D E C .

190% 1317

ON T'IIE RATE OF HEAT TRANSMISSION

BETWEEN FLUIDS AND METAL SURFACES.
-
BY MR.H. P. JORDAN, 3f.S~.(Tech.), OF THE MUNICIPAL
SCHOOL
OF
TECHNOLOGY,
R~ANCHBBTER.

[Seleeted for PubZication.]

PART I.
INTRODUCTION OF EXPERIMENTAT,
AND T~SCEIPTION PLANT.
The laws governing the Transfer of Heat, either from a heated
fluid into a colder metal surface or from a warm metal surface
into a colder fluid, are of importance to engineers, and of late a
considerableamount of interest has been taken in this subject.
The law until' recently generally accepted is that given by
Peclet and Rankine, namely, that the rate of transmission of heat
is proportional to the square of the difference of the temperatures of
the fluid and metal. This law takes no account of the rate of flow
of the fluid over the metal.
The effect of the speed of circulation was first pointed out by
Professor Osborne Reynolds in a Paper read by him in 1871 before
the Manchester Literqxy and Philosophical Society, in which he
+
gave a s the law of heat transmission h = A B p v,
4 ~ 2
131 8 IIEA’I‘ THANSJYIISSION BETmEEN FLUIDS AND ME’L‘ALS. DEC. 1909.

where h represents the units of heat transmitted per degree

difference of temperature of fluid and metal ;
21 is the velocity of flow of the fluid over the metal

surfacc ;
p is the deneity of the fluid ;
A and B are constants.
This law inay be writtcn h = A + I3
mhcre w is the weight of fluid flowing i n unit time and n is the
area of the channel,
or ‘0 is the rate of flow of the fluid per unit area of the
CL
channel.
This law was confirmed by Dr. T. E. Stanton, for the case of heat
flow from metal to water i n his Paper on “ T h e Passage of Heat
between Metal Surfaces and Liquids in Contact with Them,”
communicated to the Royal Society i n 1897. Professor Weighton’s
experiments on surface condensers have also had the effect of
drawing the attention of engineers to the increased rates of heat
transmission which are to be obtained with high speeds of
circulation.
Experiments made by Dr. J. T. Nicolson during 1905 and 1906
with superheated steam, air, compressed air, and water flowing
at high velocities inside and outside of long pipes gave a
preliminary verification of’ the same law for these two cases. The
author was therefore glad t o accept the suggestion that a further
investigation into this problem should be made the subject matter
for research during his tenure of the Schuster Research Scholarship,
and the following Paper contains the main results of that
investigation.
Only a portion of the subject of heat transfer from fluid t o
metal is dealt with in the following pages. I n the experiments here
described the heat-carrier was compressed air and the heat-receiver
was water ; there being no change of state of the heat-receiving
liquid. I f either the liquid is being evaporated or the gas or
vapour condensed, the law of heat transfer may differ considerably
from that obtained when there is no such change of state.
DEC.1909. HEAT TRANSMISSION BETWEEN FLUIDS AND m r I L s . 1319

All the following experiments were carried out i n the

engineering laboratories of the Manchester School of Technology.
The author wishes here to express his thanks to Dr. J. T. Nicolson
for his advice given during the preparation of the plant and
throughout the whole course of the research, to Mr. J. T. Hodgson,
mechanical engineer of the School of Technology, for his assistance
in preparing the plant, and to Mr. L. Saxon, fourth year student
in the mechanical engineering department, for assistance in taking
many of the observations.

Description of Plant.-The plant used was originally constructed

to the designs of Dr. Nicolson as a part of the equipment of
the School of Technology in 1901. It was intended for the
experimental study of the laws of condensation of steam, but was
modified so as to be available for the preseii t research.
A general arrangement drawing is shown in Fig. 1(page 1320).
The general scheme is briefly as follows :-Air was supplied from an
air-compressor to the receiver R ;the pressure in the receiver being
controlled by the blow-off valve V. From the receiver the air was
passed through a coil in the furnace, after leaving which it passes
to the air-cooler or calorimeter C. The temperature of the air at
inlet to the calorimeter was measured by a platinum thermometer
and Whipple indicator ; the outlet temperature being taken by a
mercury thermometer. The air was finally discharged through a
sharp-edge orifice in the air-discharge box A. The temperature
and pressure of the air within the box were respectively taken by
a mercury thermometer and mercury U-tube.
Water was supplied to the calorimeter from the high-level
tank used to aupply the hydraulic laboratory ; a heating coil was
provided so that the water could be supplied a t any temperature
u p t o 200' F. The temperature of the water was taken a t inlet
to and outlet from the calorimeter by mercury thermometers,
and the quantity of water determined by weighing. The weighing
machine used was sensitive to within 2 oz. and weighed up to
500 lb. Before being weighed the water was passed t,lirough a
cooling coil. The temperature of the ah-cooling surface of the
1320
DEC.1909. HEAT TRANSMISSIOB BETWEEN FLUIDS AND METALS. 1321

calorimeter was determined by thermo-electric couples. The

greatest care was taken to lag all parts of the apparatus and pipes
to prevent any gain or loss of heat from the surroundings.

Tke Air-Heater.-The arrangement for heating the air consisted

of a coil of ;-inch hydraulic piping immersed in the salt bath of a
Brayshaw patent hardening furnace. It was found that the heat
could be very easily controlled, and that the temperature of the
air leaving the furnace was very constant.

The Air-Cooler or Calorimeter.-This consisted essentially, sea

Fig. 2 (page 1322), of a copper pipe through which the air was
passed, surrounded by a cast-iron casing, the space between this
latter and the copper pipe forming the water-jacket. The greatest
care was taken in the design to eliminate all possible sources of
error due to heat conduction. The copper pipe was screwed to
steel ends which were made very thin to avoid end effect due to
conduction of heat along the pipe into and out of the apparatus.
Also, as may be seen from Fig. 2, there was no contact between
the pipe and the casing, the joints a t each end being made by cup
leathers ; these diminished conduction of heat from the hot ends of
the pipe through the casing. I n order to test what effect, if any,
change of dimensions or section of channel would have on the rate
of heat transmission, five series of experiments were made with
different air-passages. Two series were made with annular air-
passages and three with open pipes. The design of the calorimeter
allowed this alteration of the air passage to be made in a very
simple manner.
The changes of air passage for the open-pipe series were made
by unscrewing the copper pipe from the steel ends and fitting to
the same ends a pipe of different diameter, the casing being bushed
up to whatever extent was necessary to give the required water-
space. The experiments with annular pipes were made with
the largest copper-pipe partially stopped by a wrought-iron core
held centrally in the pipe by means of a wire spiral. Two cores
of different diameters thus gave two series of oxperiments with
1322 HEAT TRANSMISSlON BETWEEN FLUIDS AND METALS. DLC.1909.

Bra. 2. FIG. 3.
Section thozgcgli Calorimeter. Copper-Constadan Jtinction.
Dec. 1909. HEAT TRANSMISSION BETWEEN FLUIDS AND METALS. 1323

different annular air-passages. These cores consisted of a length

o f wrought-iron piping welded u p at the ends and turned down
to leave the metal as thin as possible.

Meaaurement of Metal Temperatures-Thermo-Electric Couples.-

As previously mentioned, the temperature of the cooling surface of
the calorimeter was measured by means of thermo-electric couples.
Four junctions were used a t different points along the length of
the air-cooling pipe, Fig. 2, the metals used for the junctions
being constantan and copper. Some little trouble was a t first
experienced in making these junctions in a satisfactory and reliable
manner, hut, as a result of some preliminary trials, these difficulties
were overcome and the junctions made in the manner indicated in
Fig. 3.
First, a steel, or in some experiments a gun-metal sheathing S
was screwed up tight, till its knife-edge end had made a water-
tight joint with the internal copper air-pipe. Into this sheathing
was screwed a long copper plug P, with two longitudinal slots i n
which were laid the insulated copper and constantan wires W. To
tlic copper plug was keyed n hollow vulcanite end V, so that copper
and vulcanite turned together. The end of the vulcanite was
solid and turned down to fit into a small recess of 4 inch diameter
drilled to a depth in the metal of the copper pipe equal to half its
thickness. Two small pin-holes were drilled through the small
end of the vulcanite. The insulated copper and constantan wires
were laid in the slotti in the copper plug, and then brought inside
the vulcanite end. For a short distance (about. 4 inch) a t their
ends the wires were strippcd of their insulation, and this length
was pushed through the pin-holes in the small end of the vulcanite,
thus leaving the two ends standing out from about 2% to & inch
and about & inch apart. The wires were then fixed in the slots
by shellac. To make the junction, the copper plug was carefully
Rcrewed up until it was felt that the end was fitted well up against
the internal pipe. Great care was taken that the small vulcanite
end should fit into the recess in the metal of the pipe; this was
done by testing, with a gauge and cutter, the exact size of the
1321 HEAT TRANSMISSION BETWEEN FLUIDS AND METALS. DEC.1909.

vulcanite end before fitting in the junctions. The cold junction

was kept a t a constant known temperature b y means of an oil
bath.
The direct deflection method was used to measure the
electromotive force produced by the difference of the temperatures
of the hot and cold junctions, the galvanometer being of the
D'Arsonval mirror type supplied by the Cambridge Scientific
Instrument Company. By means of mercury contacts, the
galvaiiometer could be put into circuit with the different cokples
in rapid succession, and thus the temperatures of the metal
measured a t the four different points along the length of the pipe.

Calibration of Thewno-Couples, etc.-The deflections given on

the galvanometer scale, due to differences of temperature of
the hot and cold junctions, were carefully calibrated by keeping
one junction i n a cold oil-bath of confitant temperature and
heating the other junction in a second oil-bath by means of a
Bunsen burner ; the temperatures of both oil-baths being taken
by standard mercury thermometers.
From these results a curve was drawn connecting temperature
difference^ of the junctions with the deflections on the galvanometer
wale. By the aid of these curves the temperatures of the metal
were inferred from the galvanometer deflections observed during
the course of the experiments. Great care was taken over these
calibrations, which were repeated several times during the course
of the experiments without any appreciable difference being found.
It was essential, however, that the galvanometer should not i n
any way be disturbed either during the experiments or in the
intervals between ; this condition was fdfilled hy fixing it i n a
recess in t h e wall, thus ensuring a solid foundation and enclosing
it in a fixed wooden case, the front of which was removed during
the experiments. The stand for the galvanometer scale was
screwed to the bench tvp. Both galvanometer and scale were
thus, during the course of the experiments, permanently fixed.
As a check upon t h e external calibration of the junctions, the
water connections to the calorimeter, Fig. 1 (page 1320), were so
z)iic. 1909. HEAT TRANSMISSION BETWEEN FLUIDS AED METALS. 1325

made that the air outlet could be put i n connection with the water
inlet pipe, and hot water a t a known temperature passed through
both air and water spaces. By this means the whole calorimeter
was kept a t a constant known temperature, and the deflections on
the galvanometer given by the different thermo-junctions were
observed and compared with the calibration curves. In no
case was any difference found either between the deflections given
by the different couples or between the observed deflections
and those which would have been expected from the calibration
curves. All the thermometers used in these experiments were
tested at boiling and freezing points and the necessary corrections
made.
DESCRIPTION
OF EXPERIMENTS.

As before mentioned, five series of experiments distinguished

by the letters B, C, D, E and F were carried out, each series
with a different area of air-passage. The following Table gives
information as to the form and dimensions of the various air
and water spaces.
TABLE 1.
Heating Surfaces and Areas of Channels for Experiments
Series B to F.

Air-Passage. Water-Pasnage.

3eriea

3q. f t
section.
! I In:hes,

Annular 0.00725 0.171 1-69 Aiiuular 0.0059 0.097i

,, 0'01103 0.26 1.69 .. 0*00311 0.0972

Circular 0.0211 0.492 1.69 ,. 0.00311 0.0972

,, 0.001397 0.126 0.43 ,, 0.00264 0.143

,, 0.00834 0.309 1'06 ., 0.003535 0.1302

q.=
Area o f passage j q u a r e inches.
Periiiieter of c o o x i auifacc in inctiea.
1326 HEAT TRANSlISSION BETWEEN FLUIDS AND METALS. DEC.1909-

The following measurements were taken during the course of

each experiment :-
(1) The temperatures of the air and water a t entry to and
outlet from the calorimeter.
( 2 ) The pressure of the air at entry to and outlet from the
calorimeter.
(3) The pressure and temperature of tho air in the air
discharge box.
(4) The deflections of the galvanometer registering the
electromotive forces given by the thermo-couples, also the
temperature of the constant temperature oil-bath.
( 5 ) The weight of the cooling water.
Care was taken before starting an experiment that a steady
state had been arrived at, regular observations being made of the
temperatures previous to the conimencement of the experiment.
When it was seen that a steady state had been finally attained, the
experiment was started at a given signal by switching the water
over from the waste outlet to the weighing tank.
During the course of the experiment, observations of quantities 1,
3, and 4 were taken every five minutes, quantities 2 were observed
every ten minutes, and the quantity of cooling water was weighed
a t the conclusion of the experiment. The usual duration of each
experiment was thirty minutes, but some experiments were taken
over a period of forty minutes, whilst others lasted only twenty
minutes. It was found that the twenty-minute experiments gave
results in every way as uniform and satisfactory a8 those which
lasted for the longer periods.

Calculation of Results.
The following is t,he LiBt of Symbols used in the following
pages :--
T = air temperature.
TI and T, = inlet and out,let temperatures of the air.
t = water temperature.
t , and 1, = inlet, and outlet temperatures of the water.
h C . 1909. HEAT TRAXSIISSION BETWEEN FLUIDS AND METALS. 1327

Om= mean metal temperature taken a t a depth equal to half

the thickness of the pipe.
0, = mean temperature of the outer surface of the pipe. -
0, = mean temperature of the inner surface of the pipe. -
OT and 8, = temperatures of the air cooling surface at the top
and bottom of the calorimeter respectively.
c = temperature gradient of the cooling surface of the

calorimeter = Brre~.
L
toi = lb. of air passing through calorimeter per second.
a, = area of air passage through calorimeter (square feet).
= lb. of air flowing per square foot of area of channel per
I”’
‘1 1
second through calorimeter.
W,= Ib. of water flowing through calorimeter per second.
az = area of water passage through calorimeter (square feet).

= lb. of water flowing per square foot per second through

a2
calorimeter.
L = length of cooling surface of calorimeter in feet.
H = B.Th.U. gained by water per second.
= B.Th.U. lost b y air per second.
= w 2 x rise of temperature of water.
h, = B.Th.U. transmitted from the air into the metal per
square foot, per second per degree Fahrenheit difference
of temperature of air and metal.
A, = B.Th.U. transmitted from the metal into the water
per square foot, per second per degree Fahrenheit
difference of temperature of water and metal.
p‘ = internal radius of the pipe.
A = thickness of the metal.
d, = internal diameter of pipe.
d, = external diameter of pipe.
S, = cooling surface for air (square feet).
S, = heating surface for water (square feet).
The first quantity to be determined is the quantity of heat
transniitted from the air, through the walls of the pipe, into the
1328 HEAT TRANSMISSIOH BETWEEN FLUIDS A N D MErALS. DEC.1909.

cooling water; this is given by the product of the weight of

cooling water and its rise of temperature. The weight of air
passing through the calorimeter was determined by equating the
heat lost by the air to the heat gained by the water, thus the
weight of air passing through the calorimeter per second will be
given by
B,T.U.~gained
~
by_water
_
per_second.
-. ~- _ _
0.238 x Fall of temperature of air.

or

A very useful check upon the quantities of air thus determined

was furnished by the pressures and temperatures observed in the
air-discharge box.
The main result to be deduced from each experiment was the
heat transmitted from the air into the cooling surface per square
foat per second per degree Fahrenheit difference of temperature ;
the data known being the inlet and outlet air temperatures, the
weight of air flowing per second, and the temperature of the
cooling surface along the tube.
This result was determined in the following manner. The
temperatures of the metal as measured by the thermo-couples
were plotted on a base representing the length of the tube and
the mean metal temperature determined.
The curves obtained were of the type illustrated in Fig. 4
(curve A-B), very nearly straight lines, and i n working out the
results, for the sake of simplicity, they were taken as straight
lines.
It follows from the description of t h e thermo-couples that the
temperatures measured by them wore those a t a depth in the
metal equal to half its thickness, and hence these temperatures
will require to be corrected for depth i n order to obtain the mean
temperatures of the air-cooling surface or of the water-heating
surfme. This correction is easily applied, thus :-Let x be the
distance of any point i n the metal froin the internal cooling
surface,
DEC.1909. HEAT TRANSMISSIONBETWEEN FLUIDS AND JIETALS. 1329

6 be the temperature of the metal,

q be the number of British thermal units transmitted per
square foot per second,
R be the thermal resistance of the material.
Then - d o =q R
rs

At any point in the metal, distant (8)from the internal surface

or e = - 2H RL log* (P + + c!
8)

When x = $3, B = 0, the temperature measured by the thermo-

couples.
1330 HEAT TRANSMISSION BETWEEN FLUIDS AKD METALS. DEC. 1909.

The mean temperature of the air-cooling surface is 1herefore a

temperature e,, Fig. 4 (pag! 1329) ; and the straight line joining
A-B will give the average slope of the temperature gradient
down the cooling eurface.
A line, therefore, parallel to the dotted line A B through a
temperature 0, a t a point half-way down the tribe will have the
same average slope :is the true curve, and will give the same mean
temperature of the cooling surface; and in working out the
results in this Paper, such a line has been taken to represent the
temperature gradient.
The problem is now a simple one :-
Consider any element of heating surface rrd, 6S,distant S froni
the top of the tube.
Then -cu,kST = h , ( T - e ) d , S X
But 01. is the temperature a t the top of the tube and 013 the
temperature a t the bottom, and by the approximation above
described the temperature gradient has been taken as a straight
line, hence
ae e, - e,
11x- L-- =

or -

Then - dX
(lT - M { T - (0, - cX)
-
1

The solution to this equation is given in the Appendix (page

1364), the result being-
DEC. 1909. HEAT TRANSNlSSION BETWEEN FLUIDS AND METALS 1331

This equation has then to be solved for M, whence h, c m be

obtained, namely-
=3
ZLE
k w,
The mean air-temperature can now be determined by dividing
the heat tiatismittetl per square foot per scourid by the coefficient
A,; this will give the iiiean difference of temperature of air and

and a8 t h e mean teinperdme of the cooling surface is known, tho

mean air-temperature can be found.
The qnantity e measures the mean of the temperatures of
the air and cooling surface, and is soine measure of the temperature
of the inore or less iiiotiotileas film of gas d h e r i n g to t h e cooling
surface.

Bent-Jocu .from JIetnZ to CVafer.--For thc case of heat transfer

ftorri metal into water, a s this is ouly :t secoiidary resiilt the
n w i ~ water-teiripe~;iture
i has bee11 taken simply at3 t h e arithmetical
iiiean ot‘ Ihe inlet a i d outlet toinperdturcs.

Thus t = t, + t,
2
= “(’2
~~
(t? - t , ‘
s, (8, - f)

PART 11.

CONSISTINGOF TABLES
OF RESULTS
OBTAIN RD FROM EXPERIMENTS.
4 n
 1332 HEAT TRAWSMISSLONBLTKEEN FLUIDS AHD METALS. DEC. 1909.

SERIES B.-TABLE 2 (continued on opposite page).

Summary of Results.

Air Passagc. Auiiular area = 0.00725 square foot. r = 0.151 inch.

Water Passage. Annular area = 0.0059 square foot r = 0.09i7 inch.
-

Air temperatures. Water temperatures.

- -
ern

- -
+5 i
f d c;
C
H
+= 2sl! E 3

H 2
0

- -. -- - - -

2:18'! 153 195 83.5 48.4 66.8 j7.6 18.4 0.165 3.035 75 * ,

260 16s 212 92 49.3 72.1 $0.7 22.8 0.140' 3'21 81

253 152 199 01 56.5 79-1 57.8 22.6 0.055 1.244 82.

267 147 200 L20 56.7 73.5 j5.l 16.8 0.051 0.855 76'

3%; 206 291 190 49 62 55.5 13 0.254 3.3 68,

363 203 277 160 43.5 63.3 53.4 19.8 0.217 4.3 72
315 202 236.4 113 52.8 96. 6 74.7 43-8 0'077 3.37 105'

354' 18.2 258.4 172.: 57.4 82.7 70.0 25.3 0.047 1.198 85-

310 189 256.8 151 61-5 92.9 77'2 31.4 0.053 1.662 94.

459 210 317'3 249 54 83 69'5 31 0.04f 1*424 87'

384 237 299 157 56 9s 77 42 0.0% 3.99 106

394 21c 292.4 184 62.8 97.1 79.9 34'2 0.05s 2.04: 99'

400. 215 298.9 185. 63'5 101,9 82.6 38.2 0.055 2.02: 104
543 23c 363.8 315 49.7 71.5 60.t 21 '5 0.09: i-16l 79.
/*1, 495 23: 348.9 260 46.5 66.5 56.5 0.17: Y.58 75,
UEC. 1909. HEBT TKAXSXISSION BET\VEEN VLUIDS A N D NETALS. 1333

(concluded from opposite p a p ) SERIES 8.--TABLE 2.

Sammary of Results.

Air Passage. Annular araa = 0'0Oi25 square foot. T = 0.171 iricll,

Water Passage. .4nnulitr area = 0.0059 square foot. T = 0.0977 inch.
-

+el 8,- t ',+t

16 I
2 2

- ___

1.79 117 3G 0.01528 28

1.9 129 47" 0.01476 2 3 . 9 .0734

0.73 115 41 0.00639 9.3 1.037t

0.50 123 3S.I 0.0041 1 8.6 1.03"

1.95 220 81 0 008S6 3.1 1.1'26

2.54 202 76 0.01263 a.5 1.112

1.99 I49 82 0'0132 3.c 1.049:

0.71 172 0.00414 8" ). 033:

0.9: 161 176'' 0.0061 8" 3.042

0.844 225 0'0036 0.035
2.36 190 203. 0.0124 O.0G::
1.21 192 196. 0'0063 0.046
1.195 193 202, 0.0062 0.042

1.28 283 222 ' 0.0045 0.05'24

2.12 27 1 213 0.0078 0.089E

4 n 2
 1334 HEAT TRANSMISSIOF BEPWEEX FLUIDS AND METALS. D r c . 1909.

SERIES C.-TABLE 3 (continued on opposite page).

Summary of Results.

Air Passage. Annnlar Area = 0.01103 square foot. r = 0'26 iilch.

Water Passage. Annular Area = 0'00311 square foot. 1' = 0.0972 inch.
-

Air temperatures. Water temperatures.

Nc -
u
'
G
.-
c

H
1 %5
0's E
2 d 42
e
H
-
G
iii

0' 2
?2

- -~ - -- - - .- -
c1 414 247 322.8 167 51 58 7
c4 393. 253 319 140 51.7 67.: 16.2
CB 358. 24.5.5298.6 113 50'2 75.1 24.9
c4 457 250.5342 !06 49.1 51.1 5
c 5 357. 253 304 104. 53.G 102.4 4S'Y
C G 49ti 286 70.2 !lo 50.8 59.8 9
3 7 470 291 373.9 70 51 71.2 20.2
3 8 364*< 265 314.5 98' 89.8 132.1 42.3
3 9 480 258 3.59 122 54.3 71.5 17.2
:10 560 287 107'9 'is 54.5 65 10.5
211 577 320 437'4 57 61-3 10.2
212 477 305 584.7 72 70.6 112.4 11.8
:13 750 355 525.1 95 52.6 63 10.4
:14 690 3i3 519.1 17 55.6 79.7 2i.I
- - -
 h C . 1909. HEAT TRANSMISSION BKTWEEN FLUIDS AND NETALS. 1335

(concluded from opposite page) SERIES C.--T ABLE 3.

Summary of Results.

Air Passage. Annular Area = 0.01103 square foot. 1' = 0.26 inch.
Water Passage. Annular Area = 0.00311 square foot. T = 0.0972 inch.

kg
3 I/
__
9. I
0
-"25 (61
T-8, T+8,
__ 111
2u_J 8" -t
2 a?
3%
J

-- __
0.0615 5.57 262 191.8 0.00552 112 1'024 3.2 56.1 1.321

0.0939 8.51 24 8 195 0.00751 62.5 1.321 7.1 63.4 1.1862

0.1566 4.2 220 188.6 0.0113 54.4 1.569 0.55 67.9 1'1675
0.0347 3.15 256 199 0.00353 109.6 0.716 2.2 52.7 1-326
0.1549 4.01 202 202.8 0.01124 25.4 1'619 8.1 87.1 1.0895

0.0615 316 321.2 0.00575 109.6 1'29 4.0 57.3 1.3225

0.0913 8.27 299 t24.4 0.00769 62 1.635 8.8 65.5 1.1859

0.14 2.7 186 t21.5 0,01055 25.2 I '392 3.25 117.6 1.105
0.0275t 2.5 287 t15.5 0 .00300t 27.2 D.612 7.2 66.5 l.0851

0.03 2.72 341 237.4 3.003381 59.7 3.82 4.65 62.1 1.1765
0.0582 5-28 372 !51.9 3.00567 112.1 1.495 4.6 58.5 1.325

0'081 7.35 274 !47*7 )*00715 25.5 1 *39 4.9 99 1.0934

0.03136 2-84 459 195.6 ) * 0038 91 1.236 4.6 60.1 1.269

0.0514 1.94 436 01.1 )*00556 *x.
8 1.722 0.25 72.8 8.168
 1336 HEAT 'PRANSMISSIOY BETWEEN FT,UIDS AND ?JETATAS. T'EC. 1909.

SERIES D.-TABLE 4 (continued on opposite page).

&marnary of Results.

Air Passnge. Circular area = 0'0211 square foot. T = 0.492 inch.

Water Passage. Annular area = 0.00311 square foot. T = 0.0972 inch.
~

Air temperatures. Water temperatures.

-
d
'3
4
a
3
-
340 235 283.7 50.7' 54.44 3.7 0.758 54.7

339 250 291.2 52.?( 56.92 4.66 1.17 57'4

337 251 290.7 .il.O 56.1 5.1 1.279 56.8

:$lo 252 283.7 52.4 99.7 7.3 1.813 60.4

413 293 348.4 53.8 59.0 5.2 1' 3 59.5

412 299 356.9 53.0 59.2 6.2 1.55 59.8

401 308 352 S 53.9 G2.5 8.G 2.19 ti3.3

360 293.9 329.5 59.6 80.1 0.5 2.735 79.6

513 336.5 41 1.2 51.9 57.9 ti.0 1.23 58.3

48 I 334. 400.5 53 8 00 * 3' (1.52 1'57s 60.9

486 S46 1412 53'7 61.3 7.6 1.917 62

444. 356.5396.3 60.0 86.5 !ti.5 3.53 S6.9

604 442.5.517.5 4s.a 55.1 7.1 3.79 56.9

582 403 478.7 59.c 75.1 6.1 2.125 75.2
DEC.1909. HEAT TRANSMISSION BETWEEN FLUIDS AND METALS. 1337

(concluded from opposite page) SERIES D -TABLE 4.

Summary of Results.

Air Passage. Circular area = 0.0211 sqiiare foot. r = 0.492 inch.

Water Passage. Snnular area = 0.00311square foot. r = 0.OST2id;.
-
e

0 -
e2
FI

2s &Z
*o
4 II 5% ?,+t
-2 %
. y; 1 - Bi l p-+-_4 111
08
"I4 8 s$ 2 2
eF4:
su
-_ -- - ~ -.

1.03035 1.438 0.449 228 169.5 0'001965 53.5

3.0554 2.625 0.69? 233 174.7 0' 002975 81 0.*925 2.11 55.7
D'OG26 2.965 0.75i 233 174.5 0,003245 80.7 0,538 2.45 54.8

0'1137 5.39 1.0Z 222 172.5 0.0018* 79.9 0.762 3.25 57.7
0,0456 2.16 0.77 288 204-4 0' 00268 80.5 0.547 2.3 57.6
0.0577 2.735 0.912 296 208.1 0.0031 80.5 0.652 2.8 57.5

0.096 4.55 1.291 288 208'1 0'004.5 82 0'921 3.8 60.1

0.1728 8'2 1.62 248 205.! 0.00655 42.9 1.149 8.15 73.9
0.0293 1.389 0.72! 352 235.' 0.00207 66 0.5175 2 . 7 56'3

0.0451 2.14 0.93 338 231 0'002i6 77.8 0.664 2.94 58.5

0.0576 2.73 1.13 348 238 ,0'00325! 81.2 0.806 3.4 59.2

0.1685 7.99 2.09 308 242.3 0.00678 42.9 1.481 10.55 78.5

0.0986 4.67 2.24 458 288.5 0.0049 171.8 1.592 3'15 53.1

0.05 2.37 1.25 402 277.7 0.00312 42.5 0.894 6'85 70.5
 1338 HEAT WANSMISSIOX BETWEEN FLUIDS A N D MSTALS. DKC.1309.

SERIES E.-TABLE 5 (continued on opposite page).

Summary of Results.

Air Passage. Circular area = 0.001397 square foot. r = 0.1265 inch.

Water Passagc. Annular area = 0.00261 squsie foot. r = 0.143 inch.

Air temperatures. Wuter temperatures.

NO.
c

4 3
c+ 4
c
k
C
-
d
+

0
- -_- __ __ - __
E l 242.3 218 61 72.5 11.5
E 2 251.1 200 56.7 66.6 9.9
E :1 243 208-, 6.4.3 78.5 14.2

E 4 319.6 264 54.9 69.5 14.6

I35 2 9 5 . 3 263 GH .9 S0.6 16.7

E 6 311.8 269 57.8 71 13.2

&7 359'9 305 57.9 75.3 li.4

E s 335 322 84.8 104 19.2
R 9 337.6 320 72.4 87.G 15.2

El0 406.2 358 i7.9 98.1 20.2

Ell 3G2.3 332 07.2 130-3 23.1

El2 402.7 892.8 62 99 17
- -
 h c . 1909. HEAT TRANSMISSION BETWEEN VLUlDS AND METALS. 1339

(concluded f r o m opposite page) SERIES E.-TABLE 5.

Summary of Results.

Air Passage. Circular area = 0.001397 square foot. r = 0.1265 inch.

Water Passage. Annular area = 0'00264 square foot. T = 0.143 inch.

T+B, e,+t__
r - e, -
2 h1 e,- t -~
2

- -
.

3.00831 1-01: 165 159.6 0.00G14 4.54 8.85 71.2

1.020 j 2.25: 158 162.1 0.01266 17.55 8.95 66.1
). 00865 O-Y& 161 162.5 0*011612 1.43 9.5 76.2
1.028 4.04 240 199.6 0.01C8 5'6 2.9 68.7
).00969 1.39: 209 L90.8 0.00669 3.55 2.55 78.5
).0202 2.97! 232 195.6 0.0128 ,7'15 1.6 70.2
1.028 4.67 273 !23.4 0.0171 4'5 5.2 74.2
I ' 00803 1.581 226 !22 0.00701 3.511 2.9 100.9
1'01957 3.42: 261 !27*1 0'01312 7'15 2.8 86.4
1' 0273 5.35 29 5 !58-7 0.01815 3.6 7.2 96. G
1.00584 1'216 232 !47 0.00513 8.7 0.95 124.2
1.01785 3.83 294 158.7 0.01305 7.1,; 3.9 07.5
 1340 HEAP TRANSMISSION RhTtVEEN FLUIDS A N D METALS. DRC. 1909.

SERIES F.-TABLE 6 (continued 0% opposife page).

Summary of Results.

Air Pnssage. Circular Area = 0.00834 square font. r = 0.309 inch.

Water Passage. Annular Area = 0.003535 square foot. T = 0.1302 inch.
-
1
Air temperatures. Water temperatures.
54
No. - s3
-%
ern

- -g
d
El
H
G
+ -6
3
+
i
i
.C

pi
5%
OLI

--- - - - - -
F l 415 260 12.6 77.5 0.2 10725 84.6

F 2 431 2i9 i48 152 i4.1 ;1.8 4.6 .32G 66'9

F 3 389.1 280 132.1 109'5 i3.3 :O. 85 4.9 1.596 67

F 4 381.1 283. : $29.4 98 72.4 x.5 9'8 1.346 78

F 5 4 i2 298 378.5 174 76.1 S3.05 6'1 1.2375 69-5

P G 475 308 386.2 167 55.7 62'9 5.6 )-3345 68.9

F 7 475 330 397.8 145 64.8 61 .81 5.9 1,597 69.3

F 8 446 322 381'2 124 74-1 68.5! 11.1 1.35 80.4

F 9 669 4 00 528 269 71 66 LO 1.245 76.2

FLU GO7 383 183.4 224 86 82.5 7 3.341: 89.7

Fll 676 421 535.6 255 66.3 63'1 6.4 0.555 71

F12 633 426 518'5 207 77'9 73.7< 8.2 0.594 8R.9

Fl3 590 411 497'1 170 94.2 87.1 14': 0'3551 101'3

F14 736 443 583.1 293 88.4 P3.C 8.: 0.333 93. I

F1[ 715 442 569.9 273 66.7 63.1 7.: 0.558 72

F1( 725 468 588 257 78.5 73.6 9.: 0.605

I 85
F1; 637 430 520'1 207 94.i 87'8 13.: 0.355 1101.5
- -
DEC. 1909. HEAT TRANSMISSION BETWEEN VLUIDS AXTD XETALS. 1341

(concluded from opposite page) SERIES F.-TABLE 6.

Summary of Bern&.

Air Passage. Circular Area = 0.00834 square foot. T = 0.309 inch.

Water Passage. Annular Area = 0.003535 square foot. T = 0'1302 inch.

u
eJ
0)'

C @
3
" 0
:- ei h1
3 $2 eo - t '"+t
- h,
a? y; 2
b 8
W " FQ8
__ -- - --

~ 0 2 n n : 2.4 216 ins. 1 ).oms3 20.5 0.489 6.7 80.9 1.073

I 0.696
-
I). 0415 4.98 1'41 280 308 1 . 00504 92.2 0.993 4 . 2 6 3 .9 ).2365

0.112 w
2 2.75 '263 200.6 ).01046 ;6s,s 1.933 4.35 r3 ).444

0.145:: -
7.45 8.19 249 204.7 ).0128 -
9s 2.245 8 . 5 71.7 ).264

0.0351 4.21 1.367 308 224.5 1.00445 67'4 0'961 5.55 65'8 1.1731

0.0471 5.65 1.76 316 228. Z 1.00558 94-7 1.239 4.9 G5.4 1.253

0.1018 2.23 3.31 326 231.t 3.01015 168.9 2'33 5.35 64.5 3.436

0.1316 5.79 3.66 298 232.: 3.0123 99.2 2,575 9.55 73.3 1-27

0.0383 4.6 2-30: 450 303 3'00513 69.5 1.621 8.7 70-4 3.1S6E

0'0449 5.38 12.25 392 287.4 3.00575 96.7 1.582 5.8 85.4 0,273
304.t 157 2'35 5.8 66 0.405
0.0585
I
7.01 3.34
4 62
303
0.00723
3.26; 7.15 77,3 0.457
0-1000 431 0.01075 168

0.1181 392 301. 0.0121 100.4 3.34 11.4 92.6 0.293

0.0425 490 338' 0.0057 94.1' 1.96 7'35 87.6 0.223;

0.061 495 322" 0.00754 158 2-62 6-45 66-4 0.406

0.091 4 99 338., 0'0105 171.2 3.69 8.0 77'6 0.462

0.096 415 322'1 0'01072 100.4 3.13, 10.65 93.2

__ __ __ ___
 1342 HFAT TRANSMISSION BETWEEN FLUIDS AND METALS. DEC.1909.

P A R T 111.
RES~JLTS
FRON EXPERIMENTS.

1. On the Rate of Heat Transfer from Air to Metal.-An

examination of the experimental results given in Tables 3
to G (pages 1332-1341) will show clearly that tlie rate of heat
transmission increases with the speed of' flow, and it will further
confirm what has always been hitherto accepted, namely, that in
the case of a tube with liot gas on one side and water on the other,
the drop of temperature from gas to metal is very much greater
than that from metal to water.
This greater thermal resistance of the gas side is clcarly shown
in Fig. 12 (page 1355), in which the temperature gradient from
hot air to cold water is plotted for the two experiments Nos. F1
and Fa. Thus, in experiment No. Fl, the mean drop of
temperature from air to metal is 2-1-6" and the drop between the
tube and tho water is 6.7.
In these two tests, namely, F1 and F4, the differences of
temperature of air and metal are practically the same, but the
speeds of flow are very different; thus, in experiment F1, the
rates of flow of air and water are respectively 2 . 4 and 20.5 Ib.
per square foot per second, the total heat transmitted being
0 , 7 3 9 B.Th.U. per second; whereas, in expeiiinent No. F 4 with
rates of flow of 17.45 and 98 fb. per square fuot per second for air
and water respectively, the heat transmitted is 3 - 3 9 I3.Th.U. per
second, the mean difference between the temperatures of the air
and the water being 253.6" F. and 361.9" F. respectively for the
two cases. That is, with practically the same difference of
temperature the rate of heat tranafer is in the latter case 11101e
than 4 . 5 times as great as in the former, this result being
attained solely by increasing the rate of circulation.
Fig. 12 (page 1355) also shows the difference in the temperature
gradients through the metal due to the difference in the quantities
of heat being transmitted in the two cases, and how insignificant
is the internal thermal resistance of the metal when contrasted
with the external or surface resistance.
 I ~ C 190.1.
. HEAT TRANSXISSION BETWEEN FLUIDS AND ,METALS. 1343

I n Figs. 5 to 9 (pagcs 134.1-1346) the heat transmitted per

square foot of cooling surface per second per degree Fahrenheit
difference of temperature is shown plotted on a base of rate of
00w of air, in pounds per square foot of area of channel, per
second. From these curves i t will be seen that the rate of heat
transfer increases with increasing speed of flow according to a
linear law, and the following figures taken from the Tables of
Results will show the importance of the effect of speed of
circulation on the rate of heat transmission :-

Experiment No. B4 B3 B1 F1 F4
--- --_-- -----__--

WI
-.~
4.14 7-15 21.1 2.4 17‘45
a1

16 1 OaO0411 0.00639 0.01528 0.00383 0.0128

These figures speak for themselves as to the effect of speed of

flow on the rate of heat transmission, and it will be of interest
here to compare them with some figures obtained from boiler tests
which will be referred to later i n more detail.
I n a locomotive boiler firing at the rate of 130 lb. of coal per
square foot of grate per hour, the rate of 00w of the hot gases
through the smoke tubes, as determined from one of Mr. Lawford
Fry’s tests,* was about 4 lb. per square foot of tube area per second,
the heat transmitted per second beiiig 0.0047 B.Th.U. per square
foot of heating surface per degree Fahrenheit difference of
temperature of gas and tube surface, 60 that in these present
experiments, in spite of the lower temperatures, a rate of heat
transmission about 3fr times as great as the maximum reached in a
locomotive boiler ha8 been attained, due to the greater rate of flow.
The same effect waH observed i n Dr. J. T. Nicolson’s
experiments t where, with a rate of flow of 104 lb. per square foot

* Proceedings, I. Mcch. E.,1908, Part 2, page 269.

tsTransactions, Junior Institution of Engineera, 1909, vot. 1‘3, pago 199.
1344 HEAT TRANSMMISSION BETWEEN FLUIDS ANT) METALS. DEC.1909.

FIG.5.-Seelies B.

Rate of Plow of-Air, Ib. per %parefoot per second.

Rate of Flow of Air, Ib. per square foot per second.

DRC. 1909. HEAT TRANSMISSION BETWEEN FLUIDS AND YPTALS. 1345

FIQ.7.-Seriee D.

Rate of Plow of Air, 11). per siliinre foot per seuoild.

FIG S.--&ria h'.

R&e of Flow of Air, Ib. per square foot per second

1346 HEAT TRANSbIISSION BETWEEN FLUIDS AND METALS. DEC.1909.

FIG.9.--Series F.
of area of channel, the heat transmitted per second was 0.0898
B.Th.U. per square foot of heating surface per degree difference
of temperature; or a rate of heat transfer about twenty times as
great as the best locomotive boiler.
These figures, obtained by different methods of experiment,
but all agreeing i n showing most unmistakably the influence of
speed of flow on the rate of heat transfer, should cause engineers
to inquire if the present methods of design of heating surfaccs
give the best possible results, or whether, considering such
questions in the light of recent data, the present designs might
not be considerably modified wilh a view to obtaining increased
speed of circulation and the consequent increase in the rate of heat
transmission.
A second point of very considerable interest i n questions of
boiler design is clearly shown by the experiments here described.
It has been already pointed out that the drop of temperatare on
the water side of the tube is very much less than on the air side,
and a t quite moderate rates of' flow the water has no difficulty i n
absorbing all the heat transmitted, and the temperatures of water
and cooling surface differ only by a few degrees. Clearly in such
a case any further increase in the speed of Circulation of the water
will have :o very material effect on the rate of heat transmission,
nearly all the resistance being on the air side.
I n the case of a boiler it is naturally a matter of the greatest
difficulty to make even the roughest estimate as to the tipeed of
circulation of the water ; but i t appears probable that in the case
of the water-tube boiler the water may be circulating sufficiently
rapidly to absorb easily all the heat transmitted, and that
consequently the water and the tubes being nearly a t the same
temperature, nothing would be gained by a further increase in
the speed of circulation of the water. I f however the .hot gases
be made to circulate more rapidly and to scrub the heating
surface better, then ,the heat transmitted will increase, and the
metal temperature will rise, and the bpeed of circulation of the
water may then economically be increased.
1348 HBAT TRANSMtSSION BETWEEN FLUIDS AND METALS. ihc. 1909.

This is shown by the following figures taken from experiments

Nos. F 3 and F 4 ; thus in experiment No. F 3 t h e rates of flow
of air and water are respectively 13.42 and 168.8 lb. per square
foot per second, the heat transmitted being 2 - 9 2 B.Th.U. per
second; whereas i n experiment No. F 4,with rates of flow of 17 *45
and 98 lb. respectively, the heat transmitted is 3.39 B.Tl1.U.
per second, thus showing that a small increase i n the speed of
circulation of the air will more than counterbalance a considerable
decrease in the rate of‘ flow of the water. It would thus appear
that in questions of boiler design attention may be devoted more
profitably to securing good circulation of the gases than to
considering a t too great length the water circulation.
Obviously when the fall of temperature on the gas side of the
plate may be twenty times or more as great as that on the water
side, greater econoiny is to be obtained by reducing the larger
fall of temperature.
Returning to the curves on Figs. 5 to 9 (pages 1344-1346), those
points which have approximately the same d u e of all lie on
the same straight line ; but with increasing values of this term, the
rate at which the coefficient of heat transfer increases with velocity
becomes greater: that is, the greater q2,the steeper the
gradient of the ccrve. This quantity is some measure of
the temperature of the more or less motionless film of air which
adheres to t.he cooling surface, and it will hereafter be generally
referred to a8 the ‘‘ mean film temperature.”
The equations to the curves under consideration will therefore
be of the form h = A + B -at,which is clearly the law enunciated
W

b y Osborne Reynolds, and referred to in the introduction to this

Paper.
The primary object of these present experiments was to prove
the truth of this law by direct experiment, and to determine as far
as possible what influence, if any, temperature and dimensions
and sectioii of‘ channel would have upon the rate of heat
1)Pc. 1909. HEAT ‘PRANSYISSION I%ET\VlCEX FLUIDS AND XEPBLS. 1319

transmission. A glance a t Figs. 5 to 9 (pages 13461346) will show

that the rate of heat trausmission increases with temperature, the
effectof temperature being more marked at high than a t low speeds.
Table 7 (page 1350) gives the equations to all the curves
reproduced in Figs. 5 to 9.
Table 7 shows that in the equation
A, = A + B -it
w

the term il is a constant and does not vary with temperature or

section of channel ; the small differenccs observed being easily
accounted for by errors in drawing the mean straight lines.
A may however vary with the condition of the heating surface ;
thus in the results deduced from the experiments on locomotive
boilers, the values of this term are uniformly lower * than those
obtained by the author, and this difference may most easily be
explained by the doubtlessly dirtier surface of the smoke tubes of a
boiler than that of the calorimeter tube. The term B is a factor
depending both on temperature and on the dimensions of the
channel through which the air is flowing.

2. Effectof Temperatwe.-Fig. 11 (page 1351) shows the values

of the factor B plotted on a base of film temperature, and with the
exception of series F, which appears to be slightly out, the points
lie on a number of parallel straight lines, thus giving B = X Y +
‘JI where X is a function of the dimensions of the channel.
(‘2 )
The values of X and Y for the different series of experiments
are :-

D. E.
________-I_-

0.00029 0.000454 O*OO’~\~

0*00000160 0*00000163 0*000001

The average value of Y is 0*00000168.

* Average value deduced from Mr. Fry’s experiments = 0’001. See Fig. 10
(page 1346). 4 x 2
1350 HEAT TRANSMISSION BP: I'\V&EN FLUIDS A N D METALS. DEC. 1909.

TABLE 7.-Equations to Curves.

Series. Curre No. Equation.

110 I / , = 0*0015+ 0.000656 "'I

a,
isn h i = 0.00148 + 0.000715 :
205 A, = 0.0015 + 0.00075 '"'
a1

220 It, = 0.0015 + 0.00079 If'

(1 1

I&, = 0.0014+ 0'0007 -' W

c
= 0.26 in.
h, = 0.0014 + 0.000745
(11

r:
IL] = 0.00143 f0.000834 "'
"I

It, = 0.0014 + 0.000884 w L (1 1

173 k, z 0'00145 + 0.000564 y:

D
. = 0-492 in.
205 h, = 0.00144 4-0.000625 z:
2::: 11, = 0'00145 + 0.000675 -a UJ1

280 b , = 0.00145 -+ 0.00073 w1 a1

1ti0 1 1 , = 0.0015 $. 0.00075 W1

01

E
200 IL, = 0'0015 + 0'00077 W
-'
a,
r = 0.1265 in 230 h , = 0.0015 + 0.000829 '"' UI

25.5 It, = 0'0015 + 0.000862

u.1

235

230
F
I. = 0.303 in. 295

330
DEC.1909. HEAT TRANSMISSION BETWEEN FLUIDS AND META1.S. 1351

In the author’s experiments the maximum film-temperature

attained did not much exceed 300’ F., and in order to show that
the rato of heat transfer continues to increase with temperature
a t temperatures greater than those here recorded, the dotted line
011 Fig. 7 (page 1345) is drawn, showing the average results
deduced from the boiler tests of Mr. Lawford Fry.* Here the
tube diameters for calorimeter and smoke-tube are very nearly

the same, and hence the difference in the slopes of the lines is
due wholly to temperature effect. The mean film temperature
represented by the dotted line will be about 700” F.

3. Efect of Dimensions qf Channel.-It has been previously stated

that i n the law h = A + B (:-), H is a function of the temperature
and of the dimensions of the channel, and it has just been shown
t h a t it may be expressed by the equation :-
+
B = X 0*00000165( T4-0
7).
* gee alno Fig. 10 (page 1346),
 1352 HEAT TRANSMISSION BETWEEN FLUIDS AND METALS. DEC.1909

The values of “ X ” as determined from the five different

series of experiments are given on page 1349. These values are
shown in Fig. 13 (page 13551, plotted on a base of:-
= -Area of flow of__
channel
Perimeter of cooling surface inches,
the equation to the curve being :-
X = 0.000506 - 0.00045 T.
From this it becomes evident that the smaller the tube through
which the gas flows, the more rapidly will it give up its heat.
The results obtained in Series B, C, and I) show this point very
clearly. These three series of experiments were carried out with
the same copper tube in the calorimeter, but in Series B and C it
mas partially plugged up by two cores of different diameters, and
in Series D the tube was fully open. There can thus be no
question of different conditions of heating surface, and these three
series’should show any effect which may be due to dimensions of
channel.
The values of the factor B deduced from these experiments for
as nearly as possible the same film-temperatures are as follows :-

Series. . 7. B.

inch.
B. 0.171 220 0.00079
c.
D.
0.26
0.492 I 235
0.000’745

0’000675

These figures show very clearly the lower rates of heat

transfer obtained with the larger tube areas; a.nd confirm the
above statement.
Hence t o sum up, the complete law of heat transfer will be of
the form h = A +B where A is a constant for a clean metal
surface equal to 0.0015.
DEC.1909. HEAT TRANSMISSION BETWEEN FLUIDS AND XETALS. 1353

B is a function both of dimensions of channel and of temperature,

and is given by the equation
+
B = 0.000506 - 0.00045 ( r ) O*OOOOOl65 (---)
T+e

or 76 = 0.0015
+ e :;
+ [0.000506 - 0*00045 r + 0*00000165T3--] lfl

\
The laws of heat transfer from a gas into a metal surface may
therefore be summarized as follows :-
(1) For a constant rate of mass-flow the rate of heat transfer
is proportional to the temperature difference.
(2) For a given temperature difference,the rate of heat transfer
increases with increasing speed of flow according to a linear law.
(3) For a given rate of flow and temperature difference tho
rate of heat transfer increases with the absolute value of the
temperature.
(4) The rate of heat transfer depends on the condition of the
heating surface.
( 5 ) The rate of heat transfer depends on t-he size of the
channel through whioh the gas is flowing, the smaller the ratio
~ _ _Area
Perimeter channel
_ _ ~ofof_cooling-surface
_ the greater the rate of heat transfer.

4. Heat Transfer from Metal to Water.-The values of 5, the

coefficient of heat transfer from metal to water as given in
Tables 2 to B (pages 1332-41), are shown in Figs. 14-16 (pages
1355-56) plotted on a base of rate of mass-flow of water. The
results are similar to those obtained for the flow of heat from air
into a metal surface, inasmuch as they show that the rate
of heat transmission increases with the speed of circulation (for a
given temperature difference) according to a linear law. They
agree qualitatively with the results of Dr. Stanton.
One very interesting conclusion as to the effect of the form of
the heating surface may be drawn from these results. Ih
experiments B to D, a fine screw-thread was cut on the outer
surface of the calorimeter tube, with the object of breaking up the
flow of the water and thus, it was thought, increasing the rate
of heat transmission. In series E and F the tube surface was not
so broken. Comparing the curves on Fig. 14 (page 1355) with those
1354 HEAT TRANSMISSION BETWEEN FLUIDS ANT, METALS. DISC.1909.

on Figs. 15and I 6 (page 1356) the rate of heat transfer is, however,
riot noticeably greater in series B to D than in series E and F ,
and thus the broken surface of the tube has no marked effect on
the heat transmission. This is probably due t o the bottom of the
thread being filled with nearly stagnant water, which neutralizes
the good effect produced by the water scrubbing hard over the
tops of the thread.

The Paper is illustrated by 16 Figures in the letterpress and

is accompanied by an Appendix.

APPENDIX
On page 1330 tho following equation is given for the drop of
temperature 6 T of the air in passing through an element of tube
6 X long-
- w1 k 6T = 7~~(T - 0) a d 6X,
where w, = Ib. of air flowing per second ;
7c = specific heat of air ;
h, = heat transmitted per second per square foot per degree
Fahrenheit difference of temperature of air andnietal;
0 = metal temperature ;
d = diameter of heating surface.
ae '
0 varies along the tube ; = c, a constant, the metal temperatures
at top and bottom being 0~ and OB, and hence :-
c=-
OT - OH
L
.*. 0=6T-CX
or - k w, 6T = h, {T - (OT - cx)}ad 6 s

h, 7r a
Write M = e, w,
WIQ. 12.-Ternperature Gradient FIG.13.
through Calorimeter Tube. Dimensioias of (hnnel (Function x).
1356 HEAT TRANSMISSION BETWEEN FLUIDS AND METALS. DEC.1909.
DEC.1909. HEAT TRANSMISSION BETWEEN FLUIDS AND METALS. 1367

or ET+ M T = MjOT - C X )
dX
w e may write T = u e - f M ~ =Yu e - M x
which leads to the equation

ax
%-MX = (6, - &)ill
or d‘ - Cx>Me M X
ax = (fh
Integrating u = D +&€T - cS)e M dS
=D +M e ~ f i ~ ~ Mc-
d J’Se’lXdX
e
Therefore the mlution