147

The Circulation of Water and Steam in Water-Tube

Boilers, and the Rational Simplification of Boiler Design
By W. Yorath Lewis, M.I.Mech.E.,* and Struan A. Robertson, B.Sc. (Eng.)?
The authors stress the need for a sound theory of circulation in water-tube boilers, to enable weight
and space to be saved and efficiency and reliability to be increased. They establish, as a standard of
excellence with which other boilers may be compared, a simple boiler consisting of a single drum and
a single U-tube.
The paper consists of three parts. Part I describes the nature of the change from water to steam j
in part I1 the flow in the simple U-tube boiler is calculated ; and in part I11 the features of various
departures from the standard U-tube, met with in practice, are discussed.
An improved type of tube, approaching very nearly to the standard U-tube, is evolved, allowing
much simplification of boiler design, with the advantages of excellent circulation, high gas speeds,
and higher heat transmission rates. Other rational simplifications and improvements are also rendered
possible, which enable substantial reductions to be made in the cost of manufacturing, installing, and
maintaining boilers of any pressure and capacity, for any service, ashore or afloat.

Introduction. In the design of water-tube boilers, a With most types of boilers, it is impracticable to
knowledge of the mechanics of circulation is indis- attempt any analysis of the circulation, due to the in-
pensable, if the arrangement of the tubes and the fixing determinate division of the flow between tubes connected
of their dimensions is to be carried out satisfactorily. in parallel between drums and headers. This difficulty
Faulty circulation, leading to overheating and blistering is probably the chief reason for the failure of many
of the tubes, is one of the principal causes of defective attempts to establish a theory of circulation.
operation of modern boilers. Due to bad circulation, The most useful line of approach would be to divide
the average heating rate per square foot of heating the subject into its quantitative and qualitative aspects,
surface in most boilers requires to be severely limited, and deal with each separately. Some simple type of
with the result that the boilers are unduly large, heavy, boiler, in which the flow can be calculated accurately,
and costly. requires to be established as a standard of excellence.
Even to-day, in spite of many attempts by investi- Using this standard, the performance of more compli-
gators in all countries, the mechanics of water and steam cated designs can be estimated, the effect of deviations
circulation is a subject not fully understood. Many from the standard studied, and guidance obtained in
ideas on these matters are somewhat superficial in their designing a boiler which is as near the standard as
conception, and some indeed on detailed examination possible.
appear to be in fact unsound. The standard proposed by the authors consists of
In the past, boilers have operated apparently in a the U-tube boiler shown in Fig. 1. This comprises a
satisfactory manner in spite of these errors. To-day, single water and steam drum, with a single tube which
however, with increasing pressures and capacities, forms a “U”. The right-hand leg of the “U” is the
there can be little doubt that these faulty theories are upcomer, and discharges its contents above the level
having a very pernicious effect. Without correct know- of the water in the drum. The tube is considered to be
ledge the design of the boilers is merely empirical, and capable of being heated to any required intensity,
bad features of former designs are carried forward and over any required sections of its length. The actual
uncorrected in new designs. The designer is not in a method of obtaining this heat is immaterial, since the
position to detect possible improvements, whereby standard is concerned with circulation only.
weight, space, and complications may be reduced; I n order to be able to make detailed calculations to
output, efficiency, and reliability increased; and the determine the efficacy of the proposed standard, some
frequency and duration of shut-down periods for cleaning dimensions have been allocated to it. The drum is
and overhaul reduced to a minimum. taken as being 4 feet in diameter, with the water level
half way up. The tube is 13 inches outside and 1) inches
The MS.of this paper was received at the Institution on
10th November 1938. The paper was presented for discussion inside diameter, and 27 feet long, forming a “U”
at an Extra General Meeting of the Institution on 1st March 123 feet long, with 2 feet of tube inside the drum.
1940. The radius of the bend at the bottom is immaterial, and
Formerly Managing Director and Engineer of The British its effect is neglected in subsequent calculations. The
Niclausse Boiler Company, Ltd.
t Lecturer in mechanical engineering, Battersea Polytechnic, dimensions are shown in Fig. 1, p. 148.
London. The actual dimensions adopted, whilst quite arbitrary,

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

148 T H E C I R C U L A T I O N OF WATER A N D STEAM I N WATER-TUBE BOILERS
should not affect the theory of circulation. Possibly, between the steam in the inner tube and the water
with other dimensions, a standard tube could be designed in the annulus is sufficient to cause violent ebullition,
to give adequate circulation under a wider range of simulating the action in a tube of a water-tube boiler.
pressures and rates of heating than the tube selected The pressure of the steam in the inner tube, measured on
here. For the present, however, all that is required is to the pressure gauge, enables the temperature to be known,
show that the selected tube is a possible standard. from steam tables, provided there is always a discharge
of steam to waste. This gives a rough indication of the
heat transfer rate, though of course the heat transfer
P A R T I . T H E C H A N G E F R O M WATER T O S T E A M varies with the velocity of the steam and of the contents
of the annulus. A short projecting pipe may be fitted
Before any calculation of the flow in the standard to the top end of the main tubes, to enable them to
U-tube may be attempted, a brief survey is required discharge above the water level in the upper drum.
to obtain a correct understanding of the mechanism
of boiling of water in a tube. The change from water to
steam is not a simple one. A certain amount of research C

,,CDIA. STEAPl PIPES STEA?l

SINGLE WATER A N D EXTERNAL DIAMETER IN
APPROX.
U
I .

WATER LEVEL - . .--. . .,"_

WATER LEVEL

&TUBE { I! INTERNAL DIAMETER

$'EXTERNAL
" DIAMETER

F-UPCOMER

IDOWNCOMER

+LOWER DRUM

TO WP

Fig. 1. Standard U-tube Boiler STOP VALVES A2AND Cz

has been undertaken in recent years on this problem, Fig. 2. Demonstration Boiler
particularly in America and Germany. Most of the
investigators have been concerned chiefly with evapora- The central tube B, which is unheated, may act as a
tion at or below atmospheric pressure, in steam-heated downcomer, or it may be stopped up, in which case
evaporators. Not a great deal has been done at the pres- the two remaining tubes form a U-tube boiler similar
sures and rates of heating usually required in water-tube in general principles to the standard U-tube. An estimate
boilers. Considerable guidance, however, can be derived of the effect of carrying out the experiments at atmo-
from the results published by these investigators. spheric pressure, as compared with normal boiler pres-
In addition, experiments have been made by the sures, is made at a later stage.
authors, using the apparatus illustrated in Fig. 2. This The evaporation from water to steam in a vertical
consists of an upper and lower drum, connected by tube, in which the contents are flowing upwards, takes
three glass tubes, each 14 inches internal diameter place as a result of two causes, first, the heat absorbed
and 6 feet long. The upper drum has glass sides. Two from the tube wall, and second the reduction in pressure
of the glass tubes, A and C , have within them a steel as the liquid rises in the tube. Consideration of the
pipe 4 inch diameter, through which saturated steam second cause, and of evaporation when the contents
may be passed at pressures up to 120 Ib. per sq. in. are flowing down the tube, as when in the back leg of
The steam supply to each inner tube may be regulated the U-tube, is deferred to a later part of the paper.
independently of the other. The temperature difference The first description of the general mechanism of

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS 119
boiling in a tube, was given by Barbet (1914).* More SUPERHEATED STEAM STAGE
recent investigations are given in papers by Brooks and
Badger (1937), Boarts, Badger, and Meisenburg (1937), Y SATURATED STEAM
and Stroebe, Baker, and Badger (1939) in which different D OF LATENT HEATING
sections of the whole process have been described.
Fig. 3 has been prepared by the present authors to
illustrate the findings of these investigators, together WET STEAM STAGE
TUBE WALLS DRY
with results of their own experiments.
Up to section A (Fig. 3) there is water only in the tube.
At the velocities usual in water-tube boilers, the flow END OF WATER FILM
will be turbulent. On any cross-section there is a thin
boundary layer, next the tube wall, moving with laminar
or viscous flow, then a transition zone, then the turbu-
lent core. The temperature drop, between the inner
tube surface and the water core, is concentrated almost
entirely in the boundary layer, across which heat moves
mainly by conduction. In the turbulent core, heat is
quickly distributed by eddy currents, so the temperature WATER FILM STAGE
is approximately uniform. Above A, when the tem-
perature of the water immediately in contact with the
tube is high enough, small bubbles of steam form on FORMER SLUGS OF WATER
the tube wall.
Due to the surface tension of the water enclosing a WATER FILM
spherical bubble of steam, the pressure in the interior
of the bubble is greater than the pressure in the water,
the excess being inversely proportional to the radius NO OF SLUG STAGE
of the bubble. If no particles of air or other nuclei LUGS OF WATER
are present to start the formation of bubbles, boiling
will not begin until the temperature is raised much
above the point corresponding to the external pressure,
that is, the liquid becomes superheated, above satura- LARGE BUBBLE DERIVED FROM COALESCENCE
tion temperature. Once formed, the bubble would be Of CLOUD OF SMALL BUBBLES
highly unstable, for, as the radius increases, the tension
in the envelope becomes less and less able to balance COALESCENCE BEGINS
the excess of pressure within it, and the expansion
into vapour would take place with explosive violence.
This happens to a certain extent when water is boiled CLOUD FORMATION
after being freed of air in solution, and it is said to boil
with “bumping” (Ewing 1922).
The superheating of the water above the saturation SMALL BUBBLE STAGE
FORMATION OF PERMANENT BUBBLES
temperature has been noticed by many investigators.
Jakob (1936) found, however, that water can be evapor-
. ated without considerable superheating, if there are in
the liquid sufficiently small curved surfaces on which
steam molecules can collect. Such curvatures exist, INT OF COMMENCEMENT OF EVAPORATION,
for example, on the rough surfaces of the walls. The OR “POINT OF EVAPORATION”
greater the heating rate, the greater will be the tempera-
ture of the water next the tube wall, and the smaller FORMATION OF TEMPORARY BUBBLES
becomes the radius of the tube surface irregularities
on which bubbles can originate. These irregularities
form nuclei, and this form of boiling is termed nucleate
BOUNDARY FILM
boiling. Bosnjakovic (1930), developing the theory of
the growth of steam bubbles, estimated that most of the TURBULENT CORE
heat flowing from the tube wall to the bubble was not
transmitted directly, but was transmitted first to the SENSIBLE HEATING STAGE
water and then from the water to the bubble.
A bubble formed on the heating surface will grow till
it is swept off by the moving liquid film, or till it bursts
through the film. The bubbles, once dislodged, travel in- Fig. 3. Change from Water to Steam in a Vertical Tube
* An alphabetical list of references is given in Appendix 11, The vertical scale is much smaller than the horizontal
p. 174. scale.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

150 T H E CIRCULATION OF WATER AND STEAM I N WATER-TUBE BOILERS
ward towards thi: core of the tube. A chain of bubbles from the climbing film of water on the tube walls. At
may originate from any particular nucleus. At first, this still higher rates of heating (Stroebe, Baker, and Badger
core is not sufficiently hot to enable the bubbles to exist, 1939) the coarse spray vanished altogether, showing
and the bubbles collapse after a very short journey. This that the film of water had evaporated.
occurs in section AB of the tube. By condensing, and No experimental information is as yet available as to
giving up their latent heat, the bubbles contribute towards the actual dryness fraction of the steam represented
raising the core temperature to saturation temperature. by the different sections in Fig. 3. The length of tube
Above B, the core temperature is sufficiently high to occupied by the different stages depends on the intensity
enable the bubbles to continue to exist. The bubbles of heating, and on the velocity of water supply at the
are permanent, and increase in number and size. They bottom. It will be governed also by variations in viscosity
rise through the water, that is, their upward velocity and surface tension, as affecting the possible size of
is greater than the upward velocity of the water. Al- bubbles. T h e diameter of the tube has an influence,
though, in the earlier sections of the bubble stage, the as it is easier for slugs of water to persist in a small
distribution of the bubbles is fairly uniform, after a tube than in a large tube, for the same value of surface
point they seem to gather in clouds, as at C, separated tension.
by sections in which the bubbles are less numerous. I n the authors’ experiments, with the maximum rate
Various observers have described how, after a certain of heating, the froth stage appeared quite short, as also
point D, the bubbles of each cloud coalesce, to form a did the slug stage, whilst the stage in which steam con-
single large bubble, occupying the whole area of the tained drops of water appeared to persist throughout
tube. The water displaced during this action forms a for the remainder of the tube. The film of water on the
slug of water above the large steam bubble, and the wall of the inner tube, from which the heat was being
steam pushes this slug ahead of it. Between D and H, derived, appeared to be quite thin, similar to the wetness
there is a section of “slug action”, the tube being occu- on an ordinary wet pipe. The possibility of this wet film
pied by alternate slugs of water and steam, travelling evaporating, however, was shown in other experiments
with equal velocity. There is still a film of water on the in which the flow was purposely arranged to be defective,
tube wall between the steam and the metal. As they and the film could be seen drying up.
travel upward, the water slugs become thinner, due to Before leaving this subject; it should be noticed that
continued evaporation within them, until they finally the slugs of water and steam formed during evaporation
break up as at K. are not the same as might occur in a water-tube boiler
From H to M, there remains only a film of water having bad circulation and stagnation, or due to water
on the tube wall, with occasional bulges, as at K and L, entering the tube from the top end. If the upper part
which remain from former slugs of water. This water of a tube is cool, steam formed at the lower parts may
film travels upwards, but at a speed less than the steam condense again and flow down, interrupting the processes
speed. It evaporates by nucleate boiling, the small of evaporation.
bubbles of steam forming on the tube wall, and being
either swept off by the movement of the film, or bursting The Maximum Permissible Dryness. An important
through the film. During the breaking up of the slugs, consideration arising out of the division of the evapora-
and during boiling of the film of water on the tube wall, tion into the different sections, is the effect on the heat
the bursting of steam bubbles into the general column transfer coefficients. This, in turn affects the permissible
of steam entrains particles of water into the core. These heating rates, and also may have an influence on the
travel upwards with the steam, and form one of the .permissible maximum dryness fraction of the steam
principal elements in the wetness of the steam. Above M issuing from the tube, if the temperature of the metal
there is wet steam, that is, steam containing drops of of the tube is to be kept within safe limits.
water, passing up the tube. These gradually evaporate T h e heat transmitted, h, expressed in British Thermal
until the steam becomes dry saturated, at N. .4bove this Units per square foot per hour per degree Fahrenheit
point superheating begins. difference in temperature between the inside tube
Barbet (1914) described the stages of coalesence, slug surface and the contents of the tube, where these
action, collapse, and of film movement with annular contents are either water (up to Section A), or dry
rings. Brooks and Badger (1937) repeated the experiment saturated steam, may be calculated from McAdams’
using a glass tube 4 inch outside diameter in a steam formula (1933),
jacket, and making the same observations. When their
rate of heating was such that the top of the tube was
equivalent to being at C, froth appeared. On increasing
the rate, the liquid came out in slugs at high velocity, where D denotes the diameter of the tube, assumed
uTith quiet intervals between the slugs. When still li) inches or 0.104 feet, K the thermal conductivity, V
hotter, the slugs ceased and two types of spray were the velocity, S the density, Z the viscosity, and C, the
seen, coarse and fine. The fine spray seemed to be at specific heat.
the centre of the tube, travelling straight up. The The properties of water and steam at various pressures
coarse spray seemed to be travelling at a much lower and temperatures are tabulated in Appendix I, p. 174.
velocity then the fine spray, and seemed to be at the In later calculations it is shown that in the standard
outside and was blown outwards, as if it originated U-tube, for a wide range of pressures and heating rates

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E C I R C U L A T I O N O F WATER A N D STEAM I N WATER-TUBE BOILERS 151
the velocity of the water in the tube is in the neighbour- hence metal temperature, and also the film thickness
hood of 5 ft. per sec. For this rate of flow, the value of h corresponding to any given dryness fraction. It should
for various pressures has been calculated (Table 1). be noted that the metal temperatures referred to in
this section are for the inside tube surface. The tempera-
TABLE OF h FOR VARIOUS
1. VALUES PRESSURES
- ture at the outside tube surface will be greater.

Pressure, Ib. per sq. in. Effect of “Steam Film Boiling”. In the section of
. .
abs.
Value of h, B.Th.U. per
--- -
loo I I
3oo
1 the tube containing water, if the temperature drop
between the tube wall and the water is very high, a
blanket of steam may form next the tube wall, offering a
hr. per sq. ft. per
deg. F. difference in
temperature :-
(1) Inside tube wall to
I I high resistance to the flow of heat to the interior of the
tube. Heat can only pass this layer by conduction.
The coefficient of thermal conductivity of steam is con-
water 1,840 1,950 2,000 1,970
( 2 ) Inside tube wall d siderably less than that of water. The phenomenon is
, I

dry steam . 680 802 915 1,170 comparable to the cushion of steam which supports a
- drop of water on a very hot plate, when the drop assumes
the “spheroidal state”.
Table 1 shows that the tube wall temperatures, for S. Nukiyama (1934), investigating the problem,
any given rate of heat absorption by the contents of found that if the temperature difference between metal
the tube, are greater at those sections of the tube con- and water exceeded 72 deg. F., this blanket formed.
taining steam than at the sections containing water. A very large increase of about 1,200 deg. F. in tempera-
No data are available regarding the heat transfer to ture difference was necessary before any increase in heat
wet steam, as in the section MN of the tube. It can flow could be obtained. Once the film was formed, it
hardly be less than for dry steam, and may be more, tended to persist, even if the temperature of the metal
due to the presence of water particles in the steam. were reduced to 22 deg. F. above the water temperature.
In the boiling section from A to M, the conditions At the point of maximum heat transfer coefficient, the
are complex. Various investigators (Boarts, Badger, and rate of transmission was 265,000 B.Th.U. per sq. ft.
Meisenburg 1937; Stroebe, Baker, and Badger 1939) per hr.
have found that the heat transfer coefficients are greater In Nukiyama’s experiments the metal was in the
in the boiling section than in the water section. This form of a very fine wire, electrically heated. Drew and
may be due to extra scrubbing, and to eddies caused by Mueller (1937) used a copper tube t inch in diameter
violent ebullition reducing the thickness of the viscous immersed in liquid, with steam passing through. They
film of water on the tube wall. The estimates of the found that the critical temperature difference between
increase in heat transfer in the boiling section vary con- the steam inside the tube and the liquid, at which the
siderably. King (1934) showed that the coefficient in- steam blanket formed in the outside of the tube, was
creased with the heating rate according to a straight-line 55-90 deg. F. The temperature difference between the
law. Cryder (1933) found the coefficient to be propor- metal surface and the liquid would be a fraction of this
tional to the 2-4th power of the temperature difference. only.
No information is available as to whether the coefficient At a temperature difference of 144 deg. F., the steam-
of heat transfer varies in the different stages-froth, slug, heated tube was seen to be enveloped in a film of vapour.
and film-in the boiling. The film was quite smooth except on the upper side,
Although the information is so imperfect, it is likely where there were oscillating irregularities and occasional
that the tube wall temperature increases considerably bulges that developed into bubbles. Much vapour
above the point M where the water film terminates. escaped, not by bubbling but by flowing along the heating
This may set a limit to the permissible rate of heating. surface to the place where the tube emerged from the
It may also fix the maximum permissible dryness fraction liquid.
in the tube, if it is desired that the tube shall always Heat transfer across the vapour layer took place largely
have a water film on it. by conduction and convection. Variations in the surface
There is little practical experience of established tension and in the concentration of dissolved substances
practice or experimental data on which to form a affected the results considerably. The nature and con-
judgement. For most boilers, due to indeterminate flow, dition of the heating surface had a profound effect on
it is quite impossible to estimate the dryness of the both the magnitude of the maximum rate of heat transfer
steam issuing from any particular tube. In certain and on the critical temperature difference. Corrosion
forced-circulation boilers, in which the makers claim raised the maximum fourfold.
to be able to predict the flow in any tube and the heat The experiments of Nukiyama, and of Drew and
applied to it, a dryness fraction of 0.17, that is, a “six Mueller, were made at atmospheric pressure and with
time round” boiler, is considered to give satisfaction. stationary liquid. The effects of high pressures, and of
The subject merits further investigation. In this the movement of water through a tube, still require
paper, a dryness fraction of 0.20 at exit will be looked study. At high pressures it would be possible that the
on as satisfactory, though the value should really vary steam film might be thinner, and hence the resistance
with pressure, as affecting the saturation temperature and to heat flow less.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

152 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS
A temperature drop from metal to water of 72 deg. F., If the feed water supplied from external sources to the
as found in Nukiyama’s experiments, would give a heat drum is at saturation temperature, then it is possible
transfer of approximately 130,000 B.Th.U. per sq. ft. to say at once that the water in the downcomer tube is
per hr. in the U-tube, as the upper limit. This figure is at saturation temperature, and the point of evaporation
much less than that obtained in boilers in practice, is at P (Fig. 4). The effect of variation in pressure,
especially in front-row boiler tubes exposed to both and hence of saturation temperatures at different parts
radiant and convectional heating. Dr. Docherty (1928) of the “U” is examined later.
mentioned recent experiments which indicated a heat If the external feed water supply is at a temperature
transfer rate in fire-row tubes of 143,000 B.Th.U. less than saturation temperature, a rather indeterminate
per sq. ft. per hr. McAdams(1937), in experimentswith state of affairs in the drum will exist. Where does this
a copper tube, obtained heat transmission rates of 400,000 feed water receive the extra sensible heat which is
B.Th.U. per sq. ft. per hr. required to bring it to boiling point ? Some guidance
The possibility of a blanket layer of steam existing
in a boiler tube, setting a maximum value to the rate
of heat transfer to the contents of the tube, merits further
experimental consideration. The effect of such a blanket
would apply only to the tube up to section M, where
the water film ceases. Whether or not it has an appre-
ciable effect when the contents of the tube are in motion,
it must have a very serious effect if there is stagnation
in the tube. With stagnation, the tube temperature may
rise to a very high value if the water in the tube is
protected by a steam blanket. Once the water is evapor-
ated, the steam temperatures may rise to heights ap-
proaching the flame temperature of the external gases.

PAR?’ 1 1 . T H E S T A I S D A R D U - T U B E B O I L E R
The flow in the U-tube boiler, which it is proposed
to establish as a standard of excellence, may now be
examined in detail. The first requirement is to determine
the point at which evaporation commences. This point
corresponds to B in Fig. 3, when the steam bubbles
form and are permanent. It is referred to in later parts
of the paper as the “point of evaporation”, though in
reality it is the point of commencement of evaporation.
Let PR (Fig. 4) be the heated part of the tube.
The rate of heat absorption by the contents of the tube
will be considered uniform over the length PR. The Fig. 4. Heating of Standard U-Tube
source of this heat is immaterial. Due to the variation
in the resistance to heat transfer in the different sections in answering this may be derived by examining two
of water flow, boiling, and wet steam flow, uniformity
of heat absorption will imply varying tube temperatures, extreme possibilities, between which the truth must lie.
and hence various temperatures of the outer heating On one hand it is possible that the sensible heat is
derived directly from the fire, as the water passes round
medium. In a boiler, however, the heat resistance the U-tube. If this is the case the water in the drum must
between the inside tube wall and the tube contents is a evidently be at less than saturation temperature. The
very small fraction of the total resistance, which is apparent anomaly of having steam in contact with water
proyided mainly by the gas film on the outside, so that
variations in it will have little overall effect. If conditions at a temperature lower than saturation temperature
are uniform outside the tube, there is no great error may be explained by the presence of layers of water
in assuming the rate of heat absorption by the contents in the drum at different temperatures, the water at the
of the tube as being uniform, whatever the state of these
surface being at saturation temperature.
contents. The point of evaporation will now be at Q, some
It is also immaterial, for the present, whether the distance above P (Fig. 4).The water passing I? will flow
heat is applied uniformly over the tube surface, or for some distance to Q, taking up sensible heat, before
starting to evaporate. The calculation is as follows.
whether it is concentrated on one face, as would be the The symbols used are given below :-
case with radiant heat from the combustion chamber.
The maximum intensity of heating may be much more t Saturation temperature corresponding to the steam
than the average. For convenience, however, calculations pressure in the boiler.
are based on the average heat absorption per unit ,t Temperature of the feed water supplied to the
area of the total outer circumferential surface of the tube. drum.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E C I R C U L A T I O N O F WATER A N D STEAM I N WATER-TUBE B O I L E R S 153
t , Average temperature of the water in the drum. begins to circulate at all. Cold feed water entering the
L Latent heat of steam at boiler pressure. drum will tend to reduce the temperature of the water
x Dryness fraction of the steam passing up the in the drum. This in turn will tend t6 cause condensation
tube at R. of some of the steam in the drum, and to lower the pres-
W Weight of water plus steam circulating in the tube sure. A state of equilibrium will be obtained when the
and passing any point per second. latent heat from the condensed steam has been enough
Then there are xW lb. of steam formed per second. to raise the fresh feed water to saturation temperature.
This is taken away from the boiler by the steam main Thus, of the mixture of steam and water issuing from
to the engine. The feed water supplied per second is the upcomer tube into the drum, the water returns
equal to this XW lb. at temperature t,. The water in the immediately to the water in the drum, some of the steam
steam issuing from the top of the upcomer is W (1 -x) lb. condenses, and the remainder of the steam flows out
per sec. at saturation temperature t. Hence the water in through the steam main. .
the drum is composed partly of feed water at temperature Were conditions static, such an equilibrium might
tr flowing in at XW lb. per sec., and circulating water at indeed be reached. Actually there is a continuous flow
temperature t flowing in at the rate of W (1-X) lb. of feed water, so that there must be a continuous flow
per sec. These mix to form xW+ W [ 1-2) lb., that is, of heat from the steam to the water, by condensation.
W lb. at temperature td. This implies a difference in temperature between steam
and water. How much this difference amounts to would
Therefore +
xWt, W(l-x)t = Wtd he difficult to estimate.
Hence t , = x$+t--xt If the difference were zero, then conditions would
and t - t , = r(t--t,) . . . (1) exist similar to those already discussed at first, when
Betn-een P and Q, W lb. of water pick up sensible feed water is supplied at saturation temperature. The
heat amounting to (t-td)W B.Th.U. From equation (l), flow of the water down the downcomer would be at
.r[t--tf) may be substituted for (t-td), giving saturation temperature, and the point of evaporation
would be at P. If the difference is appreciable, then
Sensible heat = x(t-$)W B.Th.U. the conditions existing in the second case are approached,
Between Q and R, x W lb. of water pick up latent heat when all the sensible heat is derived direct from the fire.
amounting to WxL B.Th.U. Assuming that the absorp- The truth must be between these two. Part of the
tion of heat is uniform between P and R, then sensible heat is no doubt derived directly from the fire,
and part by condensation of steam in the drum. The
PQ - Wx(t-t,)
_ error in assuming Q (Fig. 4) as the point of evaporation,
QR- WxL dividing PR in the ratio of sensible heat to latent heat
t-tf will be small. The rate of heating and the rate of heat
=-
L absorption in an actual boiler tube vary considerably
- Sensible heat acquired in boiler per pound along the length. The position of the point of evapora-
- tion can only be determined roughly by taking into
Laterit heat of steam per pound
account all the variable factors, and estimating their
Thus, if the feed temperature, the saturation tempera- effect. It should be possible to find it sufficiently nearly,
ture and the latent heat are known, the point of evapora- however, to enable an adequate calculation of flow to
tion Q can be determined. This point is independent be made.
of the intensity of heating, or of the dryness of the
steam at R, or of the weight of water circulating per Head Causing Flow. Having determined the position
second or of the temperature of the water in the drum. of the point of evaporation, the head causing flow
For example, if the feed water temperature t, is 250 deg. in the tube may be calculated. The dimensions of the
F., and the pressure is 300 lb. per sq. in. abs., then tube are as stated in Fig. 1. Let X be the dryness
t = 417.4 deg. F., and L = 815.2 B.Th.U. per lb. fraction of the steam at R. Let h be the height of R
PQ 417.4-250 1674 o.205 above the point of evaporation Q, k the height from
Then -- - R to the water level at S in the drum, and a the cross-
QR = 815.2 815.2 - sectional area of the tube.
In the example just given, it is assumed that the rate The force causing flow will be due to the difference
of heat absorption between P and R is uniform. If this is between the weight of a column of water (h+k) feet
not the case-for instance, if P is at the top of the down- high, and the weight of a column of wet steam, whose
comer tube, and the downcomer tube is heated less dryness increases uniformly from zero at Q to X at
intensely than the upcomer tube-the ratio PQ/QRwould R, and continues at the value X up to S. The weight
be increased accordingly. If an economizer is fitted the of the column of water is
feed temperature will be very nearly saturation tempera-
ture, hence PQ/QR will be very small. The point of
evaporation will be very near the point at which heating
begins. where V, is the volume of 1 lb. of water.
The other extreme possibility is that all the sensible If the steam and water at any cross-section of the
heat is obtained in the steam drum, before the water upcomer tube are travelling with the same velocity,

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

154 T H E C I R C U L A T I O N O F W A T E R A N D STEAM I N W A T E R - T U B E B O I L E R S
the volume of 1 Ib. of the mixture at dryness x, by between sections T and U, which are a very small
weight, is xVr+(l-x)Vt,, = x(V,-Vzu)+V,. Hence the distance dy apart, is
density is adv
1 V
-
x vs+( 1 x) V , ps+(l-~x)Vw
and the weight per foot length of tube is The total weight of the column QR is therefore

a
fL adY
.UV,+( 1 -x) v,
U
This value for the density will apply in that section
of the tube where the water and steam pass upwards - aL vw+a v,- VUJ
in slugs having equal velocity, i.e. DH in Fig. 3. It X ( v*- V,) loge V
W
will also apply fairly accurately in the wet steam stage, The average density of column QR
above M (Fig. 3), although it is possible that the water
drops in the steam may have a slightly lower velocity =-Weight
Volume
than the steam. In the froth stage BC, the steam bubbles - 1 v,+-V,- Vw)
rise faster than the water. For any given dryness fraction
the cross-sectional area of the steam stream will be less, X ( v,- V,) loge Vw
and the weight of the contents of the tube, per unit T o illustrate this point, Fig. 6 shows
length of tube, will be greater, than if steam and water a graph AB, of the density of steam
have the same velocity. of various dryness fractions, for a
Several investigators have examined this point very pressure of 300 lb. per sq. in. abs.,
thoroughly to determine the drag, or reduction in head, also a graph AC of the average density
due to the steam bubbles running ahead of the water. of a column of steam whose dryness
Schmidt (1933) made experiments at Danzig with increases uniformly from 0 to X. The
tubes 20.3 feet high, and various diameters up to 3-25 part of the graphs for values of X
inches, and pressures up to 70 Ib. per sq. in. The relative between 0 and 0.1 is plotted again on
velocity was found to decrease considerably as the pres- a vertical scale ten times larger (graphs
sure increased. The effect in the U-tube, however, AD and AE), to facilitate accurate
under conditions approximating to those in boiler tubes, measurements of the densities for low
must be quite small, due to the very limited extent of values of the dryness fraction. T
the froth stage and to the high pressures. No great The mathematical calculation of the
error is introduced by neglecting the increased velocity average density might be avoided by
of the steam relative to the water at this stage. the use of a planimeter. For example,
A similar difficulty arises in the water film stage HM if the top dryness is 0.10, the top
where the steam in the core rises with greater velocity densitv is
than the water film on the wall. The thickness of the 1
film, however, appears to be very small, and little error = 5.82
(1.552-0-019) ~ 0 . 1 + 0 4 1 9
will be introduced in neglecting this effect.
The approximation is therefore made, that neglecting as shown at c, on graph AB. Q
the effect of the short sections where the steam velocity The average density may be obtained
and water velocity are different, the weight per foot by measuring the area abcde,and divid-
length of tubes is given by ing it by the height ab. This gives
a the value of bf, the average of all the Fig. 5. Integra-
densities over the range ab, and equal tion of Weight
v,+
x (1-x) v, to 14.4. This same value would be of Column
To obtain the average density of a column of steam obtained by substituting in the general
whose dryness increases uniformly from zero at the equation for average density as follows :-
bottom to X at the top, it is necessary to integrate the Average density
weights of small lengths of the steam column in the -
- 1 0~019+0~1(1*552-0*019)
tube. 0.1(1 -552- 0.019) loge 0.019
Referring to Fig. 5, the distance QR = L. At T, = 14.4
distant y above Q, the dryness x = yX1L and the density
is Frictional Resistance to Flow. The pressure drop,
1 due to friction, in the tube already considered, may be
.21xvs+(l-;Y)v,,
L calculated most conveniently by the equation
4p*s
T h e weight of the small volume of steam in the tube
P r = - 2gD

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS 155
where pf is the pressure drop in pounds per square friction calculations, whether for water or steam. The
foot, per foot length of tube, D the diameter of the tube error will be small, and on the safe side, and the simpli-
in feet, V the velocity in feet per second, S the density fication appreciable.
in pounds per cubic foot, g 32.2 ft. per sec. per sec., and If v is the water velocity in the downcomer, and V
f a friction factor. The relevant value off, for smooth is the velocity of the mixture at any particular point in
clean commercial steel pipes 3 inch to 2 inches diameter the upcomer, s the density of water, and S the density
are given in Fig. 7, for various values of the Reynolds of the mixture, then
vs
V=%

Substituting in the friction equation, the pressure drop

at the particular point mentioned is

4f v2s2
=-
2gDS

Therefore the friction head of water per foot of pipe

=&
S

For any given value of v, the friction head per foot

is proportional to x. Hence, in calculating the total
friction head for the whole tube, in which the dryness
fraction x varies, the arithmetical average dryness
fraction may be used to find the average friction head
per foot of tube, and then this average friction head may
be multiplied by the total length of the tube.
I n this calculation, the effect of differences in the
water and steam velocity at any particular section has
0 10 io 10 0
DENSITY-LB. PER CU. FT. been neglected. In the froth stage, the effect must be
very small, since this stage is very short. In the film
Fig. 6. Top Density and Average Density of a Column stage, with the thin water film travelling at a fraction
of Steam of the speed of the steam, the correct friction effect
The dryness of the column increases uniformly upwards from could possibly be calculated. Such calculations would
zero at the bottom. Steam pressure 300 lb. per sq. in. abs.
be very difficult.
numbers. The Reynolds number for any fluid flowing Another approach would be to compare the friction
in a pipe is with the heat transfer coefficients. Reynolds in 1874
DVS showed that the flow of heat and the pressure loss due
RE=- to friction, of a fluid flowing in a tube, are both intimately
z connected with the turbulent particle paths and eddy
where Z is the viscosity in pounds per hour-foot. currents in the fluid. An increase in the heat transfer
In the U-tube, it is shown later that the water velocity coefficient should indicate an increase in the pressure
in the downcomer lies between 3 and 7 ft. per sec. drop due to friction.
For any particular water velocity, the value of V S It has been pointed out, that in the boiling section of
for steam is of course the same as for the water. The the tube, the heat transfer coefficients are greater than
diameter D is 0.104 feet. The value off for water and for for the non-boiling water section. Without attempting
dry steam has been calculated for water velocities of to evaluate the increase, it would seem that the friction
3, 5, and 7 ft. per sec., for a range of pressure of from in the boiling section should generally be more than in
100 to 3,000 lb. per sq. in., and lies practically always the water section. In view of the uncertainty of the
between the limits 0.004 and 0-005. An approximation actual conditions, however, the simple formula already
is therefore made, in taking f = 0.005 throughout the derived will be used throughout the whole tube.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

156 T H E C I R C U L A T I O N O F W A T E R A N D STEAh'I I N W A T E R - T U B E B O I L E R S
The values of the coefficient of friction shown in 7 4L -4)
Hence pressure per unit area = where is
Fig. 7 are for isothermal flow, when no heat is being
~

ag
transmitted through the tube wall to or from the contents the cross-sectional area of the tube.
of the tube. McAdams (1933) showed that if heat is being w(v-T)
transferred from the tube to the contents, the general Head of water to produce acceleration = -
effect is to reduce the coefficient of friction. W. F. Cope asg
S
(1936) brought forward the theory that the friction co- Since V = v- and TL' = as^, this may be written
efficient under heat transfer conditions is the same as S
that under isothermal conditions if the viscosity in the Head =-7-( I '- 7')
former case is that pertaining to the wall temperature g
of the pipe, which is of course greater than the core
temperature, with corresponding reduction in viscosity.
Hence the values of the friction coefficient used here,
which neglect this reduction, are on the safe side, that is,
they exaggerate very slightly the friction effects.

Again, the values o f f are given for clean pipes. I n 1

and since -=
A
S
?'&Y+ Vfu
modem boilers, fed with scrupulously clean and air-free
water, no scaling takes place, and the tube remains clean 1
indefinitely, except for slight corrosion due to inspection and - = I/?'
S
at regular intervals. There should therefore be little STi.2
reason to expect any appreciable increase in the friction Head = -' ,( I - - [ ',,)-U+
[/-?& - T;:,;
coefficient with age. ,g
_-
-
S'c.2
(Vs-Vw)X
Pressure Drop to Produce Acceleration. During the 6
process of evaporation, when water of density s and This head will be referred to as the "acceleration
having velocity v evaporates to steam of density S head".
and velocity V , the contents of the tube accelerate. I n view of the considerable uncertainty which exists
-4 pressure drop is required to produce this accelera- regarding the head required to produce acceleration,
tion. The simplest method of calculating the pressure this expression will be obtained by another method,
required is to find the change of momentum per second. by considering the force required for the acceleration
Let w be the weight of water flowing per second. Then of each particle of steam. For this calculation, reference
may again be made to Fig. 5, p. 154, and the symbols
Accelerating force = Change of momentum per second used are given below :-
=-
zu( v- v ) v Velocity of water at Q.
g I/' Velocity of steam at R.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E C I R C U L A T I O N O F WATER A N D S T E A M I N W A T E R - T U B E B O I L E R S 157
0 Dryness at Q. circulation, that the velocity head of the steam issuing
X Dryness at R. from the tube is a measure of the acceleration head is
V,, Volume of 1 lb. of water. entirely erroneous. I t is true that if water issues from,
1: Volume of 1 lb. steam. say, a jet in the side of a tank, the velocity head of the
x Dryness at T. water, v2/2g, equals the head of water in the tank
s+dx Dryness at U. above the jet. The kinetic energy of w Ib. of the issuing
densitv at Q water is wv*/2g. This exactly equals the loss of potential
The velocity at T = o x energy, by w lb. of water falling a distance h. The kinetic
density at T
energy is derived from the loss in potential energy,
and wh = w 4 2 g ; or h = v2/2g.
In the steaming tube, however, the kinetic energy
V
Velocityat U=- {( Vs- V , )(x+dv)+ V,> of the steam issuing from R is only in part derived from
Vw the pressure of the water at Q. The acceleration head,
per pound of water changing to steam, is v(V-v)/g.
Hence the increase in velocity = - ( Vs- V,) dx
V

VZ” The increase in kinetic energy is

The time required for a particle of steam to travel
TU
from T to U =
Velocity at T or U This increase in kinetic energy exceeds the acceleration
- dY head by
-
V
- W s - VWb+ V,> v 2 - v2
---- v( v - v ) - 1’2-29- 3, V+ 22,2
1
W
’
2g g %
Hence acceleration - v2-2vv+v2
-
- Increase in velocity 2g
Time
(V-V)2

- vW
-(V,-
V
V,) dx
V
-
Vw
{( Vs- V,)X+ L’w} =-
2g
dY This excess energy is derived primarily from the
The weight of steam in the section TU is given by pressure energy of the steam, which has been obtained
its volume multiplied by its density. directly from the fire, as a component of the latent heat
of evaporation.
UdY Experimental demonstration of the effect of the
Weight of steam in T U = acceleration may be made with the demonstration boiler
Vs- vw>x+ V w
(Fig. 2, p. 148). If the water level is arranged about
half way up the glass tubes (the top tank being empty),
and heat applied to the water in one leg -4, causing
evaporation, there will be a surging upwards of wet
steam in that leg. The reaction downmards increased
the pressure in the bottom drum, indicated by the water
level in the unheated tube B surging upwards also.
Another elementary experiment to show the need
for acceleration head in a tube may be made with the
V2 apparatus shown in Fig. 8. This consists simply of a
Pressure per unit area =-( V,- Vw)dx
Vw2g measuring jar A, to the bottom of which a length of steel
tube 4 inch outside diameter is fitted. The tube is bent
Head of water to form a “U”, the upcomer leg terminating as shown,
well below the level of the jar. Cold water is allowed to
flow from the jar, discharging to waste at B. The time
for the water level in the jar to fall between two marks
is noted. The whole head AB is used up almost entirely
Over the whole tube from Q to R, this head is required in overcoming friction in the tube.
for every small increment dx in the dryness fraction. An intense flame is applied to the tube at B, convert-
Hence, between Q and R, ing the water into steam which issues with considerable
m2
velocity. The use of a concentrated heat is intended to
Total acceleration head = -(V,- Vw)X minimize the extra friction which could be set tip if this
i? steam had to flow for any appreciable distance along
which is the same as that,obtained previously. the tube. T h e time now needed for the water level to
The theory, frequently put forward in discussions on fall between the same two marks is found to be 50 per

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

158 T H E C I R C U L A T I O N O F WATER A N D STEAM I N WATER-TUBE BOILERS
cent more than previously. With the reduced velocity points in the U-tube will be greater. If there is no flow,
in the tube the friction head is less, the difference being the excess pressure at the bottom of the U-tube will
accounted for by the acceleration head required to give equal the head of water multiplied by the density of
the evaporating steam its extra velocity. water. If there is flow, the excess pressure will be less,
Another way to show that the kinetic energy of the as some of this head is used up in overcoming frictional
steam is not entirely derived from a loss of potential resistance in the downcomer.
energy of, say, 1 lb. of water falling through a height In the upcomer, there is a pressure drop due to the
equal to the velocity head, is to notice that in the elemen- increasing height above the bottom of the “U”, to
tary experiment just described, the jet of steam is found frictional resistance and to the necessity for providing
to rise much higher than the level of the water in the the accelerating force for the evaporation. As the pressure
jar. Neglecting friction, a drop of water falling from A drops, the saturation temperature falls. If the water
to B (Fig. S), and losing a corresponding amount of at the bottom of the tube is at the saturation temperature
potential energy, would have an equivalent kinetic corresponding to its pressure there, evaporation of
energy and velocity. As soon as this drop in the steam part of the contents of the tube will take place by reason
jet reached the level of A, its net loss of potential energy of this pressure drop alone, independently of any heat
is zero, so that its kinetic energy would need to be zero. from external sources. This form of evaporation has
been termed “self-steaming”, or ‘%elf-evaporation”.
The steam is not generated at the tube wall, but within
the body of the fluid, just as when steam is drawn down
from an accumulator.
Schmidt (1933) made an experiment to illustrate
MEASURING FLOW self-steaming, using a U-tube in which the water in the
downcomer was heated sufficiently to cause it to rise
in temperature from that of saturation in the top drum
to that of saturation in the lower drum, without gener-
STEEL TUBE ating any steam; it then rose in the unheated riser
EXTERNAL DIAMETER with considerable steam generation. He made calcula-
S L O “ LONG tions to show that the influence of self-evaporation is
considerable, and may materially exceed the steam
generation due to heat transmission through the walls
at very low and intermediate pressure. He considered
that circulation is aided by self-evaporation.
At normal boiler pressures, however, it would appear
that self-evaporation, though present, has but little
appreciable effect. In the calculations made in a later
section for flow in the standard U-tube, it is shown that
the excess pressure at the bottom of the “U”, over the
pressure in the drum, is of the order of 4 or 5 lb. per
Fig. 8. Apparatus to Demonstrate Acceleration sq. in. At 300 lb. per sq. in., in the example calculated
Head in detail on pp. 159-61, the excess pressure at the bottom
of the “U” is 4.65 lb. per sq. in. The total heat of water
at 304.65 lb. per sq. in. is 1.54 B.Th.U. more than the
But the high velocity of the actual jet, at the level of total heat of water at 300 lb. per sq. in. Dividing by the
A, shows that there is some source of kinetic energy latent heat, 815.2 B.Th.U., this extra heat would
other than the head AB. suffice to evaporate 0.00189 lb. of water ;that is, the extra
It is, of course, true that all the energy of circulation dryness at the top of the upcomer, due to self-evapora-
comes from the fire. The kinetic energy of the steam, tion, is 0-00189. Thus, at first glance it would appear
derived from the pressure energy of evaporation, comes that there would be extra buoyancy and greater circula-
directly from the fire, without reference to the forces tion a t the lower pressures due to this self-steaming.
in any other part of the tube. The various resistances The correct viewpoint, however, to be taken in later
to flow, friction, and pressure drop to accelerate the tube calculations, is that dryness of steam issuing from the
contents on evaporation, are derived first from the tube has been calculated in terms of the properties of
potential energy of the water in the drum, that is, from steam at a nominal pressure in the drum. The extra
the head causing circulation. This potential energy pressure in the lower parts of the tube means that the
itself comes from the pressure energy of the steam, point of commencement of evaporation Q (Fig. 4, p. 152)
which has caused the tube contents to rise again into the is raised slightly, since the water has to flow farther
drum by the expansive forces of evaporation. along the heated part of the tube from P, to receive
extra sensible heat. This extra sensible heat is converted
Pressure Vuriutions in the Tube. Whilst the pressure to latent heat as the steam rises to R. The dryness of the
of the steam in the drum is the nominal pressure as steam is everywhere slightly less than if this pressure
shown on the steam gauge, the pressure at various effect did not exist. Its buoyancy is slightly less. T h e

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS 159
velocity is slightly less, so that the friction is slightly of the water and steam. I t will not affect results appre-
less in consequence, helping to balance matters. ciably if the friction loss at the bend at the bottom
Not only does water give up some sensible heat on of the “U” is neglected. The pressure drop due to
reduction of pressure, but steam formed at the higher friction, per foot length of tube, is given by the formula
pressures near the bottom of the “U” will expand already derived.
adiabatically as the pressure drops. The magnitude of
these effects is, however, so small at normal boiler
pressures, that they may be neglected here.
Another small point is the question whether there is This involves a knowledge of the dryness fraction x
any increase of pressure in the drum, over the pressure of the steam, and of the velocity v of the water.
at the top end of the upcomer tube, due to the re- (4) It must provide the necessary pressure drop to
conversion of the kinetic energy of the wet steam into produce the acceleration as the water evaporates into
pressure energy, when the steam comes to rest in the steam. This has been conveniently termed “acceleration
drum. There will indeed be a pressure on the elastic head”.
cushion of steam in the drum, or on the shell of the steam The equation for this, as already derived, is
drum itself, due to the impact of the column of rapidly
moving steam. sv*
I n hydraulics, it is found that the kinetic energy of Acceleration head = -( V,- VW)X
g
water issuing from a tube into a drum, below water
level, is not reconverted into pressure energy, but is where s is the density and v the velocity of the water,
wasted in forming eddies, and is thus converted to and X is the dryness at the top of the upcomer. It must
heat. The only practical means of securing the conversion be repeated that a large part of the kinetic energy, or
of a part of the kinetic energy into pressure energy, the velocity head, of steam, is not derived from the
is to use a gradually diverging tube, which is quite available head of water, but is part of the heat energy
unlike the method by which the steam is discharged obtained from the fire.
into the drum. Thus to calculate the head causing flow, and the
With steam there are further reasons why it is unlikely resistance to flow, the rate of heat absorption and the
that any appreciable rise in pressure, due to loss of water velocity must be known. Equilibrium will be
momentum, can be obtained. The jet of steam issuing attained if the available head equals the sum of the four
from the pipe loses its kinetic energy by forming eddies requirements : entrance loss, velocity head of water,
in the steam drum. This kinetic energy is changed to friction head, and acceleration head.
heat, which serves to dry the steam an infinitesimal T o show how the head available is utilized in these
amount, so small is the heat equivalent of the kinetic four ways, a numerical example will be calculated for
energy. The infinitesimal increase in volume, due to this the standard U-tube. The tube is shown in Fig. 1
increased dryness, would cause a very slight rise in (p. 148). As already described, the drum is 4 feet in dia-
pressure, and the whole effect may be neglected. meter and is half full of water. The tube is 27 feet long,
each leg being 12t feet, whilst the upcomer projects 2 feet
Calculation of Flow. For any assumed rate of into the drum. The tube is l-) inches external diameter
absorption of heat by the contents of any part of the and 14 inches internal diameter. The cross-sectional
U-tube, the head producing flow may be calculated by area is 1.23 sq. in. For this example, the drum pressure
the .method described above. The requirements which is taken as 300 lb. per sq. in. abs. The average rate of
this head has to fulfd are four in number. heat absorption over that part of the tube where latent
(1) I t must overcome the pressure drop at entrance heating is taking place is assumed to be 50,OOO B.Th.U.
to the tube. This may be written 3V2,/2g, where CL, is the per sq. ft. per hr., based on the outside area of the tube
velocity of the water in the downcomer. surface.
(2) It must provide the water with its kinetic energy. For this example, the point of evaporation has been
If v is the velocity of the water, then the kinetic energy assumed to have been located at Q (Fig. 4), 10 feet
is m 2 / 2 ~ and
, the head to produce this energy is $12~. below the drum on the upcomer leg. The point R, at
There is, of course, no relation between this kinetic which the steam ceases to absorb heat from the fire,
energy of the water, which must necessarily be derived is taken at the underside of the drum. The heat absorbed
from the available head, and the kinetic energy of the between Q and R is
steam, which, as already explained, is derived mainly
from the fire. There is a definite pressure drop between
the water in the bottom of the drum and the water in
3,600 ’-
12
~ = 54 B.Th.U. per sec.
509000 It’ 3 ’ 1 4 10

the immediately adjacent part of the downcomer tube, This is sufficient to evaporate 541815.2 = 0.067 lb. per
to give the water its velocity in the downcomer. This sec.
pressure drop is not regained anywhere. The kinetic It is required to find the value of v , the velocity of
energy of the water is finally dissipated, along with the water in the downcomer, which will make the available
kinetic energy of the steam, in forming eddies of steam head at that speed, equal the head required (also at that
in the steam drum. speed). As there are so many variables, whose relation-
(3) It must overcome the frictional resistance to flow ship is so complicated, the most feasible way to solve

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

160 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS
the problem is to assume various values of z', say, 3, 5, sv2
and 7 ft. per sec., and calculate, for these values, the The acceleration head is -( V,- TTr,)X.
R
available head and the required head. Then, by plotting If r = 3 ft. per sec., this equals
two graphs, showing how available head and required 52.63 x 32
head vary with the velocity it is possible to obtain, at (1*552-0*019) 0.05
the intersection of the graphs, the value of 'u which 32.2
produces the required equilibrium. = 1.12feet
If v = 3 ft. per sec., the weight of water flowing per If v = 5 or 7 ft. per sec., the acceleration head is 1.87
second is or 2.63 feet respectively.
1-23
52.63 x -x 3 = 1.34 lb.
144
The evaporation, as already calculated, is 0.067 lb.
per sec. Hence the dryness fraction at the top of the
upcomer is 0.067/1.34 = 0.05 = X . The average density
of the column, from Fig. 6, is 21.2 lb. per cu. ft. The
density at the top is 10.5 lb. per CU. ft.
The available head is due to the difference between
the weight of a column of water 12 feet high, and a
column of steam 12 feet high. T h e lower 10 feet of the
steam column has average density of 21.2 Ib. per cu. ft.
and the upper 2 feet has density of 10.5 Ib. per cu. ft.
Difference = [ 1 2 ~ 5 2 . 6 3 - ( 1 0 ~ 2 1 * 2 + 2l0*5)]
x area
= 399 x area.
Therefore head available = 399/52*63= 7.60 feet.
For values of v of 5 and 7 ft. per sec., the available
head, calculated in the same way, is 6-32 feet, and 5.50
feet respectively. These values are plotted in Fig. 9 as
curve AA.
Considering now the four resistances which this head
has to overcome, the first-loss at entrance-may be
taken at +2/2g. For a value of v , equal to 3 ft. per sec.,
this amounts to 0.07 feet. For values of v of 5 and 7
ft. per sec., it is 0.20 and 0.38 feet respectively.
Second, the velocity head of water is given by
'u2/2g. Hence for velocities of 3, 5, and 7 ft. per sec.,
the velocity head of the water is 0.14, 0.39, and 0.76
feet respectively.
The friction resistance per foot of pipe is
4X0.OO5 v2[l+x(
'/= 6 4 . 4 ~ 0 . 3 0 4 1-552-0.019
0.019 )I Fig. 9. Available Head and Required Head, Plotted
+
= 0.00298 9 2 (1 80.5~) Against Velocity
If X is the dryness fraction at the top of the upcomer Adding together the various losses :-
it follows that the dryness fraction from the top of the Value of 'o,
downcomer to Q, a distance of 15 feet, is 0, from Q ft. per sec. Head required, feet
to R, 10 feet, it averages & X ,and from R to S, 2 feet, 3 . . 0~07+0.14+1.49+1*12= 2.82
5 . . . 0.20 +0.39 +3.28 +1-87 = 5.71
it equals X . Hence over the whole tube the average 7 . . . 0.38+0.76+5.72+2.63 = 9.19
1 7
dryness is -(lOX/2+2X) - X .
27 -zi These values are plotted in Fig. 9, as curve BB.
For the water velocity of 3 ft. per sec., X has been The intersection of the two curves at C shows that when
found to be 0.05. Hence, for the whole tube the friction the water velocity is 5.3 ft. per sec., the head producing
head is flow equals the head required, and amounts to 6.20 feet.
7 This is the water velocity for equilibrium.
27 x 0.00298 x 321 1 +(SO3 x zix 0.05 ); The weight of water circulating per second is
= 0.725 (1 +1.04)
= 1.49 feet
If el = 5 or 7 ft. per sec., the friction head, calculated The dryness fraction at the top of the upcomer is
in the same way, is 3.28 and 5.72 feet respectively. 0*067/2*40= 0.0278.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS 161

The entrance loss is +(5‘32)= 0-22 feet = 3.5 per cent. the top of the upcomer indicates that the dryness is
64.4 very low for low rates of heat absorption. Throughout
5.32 the range of heat absorption shown on the graph,
The velocity head of water is - = 0.44 feet = 7.1 per the tube should operate satisfactorily, and with adequate
cent. 64.4
flow to prevent overheating.
The total friction head A second analysis of the U-tube boiler may be made
by calculating the flow for various pressures, assuming
7
+
= 27 x 0.00298 x 5.32( 1 80.5 x 27 x 0.0278) the rate of heat absorption to be constant, and the posi-
tion of the point of evaporation to be steady. Values
= 2-26(1+0-58) have been calculated, for pressures between 15 and
= 3.64 feet = 55.7 per cent 3,000 lb. per sq. in. abs. The point of evaporation is,
as before, taken to be on the upcomer, 10 feet below
The acceleration head is 52.63 5*32x 1.533 x 0.0278 the drum. The rate of heat absorption over that part
32.2
of the tube absorbing latent heat is taken as 50,000
= 1.90 feet = 30.7 per cent B.Th.U. per sq. ft. per hr.
The total head is therefore 0.22+0.44+ 3*64+ 1-90
= 6.20 feet. I I I I I I”
The friction head required, if water only had been
flowing through the 27 feet of pipe at 5.3 ft. per sec.
would be 27 x 0.00298 x 5.32 = 2.26 feet. T h e extra
friction due to evaporation is 1.38 feet, or 22.3 per cent
of the total head.
Since the dryness at the top of the upcomer is 0.0278,
or &, this is therefore what is popularly termed a
“36 times round” boiler, that is, any pound of water
has to circulate 36 times round the tube before it is
completely evaporated. This shows more than ample
flow to prevent any danger of overheating the tubes,
even though only natural head has been available,
and though the rate of heat absorption has been quite
high.

Analysis of U-tirbe for Varying Conditions. Having

derived the above method of computing the flow in the Fig. 10. Flow in Standard U-tube Boiler, with Various
U-tube, for a definite rate of latent heat absorption Rates of Heating
over a definite length of tube, an analysis may be made Pressure 300 Ib. per sq. in. Point of evaporation on upcomer
of the operation of the U-tube boiler for other rates 10 feet below drum.
of heating, for other pressures, and for alternative
positions of the point of evaporation. Each of the values The results are plotted in Fig. 11. T h e head causing
for water velocity, in the following graphs, has been flow is given in terms of feet of water, having a density
obtained by a method similar to the method just corresponding to the relative pressure. T h e head dimin-
described. ishes throughout the range of pressures, due mainly to
Investigating first the effect of varying the rate of the decreased difference of density between water and
heat absorption, i t is assumed that the point of evapora- steam at high pressures. The velocity of water in the
tion remains on the upcomer at a point 10 feet below downcomer is a maximum at 300 lb. per sq. in. At
the drum, as in the example worked above. The values higher pressures, the velocity falls off slightly, though
of the water velocity, the dryness fraction at the top of not so rapidly as the head causing flow, since the accelera-
the upcomer, and the head causing flow have been tion head and the extra friction due to increased speed
calculated for rates of heat absorption of 1,000, 5,000, of steam are less for high pressures than for low pressures.
10,000, 30,000, 100,000, and 150,000 B.Th.U. per It would appear, from the graph, that the tube would
sq. ft. per hr., whilst the calculation already described give satisfactory operation, even up to pressures of
deals with a heat absorption of 50,000 B.Th.U. per sq. 3,000 lb. per sq. in. T h e increased dryness of the steam
ft. per hr. The results are shown in Fig. 10. is mainly due to the reduction in latent heat at high
The first notable point is that the velocity of the water pressures. Although the curves are not extended for
supply to the tube has a maximum value of about pressures above 3,000 Ib. per sq. in., rough calculations
5.4 ft. per sec. when the rate of heat absorption is 45,000 show that even at pressures above the critical pressure
B.Th. U. per sq. ft. per hr. T h e velocity falls off slightly (3,226 lb. per sq. in.) the velocity of flow does not fall
with increased heating, though even with three times off seriously.
the heat the velocity is still 4.1 ft. per sec. This is in direct contradiction to the common theory
With lesser heat the velocity is less, but this is not that natural circulation cannot exist at pressures equal
important. The graph showing the dryness fraction at to, or above, the critical pressure. At the critical pressure
14

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

162 T H E C I R C U L A T I O N O F W A T E R A N D STEAM I N W A T E R - T U B E B O I L E R S
the steam has the same density as water. But, since for points of evaporation 24, 20, 15, 124, 10, 5, and 0
the latent heat is zero, the steam on formation will feet from the top of the downcomer tube. The results
immediately become superheated, and its density are plotted in Figs. 12-14. These graphs enable the
would diminish below that of water. For example, at stability of the flow to be examined.
3,400 lb. per sq. in., the density of water at 720 deg. F.
is 17.04 lb. per cu. ft., and the total heat per pound is
1,001.2 B.Th.U. With a rate of heat absorption of
50,000 B.Th.U. per sq. ft. per hr., the total heat absorbed
per second, over the 10-foot length QR, is 54 B.Th.U.
If the water velocity is 3 ft. per sec., the flow
is 0.435 lb. per sec., so that the heat absorbed is

~ POINT OF EVAPORATION, /-FEET

Fig. 12. Analysis of Conditions in Standard U-tube
Boiler
Various positions of point of evaporation. Pressure, 300 lb. per
sq. in.Rate of heat absorption, 10,000 B.Th.U. per sq. ft. per hr.

PRESSURE-LB. PER SQ. N A B S .

Fig. 11. Analysis of Conditions in Standard U-tube

Boiler : Various Pressures.
Point of evaporation on upcomer 10 feet below drum. Rate of
absorption of latent heat, 50,000 B.Th.U. per sq. ft. per hr.

54/0*435= 125 B.Th.U. per lb. This increases the heat

per pound to 1,126 B.Th.U., with density 9.6 lb. per
cu. ft. The average density over QR is taken as 13.0.
The head causing flow is found to be 3.2 feet. With
water velocities of 5 and 7 ft. per sec., the head is found
to be 2.23 and 1.70 feet respectively. The various
resistances are calculated as before, and the equilibrium
is found to occur when the water velocity is 4.5 ft. per
sec., which is slightly more than for a pressure of 3,000
lb. per sq. in.
Values for pressures down to 15 lb. per sq. in. are
POINT OF EVAPORATION, /-FEET
shown, mainly to indicate the effect of running experi-
ments at atmospheric pressure. The velocity in the Fig. 13. Analysis of Conditions in Standard U-tube
downcomer at this pressure is about one-quarter of that Boiler
at 300 lb. per sq. in. It would appear that whatever Various positions of point of evaporation. Pressure, 300 lb. per
water velocities are noticed in the downcomer of the sq. in. Rate of heat absorption, 50,000 B.Th.U. per sq. ft. per hr.
demonstration apparatus, a velocity three or four times
as great might be expected in a normal high-pressure For example, if the point of evaporation is 15 feet
boiler. from the top of the downcomer at Q (Fig. 15), and the
A third analysis of the U-tube boiler is made to find rate of heat absorption is 50,OOO B.Th.U. per sq. ft.
how the water velocity varies, if the point of evaporation per hr., the velocity as shown by point M on Fig. 13
is located at different parts of the tube. Values have is 5.3 ft. per sec. The rate of heat absorption by the
been calculated for rates of heat absorption of lO,OOO, contents of the tube is 5.4 B.Th.U. per foot length of
5O,OOO, and 100,OOO B.Th.U. per sq. ft. per hr., and tube per second. The flow is 2.38' lb. per sec., hence the

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS 163
rise in temperature per foot is 5.4/2.38 = 2.26 deg. F., of evaporation will rise above Q again. This surging
for that part of the tube where sensible heating is taking will go on until the point of evaporation settles again
place, i.e. PQ. Suppose that the temperature of the at Q. The position of the point of evaporation is in
water in the drum is, say, 6 deg. F. below saturation, stable equilibrium. If any temporary displacement
then PQ will be occurs, forces are set in action tending to restore the
612.26 = 2.7 feet original position of the point of evaporation. There may
indeed be a continuous surging or fluctuation in the
Therefore PQ = 5.9/2.18 = 2.7 feet. level of the point of evaporation. Yet if the average point
Time to flow distance PQ = 2.7/5.3 = 0.51 second. of evaporation is on the upcomer, the flow is stable.
The water thus requires to flow for 0.51 second If the point of evaporation is on the downcomer,
through the heated part of the tube, in order to receive at K, say, a different result is obtained. The vital point
its sensible heat, before evaporation begins. of the analysis is revealed in Fig. 14. For values of 1
Now, if some temporary reduction takes place in the between 6 and 10, the velocity graph is a straight line
rate of heat absorption for this tube, the level of the pointing directly to the origin. if i n i
slight displacement of the point of
-0.32
evaporation takes place, there is no
tendency to return. If K rises to K,,
-0.28-
the velocity decreases in proportion
(Fig. 15). T h e water still has the
same time, before reaching K,, to
--R
- 0.24 I2 obtain its sensible heat, as it pre-
viously had to reach K.
Thus for this rate of heating, the
- 0.20 10 flow is unstable for any position of
VI
In
W t; the point of evaporation on thedown-
z W
comer. Any slight displacement is
20-16
0
*: s immediately extended. Any surging
n 9 tends to blow the water out of the
2 tube altogether and circulation ceases.

E
(I

-0.12 6 U
z The same tendency is shown in Fig.
VI
3 13. Though the graph does not point
6
4u
directly to the origin, the restorin?
n tendencies are so weak that danger -Q I
6 -Q
Y
I
exists that circulation will cease.
1 Thus, for high rates of heating, it is
dangerous to have a heated down- -P
comer. The circulation will be faulty.
0 T o ensure that this does not take
5
DISTANCE FROM TOP OF DOWNCOMER TO place, it is advisable that the down-
POINT OF EVAPORATION, I-FEET
comer tube should not be heated at Fig. 's. Of
-11
all.
Flow in Standard
Fig. 14. Analysis of Conditions in Standard U-tube Some further reasons for not heat- U-tube
Boiler ing the downcomer are discussed
Various positions of point of evaporation. Pressure, 300 Ib. per later, but results obtained with the demonstration boiler
sq. in. Rate of heat absorption, 100,000 B.Th.U. per sq. ft. (Fig. 2, p. 148) may be mentioned here. For this experi-
per hr. ment the central tube B was plugged up and not used.
With tube A only being heated, circulation is absolutely
point of evaporation will rise to Q1-say, 6 inches satisfactory. If heat is now applied gently to tube C,
above Q. The stability of the flow will be decided by bubbles of steam are seen to form on the inner tube,
what happens when the original heating is resumed. and to be dragged down by the current of water. If
Once again the rate of heat absorption is such, that the greater heat is applied the bubbles increase until their
water gets all its sensible heat after flowing for 0.51 buoyancy is sufficient to stop the flow. A very slight
second. further increase is enough to upset the flow completely,
The water velocity corresponding to Q1, from the with violent surgings in both tubes. At times the inner
graph, at M, is 5.2 ft. per sec. The water will now tube is seen to be quite dry. Though only rough
obtain its required sensible heat after moving 0.51 x 5.2 measurements were possible, this interruption was found
= 2-65 feet. The point of evaporation will sink to a level to take place if the downcomer is heated even to one-sixth
below Q. of the amount by which the upcomer is heated. This
The equilibrium velocity for this new point is higher figure is of the same order as the ratio between rates of
than for Q, so that the water would now have further heating of fire row and back tubes of an inclined tube
to flow before receiving its sensible heat, and the point boiler.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

164 THE CIRCULATION O F WATER A N D STEAM I N WATER-TUBE BOILERS
Throughout these calculations it has been assumed than if it rises in a dry tube, owing to the momentum
that the position of the point of evaporation has been given to these entrained drops. T h e flow in the tube
known. I n practice, this could be determined only will be adversely affected.
approximately, by the rules already described, having Moreover, a violent turbulence is set up in the drum,
in mind the ratio of sensible to latent heating and the causing steam and water to mix up into a froth. Engineer
variations in the rate of heating at different parts of the Captain Dight (1933, 1936), using several water gauges
tube. Yet the various approximations should be near fixed to different parts of the upper and lower drums
enough to enable a reasonable estimate of the flow of an Admiralty boiler, estimated that at high rates
in the tube to be made. T h e dimensions of the tube the density of the contents of the drum was only three-
taken in the above example have been quite arbitrary, fifths the density of water. This has an important effect
and no attempt can be made to analyse all the possible on the supply of water to the downcomer, since the
variations in diameters and lengths. With other sizes reduced density reduces the weight of water flowing,
of tubes other outputs could be obtained, adhering and increases the dryness of the steam in the upcomer.
to the same limit of maximum dryness ; the object of These effects were clearly shown in the demonstration
this whole section is, however, to show that it is possible boiler (Fig. 2). With a drowned discharge to the upcomer,
for a U-tube to be designed to give adequate flow and turbulence in the top drum was excessive ; with a dry
safe operation with natural head. discharge there was no turbulence.
A further experiment was made in which R slight
restriction was placed at the top of the downcomer,
PART 111. S O M E M O D I F I C A T I O N S O F T H E to act as a rough metering device. With a drowned
STANDARD U-TUBE BOILER
discharge it was possible to evaporate more water than
It has been shown in parts I and I1 that the circulation could flow through the restriction. But with dry dis-
in the standard single U-tube is perfectly satisfactory charge and the same evaporation there was always an
for a very wide range of pressures and rates of heating, adequate supply of water through the restriction to the
with only one proviso, namely, that the downcomer downcomer.
should not be heated. T h e danger of drowned discharge arises again at high
Several modifications of the standard U-tube will now ratings when the downcomer is heated, and when the
be considered in turn, in order to examine the effect on water in the downcomer may be blown out of the tube.
the flow. Modifications are introduced in practice, due Consider what then happens. T h e tube is empty
to practical requirements for heating and construction. of water and full of wet steam. There is an up-current
Existing boilers consist of various arrangements of drums, of steam in both legs of the “U”, which continues
headers, and tubes, in which a simultaneous solution until all the steam is dry saturated. When the steam flow
of the problems of circulation, heating, and construction ceases there will be a flood of water down the tubes,
is aimed at by trial and error. and very shortly this will have received its few degrees
of sensible heat and will explode into steam.
( 1 ) Use of Several U-tubes Attached to One Drum. If the discharge of the upcomer tube is dry, this newly
If a U-tube boiler consists of a drum with several formed steam has an assuredly free escape in one direc-
U-tubes in parallel attached to it, and if these tubes tion, so that the effects of the rapid explosive forces
are heated at different rates, the position of the point are reduced, and there is more likelihood of stable
of evaporation will vary in each. T h e temperature flow being once again achieved, should the conditions
of the water in the drum will be uniform. Hence the of heating alter to favour this. But if both upcomer
water has to travel farther in some tubes, which are less and downcomer have a drowned discharge, there will
intensely heated, than in others, before obtaining be a flood of water in both tubes. T h e esplosice forces
all its sensible heat. T h e effect will be to raise the of evaporation will be doubled in intensity, and any
point of evaporation higher on the upcomer for the forces tending towards resumption of a steady flow
lesser heated tubes, and to lower it on the intensely will have the greater disturbing forces to overcome.
heated tubes. T h e danger of interrupted circulation This state of affairs was frequently observed in the
would be aggravated when a boiler consists of several demonstration boiler. Once circulation was upset
U-tubes in parallel, heated at different rates. I n the by overheating the downcomer, it could be restarted
intensely heated tubes, the point of evaporation might only by shutting off heat to the downcomer completely,
easily be in the downcomer, with consequent instability. and by making a fresh start to the experiment.

(2) Use of (‘Drowned” Discharge for the Steam. By (3) Use of a Bottom Drum. By replacing the U-bend
this is meant the discharge of steam into the steam at the bottom of the U-tube by a drum, as in Fig. 16,
drum below the water level. I n all discussions hitherto, though still retaining only the single downcomer and
this point has been avoided by stipulating that the dis- the single upcomer, the resistance to flow is increased.
charge should be above the water level. T h e kinetic energy of the water in the downcomer
There are practical objections to a drowned discharge, is dissipated in eddies in the lower drum, and at the
in that the current of steam passing upwards entrains entrance to the upcomer tube there must be a pressure
particles of water, and becomes wetter. T h e resistance to drop to restore this kinetic energy and to overcome the
the flow of steam must be greater if it rises through water loss at entrance.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS 165
If steam is being generated in the downcomer, the
kinetic energy of the steam is destroyed in the same
way. Although the kinetic energy of this steam, when
it was actually in the downcomer, was derived mainly
from the fire, and not from the head of water, the replace-
ment of this kinetic energy, when the steam commences
to flow up the upcomer, will entail a pressure drop.
If V is the velocity of the steam entering the upcomer
tube, the velocity head is V2/2g and the entrance loss
+V2/2g, in terms of fluid of the density of the steam S.
The head of water, of density s, is l+(V2/2g)S/s, and
since V = vs/S the head of water is l+(v2/2g)s/S.
The effects of this extra friction have been calculated
for the examples in which the rate of absordon of
latent heat is 50,000 I3.Ti.U. per
sq. ft. per hr., the pressure 300 lb.
per sq. in., and points of evaporation
at various points on the tube.
Fig. 17 shows the flow in the
standard U-tube compared with the
arrangement of two tubes and a
bottom drum. The reduction in flow Fig. 17. Analysis of Variations from Standard U-tube
is most serious when the point of Boilers
evaporation is in the downcomer, and Pressure 300 Ib. per sq. in. Rate of heat absorption, 50,000
the effect is to increase still further B.Th.U. per sq. ft. per hr.
the likelihood of instability in the
flow, since the curves point more and up above water level. Tf, for the moment, tube C is
more towards the origin, when the isolated, the arrangement of A and B is the same as in
resistance is thus increased. Fig. 16. Assuming the point of evaporation to be at E,
with a rate of absorption of latent
(4) Use of a Bottom Drum with heat by B of 50,000 B.Th.U. per sq.
Inclined Tubes. If the tubes are in- ft. per hr. and a pressure of 300 lb.
clined, say, at 45 deg., as in Fig. 18, per sq. in., the velocity of flow in A
the head causing flow is reduced, is found to be 5.4 ft. per sec., and
although the resistances are the same the hydraulic losses in A to be 1.77
as before. T h e effect of variations in

U the position of the point of evapora-

tion is shown on the lowest curve of
Fig. 17. The rate of heat absorption
and the pressure are taken as in the
Fig. 16. Two previous calculation.
Drums with When the point of evaporation is
Vertical Tubes in the downcomer, the graph points
still more directlv towards the oriein. - I

showing an increased tendency towards instability. Drums with

The head of water in the drum, above the entrance to Inclined Tubes
the downcomer, is greatly reduced in this design,
so that the liability of the contents of the tube to be
blown out is increased. W
The diagram shows the upcomer tube discharging between top and bottom drums is equivalent to
above water level. If, however, the steam discharges 14.5-1.77 = 12-73 feet of water.
below water level, causing the water and steam in the A momentary condition might exist in tube C, whereby
drum to be in violent turbulence, the adverse effect the tube contained 1.77 feet of dry steam (of negligible
on the supply of water to the downcomer would be weight), whilst the remainder of the tube is full of water.
greatly increased by the nearness of the downcomer The weight of the contents would just balance the
entrance to the water level. difference in pressure between the drums, so that the
circulation in A and B might continue as before, yet C,
(5) Use of a Bank of Tubes in Parallel. Fig. 19 though in equilibrium, would have no circulation at all.
shows an upper and lower drum, connected by a down- T h e 1.77 feet of steam in C need not be at the top.
comer A and an upcomer B, as in Fig. 16, together It may be locked in the middle, with water above and
with a third tube C, which may or may not be continued below, especially if the tube C has a drowned discharge.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

166 T H E C I R C U L A T I O N O F WATER A N D STEAM I N WATER-TUBE BOILERS
Even if a fourth tube D, shown dotted in Fig. 19, is are inclined, or even if they are nearly horizontal,
added, ostensibly to feed C, it does not follow that it joining headers at each end. T h e arrangement of water-
will do so. T h e three tubes -4, B, and D can together be wall tubes in parallel between horizontal headers is
in equilibrium and tube C still stagnant. another example. By arranging tubes in parallel, bad
If C is heated. this state of things could esist onl!- circulation results, with violent surges and possible
stagnation. Stagnation in any tube immediately permits
overheating of the tube to take place. T h e most effective
method of obviating this danger is to provide each
-++__-
\-I ~
7fb
Y-
.I
ci
steaming tube with its own individual downcomer,
as in the standard U-tube, to ensure that the head avail-
able will not be unequally divided between several
upcomers in parallel.
-l
A These conditions were reproduced in the demonstra-
I
I
tion boiler (Fig. 2). With heat on A, but no heat on
B or C, circulation is satisfactory. By adjusting the heat
D
1 supplied to C, it is possible to make the contents of C
stagnant, uncertain whether to move up or down, with
I
a large stationary bubble half way up the tube, the
inner tube surface being dry.
T h e use of bottom drums, or of headers, as distri-
butors, is thus likely to prove inefficient. I n addition,
I such drums are heavy, bulky, and costly, and increase
the overall width of the boiler. T h e tubes all require
0
.I extra expanded joints where they enter the bottom drums,
n
0
also hand-holes if they enter headers. Freedom for expan-
I sion and contraction is restricted, and racking strains
i
are set up due to unequal heating of the tubes. Cracking
iI
JI cE
!
of tube plates between the holes results. T h e drums
shield the lower ends of many tubes from the fire and
from the up-current of gases, so that the effective
heating surface is reduced. T h e original value of the
lower drums in acting as traps for sludge is of much
! less importance to-day, when practically pure feed
I water is the rule.

*
I
I
(6) High-Head Boilers. T h e principle on which
Fig. 19. Two Drums “high-head” boilers are devised is illustrated in Fig. 20.
with Several Tubes This shows the standard U-tube lengthened by the
in Parallel addition of 20 feet of tube on each leg, before attachment
to the drum. T h e latent heating section is QR, corre-
2’
momentarily, unless the heat is 0 sponding to the latent heating section of the standard
.I
concentrated on the section of C - 0
U-tube in Fig. 4.T h e rate of heat absorption is taken, as
I
containing the steam. Rut heat- in Fig. 11, as 50,000 B.Tl1.U. per sq. ft. per hr., ol-er
ing C will not necessarily cause the length QR.
an upward current. There may T h e apparent purpose of providing a high head is to
bea growth of the 1.77-foot steam increase the length of tube containing steam of maximum
bubble in both directions. T h e
top water goes upwards and the
?J.
El
dryness and minimum density, counterbalanced by an
equal length of downcomer tube containing water. The
bottom water goes downwards. .I n increase in head should increase the velocity of flow in
This will upset all the circulation i the tube, and hence improve the circulation. O n calcula-
in A and R, and as a result faulty tion, however, the benefit appears to be very much less
circulation is set up therein. Fig. 20. High-Head than might be expected. Calculations have been made
Heatine the downcomers mav
0
Boiler for various pressures up to 3,000 lb. per sq. in. These are
accentuate the conditions under plotted in Fig. 21, together with the values for the
which stagnation may exist. If the point of evaporation standard U-tube, reproduced from Fig. 11.
is somewhere on the downcomer A, the difference in Throughout the range of pressures, the head causing
head between the top and bottom drums may be much flow in the high-head boiler is seen to be approximately
less than 12-73 feet, so that tube C may be more nearly 200 per cent more than in the standard U-tube boiler.
full of steam, yet stagnant. T h e velocity of flow, however, is increased by not more
T h e conditions described in this section apply whether than 25 per cent. Most of the extra head is required to
the drums are vertically above each other, or if the tubes overcome friction in the added lengths of tuhing.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E C I R C U L A T I O N OF WATER A N D STEAM IN WATER-TUBE BOILERS 167
The adoption of high heads does not of itself increase immediately arises, Why is forced circulation necessary
the circulation appreciably, unless provision is made to at all 7 The power to drive the circulating pump map
reduce the friction in the extra connecting pipes. The amount to 1 per cent of the total power output of the
same considerations apply to water-wall tubes, which plant. The reason is that the long forced circulation tube
have a high available head, yet which frequently have is equivalent to several smaller tubes in series. More
defective circulation, due to inadequate connecting pipes water has to pass through the tube, hence the velocity
between the water-wall headers and the main boiler must be greater. The friction head increases in pro-
drum. Such conditions are known to have been portion to the square of the velocity, to the length of
ameliorated to some extent by increasing the number the tube, and to the exit dryness, all of which are much
greater than for the standard U-tube. The acceleration
head is also greatly increased. This head has usually to
be overcome solely by the pump, since natural head is
non-existent, or negligible.
The chief merit claimed for such boilers is that circula-
tion is positive. The claim may be justified in comparison
with many designs of natural circulation boilers, which
are defective by reason of their having heated down-
comers, tubes in parallel, drowned discharges, etc. Yet
the U-tube boiler is equally, indeed more, dependable
with only natural circulation and without accessory
pumping equipment.
The use of forced circulation may offer advantages in
permitting the tubes to be arranged round the floors and
walls of the combustion chamber in any desired way.
Yet boilers exist, having such tube arrangements, in
which the circulation is natural. There are, moreover,
the added difficulties of cleaning, inspection, and of
removal and replacement of the long and severely bent
and welded tubes, should these be necessary.
Any form of forced circulation, however desirable its
results might possibly be in surmounting the short-
comings of water-tube boilers in general use, cannot be
as efficient as natural circulation properly arranged, for,
as deduced from the foregoing analysis, natural circula-
tion is derived from the energy of the fire acting directly
on the water and steam. Friction losses re-appear as
heat, though at a lower pressure. Forced circulation
requires electrical energy, which in a modern power
station is seldom more than about one-quarter of the
heat energy of the fuel put into the boiler. This quarter
of the heat energy used in power for overcoming friction
in the pipes, re-appears as heat. The other three-quarters
is irrevocably lost in the condenser during the generation
of the mechanical power in the turbines.

I I I I I
(8) Field Tube Boilers. The so-called “Field” tube
0 500 1.000 1.500 1,000 1.500 3.m was invented about 100 years ago by Perkins, and irn-
PREIYJRE-LB. PLH SO. IN. ABS.
proved by Field in fire engine boilers some 80 years ago.
Fig. 21. Flow in High-Head Boilers Fig. 22 shows a fire engine boiler, of the type used in
1908, incorporating Field tubes. The Field tubes project
rather than the size of external pipe connexions between downwards into the combustion chamber. Each tube has
drum and headers. These necessities have, of course, an inner tube, which is fed with water through a conical
increased the cost and complication, and have added to funnel at its upper end. T h e steam, on formation, rises
the weight and the number of joints, enhancing thereby in the annular space between the outer and inner tubes.
manufacturing, erection, and maintenance charges. Boilers of this design gave very satisfactory results and
(7) Forced Circulation Boilers. I n the usual types of were noted for their ability to produce full-pressure
forced circulation boilers, recently introduced to ensure steam within a very few minutes of starting from cold.
better circulation, it is usual to employ very long tubes, h improved form of the Perkins-Field tube which
in which the contents may be either completely evapo- has been developed during the last 50 years, mainly in
rated, or evaporated to a fairly high dryness fraction. France, in naval, marine, and land water-tube boilers for
Comparing with the U-tube standard, the question all manner of service, is the Niclausse tube (Fig. 23). It

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

168 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS
consists of a plain tube closed at one end and leading Boilers of this design are noted for their meritorious
through headers at the other end to a single steam and performance under the most arduous conditions of firing.
water drum. Water enters each plain tube through an Captain H. Rial1 Sankey, M.1.Mech.E. (Past-President).
inner tube. The header is divided into two parts by a reporting (1914) on a test of high-duty Niclausse naval
Yertical partition. The right-hand part is full of water and boilers, of which some twenty-one were ordered for the
supplies the inner tubes. The outer tubes lead back to French battleship Beurn just before the war of 1914-18,
the left-hand side of the header which collects the steam confirmed efficiencies of 90 per cent and 78 per cent
from them, and delivers it above or near the water level respectively at normal and 100 per cent overloads,
in the drum, a spout being fitted to the top of the inside although the boilers had no superheaters and were hand-
compartment for that express purpose. fired (Niclausse 1914). The heat release per cubic foot of
combustion chamber at the overload rating was 228,000
B.Th.U., of 680,000 B.Th.U. per sq. ft. of “tube wall
area” located only 3 ft. 6 in. above the grate.
The hourly absorption rate was in that case of the
order of 185,000 B.Th.U. per sq. ft. of effective radiant
heating surface taken as the tube wall area. Reduced to

I
WATER A’iD
S T E A M DRUM the total circumferential area of two rows of the fire
’ tubes, and allowing additionally for convectional heating,
the heat absorption greatly exceeded the 50,000 B.Th.LT.

DIAPHRAGM

WATER

Fig. 23. Diagram of Niclausse Boiler

per sq. ft. per hr. rate which has been taken in the
foregoing calculations. Yet no trouble has been experi-
enced with overheated fire row or other tubes in those
boilers, although fed with salted water of sea salinity for
periods of 16 hours on test. Instances are on record of
this type of boiler in service using sea water for much
Fig. 22. Fire Engine Boiler, Showing Arrangement longer periods without distress.
of Field Tubes These remarkable performances are attained in such
boilers having a difference of level between drum and
Here there are unheated downcomers. There is a free bottom of header-for producing natural circulation in
discharge of steam into the drum. Moreover, there is an the lower tubes-of only some 5 or 6 feet.
individual supply of water to each heating tube. It may The original need for the somewhat complicated
seem possible that any water separating out at the header headers and tube joints required in this design form a
end of the upper tubes may tend to pass down the header minor disadvantage to-day, although on the other hand,
and obstruct the flow of steam from the lower tubes. it becomes exceptionally easy to inspect, withdraw, and
Observations of Aow, however, in a boiler working under replace any of the tubes, both inner and outer, without
pressure and having glass panels in the header, show that disturbing neighbouring tubes.
circulation and steam discharge are extremely satis-
factory. No chunks of solid water are carried into the (9) The Lewis Single-Drum Boiler of Rational Design.
drum-a commendable feature of this boiler which is An improved arrangement of the Perkins-Field tube is
confirmed by the fact that it delivers very dry and clean shown diagrammatically in Fig. 24. A single water and
steam under exceptionally heavy overloads. steam drum is provided. The heating element consists

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E CIRCULATION OF WATER AND STEAM I N WATER-TUBE BOILERS 169
of an outer tube projecting radially from the drum, at T h e flow may be calculated in terms of the water
some convenient angle of inclination, and closed at its velocity in the inner tube. A few variations in the equa-
lower end. The steam formed in this tube is conveyed tions are required, because of the special form of this
above the water level in the drum through a special tube. For the friction of water in the inner tube, the
compartment. An inner tube, leading from the water Reynolds numbers are lower than for the plain U-tube,
section of the drum, conveys the water to the lower end and hence the friction factor is slightly higher. T h e loss
of the outer tube. at'reversal, at the bottom of the tube, may be taken as
This design, when compared with the standard (v+ V)*/2gwhere v and V are the water velocities in the
U-tube, is seen to have the essential features which inner tube and annulus respectively. T h e friction in the
have been proved to be desirable and practical. I n a annulus is increased because of the reduction in hydraulic
boiler consisting of several such tubes attached to the mean depth, due to the presence of the inner tube. The
same drum, every tube forms an individual heating unit, head causing flow is determined by comparison of
with its circulation independent of the flow in any other
8 , I j 0.8
tube. Each tube has its own individual supply of water I I I

through its own inner tube. T h e inner downcomer tube

is unheated, and the annulus upcomer discharges its
steam above the water level in the trough. The steam on
SINGLE DRUM

DISCHARGE OF STEAM ABOVE WATER LEVEL, -I 2

-I 0

c,
-8k
I
z
OUTER TUBE OR v
9
HEATING ELEMENT
-6 4
v)
3
4
U
INNER FEEDING TUBE
-'5 0

Fig. 24. Diagram of Lewis Boiler B.TH.U. PER SQ. FT. PER HR. X lo3

reaching the upper part of the drum strikes suitable Fig. 25. Performance of Lewis Tube with Various
baffles, to make the water content separate out and to Rates of Heating
ensure that only clean and dried steam passes to the
length, 12 ft. 6 in. Pressure, 300 lb. per sq. in.
+
Tube, 14 inches outside diameter; inner tube inch diameter ;
main.
The performance of this tube may be calculated in the
same way as for the simple U-tube. A convenient size of densities in the inner tube and annulus, the height being
tube has an internal diameter of 14 inches, and an measured vertically.
external diameter of I t inches. T h e inner tube may be Values of velocity in the inner tube, dryness fraction
5 inch internal diameter and of 20 S.W.G. thickness. at the top of the annulus, and the head causing flow, are
Thus, the cross-sectional area of the inner tube is plotted in Fig. 25, for rates of heat absorption up to the
0.196 sq. in., and of the annulus 0.97 sq. in. T h e length very high figure of 200,000 B.Th.U. per sq. ft. per hr.,
of the tube is taken to be 12t feet, inclined at 45 deg. measured all round the circumference of the tube, for a
By an arrangement to be described later, it is ensured pressure of 300 lb. per sq. in., and in Fig. 26 for a
that the feed water supply to the trough shall be very pressure of 1,000 lb. per sq. in. T h e values for various
nearly at saturation temperature, so that the point of pressures from 100 to 3,000 lb. per sq. in., with a heat
evaporation is very near the bottom of the tube. The tube absorption rate of 50,000 B.Th.U. per sq. ft. per hr. are
end itself is shielded by refractories for a length of a few given in Fig. 27.
inches at the foot, so the point of evaporation may be These graphs show that this tube gives good results
located at only 13 feet from the bottom of the tube. even when the heat is extremely intense. Using the

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

170 THE CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS
criterion already discussed, bywhich a “five times round” tube in consideration correspond approximately to the
boiler should be considered satisfactory, that is, the highest rates of heating of fire row tubes at full power in
dryness fraction of the steam at the top of the annulus naval boilers. The heat would be partly radiant and partly
should not exceed 0.20, the tube would be suitable, at convectional .
300 lb. per sq. in. pressure, for a rate of heat absorption I n addition to having an adequate circulation over a
of 70,000 B.Th.U. per sq. ft. per hr. At 1,OOO lb. per wide range of pressures and rates of heating, these tubes
sq. in. pressure, the permissible rate of heat absorption offer valuable practical advantages in the general con-
would be 48,000B.Th.U. per sq. ft. per hr. struction and operation of the boiler.
These limits refer, of course, only to the tube of the A general arrangement of a boiler using these tubes is
dimensions specified. The tube lengths and sizes, and shown in Fig. 28. This boiler is designed for an electric
the relative sizes of inner and outer tubes may be varied power station, using low-grade fuel, and producing
to give suitable flow for any pressure and intensity of 36,000--15,000lb. of steam per hour at 400 lb. per sq. in.

I]rnJIjl [2

0 0 Lo
0 500 1.m 1,500 l,m 2.m 3.m
PRESSURE-LB. PER 50. IN. ABS.

Fig. 27. Performance of Lewis Tube at Various

B.TH.U. PER SQ. FT. PER HR. x 103 Pressures
Tube, 1+ inches outside diameter ;inner tube, 4 inch diameter ;
Fig. 26. Performance of Lewis Tube with Various length, 12 ft. 6 in. Rate of absorption of latent heat, 50,OOO
Rates of Heating B.Th.U. per sq. ft. per hr.
Tube, If inches outside diameter ;inner tube, inch diameter ;
length, 12 ft. 6 in. Pressure 1,000 lb. per sq. in. pressure and 725 deg. F., from feed water at 280
deg. F.
heating. It should be noted, however, that though there There is a single drum at the top, supported by trestles
are limits to the heat which should be applied to any external to the combustion chamber. The steaming tubes
tube, these limits are not due to any defect in circulation. are arranged in a bank subjected to both radiant and
The velocity curves in Figs. 25, 26, and 27 show that convectional heating, and as water-wall tubes subjected
there is very little variation in the flow, over the wide to radiant heat only. Each tube has only one expanded
range of pressures and rates of heating considered. The tube joint, namely, that at its upper end, and each tube
head causing flow is also maintained at high rates of is straight and easily inspected, externally and internally,
heating, though naturally it falls off slowly, with in- and can be replaced without disturbing any others if
creasing pressure. necessary. The inncr tube, when longer than the drum
Moreover, the calculations may be made with a diameter, is articulated in sections about 3 feet long,
reasonable degree of accuracy. There is none of the connected by flexible joints, so that any of these tubes
indeterminate character which is a feature of multi- may independently and easily be inserted from, or with-
tubular boilers, having nests of tubes in parallel between drawn into, the drum, by folding up to any required
headers or drums. The limiting rates of heating for the extent within 180 deg. The ends of the flexible coupling

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E C I R C U L A T I O N O F WATER A N D S T E A M I N W A T E R - T U B E B O I L E R S 171
project radially, and serve to centralize the inner tube little danger of sediment anywhere. Since the tubes are
within the outer tube throughout its length, though with- closed at their lower ends, they cannot be drained except
out seriously obstructing the steam flow in the annulus. by syphoning. They therefore remain full of water, and

Fig. 28. Lewis Single-Drum Stoker-Fired Water-Tube Boiler

Evaporation, 36,00045,000 Ib. per hr. ; working pressure, 300-400 lb. per sq. in. ; steam temperature, 725 deg. F. ; feed
temperature, 280 deg. F.
A Top drum. F Water-wall tubes. K Air preheater. 0 Spigots for promoting cor-
B Main steaming tubes. G Superheater. L Fan. rect feed water inflow.
C Inner tubes. P Tube support plate.
D Trough. H Superheat
per. control dam- M Hot air duct. Q Inspection door.
E Trestle. J Sensible heating tubes. N Secondary air ducts. R Flue gas.

The velocity of the water in the tubes is sufficient to are protected from the corrosion which takes place in
prevent any danger of sediment being deposited in the other types of boilers when emptied and allowed to stand
tubes, though with the exceptionally clean and de- empty for long periods during cleaning and overhaul
aerated water now in use for large boilers, there can be periods.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

172 T H E C I R C U L A T I O N O F WATER A N D STEAM I N WATER-’TUBE BOILERS
The tubes are straight, and held only at the top end, also of its heavy supporting structure, is another step of
the other being quite free; they are therefore able to great importance in the design of improved boiler plant.
expand with temperature changes, without stressing The cumulative effect of the various improvements in .
other tubes or the tube plates. The tubes can easily be design as described, namely, the simplicity of layout, the
kept clean, since external cleaning tools or blowers may safe use of higher heat-transmission rates, and the
be operated from their unobstructed lower ends, longi- abolition of a separate economizer, enable the size of the
tudinally, effectively removing all soot ,and enabling the boiler for any given duty to be greatly reduced. The
efficiency of the boiler to be maintained for very long quantity of refractory materials, insulation, and casings
periods, thus ensuring a good reliability factor. is also reduced. There is, moreover, a reduction in the
The gases cross the tubes practically at right-angles, so cost of housing the boilers, or of providing foundations,
that the maximum heat transfer may be obtained. It particularly on those riverside alluvial sites where power
becomes easy to arrange the correct variation in the stations are frequently built, and where piling is neces-
cross-sectional area of the gas passages to keep up the sary. No heavy supporting structure, such as is required
gas speeds, without resorting to inefficient baffles and for an external economizer is necessary.
multi-pass arrangements. Thus the heating rate over the Due to the reduction in size of the boiler, as well as to
back tubes may also be maintained at a high level. I I the inherent simplicity of the design, the maintenance
The possibility of greatly reduced total heating surface costs are also reduced. There are fewer starting-up,
for G given output and efficiency, by resort to higher gas standby, and radiation losses. There is less stress, strain,
speeds, as advocated by Dr. Nicolson in 1908 (Nicolson and corrosion ; and fewer other deteriorations occur. In-
1910, 1919), and more recently by Dr. Mellanby, is spection, washing-out, and, if necessary, withdrawal and
rendered feasible. Undoubtedly the hesitation of boiler- replacement of tubes, are easy. The need for withdrawal,
makers in the past to adopt such high gas speeds has however, is remote, since the circulation is good, and
been due to fear of the ill effects of heating the down- stagnation and consequent overheating is impossible.
comer tubes, with consequent interruption of circula- A Lewis boiler designed on these principles has been
tion. By using inner feeding tubes, however, this danger in operation for the last four years, giving excellent
is eliminated. Circulation is adequate, the rate of heating results (Engineering 1935). No trouble with circulation
the rear tubes can be increased, and the total area of or tube joints has been experienced, though the rate of
heating surface for any particular duty considerably firing on occasions exceeded 1 lb. of fuel oil per square
reduced. This improvement undoubtedly provides a most foot of heating surface. Annual inspections have re-
valuable means of reducing the oyerall size, weight, and vealed no tube distortions or distress, also that all interior
cost of boilers for all purposes. . tubes, etc., which have three times been drained and
The arrangement of water walls to the combustion withdrawn, and then replaced, are as clean and sound as
chamber is rendered much simpler by the use of tubes of when they were installed. The benefit of the unusually
the type described. Their certain circulation eliminates large range of feed water level has often been demon-
many of the troubles hitherto experienced with water- strated.
wall tubes, whilst the number of joints and pipes is The boiler has no superheater or economizer, yet
greatly reduced. yields efficiencies of 66 per cent and 70 per cent re-
Since each tube forms an individual heating unit, it spectively at the higher and lower calorific values of the
becomes possible to arrange that the sensible heating fuel. It raises steam at 200 lb. per sq. in. pressure within
shall take place in tubes attached to the main drum, 15 minutes. The feed water, supplied bythe Metropolitan
instead of in an external economizer. These tubes are Water Board, is of 16 deg. hardness, reduced to 2 or 3
in the rear bank, where the gas temperatures are at their deg. by a hot lime-soda softener. Concentration is kept
lowest value. They may be outer tubes with “inners”, low by .a continuous blowdown, the make-up being
similar to the steaming tubes already described, but they 100 per cent. Repeated samples of water tapped from the
would more usually be double or plain U-tubes, having bottom of the U-tubes have proved that no deposit
some legs with and some without inner tubes. The feed remains there, though in the drum, below the level of the
water from the pump can be led to a series of spigots blowdown orifice, an amount of soft mud is to be found.
pointing a short distance down the back leg of certain The high velocity of flow down the inner tubes is suffi-
U-tubes. This promotes correct water flow and prevents cient to prevent any serious accumulation anywhere
the incoming cool feed from mixing with the hot water in within the tubes or feed trough.
the drum. The forward upcomer leg discharges the A further recently completed design (Fig. 29) of a
sensibly heated water into the drum, whence it flows to portable boiler for Air Raid Precaution duties shows con-
the open ends of the central trough, from which the vincingly the great economies in weight, space, and cost
inner tubes feeding the steaming annuli of the outer of boilers using tubes of the type described. This boiler,
tubes take their supplies of water. Many efforts have been having a capacity of 6,000 lb. of steam per hour at
made to enhance the value of the rear heating surfaces of 185 Ib. per sq. in. pressure when hand-fired with coal
water-tube boilers, but they have all proved futile, be- under very low natural draught, is compact enough to be
cause admixture with the hot water in the boiler could mounted on a four-wheeled trailer; its total weight,
not be prevented. including feed pumps, etc., is less than 10 tons. The
The elimination of the external economizer and its boiler map therefore be towed at a speed of 15 m.p.h.
complicated pipework, difficult to keep clean of soot, and behind any power-driven commercial vehicle or tractor,

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 THE C I R C U L A T I O N O F WATER A N D STEAM I N W A T E R - T U B E B O I L E R S
and the weight and dimensions will not infringe the
Road Acts, or call for any special notification or per-
mission to traverse the roads (Engineering 1939).
The possibility is thus established of constructing
water-tube boilers, with perfectly reliable unidirectional

I
F9

natural circulation, of much less weight and space

than existing boilers, whether cylindrical or of present
water-tube designs. Moreover, the suggested boilers
have a wide range of application. They are suitable
either for land or marine work, in which sustained
output at full power is required, or in naval work
where efficiency at low ratings is essential, together
with ability to meet sudden increases in demand,
and even high overloads. The robustness of the
boilers reduces the amount of boiler work required
on a ship on reaching harbour, enabling the ship to
leave so much the sooner. The reduction in size and
weight of the boiler plant releases valuable cargo space
on merchant ships, or space for extra armaments,
protection, or increased range of action on warships.
The ability of the suggested boilers to use coal fuel
efficiently is a valuable asset, both in competition
with Diesel engines, and also in competition with oil
firing; moreover, it encourages the use of home-
Droduced fuel. Further, the sustained high thermal - -
national resources of materials, of construction, and of
Lfficiency, due to effective soot cleanhi, enables the man-hours and machine-hours require to be conserved
maximum useful heat energy to be abstracted from every with the care.
pound of fuel. The heat loss on shutting down a boiler
is reduced to a minimum, due to the small water content CONCLUSION
of the drum and tubes. From the theory of circulation, as derived in the
The importance of reduced construction and mainte- preceding pages, it is reasonable to conclude that, SO far
nance is especially great at a time such as the present, as circulation is concerned, the simple U-tube boiler will

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

171 T H E C I R C U L A T I O N O F WATER A N D STEAM I N WATER-TUBE BOILERS
give satisfactory results over a very wide range of overall reduction in size, owing to the possibility of using
pressures and rates of heating. Long experience with more intense heating rates, especially in the rear tubes,
boilers which approach most nearly to this standard with safety; and the very great economies due to elinii-
U-tube confirms the excellence of the circulation. ‘The nating the need for an external economizer.
U-tube boiler with unheated downcomers therefore may This form of water-tube boiler would appear to be a
be regarded as a standard of excellence, in its ample flow valuable advance in boiler technique, and to afford the
under natural head, its ability to absorb intense heat, its best means for increasing safety and reducing the capital
assurance against any interruption which might lead to and maintenance and housing costs of boiler plant ; of
overheating, and in the simplicity of its design. reducing boiler weight, work, and space on ships ; and of
-4 study of the common variations from thc U-tube improving reliability by reducing the frequency and
standard shons that these departures from the standard duration of shut-down periods for cleaning and repairs,
frequently diminish the performance of the tube con- especially in base-load electric power stations, lvhcrein
siderably, especially at the high furnace temperatures this desideratum has recently proved to be of uryent
and working pressures now becoming common. importance.
.A simplified form of boiler has been evolved whic!i A few words may be recalled, in conclusion, from
seems as near to the standard as possible, having un- James Watt, who said, “Simplicity in all things, hut
heated downcomers, individual water supply to each proverbially in mechanism, is the acme of perfection”.
tube, free discharge for the steam, and ample circulation Or, as Krishnamurti said, “Strive always for simplicity
under natural head, for a wide range of pressures and -not the simplicity of crudeness-but of culture . . .
intensities of heating. Moreover, numerous external ad- the synthesis each is able to make of all that he has
vantages accrue from the use of this rational design, learned, felt, and experienced”.
such as freedom from longitudinal temperature stresses ; And the authors would remark that in steam
ease of soot cleaning ; ease of arranging the correct area of generating apparatus, simplicity is still the acme of
gas passages; considerable reduction in size and com- perfection, and that the suggested developments i n
plexity, by the elimination of bottom drums and headers ; design form a contribution t o m r d s this desirable quality.

APPENDIX I
P H Y S I C A L P R O P E R T I E S O F W A T E R .4ND STEr\>I

Pressure, Ib. per sq. in. abs. . 100 300 500 1.000 2,000 3,000 Reference
(Appendix I I)
Saturation temperature, deg. F. . 327.0 417.4 467.1 544.7 636.0 695.5 Callendar (1931).
Total heat of water, B.Th.U. per Ib. . 298.5 394.2 450.1 544.1 676.4 816.2
Total heat of steam, B.Th.U. per Ib. . 1,190.7 1,209.4 1,213-2 1,204.3 1,148.8 1,039.3
Specific volume of water V,, cu. ft. per Ib. . 0.0177 0.0190 0.0199 0.0219 0.0263 0.0339
Specific volume ofsteam Vs,cu. ft. per Ib. . 4-429 1.552 0.940 0.458 0.193 0.083 ,
Viscosity of water 2,lb.+(hours xfeet). . 0.418 0.324 0.283 0.235 0.198 0.182 Hawkins,
Potter (1935).
Solberg, and
Viscosity of steam 2, Ib.+(hours X feet) . 0.046 0.061 0.075 0,102 0.143 0.172 Hawkins., Solberg,and
Potter (1935).
Thermal conductivity of water, B.Th.U. per
hr. per sq. ft. per deg. F. temperature
difference per foot thickness . 0.393 0.383 0.367 0.338 0.291 - King (1936).
Thermal conductivity of steam, B.Th.U. per
hr. per sq. ft. per deg. F. temperature
difference per foot thickness . 0.024 0.037 0.0498 0.0825 0.147 - Keenan and Keyes
(1936).
Specific heat of water, B.Th.U. per Ib. per
deg. F. . 1.03 1.08 1.14 1.29 1.83 - Keenan and Keyes
(1936).
Specific heat of steam at constant pressure,
B.Th.U. per Ib. per deg. F. . 0.60 0.74 0437 1.24 2.60 - Keenan and Keyes
- - (1 936).

A P P E N D I X I1 ~ O S W J A K O Y IF.C , 1930 Terhiiisrhe JIerhanik rind

REFERENCES
Thermodynamilz, vol. 1, p. 358.
~ R O O K S ,C. H., and BADGER. IV. L. 1937 Trans. Am.
BARBET 1914 Bullktin de 1’Association des Chimistes Inst. Chemical Eng., vol. 33, p. 398, “Heat
de Sucrerie et de Distillerie de France et de ses Transfer Coefficients in the Boiling Section of a
Colonies, vol. 32, p. 1 1 1 . Long-Tube Natural-Circulation Evaporator”.
BOARTS,R. M., BADGER, W. I,., and MEISENBURG, S. J. CALLENDAR, H. I,. 1931 “The Revised Callendar
1937 Trans. Am. Inst. Chemical Eng., vol. 33, Steam Tables”, Arnold, London.
p. 363, “Temperature Drops and Liquid Film COPE, W. F. 1937 Proc. I.Mech.E., vol. 137, p. 165,
Heat Transfer Coefficients in a Vertical Tube”. “Friction and Heat Transmission Coefficients”.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 T H E C I R C U L A T I O N O F WATER A N D STEAM I N W A T E R - T U B E B O I L E R S 175
CROFT,H. 0. 1938 “Thermodynamics, Fluid Flow, KING, W. J. 1932 Mechanical Engineering, vol. 54,
and Heat Transmission”, McGraw-Hill Book p. 560, “The Basic Laws and Data of Heat
Company, New York and London. Transmission”.
CRYDER,D. S., and GILLILAND,E. R. 1933 Re- 1936 Kent’s Mechanical Engineer’s Handbook,
frigerating Engineering, vol. 25, p. 78, “Heat 11th ed., vol. 2 (Power Section), p. 3-28, “Heat
Transmission from Metal Surfaces to Boiling Transmission”.
Liquids”. KING, W. J., and KNAUS,W. L. 1934 Mechanical
DIGHT,S. R. 1933 Trans. Inst. N. Architects, vol. 75, Engineering, vol. 56,p. 283, “Heat Transfer Rates
p. 171,“Naval Water-Tube Boilers. Experiments in Refrigerating and Air-Cooling Apparatus”.
and Shop Trials”.
1936 Trans. Inst. N. Architects, vol. 78, p. 54, MCADAMS, W. H. 1933 “Heat Transmission”,
“Naval Water-Tube Boilers”. McGraw-Mill Publishing Company, New York
DOCHERTY, J. G. 1928 Proc. I.Mech.E., p. 921,“The and London.
Stresses in Thick Tubes due to Temperature and 1937 Trans. Am. Inst. Chemical Eng., vol. 33,
the Effects of Temperature and Pressure Stresses p. 472.
Acting Simultaneously”. NICLAUSSE, J. 1914 Proc. I.Mech.E., p. 507, “ The
DREW,T. B., and MUELLER, A. C. 1937 Trans. Am. New Niclausse High-Duty Marine Type Boiler”.
Inst. Chemical Eng., vol. 33, p. 449, “Boiling”. NICOLSON, J. T. 1908-09 Trans. Jun. Inst. Eng., vol.
“ENGINEERING” 1935, vol. 139, p. 453, “The Lewis 19,.p.199,“Laws of Heat Transmission in Steam
Single-Drum Water-Tube Boiler”. Boilers as Deduced by Experiment”.
1939,vol. 148, p. 90, “Lewis Single-Drum Water- 1910-11 Trans. Inst. Eng. and Shipbuilders in
Tube Boiler”. Scotland, vol. 54, p. 64,“BoiIer Economics and
EWING, J. A. 1922 “Thermodynamics”, article in the Use of High Gas Speeds”.
Dictionary of Applied Physics, ed. Sir R. Glaze- NUKIYAMA, S. 1934 J1. SOC.Mech. Eng. Japan, vol. 37,
brook, Macmillan and Company, London, vol. 1 , p. S-53,“The Maximum and Minimum Values
p. 921. of the Heat Q Transmitted from Metal to Boiling
HAWKINS, G. A., SOLBERG, 14. L., and POTTER,A. A. 1935 Water under Atmospheric Pressure”.
Trans. A.S.M.E., vol. 57, FSP-57-11, p. 395, SANKEY,H. R. 1914 Proc. I.Mech.E., p. 530,reference
“The Viscosity of Water and Superheated Steam”. to fuel tests carried out.
JAKOB, M. 1936 Mechanical Engineering, vol. 58, SCHMIDT, E. 1933 Combustion, vol. 5, p. 6, “Experi-
pp. 643, 729,“Heat Transfer in Evaporation.and ments Regarding Circulation in Vertical-Tube
Condensation”. Boilers”.
KEENAN, J. H., and KEYES, F. G. 1936 Thermo- STROEBE, G. W., BAKER,E. M., and BADGER,W. L.
dynamic, Properties of Steam, including Data for 1939 Industrial and Engineering Chemistry, vol.
the Liquid and Solid Phases, Chapman and Hall, 31, p. 200, “Boiling Film Heat Transfer Coeffi-
London. cients in a Long-Tube Vertical Evaporator”.

Discussion
At the Extra General Meeting in London on 1st March 1940 *
Mr. A. F. WEBBER,B.Sc. (Eng.), A.M.I.Mech.E., con- It was now generally accepted that almost complete
gratulated the authors on having tackled so energetically water-cooling of the furnace was desirable except for
one of the hardest problems in boiler design. The num- ignition arches, and it would be seen that with the Lewis
ber of alternative designs, both of forced circulation boiler this was almost impossible as it was based essen-
boilers and of other types, showed that circulation had tially on the use of straight tubes and a single drum, so
been rather a weak point in boiler design. Nevertheless, that only the top of the combustion chamber could be
boiler design, like many other problems, had to take into water-cooled. Two points in connexion with heat trans-
account a very large number of factors, and concentra- fer seemed to be open to criticism. First, supports must
tion on one factor to the neglect of others was almost be provided to keep the inner tube central in the water
bound to lead to an unsatisfactory result. The problems tube, so that small patches of the outer tube would not
of combustion, furnace temperature, heat transfer, and be in contact with the steam and water ; the author’s
“birds-nesting” scarcely seemed to have received suffi- arrangement seemed to permit a real risk of local over-
cient attention in the design now suggested. heating in the intense radiant heat transfer of the fire-
For Minutes of Proceedings of the meeting, see p. 145. row tubes, Second, the convective heating surface
Communications will be published later. appeared to be inadequate. Many members would re-

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

176 D I S C U S S I O N ON C I R C U L A T I O N O F WATER A N D STEAM I N WATER-TUBE BOILERS
member an interesting paper+ on the first “Wood” The favourite boilers were those which might be called
steam generator. In that case an attempt was made to “real boilermakers’ jobs and not pipefitters’ night-
design a boiler which would be nothing more than a mares”!
radiant-heat combustion chamber ; but actually a con- In three-drum boilers, it was well known that the water
siderable bank of convective heating surface had to be drums usually gave most trouble. Moreover, when a
added. Returning to the present paper, Fig. 28, p. 171, tube burst it was as often as not in the middle of the
showed that where the superheater was placed, only stack-evidently one of the stagnant tubes shown by the
three rows of tubes were provided, including the fire authors in Fig. 19, p. 166. T o remove the burst tube, a
row itself. Such an arrangement might be necessary, to dozen or more adjacent tubes might have to be cut out
obtain an adequate gas temperature for superheating the and subsequently replaced. This very tiresome operation
steam to 725 deg. F., but then the only means left for was unnecessary in the Lewis boiler, in which any tube
extracting the remaining heat from the gases were the could be removed without disturbing its neighbours.
economizer and the sensible heating tubes. It was stated The Lewis boiler, in his opinion, afforded not only a
in the paper that feed water entered at 280 deg. F. The simplification but what he considered an improvement
saturation temperature would be a little over 400 deg. F. ; with great possibilities. All the desirable features of light-
little heat therefore remained to be extracted in the ness, compactness, and accessibility were there, and the
tubes to which he had referred, and it seemed likely design seemed to be entirely straightforward, with no
that exit gas temperatures would be excessive. freakish novelties. The Field tube and Niclausse
Reference was made in the paper to the Lewis boiler principles had already been long established in practice ;
installed in 1935, with which good efficiencies were U-tubes with closed ends had been widely used in super-
obtained ; but that boiler had many more rows of tubes heaters, and the design had the outstanding advantage
than were shown in the present design. Moreover, the of the elimination of water drums. One ingenious detail,
earlier boiler was not fitted with an economizer ; pre- which he missed in the illustrations in the present paper,
sumably the feed arrived at the back end of the tubes at was the U-tube itself, a clever adaptation of twin Field
a fairly low temperature. Therefore, since the refractory tubes. I t appeared in outline only in Fig. 29, p. 173, and
lining was not completely water-cooled and since the he suggested that the sectional details on a larger scale
amount of convective heating surface was small-a might be added. Possibly the authors did not now think
feature which was again inherent in the design because that the U-tubes were all-important, but considered that
of the single drum (unless the drum was made very much the same effect could be obtained with single Field
lsrger, and then a great many advantages with regard tubes. A useful application of these tubes, in which each
to weight and cost would be lost)-it was difficult to see tube or pair of tubes constituted a self-contained stearn-
that the boiler in question, in spite of its virtues of cir- ing element, was indicated in Figs. 28 and 29, pp. 171
culation, was likely to supersede more conventional and 173, where they were shown radiating as an inverted
boiler designs. fan across the brickwork at the ends of the boiler. e p -
Regarding conventional boiler designs, it was an parently such tubes could be added to many existing
interesting point that the authors’ experiments in cir- boilers with advantage.
culation rather went to show the advantages of the In Fig. 28 stack tubes marked J presented a neat form
oldest type of Babcock and Wilcox boiler as compared of economizer, which, it would seem, provided an effec-
with vertical-tube boilers, because in that design the tive form of feed heating by utilizing waste gases. The
downcomers were right at the back and presumably recent American revival of economizers in marine boilers
were less heated than some of the front rows of the back had been attended with good results, and no doubt the
banks of vertical-tube water-tube boilers. He would be practice would extend to this conservative country in due
glad to have the authors’ views on the points he had course.
raised, because, if the difficulties to which he referred In conclusion, he thought that the Lewis boiler con-
could be overcome, the feature of free circulation would stituted an improvement which was well worth pursuing.
be most valuable. It was difficult, as was well known, for a new design to
gain recognition, and he was glad that the authors had
hlr. J. HAMILTON GIBSON(London) remarked that been given an opportunity to present their case.
experience showed that if marine boilers were to give
satisfactory service they must be light in weight, occupy Dr. WILLIAMABBOTT,M.I.Mech.E., observed that
as little space as possible, be able to meet sudden varia- the first Lewis boiler was installed six years ago at the
tions in steam demand, and be easily accessible for Northampton Polytechnic Institute. Although minor
repairs. During the last twenty-five years of his working troubles were encountered at first, the performance of
career in a large shipyard he had been engaged in the the boiler was quite satisfactory ;certainly those features
design, construction, and trial performance of almost which were novel gave no cause whatever for anxiety,
every known type of marine boiler, including the Belle- and, speaking as an old Admiralty man, he felt that the
ville, the Babcock and Wilcox, the Niclausse, the Nor- simple and yet scientific Lewis boiler, devoid of heavy
mand, the Yarrow, and the so-called Admiralty type. and troublesome bottom drums, would establish itself
* SMYTH,E. W., 1930 Proc. I.Mech.E., p. 73, “General for use afloat.
Operation Experience with the First ‘Wood’ Steam One of the most attractive features of the paper was
Generator”. the use of a fundamental principle, i.e. the conservation

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 D I S C U S S I O N O N C I R C U L A T I O N O F WATER A N D STEAM I N W A T E R - T U B E B O I L E R S 177
of energy, coupled with the simplest of experiments to economizer, but was given as 90.5 per cent on the lower
investigate flow through the tubes. In papers presented calorific value of the coal fuel, on a six-hour test. There
in the past the velocity of steam had been calculated by was no superheater, but in his written reply Mr. Lewis
equating kinetic energy to available head, using the gave a heat balance showing less than one-tenth of 1
relationship v = d ?&. No one ever questioned its per cent (0.09 per cent) loss by moisture in the saturated
validity. Indeed, it was used as long ago as 1890 by steam! The efficiency of the oil-fired Lewis boiler at
George Babcock, when lecturing to Cornell University, Northampton Polytechnic Institute was 66 per cent,*
and it had appeared in the publication Steam ever since. which made the 90 per cent efficiency claimed by
The authors were convinced that that theory was wrong. Niclausse for his hand-fired boiler seem absurd.
They believed that, in equating energy to energy, the Sir Leonard Pearce’s recent lecture t showed that 90
heat energy given to the tube by the fire must be taken per cent efficiency was the maximum obtainable, even
into account, and they had illustrated their point experi- with the latest boiler units fired by mechanical stokers
mentally with, he thought, complete success. Further, and provided with superheaters, economizers, and air-
they showed that friction, which was usually ignored, preheaters. The latest units described by Sir Leonard
might absorb as much as half the total head. Pearce, each for generating 550,000 lb. steam hourly,
On the experimental side, the authors had not accepted combined the header type with the three-drum type,
as conclusive the well-known demonstrations by Yarrow, but all tubes were unobstructed and accessible.
which purported to show that the heating of the down- The authors based their claims on two small units,
comer was immaterial. No one would belittle Yarrow’s namely, the oil-fired boiler at Northampton Polytechnic
work or question his integrity when interpreting scientific Institute, which generated 2,000 lb. hourly from and at
results, but the fact remained that his experiments 212 deg. F., at 66 per cent efficiency, and the 6,000 lb.
applied only to the small-scale apparatus that he used. per hour portable hand-fired boiler at Rickmansworth
I n his experiments Yarrow found no case in which heat- (Fig. 29, p. 173). As no other water-tube boiler maker
ing of the downcomer was detrimental, and he there- had followed Niclausse in using Field tubes, the evi-.
fore concluded that it was never detrimental ; but the dence so far did not justify the theory now advanced to
authors, using full-scale apparatus, had shown that cases support such practice.
occurred, of practical importance, in which heat applied
to the downcomer was most detrimental. He himself Mr. R. F. DAVIS,M.Sc. (Eng.), A.M.I.Mech.E., said
could testify to the authors’ results; the flow in one that in Part I of the paper, dealing with the change from
direction became completely unstable when the down- water to steam, it was of the greatest interest to note that
comer was heated ;in fact it became unstable long before the authors had adopted the suggestions made by him
the heat applied to the downcomer approached that to the Council when he examined the paper on its
applied to the upcomer. The authors had proved that behalf. I n his own paper on “The Physical Aspect of
stability of flow must be in doubt when the downcomer Steam Generation at High Pressures and the Problem
was heated, and a notable feature of the Lewis boiler of Steam Contamination”, which had been accepted
was that the downcomer could not be heated. There was by the Council for possible presentation next session, it
still need for more investigation into the subject, and had been necessary for him to cover to a large extent the
he hoped that the authors would be encouraged to carry same ground, in order to carry the analysis to a further
their work further into this fascinating problem. development. It was therefore important that he should
make the foregoing statement. For the present he would
Mr. H. CHARLES ERITH, M.I.Mech.E., likened the only say that experiments with pure or nearly pure
design of the Lewis boiler to that of the Yarrow boiler water at atmospheric pressure might not represent what
but without the lower or mud-drums. The Yarrow boiler happened with chemically treated boiler water and at
had always seemed to him to make nonsense of boiler high operating pressures.
circulation theories. It had no unheated downcomers, He congratulated the authors on having analysed in
and no one could say exactly at what part of the tube such detail the circulation in a U-tube boiler. He took
banks the flow ceased to be upwards ; yet it was the exception, however, to the statement (p. 147) that it was
favourite boiler for destroyers, which were subject to not practicable to analyse the circulation in modern
severe fluctuations in steaming duty. The authors pointed types of boilers. A method had been described by Mr.
out that the Perkins-Field tube was 100 years old and Timmins and himself in a paper read before the Insti-
had been used for fifty years in the Niclausse boiler. Mr. tution in Manchester, and he considered that that method
Lewis’s adaptation of the Yarrow boiler was ingenious, gave a very good working approximation. Also, more
but few would agree with his view that the fact that his recently Lodinegg $ had published an analysis of the
tubes could not be drained was a merit. conditions for unstable flow in both natural circulation
He himself was present at the discussion at the Insti- and forced circulation boilers.
tution in 1914 on the Niclausse boiler * and he remem- He could not agree with the statement (p. 147) that it
bered the scepticism expressed by Mr. Patchell and was necessary to establish a simple type of boiler as a
others with regard to the claim for 90 per cent thermal
Engineering, 1935,vol. 139,p. 453.
efficiency on a test by Captain Sankey of a hand-fired t PROCEEDINGS,
1939,vol. 142, p. 305.
boiler ; the figure was actually 87.3 per cent, including 1 PROCEEDINGS,
1933, vol. 125, p. 732.
5 Die W h e , 1938, vol. 61, No. 48, “Unstable Flow in
PROCEEDINGS,
1914,p. 507. Natural and Forced Circulation”.
1s

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

178 D I S C U S S I O N O N C I R C U L A T I O N O F W A T E R A N D S T E A M I N W A T E R - T U B E B O I L E R S
standard of excellence. A standard of excellence for pur- Fig. 3, p. 149, was of considerable interest. He asked
poses of calculation was not necessarily a standard of the authors whether such a condition always existed or
excellence for performance. For instance, it was not whether the physical formation of steam was not affected
clear how the inner surface of the Lewis boiler tube by many factors, e.g. the viscosity and surface tension of
could be effectively inspected and cleaned. The authors the liquid, the diameter of the tube, the condition of
also stated (p. 147) : “Faulty circulation, leading to over- the tube wall, and the rate of circulation. Actually, tests
heating and blistering of the tubes, is one of the principal carried out by Vorkauf under low rates of forced cir-
causes of defective operation of modem boilers.” I n the culation, and tests that he himself had carried out under
comparatively few cases of overheating of tubes which natural circulation conditions (with unheated down-
he had encountered, scale had always proved to be the comers), indicated that, even with constant heating, the
main cause of the trouble. Zancanaro came to the same surface of the water in a tube oscillated with varying
conclusion in a recent article.* amplitude and frequency ; this was caused by the for-
mation, expansion, and collapse of large steam bubbles.
Mr. G . A. PLUMMER, M.I.Mech.E., congratulated the In some cases a portion of the water content of the
authors on their review of auto-circulation in steam tubes might be violently discharged; in other words,
boilers. The Field tube boiler was almost a perennial ;it local priming might take place.
was embodied in various Continental designs, and was He would welcome some information from the authors
developed in the Niclausse and Durr boilers, whilst as to the quality of steam delivered by their boiler, par-
mercury boilers also incorporated an intricate form of ticularly at high pressures. He would also like to know
Field tube. the general dimensions and weight of the boiler shown
He fully appreciated the difficulties encountered by in Fig. 28, p. 171. The design was interesting, par-
the authors in attempting a complete analysis of the cir- ticularly in view of the small amount of valuable radiant
culation in natural circulation boilers. The many failures heating surface. He also asked how the convection heat-
of tubes in such boilers, some of which had been ing surfaces were cleaned, as the combined depth of
recently reported in the Transactions of the American front bank, superheater, and rear bank seemed to be
Society of Mechanical Engineers,t seemed to show that considerable, and there appeared to be no space between
circulation in such boilers was frequently defective. It the banks for the insertion of any cleaning apparatus.
was probably for that reason that the forced circulation
boiler, with correctly proportioned water flow, controlled Mr. F. E. EVANS,M.I.Mech.E., said that he was par-
by nozzles at the entrance to each tube, had achieved ticularly interested in the Lewis boiler because he had
such popularity. He could not, however, subscribe to had the privilege of attending nearly all the trials on the
the authors’ view (p. 167) that “any form of forced cir- portable boiler shown in Fig. 29, p. 173. He asked the
culation . . . cannot be as efficient as natural circula- authors whether they would give the results of those tests.
tion properly arranged”, by reason of the undoubted
loss of a portion of the power supplied to the pump.
The authors appeared to overlook some of the benefits Captain R. E. TREVITHICK, M.I.Mech.E., said the
derived from the efficient dispositionof boiler heating sur- authors had given a wealth of technical data in their
face made possible by using forced circulation which en- paper, but he did not think that they had included
abled a saving in power required for mechanical draught, sufficient details of practical experience to substantiate
more than counterbalancing expenditure in pumping their case, as in the description of the boilers references
power. And added to this saving were advantages ac- were made only to constructional details. Moreover,
cruing from considerations of weight, space, and so forth. Fig. 28, p. 171, would appear to be a projected design
Regarding the Lewis boiler, he asked for further in- only. He asked whether the authors could give some
formation on the effect on the water level of a sudden operating and test results of any Lewis boilers of
change in load. The authors said on p. 172, “The benefit reasonably high capacity, say 100,000lb. per hr., with a
of the unusually large range of feed water level has often working pressure of about 500 lb. per sq. in. and a final
been demonstrated”. Assuming the circulation to be steam temperature of about 800 deg. F. In the absence
active, if a sudden rise in load occurred, resulting in a of such information, he considered that the authors in
pressure reduction in the drum, the change in pressure their paper did not do justice to forced circulation.
would be transmitted throughout the system with the There were approximately 1,OOO boilers with forced cir-
velocity of sound, but the change in temperature of the culation working at the present time, with an aggregate
fluid in circulation would be a function of the heat con- hourly evaporation of over 7,500 tons of water. Individual
tent. Thus self-evaporation might occur within the Field boiler capacities ranged up to 300,000 lb. per hr. and
tube itself and the entire contents of the tube might be working pressures up to 1,500 lb. per sq. in., whilst
discharged suddenly into the drum. Conversely, with combustion chamber heat releases were of the order of
a sudden reduction in load and consequent rise in pres- 700,000 B.Th.U. per cu. ft. per hr.
sure, the steam bubbles in the Field tubes might collapse, The authors stated that the boiler described in
thus considerably lowering the water level in the drum. Engineering * was without superheater or economizer
but had an efficiency of 66 per cent with a working
* I1 CaZore, 1938, Nov. pressure of 200 lb. per sq. in. He asked the authors what
t PARTRIDGE,E.P.,andHALL,R.E. 1939 Trans. A.S.M.E., the capacity of this boiler was. I n view of the relatively
vol. 61, p. 597, “Attack on Steel in High-Capacity Boilers as a
Result of Overheating due to Steam Blanketing”. * Engineering, 1935, vol. 139, p . 453.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 DISCUSSION O N C I R C U L A T I O N O F W A T E R A N D S T E A M I N W A T E R - T U B E B O I L E R S 179
small size of this boiler, he thought that the authors’ and a quick shut-down occurred when manoeuvring ?
comparison of the U-tube standard with forced circula- Would not the solids separate out at the lowest points
tion, and their query (p. 167), “Comparing with the immediately circulation ceased? After all, one of the
U-tube standard, the question immediately arises, Why functions of the lower drum or drums, or alternatively
is forced circulation necessary at all ?” were not justified of the mud-drum, of the established types of marine
on the evidence they presented. water-tube boilers was to collect such solids.
The authors’ claim for the “simple” boiler without
Mr. W. A. STANIER, M.1.Mech.E. (Vice-President) economizers was interesting, but such an arrangement
said that a water-tube boiler had been tried in a loco- formed the basis of the highly efficient water-tube designs
motive designed by Sir Nigel Gresley. The boiler was which were to be found in operation aboard most of the
built by one of the most prominent water-tube boiler modern British passenger and mail liners.
manufacturers in this country, and he believed that one A high air temperature was quite satisfactory at sea
of the major difficulties was the doubt as to which way with oil firing ;but on land with coal-burning mechanical
circulation would take place in the boiler. One of the grates, it was a different matter. What air temperature
greatest troubles in the water-tube design, so far as loco- did the authors expect below the grates on the design in
motives in this country were concerned, was the accom- Fig. 28? It must be high if they were aiming (without
modation of the boiler within the loading gauge, namely economizers) at the usual efficiency offered by competi-
13 ft. 1 in. from rail level and the provision of a firebox tive designs and he himself would expect considerable
sufficiently large for the proper combustion of the fuel. trouble, particularly from slag formation.
Lastly, regarding a saving in weight, which was one
Major WILLIAM GREGSON,M.Sc., M.1.Mech.E. of the main claims of the Lewis boiler, and naturally
(Menzber of Council) said that the authors had made out one of particular interest for marine work: assuming
a very clear case for the non-heated downcomer. He equal ratings when comparing with other designs, did
considered that no natural circulation water-tube boiler the authors think that the elimination of, say, a single
-unless operating at low ratings-was really satisfactory bottom drum really meant an appreciable percentage
from a circulation point of view unless it had a down- reduction on the boiler installation as a whole, after cor-
comer system of some sort which was not subjected to recting for the weight of the inner tubes of the Field
heating as a portion of the boiler evaporative surface. tubes and the extra drum fittings involved?
He agreed with the authors as to the necessity of
adopting the principle of high gas speeds over convection Mr. W. YOUTHLEWISand Mr. STRUAN A. ROBERTSON
heating surface to increase the effectiveness of heat trans- wrote in reply that the general trend of the discussion
fer and so to reduce the size and cost of this surface. But showed little criticism of the principal subject of the
these precepts-as advocated by Osborne Reynolds and paper, namely, the theory of circulation. I n fact, only
Nicolson-had for many years now been in use by boiler one speaker expressed difference of opinion. Their own
designers. The firm with which he was associated conclusions, that downcomers should not be heated,
regularly used gas speeds of well over 100 ft. per sec. in that the steam should have a dry discharge, and that each
certain types of boiler where the bulk of the heat transfer tube should have its individual water supply, might
took place by contact effect. therefore be considered as upheld.
Referring more particularly to marine water-tube The paper was written mainly on circulation. NO
boilers, the real problem in modern boiler design was to attempt could be made to cover all aspects of boiler
burn the fuel at the rates called for by the restricted com- design. It would be wrong to assume, however, that
bustion chambers associated with reduced heating sur- though little space was devoted to other problems, they
face and consequent small boiler size. His firm was burn- had been neglected. Again, it was suggested that the
ing oil in highly rated marine boilers at figures approach- claims made for the Lewis boiler were based on insuffi-
ing 400,000 B.Th.U. per sq. ft. of heating surface per cient evidence, that is on results obtained with the two
hour; such rates of heat liberation in the furnace not small boilers already in existence. As a fact, however, the
only demanded special technique in fuel-burning equip- claims were founded on the whole study of Circulation,
ment, but also a complete break-away from brick linings, as given in the paper, which showed that the Lewis boiler
and their replacement by some form of completely water- approached more nearly than any other boiler to the
cooled radiant-heat-absorbing furnace walls. ideal form, in having unheated downcomers, dry dis-
In the boiler illustrated in Fig. 28, p. 171, the U-tubes charge, and individual water supply to each tube.
forming the convection heating surface after the super- Moreover, the performance of the Lewis boiler in practice
heater did not appear to follow the precepts proved in might well be judged by the outstanding success of the
the earlier part of the paper ;the temperature drop of the Niclausse boiler, the type nearest to it in principle.
gases from the superheater zone to the air heater inlet I n reply to Mr. Webber, it was found that in the Lewis
divided by the number of rows of U-elements did not boiler higher furnace temperatures and higher rates of
allow much gas temperature difference across each lateral heat transfer than in other types could be used with
row of U-tube elements. safety, due to the assured circulation in the tubes.
He also doubted the complete faith which the authors Water-walls were easily arranged with Lewis single or
had in their ability to keep the lower ends of the Field double tubes, or double rows of tubes, at the end of the
tubes clear of deposits. What would happen in a marine in- combustion chamber, and, if necessary, with radiant
stallation, assuming impure water had got into the system superheater tubes on the ends or on the floor. Soot

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

180 DISCUSSION O N CIRCULATION OF WATER A N D STEAM I N WATER-TUBE BOILERS
cleaning, and removal of “birds-nesting,” were easily boiler was out of use reduced corrosion. The 1913 tests
secured by end access to the tubes. There were no lower on Niclausse boilers, whose results Mr. Erith doubted,
drums or headers to interfere with the effective use of consisted of a series of 16-hour tests, lasting over several
the necessary appliances, which were best operated in months.” Mr. Erith’s last sentence seemed to be point-
the direction of the tube length. The high efficiency of less, and merely called attention to the failure of
the tubes could therefore be maintained, and the boiler exponents of other boilers to awaken to the excellent
kept on load for longer periods. characteristics of the Field tube. As now applied in the
The danger from overheating at the points where the Lewis boiler, the Field tube had been shown to be as
inner tube supports touched the outer tube seemed re- nearly as possible the ideal tube.
mote, since the heat would be rapidly transmitted through They wished to acknowledge some valuable sugges-
the metal of the tube, or along the metal of the support. tions made by Mr. R. F. Davis, during the preparation
No difficulty had been found, in designing Lewis boilers, of the paper, when he directed their attention to the
in providing adequate radiant and convective heating work of Nukiyama, Jacob, Brooks and Badger, Drew and
surface with drums of normal size, or in securing effi- Mueller, and some others. These references were helpful
cient exit gas temperatures. It should be pointed out in studying the change from water to steam, though
that less convective heating surface was required in the some of the effects investigated by these authors seemed
Lewis boiler than in most conventional types, owing to to them to be of small importance in actual boiler
the higher gas speeds used. T h e gas passage areas were circulation. They looked forward with interest to
easily contracted, to suit the reduced volume of the gases Mr. Davis’s own presentation of the problem.
as they became cooler, so as to maintain the gas speed. They had closely studied the paper by Mr. Davis and
The downcomer in the Babcock boiler mentioned might Mr. Timmins, in particular the section on circulation,
have been less severely heated, but the inherent defect but were unable to agree with many points in it. In
in this design was the use of a number of tubes in parallel calculating the frictional resistance, Mr. Davis and
between the headers. Much evidence had been forth- Mr. Timmins used an expression involving the reciprocal
coming in recent years as to the uncertainty of circulation of the average density of the steam-water mixture in the
in such tubes, and the likelihood of stagnation and over- tube, but did not show how this average was computed.
heating, even in tubes in the middle of the tube bank, The correct method was to use the arithmetic mean
whether of the old type or of later designs. value of the reciprocal of the density that is, the arith-
They welcomed Mr. Hamilton Gibson’s remarks as metic mean value of the specific volume, which depended,
coming from one of the most able authorities on marine as shown in their own paper, on the arithmetic mean
boilers. The double Lewis tube which he mentioned was dryness fraction. I n obtaining the average density for
used in the first Lewis boiler. It had some advantages, calculating the available head (termed “static head”)
in that, should the inner tube in one leg of the U-tube Mr. Davis and Mr. Timmins used the arithmetic mean
become choked, the other could continue to supply both between the densities at entrance and exit. This was
legs. Again, the loss at sudden reversal in the single tube incorrect, except in one very special case in which there
was reduced. Single tubes were somewhat simpler, was a most unlikely distribution of the heat. Certainly
however. Both types might therefore be used, according it was incorrect in the case of uniform distribution of the
to circumstances. Double tubes were of particular value heat. No allowance at all was made for what they them-
in end wall tubes, and in intensely heated fire row tubes. selves termed “acceleration head.”
Dr. Abbott’s comments on the erroneous expression They agreed with Mr. Davis that it was quite possible
for calculating circulation, published fifty years ago, and to calculate the circulation in modern boilers. In the
perpetuated in the catalogue Steam, should be em- paper, however, they considered that the most useful
phasized. This error had undoubtedly misled investi- procedure was to study a simple type of boiler, in which
gators too long; even Admiralty boiler authorities had the flow could be calculated easily, and in which the
fallen into the trap in papers recently published. The effects of variations in pressure and in rates and condi-
minor troubles encountered with the first Lewis boiler, tions of heating could be analysed. By doing this, and by
mentioned by Dr. Abbott, were solely concerned with establishing the excellence of this boiler from the
the operation of the oil fuel burning equipment. circulation viewpoint, it was possible to continue the
Mr. H. Charles Erith’s contention as to the advantages analysis to elucidate the effect of deviations from the
of the Yarrow boiler seemed to be based on a mis- simple type commonly met with in practice, and to
apprehension. As mentioned by Dr. Abbott, Yarrow’s arrive at a reasoned conclusion on the principles
theories, based on his experiments, had been found to governing good circulation. Mr. Davis and Mr. Timmins
be untenable. The proposals made in the paper enabled themselves used a simple boiler in order to analyse the
satisfactory circulation to be assured at much higher effects of varying pressures, though the value of their
heating rates than were commonly adopted in Yarrow results was reduced by the errors mentioned above.
boilers. The saving of weight in the Lewis boiler would They also fell into the error of assuming that natural
be of inestimable advantage in destroyers. T h e ease of circulation ceased at the critical pressure, an idea which
access to the tubes, the absence of racking and tempera- was shown in the paper to be incorrect.
ture stresses, and of stresses tending to crack the tube The tubes of the Lewis boiler, being straight, could be
plates, were added advantages. The tubes could be * For a full answer to the late Mr. Patchell’s comments on
drained readily if required, by siphoning with a rubber these tests, see Proc. J.Mech.E., 1914, p. 528, and The
tube. The fact that they normally remained full when the Engineer, 1914, vol. 117, pp. 280 and 463.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

 DISCUSSION O N CIRCULATION OF WATER AND STEAM I N WATER-TUBE BOILERS 181
inspected throughout by an electric light bulb, or even was expected in obtaining good C 0 2 percentages and
photographically over their entire internal surface by a considerably higher efficiency, at this rating, g’iven a
well-known periscope. Internal cleaning was, however, suitable grate.
unnecessary, since no stagnation could occur in the tubes, It might be pointed out that these results were remark-
and deposits or scale were unlikely if concentration was ably good for a small boiler having only nine rows of
reasonably low. Mr. Davis’s final remarks as to the tubes, with low gas speed to suit the low draught (only
comparatively few instances of overheated tubes were 0.04 inch of water) available by short chimneys reaching
hardly in accord with the American experience mentioned only 15 feet above the grate level. The fuel bed resistance
by Mr. Plummer. was overcome by means of steam jets under the grate.
In reply to Mr. Plummer, Field tube boilers were Captain Trevithick’s interesting data on the number
hardly a “perennial,” the last paper on a boiler of this and capacity of forced circulation boilers seemed to be
type being that by M. Niclausse in 1914. It would seem irrelevant. T h e Lewis boiler was admittedly in the
that the advantages which Mr. Plummer claimed for the initial stages of its development, but its design was based
forced circulation boiler, due to reduced draught, space, on well-established principles, and they had no hesita-
and weight, when compared with older types, could all tion in anticipating its ability to compete with forced
be obtained with the Lewis boiler, without expending circulation boilers of any capacity, once the advantages
considerable power in driving circulating pumps, and of its simple design, and economy of capital cost, main-
without the complexities of the forced circulation boiler. tenance, and working became more widely recognized.
With regard to the effect of a sudden change in load, Major Gregson’s statement on the need for complete
and the possible disturbance of normal circulation by water-walling was one with which they partly agreed.
self-evaporation or fall of water level, no trouble from There was no difficulty in arranging this in the Lewis
this cause had been experienced with the Niclausse boiler, as already mentioned in their reply to Mr. Webber.
boiler, which responded magnificently to severe load T h e temperature drop of the gases across each row of
changes of the order of 80 per cent, instantaneously, and tubes was calculated according to generally accepted
without appreciable fall of pressure. In the Lewis practice, and was adequate for the heat transmission
boiler, demonstrations showing pressure drop to be required. No danger was anticipated of the lower ends
harmless had been made by suddenly opening the stop of the Field tubes becoming choked. Certainly nothing
valve to atmosphere, when carrying heavy fires, and could be deposited whilst the boiler was working, due
dropping the pressure from 185 to about 50 Ib. per sq. in., to its assured rapid circulation at every point, and the
without detriment to the tubes or to the apparent dry- amount deposited on shutting down could not be more
ness of the steam discharge. Should circulation be than the infinitesimal quantity in the water actually in
upset momentarily, it was rapidly re-established, by the tube at this instant. The turbulence at reversal was
reason of unheated downcomers, and dry discharge to enough to clean out this minute amount again very
the upcomer. The water level had also been allowed to rapidly on restarting. In practice, even with 100 per cent
disappear for several minutes from the gauge glass sight make-up feed, any sediment was found to be deposited
without apparent harm. on the top of the cover plate over the trough in the drum,
No doubt viscosity, surface tension, etc., affected the this being the first place where the steam velocities
extent of the various sections in the change from water diminished enough to allow any settlement to occur.
to steam. The authors were now carrying out further Moreover, this was the place from which it was most
experiments on this point. Experience with Niclausse easily removed. Experience with Niclausse boilers bore
and Lewis boilers showed that no priming whatever out these results, some Niclausse boilers having been fed
occurred, if concentration was kept down to a reasonable on test and in emergencies with sea water for consider-
limit. The quality of the steam from the mobile boiler able periods, without disadvantage. T h e elimination of
was demonstrated by the fact that it was found quite bottom drums and their water content secured a saving
satisfactory for use in a steam turbine. The dimensions of weight much greater than the weight of the internal
and weight of the boiler shown in Fig. 28 might be parts, since drums had to be made to withstand steam
gauged from the fact that the drum was 4 feet in diameter. pressure, whereas the internal parts were of much lighter
T h e results of the tests on the Lewis mobile boiler, material. Moreover, the space saved by eliminating the
for which Mr. Evans asked, were as follows. When bottom drums was an important advantage.
firing at the rate of 20 lb. of coal per square foot of grate Few speakers seemed to have realized the full extent of
per hour, and getting 14 per cent C 0 2 in the flue gases, the savings accruing from the Lewis design. The port-
the equivalent rate of evaporation was 8.5 lb. of steam able Lewis boiler was mounted on a four-wheeled trailer
per square foot of total heating surface, or 5,600 lb. of which could be used without trouble on common roads.
steam per hour, from and at 212 deg. F., giving an effi- Other water-tube boilers of the same capacity would
ciency of nearly 80 per cent. When firing at 30 lb. of coal have required a six- or eight-wheeled crocodile truck,
per square foot per hour, the output was 6,600 lb. per hr. outside the limits set by the Road Acts, and too high to
from and at 212 deg. F., giving an efficiency of about 64 pass under bridges. T h e economies in man-hours and
per cent. The drop in efficiency was largely accounted for machine-hours, in construction, transport, and erection
by the low C 0 2 value of 8 per cent in this test, due to of boilers of this type were also of vital importance.
difficulties with the grate. Apart from this, however, the These effects formed a contribution to the solution of the
boiler was in every way satisfactory, without any modifi- problem enunciated in the second part of the title of the
cation whatsoever of the original design, and no difficulty paper, namely, the rational simplification of boiler design.

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016