Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Reheat Factor - Mollier Diiagram

Download as pdf or txt
Download as pdf or txt
You are on page 1of 40

THE STEAM-TURBINE EXPANSION LINE ON THE

MOLLIER DIAGRAM. AND A SHORT METHOD


OF FINDING THE REHEAT FACTOR
By

Eldgar

Buckingham

CONTENTS
Page
cyg

Introduction
2.

The
The

3.

Adiabatic flow

^83

4.

Dissipative flow

5.

Isentropic changes of state

^84
^86

6.

The

1.

7.

8.
9.

total-energy equation for steady flow of a fluid


total

580

heat of steam

582

H <poT Mollier diagram

589

Combined efficiency of similar stages in series


H

cp plane
Representation of expansion on the
efficiency
Form of the expansion line graphical construction

^the

approximation for the reheat factor R


Practical working method for finding the value of
Distribution of the isopiestics along an isentropic
Second approximation for R
Tests of the rule for finding the value of i?
Uses of the reheat factor
General remarks

593

595
598
600

reheat factor

10. First
11.
12.

13.
14.
15.
16.

Note

(to sec. i).

602

i?

604
606
609
611
612

Limitations of the theory

615

INTRODUCTION

The

"heat diagram," or
(p diagram, in which the
state of a mass of steam is represented by a point on a plane with
absolute temperature {&) and entropy (^) as rectangular coordinates, has in recent years been much used by writers on technical thermodynamics, and for many purposes it is most instrucBut for the
tive, though some caution is needed in interpreting it.
quantitative solution of problems in steam-turbine design it is by
no means comparable in convenience with the Mollier diagram"
or total heat entropy diagram, in which the representation is on a
so-called

**

579

58o

Bulletin of the

Bureau

of

Standards

[Vol. 7.

No. 4

plane with the total heat (H) and entropy (^) as rectangular
The idea of using a surface with H, cp, and p (prescoordinates.
being
sure) as rectangular coordinates is due to Willard Gibbs,

the same as his "heat function" x- The


q) diagram may be
cp plane, and
regarded as the projection of this surface on the

was introduced to the notice of engineers by


Dresden, by whose name it is commonly known.

Prof. Mollier, of

it

The

an introduction, for those not familiar with the subject, to some of the
technically important properties of quantities
and cp and of the
first six

sections of the following paper serve as

The remainder

paper contains a discussion of the form, on this diagram, of the expansion line for wet
steam flowing through a multistage turbine of known stage efficiency, and the development of a practical method for use by
cp diagram
designers for drawing the expansion line on the
without the use of the laborious step-by-step method.
Mollier diagram.

of the

1.

THE TOTAL-ENERGY EQUATION FOR STEADY FLOW OF A FLUID

Let a

CoPC

fluid of

(Fig.

which we

i).

any

be flowing steadily along a channel


Let Ao and A be two sections of the channel
sort

shall call the entrance

and

exit sections.

A
Ao

Po

'"o

Oo Eo

X
^

To

..>

Fig. 1

Let poVoOoEo be the pressure, specific volume, absolute temperature, and internal energy per unit mass of the fluid as it
crosses the entrance section, and let pv6E he the corresponding
quantities at the exit section. These are to be averages over the
section, and the variations from one point to another of the section
are to be small.

The Reheat Factor

Buckingham]

581

Let To be the kinetic energy per unit mass at Ao, of the axial
component ^ of the velocity, and T the corresponding quantity at
A. The channel at ^o and A shall be varying in cross section so
slowly that the kinetic energy of the radial velocity is negligible.
If I per cent of the total kinetic energy is a negligible quantity, a
total taper of one in four, for a cone, is permissible, so that no
severe demands are made on the constancy of cross section at
Ao and A.
Subject to the foregoing restriction in the immediate vicinity of
Ao and A, the shape of the channel between Ao and A is a matter
There may, if we please, be included,
of complete indifference.
as forming a part of the channel or completely inclosed within it,
a motor actuated by the flow of the fluid and delivering work
outside the channel, or a pump actuated by the appHcation of
power from without and doing work on the fluid. The walls of
the channel must be tight so as to prevent leakage of fluid, but
they need not be thermally insulating. The fluid may be any
liquid, vapor, or gas, but for concreteness we shall usually refer to
it

as steam.

The first law of thermodynamics, if applied to the passage of one


poiuid of steam from Ao to A, gives us the following statement;
the total energy per poimd internal plus kinetic is increased by
the amount of the work done on the steam in crossing the entrance
section by the steam behind it, and is decreased by the amount of
the work it does against the steam ahead of it in crossing the exit
section, by the work given out by the motor, and by any heat
which may have been lost by conduction or radiation through the

walls of the channel, which

We

of the motor.

may

in places coincide with the walls

thus have the equation

{E + T)

- (Eo + To)

=-

poVo- pv

-W -Q

(i)

is the work done outside the channel by the motor,


Which
and Q is the heat loss, both measured per pound of steam. The
work
includes work done against friction at any bearings which
are outside the channel. All the terms in the equation are to be

in

understood as expressed in the same units,


units (B.

t.

e. g.,

u.).
^

See note at the end of this paper.

British thermal

Bulletin of the

582

Rearranging equation

(i),

Bureau

of

Standards

[V01.7.N0.4

we have

T -To + W = (E + pv)o- (E + pv) -Q

(2)

fundamental equation of the theory of fluid


motors. It is appHcable to regularly acting periodic motors as
well as to continuous-flow motors such as turbines, if the quantities
in the equation are averaged over an integral number of periods or
over any very long time. Since no restriction has been imposed

which

is

the

first

upon the sign


ically acting

pumps

valid for all cases of

averaging

W or Q, the case of continuously or periodalso included, and the equation


in
steady orunder the above condition as to

of either
is

is,

periodically varying flow of any sort of


2.

fact,

fluid.

THE TOTAL HEAT OF STEAM

The quantity (E+pv) has a definite value for every state of


the fluid in question and has been designated as ''heat of formation at constant pressure," "heat contents," and "total heat."
We shall adopt the name total heat and write

E+pv = H
The quantity

(3)

to which Regnault gave the

and which he measured

for dry saturated

ing purposes, sensibly identical with

name

steam

is,

"total heat,"

for all engineer-

H as just defined.

The small

outstanding difference is due, first to the fact that the internal


energy of water at the ice point is not absolutely though nearly
the same at

all

and second to the fact that the volume


point under any given pressure is not zero

pressures

of water at the ice

though usually quite negligible in comparison with the volume,


at the same pressure, of an equal mass of steam dry enough to be
suitable for use in a steam motor.
Values of
for water and for dry-saturated steam are given in
the steam tables. Their difference is evidently the heat of evaporation, and the value of
for any degree of dryness, x, may be
found by linear interpolation between the values for water and
for dry steam at the given pressure or temperature.
For superheated steam, the value of
is greater than for dry-saturated
steam at the same pressure, by the amount of heat needed for the
superheating at constant pressure, which depends on the specific
heat of superheated steam and is not so well known as might be

The Reheat Factor

Buckingham]

583

Convenient tables for saturated and superheated steam


are given in the "Steam Tables and Diagrams" of Marks and
Davis 2 from which all the steam data used in this paper have
been taken.
We may now write equation (2) in the form
desired.

{T-To)+W + Q = Ho-H

(4)

The quantity (Hq H) will, for short, be called the "heat-drop"


of the pound of steam during its passage from ^o to A.
The first two terms of equation (4) represent a quantity of
mechanical energy, hence the equation may be read as follows:
The mechanical energy produced plus the heat lost to the surroundings, is equal to the heat-drop of the steam. It should be
noted that it has not been stipulated that the fluid shall flow without encoimtering passive resistances such as viscosity. There
may be as much internal dissipation of mechanical energy into
represents work actually
heat as we please, provided that
delivered outside the channel, including work against friction at
any outside bearings, and that (T To) represents the actual
excess of kinetic energy of the axial component of velocity in the
exhaust at A over that in the feed at Ao.

3.

An

ADIABATIC

adiabatic change of state

is

FLOW

defined as one during which the

substance in question neither takes in nor gives out heat through

bounding surface. If, therefore, there is no heat leakage


through the walls of the channel, the change of state of the steam
between Ao and A is adiabatic. We may then set Q = o and
its

equation

(4)

reduces to

iT-To)+W = Ho-H

(s)

an equation which has several immediate applications to familiar


facts.
For example, in the case of steady flow, without heat loss,
through a simple tube not containing a motor, the work
is
zero and equation (5) says that the increase of kinetic energy is
equal to the heat-drop. This is the case of flow through a turbine
nozzle or through fixed guide blades, if the flow is so rapid or the

Longmans Green and Co.

1909.

Bulletin of the

584

Bureau

of

Standards

[Voi. 7,

No.

channel so well protected that the heat leakage is negligible. In


the case of a properly designed, velocity-compounded, impulse
turbine, working with a constant pressure within each stage, the

and

velocity

kinetic energy decrease during flow through the

Hence T, at exit from these


at entrance to them, and H>Ho.
The

intermediate reversing guide blades.


blades,

is

less

than To,

kinetic energy dissipated in the guide-blade channels thus appears


in the

steam as "reheat" making

The simplest
drawing

in

possible case

which there

is

is

now

that of adiabatic throttling or wire-

neither outside

The whole

increase of kinetic energy.


(5)

H larger than Ho.


work

first

W nor any sensible

member

of equation

vanishes, so that the total heat of the steam remains

unchanged during

its

fall

of presstue.

Since for steam which

remains dry-saturated the value of


decreases with falling pressure, wire-drawing thus tends to superheat dry steam or to dry
wet steam, a familiar fact upon which the action of the "throttling calorimeter"

is

based.
4.

DISSIPATIVE

FLOW

The passage of heat by conduction or radiation from one point


in a body to another involves a waste of availability of the heat
thus uselessly let down from a higher to a lower temperature.
The internal heating of a body by friction or viscosity involves a
waste of the mechanical energy dissipated as work done against
the passive resistances of friction or viscosity. In either case, the
original state of the body, existing before the waste took place,
can not be reestablished except by interference from without;

and

if

we take

into account all the bodies involved, the outside

bodies used as well as the one concerned in the original process,


the initial state of all of them can never be reestablished at all by

any means whatever.


Such processes are known,

thermodynamics, as " irreversible "


processes.
No physical process is entirely free from such elements,
and all real changes are therefore irreversible. But if these wasteful elements of the change are relatively so insignificant as to be
of negligible importance, the change is sensibly though never
exactly "reversible." Evidently any process from which we
in

The Reheat Factor

Buckingham]

desire to get as

much work

may

causes of waste,

be from

all

The wire-drawing

of

585

as possible should be freed as far as


i.

e.,

of irreversibility.

steam through the ports during admission

to the cylinder of a reciprocating engine, especially at the cut-off

point and at the beginning of admission with incomplete cushioning, is

an

may be.

and wasteful action to be avoided as far as


The expansion between complete cut-oft' and the opening
irreversible

of the exhaust valve

is,

so far as the steam itself is concerned, sensibly

from internal irreversible actions and therefore internally


reversible, in the thermodynamic sense, though on account of the
effect of the cylinder walls it is usually far from adiabatic.
The expansion of steam through a turbine nozzle or through
blade channels of decreasing section would, in the ideal case, be
free from dissipation, i. e., not retarded by frictional or viscous
resistances, and would be in the thermodynamic sense reversible
if so rapid that no sensible interchange of heat took place between
In practice, there are always resistdifferent parts of the steam.
ances due to skin friction and eddy currents; some mechanical
energy that might otherwise be produced is thus dissipated into
heat and the process is irreversible. Since these resistances are
wasteful, it is evident that the smaller they are and the more
nearly reversible the expansion is, the more closely the gain of
kinetic energy (T To) approaches the theoretically possible niaximtmi which is determined, if there is no leakage of heat, solely by
the initial state of the steam and the final pressure to which
expansion takes place in the space into which the steam jet issues.
During the expansion of steam between any two points in its
path through a turbine, the whole mechanical energy produced,
or {T To + W), also approaches its ideal maximum value as all
the internal losses due to skin friction and eddy currents in nozzles
or blades, windage, and wire-drawing of steam which leaks past
The only difference
blade tips or through bushings, approach zero.
between this case and that of flow through a nozzle is that in the
free

nozzle or in fixed blades the


in the
is

more general case

work

is

necessarily exactly zero, while

W may have a

finite

often negligibly small in comparison with

W.

value and (T To)

Bulletin of the

586

5.

Bureau

of

Standards

[Voi. 7,

No. 4

ISENTROPIC CHANGES OF STATE

during a reversible isothermal expansion at the absolute


temperature 0, a body e. g. a poimd of steam takes in from
If,

without a quantity of heat Q, the quantity

is

known

as the

entropy " of the body. Thus, during the evaporation


of one pound of water at constant pressure and temperature,
ending in its conversion into a poimd of dry-saturated steam, the
increase of entropy is equal to the latent heat divided by the absoThis quantity is given in the steam tables as
lute temperature.
" increase of

"entropy of evaporation."
If a reversible change of state occurs during which the temperature of the body is not constant, we may cut the process up into
a mmiber of small steps; for each of these divide the heat taken
in by the average temperature of the body during that step;
and finally add all these small quotients. If we then reduce the
length and increase the

number

of the steps indefinitely, the

sum

B
approaches a definite limit expressible in the form

-^ where A

A
and

sion

are the initial

the same for

is

state

of the

and
all

The value

final states.

reversible changes

of

tliis

which lead from the

known as the increase


body during the change A B. The elementary
to the state B,

and

more general

of entropy

it is

isothermal change, considered above,

is

expres-

case of an

evidently included in

change of entropy.
If, as is usual, we take the temperature of the ice point and
the pressure of one atmosphere as our standard conditions, the
entropy of a mass of fluid at any other temperature and pressure,
this

definition of

referred to this standard state,

is

the value of

- from the

standard to the actual state along a reversible path


and this
value depends only on the end state reached and not on how the
body actually reached it, if we assume, as is permissible, that
the end state could have been reached by a reversible process.
;

In an adiabatic change of state,


if

an adiabatic process

is

dQ

everywhere zero. Hence


also reversible, the entropy of the body
is

The Reheat Factor

Buckingham]

in question does not change.

587

Any change of state in which the


constant is known as an isentropic

entropy of the body remains


change.
The ideal adopted for the expansion of steam in the cylinder of
a reciprocating engine is that of expansion without wire-drawing
or other irreversible internal losses, in a perfectly nonconducting
cylinder.
Such an expansion would be a reversible and adiabatic,
and therefore an isentropic change of state. From this it has come
about that the term ''adiabatic" is very often loosely used in
engineering works with the meaning "isentropic." In reality,
however, an adiabatic process need not be isentropic and an
This becomes so
isentropic process need not be adiabatic.
evident in studying steam-turbine theory that there is ground for
hope that this confusion of terms may eventually be eliminated
from our thermodynamic literature.
If, for example, a pound of steam passes from a state A to a
state B, the change of its entropy is definite and depends only
on A and B. If the actual process is irreversible because of
internal dissipation by which heat is produced within the steam,
the increase of entropy is not to be found by taking the value of

B
-^ along the

actual irreversible path, but

is

greater than this.

A
For heat generated internally has the same

changing the
temperature or otherwise influencing the state of the steam as an
equal amount of heat added from without, so that the final state
effect in

reached is not the same as if there had been the actual addition
of heat from without but no internal generation of heat.
In any actual expansion between two given completely defined
states, the final entropy of the expanding substance is always

B
A
with

representing only heat added from without, for even in

always some internal dissipation. This


excess represents heat produced by dissipation of mechanical
energy which might, with ideally perfect arrangements, have
the best case there

74356 12

is

Bulletin of the

588

been saved and used.

Bureau

It is therefore

of the process of expansion.

of

Standards

[Voi. 7.

No. 4

a measure of the wastefulness

IrreversibiHty, wasted availability

and needless increase of entropy, are merely


One main reason why
different aspects of the same thing.
entropy, which has been somewhat of a stumbling block in technical thermodynamics, is retained, is that its changes give us the
most convenient quantitative expression of the wastefulness or
the efficiency of thermodynamic processes from which mechanical
work is to be obtained.
The expansion of steam in the cylinder of the reciprocating
engine is approximately reversible, so far as the steam itself is
concerned; the internal losses due to dissipation inside the steam
are small, and poor indicated efficiency is due largely to the fact
of energy, dissipation,

that the expansion

not adiabatic, the influence of the cylinder


walls causing the course of the expansion to be different from that
The expansion though nearly reversible is not adiabatic
desired.

and not

is

isentropic.

The expansion

steam in a turbine

may be

regarded for most


purposes as adiabatic, the external losses of heat from the steam
But the process is subject to a great
to the casing being small.
of

deal of dissipation

by eddy

currents, etc., so that

though nearly

from being either reversible or isentropic.


The ideal of the steam turbine is thus the same as that of the
reciprocating engine, namely, isentropic expansion.
In the one
adiabatic

it is

case the ideal

far

is

not attained because the changes of state of the

steam, though nearly reversible, are far from adiabatic; in the

though nearly adiabatic, they are far from reversible.


The assumption that expansion through a steam turbine is
adiabatic and fails of being isentropic and ideally efficient only

other, because

because of internal dissipation losses of various kinds is, of course,


only an approximation. For, except possibly in some very unusual
cases, there is always some heat lost by conduction and radiation
from the turbine, and there is always some longitudinal conduction

between the

different stages, tending to

make

separate parts

even though it might be so


as a whole.
External losses might be reduced or even made
negative by jacketing; but while this would improve the efficiency
of the turbine, considered by itself, the jacket steam used would

of the expansion not quite adiabatic

The Reheat Factor

Buckingham]

probably

much more than

offset the gain

increase in the total water rate.

589

and cause a considerable

It is possible

that with small

turbines not well protected the external heat losses might be as


great as the internal dissipation and so just balance the reheat.

In such a case the total result would be the same as if the mechanical energy actually wasted inside the turbine had been used up
on bearing friction outside, with no internal heating and no conduction and radiation loss. The expansion might thus be isentropic without being, even approximately, either adiabatic or
reversible.

We

however, treat the turbine problem as one of adiabatic expansion, having therefore an isentropic expansion as its
ideal, and shall treat the external heat losses as negligible and
shall,

consider the flow as subject to equation

(5).

Such a

simplifica-

would not be permissible in precise physical work, but for


the pmposes of steam-turbine design, the errors thus introduced
are probably always less than those due to the uncertainties in
the values of some of the quantities that have to be assumed in
the computations notably, velocity losses in nozzles and blade
channels and windage resistance, both of which influence the efficiency, of which an estimate must be made in order to design
tion

at

all.
6.

The

THE H ^ OR MOLLIER DIAGRAM

total energy equation for adiabatic flow

{T-To)-^W = Ho-H
gives us information in terms of the total heat

seen that the ideal expansion


to absence of dissipation

is

(5)

We

H.

have also

isentropic, since that corresponds

and therefore

to 100 per cent efficiency

mechanical energy. It is. therefore evident


that for the graphical solution of problems in steam flow turbine
design in particular a chart showing jthe properties of steam on
and entropy cp as rectangular coora plane with the total heat
in the production of

dinates will be very convenient.

On

this plane,

if

is

made

ordi-

an isentropic expansion is represented by a


straight line drawn vertically downward from the point representing the initial state; and the heat-drop during the expansion
is represented by the length of this line from the initial to the final

nate and

g)

abscissa,

Bulletin of the

590

position of the state point.

Bureau

of

Standards

[Vol. 7,

No. 4

In any adiabatic expansion whatever,

the difference of ordinate of the initial and final points gives us, at
once, the heat drop and therefore, by equation (5), the mechanical

energy developed.

We

have now to mention some of the geometrical properties of


cp diagram for steam and may refer to Fig. 2, which gives a
the
qualitative idea of the diagram with the isopiestics, or lines of
constant pressure, as well as lines of constant dryness and super-

An

exact plot will not be attempted because it is easily


available in the above-mentioned tables of Marks and Davis as

heat.

well as in other books to

which

it

may be presumed

that the

reader has access.


Vertical lines are isentropics,

and horizontal

lines are lines of

constant total heat or "throttling lines." A vertical distance


read off on the scale of ordinates is a difference of total heat or a
heat-drop.

entropy.

horizontal

distance

reprCvSents

difference

of

The Reheat Factor

Buckingham]

The steam saturation

line SS" is nearly straight

of pressures used in practice.

simultaneous values of

steam

table.

The

591

lines

be plotted by taking
for dry saturated steam from the

It

H and

may

within the range

easily

xx etc. are lines of constant dryness;

points on one of these lines represent the possible simultaneous

values of

and

cp

for

wet steam of the given dryness factor

x.

They run in the same general direction as the saturation line.


The lines tt etc. rimning also in this same general direction but
above instead of below the saturation
stant superheat

line

55, are

lines of con-

t.

The lines we are most


upward toward the right,

straight for

wet steam,

saturation field below 5S, but curving

heat after crossing

shown sloping

interested in are those

55 and

i.

e.,

upward with

entering the superheat

within the

rising superfield.

These

Within the saturation field they are also isothermals, since the temperature of wet
steam is fixed by its pressure; but above the saturation line they
cease to be isothermals, because superheated steam at a given
pressure may have any temperattue higher than its saturation
are the constant-pressure lines or isopiestics.

temperature.

pound of wet steam in the condition of pressure


and dryness represented by the point a, let us add to it at constant
pressiure and temperatm-e a quantity of heat represented by the
length ac. The state point moves to b, the entropy increasing by
an amount represented on the entropy scale by cb. But since the
temperature 6 is constant we have, by the definition of change of
Starting with one

entropy,
length of cb

= length

oi ac-^6

these lengths being measured on the scales of

and

cp

used in

drawing the chart. If these scales are the same, = ac-v-cb = tan a.
If as is usual the scales are different, we have
6 = k tan a
(6)

where ^ is a constant depending on the scales and equal to unity


when they are the same.
The slope of the isopiestics for wet steam is therefore proportional to the absolute temperature the lines have greater slope as
the pressure rises and have the familiar fan-shaped arrangement
shown in Fig. 2. The reasoning just used is applicable even when
:

Bulletin of the

592

the isopiestics are curved,

i.

Bureau

e.,

of

Standards

iVol.7,No.

for superheated steam,

so that equation (6)

if

ac and be

understood
to be the angle between the cp axis and the tangent to the isoThere is no sudden change of direction
piestic at the given point.
at the saturation line, because there is no discontinuity in the temperature but with increasing superheat the temperature and therefore, by (6) the slope of the isopiestics increases and they are concave upward, as shown.
Lines of constant dryness drawn for regularly changing values
are infinitesimal

is

general

if

is

of X, e. g.,

any given

:Jf

= o.9,

:x:

isopiestic,

= o.8, x^o.j,

etc.,

and the lengths

cut off equal segments on

of the

segments on two
/

ent isopiestics are proportional to the values of ^yi

+^

differ-

where

is

the latent heat, and ^

is

the quantity tabulated as "entropy of

evaporation.

we

which 5*5 represents


the steam saturation line and S'S^ the water line, i. e., the
constant dryness line for x = o. The whole length of any isopiestic AB between the two saturation lines is given by

To prove

this

AB = -ylAO + cW.

consider Fig.

But

CB = = the
/

given pressure and temperature p

same

as that of

H,

0;

AC=n=the
e

3,

in

heat of evaporation at the


while

if

the scale of

(p is

entropy of evaporation.

the

We

therefore have

AB = yJ+^^=iVi+^^
If

the scales are not the same there will be a proportionality factor

different

from unity.

The Reheat Factor

Buckingham]

Any

line of

constant superheat crosses

points where their slopes are greater


their slopes at
7.

593
all

the isopiestics at

by a constant amount than

and below the saturation

line.

REPRESENTATION OF EXPANSION ON THE H ^ PLANEEFFICIENCY

Let us start with one pound of steam at the pressure p^. If the
steam is dry-saturated, its initial state is represented by the
point

(Fig. 2), otherwise

by the

intersection of the

same

iso-

with the appropriate dryness or superheat line. Let


Then
p2 be the final pressure to which expansion takes place.
C represents the final state reached in isentropic expansion,
piestic

and

AC=H^H^

is

the

ideal

maximum

heat-drop available

for conversion into mechanical energy during adiabatic expansion.

may

once from the scale at the


side of the chart, and the ideal water rate computed for a steam
motor in which steam expands adiabatically from the initial
It is necessary to say *'the
state A to the final pressure p^.
state A'' and -not simply "the pressure />i"; for it is evident
that on accoimt of the non-parallelism of the isopiestics, this
isentropic heat-drop is greater for dry or superheated than for wet
steam. This is qualitatively in accordance with the observed
fact that, with given Hmiting pressures, a steam engine or a turbine has a lower water rate when the initial superheat is raised.
In the actual expansion of steam through a turbine to the final
pressiu'e p2, there is internal dissipation and the working of the
machine is not ideally efficient. By reason of this dissipation,
part of the mechanical energy which might have been obtained
with a perfect machine is either not produced at all or if produced
The whole of this lost mechanis again immediately dissipated.
ical energy appears as "reheat" in the steam, diminishes the
actual heat-drop, and increases the entropy just as much as if it
had been heat added to the steam from without. The final
state of the steam at p2 will therefore be represented by some
point B, to the right of and higher than C. The greater the
dissipation the farther B will be from the ideal final state C.

The

ideal yield

The
is

thus be read

off at

H^H^ = AC, the actual heat-drop


reheat is H^ H^^CD. The efficiency

ideal heat-drop being

H^H^^AD,

and the

of the process in converting available heat into mechanical energy

Bulletin of the

594

Bureau

AD /AC.

of

Standards

[Voi.

?,

No. 4

not necessarily the same as


what would ordinarily be called the efficiency; for, by equation
(5) (Hi H2) goes to the production of the kinetic energy (T Tq)
as well as to the production of the outside work W, which is the
is

evidently equal to

This

is

important result of the process and commonly the only result


But in general the quantity
considered in computing efficiency.

(T To) is negligible in comparison with the useful work


delivered by the steam to the rotor inside the glands.
If this is
true and if A and B (Fig. 2) represent the state of the steam in
the steam chest and in the exhaust chamber, the efficiency of

the machine in the usual sense

given

and therefore the


would thereby be fixed.

given a priori,
sented by

is

by e = AD/AC.

final

If

were

steam condition repre-

In order to discuss the efficiency of the separate parts of a multiIn a multistage turbine we have first to define the term stage.
stage impulse turbine, of the Rateau type for instance, points at

the entrances to the various nozzles and at exit from the last com-

In a turbine
of the Parsons type, points in the clearance spaces at entrance to
the fixed blades, together with a point at exit from the last moving row to the exhaust space, are "similarly situated. " We shall
define a stage as the part of the turbine between any two such
This agrees with the usual
adjacent similarly situated points.
definition of a stage for the impulse turbine, but not with that
sometimes adopted for turbines of the Parsons type in which each
row of blades whether fixed or moving is regarded as a separate
stage.
Our definition makes a " stage" consist of a fixed row and
the next following moving row. This use of the term seems more
rational than that which divides the turbine into two kinds of
stages the moving rows and the fixed rows in one of which no
work at all is done by the steam on the rotor. At all events, it is
convenient for the purposes of this paper and will be adopted.
If, then, we consider not a whole turbine but a single stage, we
may say that the change of kinetic energy (T To) through any
stage is negligible.
In an impulse stage T and Tq are usually separately negligible in a Parsons stage the change of steam speed is so

partment to the exhaust are "similarly situated.

"

small as to

make {T T^

negligible except in the first stage,

since turbines of this type always

and

have a great many stages, an error

The Reheat Factor

Buckingham]

595

no importance in regard
to the turbine as a whole.
If the kinetic energy term is negligible
in comparison with
for each separate stage, it is so for the whole
turbine, and if e is the efficiency in the usual sense we have (Fig. 2)
of this sort regarding a single one

is

of

AD
'=AC
When

for, as

we shall see in section 9,

8.

refers to a single stage

it is

called the

*'

a distinctive term

stage efficiency,
is

'*

needed.

FORM OF THE EXPANSION LINE GRAPHICAL CONSTRUCTION


expansion of steam in passing through a
not a perfectly regular and continuous process

It is evident that the

multistage turbine

is

>

*R

Fig.

can therefore not in reality be a


Fig. 2.
Let us consider the simple case
of a three-stage impulse turbine working with wet steam between

and that the expansion


smooth curve like AB in

line

the pressures p^ and p^, the intermediate pressures at entrance to


the second and third stage nozzles being p^ and p^.
In the first-stage nozzles the pressure drops from p^ to p2 if the

machine

is

properly designed.

There

is

some reheat due

pation in the nozzles before the pressure has fallen to

p2,

to dissi-

but by

far

Bulletin of the

596

Bureau

of

Standards

\voi. 7,

no.

the greater part of the reheat is due to blade losses, windage, leakage, and " carry-over " i. e., the kinetic energy of the steam leaving

the last set of moving blades which is in general wasted and not
available for driving the issuing steam directly into the next set
of ncfzzles.

All these except the nozzle loss occur at the constant

The true expansion through this stage will therefore


be represented by a line something like ABC (Fig. 4) and for the
second and third stages, if of approximately the same design as the
first, the true expansion lines will be of somewhat similar shape, as
shown at CDE and EFG. It would evidently be difficult to prepressure

p2.

form of the expansion line ABCDEFG in all its


small details, but fortunately this is not necessary; for in practice
we do not need to know the steam condition exactly for every point
but only for a few sets of similarly situated points.
Let us suppose, for example, that we want to know the steam
condition at entrance to the three sets of nozzles and in the exdict the precise

haust; this information

A, C, E, G.

Now

ciency of the

first

is

given by the positions of the points

the position of

stage:

For

if e is

is

this

we know the
efficiency, we have

fixed

if

effi-

which Ha is the value of


that would be reached if the expansion were isentropic and the final state represented by the point a.
Another similar step from C to the next pressure p^ fixes the
point E, and a third the point G.
The points C, E, G can thus be found if the stage efficiencies
and 3 are known. Before designing can begin, they
1,
2,
must be known, either by experiment on single stages similar to
the ones in question, or by computation from a speed diagram;
for the stage efficiency is a datum which is fundamental to the
design and can not be dispensed with.
If there were more than three stages the same method might
be continued, and for any moderate number of stages this would
not be a laborious operation; but the practical problem is not

in

so simple, for in general only the terminal pressures are given

while the intermediate pressures are not given but are to be de-

termined in accordance with some further condition. It may,


foi instance, be desired that the work done on the rotor shall be

The Reheat Factor

Buckmgkam]

the same in

all

No

the stages.

597

general solution of the problem

can be given in such a case, and a graphical solution by trial and


error might with a large number of stages involve a great waste
of time which must be avoided if possible.
At this point the problem simplifies itself. Usually we are
required to distribute between two limiting pressures a given
number of similar stages which may be assumed to have the
same stage efficiency if so designed that there shall be the same
heat drop in each. The points A, C, E, G, etc., will then lie on
a certain smooth curve of which the form is determined by the
constant stage efficiency and the initial state. If this curve can
be constructed, we may satisfy the requirement of equal heat
drop by distributing the points A, C, E, etc., on this line so as
to be at equal vertical distances apart. The points having been
thus determined, pressure and quality may be read off from the
chart.
It will be shown how such a curve may be constructed
more easily then by the step-by-step method, if the number of
stages

Up

is

large.

to this point

points A, C, E, G.

we have
But

considered only the similarly situated

let

us suppose that

what

is

wanted

is

not the steam state at entrance to the nozzles but that at exit
from them, after the reheat in the nozzles but before the further
reheat at the constant lower pressure has occurred. This will
evidently be given by the points B, D, F (Fig. 4). If the stages
are all alike and the intermediate pressures have been so determined, by distributing A, C, E, G, etc., at equal vertical intervals, that the heat drop is the same in each stage, the steam
speed and the reheat will be very nearly the same in each set of
nozzles.
The reheat after passing the nozzles will then be the
same in all the stages and B will be as far below C, measured
vertically, as D is below E, F below G, and so on.
It is true,
this amount can not be determined exactly, but it can be estimated with sufficient accuracy, and furthermore if there are
many stages so that the total heat drop in each is itself small, the
vertical distance between C and B will be still smaller and no
very great error can be made in the position oi B. If, therefore,
the cur\^e AC EG has been drawn the points BDF, etc., may all be

Bulletin of the

598

found very easily

any other

if

Bureau

of

Standards

any one can be found.

[Vol. 7,

The same

is

No. 4

true of

set of similarly situated points.

We

have referred particularly to the impulse turbine, but in


designing a turbine of the Parsons type a knowledge of the form
of the curv^e through any set of similarly situated points is equally
In either case

useful.

it is

essential that

we should

estimate the

steam quality at various points in the turbine, for otherwise the


cross-sectional areas can not be properly proportioned to pass the
required amoimt of steam at the velocities for which the blading
and nozzles have been designed, the desired pressure distribution
will not exist, and the whole working of the machine will depart
from the intentions of the designer, presumably to its disadvantage.

Whatever the

set of similar points chosen,

it is

evident that the

smooth curve drawn through


these points more and more closely as the number of stages between
the two limiting pressures increases and the heat-drop through
true expansion line will approach a

each stage diminishes. We shall speak of this limiting curve, for


an infinite number of stages of equal stage efficiencies, as "the
expansion line."
9.

COMBINED EFFICIENCY OF SIMILAR STAGES IN SERIESTHE


REHEAT FACTOR

Let us consider a single stage, of efficiency e, working with either


wet or superheated steam between two pressures which differ very
little.

^<p
Fig. 5

Let

(Fig.

5)

be the

initial

steam.

To

follows:

Measure a short distance

and

construct the expansion

AC

the final state of the

we may proceed as
downward from the initial
line

The Reheat Factor

Biukingham]

Divide this at

point.

AB

ADJAC =

so that

599
,

the given

stage

be an element of the expansion line.


Go on in the same manner from B, and continue till the lower
pressure limit has been reached. The smaller the steps the more
approximate the result if the graphical work is exact.
From the geometry of the figure we have

efficiency;

then

will

tan

/8

tan

i-e

or by equation

(7)

(6)

k tan

yS

i-e

As the pressure falls, the temperature


and the expansion

6 also falls;

(8)

hence ^ decreases

concave upward.
Let us next consider expansion of wet steam through a number
of successive stages, of equal stage efficiency e, between the
line is

9>

Fig. 6

pressures pi and p2 which differ so that the pressure-ratio pjpo


a rather large number, 30 for example.

is

be the initial state of the steam and C the state


Let ^C be divided at D' so that
after isentropic expansion to p2.
Let

(Fig. 6)

AD'IAC =
stages

Then if the combined efficiency of the whole set of


were the same as the stage efficiency, the point J5' would
.

represent the final state of the steam.

In reality, however, the expansion line starts from A with a


sharper slope than the line AB^ because the isopiestic p^ has a

Bulletin of the

6oo

Bureau

sharper slope than the isopiestic

Standards

of

p.^.

\Voi.7,no.4

The expansion

line

then

cun^es gradually, crossing the intermediate isopiestics at such


angles that equation

and at a point
actual heat-drop

bined efficiency,
the ratio

in

be

satisfied,

such an angle that

isopiestic p2 at

shall

(8)

which

AD
c

is

is

its

and meets the

tangent

parallel to

evidently somewhat below B'

therefore greater than

= ADIaC,

ADJAD'.

is

is

AB'
The

the com-

greater than the stage efficiency

This ratio

factor" and will be denoted

AD' and

final

by R.

is

known

Its value,

as the ''reheat-

which

in practice

seldom greater than i i depends on the ratio of the initial and


It could be found
final temperatures and on the stage efficiency.
in any particular case by step-by-step construction of the expansion line which would give us the position of B, while that of B'
is

on the contrary, the value of R is


known, the point B may be foimd without this graphical work.
We have now to show by a consideration of the form of the
expansion line how the value of R may be determined a priori,
for wet steam, i. e., for any part of the expansion line which
is

by AD'lAC = e.

given

lies

within the saturation


10.

If,

field.

FIRST APPROXIMATION FOR

THE REHEAT FACTOR R

The total curvature of the expansion line AB (Fig. 6) is fixed


by the stage efficiency and the difference in slope of the limiting
isopiestics; but the distance BB' and the value of {R-i) depend
on how

this curvature

is

on how the isopiestics


they change rapidly in

distributed, hence

are distributed between p^ and p^.


If
direction in the vicinity of p^, the expansion line
will curv^e
In
rapidly at first, B will be close to B' and {R-i) will be small.

AB

the opposite case {R-i) will be large.

We

have therefore to

consider the distribution of the isopiestics.

be found, upon examination of an accurate chart, that


the isopiestics if produced do not meet in a point; but that their
intersections are so far to the left and their whole divergence
within the range of pressures used in practice is so small that
only a very slight change would be needed to make them all
meet in a single point. We shall first proceed upon the assumption
It will

The Reheat Factor

Buckingham]

60

that they do thus meet in a single point, comparing the results


later with those obtained from a more exact assumption.

Let
let

be the point of intersection of the isopiestics;


be measured from this point, and let them be plotted

(Fig. 7)

H and

cp

Fig. 7

on equal

scales so that ^

substitute in equation
,

tan

=1

then have to

(7)

dH

p= acp

r.

-1

a=~
cp

dH
dcp

tan

which gives us

or, after integrating

We

in equation (8).

H=
cp

between any two points on the expansion

line,
e

^2^2^^' =H^Cp^~-

(9)

Noting that by the geometry of the figure


(Pi^Hs
(p2

equation

(9)

may be

^^^
"

H2

//a

_ tan

H^

A-2

^ 62

Ian a^

6^

reduced to the form

H,
and the heat-drop from

to

\e,)
is

therefore given

//.-//.^hIi-^uj

by the equation
(.0)

Bulletin of the

6o2
If

Bureau

of

Standards

[Voi. 7.

No. 4

the combined efficiency of the stages were the same as the

e{H^H^=eHii^\

stage efficiency, the heat-drop would be

Hence the value

of the reheat factor

is

(A)

From
of {R

this

we

equation

i), which

obtain, for example, the following values

give an idea of the order of magnitude:

TABLE

11.

p\

Pi

e=0,2

=0.6.

300

50

0.072

0.034

50

.096

.047

300

.168

.080

300

20

.100

.048

20

.083

.040

300

.186

.088

PRACTICAL WORKING

METHOD FOR FINDING THE VALUE OF R

While equation (A) is not very complicated, it is inconvenient


because it requires our working with absolute saturation temperatures instead of with pressures, and it would not in this form
be of any practical value. We therefore proceed to develop a
simpler method for obtaining values of R, which shall be sensibly
the same as those given by equation (A).

i) were computed by

equation (A) for expansion


from 277.4 pounds to eleven lower pressures, the lowest being 0.31
pound, with stage efficiencies 6 = 0.2, 0.3.... 0.7. When these

Values of

(i^

i) were

they gave sensibly a


straight line for each value of e.
It was found that all the values
could be represented very well by the equation

values of

(i^

i?

-I=

plotted against log

.207 (0.975

- e)

d^,

{log,, 0,

- log,,

equation represented equation (A)


evidently have the relation

If

this

(i?,,-l)+(i?,3-l)

6^)

exactly,

(-??n-l)

(B)

we should
(C)

The Reheat Factor

Buckingham]

where

603

and Ri^ are the values of the reheat factor for


expansion between d^ and 62, 62 and ^3, and 0^ and 6^, respectively.
It will be seen by reference to Table I that this relation is in fact
Ri2, R23,

nearly satisfied in the four cases there given.

Taking equation

(C) as sufficiently exact,

sent equation (B)

by a

find the value of

(R i)

straight line for

we could then

any given value

of

repree

and

for expansion with this stage efficiency

between any two absolute saturation temperatures, by taking the


If, finally,
difference of the ordinates at these two temperatures.

instead
of
against
the
inconwe plot (R i) against p as abscissa
venient log d, we shall have reduced our method for finding R to
practical shape.
A single curve is appHcable only to a single value
of e, but we may either work to other values of e by means of
equation (B), or plot a number of curves for different values of e
and interpolate when making the readings of (R i).
In Table II are values of R computed by equation (B) for expansion from 350 pounds to 18 lower pressures, with the stage efficiencies

= o.i,

0.2,

etc.

0.8

TABLE

II

=0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

300

1.0075

1.0066

1.0058

1.0049

1.0041

1.0032

1.0023

1.0015

250

1.0161

1.0143

1.0124

1.0106

1.0088

1.0069

1.0051

1.0032

200

^2

1.0265

1.0235

1.0205

1.0174

1.0144

1.0114

1.0083

1.0053

150

1.0394

1.0349

1.0304

1.0259

1.0214

1.0169

1.0124

1.0079

100

1.0569

1.0504

1.0439

1.0374

1.0309

1.0244

1.0179

1.0114

70

1.0717

1.0635

1.0553

1.0471

1.0389

1.0307

1.0225

1.0143

40

1.0936

1.0829

1.0722

1.0615

1.0508

1.0401

1.0294

1.0187

30

1.1047

1.0927

1.0808

1.0688

1.0568

1.0449

1.0329

1.0209

20

1.1191

1.1055

1.0919

1.0783

1.0646

1.0510

1.0374

1.0238

15

1.1292

1.1144

1.0997

1.0849

1.0702

1.0554

1.0406

1.0258

10

1.1429

1.1265

1.1102

1.0939

1.0775

1.0612

1.0449

1.0285

1.1545

1.1369

1.1192

1.1016

1.0839

1.0663

1.0486

1.0309

1.1651

1.1462

1.1274

1.1085

1.0896

1.0708

1.0519

1.0330

1.1808

1.1601

1.1395

1.1188

1.0981

1.0775

1.0568

1.0361

1.1926

1.1706

1.1486

1.1266

1.1045

1.0825

1.0605

1.0385

1.5

1.2102

1.1782

1.1552

1.1322

1.1092

1.0862

1.0632

1.0402

1.0

1.2120

1.1878

1.1636

1.1394

1.1151

1.0909

1.0667

1.0424

0.5

1.2309

1.2045

1.1781

1.1417

1.1253

1.0989

1.0725

1.0460

74356 12

Bureau

Bulletin of the

6o4

these figures any one

From

differ

from

= o.i;

this only in

ing this curve for

by the

Standards

desires

[Voi. 7,

can plot the

The curve shown on Plate

curves for himself.


the values for

who

of

No. 4

set of

was plotted from


for larger values of e the curves would
a linear reduction of the ordinates. Hav-

= o.i we may then

find the value of

(Ri)

following practical

Rule Read from

the curve the values of

(R i)

at the

two pres-

sures between which the expansion takes place; their difference is the
desired value of (R

i)

for

= o.i.

For any

other

value of

e,

'^'^

multiply by the factor

numerical tests have shown that values of R found by this


rule agree with those found from equation (A) much more closely
than any stage efficiency is ever known a priori, the discrepancy
being seldom over o. i per cent.

Many

The

rule

may

therefore be regarded as a satisfactory substitute

but it remains to be shown that


obtained from an admittedly only approxi-

for the use of equation

equation (A)

itself,

(A)

mate assumption regarding the

distribution of the isopiestics,

gives sufficiently correct values of the reheat factor.

has

now
12.

This question

to be taken up.

DISTRIBUTION OF THE ISOPIESTICS ALONG AN ISENTROPIC

be the value of the entropy at some fixed isentropic


Let 2 be used to denote the value of // at a point on this line.
Let

cpQ

line.

In

was assumed that the isopiestics all met in a single


point.
If H and cp were both measured from that point we should
then have Zjcp^ = 0. li H and (p were measured from some other,
point, we should have in general
section 10

it

where

and Z

= a + bZ

(11)

the absolute saturation temperature of any isopiestic


the value of
at the point Vv^here this isopiestic crosses the
is

is

isentropic

^ = (p^.

Equation

assumpReadings on the

(11) therefore represents the

from which equation (A) was deduced.


chart show that it is by no means exactly fulfilled.
of Z from the Hep chart of Marks and Davis at 9>o =

tion

By

readings

1.384, 1.544,

The Reheat Factor

Buckingham}

and

605

was found that in each case the values could be


represented by an expression of the form
1.704,

it

\oge = A-\-BZ

(12)

nearly though possibly not quite as closely as the readings could

The values

be made.

of

A and B

varied with the value of

number of readings were made between the isopiestic


This empirical equation
/> = I pound and the saturation line.

H during isentropic expansion

(p^.

for
for

one which I do not


remember to have seen given and which might prove useful in
other work than the present.
Since readings from the chart can not be made with any great
the change of

accuracy, recoiurse was

now had

is

to interpolation in the tables.

were computed at ^0 = 1.384 for 21 presValues of Z for the same


sures from 0.5 poiuid to 513 poimds.
temperatures were then computed by the equation

Values of

(i.

e.,

of Z)

Z = 1645

(13)
%,o<^- 3750.2
The greatest difference between Z obtained from the table and Z
calculated by equation (13) was 0.7 B. t. u. and the average
The differences showed a sysdifference was only 0.3 B. t. u.
tematic nm, but it was wave-like and not progressive, the value
passing through zero in the vicinity of 550, 700, and 900

absolute F.

Equation

a quite exact empirical representation of the facts, as given in the tables, over the whole range of
pressures used in present steam practice. At any other isentropic
where the entropy has the value (p, equation (13) leads evidently
(13) is therefore

to the equation

z = 1645

log,^

- 3750.2

e{(p

- 1 .384)

(14)

which must diverge from the tabulated values by precisely the


same amounts as equation (13). For security this was checked
through by interpolation in the tables for ^ = 1.684, from the
isopiestic p =0.505 Ib/in^ up to the saturation hne.
Since equations (13) and (14) are quite exact, it is evident that
equation (12), which is equivalent to (13) must be less exact when
(po differs much from the value i .384, though it is an approximation
nearly or quite sufficiently close for use with the chart alone.
,

6o6

Bureau

Bulletin of the
13.

of

Standards

[Vol. 7.

No.

4.

SECOND APPROXIMATION FOR R

Since equation (14) or its special case (13) is a much closer


representation of the facts then equation (11) from which (A)

was deduced,

it

follows that

an expression deduced from

(13) or

more nearly correct values of the reheat factor than


equation
(A).
We now proceed to deduce such an
are given by
(14) will give

expression.

/
A,

\y
/

i
E

<p

<Po

Fig. 8

AB

Let
line.

(Fig. 8)

Then,

if e is

be an infinitesimal element of the expansion


the stage efficiency, we have

CD
(15)

DA
Now we

have

DB

"P/p

Also

DA
DB
where the symbol d

refers to

-dH
d<p

change along the expansion

line.

Substituting these values in (15) gives us

i-e

dcp

^dH^^-'T

(16)

an equation which must be satisfied at all points of the expansion line and in all cases, regardless of the distribution of the
as

isopiestics.-^
^

The

validity of this general equation

is

not limited to the case of wet steam.

The Reheat Factor

Buckingham]

607

Let Ao be the intersection of the isopiestic through A with the


isentropic (p = (po, and let Z be the ordinate of this point.
Then

we have

CD H-Z ^^
AoE DB 9-90

EA

Whence

H-Z

9-<Po=^
Differentiating

by

H and muhiplying by
dZ

ndcf)

.^

gives us
.1

dO

another general relation, which must hold for all points within
the saturation field and for motion of the state point in any direction whatever.
By comparison of equations (16) and (17) we get

7\d

fu
If

we now

let <^o

and equation

log 6

dZ

= 1.384, we may, as shown


loge = A-\-BZ

(18)

may

in section 12, set


(19)

be reduced to the form

^ = e\^i+B{H-Z)\

(20)

a differential equation for the expansion hne in terms of


This equation is satisfied by setting

H = Me^''^+Z +

o\

H and Z.
(21)

eB

the arbitrary constant, to be adjusted so that the


curve shall pass through the given initial position of the state
in

which

is

point.

From

equation (19)

and equation

(21)

we

may

get

BZ = log 6 -A,

whence

be written

^ = ^Vz
+ L^
eB
e^

(22)

Bulletin of the

6o8

Let dy Hy Zy refer to the


expansion line, and let

Bureau

of

initial state,

Standards

[Voi.

(22) takes the

no. 4

the starting point of the

-^ = H,-Z,-^=L
Then equation

?.

(23)

form

and the heat-drop along the expansion line between two points at
the absolute temperatures d^ and ^2 is given by

H.-H.^Z.-Z.+Lli-f^y]

(25)

For isentropic expansion e = i


Hence by setting e = i in equations (22), (23), and (24) we have, along the isentropic through
.

the starting point of the expansion

line,

H = N^+Z

(26)

N = H,-Z,

(27)

where

and the isentropic heat-drop from the


final pressure where 6 = 6215 given by

initial

point at 6

H,-H, = Zy-Z, + N(^i-^j


The

in

= 6^

to the

(28)

reheat factor therefore has the value

which

N = H,-Z,

(D,)

L = N-'-^

(D,)

Beside necessitating our working with absolute temperatures, the


determination of R by equation (D) requires our finding the values of Zi and Zj either graphically or by computation from equation (13).
As a means for the practical computation of R for

use in designing, equation (D)

be entirely worthless; but

it

is

evidently so inconvenient as to

has a value as confirmatory of the

The Reheat Factor

Buckingham]

results obtained

by the

609

practical rule given in section

1 1

for

it

has been shown that that rule gives sensibly the same results,
within the range of present steam-turbine practice, as equation

and furthermore that equation (D), based on a much more


accurate assumption than that which underlies the deduction of
(A) must give more accurate values than (A) and therefore than
the rule. If, therefore, equation (D) gives sensibly the same values
as the rule, the values obtained by the rule mav be relied upon
(A),

as sensibly correct.

TESTS OF THE RULE FOR FINDING THE VALUE OF R

14.

The

paring the results


in section

of

plotted against

1.5

computing

(D)

and com-

rule given in

1 1

The values
points;

by equation
with values obtained by the

tests consisted in

log^Q

and a

The values

cp

chart at

straight line

the scales w^ere 100 B.

mm=5o.oi.

read from the


t.

of Z^

tations Tvere read from this plot.

u.

=75

^=

1.384 were

drawn through the

mm

and, for

and Z2 needed
The slope of

in the

log^^^

6,

compu-

this line

when

reduced to terms of naperian logarithms for use in equation


The readings of H^ for use in equation
(D2) gave 1/5=723.
cp chart.
More exact results might
(DJ were made from the
have been had by obtaining all these values by computation
instead of graphically, but the precision of the graphical method
It will not do, however, to compute
is sufficient for our purpose.
H^ from the table, but take Zj and Z2 from a plot or from an
equation derived from readings on the chart, for considerable
The chart, while self-consistent
errors are thereby introduced.
and quite accurate enough for all ordinary purposes, is not an
exact representation of the tables, owing doubtless to difficulties
in printing.
This, however, need not occasion any lack of confidence in the chart, for the test of its accuracy is a severe one, as
will be found upon making a few computations by means of
equations (D), (DJ, and (D2).
A number of test computations covering the part of the saturation field, which is of practical importance, were made by the
foregoing method. Values of R for the same cases were computed
by the rule of section 1 1 and in many instances also by equation
(A).
On the whole the values obtained by the rule agreed with

||

6io

Bulletin of the

Bureau

of

Standards

those obtained by the use of equation (D)

[Vol. 7,

somewhat more

No. 4

closely

A summary of the results of


than did values from equation (A)
these tests is given in the following table
.

TABLE
Value of

X[R

by equ.

(22)

/? by

III
rule]

expressed as percentage of

(^1-/^3)

pi

P\
lbs/in

Xx

6=0.1

277.4

12.7

0.981

45.5

277.4

.981

185.0

277.4

.981

44.0

66.2

1.00

33.7

21.9

.99

33.7

21.9

.88

33.7

21.9

.77

20.0

20.0

.99

+0.04

20.0

20.0

.84

20.0

20.0

.79

20.0

120.0

.99

20.0

120.0

.84

+
+
+
+

300.0

300.0

1.00

110.0

110.0

1.00

35.0

35.0

1.00

8.0

8.0

1.00

In the
ratio,

6=0.2

6=0.4

6=0,6

+0.03

-0.02

-0.02

6=0.7

abs

pi

the

first

+
+
+

.04
.03

.06

.09

.20
.09

.06

+
+

.07

.01

.04

.04

.04

.02

+
+

.03

.07

.08

+
+
-

+
+

.04
.04
.06

.05

.04
.01

.06
.07
.03

+
+

.04

.10
.22

.10
.05

.07
.08

+0.03

+
+

.03

.19
.13

.01
.01

.02
.01

.04
.04

.25

.10
.01

.02

three columns of the table are given the pressure

initial pressure,

and the

initial

dryness factor; and

it

be seen that the tests are sufficiently varied.


^
The object in using the reheat factor is to determine the position
of a point B with an error in vertical position which shall be
negligible in comparison with the isentropic heat drop {H^ H.^)
This difference of position, in percentage of (H^ H^), is found by
will

multiplying the difference of the two values of R found by the


two methods by the value of e, since R is always nearly unity.
Accordingly, the values in the table may be regarded as the changes

of height of the point B, in per cent of

{H^-H^, caused by

The Reheat Factov

Buckingham]

6ii

changing the method of computing R. The positive sign means


that the rule gives B a higher position than it should have according to equation (D)
On account of errors in reading H, Z, and i^ i from the curves,
differences of less than o.i have little or no significance unless
pjp2 is very large. It will be seen that the discrepancy never
exceeds one-fourth of i per cent and attains the value o.i per
cent only when the pressure ratio is large.
Since one-fourth of i per cent is, for the designer, a negligible

we may conclude that since one of the methods cergives a much more accurate value than the other, the error

quantity,
tainly

involved in either method

is

negligible

and the value

of

is

accu-

enough when found in either way. It follows that the values


of R obtained by the working rule given in section ii may be
relied upon for wet steam as being sufficiently accurate for use in
rate

designing.
15.

Having thus a

we

USES OF THE REHEAT FACTOR

practical

method

are able to plot the expansion

of finding the reheat factor,


line.

If

p^^

and

terminal pressures, the position of the final point

p2 are the

may

be found

by the equation

H,-H, = Re{H,-H,)

We

(29)

next take, instead of />2, the pressure corresponding to some


isopiestic about halfway (on the diagram, not numerically)
between the two terminal isopiestics. Using the rule to find R
for expansion from the initial to this intermediate pressure, we
compute the position of a point on the expansion line where it
Having now three points
intersects this intermediate isopiestic.
It
of the expansion line, we may sketch it in as a circular arc.
will then be evident whether or not it is worth while to go on to
determine still more points. Unless the expansion ratio pjp2 is
large, even a single intermediate point will usually be needless
and, after the final point B has been found, a straight line AB will
be close enough to the true expansion Une, even in the middle.
This, however, is a matter for the discretion of the individual
designer, who knows what reliance he may place on the value of
the stage efficiency on which his work

is

based.

6i2

Bulletin of the

Bureau

of

Standards

[V01.7.N0.4

In designing a turbine for new conditions, it may be worth


while to make a number of preliminary sketch designs which need
not be worked out in detail further than to compute the probable
influence on the water rate of variations in the arrangement of
In such work, what is needed is not the form of the
the stages.

expansion line but merely the actual heat drop {H^ H^, which
may be expected in a given set of similar stages. If the expansion
ratio is large, neglecting the reheat factor may introduce an error
of 5 per cent or more, which is undesirable and may be avoided
by a single computation of the reheat factor for use in equation
(29).
16.

GENERAL REMARKS

The foregoing methods

are

vaHd only when the

isopiestics are

straight lines; they therefore fail for superheated steam.

The

curvature of the isopiestics in the superheat field depends on


the specific heat of superheated steam at constant pressure.

were exactly known and expressible in simple mathematical form, the development of a general differential equation
for the expansion line would be a difficult matter and the task
of integrating it would very possibly present insuperable mathematical difficulties. But without attempting this general solution a few pertinent remarks may be made.
(a) Consideration of the Hep chart shows that an isentropic
expansion between two given isopiestics involves a greater change
in temperature for superheated than for wet steam, as may
also be seen from the pv chart.
A dissipative expansion line

Even

if

this

starting at a point in the superheat field will therefore fall

more

than if it started at the same pressure within


the saturation field and the reheat factor for a given pair of
terminal pressures and given stage efficiency will be greater
than for wet steam between the same pressures. Hence, if we
use the value of R computed for these pressures for wet steam,
we shall underestimate the actual heat-drop and combined
efficiency and be on the safe side as regards the economy
sharply at

first

expected.

is

Up

to 100

superheat the curvature of the isopiestics


so small that the error can at most not be important if we simply
(6)

The Reheat Factor

Buckingham]

use the value of

for

613

wet steam between the same pressure

limits.
(c)

The value

of the quantity (R

i)

decreases with the pres-

sure ratio pjp2 and decreases as the stage efficiency e increases.


Examination of the H(p chart in connection with Table II shows

very poor, with superheats up to


150 or 200 F, which are unusual as yet, the expansion ratio,
and therefore (R i), can not be large for expansion entirely
within the superheat field, so that (R i) can not be subject to
a large error. With poor efficiency the reheat might be enough
to keep the steam dry a long way down in the turbine, but such
a case is not of great commercial interest and does not demand
that, unless the efficiency

is

close designing.

superheat has not as yet been much used


in the Parsons type of turbine, so that the superheat field is of
interest mainly to the designer of turbines in which the first
(d)

few

high

initial

stages, at least, are of the impulse type, not requiring fine

radial clearances.

In such a turbine, the

first

one or two stages

and will get rid


of the superheat, so that the remaining stages are working with
wet steam. It is not a serious matter to treat one or two stages
separately, by determining the state-points graphically, and
For the
it is usually necessary to do so for at least one stage.
first stage is usually not quite Hke the later stages and has not
the same stage efficiency, so that an expansion fine drawn for
constant stage efficiency could not fit the facts in any event.
(e) The whole idea of the reheat factor and the need of drawing
the expansion fine arises from the fact that the combined efficiency of a number of similar stages in series is greater than the
The method developed for
efficiency of each stage separately.
finding the reheat factor was based on the assumption that the
heat-drop in each stage was small, for it consisted in developing
and integrating a differential equation for the expansion fine.
For a few stages with very high steam velocities, the results
would not be exact, but the whole matter is also, in that case,
of no importance; for the graphical construction for a few stages
is a simple operation and involves no great waste of time.
will usually

have rather large pressure

ratios

6i4

Bulletin of the

Bureau

of

Standards

[Vol.?, No. 4

term has been used in


Let p^ and p2 be
this paper, is not applicable to a single stage.
the limiting pressures of a given single stage, and A (Fig. 9) the
(/)

The notion

of a reheat factor, as the

initial state.

^H

Fig. 9

Then

if

is

the final state

AD H.-H^
AC~H,-H,
is

the ratio of the mechanical energy produced to the

maximum

may

be assumed in an
impulse stage, the difference of kinetic energy between similarly

possible in adiabatic expansion.

situated points
stage.

This

is

is

is

i.

initial state

AD/AC = e

is

the efficiency of the

at p2,

are

whether the steam is wet or superheated,


AC is vertical and DB is horizontal. If the
e.,

net efficiency of the stage

the

as

true regardless of whether the isopiestics

straight or curved,
so long as

negligible,

If,

is

given beforehand,

is

thus fixed by

A and the final pressure p2 and there is no

question

of a reheat factor.

But the

efficiency

may have been

obtained in either of two

ways. If it was obtained by experiments on a stage similar to


the one in question and working under similar conditions, there
is

nothing more to be said.

But

if it

was computed from a velocity

diagram drawn with proper allowances for velocity losses in nozzles and blades, with subsequent correction for windage and
leakage, the case may be a trifle different in theory though hardly
in practice.

The Reheat Factor

Buckingham]

A and therefore C
make AD/AC = where
Let

be given, and
e

is

615

let

be so placed as to

the computed efficiency.

It

now

becomes a question of when the various elements of the total dissipation occur and of the temperature of the steam when the
In an impulse
various elements of the reheat are added to it.
stage, in general, only a small fraction of the dissipation occurs

almost all added to the steam


after the lowest pressure and temperature have been reached.
The actual expansion line is therefore nearly coincident with the
broken line ACB, which would be the expansion line if all the disHence, if e were computed
sipation and reheating took place at p2.
by some method which treated the reheat as all added at the low
in the nozzles,

hence the reheat

is

added at a higher pressure


and temperature in the nozzles, the efficiency thus computed
would be a trifle too small and should be multipHed by a factor
pressure, while in reality a little of

it is

analogous to the reheat factor which, in

effect, is

allow for the fact that in a series of stages the reheat

introduced to
is

distributed

The

instead of being concentrated at the end of the last stage.

improvement

in accuracy

due to the use

of such a factor would,

however, be altogether illusory until our experimental knowledge


is sufficient to permit of a far more accurate a priori computation
of the stage efficiency than is at present possible.

To apply a

reheat factor computed for the case of continuous

expansion through an infinite number of stages, to expansion


between the same pressure limits through a single stage, is altogether erroneous.
(g)

In conclusion

we may say

that, so far as present practice

is

concerned, the fact the method given in this paper for finding the

value of the reheat factor

an important

is

on

applicable only to wet steam,

is

not

In the designing of
turbines with many similar stages, its use may save a great deal
of time that would otherwise have to be spent over the drawing
board.
NOTE (TO SEC. 1).LIMITATIONS OF THE THEORY
restriction

its

usefulness.

In the total energy equation, the terms To and T are taken as


including only kinetic energy of the axial component of the
velocity of the fluid.
This is done because the only kinetic energy

Bulletin of the

61

Bureau

of

Standards

Voi.

7,

No. 4

which can be utilized mechanically is that due to the general


motion of the fluid as a whole. Kinetic energy which is not
utilizable, such as that of eddy currents, is for our purposes not to
be counted as "mechanical" energy. It must ultimately be dissipated into heat; and motions which are already so unordered as
to be incapable of mechanical utilization may be considered as
It is of no
already, for our purposes, completely dissipated.
importance to the practical theory of steam flow whether they
have already become completely unordered in the molecular sense
or not.

meant by the " state " of a


mass of fluid when it is the seat of eddy currents which are still on
a relatively large scale though already beyond our power to utilize
question then arises as to what

is

In going on to complete dissipation,


which is merely our convenient way of say-

their kinetic energy directly.

they

"

produce heat,

"

ing that the dissipation tends to raise the temperature of the fluid.

We are thus led to ask precisely what is meant by the temperature


and a very little considerastate " of a mass of fluid can

of the fluid at a given point in its course,

tion at once

makes

it

clear that the

''

not be precisely defined at all, unless it is a state of quiescence.


Turbulent, states are not capable of precise description, and the
terms pressure and temperature have no precise meaning when
referring to a fluid in a condition of turbulent motion nor even
when referring to a volume element of a fluid moving as a whole
with a rapid acceleration.
But though the precise treatment of the vastly complicated
process of steam flow through a turbine is quite beyond our powers, the only important question is whether our theory is nearly
enough correct for practical purposes whether it represents the
facts sufliciently well to be useful in predicting what will happen
practically in a future case.
This question may without hesitation
be answered aflirmatively. We neglect certain recognized errors
in the theory and while we can not say, a priori, just how large

these errors are,

we

find a posteriori that they are negligible

if

we

use the theory with good judgment and do not throw common
sense to the winds.
If one attempted to measure the temperature

steam issuing from a de Laval nozzle by putting the bulb of


a mercurial thermometer directly in the jet, and then assuming
of the

The Reheat Factor

Buckingham]

that the thermometer could endure such treatment

617

compared

the readings with the theory, one might find very large discrepan-

But no sensible person would expect anything else.


Every physical theory is built upon a simplified ideal picture of
the main outlines of the known facts. The simplifications, when

cies.

recognized, constitute the assumptions or hypotheses stated as a


basis for the theory.

The degree

of simplification permissible

depends on the acctiracy expected in using the theory for prediction and is limited by the accuracy attainable in the experimental
In these respects the
verification of the results of prediction.
theory of the steam turbine does not differ at all from other physical
theories only it happens that the accuracy required of the theory
is low on accotmt of experimental difficulties, and that the things
which the theory neglects in making its ideal picture of the facts
are so obvious that anyone can see them.
;

Washington, March

2,

191

1.

!--l

i\

200

PRESSURE

LBS./IN^
PIntc

I.

ABSOLUTE

MM^

220

240

260

280

300

320

340

You might also like