTER IV

THE EXPERIMENTAL SET-UP

To obtain accurate experimental data, to compare the data

with the results of previous experimental and analytical study and

to use this data for further analytical study which is described

in Chapter VI, the following experimental set-up is designed by the

author. The set-up is shown in Figure 11. The details of the

different design parts are described throughout this chapter. The

detailed drawings for the design of the cylinder are given in the

Appendix.

Cylinder

The cylinder consists of an aluminum core, a nickel-plated

copper tube and the insulation caps, as shown in Figures 16 and

17.

The aluminum core (Figure 18) has a length of 16.2 inches

and a diameter of 2.6 inches. An electrical heater rod, of 400 W

power, crosses the aluminum core in the axial direction. The 400 W

power is evenly distributed over the 15 inch long middle portion.

Therefore, we have a 66.7 W/inch power input per unit length.

The copper tube (Figure 19) has a length of 17.4 inches, an

inside diameter of 2.6 inches and an outside diameter of 3 inches.

31
 32

5
I ~I 6
0: LJ 7

nn 1
2 t3 9

03
4
10

1 - Cylinder

2 - Electrical heater

3 - Inf rared pyrometer

4 - Electric motor

5 - Signal scanner

6 - Cold junction

7 - Reading device

8 - Ammeter

9 - Voltmeter

10 - Variance

Figure 11. Experimental set-up showing the various equip-

ment.
 33

It is nickel-plated in order to reduce the radiation losses. The

copper tube is mounted on the aluminum core by shrink fit and screws.

The insulation caps, shown in Figures 20 and 21, are made of

fiberglass and prevent (actually minimize) heat transfer in the

axial direction. Therefore, the heat is conducted only in the ra-

dial direction (for the horizontal cylinder).

Supports

The support plates are shown in Figures 22 through 26. The

cylinder is carried by self-aligned roller bearings. The weight

of the construction (which does not exceed 200 N) and the life of

the bearings (considering that this is an experimental apparatus)

are not important for the selection of the bearings. For design

purposes, the inside diameter of the bearing is dominant. At the

support point, a shaft diameter of 0.8 inches is found to be safe

against any shear yielding and against excessive bending deforma-

tion which might cause eccentricity during rotation and might

affect the flow characteristics around the cylinder.

The support plates and the support legs are designed in such

a way that the cylinder can be inclined making 30, 45, 60 and

90 with the horizontal as seen in Figure 12.

Rotation

The cylinder is rotated by using an adjustable speed electric

motor with a belt and pulley arrangement. The electric motor is

chosen to have around 2500 rpm, which can be easily found in the
 34

a = oo
b = 30

c = 45

d = 60

e = 90

Figure 12. Inclination angles.

 35
market. Electric motors with higher outputs are only constructed

for special purposes and are found to be extremely expensive.

High rotational speeds (more than 10,000 rpm is required) are in-

tended to be reached by appropriate pulley design. The horsepower

requirement is low, so it is chosen to be 1 HP. The location of

the pulley on the shaft is seen in Figure 22. Figures 27 and 28

show the shaft flanges.

Temperature Measurement

In order to find the average heat transfer coefficient, H, the

surface temperature of the cylinder is measured and H is calculated

from:

H g - CHAD (34)
(T - T )A
w 00

where: 'kAn is the rate of heat radiated to the environment.

For the surface temperature measurements, two different tech-

niques are used:

(a) Inf rared pyrometer

(b) Thermocouples

As the infrared pyrometer, the omegascope with catalog number

OMEGASCOPE 20008 [15] is used. This instrument has an accuracy

below 1C; it can take four measurements per second, store them in

its memory and display the maximum, minimum and the average temper-

ature at the end of the measurement time. The advantage of it is

that it can be held far from the target; therefore, it does not

disturb the flow field around the cylinder. It also avoids the
 36
problem of noise, defined as the voltage drop due to contact resis-

tance, which is the most important problem with the thermocouple

measurements through slip-ring and brush arrangements.

For the measurement using thermocouples, four thermocouples

are placed along the cylinder just below the copper tube. The

thermocouples make 90 angles with adjacent ones around the peri-

phery. The distances in the axial direction are shown in Figure

13.

Chromel vs. Constantan thermocouple wires are selected because

they have the highest emf output of any standard metallic thermo-

couple, as shown in Figure 14. The highest possible emf is desired

because of the signal reduction through the slip-ring and brushes

due to noise. The higher the emf, the lower the percent error in-

troduced by noise.

The emf produced in the thermocouple wires is transmitted

through a slip-ring brush assembly, the details of which are shown

in Figures 22, 29, 30 and 31. The slip-rings and the brushes are

manufactured from a special alloy of gold and platinum.

Through the brushes, the signal is transmitted to the signal

scanner. For this purpose, the DATAPLEX 10 from the Omega Catalog

[15] will be used. This will enable an automatic scanning among

the four thermocouple outputs. The signal scanner is connected

to the readout device through a cold junction. The readout device

will display the temperatures directly in degrees Celcius or in de-

grees Fahrenheit. The thermocouple circuit is shown in Figure 15.

 11.2
92
El2
5.21
-::::::::!/ ----
~~-L-~~~~~~~ w
-......J

-----------------------------~-----------

t.
+
17.4 _ _ ___..,
COPPER TUBE ---lo.

Figure 13. Thermocouple locations along the cylinder.

 38

80

70

60
EMF
(mV)
so

40 G
30

20

10

0
1000 2000 3000 4000 5000

Temperature (F)

E: Chromel vs. Constantan

J: Iron vs. Constantan

T: Copper vs. Constantan

K: Chromel vs. Alumel

G: Tungsten vs. Tungsten 26% Radium

C: Tungsten 5% Radium vs. Tungsten 26% Radium

R: Platinum vs. Platinum 13% Rhodium

S: Platinum vs. Platinum 10% Rhodium

B: Platinum 6% Rhodium vs. Platinum 30% Rhodium

Figure 14. Graph of temperature vs. emf produced for various

thermocouple wires.
 39

6
4 5
l 1

~ l l t _____ _( ~ n
1
3
I I I I I
I
I
L------------------.J II II

II ~------------------------~ II

L-------------------------------------~

1 Thermocouple junctions

2 Thermocouples (Chrornel vs. Constantan)

3 Slip-rings

4 Signal scanner

5 Cold junction

6 Reading device

Figure 15. Thermocouple circuit.

 40

Heat Input

The cylinder is heated by a 400 W heater. The voltage input

is adjusted by a variac and the voltage (E) and intensity (I) are

read by a voltmeter and an ammeter. The . heat input to the cylinder

can be calculated by the simple relation:

q = IE
I

The power is supplied to the heater through copper slip-rings

and carbon brushes. This assembly is shown in Figures 23, 32, and

33. Because the applied voltage is high, the voltage drop due to

noise is neglected. If the results show that the voltage drop is

not negligible, the slip-ring assembly will be replaced by a spe-

cial (gold-platinum) alloy. This will not be tried first due to

economical considerations.
 CHAPTER V

EXPERIMENTAL PROCEDURE

By using the experimental set-up shown in Figure 11, the fol-

lowing procedure is applied to the experiment. The procedure is

given in algorithmic form. In the first part, a stationary cylin-

der is used.

1. Set the cylinder in the horizontal position and make

a careful alignment.

2. Set the variac to a certain value.

3. Read the intensity and the voltage; and calculate

the heat input from equation 34.

4. Read the steady-state* surface temperature when the

cylinder is stationary.

5. From the natural convection relation (equation 16),

for a stationary cylinder, calculate the heat trans-
fer coefficient, H. The?, substitute it into the
following formula to find the surface emissivity,

4 4
_g_
A
= H (T
w
- T00 ) + OE (T
w
- T00 ) (35)

Since H is calculated for the stationary cylinder,

Tw, Too, q/A are measured and E can be found. In
this equation, a is the Stephan-Boltzman constant
and is equal to 5.67 x 10-8 W/m2-~K4.

6. Increase the variac value, and repeat the steps from

2 through 6 until the heat input reaches 400 W, in
20 W increments.

*The steady-state is assumed to be reached when 3 consecutive

readings taken with 10-minute intervals are the same.

41
 42
Through this first part of the experiment, a set of e values

for different surface temperatures will be tabulated. Assuming

that the rotation does not affect the radiation characteristics,

this table will be used in the second part where the cylinder will

be rotating.

1. Set the variac to a certain value.

2. Read the intensity and the voltage, and calculate

the heat input from equation 34.

3. Set the electric motor to a low speed.

4. Measure the steady-state* surface temperature of the

cylinder.

5. Measure the ambient temperature, T .

6. Calculate the radiation heat loss by the following

equation:
4 4
= 0(T ) A (T
w w
- T
00
) (36)

E(Tw) can be found from the table constructed in

part 1.

7. Calculate the average heat transfer coefficient, H,

by:

q/A - q
ra d/A (37)
H =
(T - T00 )
w

8. Change the speed setting. (The speed is adjusted

from O to 12,000 rpm in 100 rpm intervals, by using
the speed adjustment dial of the motor and the pulleys
of different sizes.) Repeat the procedure from steps
3 through 7.

*The steady-state is assumed to be reached when 3 consecutive

readings taken with 10-minute intervals are the same.
 43
9. Change the variac setting. (The heat input is
adjusted from 0 to 400 W in 20 W increments.)
Repeat the procedure from steps 1 through 9.

In the third part of the experiment, the same procedure followed

in part 2 is repeated for an inclined cylinder, using 30, 45, 60

and 90 inclination angles, as shown in Figure 12.

During the surface measurement process, the readings from the

thermocouples and the pyrometer are checked against each other;

the reading from the pyrometer is considered to be more reliable.

If the resulting thermocouple readings are much different than the

pyrometer readings, the conclusion is that the voltage drop through

the slip-ring brush arrangement is not negligible. An amplifier

is intended to be mounted on the shaft, in this case. This ampli-

fication is expected to reduce the percent error due to noise.

Table 3 shows the surface temperatures expected to be measured

for different heat input and rotational speeds. These values are

on 1 y approximate, t h e experimen
t a 1 va 1 ues are expec t e d in a +_ 3 C

range.

Table 4 shows the expected Nusselt number values. The criti-

cal Reynolds number value is calculated by the equation:

For the NuD' the empirical equations of Etemad [10] are used for

rotational speeds above the critical value:

2 0.35
N~ = 0.097 (0.5 R~ + GrD)

. )0.7
N~ = 0.076 (R~
 TABLE 3

EXPECTED SURFACE TEMPERATURES (WITHIN 5C ACCURACY)

. Surf ace Temperature (C)

q
(watts) RPM RPM RPM RPM RPM
100 500 1000 5000 10,000

20 40 30 30 25 25

40 60 40 35 25 25

80 90 60 55 30 25

160 150 150 90 40 35

~
200 170 170 110 50 40 ~

240 200 200 125 50 50

I
I

280 220 220 140 60 50

320 240 240 160 65 50

360 260 260 180 80 55

I

400 290 290 195 75 60

 TABLE 4

SOME EXPECTED NuD AND ReCRIT VALUES

. NuD
q ReCRIT
Gr
(watts) RPM RPM RPM RPM RPM
100 500 1000 5000 10,000

20 7 4250 27 53 81 245 400

1.5 x 10

40 2.5 x 10 7 5500 30 56 82 241 401

7 240 401
80 3.6 x 10 6500 34 57 75

160 7 7300 36 70 238 391

4.5 x 10 36 ,.i::-..
Vl

200 4.6 x 10 7 7400 36 36 65 229 386

240 4.65 x 10 7 7400 36 I 36 64 229 387

I

280 7 35 35 64 223 379

4.3 x 10 7100
7 I
320 3. 7 x 10 6600 34 34 63 228 379

360 4.4 x 107 7300 35 35 60 220 371

400 7 35 35 58 220 371

4.2 x 10 7100
 46

Below the critical Reynolds number, the relation for a stationary

cylinder is used:

The values in Tables 3 and 4 are only for a horizontal cylin-

der where radiation losses are neglected and steady-state has been

reached.
 CHAPTER VI

CONCLUSION

Previous work has been analyzed for the horizontal cylinder.

It is seen that most of the results of both experimental and theo-

retical approaches are far from being satisfactory. The theory in

the three different regions mentioned in Chapter III must be re-

viewed; especially more emphasis should be given to analytical solu-

tions in the transition and the fully developed regions.

In the transition region, there is no analytical solution

available in the literature. If the same order of magnitude analy-

sis is applied to the transition region as applied to high rota-

tional speeds, the following conservation equations will be ob-

tained:

a (38)
ar

vr
(39)

+ pg cos <P

vr aar (V'T')
r
(40)

where:

47
 48

V
r
V
r,w
+ V
r,c
+ V'
r,w
+ V'
r,c

T = T + T'

P = P + P + P' + P'
c w c w

Assuming that in the transition region the turbulence starts

in the rotational components, the terms Vl and V' can be dropped.

'+''c r,c
This assumption will make the use of the apparent mixing length,

1 , used by Teng and Grief [18] possible. For V~ , V and P ,

c '+'' w r, w w
the solution for laminar flow can be used.

wR2
r

vr, w = 0

Substituting these expressions in equations 38, 39 and 40, the fol-

lowing set of equations is obtained:

0 (41)
 49

av
WR 2 R2 avcp
vr,c [ cp'
ar
c
-2] + (Vcp + ~) ,c =
r ,c r2 acp

(42)

a~r (Vcp' ,w V'r,w )} + pg cos cp

a

(43)

- v 2
+ (~ + a i a - a
V dT
r,c ar r
wR )
r
2
dT
acp = a ar [r ar (rT)] - ar (v;,w T')

Now these equations, equations 41-43, must be solved numerically to

obtain the temperature and velocity profiles around the cylinder.

For the solution, the same apparent mixing length relation proposed

by Teng and Grief [6] will be used:

1 = 1(1 - BRi)
c

B, for transition region, must be determined experimentally. In

fact, the experimental results will determine the validity of the

assumption that the turbulence starts in rotational components. If

the experiment gives a value close to 0 for B, this will show that

the rotation has a minor effect on the turbulence and the turbulence

is mainly due to natural convection.

For high rotational speeds, in the fully developed turbulent

region, the analytical solution of Teng and Grief [6] is based upon

the proportionality constant, B, used in the correlation of the

 50
apparent mixing length. The experimental re$ults obtained by the

described experimental set-up will be checked against analytical

solutions using different values of B. Although B = 5 represents

well the experimental velocity distribution, there is a deviation

for the temperature distribution, especially for low Reynolds num-

ber values, in the graphs obtained by Teng and Grief [6]. B = 7

is also suggested by Bradshaw [6].

In order to find a good mathematical model which will be appli-

cable to a wide range of Reynolds numbers, two approaches will be

tried. First, the contribution of viscous dissipation will be

analyzed. This can be an important factor, especially for high

Reynolds numbers, and this possibility is not considered in the

study of Teng and Grief [6]. Then, instead of taking Bas a con-

stant, the possibility of B being a function of Reynolds number will

be considered, and the correlation B = f (Re) will be found.

These constitute the proposed study for the horizontal cylin-

der. The results for high rotational speeds will also be applica-

ble to the inclined cylinder case. But for low rotational speeds,

where natural convection effects still exist, the flow becomes

three-dimensional and the theoretical analysis is much more complex

than the case of the horizontal cylinder. The discussion for the

inclined cylinder will be left until the end of the experimental

study, considering the fact that so far there is no information

available in the literature.

 APPENDIX

The appendix shows the details of the design parts for the

experimental set-up. It contains Figures 16 through 33.

51
 copper tube
aluminum
core

V1
N

electrical
heater
end insulation

LEFT PART OF THE CY~INDER

thermocouole holes not included
Figure 16. Section of the left half of the cylinder.
 V1
w

RlGHT PART OF THE CYLJNDER

(thermocouple holes not included)
Figure 17. Section of the right half of the cylinder.
 (4+4)
12-24 UNF
4 TCBED
<!> 2/10in .

'1I I
l'l
~;i

-----------------
---ct>
_J~~
--------------------~~-
ft-i
I I.
\J1
+:'-
I .. "

..__--16..2in /1 ~

heater hole ALUMlNUM COR.E

Figure 18. Aluminum core.
55

Q)
.
,..c
::s
.

J..j
Q)
p..
p..
0
u

O'\
.
........
4 holes ct> 2/10 in.
(4) 12-24 UNF
-;:..;;:;;;.~-:.. -"

--,
=--=--='--...:r.a - -.J~--"'f..:a:
I I :(1~f I
.
l f~ 4'-t.
:..::...-=r---:.- ,.--tt-.,.+.+-1.~~--
- -~ ,_= --
1
I r' I Lrr
I I
I 111 ,, '
ct> 6/10
1n

1.5 '6~0

LEFT END JNSULATJON

Figure 20. Left end insulation cap.

~------------------------------------~----~------------------------------~------------------------------------------------------
...
heater hole
5/16-18 UNF (4)12-24 UNF
( 4)

17 in

I
I
I

L__J I ~
-------

section A-A
</) 3.0 ill
RIGHT ENO JNSULATJON
Figure 21. Right end insulation cap.
bearing
support----+t

pul ley_...u

bear;ng __:___"
IJ1
co

bearing slip rings

casing
insulation rings I f lafige
insulation envelop

LEFT SHAFT
( wi~h accessories)
Figure 22. Section of the left shaft (with accessories).
 59

RIGHT SHAFT

Figure 23. Section of the right shaft (with the accessories).

 II It
11 11
,, ,.
11 11 :1 1:
11 1 ,1 1,
ti 1 ti 11
1
" h 1 I~
'I
.....,,,,.-::
4 holes I
't::: :1 .
I

~ 0.3

0\
0
7in A

</) 0.161
.....
,,..
I I 0.251n ,--,
:1 11 ,1 1,
' 1, ,1
l~r 1,
~I~
~ c:r 11 t II
I 11 11
" II 1 1: :1 1,
" '1 ti 11
1l1 I. I '

Figure 24. Front view of the bearing support.

 ,.

~~~~~-7~~~~~

0.4
5/16-1BU~F
. { 6)
O"I
f-1

BEAR.ING SUPPORT (2)

(sec ti on A-A)

Figure 25. Section of the bearing support.

outer diam~ter ~ 1in
of the b~aring 4 holes
~ 3/10 in

--+-
N
'

J l.~3/10in
<t>1.8 in <t> 3.4 in 4/10in
BEARING CASING (2)
Figure 26. Bearing casing.
 4holes f tange </J 3in
r/J 4/10 in

</J 6/10 in

w
'

4 holes shaft </J 1in.

</J 3/10 in
LEFT SHAFT FLANGE.
Figure 27. Left shaft flange.
 2 holes 3/20in
ct>1.7in ~

0\
~

section A-A
4 holes 4/10 in
shat t 1. 0in
R1GHT SHAFT FLANGE
Figure 28. Right shaft flange.
 1.2 in _a_
2/10in THERMOCOUPLE
SLIP RINGS (8)

1/1

( 4} ~1/10in _b_ O'\

V't

A
INSULATION RINGS
(12)

--.i '4--1/10 in 1/1

<t.>1.21 n -1. S 1n
'
Figure 29a. Thermocouple slip-rings.
b. Insulation rings.
 66

ct> 1.2 in i> 8/10 in

A

(4)
1/10 in A.

JNSULATlON ENVELOPE

Figure 30. The insulation envelope for the slip-rings.

 BRUSH ARM
A .. 3.8

A -12-24 UNF 12-24UNP (2)

3.9 0\
A -....J

lJ ;o.s
0.61 A --- ~ ~ -- c/J 0 2
0.2 0.2 (2)
BRUSH HOLDER
Figure 31. Brush arm and brush holder.
 <t>1.2 A <t>0.1 (4)
:-. :.: ::.:<-: ';' ."- ::-.

1:-------
.... ------ _a_
. :. :'.:.... .. :: :.: :... :::

<t>0.8 - 1. Q. --
A
<1>1. 6

0\
00

_b_
~0.1 (4)
~o.2

Figure 32a. Insulation envelope.

b. Copper slip-rings.
 12-24 UNF
(2)

.~ Jl0.4
1 ro
BRUSH ARM
A .. 1.6 .
0.5 t :c \.0
'
0.6LA 2 holes
ct>0.25
BRUSH HOLDER

Figure 33. Brush arm and brush holder.

 REFERENCES

1. F. Krieth, "Convection Heat Transfer in Rotating Systems,"

Advances in Heat Transfer 5 (New York: Academic Press, 1968): 129-
151.

2. S. Ostrach and W.R. Braun, "Natural Convection Inside Flat

Plate Rotating Container," NACA Technical Note 4323 (1958).

3. F.E. Fendell, "Heat Transfer to Rotating Cryogenic Fuel

Tanks in Orb it," AIAA Paper No. 66 (1960) .

4. D.W. Kern and H.J. Karakus, "Mechanically Aided Heat

Transfer," AIChE Preprint No. 17 (New York: Inst. Chem. Engrs,
1958).

5. L.L. Goettler and J.A. Fillo, "Free Convection From a

Slowly Rotating Cylinder," N.C.5.8 Tokyo: Heat Transfer 1974,
5th I.H.T.C. (1974).

6. J. T. Teng and R. Grief, "Turbulent Transport to a Rotating

Cylinder," Letters in Heat and Mass Tr. 3 (1976): 365-374.

7. J. T. Anderson and 0. A. Satmders, "Convection from Isolated

Heated Horizontal Cylinder Rotating About Its Axis," Proc. Royal
Society A. 217 (1953): 555.

8. P. D. Richardson, "Transition on a Heated Horizontal Rota-

ting Cylinder," J. of Heat Tr.. (Aug. 1961): 386.

9. D. Drapkin and A. Carmi, "Natural Convection Heat Trans fer

from a Horizontal Cylinder Rotating in Air," Trans. ASME 79 (1957):
741.

10. G.A. E:temad, "Free Convection Heat Transfer From a Rotating

Horizontal Cylinder to Ambient Air with Interferometric Study of
Flow," Trans. ASME 77 (1955): 386.

11. Y. R. Mayhews, "The Reynolds Analogy Applied to a Cylinder

Rotating in Air," J. of Royal Aeron. Society 58 (1954): 205.

12. W.R. McAdams, Heat Transmission, . 2nd ed. (New York:

McGraw-Hill Book Company, Inc., 1942).

70
 71

13. T. Theodorsen and A. Regier, N.A.C.A. Report No. 793 (1944).

14. N. Kasagi and M. Hirata, "Transport Phenomena in Near Wall

Region of Turbulent Boundary Layer Around a Rotating Cylinder,"
ASME 75-WA/HT-58 (1975).

15. Omega Engineering, Inc. Temperature Measurement Handbook

(1982).
 ADDITIONAL BIBLIOGRAPHY

Dorfman, L.A. Hydrodynamic Resistance and the Heat Loss of Rotating

Solids. Edinburg, London: Oliver Boyd, 1963.

Golstein, S. Modern Developments in Fluid Dynamics, Vol. 2. New

York: Dover Publishing, Inc., 1965.

Holman, J.P. Heat Transfer, 4th ed. New York: McGraw-Hill Book
Company, Inc., 1976.

Jacob, M. Heat Transfer, Vol. 1. New York: John Wiley and Sons,
Inc., 1949.

Kays, W. M., and Bjorklund, I. S. "Heat Transfer from a Rotating

Cylinder With and Without Cross-Flow," Trans. ASME 80 (1958):
70.

Kays, W.M., and Crawford, M.E. Convective Heat and Mass Transfer,
2nd ed. McGraw-Hill Book Company, Inc., 1980.

Schlinchting, H. Boundary Layer Theory, 7th ed. McGraw-Hill Book

Company, Inc., 1979.

72