AE-338

UDO 532.542.4.082
AE-338

Calibration Experiments with a DISA Hot-Wire

Anemometer

EL Kjellstrom and S. Hedberg

AKTIEBOLAGET ATOMENERGI
STOCKHOLM, SWEDEN 1968
 AE-338

CALIBRATION EXPERIMENTS WITH A DISA

HOT-WIRE ANEMOMETER

Bjorn Kjellstrom and Stellan Hedberg

ABSTRACT

From Collis' law and the direction sensitivity relation pro­

posed by Hinze it is possible to derive the following relation be­
tween the voltage over the wire, the velocity and the angle of in­
cidence of the flow;

= b(pu)c(sin2 Y + k2 cos2 ¥)c^2

The values of the exponent c and the direction sensitivity coeffici­

ent were determined experimentally in the range 20 < pu < 1 80
kg/sm^. It was found that, if V is the voltage measured with no
2° ...
net flow past the wire, c and k are decreasing with increasing
values of pu.
In order to check these calibration experiments, shear stress
and turbulence measurements were made in a circular channel. For
this geometry the shear stress distribution can be estimated theoretic­
ally and several earlier experiments can be used for comparisons.
These experiments were made at Reynolds numbers 3 • 10^- 10^,

Mach numbers 0.1 -0.3 and a channel length of 6l diameters. Excel­

lent agreement with the theoretical shear stress distribution (corrected
for compressibility effects) and earlier data for the axial and radial
turbulence components was obtained when the results of the calibration
experiments were used for the evaluation of these measurements.
2
Evaluation with a constant value of c or with k equal to zero (as
often recommended) gave less good agreement.

Printed and distributed in November 1968.

 LIST OF CONTENTS

Page

1 . INTRODUCTION 3

2. THEORETICAL CONSIDERATIONS REGARDING

TURBULENT FLOW IN A CIRCULAR CHANNEL 3
2.1 General Assumptions 3
2.2 Basic Equations of The Flow 4
2.3 The Shear Stress Distribution for Pseudo-developed
Flow 5

3. EVALUATION OF THE HOT-WIRE DATA

THEORETICAL ASPECTS 7
3. 1 Basic Equation 7
3.2 Calculation of The Turbulence Components and
The Reynolds Stress 8

4. EXPERIMENTAL EQUIPMENT 10
4. 1 The Experimental Channel 10
4.2 The Atmospheric Air Rig 1 1
4.3 The Traverse Device and Probe Support 1 1
4. 4 Stagnation Pressure Probe 1 2
4.5 The Hot-wire Anemometer 12
4.6 Electronic Measuring and Data Recording Equip­
ment 1 2

5. EXPERIMENTAL PROCEDURE 12

6. EVALUATION OF THE EXPERIMENTAL DATA 1 3

6. 1 General 1 3
6.2 Basic Parameters of The Flow 1 3
6.3 The Velocity Distribution 14
6.4 Evaluation of The Anemometer Measurements 1 5

7. DETERMINATION OF THE FRICTION FACTOR AND

VELOCITY DISTRIBUTION IN THE CHANNEL 16
7. 1 The Friction Factor 1 5
7.2 The Velocity Distribution 15
 Page

8. DETERMINATION OF THE EMPIRICAL CONSTANTS

IN THE ANEMOMETER EQUATION 19
8. 1 The Exponent in Collisz Law 19
8.1.1 Own measurements 1 9
8.1.2 Comparison with Earlier Works 20
8.1.3 Choice of correlation method 20
8.2 The Direction Sensitivity Coefficient 21
8.2.1 Own measurements 21
8.2.2 Comparison with earlier works 21

9. MEASUREMENTS OF SHEAR STRESS DISTRIBUTIONS

AND TURBULENCE INTENSITIES 24
9• 1 The Shear Stress Distribution 24
9.1.1 General 24
9.1.2 Comparison of Evaluation Methods 24
9.1.3 Influence of the Compressibility Correction 26
9.2 The Axial and Radial TurbulenceIntensities 26
9.3 Observed Standard Deviations 26

10. COMPARISON WITH EARLIER TURBULENCE AND

SHEAR STRESS MEASUREMENTS IN CHANNELS 27
10.1 Earlier Investigations 27
10.2 Shear Stress Distributions 27
10.3 Turbulence Intensities 28

11. CONCLUSIONS 28

NOMENCLATURE 31
Subscripts 33
Special Signs 33

REFERENCES 34

ACKNOWLEDGEMENTS 38

APPENDIX 1 39
APPENDIX 2 54
APPENDIX 3 68
APPENDIX 4 74
APPENDIX 5 80
APPENDIX 6 81

FIGURES
 3

1. INTRODUCTION

The high heat fluxes and narrow flow channels of fuel elements
for fast reactors, has increased the demand for information on the
local turbulent transport processes for flow in such geometries. It
has therefore been decided to make measurements of shear stress
and turbulence distributions in a large scale rod bundle.
Such measurements can be made with a hot-wire anemometer.
Earlier experience from hot-wire measurements [39] indicated, however,
that the accuracy is not so good. It was believed that this may have been
caused by shortcomings in the technique or evaluation method. It was
therefore decided that prior to measurements in a rod bundle measure­
ments in a circular channel should be made. This has the advantages
that the shear stress distribution is known theoretically and that several
earlier investigations can be used for comparisons.
For the measurements in the circular channel different experimental
techniques and evaluation methods should be used and the results compared.
The first results of these measurements have been communicated
in an earlier report [40 ]. It was found that repetition of the measurements
with a modified inlet to the test channel and direct determination of the
exponent in Collis ' law and the direction sensitivity coefficient were desir­
able .
This report contains the results of the later measurements, but
summarises also the results from the earlier experiments.

2. THEORETICAL CONSIDERATIONS REGARDING

TURBULENT FLOW IN A CIRCULAR CHANNEL

2. 1 General Assumptions

In the following, the fluid will be assumed continuous, isotropic and

with constant viscosity. Compressibility will be accounted for, but it will
be assumed that a pseudo-developed flow is obtained where the velocities
and the density, made dimensionless by division by the cross-section mean
values, do not vary in the axial direction. Only steady flow with perfect cir­
cular symmetry in a channel with constant cross -section area will be
considered.
 - 4 -

A detailed discussion of the shear stress distribution can be found

in appendix 1 . Only a summary will be given here.

2.2 Basic Equations of The Flow

In turbulent flow, the instantaneous value of any of the irregularly

varying quantities such as the velocity, the density, the pressure or the
temperature can be regarded as the sum of a time mean value and a
fluctuation. For instance, the instantaneous value of the axial velocity
can be written as:

u = u + u 1. (t)

where the mean value is defined as

t

Navier-Stokes equations are valid also for turbulent flow if the

motions of the fluid are studied in sufficient detail.
If the instantaneous values for the different quantities which occur
in these relations for a compressible fluid with constant viscosity,
are written according to eq. (l) and inserted, the time mean values are
taken and all the simplifications are made which are possible for steady
axi-symmetric flow at a low turbulence Mach number (see appendix l)
one arrives at the following set of equations:

in the axial direction

(3)

in the radial direction

(4)
r

and in the peripheral direction

 - 5 -

-^-r r'cp 1 - 2 r 1 cp* = 0 (5)

where

A 3u 1 drr 1 3rcp
0 = v— +---- r— +---- T-^- (6)
ox r or r ocp

The continuity equation for the mean values, which can be obtained
in a similar way, starting from the continuity equation for the instantaneous
values, reads

It is well known that these equations can not be solved even for
the simpler case of incompressible flow where 9=0. For such flow,
however, many important conclusions can be drawn from these equa­
tions without further approximations or assumptions. This does not
seem to be possible for compressible flow.

2. 3 The Shear Stress Distribution for Pseudo-developed Flow

Attempts to treat the flow of a compressible fluid in a channel

have been made by Nazarchuk [1 8 ], Deissler [7 ] and Ward Smith [33 ].
Ward Smith [33 ] made an integral analysis and did not consider the
radial shear stress distribution. Nazarchuk [18] assumed Prandtl's
mixing length hypothesis to be valid, Deissler [7 ] employed
v. Kh.rmh.ns relation for the eddy diffusivity and succeeded by this
means to calculate the velocity distribution for compressible flow.
Both approaches are somewhat questionable, if the object is
to study the effect of the compressibility on the shear stress
distribution. An attempt to work with less far-reaching assumptions
was therefore made. Thus, it was assumed only that the flow reaches
a pseudo-developed state after a sufficient length of the channel. This
flow is characterised by the constancy of the relative velocities and
velocity correlations, for example u/u^; r/u^, u* r' /uj~ and the rela­

tive density p/p^, where u^ and p^, the cross-section mean values of
the axial velocity and the density, are defined by
 - 6 -

R
2 J u r dr
u (8)
b R
0

R
"pu r dr
0
Pb = R----------- (9)

u r dr
0

That u/u^ reaches a constant value after x/d ~ 40 has been

established by Deissler [7] up to Mach number 1.0. The assumptions
above are quite reasonable then on the basis of local similarity of the flow.
d2ub
For cases where ----- can be neglected (this is normally the case,
dx
see appendix 1 , section 6), eq. (3) may be reduced to

du r r dpf
T1 dr i 4!
pu'r : = 2 dT (10)

where the "friction11 pressure drop is defined as

fff _ dn ^b
(11)
dx ~ dx PbUb dx

and the correction term £» which is zero for incompressible flow is

1 p (u2 + u 1 -
P (12)
rUb dx ,
0

Since the shear stress is given by

xr = n If u ryr (13)
the shear stress distribution is given by

r^f
(14)
xr - C - 2 dx
T

The details of the derivation and a more thorough discussion can

be found in appendix 1.

3. EVALUATION OF THE HOT-WIRE DATA

THEORETICAL ASPECTS

3. 1 Basic Equation

Collis " law £ 53 for heat transfer from a wire reads:

N"X%y-)°' 17 = a + b R<i <15>

This equation was established for flow perpendicular to the wire.

For the directional sensitivity Hinze [ 10] proposed the following
relation for the effective velocity to be introduced in the Reynolds number
of eq. (1 5)

u2(Y) = u2(^) • (sin2 f + k2 cos2 f) (16)

with a value of k between 0.1 and 0. 3 depending on the magnitude of the

velocity.
For laminar flow past an infinitely long cylinder oblique to a uniform
velocity field the two velocity components in a plane normal to the axis
of the cylinder are independent of the axial component, as has been established
by among others Prandtl £ 24]. This leads to the sine law of directional
sensitivity for infinitely long wires, i. e. k = 0 in eq. (16). This is often
assumed to apply also for wires of finite length, cf. Patel [23] , Newman
and Leary [19] and Sandborn [29, 30].
 - 8 -

However, as has recently been established by Champagne [2]

for the short wires normally used in hot wire anemometry, (4/d<600),
the velocity component parallel to the wire is also important for the
heat transfer. It was found that the direction sensitivity coefficient k2

is dependent on the length to diameter ratio of the wire. This is dis­

cussed further in section 8. It is sufficient for the moment to notice
that k in eq. (1 6) is not zero.
Assuming the resistance of the wire to be proportional to its
temperature, and neglecting the temperature dependence of the vis­
cosity and conductivity of the fluid allows eq:s (15) and (1 6) to be com­
bined to

.<+* a_ 0.17 2
t)c/2J Ot-KJ (17)
J = fa + b (pu)C (sim2 Y + k2 cos2

If the ambient temperature is constant and the resistance ratio R/R^ is

kept constant eq. (17) can be simplified to

where of course the constants a and b assume different values from

eq. (17).

3, 2 Calculation of The Turbulence Components and The Reynolds Stress

In the following the wire is assumed to be lying in the x - r plane.

If then eq. (18) is differentiated, small velocity fluctuations, u' , r * and
rep • in comparison to the mean velocity u are assumed, the equation is
squared and the time mean value is taken, one obtains the following rela­
tion between the voltage fluctuations and the velocity fluctuations (cf.
appendix 4)

2—»- 2
c P ♦ (. 1 - k 1 - k r] (19)
e v* + 2
tan Y + k cot Y tan Y + k cot Y

whe re
2V
P = (20)
o
 - 9 -

Vq being the voltage at zero velocity past the wire.

This equation can be used for calculation of the turbulence compo­
nents and the Reynolds stress.
Measurements with a right angle probe Y = 90° gives the axial
turbulence

whereas measurements with a slanting wire probe at two slanting angles

Y and Y. allows calculation of the radial turbulence and the Reynolds
a b
stress from the simultaneous equations

1 - k . TT2 =
2 u* rT +
tan V + k cot Y
a a

tan Y + k cot Y
a a B»2a^-^2] (22)
1 - k2 a

1 - k2 ~Z
2 xTF + ------2------------ r
tan Y^ + k cot Y^

tan Y + k cot Y, ----- 2 00 0 -.

b ~2 ~ [ ~Z ®b ^t)2 " u* ^ ] (23)
1 - k

The two measurements can be made with the same probe, the second
with the probe rotated 180° around its own axis. Ideally Y^ = 180- Y » so
that eq:s (22, 23) can be combined to

2 2
= (■tan Y +il'V
k cot Y'X 2 f u
; (PMV’ W + s2(-)v’2 ) - iR] (24)
(-)

-2
pu 2 1
- puTF = V (+)-l (25)
4c

where the subscripts (+) and (-) refer to the angles Y and 180 - Y respectively.
 -10-

If the simple sine -law for the direction sensitivity had been employed
i. e. if k = 0 the above equations reduce to

F-2 = tan2 (e2(+) V2(+) + e2_, W2{_}) - V-2] (26)

- ^ * 4^ <6<-> ^<-> - (27)

which were used by among others Patel [23 ] in his evaluation.

In the derivation of eq. (19) the exponent c in Collis ' law and
the direction sensitivity coefficient k were assumed constant. It will
later be shown that both vary with the velocity. Fortunately, the va­
riations are weak enough to allow the variations caused by the veloc­
ity fluctuations to be neglected.

4. EXPERIMENTAL EQUIPMENT

4. 1 The Experimental Channel

The experimental channel consisted of a smooth steel pipe diam.

128. 8 mm with an overall length of 7905 mm, cf. fig:s 1 and 2.
For measurements of the pressure distribution it was provided with
carefully made static pressure holes (diam. 0.7 mm), two at 1870 mm from
the inlet, four at 2870, four at 3870, two at 4870, two at 5870, four at
6870 and four at 7 87 0 mm from the inlet. During introductory tests the
pressures measured at the holes at the same level were compared and
the surface around the holes carefully honed until the differences at the
maximum flow rate were less than 0.2 mm H^O. During the later mea­
surements the pressure taps at the same level were connected.
The test channel was supplied with air from the atmospheric air rig
FRIGGA. The inlet was formed by a wooden nozzle giving a well rounded
contraction from 440 to 128. 8 mm. Immediately before the entrance to the
tube a sharp-edged orifice plate, diam. 1 1 9 mm was positioned, with the
purpose of giving a well defined separation point and a turbulent boundary
layer starting at the entrance of the test channel.
 In the first measurements [40 ] the air entered the test channel
directly from the tank. A flow rectifier, fig. 1 , was later introduced.
This consisted of a cylindrical tube diam. 440 mm with a honeycomb.
In the lower end of the tube, 200 holes diam. 1 5 mm allowed the air to
flow into the tube from the tank. The honeycomb (with a cell area of
2
7 mm ) was positioned 450 mm upstream of the nozzle, 1 00 mm from
the upper row of holes. The length of the honeycomb was 75 mm. The
introduction of this rectifier had an appreciable effect on the velocity
and turbulence distributions in the centre of the channel as will be
shown later.

4, 2 The Atmospheric Air Rig

The atmospheric air rig FRIGGA has been described in detail by

Lars son [14] and it may be sufficient to describe only its most important
features here.
Air is supplied from a blower to a tank, diam. 940 mm, provided
with a cooler and three large outlets which can be used for connection of
test channels. The second outlet was used in this investigation. The mass-
flow was measured by a calibrated Foster flow meter in the pipe feeding
the tank.
The pressure in the tank and the temperature at the outlet nozzles
could be measured.
The velocity distribution over the outlet nozzles was measured by
Larsson [14] before introduction of the rectifier. The velocity was found
to be practically uniform. The turbulence level in the outlets was found to
be 1.7 %.

4. 3 The Traverse Device and Probe Support

The traverse device employed was identical to that used during the
earlier measurements at the laboratory, see Kjellstrom and Hedberg [39].
The probe support was identical to the modified design described in
the same report.
A better device for zero adjustment of the screw scales was used,
however, consisting of a measuring microscope, type VM-3 from Svenska
Ackumulator AB Jungner. This was also used for measurement of the
wire inclination.
 -12-

An overall view of the microscope, probe support and outlet of

the test channel is given in fig. 2. The distance between the probe support
and the outlet of the pipe was 105 mm. The traverses were made 3 5 mm
upstream the outlet.

4, 4 Stagnation Pressure Probe

The stagnation pressure probe consisted of a thin-walled stainless

steel tube (diam. 1. 0 mm) which was mounted in the probe support with
its axis parallel to the flow direction.

. 4. 5 The Hot-wire Anemometer

A constant temperature anemometer DISA 55A 01 was employed.

The experiments were made with a single anemometer and miniature
probes (tip diam. 2 mm) 55A2 5 with the wire perpendicular to the flow
(R-probes) and 55A29 with the wire at about 45° angle to the flow (V-probes).
The technical data for the anemometer and the probes can be found
in the data sheets supplied by DISA [3 5] . The length to diameter ratio of
the hot-wires was established by measurements in a microscope. For all
the probes used the value was found to be between 220 and 224.

4. 6 Electronic Measuring and Data Recording Equipment

All measurements except those with the anemometer were recorded

on paper tape by a data logger using a four-figure digital voltmeter.
The RMS-value of the voltage fluctuation was measured by the instru­
ment on the anemometer. The mean voltage during the first experiments
(without rectifier at the inlet) was measured with the instrument on the
anemometer. Later a four-figure digital voltmeter together with an
adjustable constant voltage source "Dial-a-volt" were employed. This
allowed five-figure measurements of the mean voltage.

5. EXPERIMENTAL PROCEDURE

After calibration of the pressure tranducers against Betz micromano­

meters, the blower was started and steady state conditions awaited.
Pressures, temperatures and pressure differences in the test section
were then registered twice by the data logger, after which a stagnation
pressure traverse was made across the channel. Measurements were made
at each 5 mm along a diameter, starting at the radius 62. 5 mm.
 -13-

After preparation of the anemometer according to the manual [35],

and measurement of the angle between the hot-wire and the flow direction
by the microscope, hot-wire traverses were made, similar to the
stagnation pressure traverse. Measurements were first made with a
normal wire probe, then with a slanting wire probe. With the latter two
traverses were made, the second after rotation of the probe 180° around
its own axis.
The ratio of hot to cold resistance for the wires was maintained
at a value of 1 . 8 in all experiments.
After the hot-wire traverses which lasted about 2 hours including
preparation of the anemometer and measurements of the wire angles, a
repetition of the stagnation pressure traverse was made, followed by two
registrations of all pressures, temperatures etc.

6. EVALUATION OF THE EXPERIMENTAL DATA

6.1 General

The evaluation was performed on an IBM 360/30 computer. To

check the computer programme one experiment was evaluated by manual
calculations.
Equations for the properties of air given by Lindh [ 16] and obtained
by fitting to the data in NBS tables [37] were used. Temperatures were
calculated from the thermocouple emfs by the relations of Kjellstrom,
Lars son and Lindh [38].
The air properties were evaluated at static pressure and temperature.

6. 2 Basic Parameters of The Flow

The Reynolds number for the channel was calculated from:

2m
(28)
TfflR

The friction pressure drop was calculated from the measured pressure
difference between the two last measuring stations (I and II, at 687 0 and
7 87 0 mm from the inlet) by deduction of the acceleration pressure drop:

4pf = 4pmsd - W(ubn- "bjl (29)

 -14-

The friction factor was then calculated as

i3. *,2r5^ (30)

L m

where the density p was evaluated as the mean value over the channel
half way between stations I and II.

6. 3 The Velocity Distribution

The velocities were calculated from the stagnation and static pressure
measurements by

where the density "p- was evaluated at the static pressure and static temper­
ature

The shear stress due to molecular momentum transfer, the first

term in eq. (10) was calculated from

rl (33)

where the velocity gradient at the i:th point of measurement was

calculated as

8(Ui+l " ^i-l) ' Ui+2 + ui-2

H)i - ~TW\
(34)

A modified definition for the eddy diffusivity

r dPf - du
, _-p7F
(3 5)
era du
3F 3F
 -15-

and a modified wall shear stress

i^i
T (36)
w 2 dx

were used for convenience. This modified wall shear stress was used only
for calculation of the shear velocity

u* (37)

The difference from the true wall shear stress is the term

1 % +1?2)
ew Rub dx
(38)

which is of the order of 0. 1 % of t or less.

w

6. 4 Evaluation of The Anemometer Measurements

The theoretical background for the evaluation is given in section 3.

The data were evaluated with some different values of the constants
2
c and k in the anemometer equation as explained in the table below.

Table 6.4.1 Evaluations of Hot-wire Data

Exponent c in Direction sensitivity

Method coefficient k2
Collis ' law

1 constant, equal zero

to average value
for the probe dur­
ing the particular
experiment
2 Eq. (42) zero
3 As in method 1 Eq. (43)
4 Eq. (42) Eq. (43)
 -16-

7. DETERMINATION OF THE FRICTION FACTOR AND

VELOCITY DISTRIBUTION IN THE CHANNEL

7.1 The Friction Factor

The axial variation of the friction factor was determined at three

different Reynolds numbers in the channel without inlet rectifier. The
results are shown in fig. 4. The variations over the last 15 diameters
are small which would lead to the conclusion that the flow is fully developed.
In fig. 5 the friction factors measured at x/'D = 57.2 for both inlet
arrangements are compared to the results of Nikaradse [20] , which were
correlated by the equation

— = 2. 0 log Rey/T- 0. 8 (39)

Vf7

The agreement is quite satisfactory for both arrangements. Without inlet

rectifier the mean deviation is 0.62 % and the RMS deviation 2.34 %. The
corresponding figures with inlet rectifier are -0.78 and 2.59 % respectively.
This is an indication that the friction pressure drop, which will be
used later in the comparison of experimental and theoretical shear stress
distributions, is correct.

7.2 The Velocity Distribution

The velocity distributions were checked for symmetry by calcula­

tion of the ratio between the values measured at the same radius on
different sides of the centre. For the channel without inlet rectifier the
mean value of the ratios (104 values) was found to be 0. 9936 and the standard
deviation 0. 0155. The corresponding values for the channel with rectifier
(78 points) was 0.9970 and 0.0089. The difference from unity is of no
practical importance.
The integrated mass-flows were compared to those measured inde­
pendently by the Foster meter. The agreement was excellent, with a mean
deviation of 0. 2 % and RMS deviation of 0.7 %.
 -17-

The dimensionless velocity distributions u* - u"*~ have been plotted

in fig. 6-, The results can be compared to the measurements of

Nikuradse [20] and Coantic [4] and the correlation of Reichardt [27]

u"*~ - u+ = 2. 5 In (40)

There is a definite difference between the two inlet arrangements,

the distribution obtained without inlet rectifier being flatter in the centre
region and steeper at the wall.
The best general agreement with earlier data was obtained with the
inlet rectifier. In particular the agreement with the results of Coantic [4 3
is excellent.
The differences between the velocity distributions show more easily
in fig. 7 where the eddy diffusivity distributions for the two arrangements
are plotted together with earlier data.
The agreement is satisfactory in the region 1.0 < r/R < 0.5, but in
the centre of the channel the data obtained without rectifier are almost
twice as large as those obtained with the rectifier, the latter being in
reasonable agreement with earlier data. However, the theoretical shear
stress and not the actually measured values were utilised for calculation
of e . The picture in the centre region will be somewhat changed if the
measured values are the more correct, since those are significantly lower
in the centre region for the measurements without inlet rectifier.
It seems reasonable to assume that the differences between the two
inlet arrangements are caused by not fully developed flow conditions at
least in the experiments without inlet rectifier. It is not surprising that
the inlet conditions may effect the necessary inlet length. This shows
clearly for instance in the results of Deissler [7] . The question remains
if the flow is fully developed after 60 diameters in the channel with inlet
rectifier. Nikuradse [ 20] used 50 diameters in his main tests and found
no variation between 40 and 100 diameters. Sleicher [3l] on the other
hand found a definite change of the velocity distribution between 50 and 80
diameters. This may, however, again be caused by different inlet conditions,
Nikurad.se [.20] using a turbulence promoter at the inlet, as in this
investigation, a device not mentioned by Sleicher [3l] .
It might also be questioned if measurements close to the outlet
of the test section can be influenced from the outlet. Nikuradse [20]
made his measurements 0. 1 - 0.2 mm outside the channel and found
no difference from results obtained 2 and 5 mm and even 20 diameters
upstream from the outlet. It can be concluded that outlet effects in
this experiment, where the measurements were made 0.27 diameters
upstream from the oulet, should be negligible.
 -19-

8. DETERMINATION OF THE EMPIRICAL CONSTANTS

IN THE ANEMOMETER EQUATION

8. 1 The Exponent in Collis ' Law

8.1.1 Own measurements

The earlier measurements, see Kjellstrom and Larsson [40]

indicated that, contrary to what is generally assumed, the exponent in
Collis' law is dependent on the velocity. It was decided to study this
further by special calibration experiments. Details of the arrangement,
the evaluation and the results of these measurements can be found in
appendix 2.
The results can be evaluated in two ways. In eq. (20) V"o may be
considered as the actually measured voltage at zero velocity, or may
be considered as the voltage at zero velocity obtained by extrapolation
of data in a specific velocity range. The second definition was used in
the original work of Collis and Williams [5] but the first has been pre­
ferred here for reasons given in section 8.1 .3. Using the second definition
ofV , the best agreement with the experimental data was obtained with c=0.35 .
When Vq was determined by measurements at zero velocity (free
convection only) c was found to vary with the velocity. This could be
described by the regression equation

c = A] - 0.0007782 pu (41)

where A^ is a constant which may vary from probe to probe and also
is dependent on the air temperature. The experiments were run at a
constant resistance ratio equal to 1.8.
It was found that c is not very much affected by a moderate flow
turbulence (below say 10 %), and that c is not appreciably affected by
the wire inclination.
Plots of c against Pu are shown in figs. 10 - 12.
 - 20

8.1.2 Comparison with Earlier Works

Collis and Williams [5 ], with Vq obtained by extrapolation found

c constant equal to 0.45 in the Reynolds number range 2 -44. Hoole and
Calvert [l 1 ], Newman and Leary [1 9 ] and Sandborn [29 ] found reasonable
correlation with c = 0.5, the value initially suggested by King [12]. These
results should be compared to the value 0. 35 established in this investiga­
tion with the same definition of V . With V measured at zero velocity
Patel [23] established that c =0.45 gave satisfactory correlation with his
calibration data.
Reasonable agreement between measured and extrapolated Vq was
found by Hoole and Calvert [1 1 ] and by Newman and Leary [19] for some
operating conditions.
Calculation of c from the experimental data was not made by those
authors. Variation of c with the velocity was therefore not observed, al­
though at least the results of Patel [23] indicate a decrease of c at in­
creasing velocities.
Norman [21 ] has also supplied data which may be used for calcula­
tion of c. Utilising his data at p ~ 1.2 (6 expts) the curve shown in fig.
was found [22]. The agreement in the absolute value of c is not too bad,
but the variation is more pronounced than found in this investigation. This
is not considered so serious since Normans curve is based on very few data.

8.1.3 Choice of correlation method

Since it was found that c might vary from probe to probe, and also
to some extent from experiment to experiment, it was considered valuable
if c could be established within each experiment. This was possible to
achieve since during the main experiments,the velocity distribution was
measured with a pitot probe and not with the anemometer.
The experience from the calibration experiments (see appendix 2)
showed that fitting of V , b and c in eq. (2-1) was difficult since conver­
gence was not easily obtained. In addition some experiments were run at
wire Reynolds numbers around or above the critical value (Re = 44) where
Collis and Williams [5] reported breakdown of the correlation with con­
stant c and V .
o
It was therefore preferred to use the measured value of V and to
o
determine c by a least square fit to the experimental data. The average
value of c determined in this way was assumed to be valid at the logarith­
mic average velocity. Eq. (41) was then used for calculation of c at other
velocities:

c = c - 0. 0007782 |_p u - (pu)ay J (42)

 - 21

8.2 The Direction Sensitivity Coefficient

8.2.1 Own measurements

It was concluded in the earlier report [40] that additional informa­

tion on the direction sensitivity coefficient would be valuable. Own mea­
surements were therefore made. The experiments and their evaluation
are described in appendix 3.
The results are shown in fig. 13. It was found that k2 varies with

pu. The variation can be described by the regression line

k2 = 0.0505 - 0.000415 pu (43)

with the correlation coefficient 0. 979 and a residual root mean square
0.0049.
The good agreement with the regression line does not unfortunately
mean that the determination of the direction sensitivity coefficient is very
2
accurate. The reason for this is that the values of k calculated from the
measured data are very sensitive to the values used for the exponent c
in Collis law.
. 2
This may be the reason why negative values of k were obtained
at high velocities. Negative values are theoretically unlikely since they
must be interpreted as if the cooling of the wire is reduced by the parallel
flow component. A 2.5 % increase of c would have been sufficient to in-
2
crease the value of k = -0. 01 1 1 obtained at the highest velocity, to zero.

8.2.2 Comparison with earlier works

Hinze [l 0 ] states without giving any details that k2 varies between
0.09 and 0.01 depending on the velocity (the value of k2 decreasing with
increasing velocity).
 - 22

More specific information has been supplied by Champagne [2],

Hoole and Calvert [11], Rasmussen [25] and Webster [34]. Champagne
[2] found that k is dependent on the length to diameter ratio (&/d) of the
wires. The probes used in this experiment had i/d-ratios between 220
and 224. This gives k = 0.21 by estimation from Champagne's data.
At i/d = 220 the 95 % confidence interval of his results was ±0.008
(mean value of 92 exp. ). It can be assumed that this applies in general
for his results.
2
Hoole and Calvert [1 1 ] reported k to be about 0. 04. Unfortunately
it was not stated at which velocity their measurements were made.
Measurements of the direction sensitivity of a DISA probe type
55A25 have also been made by the manufacturer [25]. If these data are
2
utilised a mean value of k = 0.0676 ± 0.0125 (95 % confidence interval)
can be calculated for the range of major interest 30° < Y ^ 60° if
c = 0.5 is assumed. With c = 0.48 k^ = 0.0434 ±0.0101 is obtained.

Webster [34] has supplied data which can be used for estimation
of the velocity dependence of k although they cover a very narrow range
of velocity.
The measurements were made for two groups of the velocity,
around 4 m/s and around 6 m/s. Several values of the length to diameter
ratio were tested, but exactly the same values were not used in the two
groups and more values were run in the second group. In order to avoid
a bias from the variation with 't/d, some values from the second group
had to be excluded. The values from each group with approximately the
same 't/d for the wire (6 expts at each velocity) were used.
The results are summarised in the table below.

Table 8.2.1 Summary of measurements of k by Webster [34 ],

't/d 213 - 1456.

Average Average 95 % confidence

interval of k^
, 2
pu k
4. 92 0.046 ±0.018
7.40 0.037 ±0.016
 - 23 -

A comparison of the results from these three investigations with

those of this work is made in fig. 13. The agreement is quite satisfactory
except perhaps for the result of Rasmussen [25]. The large influence of
the value of c used in the evaluation must, however, not be forgotten.
Considering this the agreement is surprisingly good.
Measurements of the directional sensitivity of hot-wires have also
been made by Kronauer [13], Sandborn and Laurence [28], Chu [3],
Davies and Fisher [6], Spangenberg [32], and Newman and Leary [19].
The former five references were discussed by Champagne [2],
Chu [3] and Delleur [8] found no appreciable deviation from the sine-
law in the interesting range of Y. Sandborn and Laurence [28],
Kronauer [ 13] and Davies and Fisher [6] suggested different forms of
the direction sensitivity relation. Newman and Leary [19] used the
sine-law of direction sensitivity but found that different exponents should
be used on the velocity and the sine term in the characteristic equation.
For the velocity c = 0.5 was employed whereas the exponent 0.457 was
used for the sine term. Time has not allowed further studies of this,
but a comparison of all the available experimental data on the same
basis would certainly be of great value.
Spangenberg's [32] results can unfortunately not readily be utilised
2
for a comparison since they do not allow calculation of k due to missing
information. Also the measurements at different inclinations were made
with wires of different -t/d ratios, which according to Champagne's [2],
findings would give a bias.
 - 24 -

9. MEASUREMENTS OF SHEAR STRESS DISTRIBUTIONS

AND TURBULENCE INTENSITIES

9. 1 The Shear Stress Distribution

9.1.1 General

The symmetry of the distributions of - p uT r1 were tested in a

manner similar to that employed for the velocity distributions. For
the channel without inlet rectifier the mean value of the ratios of
- p u’ r 1 measured at the same radius on different sides of the centre
was 0. 978 with the standard deviation 0. 1 72. The corresponding values
for the channel with inlet rectifier were 1 . 009 and 0. 1 61 respectively.
The largest deviations (up to 50 %) occur at small values of r/R where
the relative accuracy of the measurements is low.
Where the opposite is not specifically stated, the results discus­
sed below refer to the measurements made in the channel with inlet
rectifier. Tabulated results from the experiments can be found in
appendix 6.

9.1.2 Comparison of Evaluation Methods

The experimental results were evaluated by four methods, see section

6,4. To judge which method is to be preferred, the experimental data were
compared to the theory.
According to this, cf. section 2 and appendix 1 , the shear stress
distribution is given by

dp
xr = 2 dx C

This can be written in dimensionless form as

T* (45)
R

where

T* (46)
2 dx
 - 25 -

In order to make the comparison the mean value of the data for
T* obtained in all experiments was calculated for each value of r/R.
T*
The results are shown in fig. 17, where is plotted as a
function of r/R for the different evaluation methods.
According to the theory this ratio should be constant and equal to
unity.
There is no doubt that the best agreement is obtained with method
IV (which considers the cooling by the parallel velocity component also)
and that the constancy of the ratio is best obtained with methods
II and IV which take the variation of the exponent in Collis' law with
the velocity into account.
More specific information can be obtained from table 9. 1 where
the mean value (for all r/R) of the 95 % confidence interval, and
the inclination of the line

T# r»
^ = *1 " R + %2

are given for the different evaluation methods.

Table 9. 1 C omparison of Evaluation Methods

Method I II III IV

mean value of
T*
0. 945 0. 967 0. 972 0. 994
r/R
95 % conf.
limits ±0.030 ±0.019 ±0.034 ±0.021

K1 0. 1 32 0.033 0. 160 0.060

Even if method II shows better linearity (smaller K^) and a

smaller scatter, method IV with its excellent mean value must be
preferred.
A plot of T% evaluated by method IV, versus r/R is shown in
fig. 18, where also the data measured without inlet rectifier are
included. The agreement of the latter with the theory is less good,
which may be due to an insufficiently developed flow.
 - 26 -

It was also checked whether a correlation between -y^ and the

channel Reynolds number could be found. The data are plotted in fig, 19.
The correlation coefficient was found to be 0.131, which is not significantly
different from zero even at the 0. 8 level.

9.1.3 Influence of the Compressibility Correction

The distributions of the correction term Q for each experiment are

included in the tables of appendix 6.
At low velocities the influence is negligible but at the highest
Reynolds number run, exp. V52 it is quite important, varying from
about 2 % of the shear stress in the wall region to about 7 % close to
the centre of the channel.

9. 2 The Axial and Radial Turbulence Intensities

The components of the axial and radial turbulence intensity were

evaluated by method IV. The results are tabulated in appendix 6.
Mean values of Vu*"^/u* and Vr * ^/u* for each radius are plotted

in fig:s 14 and 15. The variation of the axial turbulence in the centre
number is shown in fig. 1 6
It can be observed that the measurements in the channel without
inlet rectifier gave higher values particularly in the centre region.
A further discussion of the results can be found in section 10.

9.3 Observed Standard Deviations

For planning of further experiments knowledge of the standard

deviations which can be expected is very valuable. The necessary
number of experiments for specified risks for errors of the first and
second kind can then easily be calculated. .
The standard deviations for Vu1 ^/u*; r1 ^/u* and T* estimated

by the well known relation

(47)

are shown in fig. 20.

 - 27 -

10. COMPARISON WITH EARLIER TURBULENCE AND

SHEAR STRESS MEASUREMENTS IN CHANNELS

10.1 Earlier Investigations

Comparable turbulence measurements have been made by

Reichardt [26], Newman and Leary [19], Laufer [15], Sandborn [29],
Coantic [4], Martin and Johanson [17] and Patel [23].
Data about the experimental conditions and evaluation methods
have been collected in table 10.1.1, page 29.
The main part of the measurements have been made with air and
at L/D of the order of 50. Reynolds numbers up to 5 - 10 have been
run but there is no indication that the Reynolds number has any impor­
tant influence in the investigated range. Most measurements have been
made at pu < 40 kg/m
z s. Patel [2 3] who has run up to ~ 90 kg/m
z s
forms an exception.
The l/d-ratio for the hot-wire probes is of the same order as
in this investigation except in the investigation of Newman and Leary [19]
where a much larger value was used.
The general conclusion is that the conditions in this investigation
seem to agree very well with those in earlier works, except for the
velocity where higher values were run in this investigation.

10.2 Shear Stress Distributions

The agreement of the measured shear stress distributions with

the theory is in general excellent, although most investigations show
a slightly curved distribution, see for instance fig. 21, probably due
to neglect of the compressibility effects on the shear stress distri­
bution.
The excellent agreement may seem surprising considering that
the evaluation methods used in most earlier investigations were found
less satisfactory here, cf. section 9. 1 . This may be explained by the
generally lower values of c used by earlier authors for slanting wire
2
probes, which may compensate for neglecting k .
 - 28 -

Newman and Leary [1 9], Laufer [15] andSandborn[29]used c = 0. 5

2
and worked at pu ^ 20 kg/sm , where for probe V 81 in this investigation
a value 0.526 was used, which would give a 1 1 % lower shear stress.
Patel [23] used c = 0. 45 at "ptf ~ 75 kg/sm^, where probe V 81 gave

c = 0.482 which would give a 15 % lower shear stress.

The accuracy of the earlier investigations is not easy to judge
since the mean values but not the standard deviations are given. Sandborn's
[29] data fig. 22 can, however, be used for calculation of standard devia­
tions. They are found to be between 0. 1 and 0. 05 in in close agreement
with the results of this investigation, cf. fig. 20.

10.3 Turbulence Intensities

The axial and radial turbulence relative the shear velocity are
compared to earlier results in fig:s 14 - 15.
The agreement is most satisfactory for the axial turbulence and
also good for the radial turbulence if the data of this investigation are
evaluated by method IV.
____The axial turbulence relative the axial velocity at the tube centre
(Vu 1 ^/u) is compared to earlier results in fig. 16.

The correlation of Sandborn [29]

-0. 146
=0.1 44 Re (48)
' u 'o

5
seems to give too low values when extrapolated above Re = 2 • 10
Martin and Johanson^s [17] data are not shown in fig. 1 6, but they
5
reported excellent agreement up to Re = 1 . 6 • 10 .

1 1 . CONG LUSIONS

From the comparison of the experimental turbulence intensities

and shear stresses with earlier measurements and theoretical values
it must be concluded that evaluation method IV using a varying value of c
2
calculated from eq. (42) and a varying value of k calculated from eq. (43)
is the best of the tested methods. The agreement of the turbulence intensi­
ties and shear stresses is then in every respect satisfactory.
 - 29 -

It can also be concluded that:

]. if the voltage measured at zero velocity is used in Collis " law,

the exponent c will vary with pu.

2. this variation can be described by a straight line, in the range

2
30 < pu < 160 kg/s m the inclination of which seems to be very
little affected by the particular probe chosen or the flow condi­
tions.

3. the cooling of the wire also by the parallel component of the

flow should be considered. This can be made by use of the
gelation suggested by Hinze [10] eq. (16).

4. the direction sensitivity coefficient varies with the velocity.

this variation can be described by a straight line in the range

1 0 < pu < 1 50 kg/s m^, eq. (43).

6. the determination of the exponent in Collis ' law is not

affected significantly by a moderate turbulence in the flow
(fully developed channel flow).

7. if the extrapolated voltage at zero velocity is employed in

Collis' law, which is not recommended, the best general
agreement was obtained with an exponent c = 0.35. This is
considerably less than the values between 0.45 and 0.5
normally recommended.

8. the shear stress distribution in a circular channel should

be corrected for the influence of compressibility even at the
low Mach numbers (up to about 0.35) run in these experiments.
The estimation of the correction made in appendix 1 gives
satisfactory results.
 - 30 -

Table 10.1 Earlier turbulence measurements in channels

Author ref. Anemometer Experimental conditions Evaluation

and method
year
Medium l/d pu 10"5Re
System and probes i/d

Reichardt ^ 40.7
26 Three-wire probe . A vl -0.25 No details given
1938

Newman and 19 Constant current ~400 A 115 12-22 1.15 Constant c = 0. 5

Leary 2) V-probe Sine direction sensitivity
1950 but with the exponent
0. 457 instead of 0. 5

Laufer ^
Constant c equal to
15 Constant current -250 A 51 ~3-36 0. 5-5. 0
1954 (?) R- and X-probes 0.500. Sine direction
sensitivity (?)

Sandborn ^ A 55. 5
29 Constant current and ~250 -4-40 0. 2-2.0 Constant c = 0. 500
1955 30 Constant temperature Sine direction sensi­
R- and X-probes tivity

Coantic ^ A 0. 50
4 Constant current, R- 100­ 53 10-15 The calibration function
1962 V- and X-probes 200 V = f(u) and the direc­
tion sensitivity were
established by special
experiments

Patel 23 Constant temperature -220 A 57 60-90 2.75­ Constant c = 0. 45

1963 (?) 3.64 Sine direction sensitivity

Martin and . 17 Hot-film W 52. 5 0.2-1.6

Johanson
1965

This investiga­ Constant temperature ~220 A 61.1 30­ 3.1­ cf. section 6
tion 2) R- and V-probes 170 10. 6

A Air R-probe Right angle probe 1) Rectangular channel

W Water V-probe Single slanting wire probe 2) Circular channel
X-probe Cross-wire probe
 - 31

NOMENCLATURE

A = flow area
a = constant
b = constant
c = exponent in Collis ' law
d = wire diameter
f = friction factor
K : constant
k = direction sensitivity coefficient
L = axial length in the test channel
l = wire length
m = mass flow
n = number of measurements
P = probability

P = pressure
Ap = pressure difference
R = radius of test channel
» = electric resistance
r = radius
r = radial velocity
Re = Reynolds number
T = temperature
t = statistic quantity, or time
U = perimeter
u = velocity
u* = friction velocity u* = y ——

u = dimensionless velocity u
+ _
u/-u ■it
u = bulk velocity
V = voltage
V = voltage at zero velocity
X = axial coordinate
3 = defined by eq. (20)
n = dynamic viscosity
p = density
 - 32 -

= bulk density
O' = standard deviation
t = shear stress
T* = dimensionless shear stressdefined by eq. (46)
Y = slanting angle of hot wire
£ = correction term defined byeq. (12)
 - 33 -

Subs cripts

= refers to the two hot -wire traverses made with V -probe

av = average
f = friction
l = laminar
msd = measured
tot = stagnation conditions
w = wall
wi = wire
+ = slanting angle Y
= " " 1 80- Y

Special Signs

i = fluctuating ex. u 1
= mean value ex. u
= maximum ex. u
 - 34 -

REFERENCES

1 . BROCK T E and MOON C J,

A bibliography on hot wire anemometry. 1 965.
(BHRA-BIB. -1 8).

2. CHAMPAGNE F H,
Turbulence measurements with inclined hot-wires,
Seattle, Wash. , 1 966 (Thesis, Washington Univ. ).

3. CHU W T,
Ann. Prog. Rep. Inst, for Aerospace Studies, Univ. of Toronto
42 (1964) ref. cit. par Champagne [2 ].

4. COANTIC M,
Contribution a Hetude theorique et experimental de Tecoulement
turbulent dans un tube circulaire,
Publ. Scientifiques et Techniques du Ministhre de L/air, Paris
1 962.

5. COLLIS D C and WILLIAMS M J,

Two dimensional convection from heated wires at low Reynolds
numbe rs,
Journ. Fluid. Mech. 6 (1959) 357 - 84.

6. DAVIES P O AL and FISHER M J,

Heat transfer from electrically heated cylinders.
Proc. Roy. Soc. A 280 (1964) 486 ref. cit. par Champagne [2 ].

7. DEISSLER R G,
Analytical and experimental investigation of adiabatic turbulent
flow in smooth tubes. 1 9 50.
(NACA-TN-21 38). 41 p.

8. DELLEUR J,
Les echanges thermiques de Vanemomhtre h fil chaud place
oblique me nt dans un ecoulement,
C.R. Acad. Sci. Paris 259 (1 964) 712 ref. cit. par Champagne [2 ].

9. HALL A A,
Measurement of the intensity and scale of turbulence. 1 938.
(A. R. C. Tech. Rep. R & M 1842) ref. cit. par Newman and
Leary [l 7 ].

10. HINZE J O,
Turbulence.
McGraw-Hill, New York. 1 959.

11. HOOLE B J and CALVERT J R,

The use of a hot-wire anemometer in turbulent flow.
Journ. Royal Aeronaut. Soc. 71 (1 967) pp. 511 -13.
 - 35 -

1 2. KING L V,
On the convection of heat from small cylinders in a stream of
fluid. Determination of convection constants of small platinum
wires with application to hot-wire anemometry.
Phil. Trans. Roy. Soc. (London) Ser. A 21 4 (1914) pp. 373 - 432.

13. KRONAUER R E,
Survey of hot-wire theory and techniques. 1 953. (AD-87948).
(Pratt and Whitney Res. Rep. No. 1 37) ref. cit. par Champagne
[2].
14. LARSSON A E, '
The modified atmospheric air rig FRIGGA.
1 967, AB Atomenergi Sweden (Internal report RTL-883).

1 5. LAUFER J,
The structure of turbulence in fully developed pipe flow. 1 953. 53 p.
(NACA-TN-2954. )

1 6. LINDH R,
1 965, AB Atomenergi, Sweden (Internal report RFN-1 97).
(In Swedish).

1 7. MARTIN G Q and JOHANS ON L N,

Turbulence characteristics of liquids in pipe flow.
AIChE Journ. 11 (1 965) 29 - 33.

1 8. NAZARCHUK M M,
Velocity profile and friction in a plane-parallell channel with a
developed turbulent gas flow.
Int. J. Heat and Mass Transfer 9 (1 966) 1285 - 89.

19. NEWMAN B G and LEARY B G,

The measurement of the Reynolds stresses in a circular pipe as
means of testing a hot wire anemometer. 1950.
(ARL/A 72).

20. NIKURADSE J,
Gesetzmassigkeiten der turbulenten Stromung in glatten Rohren.
VDI Forsch. Heft 356 (1 932).

21 . NORMAN B,
Hot-wire anemometer calibration at high subsonic speeds.
DISA Information 1967:5, 5 -19.

22. NORMAN B,
Division of Steam Engineering, Royal Institute of Technology
Stockholm, Sweden.
Personal communication 1 967 Dec. '

23. PATEL R P,
Measurements of the Reynolds stresses in a circular pipe as a
means of testing a DISA constant-temperature hot-wire anemo­
meter. 1 963.
(Mech. Eng. Res. Lab. , McGill University, Montreal T N 63 - 6).
 - 36 -

24. PRANDTL L,
Ministry of Aircraft Production Volkenrode, Rep. and Trans.
No. 64, (1946), ref. cit. par Champagne [2 ].

25. RASMUSSEN C G,
Dansk Industrisyndikat A/S Herlev Denmark,
Personal communication 1 965 Dec. 1 3.

26. REICHARDT H,
Messungen turbulenter Schwankungen.
Die Naturwissenschaften 26 (1 938) 404 - 8.

2 7. REICHARDT H,
Vollstandige Darstellung der turbulenten Geschwindigkeitsver-
teilung in glatten Leitungen.
ZAMM 31 (1 951 ) 208 - 1 9.

28. SANDBORN V A and LAURENCE J C,

Heat loss from yawed hot wires at subsonic Mach numbers. 1 955.
44 p.
(NACA-TN-3563) ref. cit. par Champagne [2 1.

29. SANDBORN V A,
Experimental evaluation of momentum terms in turbulent pipe
flow. (1 954).
(NAC A-TN-3 26 6).

30. SANDBORN V A and SLOGAR R J,

Study of the momentum distribution of turbulent boundary layers
in adverse pressure gradients. 1 954.
(NAC A-TN - 3264).

30 B SHAPIRO A H,
The dynamics and thermodynamics of compressible fluid flow.
Vol. 2. New York 1954, p. 1123.

31 . SLEICHER C A jr,
Experimental velocity and temperature profiles for air in tur­
bulent pipe flow.
Trans. AS ME 80 (1 958) 69 3 - 704.

32. SPANGENBERG W G,
Heat-loss characteristics of hot-wire anemometers at various
densities in transonic and supersonic flow. 1 955. 82 p.
(NAC A-TN- 3381 ).

33. WARD SMITH A J,

Subsonic adiabatic flow in a duct of constant cross -sectional area,
Roy. Aeronaut. Soc. J. 68 (1 964) 117 - 26.

34. WEBSTER C A G,
Note on Sensitivity to yaw of a hot-wire anemometer.
J. Fluid Mech. 1 3 (1 962) 307 - 1 2.
 - 37 -

35. DISA ELEKTRONIK A/S HERLEV DENMARK,

"DISA Constant temperature anemometer Instruction manual", 1966.
(Reg. No. 91 50 A 021 3).

36. Ibid.,
"Instruction manual for DISA random signal indicator and correlator
type 55A06", 1 962.
(Reg. No. 91 50 A 031 1 ).

37. NATIONAL BUREAU OF STANDARDS,

Tables of thermal properties of gases,
Washington 1 955 (NBS Circ. 564). "

38. KJELLSTROM B, LARSSON A and LINDH R,

1 964. AB Atomenergi, Sweden (Internal report RPL-665).

39. KJELLSTROM B and HEDBERG S,

On shear stress distributions for flow in smooth or partially
rough annuli. 1 966.
(AE -243).

40. KJELLSTROM B and LARSSON A E,

1 967. AB Atomenergi, Sweden (Internal report RTL-958).
 - 38 -

AC KNO W LEDGE ME NTS

The authors want to express their gratitude to Mr. Bror Hedberg

who prepared the test equipment and assisted during the measurements,
to Mr. Rolf Lindh who coded the computer programme for evaluation of
the data, to Messrs. Andreas Borsos and Goran Karlsson who made
some of the measurements and to Mr. Peter Lundborg who carried
out many of the necessary manual calculations.
 - 39 -

APPENDIX 1

THE SHEAR STRESS DISTRIBUTION FOR COMPRESSIBLE

AXISYMMETRIC FLOW AT LOW MACH NUMBERS IN A
CHANNEL WITH CONSTANT CROSS-SECTION

1 . BASIC EQUATIONS

Navier-Stokes equations for compressible flow with constant

viscosity written in cylindrical coordinates (x, r, cp) can as for the
equations for incompressible flow, see Kjellstrom and Hedberg [39 ],
be transformed into the relations for the mean values of turbulent flow,
known as Reynolds equations.
The difference from the equations for incompressible flow will
be terms containing the compressibility

du 1 drr J_ dr cp 0-1)
6 = +
dx rdr r ™dtp

and correlations containing the density fluctuation p 1 .

These correlations may be neglected if the Mach number of the
turbulence

(1-2)
Mat

is small, see Hinze [10]. For subsonic channel flow, which will be
discussed below, this is definitely the case. The ratio p'/jTis of the
2 .
order of Ma^ which for subsonic channel flow is less than 0.001 .
Utilising the condition of symmetry that all derivatives of mean
values with respect to cp are zero, the Reynolds equations will read:
in the x-direction,
 - 40 -

—— Su du dp , d / - du
Pu d7 + P r Pu
d7 ~ & + '§x V d7 "

l d_
r dr [ Hr dud7 - r p u' rT (1-3)

in the r-direction,

---- df di Pu’f0+ r feC71 dr

pu 37 + ^I = 37 "57
r P

r_ T[ d_0
- T1 2 (1-4)
3 dr
r

and in the tp-direction

d_ 2"p r’tp1 = 0
- r p u* 5pf (1-5)
dx

The continuity relation for the mean values reads:

d —— 1 d_
rp r = 0 (1-6)
37 Pu + 737

. For incompressible flow p“ is constant and thus 0=0. Then it can

be concluded that ^ will approach zero for large values of x in a channel

with constant cross -section independently of the velocity distribution

at x = 0. The flow is said to be fully developed when ~ 0. For such
flow it is easy to show that the radial velocity r will be zero. Since the
local Reynold's number will be constant in the axial direction for the
fully developed flow, the similarity laws require that all derivatives of
velocities and velocity correlations with respect to x are zero.
 - 41

It can then be shown, see Kjellstrom and Hedberg [39], that the
shear stress distribution across a circular channel will be linear

du r dp
T = T1 37 - Pu'r1 (1-7)
xr R dx

The effects of compressibility on this equation will be discussed below.

2. SIMILARITY CONSIDERATIONS

Experience tells that stationary incompressible flows in constant area

channels can be fully described if a characteristic velocity, a character­
istic length describing the cross-section, the fluid density and the
viscosity are given. It will be assumed that a compressible flow can be
Characterised by a characteristic velocity u^> a characteristic length of
the cross-section, a characteristic density, a characteristic sound
velocity of the fluid and the viscosity.
We define the characteristic velocity u^ as .

R
ub = \l urdr (1 -8)
R 0

the characteristic density p as

R
f Purdr
0
Pb (1 -9)
R
J*u rdr
0

We define also the characteristic length as 2R and the characteristic

sound velocity as aQ, the reservoir value which for adiabatic flow should
be constant along the channel.
 - 42 -

By dimensional analysis it is then found that for instance the

axial velocity u at a position (x, r) will be given by

Pbub2R x
Tj ' 2R (1 -10)
ub

Similar equations for the other velocity components and the velocity
correlations can easily be written.
The relation for the density reads:

P f Pb Ub 2R %
JL. O (1 -11)
P
b
v n a
o
2R ’ R J

It is now interesting to notice that Deissler [7 ] by experiments

up to outlet Mach numbers of 1.0 has found that at any r/R, vf/u^
reaches a constant value above a certain value of x/2R. This means
that the effect of u^/aQ on u/u^ is negligible' ^.

It seems reasonable then to assume that _

a) all other velocities and velocity correlations made dimensionless

by division by u^ or u2 are independent of x/2R and u^/aQ for

values of x/2R where u/u^ is independent of-x/2R.

b) also p/p^ is independent of x/2R and u^/aQ for that part of the
channel.

Similar assumptions were made by Ward Smith [33] in his

theoretical study of compressible flow in ducts.
It should be realised that the following considerations are based
on these assumptions, which although they seem reasonable in view of
Deissler 's [7 ] results are still assumptions and should be tested by
further experiments.

1)
PbUb2R
is constant in the channel.
11
 - 43 -

3. CONCLUSIONS FROM THE CONTINUITY EQUATION

The continuity equation reads

d —— 1 d —™ „ (1-12)
pu + 7'57rpr = 0

Introduce dimensionless velocities and a dimensionless density de­

fined by

U = — (1-13)
Ub

V = (1-14)
U,

Y = ^
(1-15)

It should be observed that

J* P u rdr = m (1-16)
pbub = TZ
R 0 TTR

is independent of x and that and u^ are independent of r.

Eq. (12) can then be written as

ul^+Ylr + 7fe rYV = 0

(1-17)
 - 44 -

But according to the assumptions above ^ and are equal to

zero after a certain length. Then apparently

(1-18)

But since y V is zero at the walls it must be zero throughout the flow.
The dimensionless density y is definitely non-zero and thus V = 0.
If the dimensionless velocity and density are introduced in eq. (16)
one obtains

J* y u rdr = Mj-
(1-19)
0

. 4. CONCLUSIONS FROM REYNOLDS 'EQUATION

IN THE RADIAL DIRECTION

Introduce the dimensionless velocities and the dimensionless

density defined by eq:s (13 - 15) and further the dimensionless correlations

2 (1 -20)
u’

2 (1-21)
VT

—T7T
(1-22)
u.

in eq. (4). After division of both sides by (p^ u^)/2R one obtains
 - 45 -

du, V
e[vu^l+vv|X] = 2R dp.
2 dr
pbub

, Tl r 4R2 ( 8 ubV . 1 8 8ubV ubV

pbub2R I- ub ^ 8xZ r 3r ' dr " ,2 )]-

2r rs
pbubL3x [ fi^vtrr
b b + 7l?rpb%’' v2] +

. iRis.r^V , i 8rubv i
+
n (1-23)
Py u^ 2R 3 dr L dx r dr J

This can be simplified considerably by utilisation of the results

of sections 2 and 3, namely that derivatives of dimensionless velocities
and correlations with respect to x are zero and that V = 0.

2R £- 2r(\ UW" 1_ d ■ 2 'Y

0 = - + - r Y V
u. dx r dr J
pbub

1 4R2 dU
(1-24)
Re 3 u, dx dr
b

If this equation is multiplied by u^/2R and differentiated with __

respect to x one obtains

d2u
1 d dp b 2R dU
Y Urvrr>
V (1-25)
Pb% ^ ^ 3Re dr
dx

Or after rearrangement

d dp d ub /"2R dU
Y arv'') (1 -26)
Fr dx Vb dx2 ^
 - 46 -

d2 u,
For cases where ----- is zero or neglectable (this is discussed
dx d d_
in more detail in section 6 of this appendix), is zero or small
which means that the variation of ^ across the channel is zero or

small. This will be utilised in section 5 below.

5. CONCLUSIONS FROM REYNOLDS' EQUATION

IN THE AXIAL DIRECTION

After introduction of the dimensionless velocities, velocity correla­

tions and the dimensionless density in eq. (3) and division of both sides
by (p^ u2)/2R one obtains

2R
dp
2 dx
pbub

2 r a Bubu 2 -,2^
2 dx x dx - pbubYU J+
pbub

dUb°
r Pb^u^r) +
r or v dr
pbub

2RTI 3 f 8 UbU . 1 8rubV>

+
2 dx V dx r dr J 0 -27)
-3PbUb

Simplification according to the conclusions from sections 2 and

3 gives

2R_ 3£ _ 2RX-(u2 + U'2)Tk.

0 = -
pb-b 6X “b dx

+ r & 37 Y UrVr 1 + 4__ U_ R2 d ub

2R (1-28)
3 Re u, , 2
b dx
 - 47 -

For reasons of symmetry, at r = 0

= 0 and U* V* = 0
or

Integration of eq, (28) therefore gives

r
r dU &YU’yT i
rdr +
Re dr
2Rpb^

J* y(U2+ U' 2) rdr J_ J JUrdr

2Rub dx 3 R^b
0 0

(1-29)

If the integration is extended over the entire channel, to r = R

where U 1 V1 =0 one obtains

R R
du.
d
E(^ —^ f If *o dx + =i-
Rub ^
dx I Q ,U2rdr +

+ d2"bR
J "U rdi (1-30)
3 Reub dx2 0

According to the definition of u, the last integral is equal to

R
J* U rdr (1-31)
0

Define a "bulk" static pressure by

 - 48 -

R
Pb J P rdr (1-32)
0

so that

(1-33)

Eq. (30) can then be written

R
2R f dU'X 1 ^b
+
/ y(U2+U’ 2)rdr -
Re Vdr✓R Rub dx
0

I 4 (1 -34)
3 Re^

This equation will be discussed further in section 5.

Transformation of eq. (29) into dimensional form again gives

"i? ' P u' ft ) = <f ^ rdr + J- f p(u2 + u’ 2) rdr

% 0

.2 r
T1 d Ub
J* u rdr (1 -35)
3 u.
b dx 0
 - 49 -

which after some rearrangement may be written as

du
rC du - - 0= {+ pbub "d^)
rdr (1-36)

whe re

du^
C = 7 [ S t P(^2+u’2) - Pb^ } rdr] • dx

d2u' »_
_ 1JL “ b
3% <^hdr (1 -37)

d2 u
b . .. . , , d£
For cases where ----- — is neglectable ^ is independent of r,
dx
see section 4 and eq. (36) may be written as

11 i7 ■ pu - C = ! ' "dbf (1-38)

dpf
where —:— is the friction pressure gradient defined by
dx

dp %
di = + pbub dx (1-39)

The first two terms on the left hand side of eq. (38) are equal to
the shear stress T in the flow, which means that the theoretical shear
xr
stress distribution is given by

, r dPf
T " C = 2 ±T (1-40)
xr

For slug - like flow where u = u^ eq. (40) reduces to the form
used in the earlier reports [39,40],
 - 50

6. THE AXIAL VARIATION OF THE MEAN VELOCITY

According to the continuity equation

m
P bUb (1-41)
A

We define a static bulk pressure and a bulk temperature by

Pb = pb ' RT1 (1-42)

2
_^b
Tb = To (1-43)

where T , the stagnation temperature is independent of x.

We assume that the bulk pressure is equal to the pressure defined

by eq. (32)

R
■^2 S P rdr 0-44)
Pb
R 0

Combination of eq:s (41 - 43) gives

2
% K PbubA
(1-45)
2
+
K-l CP To
m

Or after differentiation with respect to x and rearrangement

^Pb _ h- 1 m (]_ Cp ^o ^Ub

(1-46)
dx H A n. 2 2 y dx
u.
 - 51 -

Further differentiation gives

d2u. c T d2u.
h-1 m P °
H A dx2
dx dx

- 2 ^VC^)2] 0 -47)
ub

Eq. (34) may be rearranged thus

R
^Pb = 2 % / Y(U2 + U’ 2) rdr -
dx 2 b ub dx
£V 0

2_
^b (1 -48)
R R

After differentiation with respect to x one obtains

d2Pi d2 u. i-X

T = 2 Pb% J
dx dx

2TJ ZdUX dub

R v5r/R dx (1 -49)

where

R
J = -W y(U2 + U’ 2) rdr
(1-50)
R 0
 - 52 -

Eq. (49) may be combined with eq:s (48) and (46) to give

d2Pl 2 3
d u , d u. 2pbubJ /-dn
= 2
dx2 u.

, 1 . % _ K-l Pb"b
G+^X^)V5o
3 u, t 2 dx n u, V2
b dx %

Elimination of by means of eq. (47) and some rearrangement

dx2
d3u
b .
gives if----- =— is neglected and the dimensionless axial coordinate
dxJ

§ = rp=r- is introduced
ZK

2 - K.
du., 3 Re u, d§
KG?>G- ^20 + b

w
"b dT o i M , du,
1 +KM3(2J + L)+iL_°i__b

(1 -52)

whe re

PbUb2R
Re = (1-53)
T1

u.,
M o == 7^ (1-54)

d2u
It can now be concluded that generally ----- can not be zero if
dx
d2 u, du.
1Ub - “b —h
-Q- is non-zero, but that----- may be neglected if is small
dx
d2u,
IC. f* "'N 2
since ------7- is of the order of —v —— > where K is between 3 and 1.5
dx^ U^VdxV

for M between 0 and 1 and Re ~ 5 * 10 .

o
 -53-

7. CONCLUSIONS

1 . The theoretical treatment of compressible flow in an axi-symmetric

channel with constant cross-section can not be carried far without
approximations and simplifying assumptions.

2. If the Mach number of the turbulence is small, the correlations

containing the density variation may be neglected. This is de-
monstrated by Hinze [10].

3. Experimental results of Deissler [7] indicate that the dimensionless

velocity u/u^ becomes independent of x after a certain inlet length.
This makes it reasonable to assume that all other velocities and
2
velocity correlations made dimensionless by division by u^ or u^,
as well as the dimensionless density p/p.^ are independent of x
after a certain inlet length.

4. For that part of the channel it may be shown that:

a) the radial mean velocity is zero

b) the shear stress distribution in the channel is not linear

as for incompressible flow but given by

, r dPf
T - c = z sr (1-55)
xr

where

dp dUj
L = d£ (1 -56)
dx dx + pbub dx

and

r
/ [p(u2 + u’ 2) - pbu2]rdr
c= ru
b
dx
(1-57)

this holds when

i dub . J___ dp , since the terms containing

---- is small in comparison with
u Pb"b ^

d2u.
may then be neglected.
dx
 - 54 -

APPENDIX 2

DETERMINATION OF THE EXPONENT IN CO ELIS' LAW

1 . THEORETICAL CONSIDERATIONS

1 . 1 General

V2 - V2 = b(pu)C (2-1)

where Vq is the voltage at zero velocity.

It is not likely that the heat transfer mechanisms are the same
at very high velocities and at zero velocity. If therefore Vq is the voltage
actually measured at zero velocity, it would not be surprising if b and c
in eq. (l ) were found to vary with the velocity (or the Reynolds
number).
In eq. (l) Vq may also be considered as the voltage at zero velocity
obtained by extrapolation of data in a specific velocity range. Then V ,
b and c can be considered as constant over specific ranges of the velocity.
This is in fact the way the work of Collis and Williams [5 ] must be inter­
preted.
However, even if Vq is put equal to the actual voltage at zero velocity
eq. (l) may still be used for correlation of calibration data. This has some
advantages over the method of Collis and Williams [5] see 8.1.3 in the main
text.
The special calibration experiments which were run, were analysed
with both methods.

1 .2 The Method of Collis and Williams [5]

Since the majority of the experiments were made in the Reynolds

number range (Reynolds number based on wire diameter) 2 - 44, with
a few points at up to 53, V , b and c should be constant according to Collis
and Williams [5 ].
 - 55 -

It was therefore first tried to fit values of Vq, b and c in eq. (l)
by the method of least squares to the data from each of the experimental
runs. It was found that convergence was very difficult to obtain, probably
due to the fact that changes of the different constants may balance each
other. It was therefore decided to fit only Vq and b for different given
values of c between 0.3 and 0.5 and to choose the value of c which gave
minimum residual sum of squares.

1 . 3______ Established by Measurements at Zero Velocity

There is no reason to expect that c is constant if Vq is the actual

voltage at zero velocity. Since the value of b is unknown c can not be mea­
sured at a specific velocity. It can be established, however, by measure­
ments of the voltages Vj and V^ at two velocities Uj and u^ as a mean
value over the velocity range u^. - u^.
Inserting the values in eq. (1) gives

V, - = b(p^ (2-2)

vn-vo = b<Ki (2-3)

(where c and b are mean values over the range considered).

Division of eq. (2) by eq. (3) and taking logarithms gives
i - vg) - w* - vg
C " IntpuJj- ln(pu)n

If small increments of the velocity are used c can be assumed

valid for the average conditions between experiments I and II.

1.4 Influence of Turbulence on the Determination of c

Earlier experience indicated, cf. Kjellstrom and Hedberg [39]

that the value of c could vary not only between probes but also from test
to test, dependent among other things on the air temperature. It is
therefore an advantage if c can be determined within each experiment.
This can be achieved in the experiments described in this report, where
the velocity distribution in the channel is measured by a pitot probe. It
must then be assumed that the influence of the turbulence on the determina­
tion of c is negligible. This assumption will be checked theoretically here.
 -56-

The conclusions are compared to experimental results below.

The possible dependence of b on the velocity and variations in
the density will be neglected. Insert in eq. (1)

V = V + V' (2-5)

u = u + u1 (2-6)

square, and take the time average

(V2 - V2 + V' )2 = b(u2 + u 1 )C (2-7)

insert from eq. (21) in the main text

2 _2
u1 =^2 ' C=r~x)
v2 - v2
V1 (2-8)
O

and rearrange

(V2 - vl) (' ♦ ^-a)

v"2 - v2
- b «- [. + ^(=^b)2]c/2
c2 VV2 - V2
(Z-9)
o o

or
0 + v2 - v2
(v2 - v2)
o , —c
= b • u (2-10)

['^(=Z^)2]CA

comparison with eq. (l) shows that in turbulent flow a correction factor
must be introduced. The value of this and its influence on the determina­
tion of c is discussed further in section 3. 3 of this appendix,

2. EXPERIMENTAL ARRANGEMENT FOR MEASUREMENT OF c

In order to check the influence of the flow turbulence on the determina­

tion of c, measurements were made both in a low turbulence jet after a
nozzle and in the turbulent flow in the centre of the test channel. Details
 - 57 -

about the arrangement of the nozzle and the flow conditions in the
jet can be found in appendix 3. Except for the arrangement of the
flow the techniques were similar in the two series of experiments.
The hot-wire probe and a pitot tube were mounted in parallel on
the probe support at a distance of 20 mm. Readings were first taken
with the pitot tube. The support was then displaced 20 mm so that
readings with the hot-wire probe could be made at the same point.
Three repetitions were made. One stagnation pressure reading,
and for the majority of the experiments one voltage reading was
made in each sequence. In one experiment 100 voltage readings
were taken in each sequence in order to obtain better information
about the accuracy of these measurements.

3. EXPERIMENTAL RESULTS

The measured values are given in table 2. 1.

3. 1 Evaluation by The Method of Collis and Williams

The residual sum of squares when Vq and b in eq. (l) was fitted
to the experimental data is shown in fig. 8 for different values of c.
A minimum somewhere between 0.34 and 0.35 was found. This is
considerably below the value 0. 45 established by Collis and Williams
[5 ] in the same Reynolds number range. Fig. 8 shows the total sum
of squares for all experiments. The individual sum of squares for each
experiment show some variation of the value of c which gives a minimum,
the extreme values being 0.45 and 0.30.
A comparison between the correlation lines at c = 0. 35 and the
experimental data has been made for some of the experiments, in fig.
9 . The agreement is quite satisfactory.
This means that the method of Collis and Williams [5] can well
be used for correlation of calibration experiments. The method, however,
has some disadvantages in comparison to the other method. This is dis­
cussed in the main report, section 8.1.3.

3.2______ Vq Established by Measurements at Zero Velocity

It was found, as expected, that c was not constant but varied with
the velocity.
 - 58 -

Results obtained in channel flow for different probes are shown

in fig. 1 1 . A comparison between results for one probe in channel flow
and in a jet after a nozzle (about 0.2 % turbulence) is made in fig. 10.
Results of measurements in the jet at different air temperatures are
shown in fig. 12.
For comparison of the results of the different experiments,
regression lines

c = A.| + A^ • pu (2-1 1)

were fitted by the method of least squares to the results from each
experiment. No significant differences between the regression coefficients,
A-,, were found. The regression lines could therefore be considered, as
parallel and a weighted mean value of the regression coefficient A^ was
calculated. It was thus found that the variation of c with pu could be
described by

c = A1 - 0.0007782 pu (2-1 2)

with a residual mean square of 0.0007514 and a 95 % confidence interval

of the regression coefficient of 0.000122. The distances between the
regression lines were then tested for significance. It was found that no
significant difference could be established between the measurements in
turbulent and in low-turbulent flow. This will be discussed further in the
next section. The results for different probes were, however, found to be
significantly different and so were the results for different air tempera­
tures, the higher temperature giving lower values of c. This is in agree­
ment with the findings of Norman [ 21]. The existence of a variation of
c with pu is also in agreement with Norman's results, cf. fig. 1 1 where
a comparison is made. Norman found a much steeper variation but the
data on which his curve is based are few so this does not throw any
serious doubt on the results obtained here.

3. 3 Influence of Flow Turbulence on The Determination of c

A comparison of the results from measurements of c in the centre of the

channel (axial turbulence level ~ 3 %) and in a low-turbulence jet (axial turbulence
level 0.2%) showed no significant difference as was mentioned above. This
 - 59

is in agreement with the theoretical considerations of section 1.4 in

this appendix, since if pertinent‘data are inserted in eq. (l 0), it is
found that neglect of the correction factor
2

o
2
1 +
2
W2
J
V
o

causes an error of less than 0.5 % in c. If c is measured during a

traverse in a circular channel, as was made during the main experi­
ments, the error is somewhat greater due to the higher turbulence in
the wall region and the variation of the turbulence intensity across the
channel. The maximum error, calculated from data obtained during
the main experiments, was found to be 1.5%.

3. 4 Influence of the Wire Inclination on c

A small experiment was also made to check the influence of wire

inclination on c.
The results are shown in table 2. 2.
The variations of c which were found are negligible and within the
accuracy of the experiment.
It can therefore be concluded that c is not affected by the wire
inclination in the investigated range of angle, 34 < Y < 54° and velocity
61 < u < 72 m/s.
 - 60 -

Table 2. 1 . Results of experiments for determination of c

Experiment no: 680206-1

made in Channel with probe R8 1
T 00 = 283 K p sta = 1.019 bar
^ stat
r
V =5.00 volt ft/ft =1.8
o

pu V pu
kg/m^ s volt kg/m^s

132.17 10.547
0.3733 127.96
123.74 10.447
0.4155 122.40
121.06 10.410
0.3976 117.59
1 14. 12 10.317
0.3668 110.27
106.42 10.217
0.3858 102.52
98.62 10.103
0.3829 95.01
91.40 9. 993
0.4244 85.47
79.55 9. 777
0.4569 73.63
67. 70 9.517
0.3887 64. 30
60.89 9.377
0.4487 55.48
50.06 9. 090
0.4112 46.61
43.15 8. 900
Table 2.1. Results of experiments for determination of c

Experiment no: 680207

made in Channel with probe R82
T CO = 282 K p = 0.998 bar
“ stat
V
o
=5.45 volt */»a=1.8

pu V pu pu V pu
c c
kg/m s volt kg/m s kg/m s volt kg/m s

181.84 12.057 48.41 9.753

0.3727 165.05 0.4518 44.69
148.25 11.700 44. 69 9.507
0.4286 143. 34 0.4843 35.39
138.42 11.567 35.39 9.040
0.4240 134.06 0.4703 26.56
129.70 11.443 26.56 8. 720
0.4155 126.12
122.54 11.340
0.4226 1 18.97
115.39 11.230
0.3966 111.43
107.47 11.110
0.4476 104.04
100.61 10.987
0.4368 96.83
93.04 10.847
0.4182 89.41
85.77 10.710
0.4647 83.41
81.05 10.607
0.4187 76.01
70.97 10.397
0.4666 66. 97
62.97 10.187
0.4746 59.91
56.85 10.013
0.4701 52. 63
48.41 9. 753
 - 62 -

Table 2.1. Results of experiments for determination of c

Experiment no: 680208

made in Channel with probe V81
Tm = 285 K p = 1 .027 bar
c stat
V =5.40 volt A/% =1.8
o

pu V pu pu V pu
c c
kg/m^s volt kg/ m s kg/m s volt kg/ m s

178.81 11.000 48.59 8.953

0.4252 165.45 0.4699 45.11
152.09 10.720 41.63 8. 753
0.4236 145.98 0.5094 33.09
139.87 10.580 24.55 8.093
0.4562 135.39
130.91 10.463
0.4278 127.25
123.59 10.370
0.4171 119.88
1 1 6. 1 6 10.273
0.4286 112.46
108.76 10.170
0.4345 105.07
101.38 10.060
0.4619 97.48
93. 58 9.930
0.4561 90.05
86.51 9. 807
0.4376 82.67
78. 83 9. 670
0.4718 75.17
71.51 9.520
0.4830 67. 73
63. 95 9. 350
0.4715 60. 31
56.66 9. 177
0.4990 52.63
48.59 8.953 |
 - 63 -

Table 2. 1. Results of experiments for determination of c

Experiment no: 680405-1

made in jet after nozzle with probe V81
T CO = 282 K p . = 0.996 bar
r st at
V =5.48 volt =1.8
o

pu V c pu
kg/m s volt kg/m2s

197.18 11.210
0.4788 187.27
177.36 10.997
0.4645 164.40
151.44 10.700
0.4681 145.08
138.71 10.540
0.4729 134.24
129.78 10.420
0.4290 125.96
122.15 10.323
0.4884 118.32
1 14.49 10.207
0. 4746 110.48
106.46 10.083
0.4765 102.32
98.19 9.947
0.4649 94.49
90.79 9. 823
0.4967 88.15
85.50 9.723
0.4719 81.59
77.67 9.575
0.4872 74. 19
70.71 9.430
0.5072 66.57
62.43 9.237
. 0.4893 59.01
55.58 9.070
0.5691 54.57
53.56 9.010
 - 64 -

Table 2. 1. Results of experiments for determination of c

Experiment no: 680405-2

made in jet after nozzle with probe V81
T = 320 K p . = 0. 997 bar
” r stat
V = 5.49 volt ft/Tl =1.8
o ' a

pu V c pu
kg/m s volt kg/s

53. 28 9. 233
0.4524 56.30
59. 31 9.380
0. 4211 63.45
67. 59 9. 553
0. 4410 70.29
73.00 9. 663
0.4286 76. 34
79. 68 9.787
0.4377 84. 28
88. 88 9. 950
0.4627 92. 15
95.42 10. 065
0.4163 98. 82
102.21 10.167
0.4143 105.82
109.43 10. 270
0.3709 1 12.73
116.02 10. 350
0.4256 1 1 9. 77
123.51 10.450
0.3624 126.40
129. 30 10. 513
0. 4065 135.95
142.60 10.667
0. 3716 152.91
163.22 10. 867
0.3628 1 72.78
182. 35 11.033
 - 65 -

Table 2.1. Results of experiments for determination of c

Experiment no: 680417

made in jet after nozzle with probe V81
T = 281 K p , = 1. 028 bar
co S tclt
V = 5.42 volt ft/ft =1.8
o cL

pu V c pu
kg/m s volt kg/m^s

140. 05 10. 593

0. 4786 136.39
132.73 10.493
0.4281 128.53
124. 33 10. 387
0.4758 120.71
117.10 10.280
0.4577 113.61
110.13 10. 177
0.4868 106.09
102. 06 10. 047
0. 469? 98.49
94. 91 9. 923
0.5035 91.06
87. 20 9.777
0.4712 83.23
79. 26 9.627
0.5766 75.61
71.96 9.447
0. 47 81 67.64
63. 32 9. 257
0.4964 60. 01
56.71 9. 093
0.5206 52. 80
48. 89 8. 873
0. 5190 44. 94
40. 99 8.627
0.5023 37.41
33. 83 8. 383
0.4502 29. 97
26. 11 • 8.111
 - 66 -

Table 2.1. Results of experiments for determination of c

Experiment no: 680423

made in jet after nozzle with probe V81
T = 282 K 1.020
r stat
V = 5.42 9 volt «/* 1.8
o

Mean values of 100 voltage readings.

pu V pu
c
kg/m s volt kg/m s

139.16 10.6175
0.4390 132.74
126.31 10.4895
0.4243 122.08
117.85 10.3776
0. 4595 113.91
109. 97 10.2589
0.4545 105.68
1 01.40 10.1244
0.4548 97. 77
94. 1 4 10.0039
0.4635 90. 96
87. 78 9.8908
0.4037 84.29
80.79 9. 7763
0.4747 76.37
71.94 9.5935
0.4736 67. 78
63.61 9.4070
0.4744 59. 88
56. 15 9.2251
0.4910 52.41
48. 67 9.0183
0.4984 44.89
41.11 8.7835
0.5257 37.26
33.40 8.4980
0.5462 30. 91
28.43 8.2835
0.5734 24. 45
20.46 7.8671
 - 67 -

Table 2. 2.

Results of experiment for study of the influence of the inclination of

the wire on the exponent in Collis ' law

u1) y1) ln(V2-V2)

¥ In pu c
m/s volt

72.49 9.680 4.5324 4. 1580

34° 0.4065
61.37 9.458 4.3634 4.0893

72. 49 9.998 4.5324 4.2514

%

0.4000
61.37 9.766 4.3634 4. 1838

72.49 10.240 4.5324 4. 3188

54° 0.4053
61.37 9.994 4.3634 4.2503

l) Mean values of 5 measurements.

V = 5.455 volt,
o
 - 68 -

APPENDIX 3

MEASUREMENT OF THE DIRECTION SENSITIVITY COEFFICIENT

1 . THEORETICAL BACKGROUND

It is assumed that the governing equation for the anemometer is

*(« = u%in2 Y + k2cos2Y)C/2 (3-1)

This is supposed to hold for the instantaneous values. In a flow with

low turbulence it holds of course also for the time average values so
that
V2 -V2
(sin2 Y + k2 cos2 Y)°/2
= a (3-2)
K*{*~*J
P b

After some rearrangement this may be written as:

1 1 1
(v2-yV/c
sin Y + = 0(3-3)
-2
b 1 -k u 1 -k

If an experiment is so arranged that the mean voltage is measured

for different inclinations Y in a flow with constant temperature and constant
velocity eq. (3) reduces to

sin2 Y - K(V2 - V2)2/° +

(3-4)

where K is a constant.

This means that the direction sensitivity coefficient can be calculated

2
from the intersection of the sin Y - axis by the straight line represented
by eq. (4) if sin2 Y and (V2 - V2)2^C are used as variables. A straight line
° 2
was therefore fitted by the method of least squares, to the data for sin Y
and (V2 - V2)2^C established in the experiments, k2 was then determined

from the equation of that line.

 - 69 -

2. EXPERIMENTAL ARRANGEMENT

A view of the experimental arrangement is given in fig. 3. The

probe was mounted in a revolving support in the cylindrical jet (diam.
70 mm) a few millimeters after a nozzle.
A protractor was used for adjustment of the probe in different
inclinations and a measuring microscope could be used for measurement
of the inclination of the hot wire. The support could be displaced so that
the hot wire was in the centre of the jet independent of the inclination.
The nozzle was mounted in the test channel used during the turbulence
and shear stress measurements and preceded by a honeycomb and a net.
The velocity distribution in the nozzle was quite uniform, see fig. 23.
The centre turbulence was measured to 0.26 % at 71.5 m/s and 0.22 %
at 96. 1 m/s. This is probably sufficiently low to make the difference bet­
ween instantaneous values and average values negligible.

3. EXPERIMENTAL RESULTS

Experiments were made with one slanting wire probe V81 , for which
measurements of c had also been made. The data from the experiments
have been collected in table 3.1. Plots of (V^ - Vo^)^C against sin^ Y are

shown in fig. 24 for the lowest and highest velocity run in the experiments.
The direction sensitivity coefficient k plotted against u is shown in fig. 1 3 .
For discussion of the results see section 8.2 in the main text.
Table 3. 1 . Results of experiments for determination of the direction sensitivity coefficient

Experiment 680402 -1 Experiment 680402 -2

Measured values of V Measured values of V

Y° volt Y° volt
Reading 1 Reading 2 Mean value Reading 1 Reading 2 Mean value

u = 1 0.86 m/s 35.0 7.41 7.40 7.405 u = 20.69 m/s 35.0 8.03 8.03 8. 030
p - 1.21 kg/m2 37.0 7.44 7.44 7.440 p = 1.21 kg/m2 37.0 8.08 8.08 8.080
pu = 13.132 kg/m 2 s 39.0 7.49 7.48 7.485 pu = 24. 937 kg/m2s 39.0 8. 1 3 8. 1 2 8. 1 25
41.0 7.52 7.52 7.520 41.0 8. 1 7 8. 1 7 8. 1 70
=1.8 43. 0 7.55 7.56 7.555 #/£o =1.8 43. 0 8.21 8.21 8.210
o
45.0 7.58 7.58 7.580 45.0 8.25 8.26 8. 255
V =5.45 volt V = 5.45 volt
o o
47. 0 7.61 7.61 7.610 47.0 8.30 8. 30 8. 300
c = 0.5321 49.0 7.64 7.64 7.640 c = 0.5226 49.0 8.34 8. 34 8. 340
51 .0 7.67 7.67 7.670 51 .0 8.37 8. 37 8. 370
53.0 7.70 7.70 7.700 53.0 8.39 8.39 8.390
k2 = 0.0473 k2 = 0.0371

55.0 7. 71 7.71 7.710 55.0 8.42 8.42 8.420

Table 3.1. Results of experiments for determination of the direction sensitivity coefficient, (contd. )

Experiment 680403 -1 Experiment 680327 -1

Measured values of V Measured values of V

Y° Y°
Reading 1 Reading 2 Mean value Reading 1 Reading 2 Mean value

u = 41.55 m/s 35.0 8.84 8.84 8. 840 u = 59.59 m/s 35.0 9.40 9.40 9.400
p = 1.24 kg/m2 37.0 8. 90 8.90 8. 900 p = 1.28 kg/m2 37.0 9.47 9.47 9.470
pTT = 51.574 kg/m2s 39.0 8.96 8.96 8. 960 pu = 76. 653 kg/m2s 39.0 9.53 9.54 9.535
41.0 9.01 9.02 9.015 41.0 9.60 9.60 9.600
#/#o =1.8 43.0 9.07 9.07 9.070 #/#c =1.8 43.0 9.65 9.65 9.650
45.0 9.12 9.12 9. 1 20 45.0 9.71 9.71 9.710
V = 5 ♦ 44 volt V = 5.44 volt
o o
47.0 9. 1 7 9.1 7 9. 1 70 47.0 9.76 9. 76 9.760
c = 0.5011 49.0 9.21 9.22 9.215 c = 0.4809 49.0 9. 81 9. 81 9. 810
51 .0 9.25 9.25 9.250 51.0 9.86 9. 86 9. 860
53.0 9.29 9.29 9.290 53.0 9.90 9.90 9. 900
k2 = 0.0276 k2 = 0.0204

55.0 9. 33 9.33 9.330 55.0 9.94 9.94 9.940

Table 3.1. Results of experiments for determination of the direction sensitivity coefficient, (contd. )

Experiment 680327 -2 Experiment 680404 -1

Measured values of V Measured values of V

Y° Y°
Reading 1 Reading 2 Mean value Reading 1 Reading 2 Mean value

rr = 87.47 m/s 35.5 10.01 10.01 10.010 u = 96.69 m/s 35.0 10.10 10.10 10.100
p = 1.29 kg/m^ 37.5 1 0.08 10.08 10.080 p = 1.26 kg/m2 37. 0 10.18 10.18 10.180
ptr= 1 1 3.01 5 kg/m2s 39.5 10.15 10.15 10.150 pu = 1 22. 1 98 kg/m2s 39.0 10.25 1 0. 25 10.250
41.5 10.21 10.21 10.210 41.0 10.32 10. 32 10.320
= 1.8 43.5 10.27 10.28 10.275 f£/$ =1.8 43. 0 10. 38 10.38 10.380
O
45.5 10.33 10.34 10.335 45.0 10.44 10.44 10.440
V =5.45 volt V =5.45 volt
o o
47.5 10.38 10.39 10.385 47.0 10. 50 10.50 10.500
c = 0.4516 49.5 10.44 10.45 10.445 c = 0.4442 49.0 10.56 1 0.56 10.560
51 .5 10.49 10.50 10.495 51.0 10.61 10.62 10.615
53.5 1 0.54 10.54 10.540 53.0 10.66 10.66 10.660
k2 = 0.0113 k2 = -0.0063

55.5 10.59 10.59 10.590 55.0 10. 70 10.70 10.700

 - 73 -

Table 3. 1 . Results of experiments for determination of the

direction sensitivity coefficient, (contd. )

Experiment 680327-3

Measured values of V

Reading 1 Reading 2 Mean value

u = 112.95 m/ s 35.0 1 0.40 10.40 10.400

p = 1.30 kg/m2 37.0 1 0.48 10.48 10.480
pu = 146.694 kg/m2s 39.0 10.55 10.55 10.550
41.0 10.62 10.62 10.620
$/fi0 =1.8 43.0 10.69 10.69 10.690
45.0 1 0. 75 10.74 10.745
V = 5.44 volt
o
47.0 1 0. 80 10.80 10.800
c = 0.4244 49.0 10.86 10. 86 10.860
51.0 10.92 10. 92 10.920
53.0 10.97 10.97 10.970
k2 = -0.0111

55.0 11.01 11.01 11.010

 - 74 -

APPENDIX 4

DERIVATION OF A RELATION BETWEEN THE VELOCITY

FLUCTUATIONS AND THE VOLTAGE FLUCTUATIONS

Eq. (18) in the main text reads

V2 = a + b(pu)C(sin2 Y + k2 cos2 Y) C/^2 (4-1)

If Vq is the voltage corresponding to zero velocity eq. (l) can be

written

—*----- (V2 - V^) = uC(sin2 Y+ k2 cos2 Y)C/2 (4-2)

pC • b
Differentiate.'' ^

-2VdV = c uC_1 1” sin2Y + k2cos2Y j C/2 du +

pC.b L J

+ uC • ^sin2Y + k2cos2Y^ 2 ^ * ^2 sin Y cos Y - 2 k2sin Y cos yJ dY

(4-3)
which may be divided by eq. (2) to give

2VdV f du sinY cos Y - k2sinY cosY 1

—------— = c ---- +---------- ~-------- =------ =------------- dY (4-4)
V -V Lu sin Y + k cosZY J
o

Squaring this equation gives after some rearrangement

(2 V (dV)2 ^ [(du)2 + 2----- ------------------ udYdu +

V u tan Y + k cotY

+
c 1-k
tanY + k cotY
-) (u dY)2] (4-5)

^ It is assumed that the density fluctuations are negligible (cf. ap­

pendix 1). The density, p, can therefore be treated as a constant
during the differentiation.
 75 -

Consider now a wire positioned in the x-y plane at an angle Y to the

x-axis in a flow with the main flow in the x-direction. The mean
velocity is u . Turbulent fluctuations u1, v1 and w* in the x, y and z
directions are superimposed on the main flow.

The instantaneous velocity past the wire will be

u + u' )2 + v ,2 + w 2
(4-6)

or after rearrangement

u (4-7)

or approximately

]_ u
. 2
u u (4-8)
8
u u

For small fluctuations, u' ~v* ~ w* ^ 0.1 u, as in shear flow in

channels, the quadratic terms may be neglected so that

u ~u + u* (4-9)

The velocity fluctuation is apparently equal to

du = u - u = u' (4-10)
 - 76 -

Estimation of the effect of the fluctuations on

the angle between the flow and the wire is
more complicated.

With reference to the adjacent figure

the change of the angle Y is

dY = cp - Y (4-11)

The angle cp can be calculated from the cosine theorem applied to

the triangle OCD.

The sides of this triangle are

(4-12)

CD = u tan Y

(4-13)

(4-14)
 77 -

or if quadratic terms, as earlier, are neglected

OC ~ u + uf (4-15)

CD~ u tan if + uf tan Y + v* (4-16)

1
OD = (u + uf ) (4-17)
cos Y

The cosine theorem then gives

CD2 = OC2 + OD2 - 2 OC OD cos Y

(4-18)

or
u 2 tan2Y + 2 u uf tan2Y + 2 u v* tan Y = u 2 + 2 u u* —---- ^
cos Y

(4-19)

after neglect, of quadratic terms.

We may write

cos cp = cos(Y + dY) = cosYcosdY - sinYsindY (4-20)

or

cos cp ~ cosY - dY sinY (4-21)

cos cp
= 1 - dYtanY (4-22)
cos Y
 - 78 -

The right hand side of eq. (19) may be written as

( u 2 + 2 u i/) (*[ +-----L------ 2 + 2 dY tanY) =

^ ^ ^ cos^Y ^

= (u 2 + 2 u uA (tan2Y + 2 dY tan Y^ (4-23)

Eq. (19) can then be reduced to

= Qu. + 2 u u* dY (4-24)

so that

dY = (4-25,
u
u 1 + 2 u u u
u

or if quadratic terms of the fluctuations are neglected as earlier

dY = (4-26)
u

Introduction of eqs. (9) (l 0) and (26) in eq. (5) gives if Vr is employed

for notation of the voltage fluctuation, which for small fluctuations can
be put equal to dV

4(V+V')Z yt 2 2
c
1-k u’ v* +
r(v+v)2-vy2 u 2 + 2uu' +u’ 2 tan Y + k cotY

<—!^----- )2 (4-27)
tanY + k cotY
 - 79 -

Taking the time mean value and neglecting again quadratic terms of
the fluctuations gives

)2' V'2=^2[u
Q 2 V 1 2+ 2------ L* u’v1 +
v 2-v2 u tanY +k cotY
o

.)V2]
( 1 -k
tanY + k cotY
(4-28)

which is identical to eq. ( 1 9 ) in the main text if v* is replaced by the

radial velocity fluctuation rr .
 - 80 -

APPENDIX 5

CHECK OF THE RECOVERY FACTOR FOR THE WIRES

It was assumed in the theoretical considerations in appendix 2,

3 and 4 that the effective temperature difference between wire and
fluid is constant for adiabatic flow and independent of the velocity
and the inclination of the wire. This requires a constant recovery fac­
tor which is close enough to unity to make the difference between the
stagnation temperature and the adiabatic wall temperature negligible
at the velocities used in the experiments.
The assumption of a high recovery factor was supported by
Shapiro [30B ] who gives values 0.84 - 0.87 at Mach numbers below 1 . 0.
Nevertheless it was decided to check this assumption by own
experiments.
This was achieved by measurements of the cold resistance of one
probe positioned in the jet after the calibration nozzle described in
appendix 3 for the following cases:

a) probe with cover.

b) probe with arrangement to give stagnation temperature

at the wire, see adjacent figure.

c) probe arranged as for turbulence measurements

¥ ~ 45°.
_ 4 A*/es
d) probe inclined to give ¥ = 65° and 25
respectively.

The velocity in the jet was ~ 100 m/s. No difference in

the cold resistance could be detected, this being con­
stant and equal to 3. 20. This means that the assumption
of a constant recovery factor close to unity is justified.
 - 81

APPENDIX 6

TABULATED RESULTS FROM THE MEASUREMENTS IN

THE CIRCULAR CHANNEL WITH INLET RECTIFIER

Experiment V 40
Re = 7. 46 • 1 05 f= 0.0125 u^ = 79. 5 m/s (|^)f = -389. 7 N/m2

^/u* V r 1 2/u
u m/s T* C

0.970 65.9 2. 10
0.893 75. 1 1.92
0.815 79.7 1.76
0.738 82.5 1.62
0.660 85.0 1.48
0.582 87. 1 1.41 not
0.505 89. 1 1.28 measured
0.427 90. 9 1.19
0. 349 92.1 1.09
0.272 93.6 0.98
0.194 94.6 0. 91
0.116 95.4 0. 82
0.039 95.7 0. 79
 82 -

E xpe_ri_me_nt_V_42_

Re = 7.46 • 105 f = 0.0125 ub = 79.5 m/s (j&)f = -389.7 N/m2

4^T/n
r/R u m/s

0.970 65. 9 2. 1 9
0. 893 75. 1 1.95
0. 815 79.7 1.78
0. 738 82.5 1.64
0. 660 85.0 1.51
0.582 87. 1 1.42 not
0.505 89. 1 1.30 measured
0.427 90.9 1 .21
0.349 92. 1 1.09
0.272 93.6 1.00
0. 194 94.6 0.92
0.116 95. 4 0.84
0.039 95.7 0.79

E xpe_rJ1me_nt_V_42_
Re = 5.42 * TO5 f = 0.0129 u^ = 57. 8 m/s (^)f = -2 1 2. 9 N/m2

r/R u m/s Vu'2/u

0.970 47.4 2.23

0.893 54.6 1.99
0.815 58.0 1.83
0.738 59.9 1.68
0. 660 62.0 1.57
0.582 63.3 1.46 not
0.505 64.9 1.35 measured
0.427 66. 0 1.24
0. 349 67. 1 1.15
0.272 68.1 1.06
0. 1 94 68.7 0.95
0.116 69.3 0.88
0.039 69.6 0.83
 - 83

Experiment V 43
(£)f = -141.3 N/m2
Re = 4. 33 • 105 £ = 0.0132 u, = 46. 9 m/s
b

u m/ s Vu12/u* V r 12/u*
r/R T* C

0. 970 38.2 2.24

0. 893 44.2 2.01
0. 815 46.9 1.84
0.738 48.6 1.70
0.660 50.2 1.58
0.582 51.5 1.47 not
0.505 52.9 1.37 measured
0.427 53.8 1.26
0.349 54. 6 1.17
0.272 55.5 1.06
0. 194 56.0 0.97
0.116 56.5 0.89
0.039 56. 8 0.85

Experiment V 44

Re =2.62 • 105 f =
0.0139 ub = 28. 3 m/s
(§)f = -54. 1 3 N/m2

r/R u m/ s Vu12/u* Vr12/u* T* C

0. 970 22.9 2.16

0.893 26.6 2.00
0.815 28.3 1.83
0.738 29.5 1.68
0.660 30.4 1.56 not
0.582 31 . 1 1.47 measured
0.505 31.9 1.33
0.427 31.5 1.21
0.349 33.2 1.16
0.272 33.6 1.05
0. 1 94 34. 1 0.96
0.116 34.3 0.88
0.039 34. 3 0.83
 - 84 -

E xpe_riment_V_45
Re = 11.24 • 105 f = O.Ol 16 u^ = 121.7 m/s (|^)f = -81 9. 4 N/m2

r/R u m/ s T*
c
0. 970 100.8 2.17
0.893 115.2 1.94
0. 815 121.4 1.75
0.738 126.4 1.63
0.660 130.2 1.52
0. 582 1 33. 4 1.41 not
0.505 136.6 1.32 measured
0.427 139.0 1.20
0.349 141.3 1.10
0.272 143.2 1.01
0. 1 94 144. 8 0.91
0.116 145.7 0.84
0.039 146. 3 0.78

Experiment V 46

Re = 4.37 • 105 f = 0.0128 u, = 62.2 m/s 8 N/m2

b

Vu12/u* V r1
r/R u m/s T*
G

0. 970 50.9 2.41 0.78 0.911 0.018

0.893 58.2 2.20 0.97 0. 925 0.030
0.815 62.0 2.03 1.01 0.849 0.037
0.738 64.8 1.90 1.04 0.832 0.040
0.660 66.3 1.74 0.99 0.689 0.042
0.582 68.5 1.64 0.92 0.631 0.042
0.505 69.9 1.51 0.89 0.534 0.040
0.427 71.2 1.39 0.87 0.436 0.037
0.349 72.8 1.30 0.83 0.385 0.032
0.272 73.8 1.17 0.83 0.293 0.026
0. 1 94 74.5 1.08 0.74 0.223 0.020
0.116 74.9 0.99 0.74 0.138 0.012
0.039 75.7 0.95 0.75 0.041 0.004
 - 85 -

E xpe_riLment_’V_47
Re =2.43 • 10^ f = 0.0153 = 36.5 m/s = -84.36 N/m2

r/R u m/s Vu12/u* V r 12/u* T*

£

0.970 29.3 2.28 0.76 1.020 0.002

0.893 33.7 2.15 0.63 0.893 0.003
0.815 36.4 2.00 0.84 0.791 0.004
0.738 37.8 1.81 0.81 0.668 0.005
0.660 39.5 1.68 0.82 0.574 0.005
0.582 40.2 1.55 0.78 0.494 0.005
0. 505 41.3 1.44 0.77 0.435 0.005
0.427 42.0 1.33 0.71 0.381 0.004
0.349 42.8 1.23 0.72 0.314 0.004
0.272 43.6 1.13 0.70 ' 0.251 0.003
0.194 44. 1 1.03 0.68 0.174 0.002
0.116 44.5 0.96 0.65 0. 108 0.001
0.039 44.4 0.90 0.64 0.015 0.000

E xp e_rim e_nt_V_48
Re = 5.62 • 105 f = 0.0131 u^ = 77.5 m/s (|^)f = -343.2 N/m2

Vu1^/u* V r12/u*
r/R u m/s T*
£

0. 970 62.9 2.24 1.06 0. 920 0.044

0.893 72. 8 2.04 1.05 0. 885 0.072
0.815 77.0 1.86 1.10 0. 805 0.088
0.738 80.6 1.73 1.09 0.727 0.097
0.660 83.2 1.61 1.05 0.625 0. 101
0.582 85.3 1.49 1.04 0.586 0. 100
0.505 87.5 1.38 0.485 0.095
NO
00
o

0.427 89.1 1.28 0.93 0. 406 0.087

0. 349 90. 3 1.16 0. 87 0. 345 0.076
0.272 91.6 1.06 0.82 0.265 0.062
0. 1 94 92.5 0.97 0.75 0. 1 82 0.046
0.116 93.5 0.90 0.73 0.114 0. 023
0.039 94. 1 0.85 0.72 0.045 0.010
 - 86 -

Experiment V 49
Re = 3.22 • TO5 f= 0.0142 u^ = 47. 9 m/s (^)f = -1 35. 6 N/tn2

. r/R u m/s Vu12/u* V r12/u* T* c

0. 970 38.7 2.24 1.06 1.045 0.005

0.893 44. 8 2.07 1.07 0.91 1 0.007
0.815 47.7 1.90 1.05 0.849 0.009
0.738 49.9 1.76 1.05 0.738 0. 010
0.660 51.5 1.61 1.11 0.687 0.010
0.582 52.6 1.51 1.02 0.564 0.010
0.505 53. 7 1.40 0.97 0.480 0.009
0.427 55.4 1 . 31 0.94 0.432 0.009
0.349 56. 3 1.20 0.87 0.351 0. 008
0.272 57. 1 1.09 0.88 0.278 0.006
0.194 57.5 1.01 0.80 0. 1 95 0.004
0.116 57.7 0.92 0.78 0. 108 0.003
0.039 57.8 0.86 0.76 0.035 0.001

Experiment V 50

Re =6.87 • 105 £ = 0.0123 u. = 94. 5 m/s <S>f = -475-8 N/m2

b

r/R u m/ s Vu12/u* V r 12/u* T*

c
0.970 78.2 2.27 0.95 0.979 0.124
0.893 88.8 1.99 1.01 0.905 0. 1 88
0.815 94.3 1.83 1.17 0.757 0.224
0.738 97.9 1.68 1.03 0.734 0.244
0.660 101.3 1.57 0.97 0.630 0.251
0.582 103.7 1.45 0.96 0.580 0.248
0.505 106.3 1.34 0.92 0.498 0.236
0.427 108.2 1.24 0.85 0.415 0.215
0.349 109.8 1.13 0.81 0. 358 0. 187
0. 272 111.4 1.03 0.77 0.260 0.154
0. 194 112.6 0.94 0.74 0. 1 91 0.114
0.116 113.6 0.87 0.72 0.114 0.071
0.039 114.1 0.82 0.71 0.036 0.024
 87

E xperiment_V_5 J_
Re = 6.60 • TO5 f = 0.0126 = 72.0 m/s (^)f = -323.0 N/m2

Vuf2/u* Vr'2/u*
r/R u m/s T*
c
0. 970 59.5 2.17 1.59 1.088 0.046
0.893 68.0 1 . 93 1.52 1.024 0.071
0.815 71.7 1.75 1.48 0.896 0.085
0.738 74. 8 1.64 1.36 0.831 0. 092
0.660 77.0 1.50 1.38 0.740 0.095
0.582 78.8 1.41 1.23 0.638 0.094
0.505 80.6 1.30 1.17 0.552 0.090
0.427 82.6 1.21 1.10 0.451 0.083
0.349 83.7 1.10 1.08 0.400 0.072
0.272 84.8 1.00 1.00 0. 303 0.058
0. 1 94 85.6 0.92 0. 94 0.144 0.043
0. 116 86.4 0.83 0.91 0.140 0.027
0.039 86.6 0.79 0.89 0.038 0.009

Experiment V 52
Re = 9. 35 • 105 f = 0.0119 u. = 1 04. 0 m/s 9 N/m2
b = -620’

r/R u m/s Vu'2/u* Vr’2/u* T*

c
0. 970 86.4 2.21 0.76 0. 895 0.310
0.893 98.0 1.96 0.90 0.861 0.434
0.815 103. 9 1.80 0. 92 0.804 0.501
0.738 108.2 1.67 0.89 0.725 0.535
0.660 111.1 1.52 0.89 0.627 0.545
0. 582 114.1 1.43 0.83 0.554 0.535
0.505 116.8 1.32 0.81 ' 0. 485 0.506
0.427 1 18.8 1.21 0.77 0.421 0.460
0. 349 120.7 1.11 0.73 0. 343 0.400
0.272 122. 5 1.01 0.68 0.263 0. 328
0.194 123. 9 0.91 0.64 0.186 0.244
0.116 125.0 0.83 0.63 0.113 0. 151
0.039 125.4 0.79 0.63 0. 036 0.051
 Principal arrangement of the
experimental equipment Fig. 1

(T) Air from blower

Flow rectifier

Foster flow meter

Cooler

Wooden nozzle

Test channel 128.8

Probe support

Hot wire -or stagnation

pressure probe Static pressure
tappings each
Static pressure tappings 1000 mm.

Honeycomb

Flow rectifier
positioned before
the inlet nozzle
during the second
test series
 Principal arrangement of the experiment
for determination of the direction Fig. j
sensitivity coefficient.

1. Hot wire probe 5. Circular channel

2. Revolving probe holder 6. Net

3. Protractor 7 Honeycomb

4. Nozzle 8. Probe support

7
 Axial variation of the
friction factor. Fig. 4

a
X

a> o
or CD

[>•

o ©
o
<D in

O •©
-O

o ©

,o
co

,o
CM
© o

-O
57.2

0.954

T
in o
o p p
f

O without rectifier before the inlet nozzle.

# with rectifier before the inlet nozzle.

Friction factors measured

for x/D =57.2
2.0 log Re Vf~- 0.8

Fig. 5
 Comparison of velocity distributions
in circular channels.
Fig. 6

u+
0

3
■/

6 f
n This
O without rectifier before the
inlet nozzle. 18 experiments.
ie investigation with rectifier before the inlet
7 nozzle. 13 experiments

V Measurements by Nikuradse
8 [20]. 16 experiments.

0 Measurements by Coantic [4]

9
J 95% confidence interval

10

11

12

13 “1------------1------------ 1------------ 1------------1------------1------------1------------ 1------------ 1------------ r

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
r/R
 Distribution of the eddy diffusivity;
calculated from the velocity distribution. Fig. 7

without rectifier before the inlet nozzle

This investigation
with rectifier before the inlet nozzle

Calculated f+ from results of Nunner

Ru
by Hinze [l0] from results of Laufer

0.10 -
O

—O—
x /<"
+ /x o * X
/z 6 +' +
\
/o
+
V+
\
\
0.05- / \
\
A
Curve representing the results \
/ of Nikuradse [20] \
+
f \

Q
A

0.0 --------- ,------------ ,------------1------------ 1-----------1——i------------- 1------------ 1 i ------ 1

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
r/R
 Residual sum of squares for different
values of c bv correlation of the
experimental data with Collis* law. Fig. 8
Res. sum of squares
O
 CoilisJ law with a constant value
of c = 0.35 and Vn chosen for best
fit with experimental data. Fig. 9

v2-
140

130

120

110

100

90­

80

70

Calculated Measured
Experiment V0 V0
60­ 680206-1 O 3.47 5.00
680206-2 V 3.04 5.47
680405-2 0 4.26 5.49

50- T ~T~ T T~ T I
4.0 5.0 6.0
 The exponent c in Collis' law
determined by measurements in the Fig. 10
test channel and in the jet after a nozzle.
CM

kg/sm
D
282
286
Channel
Nozzle

I--------------1--------------1--------------1--------------1--------------1--------------r o
CO Lf) pr)
O O O O
Influence of pu on the exponent
c in Collis' law.
31
IQ
c
Line 1 correlation of measurements at

2 correlation of measurements at

measurements at

temperature on the exponent

Influence of the sta g nation
c in Collis'law.
T ~ 320 °K

Fig. 12
kg/sm 2
 k2
• This investigation
0.08­ 4 Champagne [2J
0 Rasmussen [25]
0.07- ▼ Webster [34] mean valve, 213 < l/d <1456

Comparison of measurements of
the direction sensitivity coefficient-
V ------ n----- l/d^220
0.06 J 95% confidence interval

0.05
c= 0.48
6
0.04

0.03

0.02-

0.01

31
0.0 T t--------r (5

oo
0.01
— -
 Measured distributions of the axial
turbulence. Comparison with earlier Fig. 14
results.
Evaluation by method IT

2.6­

2.4­

2.2- fv

2.0­

1.8 -

V
1.6­ Q

1.4­

1.2­ O
V
This investigation: o
1.0­ Measurements in channel
Mean value
13 exp.
with inlet rectifier f 95% confidence
1 interval.
0.8­
Measurements in channel
O Mean value. 8exp.
without inlet rectifier
0.6-
Earlier investigations;
0.4- V Coantic [4]
0 Patel [23]
0.2-i A Reichardt [26]
Laufer [15]
0.0- --------- -—------,--------- ,--------- T T i------------ 1
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
r/R
 Measured distributions of the radial
turbulence. Comparison with earlier pig. 15
results.
Evaluation by method H

*
u

This investigation:
0.5
Measurements in channel • Mean value. 7exp.
with inlet rectifier J 95% confidence interval.

Earlier investigations:
Explanations are given in fig. 14.

o.o H------1------1------1------1------1------1------1------1------1------1
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
r/R
Intensity of axial turbulence at the
centre of a circular channel- Fig. 16
Comparison of different investigations.

O <3>
 Comparison of experimental shear
stresses, evaluated bv different
Fig. 17
methods, and theoretical values.
T*
r/R Symbol Evaluation method

r/R
1.10 1 Symbol Evaluation method

A 1
V
1.05- V nr

A
A ^ v
1.00 i------------1---------- —|— — r~ —l— T-----------1
0.8 0.6 0.4 0.2 0.0
A A V
V V V r/R
A V A
0.95 A A V V
A
A A
V
0.90-
A

0. 85-
V
A

0.80-
 Experimental and theoretical shear
stress distributions Fig 18

Evaluation by method Df
# Mean value
Measurements in channel 7 exp.
with inlet rectifier T 95% conf.
1 interval

Measurements in channel O Mean value

without inlet rectifier 8 exp.
t- i
R_ dp
2 dx

Theoretical distribution

0.6 0.5
 Correlation with the Reynolds number of
the ratio of the measured and theoreti
Fig. 19
cal shear stress.

10 Re
CO
LP
o I

c
4>
"o _ O
a>
o
o
c
o
a - O)
ai
i—
i—
3
- 00

- to

— LO

O CO OOO CD

O OCKD ooo o
CO

o oo caxro

- CM

* cc
"T ~T~ o
o Lf) o
O o
 Observed standard deviations for
the dimensionless turbulence Fig. 20
intensities and the dimensionless
shear stress.
VP/U *
V Vftu*
W/u

crx *

0.1 -

G
--------- 1 —i----------- 1-------------1 —i-----------1------------- 1----------- 1— “I------------1
1.0 0.9 0.8 0.7 0.6 0.5 Q4 0.3 0.2 0.1 0.0
r/R
 Shear stress distributions
measured by Coantic [4]. Fig. 21

O Result of measurement

r/rw

Theoretical distribution
 Experimental and theoretical shear
stress distributions Fig. 22
Measurements of Sand born [29j

O Mean value of measurements

95% confidence interval

V Standard deviation from mean value

Number of observations;
 Velocity distributions in the jet
immediately after the nozzle. Fig. 23
uc is the velocity at r = 0

u/ur
1.02-,

1.01 -

A®®
o o ° o o
1.00

0.99-

0.98

Velocity distribution at an average r mm

velocity of 71.5 m/s

U/Uc

1.01 -

Velocity distribution at an average r/mm

velocity of 96.1 m/s
 Verification of the validity
of eg (3-4). Fig. 24

)
(V -V 0
CN
in
E
cn
-SC

cn
cn
ii
3
Measurements at
O
Line 1
 LIST OF PUBLISHED AE-REPORTS
1-280. (See the back cover earlier reports.)
261. On the attenuation of neutrons and photons in a duct filled with a helical
plug. By E. Aalto and A. Krell. 1966. 24 p. Sw. cr. 8:-.
262. Design and analysis of the power control system of the fast zero energy
reactor FR-0, By N. J. H. Schuch. 1966. 70 p. Sw. cr. 8:-.
263. Possible deformed states in 1tsln and 11,fn. By A. Backlin, B. Fogelberg and
S. G. Malmskog. 1967. 39 p. Sw. cr. 10:—.
264. Decay of the 16.3 min. 1e,Ta isomer. By M. Hdjeberg and S. G. Malmskog.
1967. 13 p. Sw. cr. 10:-.
265. Decay properties of ,4TNd. By A. Backlin and S. G. Malmskog. 1967. 15 p.
Sw. cr. 10:—.
266. The half life of the 53 keV level in 19fPt. By S. G. Malmskog. 1967. 10 p.
Sw. cr. 10:—.
267. Burn-up determination by high resolution gamma spectrometry: Axial and
diametral scanning experiments. By R. S. Forsyth, W. H. Blackladder and
N. Ronqvist, 1967. 18 p. Sw. cr. 10:-.
268. On the properties of the sny2 ---- d3y2 transition in mAu. By A. Backlin
and S. G. Malmskog. 1967. 23 p. Sw. cr. 10:—.
269. Experimental equipment for physics studies in the Agesta reactor. By G.
Bernander, P. E. Blomberg and P.-O. Dubois. 1967. 35 p. Sw. cr. 10:-.
270. An optical model study of neutrons elastically scattered by iron, nickel,
cobalt, copper, and indium in the energy region 1.5 to 7.0 MeV. By B.
Holmqvist and T. Wiedllng. 1967. 20 p. Sw. cr. 10:—.
271. Improvement of reactor fuel element heat transfer by surface roughness.
By B. Kjellstrom and A. E. Larsson. 1967. 94 p. Sw. cr. 10:-.
272. Burn-up' determination by high resolution gamma spectrometry: Fission pro­
duct migration studies. By R. S. Forsyth, W. H. Blackadder and N. Ron­
qvist. 1967. 19 p. Sw. cr. 10:—.
273. Monoenergetic critical parameters and decay constants for small spheres
and thin slabs. By I. Carlvik. 1967. 24 p. Sw. cr. 10:—.
274. Scattering of neutrons by an enharmonic crystal. By T. Hogberg, L. Bohlin
and I. Ebbsjo. 1967. 38 p. Sw. cr. 10:—.
275. ThelAKl=1, E1 transitions in odd-A isotopes of Tb and Eu. By S. G. Malm­
skog, A. Marelius and S. Wahlborn. 1967. 24 p. Sw. cr. 10:—.
276. A burnout correlation for flow of boiling water in vertical rod bundles. By
Kurt M. Becker. 1967. 102 p. Sw. cr. 10:—.
277. Epithermal and thermal spectrum indices in heavy water lattices. By E. K.
Sokolowski and A. Jonsson. 1967. 44 p. Sw. cr. 10:—.
278. On the transitions in odd mass Pm nuclei. By A, Backlin and
S. G. Malmskog. 1967. 14 p. Sw. cr. 10:-.
279. Calculations of neutron flux distributions by means of integral transport
methods. By I. Carlvik. 1967. 94 p. Sw. cr. 10:-.
280. On the magnetic properties of the K=1 rotational band in "'Re. By S. G.
Malmskog and M. Hdjeberg. 1967. 18 p. Sw. cr. 10:—.
281. Collision probabilities for finite cylinders and cuboids. By I. Carlvik. 1967.
28 p. Sw. cr. 10:-.
282. Polarized elastic fast-neutron scattering of "C in the lower MeV-range.
I. Experimental part. By O. Aspelund. 1967. 50 p. Sw. cr. 10:—.
283. Progress report 1966. Nuclear chemistry. 1967. 26 p. Sw. cr. 10:-.
284. Finite-geometry and polarized multiple-scattering corrections of experi­
mental fast-neutron polarization data by means of Monte Carlo methods.
By O. Aspelund and B. Gustafsson. 1967. 60 p. Sw. cr. 10:-.
285. Power disturbances close to hydrodynamic instability in natural circulation
two-phase flow. By R. P. Mathisen and O. Eklind. 1967. 34 p. Sw. cr. 10:—.
286. Calculation of steam volume fraction in subcooled boiling. By S. Z. Rou-
hani. 1967. 26 p. Sw. cr. 10:—.
287. Absolute El, AK = 0 transition rates in odd-mass Pm and Eu-isotopes.
By S. G. Malmskog. 1967. 33 p. Sw. cr. 10:—.
288. Irradiation effects in Fortiweld steel containing different boron isotopes.
By M. Grounes. 1967. 21 p. Sw. cr. 10:-.
289. Measurements of the reactivity properties of the Agesta nuclear power
reactor at zero power. By G. Bernander. 1967. 43 p. Sw. cr. 10:—.
290. Determination of mercury in aqueous samples by means of neutron activa­
tion analysis with an account of flux disturbances. By D. Brune and K. Jir-
low. 1967. 15 p. Sw. cr. 10:—.
291. Separtaion of 81Cr by means of the Szilard-Chalmers effect from potassium
chromate irradiated at low temperature. By D. Brune. 1967. 15 p. Sw.
cr. 10:-.
292. Total and differential efficiencies for a circular detector viewing a circu­
lar radiator of finite thickness. By A. Lauber and B. Tollander. 1967. 45 p.
Sw. cr. 10:-.
293. Absolute M1 and E2 transition probabilities in mU. By S. G. Malmskog and
M. Hdjeberg. 1967. 37 p. Sw. cr. 10:-.
294. Cerenkov detectors for fission product monitoring in reactor coolant water.
By O. Strindehag. 1967. 56 p. Sw. cr. 10:—.
295. RPC calculations for K-forbidden transitions in 14}W. Evidence for large
inertial parameter connected with high-lying rotational bands. By S. G.
Malmskog and S. Wahlborn. 1967. 25 p. Sw. cr. 10:—.
296. An investigation of trace elements in marine and lacustrine deposits by
means of a neutron activation method. By O. Landstrom, K. Samsahl and
C-G. Wenner. 1967. 40 p. Sw. cr. 10:—.
297. Natural circulation with boiling. By R. P. Mathisen. 1967. 58 p. Sw. cr. 10:—.
298. Irradiation effects at 160—240°C in some Swedish pressure vessel steels.
By M. Grounes, H. P. Myers and N-E. Hannerz. 1967. 36 p. Sw. cr. 10:-.
299. The measurement of epithermai-to-thermat U-238 neutron capture rate (p2s)
in Agesta power reactor fuel. By G. Bernander, 1967. 42 p. Sw. cr. 10:-.
300. Levels and transition rates in 1"Au.
2 3 4 By
5 6 S.
7 G. Malmskog, A. Backlin and B.
Fogelberg. 1967. 48 p. Sw, cr. 10:—.
301. The present status of the half-life measuring equipment and technique at
Studsvik. By S. G. Malmskog. 1967. 26 p. Sw. cr. 10:—.
302. Determination of oxygen in aluminum by means of 14 MeV neutrons with
an account of flux attenuation in the sample. By D. Brune and K. Jirlow.
1967. 16 p. Sw. cr. 10:—.
303. Neutron elastic scattering cross sections of the elements Ni, Co, and Cu
between 1.5 and 8.0 mev. By B. Holmqvist and T. Wiedling. 1967. 17 p.
Sw. cr. 10:-. r 26 p. Sw. cr. 10;-.
304. A study of the energy dependence of the Th232 capture cross section in
the energy region 0.1 to 3.4 eV. By G. Lundgren. 1967. 25 p. Sw. cr. 10:-.
EOS-tryckerierna, Stockholm 1968
305. Studies of the reactivity effect of polythene in the fast reactor FRO. By L.
1. Tir6n and R. HSkansson. 1967. 25 p. Sw. cr. 10:—.
306. Final report on IFA-10, the first Swedish instrumented fuel assembly irra­
diated in HBWR, Norway. By J-A. Gyllander. 1967. 35 p Sw. cr. 10:-.
307. Solution of large systems of linear equations with quadratic or non-qua­
dratic matrices and deconvolution of spectra. By K. Nygaard. 1967. 15 p.
Sw. cr. 10:-.
308. Irradiation of superheater test fuel elements in the steam loop of the R2
reactor. By F. Ravndal. 1967. 94 p. Sw. cr. 10--.
309. Measurement of the decay of thermal neutrons in water poisoned with the
non-1/v neutron absorber cadmium. By. L. G. Larsson and E. Moller. 1957.
20 p. Sw. cr. 10:—.
310. Calculated absolute detection efficiencies of cylindrical Nal (Tl) scintill­
ation crystals for aqueous spherical sources. By. O Strindehag and B.
Tollander. 1968. 18 p. Sw. cr. 10:—.
311. Spectroscopic study of recombination in the early afterglow of a helium
plasma. By J. Stevefelt. 1968. 49 p. Sw. cr. 10:—.
312. Report on the personnel dosimetry at AB Atomenergi during 1966. By J.
Carlsson and T. Wahlberg. 1968. 10 p. Sw. cr. 10:—.
313. The electron temperature of a partially ionized gas in an electric field.
By F. Robben. 1968. 16 p. Sw. cr. 10:—.
314. Activation Doppler measurements on U238 and U235 in some fast reactor
spectra. By L. I. Then and I. Gustafsson. 1968. 40 p. Sw. cr. 10:-.
315. Transient temperature distribution in a reactor core with cylindrical fuel
rods and compressible coolant. By H. Vollmer. 1968. 38 p. Sw. cr, 10:—.
316. Linear dynamics model for steam cooled fast power reactors. By H. Voll­
mer. 1968. 40 p. Sw. cr. 10:—.
317. A low level radioactivity monitor for aqueous waste. By E. J. M. Quirk.
1968. 35 p. Sw. cr. 10:—.
318. A study of the temperature distribution in UOa reactor fuel elements. By
I. Devoid. 1968. 82 p. Sw. cr. 10:—.
319. An on-line water monitor for low level ^-radioactivity measurements. By
E. J. M. Quirk. 1968. 26 p. Sw. cr. 10:—.
320. Special cryostats for lithium compensated germanium detectors. By A.
Lauber, B. Malmsten and B. Rosencrantz. 1968. 14 p. Sw. cr. 10:-.
321. Stability of a steam cooled fast power reactor, its transients due to mode­
rate perturbations and accidents. By H. Vollmer. 1968. 36 p. Sw. cr. 10:—.
322. Progress report 1967. Nuclear chemistry. 1968. 30 p. Sw. cr. 10:-.
323. Noise in the measurement of light with photomultipliers. By F. Robben.
1968. 74 p. Sw. cr. 10:—.
324. Theoretical investigation of an electrogasdynamic generator. By S. Palm-
gren. 1968. 36 p. Sw. cr. 10:-.
325. Some comparisons of measured and predicted primary radiation levels in
the Agesta power plant. By E, Aalto, R Sandlin and A. Krell. 1968. 44 p.
Sw. cr. 10:-.
326. An investigation of an irradiated fuel pin by measurement of the production
of fast neutrons in a thermal column and by pile oscillation technique.
By Veine Gustavsson. 1968. 24 p. Sw. cr. 10:—.
327. Phytoplankton from Tvaren, a bay of the Baltic, 1961—1963. By Torbjorn
Willen. 1958. 76 p. Sw. 10:-.
328. Electronic contributions to the phonon damping in metals. By Rune Jonson.
1968. 38 p. Sw. cr. 10:-.
329. Calculation of resonance interaction effects using a rational approximation
to the symmetric resonance line shape function. By H. Haggblom. 1968.
48 p. Sw. cr. 10:—.
330. Studies of the effect of heavy water in the fast reactor FRO. By L. I. Then,
R. HSkansson and B. Karmhag. 1968. 26 p. Sw. cr. 10:—.
331. A comparison of theoretical and experimental values of the activation Dop­
pler effect in some fast reactor spectra. By H. Haggblom and L. I. Then.
1968. 28 p. Sw. cr. 10:—.
332. Aspects of low temperature irradiation in neutron activation analysis. By
D. Brune. 1968. 12 p. Sw. cr. 10:—.
333. Application of a betatron in photonuclear activation analysis. By D. Brune,
S. Mattsson and K. Lid&n. 1968. 13 p. Sw. cr. 10:—.
334. Computation of resonance-screened cross section by the Dorix-Speng
system. By H. Haggblom. 1968. 34 p. Sw. cr. 10:-.
335. Solution of large systems of linear equations in the presence of errors.
A constructive criticism of the least squares method. By K. Nygaard. 1968.
28 p. Sw. cr. 10:-.
336. Calculation of void volume fraction in the subcooled and quality boiling
regions. By S. Z. Rouhani and E. Axelsson. 1968. 26 p. Sw. cr. 10:—.
337. Neutron elastic scattering cross sections of iron and zinc in the energy
region 2.5 to 8.1 MeV. By B. Holmqvist, S. G. Johansson, A. Kiss, G. Lo-
din and T. Wiedling. 1968. 30 p. Sw. cr. 10:-.
338. Calibration experiments with a DISA hot-wire anemometer. By B. Kjell­
strom and S. Hedberg. 1968. 112 p. Sw. cr. 10:-.
List of published AES-reports (In Swedish)
1. Analysis be means of gamma spectrometry. By D. Brune. 1961. 10 p. Sw
cr. 6:—.
2. Irradiation changes and neutron atmosphere in reactor pressure vessels
some points of view. By M. Grounes. 1952. 33 p. Sw. cr. 6:-.
3. Study of the elongation limit in mild steel. By G. Ostberg and R. Alter
mo. 1963. 17 p. Sw. cr. 6:—.
4. Technical purchasing in the reactor field. By Erik Jonson. 1963. 64 p
Sw. cr. 8:-.
5. Agesta heat generating station. Summary of technical data, descriptions
etc. for the reactor. By B. Liiliehook. 1964. 336 p. Sw. cr. 15:-.
6. Atom Day 1965. Summary of lectures and discussions. By S. Sandstrom
1966. 321 p. Sw. cr. 15:—.
7. Building materials containing radium considered from the radiation pro
lection point of view. By Stig O. W, Bergstrom and Tor Wahlberq. 1967
Additional copies available from the library of AB Atomenergi, Pack, S-611 01
Nykoping, Sweden.