r~.~ H~ - .R4.C.

Technicas1 1lpvrt

.....
..... ....

C) L~ 03 ,AR
L~i~flRY ~ .. ,ION

MINISTRY OF SUPPLY

AERONAUTICAL RESEARCH COUN4CIL JN2118.

REPORTS AND MEMORANDA

Al 1ethod of Performance Est'rna~tio'.'n"

.forA Axial-Flow Turbirlea>..
7 .~ By

I D. G IN~tY _r.
- .Accesippf For.

...- * DTIC rAR

/~'X1Cowii Cop~~I:rlgh-Msved, -Ava inbility Codez

LONON:HER MAET' SAINR OFFICE iiOco

lipmm-NINIL ST1Il S NET

.TS lk . .bs

AIvd far. publli'MOA"&i

 A Method of Perfbrmance Estimation for
Axial-Flow Turbines
By
D. G. AiNL.Y and G. C. R. MATIIESON

COMMUNICATED BY THE PRINCIPAL DIRECTOR OF SCIENTIFIC RESEARCH (AIR

MINISTRY OF SUPPLY

Reports and Memoranda No. 2 9 74 t,

December, x9 5 I

Supnmary.-A method for calculation of the performance of conventional axial-flow turbines is presented. It makes
use of data derived from the analysis of a large number of turbine tests and other associated test work reported elsewhere.
The method enables the performance of a turbine to be calculated over a wide part of its full operuting range. It is
estimated that the tolerance on the absolute values of gas mass flow and peak efficiency wiU be in the region of ± 2.0
per cent for efficiency and ± 3 per cent for gas mass flow on present day types of turbine.
The method is illustrated by a worked example.

1. Introduction.-The estimation of the performance of a gas-turbine stage under different

conditions of speed and pressure ratio has until recently been hampered by lack of reliable data
regarding the pressure losses and gas deflections through rows of turbine blading. Information
from cascade tests, giving the two-dimensional performance of blade sections, is not enough.
The three-dimensional effects in a turbine are so powerful that unless they can be estimated
with a fair degree of accuracy, prediction of the performance of blading fitted in a turbine is
largely reduced to the level of guess-work. The construction and operation of an increasing
number of rigs for the detail testing of gas turbines has made available a stock of test calibrations
from which some useful experimental data has been extracted by analysis. Ref. 1 gives an
account of the work which has been carried out at the National Gas Turbine Establishment
on this analysis and the conclusions which have been drawn from it.
This present report giv, a method of estimating the performance of a turbine, making use of
the data derived in that analysis. The method proposed applies to axial-flow turbines only.
Some of the basic data presented here will be of use in the performance estimation of some
types of radial-flow turbines (such as the Ljungstr6m type) but the three-dimensional flow
picture will be different as between the axial and radial types, and so the method cannot be
used diiectly for radial-flow machines.
2. Applicalion of the Method.-The procedure adopted in this method can be split conveniently
into two parts. The first is the determination of the characteristics of pressure loss and gas
efflux angle for each blade row in the turbine under a wide range of inlet conditions. The second
is to take a fixed set of conditions at inlet to the turbine and a given turbine speed and to follow
t N.G.T.E. Report R. I1l, received 8th May, 1952.

A CPYRIGHT @
1-a---'
CONTRtl... '1
 the course of a fixed mass of the gas from one blade row to the next, calculating the new gas
conditions at each step, so building up a complete picture of the flow at all points for the given
initial conditions. Then from the final picture, the necessary information such as gas angles,
stage and overall pressure ratios, stage and overall efficiencies and the gas mass flow can be
extracted.
The procedure of the second part must be repeated for several sets of initial conditions with
the same turbine speed and further points on a constant-speed characteristic curve can be
constructed. Another speed may then be selected and the calculations of the second part repeated
to give a second constant-speed curve. This may be repeated for further constant-speed curves
but, to save calculation, it is only necessary in many instances to calculate the performance at
two turbine speeds. A method is suggested of interpolating the characteristics at intermediate
speeds.
Since the turbine outlet conditions are not initially fixed, but determined by the final result
of a row-by-row progression, the direct calculation of a characteristic curve at constant
pressure ratio is not possible. If such characteristics are required they have to be obtained
by interpolation between series of constant-speed curves.

3. Limitations of the Method.-To predict exactly the flow at all points in a system such as a
turbine an almost infinite number of variables would ultimately have to be considered. . In
F contrast to this, to achieve simplicity in a method of prediction the number of variables must
L be reduced to a minimum. Thus any working method finally arrived at must be a compromise
between simplicity on the one hand and accuracy on the other.
In the present method only those variables which calculation or statistical experimental
evidence has so far proved to be essential to achieve a final tolerance of about ± 2.0 per cent
in efficiency and flow have been employed. This tolerance is about the same magnitude as that
which must be allowed in interpreting most current experimental data and higher accuracy
scarcely seems possible at the present time.
A major simplification employed throughout the following treatment is the widely adopted
one of considering the flow path through each stage at one diameter only-this diameter being
termed the' reference diameter.' The reference diameter selected in each stage is the arithmetic
r mean of the rotor and stator row inner and outer diameters (see Fig. 2). In adopting this procedure
it is assumed that in any one cross-sectional plane of the flow between adjacent blade rows the
total pressure, total temperature, and axial velocity are equal at all points. Such an assumption,
* - though often widely divorced from fact, may be expected to yield correct overall characteristics
of a stage if the gas efflux angles and pressure loss coefficients used for each blade row at the
reference diameter in the calculation are equal to the momentum mean values over the entire
cross-sectional plane which would be determined by a flow exploration in the actual turbine.
The accuracy of the performance calculation rests entirely upon the accuracy with which these
mean loss coefficients and gas flow angles for each blade row may be predicted.
The present method owes its accuracy to the fact that the loss coefficient and gas angle data
have been derived largely from overall tests on a variety of turbine stages (Ref. 1). At the same
time this sets a limitation to the method since the data so obtained applies primarily to blades
having a 'conventional' (Ref. 1) profile shape and the quoted accuracy cannot therefore be
guaranteed for widely different profile shapes. However, the data should apply to most blade
designs in current use in gas and steam turbines in this country and the U.S.A. so that calculations
made in accordance with the present treatment should prove to be of wide use.
Since the present treatment allows loss coefficieuts and effiux angles in any blade row to vary
with gas flow conditiQns, such as incidence and Mach number, it may be expected to give greater
accuracy over a wide range of operating conditions in a turbine than earlier and simpler methods
(e.g., Refs. 2 and 3) in which constant loss coefficients and gas efflux angles in blade rows have
been assumed.
2

a.
 4. Notation and Convention of Signs.-The notation and convention of signs for the gas-velocity
triangles in each stage is similar to that used in Ref. 4 and follows directly from that already
established in axial-compressor theory in this country. A complete list of symbols and a note
on the cenventiwn of signs is given in Appendix I ; the velocity triangle notation and sign con-
vention is further illustrated in Fig. 1 ; critical dimensions used to define blade form are further
illustrated in Fig. 3. Fig. 2 illustrates the choice of the reference diameter and the reference
stations between adjacent blades together with the dimensions used to define annulus areas at
the reference stations, blade height, and tip clearances.
In each stage the suffices adopted for absolute and relative pressures and temperatures at
each station are the same as the suffices used for gas angles and velocities .in the relevant velocity (
triangles. These suffices repeat themselves from stage to stage so that care must be taken when "j
setting out the calculation to ensure that pressures calculated at any station are associated with
the correct stage.
In section 5 (dealing with the estimation of loss coefficients and gas angles) and in Fig. 3
(illustrating critical dimensions used to specify blade shape) it should be noted that the notation
relates to a rotor row. This has been done for convenience since this same notation was used in
Ref. I (dealing with the derivation of the data). The procedure for determining the loss coeffi-
cients and gas outlet angles on a stator row is precisely similar-suffices , and . (or i and , if it
is a first-stage stator) merely being substituted for . and 2 respectively.

5. Determination of the Variation of Gas Deflection and Total-PressureLoss in a Row of Blades

with Incidence and Mach Number.-The derivation of the pressure losses and gas effiux angles
for each blade row follows directly from the analysis described at length in Ref. 1. The blade
sections at the reference diameter on each row are treated as the reference sections used for
predicting the aeriodynamic characteristics of each row.

Throughout the succeeding-treatment the following assumptions are made:

(i) The pressure-loss coefficients in each blade row defined by the ratio (inlet total pressure -
outlet total pressure)/(outlet total pressure - outlet static pressure), are not influenced
by the gas Mach numbers.
(ii) The gas outlet angles from a blade row are not influenced by the gas incidence angle.

The assumption that the loss coefficient is uninfluenced by gas Mach number is unlikely to lead
to appreciable error unless the blades have a high degree of curvature on the upper surface near
tf the trailing edge. Since such blades have been shown to have high losses when the outlet Mach
number approaches unity such blades are unlikely to be used in practice at high Mach numbers.
The assumption that outlet angle remains independent of incidence does not accord precisely
. vith fact but will lead to very little error over the efficient operating range of a turbine.
Neither of these assumptions are absolutely essential to the method of calculation propounded
in section 6. They are introduced primarily because they reduce the amount of trial and error
* necessary in the calculation. If in any calculation it is felt desirable to take account of variation
of loss coefficient and gas angle with Mach number and incidence respectively the general methods
outlined in section 6 may still bL. used.
The effect of Reynolds number-is discussed in section 8.

5.1. Derivation of the Gas Outlet Angle from a Blade Row.-Following the recommendations
made in Ref. 1 it is assumed that gas outlet angle from any row remains constant over the gas
outlet Mach number range 0.5 > M 2 > 0 and given over this range by:
=Ct, - 4(sic) .. . . . . . , .. 1 /

where =C2* (cos -

=-f ols), see Fig. 5
3
Al

~1
 I
e is the mean radius curvature of the upper surface of the blade profile between
the passage throat and the trailing edge (see Fig. 3).
s is the blade pitch.

At a gas outlet Mach number of unity it is assumed that the outlet angle is given by:
S- cos - ' AI/A,, .. .. .. .. .. .. .. (2)
where A, isthe passage throat area
A., is the annulus area in reference plane downstream of the blade row.

If the annulus walls at the ends of the blades are flared the throat area may be approximately
expressed by an empirical formula:
A, = (oIs),.I [5A, 2 + A.]6 (3)
where A,, is the annulus area in reference plane upstream of row
wA., is the annulus area in reference plane downstream of row.
Between M, =.0-5 and M, = 1-0 a linear variation of a2 may be assumed with reasonable
I accuracy. Alternatively a smooth curve with a point of inflection at M2 = 0.75 may be arbi-
trarily drawn similar to those shown in Fig. 17. A variation of this latter form will occasionally
be found more convenient in practice since it will be found to ease a trial and error step required
in the calculation when the flow approaches the choking flow (see sections 6.1 and 6.3 later).
The outlet angles as deduced in the above paragraphs apply to blades having zero tip or shroud
clearance. If a finite tip or shroud clearance exists then a small'portion of the flow passes through
the row with little or no deflection and the momentum mean outlet angle is thereby reduced
slightly.P
At low Mach numbers, if .a is the mean outlet angle from a row having a small tip clearance
and oc'is the angle deduced (M12 < 0.5) for zero clearance then oc. given approximately by:
is.
I C2 = tan-' {[l - X(k/h)(cosfl,/cosoc2')] tan cc
2' + X(k/h)(cosfl1 /cos c.') tan #,} (4)
where X = 1-35 for radial tip clearance
or = 0-70 for simple shrouded blade,

For complex shroud bands the expression (k/h) (cos #f/cosa.') may be replaced by wIW where
w is the estimate leakage flow round the band and W is the total flow. In this case X = 1.35.
At M, = 1.0 the angle is deduced from equation (2) by allowing for the tip clearance area
in the evaluation of throat area. Thus if A,' is passage throat area when the clearance is zero then
A. = A[1 - (k/h)] + A, where Ak is the flow area in the clearance space, e.g., if the tip diameter
is D inches and the radial tip clearance is k inches then A., = -Dk sq in.
If the row has a shroud then the throat area should be A, = A,' + A,..
5.2. Derivation of Pressure Losses in a Blade Row.-The pressure loss is subdivided into three
components, namely (i) profile loss (ii) secondary loss and (iii) tip or shroud clearance loss.

(i) Profile Loss.-Firstly profile loss is determined at zero incidence. The stalling incidence
of the blade row is then determined, stalling incidence being defined as the incidence at which
profile loss is twice the loss at zero incidence. Profile losses at incidences other than zero are
then obtained from the assumption that the ratio of profile loss at any incidence to profile loss
at zero incidence, Y/ Ypo), is a function of the ratio of incidence to stalling incidence, i/i,.
This relationship (derived in Ref. 1) is plotted in Fig. 6.

44
 Profile-loss coefficient at zero incidence is assumed to be a function of a., #,/L%,, s/c, and t/c
and is given by:
Y,-,= {YP,., + (fi,/c,) 2[Y1 ,,.- }t-l--)
Y... ,]S .. .. (5)
where. Yp,-.)is the profile-loss coefficient of a blade having/f, = 0 and same ix, and s/c as
the actual blade (Fig. 4).
Yp,--, is the profile-loss coefficient of a blade having =-
- oc, and same cc, and s/c
as the.actual blade (Fig. 4).
On impulse or nearly impulse blades the use of equation (5) should be restricted to
0-15 < t/c< 0-25. If the actual blade has tic greater or less than the limits quoted then the
loss should be taken 'as equal to a blade having i/c either 0-25 or 0-15 respectively (Ref. 1).
Stalling incidence is approximately a function of cc, fl/cc, and s/c. Fig. 7 enables stalliig
incidences of a wide range of blades having s/c = 0-75 to be determined and also shows the.
variation of stalling incidence and gas outlet angle with s/c. The procedure for determining
stalling incidence for a given blade for which cL,, f#,,and s/c is known is then as follows:
(a) From Fi2.7 find a,(slc = 0"75)/ , for given value of s/c and hence find oc,(s/c = 0.75).
(b) Determine fl/c,(s/c = 0.75) and from Fig. 7b find i,(s/c = 0.75).
(c) From Fig. 7a find Ai, for actual value of s/c and hence determined i,from:
i,= i,(slc = 0.-75) + Ai,.
This value of stalling incidence is, of course, only approximate, but the accuracy is considered
sufficient for the present purposes.
Having now determined i, and. Y(,-(,) the profile-loss coefficients, ever a wide range of incidence
can be deduced from Fig. 6..
(ii) Secondary and Tip Clearance Loss.-It has been demonstrated in Ref. 1 that secondary
losses may be conveniently expressed by an equation of the form:
CD, = ACL'l(sIc)
where Ais dependent primarily upon the degree of acceleration imparted to the gas as it flows
through the blade row. Similarly it was shown that tip.clearance losses may be expressed by:
CD, = B(klh)CL'l(slc)
where B is a constant.
Since the equation for secondary and tip clearance losses are sirixilar in form it is convenient
to treat them simultaneously. Thus, converting the drag coefficients into the more convenient loss
coefficients', we may derive an expression for the sum of the secondary and tip clearance losses
Y , + YA = [) + B(klh)][Cl(slc)]'[cos' ,/cos' ,.] .. .. .. .. (6)
where A.=f {(A,/A,)'/(1 - I.D./O.D.)}, see Fig. 8
A1 = A, 0 cos l .
As = A., cos a.
CLI(slc) = 2(tan c.,
- tan c,) cos a,,
a.m = tan - ' [(tan c:i+ tan C,)/2]
B = 0.5 for row with radial tip clearance or 0.25 for row with shroud seal.
Equation (6) may be used to find the secondary and tip clearance loss for a wide range of gas
inlet angles. The variations of -these losses with gas inlet angle to a blade row at large positive
and negative incidences is uncertain but reasonable correlation of test and calculated turbine
S
 performance has been obtained by restricting the use of equation (6) to - 1-5 < ili, < 1.0
At values of ili, > 1 0 the secondary and clearance loss coefficients should be assumed constant
and equal to the value when ili, = l- 0. Similarly when i/i, < - 1-5 secondary and clearance
loss is assumed to remain constant and equal to the value at il, = - 1-5.
Effect
(iii) to of Trailing-Edge Thickness on Blade Loss.-The preceding pressure loss coefficients
apply 'conventional' blades having a trailing-edge th~ickness, t,, roughly equal to 2 per cent
of the blade pitch. Reeman has shown theoretically that trailing-edge thickness can markedly
influence the loss and it was further indicated in Ref. I that theoretical estimates of the effect
of trailing-edge thickness based on total loss coefficients have compared favourably with observa-
tions made on turbine and engine tests. If the ratio of trailing-edge thickness to blade pitch,
t,ls, differs from 0-02 then the total loss coefficients deduced from the previous sections should
be corrected by a multiplication factor plotted in Fig. 9, this factor being derived from the
previous mentioned references.

6. Performance of a Turbine Stage (uncloked).-Having determined the characteristics of

each individual row of blades it is now possible to proceed with the calculation of the performance
of the complete turbine. At this point the following information should have been determined:
(a) The reference diamefer of each stage
(b) Annulus areas at the reference stations mid-way in the axial clearance spaces between
adjacent blade rows
(c) Estimated variation of gas outlet angle with gas outlet Mach number at the reference
diameter of each row
(d) Estimated variation of the total loss coefficient, Y, with incidence for each row.
An inlet mass flow, inlet pressure, and inlet temperature are now arbitrarily selected and at a
fixed turbine speed the course of the chosen mass of flow is followed from row to row, gas tem-
perature and pressure being determined at the entry and exit of each row. Kp and y are assumed
to remain constant throughout expansion of the gas through the turbine. The process will be
described in detail for one stage. Further stages should be treated in a similar fashion.
For accomplishing the calculations it is essential to have at hand a set of curves relating the
parameters V/VT and M with the non-dimensional flow quantity WVTIAP, these curves
being calculated for isentropic expansion of the gas from the stagnation pressure, P, and stagna-
tion temperature, T. The relationships between these parameters have been calculated for
S 7 = I1.3330, Kp = 12,300 ft/pdl/deg C., g = 32.-173 ft/sec2- and are tabulated in Appendix IV.
- Further relationships between non-dimensional flow parameters are also required and are
described later in section 6.1, and plotted in Figs. 10, 11 and 12. It is important that these
latter relationships should be calculated for the same values of y, Kp and g as the former. Since
the computation of these curves for various values of Kp and y is very laborious it is recommended
that all performance calculations on turbines should be made using some arbitrarily defined
standard values of. gas properties. If the turbine chararteristics are required for other values
Of 1fp and y the final characteristics obtained using the ' standard ' values should be corrected
by one of the simple methods described in Ref. 4.

6.1. Determinationof Gas Flow Conditions at Outletfrom a Nozzle Row.-Suitable turbine entry
values of inlet temperature, inlet pressure, inlet gas flow direction (usually axial at entry to the
first nozzle) should be decided and a first value for gas mass flow arbitrarily selected. If the
nozzle row is preceded by other turbine stages then, of course, these inlet conditions will be
already determined.
The pressure loss coefficient for the row is determined by the incidence, and the gas outlet
angle is a known function of outlet Mach number. Since the pressure loss coefficient for the
nozzle row is defined as (P - Po)I(P, - p,o) it is seen that both the absolute pressure loss
(P. - P.) and gas outlet angle depend upon the outlet conditions from the row which are as
6 .
yet unknown. At first sight this presents a problem which must be solved by trial and error,
i.e., arbitrarily selecting outlet conditions (Mo) until conditions are found which satisfy the
requirement of flow continuity. This trial and error process may be nearly eliminated as shown
in the following paragraphs.
If a blade row was assumed for the moment to have a,constant outlet angle and a constant loss
coefficient at all outlet Mach numbers up to unity then it would be found that the non-dimensional
inlet mass flow (W/TIA.P) would increase as M. increases until it reached a maximum value
at a value of M, slightly less-than 1-0. This phenomenon has been demonstrated in Refs. 6 and 7.
For convenience the value of M, at which WiVTi/A.Pi reaches a maximum will be termed
the 'critical' outlet Mach number. The actual values of- (WVTd-4,Pi),, M, ,t, and the"ratio
of pressure loss to inlet pressure at the critical flow conditions, (/Pi),. 1t, all depend upon the
value of the loss coefficient, Y,. Values of (WVTIAPi)., and (&/P,),, are plotted in
Figs. 10 and 11. Now it is found that over, the required range of M,(0 < M. < 1-2) and
Y,(0 < Y, <1 -0)the ratio of pressure loss to inlet piessure (@(Pj), can be related to the non-
dimensional inlet mass flow (WvTdIA.P,), by a single curve by plotting (@,P)1(@/P),, against
(Wi/TiAoPi)/ (WVTdIAoPi)a. This relationship is plotted in Fig. 12. Mathematically this
relationship is not exactly unique for all values of Y,but for 0 < Y,< 0-1 the maximum departure
from the curve plotted is very small and within 1 per cent. This error is insignificant for the
present purposes since a 1 per cent error in blade loss is equivalent to less than 0-2 per cent in
turbine efficiency when the efficiency is in the region of 80 per-cent or more.
By employing these relationships a procedure for determining the outlet gas condition from
the known inlet conditions may therefore be specified as follows:
(i) Select by conjecture a first approximation to outlet angle, co
(ii) From the known mass flow and inlet conditions find WVTIA 0 PA
[N.B.; A, = A., cos j.
(iii) From Figs. 10 and 11 find for the relevant value of Y, the value of (WVTA.Pi) ,aand.

(iv) Calculate (WVTiAoPi)/(WVTiAoP)a, and hence from Fig. 12 find the corresponding
value of (@IPi)I(@IPi) r1t
(v) From the above determined values of (@/Pi),,i and (@,Pi)P(/Pi) ,, find (/P)
(vi) Find Po,, given by P, = Pi - @, and hence determine WVToIAoP°
[AT.B., for an uncooled nozzle row T", = TJ.
(vii) From the value of W/To/AP the values of M. and.V,VT may be deduced. The table
given in Appendix IV should be used for this purpose.
Thus the outlet conditions are determined. If the value of C.o corresponding to the value of 1V[°
found in (vii) differs from the initially assumed value in (i) by more than about 0.2 deg then
the process should be repeated usifng a second approximation to a,,. With a little experience,
however, it is possible to approximate-to cc, sufficiently accurately at the outset since its rate of
variations with A 0 is generally quite small.

6.2. Determination of Gas Conditions Relative to the Rotor Row at Inlet.-Knowing the exit
gas velocity and flow angle from the nozzle row and the blade speed being known at the reference
diameter the gas conditions relative to the inlet of the rotor may be determined.
From the velocity triangles we have the relationships:
= V°.Cos a,
1!,~V, . .. . . . . . . (7)
and tan o = (U/V.,) - tana .. .. .. .. .. .. .. (8)

7
 The energy equation gives
T = T. +P/2K ..... . . 9) ..
From equations (7) and (8) the gas inlet angle relative to the rotor, cci, may be found and hence
exit the From equation (9) and from the fact that the static temperature at
also fromincidence
the nozzleangle.
is equal-t0='the~statilc temperature relative to the rotor entry (i.e.,T. T,,
-- ,)
the total-head temperature relative to.the rotor inlet may be found. Thus:
T. = T. + V.
2
.. .. . .. .(10)
T, = T"., + v,,12Kp, T1,. + .. . . .. .. .. 11
T.- T 1 = (V.- V,')/2K,
2
= V. (1 - [cos' a./cos cc,])/2Kp .. .. .. .. .. (12)
Having found TL the total-head pressure relative to the rotor inlet, P,, may be deduced from
PI/P. = (TIT,)" - = {1 - (T. - T1)IT.]IY- . (13)
Thus the values of P,, T1 ,, cxi, and incidence are determined. From the incidence the rotor-loss
coefficient may be found.
6.3. Determination of Gas Conditions Relative to the Rotor at Outlet.-Since the flow is followed
through the rotor along an axial line (i.e., at constant diameter), and since it is assumed that no.
heat is transferred to or from th,. blades, there is no change of gas total temperature relative to
the row, i.e., T, = T. Knowing the flow conditions relative to the rotor inlet the flow conditions
relative to the rotor at the rotor exit may be determined by the same process as was adopted
for the nozzle row. Thus:
(i) The value of WV/TI/AP, is calculated. In this calculation A,,= A., cos OC, and a, must
be found by making a rough approximation to M2 and effiploying the relationship
between M, and ac,, deduced from section 5.1
(ii) For the value of Y, found for the rotor row find (WV/Tj1 AP,),a. and (C]P,),,!from Figs.
10 and 1.1
(iii) Calculate the ratio (WVTIAP , )I(WVTAP,).,, and find corresponding value of
(@IP)I(&/P),,t from Fig. 12
(iv) The value of @/P, may then be determined and hence P, (given oyP, = P, - @). The
magnitude of W.v/T/A 2P, may now be calculated and from the table in Appendix IV
the corresponding values of V2/ v/T, and M, may be found.
As before, if the value of cc2 corresponding to the value of M, differs from the value approxi-
mated in (i) by more than about 0.2 deg then the process should be repeated using a second
approximation to a,.
6.4. Determinationof Absolute Gas Outlet Conditionsfrom the Rotor.-Fromthe velocity triangles
at the rotor outlet:
V. 2 = V, cosc 2 .. .. .. .. .. .. .. .. (14)
3 = (UV,,) - tan a,2
tan oc .. .. .. .. .. .. .. (15)
From equations (14) and (15) the absolute gas flow angle at rotor exit, C3 , may be found. From
the energy equation (equation (9)) the absolute total gas temperature at rotor outlet, T,, may
be derived:

T, - T = V22 [1 - (cos2 a 2/cos 2 cx,)]/2K, .. .. .. .. .. (16)

and T 3 = 7 2 - (T 2 - T) .. .. .. .. .. .. .. (17)
i!, s
8
 The absolute total-Read pressure at the rotor outlet may be derived from:
P3IPI = (TT,)"' -
' = {1 - [(T. - T2)/T] 1 -
..... .. (18)
6.5. The Turbine Overall Characteristic.-Theoverall stage pressure ratio PPi,and tempera-
ture drop ratio (Ti - T,)Ti corresponding to the initially selected values of Wa/Ti/Pi and
NIVTi may now be determined. The overall stage isentropic efficiency is then given by:
71= [(Ti,- TW)TI[1 - (p3/P,)V-,,]. ... . . (19)
The process described in sections 6.1 to 6.4 may be continued through subsequent stages in
multi-stage turbines and both the overall turbine characteristics and also the individual stage
characteristics determined for the flow quantity and speed initially selected.
The process must be repeated for several flow quantities at the same values of NIVT, to
construct a constant-speed characteristic, the flows being suitably selected to cover the range
over which the characteristic is required. If the calculation is set out in tabular form as shown
in Fig. 16 it uill be found convenient to estimate the stage characteristics at several flows
concurrently.
The entire procedure must be repeated again if the characteristics at more than one speed
are required. If the characteristics at two speeds have been calculated then the characteristics
at intermediate speeds may usually be interpolated with. reasonable accuracy by a method
described in Appendix III.
7. Performance of a Choked or Nearly Choked Stage.-The procedure outlined in the previous
, sections may beW-/TiIA.Pi
flow parameter applied forinalla nozzle
flow quantities up toIAPi
row or WVT the in
flow at which
a rotor the non-dimensional
row equals the maximum
value possible (this maximum value being dependent upon the loss coefficient for the row as
given by Fig. 10).
Choking flow is usually defined as the flow at which the mean gas Mach number in the throat
of the blade passage of a,row becomes unity. At this point pressure disturbances downstream
of the choked row can no longer affect the flow upstream of the row. If the expansion ratio
across a turbine is increased.beyond the amount required to choke a row the flow Mach number
downstream of the choked row will increase to some supersonic value, the gas direction altering
slightly to preserve flow continuity. This supersonic expansion will increase as the overall
turbine expansion ratio is increased until eventually further choking occurs in another row
downstream of the initially choked row. At still higher overall expansion ratios further supersonic
expansion will no longer take place downstream of the initially choked row but will commence
3downstream of the newly choked one. This process can continue until choking occurs in the
turbine exhaust ducting. At overall expansion ratios beyond the value at which this occurs the
flow conditions in the turbine, and the power output must remain constant (as expressed non-
dimensionally).
It has been assumed in the preceding sections that the loss coefficient in any row remains
invariant with the outlet Mach number. If the outlet gas angle also remained constant then as
mentioned in section 6.1, the non-dimensional flow parameters would rise to a maximum value
at an outlet Mach number (the 'critical' Mach number) less than unity and would then decrease
as the outlet Mach number 'approached unity and the row became fully choked. Turbine test
results to the time of writing give no clear indication that such a phenomenon occurs in practice.
For the purpose of simplicity in calculation therefore, the following assumptions are made when
the outlet Mach number from any row exceeds the critical value:
(a) The flow quantity WVTdIP at the turbine inlet remains constant
(b) The loss coefficient Y, remains invariant with outlet Mach number
(c) The gas outlet angle from a blade row adjusts itself away from the estimated value at
the critical Mach number just sufficiently to preserve constancy of flow.
Thus the variation of flow angle with Mach number for each blade row deduced in section 5.1
is only employed in the calculation up to the critical outlet Mach number. At higher Mach
numbers the gas angle used in the calculation will differ from that estimated in section 5.1.
9
 The procedure is demonstrated in detail in the worked example of a turbine perlormance
calculation given in Appendix II.
It is not claimed that the above assumptions are truly representative of what physically occurs
in a turbine. The behaviour of the flow in a"turbine blade row as the choking expansion ratio
is approached and exceeded has not yet been examined in sufficient detail to allow more accurate
assumptions to be made. On the other hand the present assumptions lead to simplicity in the
calculations and appear to produce final characteristics which compare reasonably well with
characteristics acquired experimentally.
8. Effect of Reynolds Number.-The greater part of the experimental data from which the
approximate working rules for defining loss coefficients and gas angles have been deduced
appertain to Reynolds numberst in the range of 1 X 105 to 3 x 101. Thus any turbine per-
formance as calculated -ising this present data will correspond to a mean operating Reynolds
number of approximately 2 x 10s.
The variations with R, of the components of the pressure losses and also the gas outlet angle
from rows of turbine blades are as yet very imperfectly understood. If it is desired to estimate
the performance. of a turbine operating at mean Reynolds numbers very different from 2 x 10
it is recommended that the performance should first be calculated for a mean Reynolds number
of 2 x 101 and an approximate correction then applied to the overall characteristics by assuming
that the turbine overall efficiency roughly obeys the law:
(I--n) 0:R -11 .. . . . . ... (20)

It is believed that this rule may be applied with reasonable confidence down to Reynolds
numbers of roughly 5 x 10. At lower Reynolds numbers it is anticipated that the efficiency
may decrease rather more rapidly than required by equation (20).
The value of R, recommended for this correction is the arithmetic mean of the Reynolds
number of the flow through the first nozzle row and the final rotor iow.
9. Discussion of lethod.-The present method of estimating turbine performance has been
described in considerable detail and the method might consequently appear to the reader as
being very klborious. In practice, however, it has not been found so laborious as it might first
appear. At the N.G.T.E. it has been found that to calculate the performance of a single-stage
turbine about two or three manhours of work is required to set down the relevant geometric
data, estimate the variation of outlet gas angle with outlet Mach number, and estimate the
variation of loss coefficient with incidence on the rotor and stator rows. When this has been
accomplished about six man-hours of work are required to estimate each constant-speed charac-
teristic (as defined by about five flow quantities). A convenient tabular layout of this part of
the calculation for one stage and one rotational speed is shown in Fig. 16.
Comparisons of test and calculated characteristics of a variety of single- and two-stage turbines
has been made at N.G.T.E. In all the comparisons that were made the calculated efficiency in
the vicinity of the turbine design operating conditions came within ±. 2 per cent of the test
efficiency. At a given speed and pressure ratio the calculated flow was generally within about
=3 per cent. This infers that the working rules propounded for determining total-loss coefficients
were good to within about + 15 per cent of the true values around the design incidences and
Mach numbers and the rules for determining mean gas outlet angles were good to within about
1 per cent and - 1.5 per cent for gas outlet angles in the vicinity of 60 deg and 50 deg
respectively.
It must be emphasised that to achieve good agreement between test and calculated flow
quantities careful checking of blade throat areas is required. Small differences in blade thickness
and setting angles between design values and values as manufactured and measured in an actual
t Reynolds number in any row is herein defined by blade chord, outlet gas velocity (relative to the row), outlet gas
density, and outlet gas viscosity.
10 1
 turbine can make a disproportionately large difference to blade passage throat area and hence
to turbine swallowing capacity. Care should also be taken to assume. the correct tip or shroud
clearances since both theory and experiment have indicated that small variations in these can
have a marked influence upon turbine efficiency. Finally, the working rules for assessing the
lo-,s coefficients have been derived from turbine tests in which the gas pressures and temperatures
,*e reasonably uniform around the turbine inlet. The influence of non-uniform turbine inlet
,c.Iditions on overall performance has not yet received much study but it is anticipated that
non-uniformity may lead to some reduction of efficiency due to more powerful secondary ilows,
of incidence on to blade rows.
t1 mixing losses, and possibly to large variations

REFERENCES

No. Author- Tit, etc.

I D. G. Ainley and G. C. R. Mathieson .. An examination of the flow and pressure losses in blade rows of
axial flow turbines. R. & M. 2891. 1951.
2 J. Reeman, P. Gray and C. Morris .. Some Calculated Turbine Characteristics. Power Jets Report
R.1111. 1945.

3 A.B.P.Beeton .. .A method of estimating the performance of a gas turbine stage.

R.A.E. Tech. Note Eng. 330. A.R.C. 8,425. 1944.
4 D. G. Ainley .. .. .. .. Estimation of the change in performance characteristics of a turbine
.resulting from changes in gas thermodynamic properties.
R. & M. 2973. 1951.
5 J. Reeman and E. A. Simonis .. .. The effect of trailing-edge thickness on blade loss. R.A.E. Tech.
Note Eng. 116. 1943.
6 G. Hudson .. .. .. .. The physical implications of representing frictional flow in con-
vergent nozzles by a law of constant adiabatic efficiency. Power
Jets Memorandum 1202. 1946.
7 S. J. E. Moyes . .. .. .. Note on the critical flow conditions of a gas in convergent passage
and the influence of frictionallosses. R.A.E. Report E.3923. 1942.

APPENDIX I
A Gas flow area measured normal to flow direction (sq in.)
A. Annulus area (sq in.)
At Blade passage throat area (sq in..)

1.C C
A

CD1
Blade- chord (in.)
Coefficient of drag on blades created by tip clearance pressure losses
Coefficient of drag on blades created by secondary flow pressure losses
CL Lift coefficient based on vector mean velocity
I.D. Inside diameter of turbine annulus (in.)
O.D. Outside diameter. of turbine annulus (in.)
M.D. Mid-diameter of turbine annulus (in.)
11
I

9 Mean radius of curvature of the convex surface of a blade between the

throt and the trading edge (in.) (s Fig-3)
g Accelrtion due to gravity (ftlsec'
k Annulus height (equals blade height if radial tip clearance is zero) (in.)
i Lncidence angle of flow on to a blade row.gien by difference between gas
flow angle relative to blade inlet and blade inlet angle
i. Stalling incidence
k Radial tip clearance or minimum shroud band clearance (in.)
K, Specific heat at constant pressure (ft pdllbjdeg C.)
M Mach nmber
N Rotational speed
0 Blade opening or throat (in.)
P. Static pressure Oblsq in. abs)
P Total pressure (lbsq in. abs.)
Q ~on-diensional mass Bow parameter (,V/TIAP)
R Gas constant (ft pdijlbfdeg C.)
R Reynolds number
s Blade pitch or spacing (in.)
I Maximum blade thickness (in.)
t, Blade trailing-edge thickness, measured normal to camber-line at trailing
edge (in.)
T Total gas temprature (deg K)
T, .Static gas temperature (deg K)
AT Gas total temperature drop across turbine (deg C.)
AT. Isentropic gas total temperature drop across turbine (deg C.)
U Rotor blade speed at reference diameter (ft/sec)
"V Gas velocity (ft/sec)
V. Axial component of gas velocity (ft/sec)
ll Whirl component of gas velocity (ft/sec)
W Gas mass flow (lb/sec)
Yp Profile-loss coefficient
Y, Secondary-loss coefficient
Yk Clearance-loss coefficient
Y, Total-loss coefficient
c Gas flow angle, always measured relative to the axial direction
p Blade angle, measured relative .to the-axial direction
y, Ratio of specific heats
12
I
4 Pr V.4 VA.
* Stator blade stagger angle
Rotor blade stagger angle
7 ; Isentropic total head efficiency
Empirical factor defining secondary loss (. - C.,(slc)IC 2)
I
pt Gas viscosity
p Gas density
a Mean loss of total pressure through a blade -row due to friction, flow
separation, etc. (blsq in.)
- Sulkies
£ Inlet to first stator row of a turbine
i -Outlet from a stator row
Relative to a rotor row at inlet to the row
Relative to a rotor row at oulet from the row

Inlet to a stator row

Note: These suffices are applied to gas flow angles, blade angles, gas velocities,
gas temperatures, gas pressures, gas flow areas measured no-rMal to the
flow direction, and annulus areas. At reference stations between blade
rows A.. = A, 1 and A., = A53. The suffices repeat themselves from
f stage to stage in a turbine. The suffix . is also used for outlet total pressure
and temperature from a multi-stage turbine, but this fact is specifically
stated-in text when the suffix is so used. a., ' is the estimated gas outlet
angle if tip clearance were zero on a blade possessing a finite value of k.
Convention of Signs.-This has been chosen to follow directly from that already well established
in axial-compressor theory in this country. The extension of th4 compressor system to a typical
tturbine stage is illustrated in Fig. 1.
It should be noted, with the arrangement drawn, that the sign of th outlet gas angles, outlet
blade angles, and outlet swirl velocities relative to each row becomes negative for a turbine.
Furthermore the blade velocity must also be regarded as negative in order that the stage tem-
perature rise (or work done on the gas by the blading) shall be negative (i.e., a temperature
drop across the stage) for a turbine.
The sign to be attributed to a gas swirl velocity is always the same as that of the associated
vector swirl angle.
Gas axial velocity is always positive.
It is important to appreciate that the signs to be attributed to the gas swirl angles and veloci-
ties must always be determined by reference to the blade row with which these velocity com-
ponents are directly associated. !n other words, the sign attributed to absolute velocities and
angles (i.e., velocities and angles which would be registered by instruments which are stationary
relative to the machine casing) must be determined by reference to the immediately adjacent
stationary blade row. The signs attributed to relative velocities and angle must be determined
by reference to *'the immediately adjacent moving row (relative to which the velocities and
angles are supposed to refer).
Note The swirl velocity at outlet from a turbine is always negative when it swirls in the same
direction as the direction of rotation of the rotor.
13
 APPENDIX 11
A" Example of a Turbine PerformanceEstimation
To filustrate the method of performance estimation described in the preceding report the
performance of a single-stage turbine is calculated below at one speed. It is supposed that the
gas temperature at turbine inlet is to be 1,100 deg K and that the speed selected is N.IVTi = 435.
It will also be assumed that the inlet total pressure to the turbine is 40 lb/sq in.(absolute) at all
flows. It is not ,essential to choose actual values for the inlet conditions; the calculation can
be carried out using non-dimensional quantities throughout. However it is considered that
the calculation is easier to follow if actual quantities are used. If the final overall stage charac-
teristics are. plotted in terms of the familiar non-dimensional quantities, then the turbine
performance with inlet conditions other than those selected may be easily deduced.
The geometrical dimensions of the turbine required for the calculation are set out below and"
an explanatory diagram is given in Fig. 13.
The required geometrical dimensions are as follows (at an inlet gas temperature of 1,100 deg K):
Turbine Annulus Dimensions

Inner diameter of annulus - 9.50 in.

(constant throughout the stage)
Outer diameter of annulus - 13.00 in.
(constant throughout the stage)
Reference- diameter - 11.25 in.
Annulus area throughout turbine .= 61-85 sq in.

Details of Stator Row

Inlet blade angle (il) = 0 deg .
Blade chord (c) at reference diameter = 1-33 in.
Blade pitch (s) ,, = 09817 in.
0,,
Pitch/chord (s/c) at ,, = 0739

Blade opening ,, ,, ,, = 0.429 in.

Value of e ,, ,, ,, = 3.516 in.
s/e ,, ,, ,, = 0.279
t/c ,, ,, , = 0.200
4tls ,, ,, ,, = 0020
o/s ,, ,, ,, = 0"437
cos - o/s ,, ,, ,, = 64"1 deg
Tip clearance (k) 0 in.
Annulus height (h) 1.75 in.
Throat are (A,) 26.97 sq in.
Annulus area at reference station down-
stream of stator row (A..) = 61 .85 sq in.
cos-' (AIA,,) = 64.1 deg
14

AI
_______
 Details of Rotor Row

Inlet blade angle (f1) at reference diameter - 36-0 deg

Blade chord (c) ,, - 0.95in.
Blade pitch (s) ,, 0.7113 in.
Ss/c
-

, ,, ,, - 0-749
Blade opening (o) ,, ,, - 0-447 in.
e - 2-00 in.
s/e It 0.355
tc ,, ,, - 0-15
t./s ,, ,, - 0.010
o/s ,, ,, 0-629
cos - , o/s ,, ,, , 51.0deg
Tip clearance (k) 0.030 in.
Annulus height (h) - 1-75 in.
k/h - 0.01715
Throat area (A,'), if k = 0
(see section 5.1) - 389 sq in.
Annulus area at reference station 61.85 sq in.
downstream of rotor row

Calculation of Gas Outlet Angles from the Stator Row.-When the outlet Mach number (M.)
is less than 0. 5, the gas outlet angle is given by:
a.-- =* - 4(se) .. .equation (1), section 5.1
= f (cos - os)
=Xo* -62.4 deg .. .. Fig. 5
4(s/e) --1.1:
therefore cc. - 63.5 deg.
When M. 1.0 :
W nM --=1cos 1 (A'/A,,) .equation (2), section 5.1

therefore c. -64.1 deg

Between M. = 0.5 and MVr = 1.0 a smooth transition curve is sketched in as shown in Fig. 14.
As stated in section 5 it is assumed that the gas outlet angle is uninfluenced by the gas incidence
angle on to the blade row.

Calculation of Gas Outlet Angles for the Rotor Ro.-When M, < 0-5 the outlet gas angle is
first estimated from eqtiation (1),.section 5.1, and is then corrected for the finite blade tip clearance
by use of equation (4).
1
f Thus, with zero clearance,
iC2 = OC* - 4(s/e) .. .. .. .. .. .. .. equation (1)
4Cc* =f (cos- ' o/s) = - 47-2 deg .. .. .. .. .. .. Fig. (5)
4(s/ ) = 1-4:
therefore (X2= - 48.6 deg.

Correcting for tip clearance, using equation (4):

2 = tani{[1 - X(klh)(cos "cos OC2')] tan c2' + X(k/h)(cosfi,/cos c2 ) tan fl}.
1
OC
and X = 1.35 for radial tip clearance,
also k/h = 0.01715"
therefore c,.= tan-' {(1 - 0.0284)(- 1.. 134) + 0.0284 x 0.726}:

therefore OC = - 47.3 deg.

When M2 = 1 0 the outlet gas angle is given by equation (2), viz.,
C. - cos- ' (AIA.,).
If the tip clearance were zero the throat area of the blade (A,') would be 38.9 sq in. With a
tip clearance (k) of 0.030 in. the throat area of the blade is
A,= Al' [1 - (klh)] + n x k x O.D..
=389 [1 - 0-01715] + ; x 0030 x 13 = 39.46 sq in.

Therefore when M = 1.0

therefore (X = - 50.3 deg.

Between M, = 0.5 and M, = 1 .0 a smooth transition curve is sketched in as shown in Fig. 14.
Calculation of Stator-Loss Coefficient.-Since the stator row operates at one incidence only
(i = 0-deg) it is only necessary to predicf the 16ss coefficient for this one incidence.
(a) Profile Loss
,/ = 0 deg, c0 = - 63.5 deg, s/c = 0. 739, fllc = 0, = 0.20.
From Fig. 4 and equation (5) Yo 0"0288
(b) Secondary and Tip Clearance Loss.-At zero incidence: a, = 0, c.- - 63.5 deg,.
sc =0.739, kh = 0
Now, CLI(slc) = 2(tan cci - tan oc,) cos oc
where tan oc, = (1/2) (tan a,-+ tan ot.)
therefore CLI (s/c) = 2.83, and coss o,/cos* ,, = 0.566
The expression for secondary and clearance loss is given by question (6):
Y. + YA = [, + B(klh)][CLI(slc)]' [cos' ac/cos' O,]
The factor Ais a function of (A0/A)2/[1 + I.D.O.D.] and is plotted in Fig. 8.

-!6
 A, = A., cos fl= A. =61-85 sq in.
-A. A..cos o= 27.60 sq in.
I.D.O.D. = 9.5/13.0 = 0.73
therefore (Ao/Aj)'/(1 + I.D./I.D.) = 0 115
therefore ---
00065 .. . . . . . . . Fig. 8
therefore Y, + Y, 0-0065 x (2-83)Y x 0 -566 -- 0- 0295
0
Total Loss Coefficient
, Y, = Y. + Y. + Y, 0 0-0288 + 0 0295

therefore Y, = 0.0583
Since t/s = 0.02 no correction is required for trailing-edge thickness. This total-loss coefficient,
calculated from predicted gas angles for, M. < 0 -5, is assumed to remain constant at all values
of M..

Calculation of Rotor Row L'oss Coefficients.-Profile Loss.-Relevant data for determining

profile-loss coefficient at zero incidence is-:
f 1 = 36.0 deg ; a2 = -48.6 deg (Note that value of a, for zero clearance is
used for predicting loss coefficient).
s/C= 0749 ;- /#I = -0-74.; tlc=-15
Y'%/t = 0 deg) =0-00238 "
Fig. 4.
YP(/t1 = - 02) =0.0722 f.

Profile loss at zero incidence is given by equation (5):

Y, = {Yp(flj = 0 deg) + (fl,/cz) [Y(
2 - a2)- Yp(#L =0deg)]}
= {0.0238 + (- 0.74)2 (0.0484)} [0.75]0-74
= 0.0503 x 0.808 = 0.0406.
To calculate the profile-loss coefficient at incidences other than zero it is necessary to evaluate
the stalling incidence, i,, and employ-the relationship of Yp/Yp(i = 0) with.i/i, plotted in Fig. 6.
The stalling incidence is evaluated as described in section 5.2. Thus:
fl, = 36.0 deg, = - 48486 deg, s/c = 0.749.
As the pitch/chord ratio is virtually 0.75 the curves of Fig. 7b can be used directly
*p1/ 2(s0c = 0.75) = - o.74
therefore i, =-9.5 deg . . . . . . . .. Fig. 7b
Secondary and ClearanceLoss.--From equation (6):

Y + YA = [I + B(k/h)][Cj/(s/c)]'[cos2 c,/cosP x,.].

For the rotor row, A, '-A.. cos/ 1 - 50.0 sq in.
and A, = A., cos a2 = 40.9 sq in.

Therefore (A,/A,)1(1 + I.D.O.D.) = 0386

and from Fig. 8, A = 00183.
17
 Furthermore B = 0-50 for radial clearance and k/h = 0-01715.

I Therefore Y, + Y, = 00269[CL/(sc)1[cose o2/cos3 Q.

This equation must be evaluated for a number of incidence, angles. This may best be done in
tabular form as shown in the following table together with the profile loss and final evaluation
of total-loss coefficient.
Total-Loss Coefficient
ill, -3-0 -1.5 -- 8 0 1.0 1.5
S-28-5 -14-3 -7-6 0 9-5 14-3
+7.5 -+217 +28.4 +36 +45.5 +50-3
CL(SIC) - 2"88 3"22 3"65 4"30 -
cos a.2 /cos' 0C. - Q528 0"494 0-465 0"44
Y, + Yk 0.1178 0-1178 0-1378 0.1668 0.2186 0-2186
YP 0.1739 0-0853 0.0565 0.0406 0-0812' 0*1820
Y, 0.2917 0.2031 0-1943 0.2074 0-2998 0-4006
These values of Y, correspond to t/s = 0-02. Since on this rotor /s = 0-01 the total-loss coeffi-
cients as calculated above must be corrected for trailing-edge thickness. This is achieved by
multiplying the loss coefficients by-an appropriate factor obtained from Fig. 9. For ts = 0.01
this factor is 0.96. We have a final estimated variation of total-loss coefficient with incidence
therefore as below:
i (deg) -28-5 -14-3 -7-6 0 9.5 14-3
YJ 0.2800 0-1950 0-1867 0.199 -r 0-288 0-3845
These values are plotted in Fig. 14.
Calculation of the Turbine Characteistic.-Havingdetermined the loss coefficients and gas
outlet angles for the individual rows it is possible to commence the step-by-step calculation of
the gas flow conditions at entry to and exit from each row for selected entry conditions and
rotational speed.
: ,For demonstration the following entry conditions are arbitrarily selected for the turbine
P= 40 lb/sq in., absolute

T= 1100 deg K
NIVT, = 435.
These conditiofis represent a blade speed at the reference diameter (11 -25 in.) of - 708 ft per sec,
the negative sign being inserted in accordance with the chosen sign convention. The ' standard'
thermodynamic gas constants used throughout are:
K, = 12,300 ft pdl/lb/deg C.

R = 3,073 ft pdl/lb/deg C.
y = 1-3330.
It is now necessary to select arbitrarily a number of mass flows through the turbine. For each
mass flow the inlet and outlet conditions for each blade row are estimated in the manner described
in sections 6 to 6.4. The calculation can conveniently be.carried out in a tabular form, gas
conditions for several mass flow values being calculated concurrently. A table of calculation
for the present turbine is set out in full on Fig. 16 with an explanatory note attached to each step.

1-"
'1 . , -4
 In this example the first five selected flows are found to pass through the stage without the
outlet Mach numbers from either the nozzle or the rotor row exceeding the critical Mach number,
i.e., the non-dimensional flows through each row are less than the maximum flow that can be
passed by the row. The sixth selected flow is exactly equal to the maximum flow that can be
passed by the nozzle row but it is found that in this instance the non-dimensional flow relative
to the rotor row (see line 32) is greater than the maximum possible values (line 34) that can be
passed by the rotor. This latter flow is therefore an impossible value and the calculation of the
flow conditions with this initial mass flow must be discontinued. When this impasse is en-
countered it becomes necessary to select an initial flow which will just pass through the rotor
row. This flow can be selected by a-process of interpolation by plotting both the non-dimensional
flow through the rotor (line 32) and the maximum possible non-dimensional flow (line 34) against
the initial flow (W or WVTIP). The point at which these two curves intersect then represents
the maximum flow which the turbine will pass. In the present case this flow is given by
WVTIP = 10 268 and the calculation for this flow is carried through in the seventh column
of the table in Fig. 16. At this flow the relative outlet Mach number from the rotor is exactly
equal to the critical value for the relevant rotor loss coefficient. In accordance with the assump-
tions outlined in section 7 it.is now supposed that if the total pressure drop across the turbine
exceeds the critical drop calculated in column 7 then (a) further increase in pressure drop across
the turbine is not accompanied by any further increase in flow, (b) the flow conditions upstream
of the rotor remain fixed (i.e., identical to the values calculated in column 7 down to line 30)
and (c) further gas expansion takes place only downstream of the rotor row with the rotor loss
coefficient remaining unaltered and the relative gas outlet angle adjusting itself just sufficiently
to maintain continuity of flow.
To demonstrate the altered procedure after the flow has attained its maximum value columns
eight and nine are added on Fig. 16. Values of M, at the rotor outlet (line 42) are now arbitrarily
selected. Uniquely related with these Mach numbers are the non-dimensional flow parameter
at rotor outlet, WVT,/A 2 P.. (see table in Appendix IV), and the total to static-pressure ratio
P 2/P2. Also known is the rotor loss coefficient and the inlet total pressure, P,, and outlet total
temperature T2 (equal to T,). These conditions are sufficient to determine the gas outlet angle
cc., required to preserve continuity of flow. Thus"
@/P 2 = Y1[l - P,2 /P2 . . .. (21)
C = (aIlP 2)/(1 + CO/P
2) ..-. .. .. .. .. (22)
IT
romvalueo = 1 .. .. .. .. .. .. .. ..(23)
From value of WVT 2 A2 P2 in line 49 and the known values of W, T2 , and P, we can find the
necessary value of A 2. Since A2 = An2 cos X2 we can determine or, (line 51). Having thus deter-
mined the new outlet conditions relative to the rotor the absolute conditions a, the rotor outlet
can be calculated as before (lines 52 to 64).
From the initial and final pressures and temperatures the overall stage values of pressure
ratios, efficiency, etc., are calculated in lines 65 to 70 and plotted in Fig. 15.
, These characteristics correspond to a mean blade Reynolds number of 2 x 101. The overall
efficiency and temperature ratio may be. roughly corrected to other Reynolds numbers by use
of equation (20), section 7.
*, In this particular example the flow was limited or choked by the rotor row. This is not, of
course, peculiar to all turbines. If'the nozzle row is found to choke first then when the performance
at pressure ratios beyond the critical are calculated the same procedure that has been outlined
for the choked rotor must be applied to the nozzle. In such an instance, of course, the tabular
layout shown in Fig. 16 would have to be modified accordingly. Generally it may be found
more convenient to continue the calculation of the performance of the turbine under choked
conditions on a second table rather than to concentrate the entire performance into bne table
as shown here.
19
1' .
 APPENDIX III
Interpolation of Constant-speedCharacteristics
If the characteristics of a turbine are required for a large number of speeds it is only necessary
in general to perform the detailed calculations as set out in the preceding sections for two speeds-
preferably a speed close to the highest speed and a speed close to the lowest speed of the required
range.
If the gas was nearly incompressible then the turbine performance at various speeds could be
roughly related on a pair of curves by plotting efficiency and (WVTIP)/(NVT,) against
(KA T/Tj)/(N/ VT)' since (WVT,/Pj)/(NiVTj) would be proportional to the familiar velocity
coefficient V/U and (Kp T/T,)/(N/i/2)- is proportional to the familiar temperature drop
coefficient 2K.A T/ U'.
This principle can be extended to compressible gases with some degree of success if the ratio
of non-dimensional inlet flow to non-dimensional speed (WVTdPj)/(N/VT), is multiplied by a
suitable function of the overall density ratio. On single and two-stage turbines' operating with
pressure ratios up to about 0.3 it has been found that plotting (WVTj1P,)(PT3/P3Tj)uI(N/VTj)
and efficiency against (KO1T/Tj)I(NIV/T,)' brings constant-speed characteristics at different
speeds sufficiently close together to permit reasonably accurate interpolation
preceding expressions P 3 and T3 are supposed to be the outlet total pressure and(N.B. in the
temperature
from a turbine). It will be noted that the expression (PT/PT,)is the overall density ratio
of the gas and the power factor n is entirely empirical-.

The value of n is selected by trial error to bring the curves of (WVTPj)(PT3/P T,)*I(N/VT,)
3
plotted against (KpA T/Tj)I(NIVT,)' for the two calculated rotational speeds as close together
as possible. The best value for n generally lies in the range 0 5 < n < 1.0.

20
 K APPENDIX IV
Relationship Between Non-dimensionalFlow
j ParametersWVTIAP, pIP,M and V/VT
All values tabulated have been calculated for:
y=1-3 3 3 0
Kp = 12,300 ft pdls/lb/deg C.
R = 3,073 ft pdls/lb/deg C..
,- g --32-173 ft/see _
I M WVTIAP P.lP V/VT M WVTAP p./P V/VT

0-02 0.0135 0.99973 1.290 0.62 0-3342 0-7802 38.45

0-04 0.0268 0.99893 2.565 0.64 0-3404 0.7678 39-64
0.06 0.04016 0.9976 3.845 0.66 0.346 0.7555 40.80
0.08 0.05345 0.99574 5.135 0.68 0.3515 0.7427 41-99
0"10 0-06667 0.99335 6"400 0-70 0'35625 0-7307 43-06 4

0-12 , 0"0798 0.9904 7.678 0-72 0-361 0-718 44-20

0.14 0.09278 0-9870 8.95 0-74 0.3653 0-7052 45.35
0.16 0.1055 . 0"98317 10.21 0.76 0.3694 0.6914 46-49
0.18 0.1184 0.9787 11-495 0"78 0.3729 0.6791 47.60
0.20 0.1309 0-9738 12-75 0.80 0.3761 0-6665 48.70
0.22 0-1433 0-96835 14.024 0.82 0.37875 0.655 49-71
0.24 0.1554 I 0.9626 15.25 0-84 0.3814 0"6408 50.9
0.26 0.1673 0.9563 16-52 0.86 0-3837 0.6276 52.01
0.28 0.17925 0.9495 17-80. 0"88 0"3856 ' 0"6141 53.14
0.30 i 0190 0.942 '19.09 0.90 0.3872 0.6025 54-20

0.32 0.20225 0.9344 20.32 0.92 0.3885 0.589 55-20

0.34 0.2131 0.9265 21.53 0.94 0.3895 0-5755 56-39
0.36 0.2238 0"9180 22-80 0"96 0.3901 0-565 57-30
0"38 0.2343 0.9092 24-02 0.98 0.3905 I 0.551 58"25
0.40 0.2442 0.9003 25-25 1.00 0.3907 0-540 59.25
0.42 0.2543 0.8908 26-5 1.02 0,3905, I 0528 60-35
0.44 0-26375 0.8810 27-68 1.04 0.3901 0.517 61-30
0-46 0.273 0.8706 28-9 1.06 0.3895 I 0.504 62.2
0.48 0.2819 0.860 30-12 1"08 0.3886 0-491 63-29
0"50. 0.2903 0-8495 31.31 1.10 0.3875 0.480 64.20
0.52 0.2986 0-8386 32.57 1.12 0.3861 0.4682 65.20
0-54 0.3064 0.8272 33.76 1.14 0-3846 0.4565 66.20
0.56 0.3139 0.8157 34-95 1.16 0.3828 0.4453 67.12
0.58 0.321 0.804 36.16 1.18 0.3807 0.4333 68-08
0-60 0.3278 0.7922 37-31 1.20. 0.3786 0.4225 68-95

S
21
 42 Cl4

ci i ui

a) I C

a'.a

z ~ jo -

C.d

LIU

7/ 'H=.-~.~1I3, u-i
IAI

ujfw
qs
0

s
i
=2 L.

.7>' 2 4

220
 I! 83 a

z~in
00
$P
-~
urn
8
7 w~u us ~ZL
Aaa

LJ

In t

in

0
u p0

0 ~ Be I.

a::)j ~ _j 44-4

a U

I~L9

ii0

caIcl - J c

t ~ ~La 6iU

1' 23
 0

0 ~
- cn

C0 9

4)cfd

o ]o v
o ~-*'~ 0

In(

I -

C'Co

II 90.0

a -- 6
b J.3:)J'0 3.2d -1 l'LN1L30'S4:Ub

I24
.1 6.0

-0

Ju

1-59 -4.0 -3-0 *-20 -1.00 +1-0 +

ASTALL
F'IG. 6. Variation of profile loss with incidence for typical
turbine blading.

- . 4.
-99

0I:

I~C a - 0

ori
 0 TLOSO ?O=X ROMS.
x CASCAM IC59~

01,1
oo!®

1 I

0.0 oH o-Z A / . . 04 015

[I.,.0'4 0. o-
AI
FIG. 8. Secondary losses in turbine blade rows.

I-8 i

26 2
"I

0 G0 oo k/ .0 00 o &

FIG. 9. Effect of trailing-edge thickness on blade loss coefficients.

26 ...
 (40 __ '_ _

WT

0-30

0.281

0"78 _ .,_ _ _ _ _ _ _ _ _
0 0,1 .0. 0'3 04 0.5 0.6 0.7 0.8 0.9 1.0
Y
FIG. 10. Compressible adiabatic flow.

IWI

404

0 ~ 0. 1 0? 0 3. o4 0.5 0-6 0-7 0.8 0-9 10

Y
FiG. 11. Compressible adiabatic flow.
27
 IIII I II I I II

o°e _ _ . 1
0.I.

4 40
021
04 0.2 0.3 04 0-6 0"7 0-8 05 1.0
SMAX

FIG. 12. Compressible adiaba .ic flow.

"o~~ c-
,, E,,
OUl

IL
3
3, -. ,9
0

n W1

28 " II I - II
cifl I

i iii

lm
I'
 Ic MI. Ie -
kI).. * .

Ae.%
I IOf

__ ___ _I

=a

ZL)

119~

63 0

0 c.)
hii

z .2

01 cn

< 00

MLC 3 03

V 4
~~d

W Li0 5C

1- -i 29
 CALCULATED PERFORMANCE OF A SINGLE STAGE TURBINE

NI IT, 43: T I 100'K 1P -. 10 P.ia. K, 12,300 ft. pdl.j/*/C ; V 1-3330

-::'z .z C.0 0 1060 11-0 .000 60016 0006d-IT40Y10~C

0 0. 03a1~ 010 a
YI(. .9.r 0 5LS 0011. 05 06 -183 005138 O10 PK-Df.C AN 1 bOU ."
o/oA±n- 057 0 306 a. 71 . 0% 0366 0n @rm)
00 5 - F IQA 4is 7w M
~~o1~~lC~ '76
LL 081 s Cora 0 40760011roV.CI
0or 0 04
M 0 1.50023 r004 AIC " ANDWAW06(COL
.14 1700
o
S0.570 0315 010 4 0 .716 rot0 & 9
0S4 1113-4 L S 4PJV,
AND
I 301 317
OS 0348 4 3 031%4 -60 05 jt4i
1343 0101 0403 060 079 06105* ~00~
0 P U0-0*310003 .3 Cot 04-11-64
10 0 4400 __ F1134 O A ND
N_ 10o 6R 0hAUCAIJOIiL T .m .01V07t I
I's0_5____5_1-76__0_ 0-3 O . LD 0100 AN01A AP$ 0I00A1Fd70
3 1 7.1 ".4 0
%44430. 54 "0.4(0.aC1
77M SM 00IS0 0137 0604 030(3 0.3 UN...C .TL 00

41
v11. MI .5 31 41Gs .51 6 70 t 4 600
SS)0 Ce. ot 1

.001
773 00 0 01157~ 17600 0t0 007 74N. * IL - 110 0 10
39G1.51 00. 24 55
46 0 50 4*1
Ca' S *.11 .15 o*,S
-6 EG 00 -0435e -035861

(.?A l 0.01
5 . 0030040 10546 0070.17 IA o.7.,TNA Ci"f~d8
T__ .4.70. 120 Z. 47 4- 0 5 4.

P330 3 37 44-l30 61.14 18_ z _ I. CO___V f

7.
~.47.)
0016 0-02Z40 NOG
47.3 .4.)
004035 0037
47.3 .47.4
007
.00 1
041 31405
C&70qA?40tC0J7031
ool4._C..M 001?? 0083 01010 00154p 01353 O45 '*A1

037-3GI.
3 1311-1
413 S . 33 Si3114 0 6 09- IS S100 MI.R 3 00MO64VA(

T__ 074 8.13 I 076

_0___9 M A 1454 2 1lt MOC 34

C0 ! 0007IZ 0.01 .1 ' 0I400? 0.410 0

0-49- 104141 cc.0A q'040 .~ 'IC36
3L, 04, 0 .40 01 0 03 0176 0. 644 0.3w I- mo 4%0w012 m
P 3?3 031 35a 4Aquas0.76as0.3 010 5 0.33v P . OAR

z- 0647 0-15 1 34 0405o 066 11. .10 0.0 032704.004 FR ANOQO 380RO 002 C.1700.0UM

. /p .47.3
0 47 . .47.30
ZZ -4.0 -0.013 F001 400.0.f 36O0 A U C0 .1 %.000C3 ~l~ 0 0 TP 004
10-49 -410 000.
Fil~014 .00 071 # -4 00c
!!A61_01JT00 13Z-7610 s 0.;r7 ,- 005? 3M

, 1.4 0- -4)045 04 .- & 1.0 1037 F104. 6C 4 kMOk

0101 AW..tlt1AIX % C VC M, 2-' MtCA
-47-3 -473
-43 47-31 001 1 700.0.6 41A MONG VLU AWA
AU0700WM0.* N VMO MN 1 'WMrST O4

.01~~~~A VC 33.00
1.AA 06.00MTIO
31.0 37 ~ 00 0 WIT A0.3 It0,0

F . 40441-0110 7560 00 C446 o0t M x

I II 44.9.0 - *o 00.0 .1.0 040 - -0.

!36

-Z ,/T 06I .00

MO 0 74
W1 1151 11. t. 45""
67 7004 C0 w 3I&

I ____, S$ C.0 374 61 0 0 ole las lisp

4, 104

0a6..408 08 a" 7 141.07t 136 4416040,A

_3,-

000364 .1-0114 -. 06406 4113 0641 .0 01 0.-01316

SI 2. '0"a -0011 066A 0 63?4

fI70-15 .780 075 5 TAN 1740 TA1 .
1l1 ,-T 0.95 0.401 01 1 I.16 4.40 6547 61 70. FRM'IAkN1

~ ~ 00S~
23i~ ~~ -04~ 0026015 ~~~Fr
4 0 16.*O
IK D C
0031z8J/o4
0010 334 0963 L4_S C VIJN0KD GlRAt 0143I
Q-141 139
064 IS
11040 .14 .6% 30&R