009

9
http://www.developmentbookshelf.com/doi/pdf/10.3362/9781780445519.009 - Kanye G. Radabaugh <rivadeneirarafaelnazareno@gmail.com> - Wednesday, October 25, 2017 4:11:39 PM - IP Address:190.139.125.34
APPENDIX: STANDARD UNITS, FORMULAE

AND NOTATION
9.1 Standard units 9.2: Multiplication factors
Equations in this book are presented in the metric or The above units can be multiplied by the following
SI (Systeme International d'Unites) system of units. factors to make the numbers more easy write. The
To use them, quantities should be converted to the factors should be removed before putting quan-
base or derived units shown in the table below, and =
tities into equations. So, 1 MW (is equivalent to)
then the output will also be in these units. Some 1 000 OOOW, but only W will be used in equations.
books specify different units for the quantities given
in an equation, even when the book uses metric Table 9-2: Multiplication factors
units. In such cases the units given have to be used, Prefix Symbol Multiplication factor
but the author will have included the necessary con-
Multiples
version factors within the equation. Tera = 1012
T 1 000 000 000 000
In theory, basic equations may be used for any Giga G 1 000000000 = 109
self-consistent system of units, provided that both Mega M 1 000000 = 106
kilo k 1000 = 103
the input and output quantities are in the base units milli = 1Q-3
m 1/1000
for that system. So it is possible to use imperial micro 1.! 1/1 000 000 =1~
units in the equations, provided one is consistent nano n 1/1 000 000 000 = 1Q-9
pi co p 1/1 000 000 000 000 = 1Q-12
and only use base foot-pound-second units. femto = 1Q-15
f 1/1 000 000 000 000 000
However, in the author's opinion, the metric sys- atto a 1/1 000 000 000 000 000 000 = 1Q-18
tem is by far the easiest, and it is recommended that Sub-multiples
the reader uses it exclusively. hecto h 100 = 102
deca da 10 = 101
deci d 1/10 = 1Q-1
Table 9-1: Metric system of units centi c 1/100 = 1Q-2
Quantity Unit Symbol Equivalent
Base units 9.3 Other Sl units

Length metre m Table 9-3: Other Sl units are also used by conven-
Mass kilogramme kg
tion or for convenience
Time second s
Electric current Ampere A Quantity Unit Symbol Equivalent
Temperature Kelvin K
Luminous intensity candela cd Volume 1Q-3 m3
litre
Supplementary units
Note that 1 litre= 1 decimetre3, dm3
Plane angle radian rad
Solid angle steradian sr Mass tonne t 1000kg
Pressure Pascal Pa = N/m2
Derived units
bar bar 105 N/m2
Frequency Hertz Hz 1/s Dynamic viscosity centipoise cP 1Q-3 N.s/m2
Force Newton N kg.m/s2 Kinetic viscosity centistoke eSt 1~m2/s
Work/energy Joule
Power Watt
J
w
N.m
J/s
Electrical conductivity Siemens s 1/Q
Electric potential Volt v W/A
Electric capacitance Farad F CN
Electric resistance Ohm Q VIA 9.4 Basic formulae
Magnetic flux Weber Wb V.s
Magnetic flux density Tesla T Wb/m2 The following two formulae are given as they are
Inductance Henry H Wb/A fundamental to all hydro-power derivations, and
Luminous flux lumen lm cd.sr also to illustrate the use of correct units.
Illumination lux lx lm/m 2
Note also: 9.4.1 Turbine output power
Temperature Celcius oc K
The power output from a hydraulic system is:
oc
The interval of 1 is the same as 1K, but the zero point of the
Kelvin system (Absolute Zero} is at -273.2C. Despite K being Power = Pressure x Flow x Efficiency
the standard unit, Celcius is much more common. Eq. 9-1
130
Or: This equation uses the standard SI units. Note that
w is in rad/s, not rpm. It is possible to convert this
p = P Q . 1Jcotal equation to use rpm:

= (Hgross X Pwater X g).Q.1Jtotal Eq. 9-2 P =T.ro
P
p
-power (W)
- pressure (N/m2)
=T. 2n.(~)
Q - flow (m3Js) n.T.N
71totai - total system efficiency (unit-less, ---- Eq. 9-7
30
fraction)
Hgross - gross hydraulic head (m)
N - rotational speed (rpm)
Pwater - density of water (kgfm3) Note the final equation has 1t/30 in it. The 21t con-
g - acceleration due to gravity (m/s2) verts revolutions to radians, and the 60 converts
rpm to revs/s. So the unit conversion is included in
This is true for any turbine. If we convert this to the equation, and non-standard units can be used.
take power in kW and flow in litres/s, and substitute
values for Pwater and g it becomes:
9.5 Unit conversion
1000. P = Hgross Pwater .g.(~~ ).1Jtotal
Below are listed some conversion factors com-
monly used in hydro-power work. Some are be-
:. p = Hgross Pwater .g.Q.1Jtotal tween metric units, the rest are from the old
British, Imperial, or inch, units to the SI system.
106
H gross .1000 X 9 8 Q.1] total Table 9-3: Conversion factors
106 To convert
From To Multiply by Notes
=9-
8-__.:::_
Hgross _ _
Q.1Jtotal
__
1000 Eq. 9-3 Length
in mm 25.4
P (kW) in m 0.0254
ft m 0.3048 foot
where: Q - (litres/s) yard m 0.9144
Hgross - (m) Area
mm 2 m2 1~ square millimetre
cm 2 m2 1()-4 square centimetre
This form is often found in books. It is quite valid, in2 m2 645.2 x 1~ square inch
but it is not the root form of the formula, and only ft2 m2 0.0929 square foot
works with the specific units given for the yard 2 m2 0.8361 square yard
quantities. Volume
litre m3 1()-3
Below is given a form of the formula which is cm 3 m3 1~ cubic centimetre, cc
very easy to remember. It assumes a 71totai of 51%, mm 3 m3 1()-9 cubic millimetre
in3 m3 16.39 X 1~ cubic inch, cu.in
which is quite a good initial estimate for a micro- ft3 m3 28.32 X 1()-3 cubic ft, cu.ft
hydro scheme generating electricity. yard3 m3 0.7646 cubic yard
pint m3 0.5683 X 1()-3
gallon m3 4.5461 X 1()-3 Imperial gallon
p electrical =5Hgross Q Eq. 9-4
US gall. m3 3.7851 X 10-3 US gallon
Angle/angular speed
p - (W) deg rad rr/180 or
Q - (Iitres/s) 0.01745
rpm radls rr/30 or 0.1047
Hgross - (m) Mass
tonne kg 1000 metric tonne
9.4.2 Power/torque relationship lb kg 0.4536 pound
cwt kg 50.8 hundredweight (1 cwt =
For any shaft: 1121b)
ton kg 1016 Imperial ton
Power= Torque x Angular Speed Eq. 9-5 ton t 1.016 So a metric tonne is
nearly the same as an
Imperial ton!
or: (1 ton= 2240 lb = 20 cwt)
Force
P=Tw Eq. 9-6 lbf N 4.448 pound force
(the force exerted by 1 lb
mass under standard
P - mechanical power (W) gravity).
T - torque (N.m) tonf N 9964 Imperial ton force
w - angular speed (rad/s) (continued over)
131
9. 7 Accuracy and the acceleration due
To convert
From To Multiply by Notes to gravity

Torque Many of the quantities used in the calculations for
daN.m N.m 10 turbines are not known accurately. Surveying tech-
kN.m N.m 1000 niques may be approximate, and if an Abney level
N/mm N.m 1000
lbf.in N.m 0.1130 is used to measure the head then an accuracy of
lbf.ft N.m 1.3558 foot-pound (often 0.2m would be quite good. Unless values of such
incorrectly written simply
as lb.ft)
quantities as the velocity coefficient for the nozzle
and the strength of the steel being used can be
Pressure/stress
Pa N/m2 1 i.e. Pascal =N/m2 tested, average values have to be assumed. Given
bar N/m2 105 this inaccuracy in the inputs, it must be realised that
kN/m 2 N/m2 1000 the calculated figures will only have limited ac-
kgf/cm2 N/m2 98066 standard gravity
mwater Nfm2 9800 metres of water, curacy. As a rule, the results of calculations should
i.e. the pressure exerted only be given to the number of decimal places of
by a head of this many the least accurate input. It is useless to quote the
metres of water.
psi N/m2 6890 =
psi lbf/in 2 output power to five decimal places if one only
psi bar 14.5 knows the head to within a metre or so.
fiwater N/m2 2987 One quantity that is usually assumed to be
atm N/m2 1.013 x 105 atmosphere
So a 1OOm column of water generates a pressure of: known accurately is the acceleration due to gravity.
100 mwater = 10 kgf/cm 2 = 9.8 bar = 9.8 x 105 N/m2 = 9.8 x 105 Most book consider this to be a fixed and immuta-
Pa = 0.98 MPa = 142 psi ble 9.81 mJs2, or 9.80655 mJs2 (the International
MPa N/m2 106
MPa N/mm2 1 usual unit for material Standard Gravity) if greater accuracy is required.
stress. However, while these figures are true for most of
N/m2 N/mm2 106 Europe and North America, for many other parts
kN/mm2 N/mm2 109
GN/m 2 kN/mm2 1 usual units for specifying of the world they are not so accurate. The tables
material modulus below give the variation in gravity from the Equa-
properties, such as tor to the Poles, and the approximate variation with
Young's Modulus.
tonfin2 N/mm2 15.44 height above sea level (Harris, 1983).
tonfin2 N/m2 15.44x 106
psi N/mm2 6.895 X 1Q-3 Table 9-5: Variation of gravity with latitude and
Power/energy height above sea level
kW W 1000 kilowatt
MW W 106 Megawatt Latitude Gravity Height above Reduction
hp w 745.7 Imperial horsepower (mfs2) sea level in gravity
mhp W 735.5 metric horsepower (m) (m/s2)
(1 mhp = 75 kgf.m/s)
9.780 1000 -o.0031
9.782 2000 -o.0062
9.786 3000 -o.0093
9.6 Greek alphabet 9.793 4000 -o.0124
9.802
Many of the equations in this book use Greek let- 9.811
ters for variables. 9.819
9.826
9.831
Table 9-4: Greek alphabet 9.832
Lower Upper Name Lower Upper Name
case case case case
These figures are not given so that super-
0. A alpha v N nu accurate values of g can be used in calculation -
~
y
B
r
beta
gamma
~
0
-0 xi
omicron
quite the opposite - but to illustrate that even g is
I) ~ delta 1t n pi not fixed. If a micro-hydro site is anywhere south of
E E epsilon p p rho 40 latitude, using a value of 9.81mJs2 is actually
~ z zeta CJ :E sigma worse than using 9.8m/s2. A figure of 9.8 is quite
.., H eta 't T tau
e 8 theta '\) y upsilon adequate for most hydropower calculations any-
I iota <jl <l> phi where. Indeed using g = 10mfs2 only gives an error
1( K kappa X X chi of2%.
A. A lambda IV 'I' psi
j.l M mu (l) n omega
132
10
APPENDIX: DERIVATION OF FORMULAE

10.1 Nozzles
Forebay tank Penstock
---
2
Water jet
Nozzle
Fig. 1D-1: Simplified forebay tank, penstock and nozzle
Consider the simplified system shown in Figure 10- Now h 0 = 0, atmospheric pressure, v0 ""0, since the
1. A forebay tank feeds a penstock which takes water is effectively still in the forebay, and z0 - z1 =
water down to a nozzle, where it comes out as a jet. H gross> the gross head. So:
Bernoulli's equation for an incompressible fluid v21
(such as water), modified to take account of fric- ~ +- = Zo- z l - L1 hpcn = Hgross -L1 hpen = Hn
tional losses, between the surface '0' of the water in 2g
the fore bay tank and the inlet to the nozzle '1' is: Eq. 10-2
where Hn is the net head at the nozzle. We can
vz vz write a similar equation between the nozzle inlet
z. +-o-+Zo
"O =''1z. +-~ +z I+d h pen and the jet; assume for the moment that the nozzle
28 28
is perfect, with no losses.
Eq. 10-1 v2 vz
z. +-~ +z = z. +-2 +z
''1 2g I ''?. 2g 2 Eq. 10- 3
- pressure head at the surface of the
forebay tank (m) hz - pressure head in the jet (at the vena
- water velocity in the forebay tank contracta)
(m/s) Vz - water velocity in the jet (at the vena
Zo - height to water surface in forebay tank contracta)
(m) Zz - height of jet centreline
- pressure head before the nozzle (m)
- water velocity before nozzle (m) Now z 1 = z2 and h 2 = 0 (atmospheric pressure) so,
- height of nozzle centreline above using the result in Equation 10-2 this becomes:
datum (m) v2 v2
~+-~ =H = - 2
- frictional head loss in penstock and 2g n 2g Eq. 10-4
manifold to nozzle (m)
g - acceleration due to gravity (mfs2) Rearranging this:
133
Eq.10-5 Separating the sides of Equation 10-4 and using
Equation 10-6, the efficiency of the nozzle is:
This result is known as Torricelli's theorem, from

Evangelista Torricelli (1608-47) who established energy input
T/ = --=~-=---
by experiment that the flow from an orifice was energy output
~ (;~)I(~ + ;~ l
proportional to the square root of the head. In
practice the nozzle is not perfect, but by conven-
tion, the losses are accounted for by a velocity co-
efficient, Cv, so that the actual velocity in the jet is:
Eq.10-6
Eq. 10-12
There are two points to note about the definition
of Cv. Firstly, it gives the velocity of the jet at the If there is more than one jet the total flow for the
narrowest section of the jet (the vena contracta de- turbine is simply the flow for one jet times the num-
scribed in Section 2.2.1). Secondly, it calculates the ber of jets:
velocity from the total head, pressure head plus
velocity head, at the inlet to the nozzle. The
idealized experiment for measuring Cv has the
nozzle in the side of a large tank. The inlet condi-
tions are the pressure head but at zero velocity, so
the head represents the total energy going into the
nozzle. In a real nozzle, the water approaches the njet - number of jets in the turbine
nozzle down a pipe with some velocity, so both dnoz - diameter of nozzle opening
pressure and kinetic energy are involved. It hap-
pens that the definition of net head, Hn, used in Rearranging this, the nozzle diameter can be calcu-
hydropower is actually the total head, as can be lated for a given head, flow and number of nozzles:
seen from Equation 10-2. Hn is therefore the cor-
rect head to use in Equation 10-7. In practice the
velocities in the pipes in micro-hydro are usually
kept below 3m/s, and the velocity head is less than:
v2 32 Eq.10-14
h --- 0 -46
vel- 2g- 2X98 Eq. 10-7
10.2 Spear valves
So even if measured head is used rather than total The diameter of a jet emerging from a spear valve
head, the difference is not that significant. can be calculated as follows. Consider the spear
The flow through an ideal, perfectly efficient valve shown in Figure 10-2. The flow above the
nozzle would be: spear passes through the line BC, which is drawn
perpendicular to the spear across to the corner of
Q =Anoz"~2g. Hn Eq.10-8 the orifice. The spear valve is shown opened by a
distance's'. The direction of the flow is very nearly
In practice, it is less than this, and the discharge
perpendicular to BC.
coefficient, C0 , is defined as a fraction of the ideal
Now the line BC is actually part of a cone, with
flow:
its apex at A and a base diameter of dnoz- The area
of conical surface of a cone of side length l and base
Q = CD . Anoz~2g.Hn Eq.10-9
diameter d (not including the base) is:
Using the definition of the contraction coefficient
A= n.d.l
Cc in Equation 2-2: Eq. 10-15
2
Q= A2.v2
The area of a section of cone between /2 and /1 is:
= C,.A nozCv-~2g.H"
A= n.dJ-11 _ n .d2-l2
= Cc.Cv- ~2g.H" Eq.10-10
2 2 Eq. 10-16
Comparing Equation 10-9 with Equation 10-10 it Also:

can be seen that:
Eq.10-11 Eq. 10-17
134
-J f
dnozzle
I. .I
Fig. 1Q-2: Spear valve geometry
Using this result in Equation 10-16 the area of the It is now possible to write the equation for the
cone section can be written: area through which the water is leaving the spear
A= 1r.d1.!J lLdd2 valve, using Equation 10-18 and Equation 10-19:
2 2 7r
A spear =-.b.( d not+ dspear )
/~ 2
= -7r .( d,./,-dJ--) n . a a
2 II =-.s.sm-.( dnoz+ dnoz-2b.cos-)
2 2 . 2
Jr.d I (/~ -{~)
=----- n .a .a a
2 . II =-.s.sm-.(2d noz-2s.sm-.cos-)
2 2 2 2
n.d1 (l~-h).(!J+l2)
= 2 II
n a 2d ,-s.sma)
=-.s.sm-.(
00
2 2
7r
=2 (fr h). ( d I+ d 2) Eq. 10-18 -- n.sm2.(s.
. a d no,- I 2 .
2 s .sma) Eq. 10_20
Putting the length of BC as b. the following for-
mulae can also be written from the geometry in If the speed at the conical section is assumed to be
Figure 10-2: the same as the jet speed, then by continuity the
. a effective area of the spear valve must be the same
b = s.sm- as the jet area, and:
2
a A spear -- TC.d2jet
4 = TC.SIn 2a ( S. d noz - 2I S 2 .sma
. )
dspcar =dnM - 2b.cos-
2 Eq. 10-/9 Eq. 10-21
135
This leads to the result: two by the splitter ridge and each half travels
around one side the bucket, emerging at edge. The
=2.sin ~
d 2jet .(2s.d002 - s2.sina) sides are angled outwards y, and the jet leaves the
Eq.J0-22 bucket at this angle. The exit velocity is slightly less
than the entry velocity due to friction and tur-
Some books have a different formula for diet for bulence losses as the water flows around the
instance, Nechleba (1957). They argue that the flow bucket: v2 = ~.(vi - vb) where ~ is an efficiency
is not parallel to the spear, but varies between the factor for flow in the bucket.
line of the spear and the line of the nozzle. They The force on the bucket in the direction of travel
therefore calculate Aspear using a cone that bisects is found from the rate of change of momentum
the angle between the spear and the nozzle. The between the water entering and leaving the bucket.
resulting equation is much more complicated, but
gives practically the same value for diet The justi- F= ri1b.[(vi- vb)- {-~.(vi- vb).cosy}]
fication for the method used here is that the spear = mb.(vi- vb).(l +~.cosy)
has the greatest influence on the flow direction be- Eq. 10-23
cause the water continues to travel along the spear
even after it has left the nozzle.
F - water force on the bucket (N)
For a given head, flow is directly proportional to
mb - mass flow rate into bucket (kgls)
the jet area, so Equation 10--21 can be used to cal-
culate the flow. The results are plotted on the graph vi - jet velocity (m/s)
vb - bucket velocity, runner tangential ve-
in Figure 2-3.
locity at PCD (rn!s)
s - efficiency factor for flow in bucket
10.3 Turbine power and efficiency y - angle of bucket sides
10.3.1 Simple Theory

If we define x as the speed ratio between the tan-
Consider the single Pelton bucket shown in Figure
gential speed of the bucket at the runner PCD and
10--3. The jet velocity is vi, and it chases after the
the jet speed:
bucket, which is moving at the slower speed vb. The
jet enters the bucket at a relative velocity (vi - vb),
as shown in the vector diagram. The jet is split in
Eq.J0-24
Then Equation 10--23 becomes:

0) ( .,--
~ /, F = mb. vi. (1 - x) (1 +~.cosy) Eq. 10.25

Velocities relative to j I
the housing (absolute),
I The power input P to the runner is:
,
P= F.vb
= mb. vj .(1- x).( 1+~.cosy). vb
=mb. v~ .x.(l- x).(l +~.cosy)
---L Eq.J0-26
-T---
---
y ----
---
:.-~--- Now the total kinetic power in the water hitting the
bucket is:
Eq. 10-27
~ VECTOR DIAGRAM
J<t volocity ~ So the efficiency of the turbine 11 (as a fraction) is:
Bucket velocity v b (v r vb) mb. vJ .x.(!- x ).( 1 +~.cosy)

Velocity of jet relative to bucket T] I 2
:~mb.vt
Fig. 1G-3: Simplified diagram of flow in a Pelton =2x.(l- x).(l +~.cosy) Eq. 10-28
136
Differentiating this to find the maximum: 10.3.2 Water missing the runner
Figure 10-4 shows a diagram of buckets in a jet.
Though the whole jet is shown, we shall only con-
dTJ =2(1-2x).(l+s.cosy)=O
dx sider what happens to the jet on its centreline. For
.
.. x=:~
I
Eq. 10-29 the bucket in position A, the end of the splitter
ridge P is just on the centreline. This leaves a sec-
tion of the jet / 1 cut off from the main jet but conti-
So the efficiency is maximum when x = 0.5. Inter-
nuing on to the next bucket. bucket B. B' shows the
estingly, this is independent of the friction in the
position at which the rear end of the length 11 fi-
bucket, 1;, or the angle, y.
nally hits the bucket. (To simplify the calculations,
the splitter ridge is drawn as a radial line, which is
Force on a bucket not usually the case.)
To calculate the force on the bucket we need to In moving from B to B' the bucket turns through
know the mass flow entering the bucket: an angle of ( 08 + 0A -5). This takes a time:
88 +e., -8
mb = Aj.,.( vi- v" }Pwatcr tl = w Eq. 10-32
= Pwatcr .Aj vj .(1- x) w - angular velocity of runner (rad/s)
=m(l-x) Eq. 10-30 e. 8 - angles (rad)
Aict - cross-sectional area of the jet (m 2 ) Now from the definition of the velocity ratio x
Pwater - density of water (kg/m 3 ) above:
m - total mass flow in jet (kg/s)
w. R = x.vi Eq. 10-33
Substituting this into Equation 10--25: R - radius of pitch circle = PCD/2 (m)
F= m.vt<l-x) 2 .(1 + c;.cosy) Eq.10-31 So:
This is maximum when the runner is stationary, x =

0. Eq. 10-34
/
V
Jet centreline
~
Fig. 1Q-4: Angles at which buckets catch the jet
137
In this time the jet has to cover a distance in a
straight line of (R.tanOa + R.tanOA), which takes
(1 1 - PP') (vi .tA-+A'- 2R. tanoA)
( 8~R)
time: lb
Eq.10-35 x.( vi.[~]- 2R. tanOA)

Since the jet and the bucket arrive at B' at the same 8.R
(l
time, these two times t 1 are equal and:
0a + 8A- 8 =X. (tan 0a + tan8A) 28A.Rl -2R.tan8A )

x. vi.~
Eq. 10.36
8.R
As x increases, the runner turns further before the 2(0A -x.tan8A)
last bit of cut-off jet enters bucket B. There is a = 8 Eq.10-39
limiting value Of X when 98 = 9A, When:
(20A- 8) So, multiplying Equation 10-28 by this additional

X - loss factor, the efficiency for x > x 1im becomes:
lim- 2tan 8 Eq.10-37
A
e
For the bucket design used in this book, A = 36' 8
71 = 4 x.(l- x). ( l +<;.cosy) .
(oA-x.tanoJ
= 20, and xlim = 0.62. What this equation says is that 8
at high runner speeds, when x is greater than xlim a
portion of the jet never catches up with the bucket,
Eq.10-40
and this water passes straight through the runner. This is the major cause of the asymmetry of a real
Some qualification needs to be made. The analysis Pelton efficiency curve. The effect is also discussed
has been done for the jet centreline only. The lower in Section 2.3.2.
portion of the jet is cut off by the splitter ridge at a There is another factor that also causes the jet to
lower e, and starts to miss the runner at a lower 'miss' the runner. Above a certain value of x, the
value of x. The top of the jet does not pass through bucket actually travels faster than the jet at high
the runner until a higher value of x. The xlim derived values of e. In Figure 10-6 point D is on the splitter
here represents an average value for the jet. ridge. The length CD is R.tane. The apparent velocity
Note that a modified version of Equation 10-37 of the bucket along the path of the jet centreline (the
can be used to calculate the minimum number of line of CD) can be found by differentiating.
buckets required on a runner. The part of the jet
that misses the runner first is the bottom, the el- d(CD)
vb=~
ement furthest from the runner centreline, and a
modified eA for this position needs to be used. _ d(R.tanO)
The reduction in efficiency due to water missing
the runner can now be calculated. The maximum dt
amount of the jet that each bucket can receive is = R.sec 2 O.ro Eq. 10-41
the amount leaving the nozzle in a time 0/ro, which Using Equation 10-33 this becomes:
has a length:
8 8.R 8.R vb = x. vi secze Eq.10-42
/b = - .v. =--.v; =
ro ' x.v; x Eq.10-38 From this equation vi = vb when:
Up to x 1im all the water enters the bucket, and so x = cos 2
~im Eq.10-43
Equation 10-38 holds. For x > xlim some of the
water misses the bucket. Figure 10-5 shows bucket The jet cannot catch the bucket at a given x if is e
A just touching the jet centreline, and the same larger than this. This only becomes a problem when
bucket later at position A' when it is just moving 8lim < 8A, which OCCUrS at speeds higher than:
off the centreline. The particle of water at P' for Eq. 10-44
position A' was a distance /2 back when the bucket
was in position A, and /2 can be found as the jet Above this limit, the water only does useful work
velocity multiplied by the time it takes the runner e
on the bucket when is less than olim either side of
to rotate from A to A'. The length of water that has vertical. For x > x 1im Equation 10-40 holds with
gone into the bucket is 12 minus the distance PP'. ~im used instead of eA'
The fraction of the total possible water hitting the For x :S; x 1im the efficiency equation is derived from
bucket is therefore: a multiplier similar to Equation 10-39, with ~im
138
P'
v
~-----
A
'--..
-- - -- ~I --- - .--
_.., A'
Fig. 1Q-5: Water missing a bucket for x > x1;m
line. Above this speed the splitter goes faster than

the jet, and only slows down enough to be caught at
an angle less than fJA. The angle Bs starts at a nega-
tive value (before vertical); at very low speeds it
tends to a value of -18, and is only cut off when the
next bucket occludes the jet. As the speed increases,
the position at which the last water hits the bucket
moves around, past the vertical, until it reaches the
value fJA. This is at xlim and beyond here Bs has the
R
effective value fJA; water starts to pass through the
runner without being caught. At a slightly higher
speed, Bs also meets the Ot;m line and follows it
down, as the splitter end again exceeds the speed of
the tail end of the cut-off length of jet.
The conclusions thus far can be summarized in
the following formulae:
For x ::;x1;m:
11 ,,;;,,m= 2x.(1- x).(l +~.cosy).
[(e8 . +eJ- x. (tan8 8 . +taneJ]
8
Eq. 10-45
Fig. 1D-6: A bucket moving in the line of the jet For x >X 1;m:
(e A- x.taneJ
11 x;,x,im=4x.(l-x). ( l+~.cosy) . 8
substituted for ()A or Bs if Otim is less than either of
these values. Note, however, that for the bucket ge-
Eq. 10-46
ometry used here Xelim is (just) greater than xlim It
should also be noted that when Xelim is exceeded, the Where:
jet will be acting on the rear of the buckets, slowing (28A-8)
X -
them down, though this effect is ignored here. lim - 2 tanO A Eq. 10-47
Figure 10-7 shows how Bs and Ot;m vary with x. for
Using ()A '= ()A ()A:=:; f'1t;m
This is drawn for the splitter-end radius of the
()A,= ()lim for ()A> ()lim
bucket design given in this manual. The angle fJA,
at which the splitter first enters the jet, is fixed by Bs = ()B for Bs :=:;lim
the geometry of the runner until it reaches the ()lim Bs = l'1tim for Bs > l'1tim
139
100
aoo
ooo
J
al 400
c
oC
200
oo
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Speed Ratto, x
Fig. 1Q-7: Graph of the variation of 68 and 61;m with x; drawn for a radius to the splitter ridge end of 62% PCD
With ~im derived from the equation: sweeps around the curves of the bucket and
comes out some way along from where it enters,
Eq.J0-48 but it makes the calculations much more compli-
cated to include this effect. However, since as a
The implications of this are discussed in Section 2.3.2. fraction of the radius from the runner centre the
displacement is quite small, it does not have a
10.3.3 Three-dimensional theory large effect on the results.
The analysis in Section 10.3.1 is a gross simplifica- o The jet emerges at the same instant as it enters.
tion of what is really going on in a Pelton. The flow This aga\v. is not true, but it is extremely difficult to
into and out of the bucket was analysed only with allow for the difference. Not only is there a time
the bucket at 90 to the jet, and all the water in the delay, but the bucket orientation changes during
jet was assumed to follow the path of the water on this period. The runner actually turns about 7-8
the jet centreline. The real flow pattern is much (at optimum speed) while the water goes round it.
more complex. o Considering a section across the bucket, the jet
When the bucket is at an angle 8 before the bot- leaves at the angle of the side wall y. This is true
tom the jet strikes it at an angle, the water flows when the jet leaves near the centre of the bucket
around the curves of the bucket, and it emerges at a but is not so accurate for the water that comes
range of completely different angles. This is illus- out further up or down it.
trated in Figure 2-8. o The whole jet behaves in the same way as the
A relatively simple analysis can take some ac- element on its centreline.
count of the varying inlet and outlet angles as the o The splitter ridge and the edge of the bucket are
runner turns through the jet. This is presented be- both on the same radial line from the runner
low. It still assumes that all the water follows the centre.
path of the element on the jet centreline, but it o In the side elevation, the jet is effectively re-
gives some interesting results that are quite close to flected in the bucket, so the angle at which it
what is found in real turbines. It can be shown that, enters, ai> is the same as the exit angle, a2 ; a1 =
even if each part of the jet is analysed separately, a2 = a. This is probably reasonable near the cen-
the resultiilg average corresponds quite well to the tre of the bucket, but less so towards the ends.
values calculated for the centre element (Kisioka &
Osawa, 1972). The case analysed is shown in Figure The force on the bucket from the jet entering it is
10-8. The main assumptions are given below: equal to the mass flow rate into the bucket multi-
plied by the inlet velocity relative to the bucket, w 1
o Viewed in side elevation, the jet leaves from the The component of this force normal to the bucket
same point it enters. This is not true, as the water is:
140
VELOCITIES RELATIVE TO BUCKET
wi,w2
a 1 =~=a
Vj
Jet velocity into bucket

Jet velocity
leaving bucket
VECTOR DIAGRAM
normal
to
bucket
., ...... ,
w1
........
.,."
.... t .. ~
\ ... ....
.... .-f.,
....,\..oo""
i .-"'
I\
\
\
,." .,.... \ \ \
f( .-"' I I I
\ ....... .... L._\ \
I
I
(
I (X2
~--\ 'Y II
I I I I I
\ \ \ \
I
\
\
\
\
\
1
\
\
I
\
\
\
'
\
I I \ \ \
\ \ ',
'
\ \
\ \ \
\
'
\ \ \
\
\
\
\
\
\
\
'
\ \
''
\ \
\
'
Fig. 10-8: Water velocities
Eq. 10-49 v, =Rp.ro

Similarly for the water leaving the bucket. The only _ _!!__ x.~
complication is that the velocity has an axial com- cos8. R
ponent, as shown in the detail in Figure 10--8. The =x.vj.sec8 Eq. 10-53
component of the force from the water leaving the
bucket, normal to the bucket, is: Substituting from Equation 10-52 and Equation
10-53 into Equation 10-51:
Eq.10-50
Pb = rilb.x.vi.secO.w 1.cosa1.(1 +~.cosy)
The instantaneous rate of work done by both these
forces is equal to the force multiplied by the speed, Eq.10-54
which is the tangential velocity of point P, vt" Since
both the force of the jet pushing on the bucket and From the vector diagram in Figure 10-8, and using
the reaction force of it leaving add together, these Equation 10-53:
two rates of work add to give the power output vi.cos0 = w 1 .cosa + v1 = w 1 .cosa + x. vi.sec0
from the bucket, Pb. :. w 1 .cosa = vi (cosO- x.secO)
Pb = rilb.w 1.cosa.v,+ ril.w 2.cosa.cosy.v, Eq.10-55
= rilb.v,.cosa.(w 1 + w2.cosy)
Substituting this result into Equation 10-54:
Eq. 10-51
Allowing for friction in the bucket again, as in Pb = n1b.vi.x.secO.[vi.(cose- x.sec0)].(1 +~.cosy)
Equation 10-23: = mb. v2i.x.(1 - x.sec20).(1 + ~.cosy)
w2 = ~-Wl Eq. 10-52 Eq. 10-56
The velocity v1 can be derived as: Dividing this by the kinetic energy per second
141
going into the bucket (from Equation 10-24) gives found by expanding to sec28.(1 + tan28) = sec28 +
the instantaneous efficiency at 9: (sec8.tan8).(sec8.tan8), integrating the second term
2 by parts, and rearranging to make the sec48 integral

mb. vj .x.(l- x.sec 8 ).( 1+~.cosy)
2
the subject. The result is:
1'/inst = ( mb.vi_
. 2) W,= m. vrR (l +~.cosy).
2
= 2x.(I- x.sec 28 ).( 1+~.cosy)
Eq. 10-57
Note that, for 8 = 0, this is the same as Equation
10-28.
We can now integrate the formula for Pb to find
the work done on the bucket as it moves from a The energy in the jet per bucket is the nozzle power
position (}A before the bottom to 8a after, as in Sec- multiplied by the time taken for the runner to turn
tion 10.3.2. We shall also need to know the mass flow by one bucket:
into the bucket, which is found by subtracting the 2 i: 2 i: i:R
bucket speed vb (from Equation 10-42, and in the m.v1 v m.v, v.R m.vi .v.
w =-- - = - - --=_....:.___
same direction as vi) from the jet speed: ' 2 co 2 x. v j 2x
mb = p.Aiet( vi - vb) Eq.10-61
= p.Ajel'( vj - x. vi.sec 28)
= p.Ajetv/1 - x.sec2 8) The hydraulic efficiency is the work done on the
= m(l - x.sec28) Eq. 10-58 bucket, Wb, divided by the energy in the jet, Wi.
The work done on the bucket as it moves can be 1'/=-b
w.
found by integrating the power, using Equation 10- ~
..
33, Equation 10-58, Equation 10-56:
~ {88 . +8A.)
w = J~.dt 2
= x. { +".Cosy) + x.(x- 2).(tan88 + tanOA.)
1
8 2
9
dt + ~ 3 3
.(tan 6 8 . +tan 8A.)
= J Jt.d8.-
"
-9 d(}
A
Eq.10-62
o. d(}
= J ~.-
(1) The formula is completed by allowing for the water
missing the runner as derived in the previous
section.
Where (}A'= (}A for (}As; ~im
fJA,= ~im for fJA > 9lim
and 8a = 8a for 8a s; lim
8a = ~im for 8a > elim
The efficiency curves produced by this theory are
shown in Figure 2-11, and they bear a good like-
ness to a real runner. A full discussion on efficiency
6" is included in the same section.
= m.vi .R. (1 +~.cosy) J(1- x.sec 2 ef .d8
-OA 10.3.4 Windage losses
The standard formula for drag resistance for an
= m.vi .R. (1 +~.cosy). object moving in a fluid is:
2
lie D = Ko.Ar.q =Ko.Ar p.u
2
J(1- 2x.sec 8 + x 2 .sec4 8 ).d8 2
-o. Eq.10-59 Eq.10-63
Now the integral sec2(} is a standard case that can D - drag force (N)
be looked up in tables. The integral of sec4(} can be K0 - drag coefficient
142
Ar - frontal area of the object presented in for each other. K 0 of about 4 seems to give areas-
the direction of movement (m2) onable representation of experimental results.
q - dynamic pressure (N/m2) For the bucket pattern described in this book, Ar
p - density of the fluid (kgfm3) is approximately 0.093 x PCD 2, where PCD is mea-
u - velocity of the object (m/s) sured in metres. For the design jet diameter of 11%
PCD, the ratio A/Ajet = 9.8.
To adapt this to the case of a Pelton runner, we
The fluid density in the casing is basically the
assume that the velocity u is the bucket speed at the
density of air. Atmospheric density is about 1.23
PCD. Multiplying the drag force by the speed to
obtain the power lost to the drag, and using Equa- kg/m 3 at sea level and 15C, giving PairtPwater =
tion 10-24. 0.00123. This may need to be increased to allow for
)
water spray in certain cases. At low head there is
pwindage= D.U KD" A ,-.-2-.
Pc v;, Vb Jess spray in the casing, but at high heads the casing
=
is full of spray, and a higher Pcasing is appropriate.
So for an 11% PCD jet, the windage efficiency
= 2I Ko Ar Pc vb3 Joss measured as a fraction is approximately:
= f K0 .ArPc .(x.vJ 11w = 0.05 X x3 Eq. 10-67
Eq. 10-64 This gives a Joss of 0.5% at x = 0.46, which is realis-

tic as windage losses are not high at operating
- power lost to windage (W) speeds. Windage and friction together typically
- tangential bucket velocity at PCD give 1% Joss.
(m/s) There are some other minor components of loss,
Pc - density of air/spray within casing such as the drag of the runner hub and seal discs
(kgfm3) moving in the casing fluid, but these are
- jet velocity (m/s) insignificant.
Now the power in the jet is:
10.3.5 Mechanical friction losses
m.v] (Pw .vi. Aiet Jv/ The frictional torque of the bearings and seals is
PJet =-2-= 2 basically independent of speed. Representing this
as Tr the efficiency loss for friction can be calcu-
= 2I P.. A it vi3 Eq. 10-65 lated in the same manner as the windage loss:
- mass flow in jet (kg/s) Prnction T,. w T,- x. vi F,-. vi
11r = - - = - - = - . - - = - - . x
- density of water (kgfm3) pjct pjct R pjct pjct
- cross-sectional area of the jet (m)
Eq. 10-68
So the efficiency Joss due to the windage is:
Fr is the force at the PCD required to turn the
pwindage
7Jw=--- runner, measured in Newtons.
pjet This is the only efficiency equation derived so far
that does not scale. The other efficiencies are
theoretically (though not in practice) all indepen-
dent of the actual head or flow or size of the tur-
bine. Because the friction loss for a given turbine is
= K 0 .(~].(~).x 3
constant, its effect on the overall efficiency is
dependent on the turbine power. If the turbine is
Aiet P "' Eq. 10-66
operating at maximum flow under a high head, the
Note the dependence on x 3 , and the use of the two friction loss is relatively insignificant. If the turbine
ratios. is operated with a small flow under a low head, the
A runner is quite a complex shape, and K 0 is friction can be relatively large.
difficult to establish. As an approximation, an indi-
vidual bucket can be likened to a flat circular disc.
K 0 varies with the Reynolds number of the flow
10.4 Bucket stem stress calculations
past the bucket, but for the conditions within a This section shows how to check the stress in the
micro-hydro turbine (with Reynolds number in the stem of the bucket. There are two load cases that
range 10L106) a disc has a K 0 of about 1.5. can cause the bucket to break off. The first is runa-
However, K 0 for the runner cannot be found by way, when the external load is removed from the
multiplying 1.5 by the number of buckets. This turbine and the runner accelerates to a high speed.
would grossly over-estimate the drag, as the This produces a large centrifugal force in the
buckets follows each other round and 'clear a path' bucket, which can snap the stem. The second case is
143
the fatigue load caused by bending stress on the COrunaway- angular speed of runner at runaway
stem due to the water hitting the bucket every time (rad/s)
it passes a nozzle. Though the first case needs to be Pbucket -density of bucket material (kglm3)
checked, it is usually the fatigue load that causes vbucket- volume of bucket and stem (m3)
failure. Nrunaway- runner speed at runaway (rpm)
10.4.1 Runaway load If the turbine is being operated at its optimum

The calculation for the runaway case is quite point, the runaway speed is about 1.8 times the
straightforward. The centrifugal force can be calcu- optimum speed. The area of the section A-A has an
lated from the mass of the bucket and its speed, and area of 0.16 x 0.122 =0.0195 x oz. The stress in the
the stress is simply the area of the stem divided by stem is therefore:
this force. The volume of the bucket design given 2
here (without any stem) is approximately 0.0038 x Tr. N runaway )
Rg ( 30
03. The mass of the stem depends on its design. (j =F runaway
V
=______
Pbucket'
:;______
bucket'
;____
With the basic stem shown in Figure 4-3, the 1
A stem A stem
bucket and stem have a volume of 0.0069 x 03. 3 2
Consider the specific case of the single-piece X0 0040X D X0 48 X D.Tr N;_way
Pbucket
casting shown in Figure 10-9. The buckets are inte- = 30 2 X0 0195. D 2

gral with the hub, and the stress is calculated across
= 0. 0011 XPbucket' D2. N ;...,away Eq.J0-70
section A-A. The volume of the bucket outside the
section is 0.0040 x 03, and the centre of this mass is - tensile stress at runaway in bucket
at a radius 0.48 x D. The centrifugal force is given stem (Nfm2)
by: - area of stem at weakest section (m2)
A stem
D - runner PCO (m)
Frunaway = mbucket X R}l, 2
X liJ runaway
2
_ 2tr N runaway This can be multiplied by 106 to give a result in
- Pbucket Vbucket R8 (
60 ) Eq. JO-(j9 N/mm 2. This stress should be less than the yield
point of the material, with a suitable safety factor.
Frunaway - centrifugal force on the bucket at The method is similar for clamped buckets, but
runaway (N) the weakest section is at a different place. For the
- bucket mass (kg) clamping arrangement shown in Figure 10-10(a),
- radius of bucket centre of mass the fracture will occur at the section through the
from runner centre (m) outer bolt hole, B-B. The part of the bucket outside
CENTRIFUGAL FORCE
Fig. 10--9: Single-piece casting, buckets integral with the hub. Dimensions are in % PCD
144
JET FORCE
CENTRIFUGAL FORCE
(a) Clamped attachment
I
I
I
---\---
\I
I
I
/
I ;
I
" "
CENTRIFUGAL FORCE
(b) Welded attachment
Fig. 1Q-1 0: Clamped and welded bucket attachment designs for medium heads
B-B has a volume of 0.0052 x 03, the centre of mass buckets are neglected. This is because all the
is at a radius of 0.44 x D, and the area of the section buckets are being flung out at the same time, so
is 0.0071 x 02. Equation 10-70 can be used to cal- they all pull on the hub welds together. In practice,
culate the stress. Note that the holes shown in Fig- there is a 'hoop stress' contribution from between
ure 10-lO(a) are 2.5% and 3% PCD, but this may the bucket welds, and the real stress will be some-
need to be varied for a given PCD to fit actual bolt what lower than calculated here.
sizes.
The welded construction in Figure 10-lO(b) is 10.4.2 Fatigue load
based on Figure 4-11. The volume of the bucket The main fatigue load on a bucket comes from the
(including half the weld fillets) is approximately jet force. This creates a bending stress in the stem
0.0049 x 0 3 , and the centre of gravity is at a radius every time a bucket passes a jet. Most turbines will
of 0.45 x D. The force is carried by the weld con- exceed one million bending cycles on the buckets
necting the bucket to the hub, and the weld area within a few weeks, or even days, of operation. The
per bucket is 0.0108 x 0 2 . The welds between the point at which the worst fatigue stress occurs de-
145
pends on how the buckets are fixed. For the single- drawn for the maximum bending moment position
piece casting shown in Figure 10-9, it is plain that and has an effective moment arm of 0.183 x D
about the neutral axis, and the section modulus Z
the maximum bending moment is going to occur at

the section where each bucket joins the disk, A-A. for the section C-C is 0.00022 x D3. Feeding in the
This will give a maximum tensile bending stress at figures into the equivalent to Equation 10-71 gives
point 'X'. The jet force is drawn for the maximum the stress. Note that if the hub sides are not a press
bending moment position, and has an effective mo- fit into the bucket sides, the bending occurs lower
ment arm of 0.163 x D about the neutral axis. The down the stem, and the calculations need to be
section modulus Z for the rectangular section A-A done for section B-B. Here the moment arm is 0.22
is 0.16 x 0.1222/6 = 0.00040 x D3. The stress is then x D, and Z is 0.00015 x D3. Using these figures, the
found from the equation: maximum allowable head is only 56% of that for a
M Fjet X ajet Fjet X 0 163 X D 408 X Fiet clamped fit. Take care!
Gt =z = Z = 000040x D 3 = D2
For the welded design in Figure 10-10(b), crack-
ing is most likely to start from the root of the weld
Eq. 10-71 at the base of the splitter ridge, marked 'Z'. The jet
force has a moment arm of 0.121 x D to this point,
- fatigue stress in stem (N/mm2) and the section modulus Z of the bucket through
- bending moment (N.mm) the section marked 'D-D' is approximately 0.00025
- stem section modulus= IIYmax (mm) X 1Q-4 X D3.
- jet force on bucket (N)
- moment arm from jet to section (m)
- runner PCD (mm) 10.4.3 Stress limits
Defining the allowable stresses for a Pelton runner
Note that by putting D into the equation in (mm),
is difficult, because there are so many variables. A
instead of (m), the stress comes out in N/mm 2. The
high-quality casting from a known material, care-
jet force can be calculated from:
fully ground and finished, and machined to fit the
Fj.1 =PwQjetCv.~2g.H. (1- xt_(l +~.cosy) runner with no poor details or stress raisers, will be
quite strong. A rough-cast bucket made from what-
Eq. 10-72 ever scrap iron happens to be lying around a
foundry, with holes and inclusions in it, left with a
Pw - density of water = 1000 (kg!m3)
rough surface finish, and machined with some sharp
Qiet - flow from one jet (m3/s)
corners, will be much weaker. Materials too can
Cv - nozzle coefficient of velocity
vary enormously. Impurities in the mix, and care-
g - acceleration due to gravity (m/s2)
less foundry procedure, can make the material
Hn - net head (m)
much weaker than it should be. Heat treatment can
x - ratio of bucket velocity to jet velocity
improve the strength of some materials. For these
~ - efficiency factor for flow in bucket
reasons, this section takes quite conservative values
y - outlet angle of bucket sides
for material properties, and uses large safety
This formula is derived from Equation 10-31. At factors.
optimum operating speed, x = 0.46. y is 15 for the Table 10-1 gives the properties of possible
bucket here. For most calculations, Cv can be taken bucket materials described in Section 4.1.3. Note
as 0.97, and~ as 0.95. that the fatigue design stress in the table is the
Now consider the clamped attachment in Figure maximum recommended stress for a non-reversing
10-lO(a). At first sight, it looks as if each bucket bending load. Where possible, these figures are
rests on the bucket behind it, touching at point 'X'. based on published data. Safety factors for tensile
This would make a solid disk in the centre, load are between 2.5 and 4, depending on the ma-
equivalent to the single-piece casting above. In terial. Where the endurance limit, O'e, is not known
practice, this is not the case. Variations in the ma- for a particular steel, it can be estimated at 40% of
chining of each bucket will mean that there will be the ultimate tensile strength. This is the figure for a
gaps between some of the buckets. If there is no fatigue test piece subjected to a fully reversing load
gap when the runner is made, corrosion and move- that gives a stress ae. The fatigue case we are
ment will create one. So at what point do we con- considering here has the stress going from zero to a
sider the bucket to be fixed? The trick is to make maximum value of <Tc, as given by Equation 10-71.
the notch on the sides of the stem a press fit on the A Goodman diagram can be drawn to show that
hub side plates. By doing this we know that the the maximum stress is approximately 1.4 x ae when
buckets will be held at section C-C, and the max- the endurance limit is 40% of the ultimate tensile
imum bending moment will be at point 'Y'. Since strength. However, the allowable stress is much
this is a thin point in the stem, Y will also have the lower than this, because the bucket will have stress
maximum tensile bending stress. The jet force is concentrations and a much poorer surface finish
146
Table 10-1: Properties of materials for Pelton bucket castings. Figures in italics are estimated
Material Ultimate Proof or Endurance Density Tensile design Fatigue design

tensile stress yield stress limit stress stress range
ault aP or ay ae p at a,
(N/mm 2) (N/mm2) (N/mm 2) (N/mm2) (N/mm 2)
Brass, 60/40 280 90 8300 30 20

Brass, Grade SCB4 250 80 8300 30 20
Silicon Bronze, C87200 380 172 8600 60 30
Cast Iron, Grade 220 220 140 100 7150 54 25
Cast Iron, Grade 260 260 170 120 7200 65 30
Cast Steel, Grade A 1 430 230 172 7830 58 35
Cast Steel, Grade A3 540 295 216 7830 74 45
Stainless Steel, 304C12 430 215 172 7970 54 35
Welded joints as material 20
than the test piece. It is also subject to water corro- runaway speed and jet force into the equations in
sion. Putting these factors all together, the working the preceding sections. The bucket shape, including
stress is lower than the endurance limit by a factor the stem and attachment system, and the ratio of
of 4 to 5, and lower than the ultimate tensile the jet diameter to the PCD, must be kept the same
strength by a factor of between 10 and 12.5. for this to hold true. This means that one can esti-
The brasses and bronze listed in Table 10-1 do mate the maximum heads allowable for each ma-
not have defined fatigue limits, but the fatigue life terial and attachment system. The results of all the
does increase substantially as the stress is reduced above calculations for the standard buckets and at-
to low levels. The design stresses here are set low to tachment methods used here and a 12% PCD jet
give adequate fatigue life. Many stainless steels also (the largest recommended) are tabulated in terms
have no fatigue limit, and again the stresses need to of the limiting heads in Table 4-1 in Section 4.1.4.
be kept low to give an appropriate life. Welded All these calculations err on the safe side. If a jet
joints do not, in theory, have a fatigue limit, but the diameter less than 12% PCD is being used, the
prediction of life at very high cycles is uncertain head can be increased, but it may also be possible
(for further information see BS 7608:1993). The fig- to increase the allowable head by reducing the
ures below derive from experience of welded safety factors. This should only be done if one can
micro-hydro runners surviving. control the quality of the buckets. If sample pieces
Table 10-2 summarizes the bucket properties for of the material are tested and found to meet or
the bucket designs given above, for various attach- exceed the specification, if the castings are of good
ment methods. quality, with no inclusions or defects, and if the
If these figures are substituted into the formula finishing and machining are done to a high stand-
in Sections 10.4.1 and 10.4.2, the attachment ard, then the safety factors of 3 (for tensile load)
stresses can be calculated, and compared with the and 4 (for fatigue load) can be reduced. Only in-
figures in Table 10-1. crease the allowable weld stress if the weld is care-
For a given geometry of turbine, both the tensile fully controlled, crack-detection is done, and the
stress due to runaway and the fatigue stress due to weld itself is ground or peened to a smooth finish. It
the jet force are independent of the PCD. This can is also possible to increase the allowable head by
readily be shown by substituting equations for modifying the stem to lower the stresses.
Table 10-2: Summary of bucket dimensions for stress calculations for different attachment methods
Attachment method Bucket volume Radius to centre Stem area for Moment arm of Section modulus
of gravity tensile failure jet about Astem for fatigue
vb Rg Astem ajet z
(x fJ3) (x 0) (x 02) (x 0) (x fJ3)
No stem 0.0038 0.48

Stem to Figure 4-3 0.0069 0.39
Single-piece casting 0.0040 0.48 0.0195 0.163 0.00040
Clamped 0.0052 0.44 0.0071 0.183 0.00022
Welded 0.0049 0.45 0.0108 0.121 0.00025
147
10.5 Bolted buckets Assuming infinite rigidity again, the amount of the
moment or secondary shear that a bolt takes is pro-
This section gives the background for the calcula-
portional to its area multiplied by its moment arm

tion of the bolted clamping friction in Section 4.2.2.
from the centroid:
The bolt forces are calculated for a simple two-bolt
clamped case, but the general theory is presented
too, so that the reader can do calculations for more
l=_li_= ... =~
A1 .1j A2.r2 A. r. Eq.10-76
complex arrangements.
Combining Equation 10-75 and Equation 10-76:
10.5.1 Standard analysis
(AI.1j2 + ~.r22+ ...+A. r})
Consider the group of n bolts shown in Figure lO-
ll. The force F applies both a direct shear force to T; = F.a. A;.r; Eq. 10-77
the bolts, and a bending moment. A standard text-
Both S; and T; can be drawn as vectors, and the
book analysis, for example Shigley & Mischke
actual force on bolt 'i' is the vector sum of these
(1989), considers these two components separately.
two, as shown in Figure 10-11. This is a standard
If it is assumed that the structure around the bolts
method for calculating bolt forces. The assumption
is completely rigid, then the direct, or primary,
that the parts being bolted are completely rigid is
shear force is distributed among the bolts in pro-
never actually true, but the simplicity of the
portion to their area.
method has lead to it being widely adopted. It is
A. also assumed that the bolt centres and the hole
S=F(
A1 + A2+...+An centres match each other exactly, so that the loads
1 I )
Eq. 10-73
are distributed into the bolts evenly. Therefore this
S; - direct shear force component on bolt i calculation does not account for internal loads gen-
A; - cross-sectional area of bolt i erates by misalignment as the bolts are fitted and
tightened - which can be considerable. Neverthe-
less, it usually gives results that are a reasonable
approximation to the real loads, and is certainly
better than doing no calculations at all.
10.5.2 Analysis of a bolted bucket

Now consider the bolted bucket in Figure 10-12.
The jet force is drawn in the position that gives the
maximum bending moment as discussed in Section
10.4.2. From Equation 10-73, the direct shear force
in each bolt is given by:
AI
sl =F jet. ( A + ~)
1
0 As
A2
Eq.J0-78
By symmetry, the centroid C must be on the cen-
X
treline of the two bolts, so Equation 10-74 can be
simplified to:
Fig. 1o-11 : A group of n bolts of varying cross-
sectional area, with a force F acting on the
joint they are clamping
Eq.10-79
To find the effect of the moment, the centroid of
the bolts has to be found. In the x-direction, this is Equation 10-77 becomes:
given by:
AI.(Rc- Rl)
=
(A1.x1 + A2.x2+...+An.x.) ~=Fjet'a'[ 2 2]
X AI.(Rc-RI) +A2.(R2-Rc)
c (A 1 + A2+...+An) Eq. 10-74
Similarly for y c If the shear force on bolt 'i' from ~.(R2 -RJ
the moment is T;, then for equilibrium:
F.a = T1.r1 + T2 .r2 + ... + Tn.rn
Eq.10-80
Eq. 10-75
148
-Rso
\
\ \ '
Rc
\.-
~ __.-/
'
""' '
""'
a
.... ..........._
Fjet
----
JET FORCE
Fig. 10-12: A Pelton bucket clamped between the two sides of a hub by two bolts. (Dimensions% PCD)
However, for the two-bolt case, the moment is re- 10.6.1 Static loads
solved as a simple couple, and the above formulae First consider the static case. The bending and tor-
can be much reduced to: sion stresses in a solid round shaft can be found
from basic beam and torsion theory:
32M
Eq. 10-81 (J' X= - - 3 -
nd
The primary and secondary shear forces can be l6T
r xy = - - 3
Eq. 10-82
added vectorially as shown in Figure 10-12. Section nd
4.2.2 in the main text gives an example of the cal-
culation procedure. crx - bending stress (N/m2)
'txy - torsion stress (N/m2)
M - bending moment (Nm)
10.6 Shaft design T - torsion (Nm)
This section derives and explains the background d - shaft diameter (m)
of the formulae used in Section 4.8.2. It is a calcula-
tion procedure for finding the material diameter These need to be added and compared to the
required for a rotating shaft only subjected to tor- failure stress of the material. For ductile materials
sion and bending moments. such as steel, the von Mises-Hencky distortion-
149
stresses on
energy theory gives the best prediction of failure. cross-section
t
For the two-dimensional stress case being con-
sidered here, the von Mises stress is:
avm - von Mises equivalent stress

aA as - principal stresses
For torsion and bending only, the Mohr's circle can

be drawn as shown in Figure 10-13, and the princi-
pal non-zero stresses are:
Fig. 1Q-13: Mohr circle for torsion and bending only in

CJ', a shaft
CJ'A=2+rmax
CJ',
(J'B =--rmax 10.6.2 Fatigue loads
2 Consider a point on the surface of the shaft. At a
2
given torque, the stress due to torsion is constant.
rmax = -2
2 ((J'x) +r,y2 The bending moment stress, however, alternates
Eq.10-84
between a positive and negative maximum value as
the shaft rotates. This is a special case of shaft load-
Substituting into the previous equation: ing - albeit a very common one - for which experi-
mental results can be analysed very simply
(Shigley, 1986). It is found that the bending-fatigue
strength of a shaft for this case is independent of
the torsional stress.
So if the bending stress in the shaft is kept below
the fatigue endurance limit for the shaft, with an
Then substituting from Equation 10-82:
appropriate safety factor, the shaft will have infinite
fatigue life. The bending stress is given by Equation
10-82, so:
(J'cc32M
SFr- nd 3
Eq. 10-86 1
=(32M.SFr)
1r (J'cc Eq. 10-88
This stress needs to be less than the yield stress of
the material, with an appropriate safety factor: df - -minimum shaft diameter
aec corrected endurance limit for the shaft
-
material
(J' yield _
----- M +-
32 ~3T2 SFf - safety factor
SFy nd; 4
1 1
.. dy =[ 32SF'y(M + 3T ) ]
1r (J' yield
2
4
2 The endurance limit found from a standard test has
to be corrected for a real shaft. This is because the
rotating-beam specimens used to find endurance
limits have cross-sections equal to <!>75 mm, have a
Eq. 10-87 shape that is free of stress concentrations, and have
polished surfaces. Most turbine shafts are larger,
have a variety of shapes, and have machined
- minimum shaft diameter (m) surfaces, all of which reduce the fatigue life. The
- yield stress for shaft material (Nfm2) endurance limit from a rotating beam test has to be
- safety factor reduced by approximately 0.7 for an average
150
machined micro-hydro shaft (0.8 if the shaft is pipework components are assigned coefficients 'K',
ground). If the endurance limit is not known, it can and the head loss across them is equal to the velocity
be estimated as half the ultimate tensile strength. head multiplied by this coefficient. By using
Correcting this for a real shaft: equivalent values of K for all the components, the
relative size of each loss in each can easily be seen.
So, for a straight section of manifold pipework:
A word of caution, though. The estimates for ae

above are approximations only. In most cases they Eq. 10-89
will lead to reasonable designs, but they cannot be
guaranteed to be reliable in all cases. Fatigue is a Hr - head loss in a given section of pipe (m)
complicated subject, and a full treatment is beyond Kr - pipe friction head loss coefficient
the scope of this manual. If the fatigue life is crit- V - water velocity in the pipe (m/s)
ical, you are advised to consult a specialist text. g - gravity (9.8 mfs2)
Shigley (1986) gives a good introduction to the
subject. The coefficient Kr for a pipe depends on a num-
Note that the method used here to find the diam- ber of factors: the diameter, the length, the ve-
eter for a shaft is based on recent advances in the locity, and various water properties. In order to
understanding of fatigue. Many references still use show the link with conventional formulae, the
the 1927 American Society of Mechanical Engin- equation below shows Kr in terms of a friction co-
eers (ASME) code, even though this has been ob- efficient f The value of f can be calculated from
solete for many years. It uses some rules-of-thumb Equation 10-90, or looked up in a Moody diagram
for calculating the allowable shear stress within a in a hydraulics book, or in Kempe (1989).
material which are not accepted nowadays. Nev-
ertheless, it can be used to design shafts provided L
an ample safety allowance is included.
Kr =! d
10.7 Manifold loss calculations f = [

4. log
( k
3. 7d
574)]
+ R.~(i9
2
The following sections give data for calculating the

head losses in a manifold. As was stated in the V.d
discussion on manifolds in Section 4.10, it is not Re=- Eq. 10-90
v
usually necessary to do all these calculations - the
losses are generally small. If small pipes and high f - friction coefficient
water velocities are being used, or a very low head L - length of the pipe (m)
is being considered where the manifold loss could d - pipe bore (internal diameter) (m)
be quite high proportion of it, it is worth doing k - friction coefficient (m) - note units for
design calculations. It is instructive to do the full calculation!
calculations at least once just to get an idea of Re - Reynolds number for flow
where the losses come from, and to check what V - water velocity (m)
design features to avoid. v - kinematic viscosity of water (1.57 x 1Q-6
The losses in any particular route through a ma- m2 /s at 5C)
nifold are found by adding the pipe friction losses
for the straight lengths of pipe to the losses due to Typical values for k are given in Table 10-3.
the bends, bifurcations, valves and diameter They are given in millimetres, as universally found
changes. This calculation actually slightly overesti- in the literature. Remember to convert to metres to
mates the losses if the features are within a few put in the formula, or the results will be wildly
pipe diameters of each other, as the turbulence wrong! The value of k is roughly the average value
caused by the first one will not have fully dissipated for the roughness on the surface of the material. A
by the second, and so on. Nevertheless, the result pipe with heavy deposits on it can have a k of a few
gives an indication of the overall effect. For more millimetres.
detailed information on loss calculation, the reader A word of warning about using values of k or f
is referred to Miller (1990). from other literature. There are many different
friction loss formulae for pipes, and they use dif-
10.7.1 Pipe losses ferent roughness coefficients, which are not inter-
The head losses in the following sections refer to the changeable. Another common pipe loss equation,
'velocity head', V2/2g, which is the head that would the Manning equation, which is used in Inversin
appear if all the momentum of water travelling at (1986), uses a roughness coefficient normally la-
velocity V were turned into pressure. The various belled n, but n is not the same as k, and cannot be
151
Table 10-3: Roughness coefficients k for manifold and then calculate a value of Kr for each length of
pipes pipe.
Type of pipe Roughness

coefficient 10.7.2 Bend losses
k Manifold bends as shown in Figure 10-14 are usu-
(mm) ally made as mitred joints, which are much easier to
Welded steel, bare or light rust 0.02 make than radiused bends, certainly in steel. The
bend has (n-1) intermediate sections of pipe cover-
Welded steel, rough bitumen paint 0.10
or brush enamel
ing q,o each. Note the pipes on either side extend
into the bend area by r.sin (<jl/2). The head losses
HOPE, PVC with smooth joints 0.003 due to bends are calculated from the loss coeffi-
HOPE with internal beads at the 0.2 cient, Kb, in a similar way to the pipe losses above.
joints
vz
Hb = Kb2g Eq.J0-91
used in Equation 10-90. For reference, the equa-
tion used in this text, Equation 10-90 is a variant of Hb - head loss in bend (m)
the Colebrook-White equation, with this version
taken from Miller (1990). The Moody diagrams The graphs shown in Figure 10-15 and Figure
found in books are a graphical form of the 10-16 give Kb for common micro-hydro bends. The
Colebrook-White equation, and normally do use graphs are drawn for typical conditions (equivalent
the same definition of k as here. Be aware too that to a 2.5 m/s flow in a 100mm bore welded mild steel
some boo.ks use a different definition of the friction pipe). In theory, the loss coefficient varies with
coefficient f, which is still normally labelled f, but is Reynold's Number and roughness, but the graphs
a quarter of the value used here. The definition off give values of Kb accurate enough for most micro-
used here is easier to use, but the other definition hydro. For more detailed information on loss co-
arises more naturally in the derivation of the equa- efficients, see Miller (1990).
tion. It works best to calculate the value of f for Figure 10-15 shows the loss coefficient against
each section of pipe with different diameter or flow, the bend angle for mitred joints and for curved
r.sin(0/2)
= n.0
Fig. 10-14: Compound mitred eo bend with n joints

152
1.00
10.7.3 Bifurcation losses
0.90 Single Mitre~ d'I ~.
;'- .. ,
~-,a Bifurcations are where the flow is divided in two,
0.80 ~--- ;r
the junctions in the manifold to take the flow off to
0.70 I ~-:---.
"1
!
- .>r: J 9 the jets. The losses at bifurcations can be much
0.60 II higher than for valves or bends, and careful design
>2 0.50 r = Sd is essential. The following section allows the losses
I ).....~ to be calculated - or at least, estimated. More im-
t:::----
0.40
0.30 I _....-::: portantly, studying the graphs shows which sorts of

0.20 I ~~ bifurcations are reasonable and which ones are
likely to give large losses.
0.10 ~ t-- r=2d
~ ' r =3d The losses for geometry shown in Figure 10-17
0.00 are again calculated from coefficients, as follows:
0 20 40 60 80 100 120 140
fl[degrees]
Fig. 10--15: Loss coefficients for various bends (Miller,

1990)
bends with various radii. The figure is drawn for Re Eq. 10-92
= 1.5 x 105 and roughness of k = 0.25mm. (Note
that most literature has these curves in idealised H31 - head loss between inlet 3 and
form - for smooth pipes and Re = 106 - which give branch 1 (m)
rather lower values of Kb.) As would be expected, H32 - head loss between inlet 3 and
radiused bends are better than mitred joints. Al- branch 2 (m)
though it looks as if r = 5d bends have worse losses K31 K32 - loss coefficients
than the smaller bends, which is not what you v3 - water velocity at inlet 3 (m/s)
would expect, this is misleading because Kb in-
cludes the pipe losses, and larger bends are longer There is one coefficient for each branch, 1 and 2,
than short ones: if lengths of straight pipe were and the inlet is given the label 3. All coefficients are
added to the 2d bend to make the same length as referenced to the inlet velocity, V 3 . Note that for
the 5d bend, the overall loss would be higher. some branch conditions, the coefficients can be
Figure 10-16, shows the loss coefficients for 90 negative, corresponding to a head gain. The coeffi-
bends made out of multiple mitred joints. The dia- cients K are affected both by the geometry of the
gram shows a 3 x 30 compound mitre bend, with a bifurcation, and by the relative flow in each branch.
total angle of 90. It is drawn for the same flow There are so many variables for bifurcations that
conditions as Figure 10-15. The line for a smooth, it is not practicable to present all possibilities. The
radiused 90 bend is shown for comparison. Having K coefficients for joints not covered can be esti-
more segments in the bend lowers Kb, but there is mated by finding similar junctions in this text. If
no need to have too many. If r;?: 3d, mitre angles of
22Y2 are nearly as good as a radiused bend. (This is
1.0 01 Oo
1/~ ~I? ~~ v
v
~~ ~
/
one reason for recommending that bend radii should
v
I I
0.9
generally be kept above 3d in Section 4.10.1 ). I 14
v
/
Though the two graphs do not cover every bend,
I I 7 I [7
0.8 17
/
v
it is possible to estimate Kb for most bends by ex-
1/ I
0.7 t'----' V\_1.)
trapolating from values for similar types.
/ /
J7 v
-
0.6 ]:::\.?
[7
1.20
\
!7 [7 v
/
l---
v
/
1.00
1/1.~
1---
v v ~v
~Jx45-- /
0.4 v
/
~
0.80
\ /
vt;:--vI/v
~ / / /
~ 3x30" 0.3
f--./ ].?
" 0.60
4 X 22.5 ~ 1--
0.2
1\ /
\ ~~~ .
~
0.40
'~
Raldius
0.20
0.1 .
'
0
0.00 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
2.5 3.5 4.5 5.5 6
rid Flow ratio. Q,IQ,
Fig. 10--16: Loss coefficients for composite mitre bends Fig. 10--17: Loss coefficient K31 for sharp-edged 60
(Miller, 1990) branch junctions. A3 = A2 . (Miller, 1990)
153
1.0 . - - - - - - - - . - - r - - . - . - - , r - - 7 1
more detailed information is required, the reader is
referred to Miller (1990). 0.9
Figure 10-17 shows the side branch coefficient,

K31 , for a rather poor, sharp-edged junction. This is 0.8
produced by simply welding a branch pipe on to the
0.7
main pipe at the required angle. Q 1/Q 3 and A 1/A 3
are the ratios of the flows and areas of the branch
and inlet respectively. In certain areas of this graph,
the losses are not too bad, but if the design operat- ~ 0.5 f--+==-+-+:+--t+~-+:---:!f"-t-h:-:-+~<t--t-7t:-i
-.:
ing conditions are near the bottom of the graph, the
coefficients can be high.
Better are Y-junctions, shown in Figure 10-18.
The graph is based on test pieces with rounded inter-
nal comers, and standard joints will be somewhat
worse. Even so, Y-junctions are good for most flow
conditions. Note that the graph can also be used to
determine the loss in both branches, because the 0.4 0.5 0.6 0. 7 0.8 0.9 1.0
junction is symmetrical. Comparing the relevant Flow ratio, Q,IQ,
parts of Figure 10-18 and Figure 10-17 gives an in-
dication of the difference sharp edges make. Fig. 1Q-19: Loss coefficient K31 for tapered 45
junctions. A3 = A 1 + A2 (Miller, 1990)
150
~
.
135
-< .
0.40
_j_
120 3
v
0.35 2
a,
,~.
I I/ D
105 0.30
v !--- L
/ ~
v
0.25
90
1/ I 1\ 0.20
~v
v lL
~
"""
0 1.... /rv 1---- " 0.15
L
0 ,,,.
I
( '/;
0.10
IJ/
/ L
Uf
45 0.05
I\ I
\ ..... 0.10 [__...--Y/~
0.00 ~ ,...., L "'
30 "'
<:::>
"')~'>- ...........
15
T ~.05
0 0.1 0.2 0.3 0.4 0.5 0.6 0. 7 0.8 0.9

Flow RJollo Q,IQ,
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Fig. 1Q-20: Loss coefficient K32 for the straight-through
Flow ratio. Q,IQ, branch of an angled junction. (Miller, 1990)
Fig. 1Q-18: Loss coefficient K31 for symmetrical
Y-junctions. A3 = A 1 + A2 . (Miller, 1990)
Figure 10-19 shows K 31 for a good junction, that Eq. 10-93

has taper reductions into the two branches. Again,
the test pieces had radiused comers, which are time He - head loss over contraction (m)
consuming to produce in a workshop. Real joints V2 - water velocity at the outlet (m/s)
are more likely to have sharp inside edges, and will
not be so good. It can be seen that contraction losses are not too
Finally, Figure 10-20 shows the loss coefficients, significant, even for sharp, stepped reductions in
K 3 z, for the straight-through paths of bifurcations like diameter. Note that when the reduction is part of a
those in Figure 10-17 and Figure 10-19. K 32 is little bifurcation, as in Figure 10-20, the contraction loss
affected by the geometry or the flow ratio, and this is included as part of the overall junction head loss,
graph is valid for most branches off a straight pipe. and does not need to be added in again.
10.7.4 Diameter transition losses 10.7.5 Worked example

Figure 10-21 shows the loss coefficient for contrac-
tions in a pipe. Note that the head loss is calculated Problem
from the water velocity in the smaller pipe, the Consider the welded steel manifold in Figure 10-
outlet of the contraction: 22. The water has come down a 0100mm bore
154
4.0
~ vv [/
/ _,..v Similarly, when both jets are open, the velocity in
1// :;; v l/ v
v
3.0 /
the penstock is 2.55m/s, and after the bifurcation,
~.3 the velocity of 10 1/s in the 0100mm pipe is 1.27m/s.
[/
2.0 / v )....- The losses in the pipes comes from Table 10--3
~ 1.8 _L v / / v and Equation 10--90. For the 070mm pipe, using k
oy v
../
/ ... )./ v = 0.02 mm:

"""0
.E
1.6
v/ v v / v
o.oy V.d 260X007
/
I/
"' 1.4 l/
~ ~K=0.05 l/0.03 Re = ---;- = 1 57 X 10-6 = 115,900
l I I
\
v,:l
)r
1.2
-----~' J=-------
1.0
, -.1>
Ld \ \ .'1 4.[log(3.~d + ~~:~
0 0.1 0.2 0.3 0.4 0.5 2 3 4 5
Contraction length ratio, 1/d, 1
=--------------------------2
Fig. 10-21: Loss coefficient Kc for contractions. (Miller,
1990) 4. log 000002- + 574 ~~ )]
[ ( 3 7 x o-o7 [115,9oor
penstock, and enters the powerhouse through a 35

mitre bend. A 070mm pipe branches off to the = 0.0190
lower nozzle, and the main pipe contracts to
070mm for the upper nozzle. Each jet takes 101/s of
water. What are the head losses from the end of the Similarly, for the 0100mm pipe, Re = 162 400, f =
penstock to the nozzles when both jets are 0.0176 before the junction, Re = 80 890, f = 0.0197
operating? after.
These values can be used to calculate Kr values
Solution for each length of pipe. So, for example, in the
Each branch can take 10 litres/s (0.01m/s), so the 795mm long section of the 0100mm pipe:
velocity in the 070mm (0.07m) branches is L 0795
K = j - =0.0176 - - = 0 14
Q 0010 f d 01 .
V=-=(Jr )=260m/s
A - 007 2 Other values are given in Table 10--4.
4
795 229 55 200

~--------------------~~----~
Fig. 10-22: Manifold system
155
Table 10-4: Worked example results shows Kb for this type of bend. The bend has r =2d,
Component Loss Reference Head and is made of 6 x 15 segments, so it will be some-
where between the 4 x 22.5 and the radiused bend
coefficient velocity loss

K line, Kb =0.5. The 60 bend is not covered properly
(mls) (m)
by either Figure 10-16 or Figure 10-15. If it were a
radiused bend, then Figure 10-15 would give Kb as
MANIFOLD ROUTE TO TOP BRANCH: 0.25. For the top bend, it was estimated from Figure
Mitred bend, 35 ~ =0.37 2.55 0.12 10-16 that the compound mitred joint had a Kb
Straight, 0100 x 0. 795m f<t =0.14 2.55 0.05 about 0.1 higher than a similar radiused bend. The
Bifurcation K32 = 0.03 2.55 0.01
Straight 0100 x 0.229m f<t =0.05 1.27 0.00 losses incurred by 60 bends will be less, so we might
Contraction Kc = 0.05 2.60 0.02 add, say, 0.05 to the radiused Kb to get the effect of
Straight, 070 x 0.200m f<t = 0.05 2.60 0.02 the mitres, giving a total of Kb =0.25 + 0.05 =0.3.
Compound mitred bend, ~ =0.50 2.60 0.17
goo The coefficient for the contraction on the upper
Straight, 070 x 0.169m f<t = 0.05 2.60 0.02 line comes from Figure 10-21. A 1/A2 = 1002/702 =
Total 0.41m 2, and lld2 =55170 =0.79, giving Kb =0.05.
Lastly, consider the bifurcation. For the upper
MANIFOLD ROUTE TO BOTTOM BRANCH: branch, look at the straight-through loss graph, Fig-
Mitred bend, 35 ~ =0.37 2.55 0.12 ure 10-17. For Q 1/Q 3 =10/20 =0.5, K32 =0.03. For
Straight, 0100 x 0.795m f<t =0.14 2.55 0.05
the upper branch, consult Figure 10-17, Q 1/Q 3 =
Bifurcation K3 1 =0.77 2.55 0.26
Straight 070 x 0.566m f<t = 0.15 2.60 0.05 10/20 = 0.5, A 1/A 3 =702/1002 = 0.5, giving K31 =
Compound mitred bend, ~ = 0.30 2.60 0.10 0.77.
60
Straight, 070 x 0.169m f<t = 0.05 2.60 0.02
All these results shown in Table 10-4, together
with the reference velocities used to calculate the
Total 0.60m
velocity head for the coefficient. To illustrate this,
the equation below shows the calculation for the
35 mitre bend.
Next, consider the bends. To find the loss in the
35 bend, read Kb from Figure 10-15; Kb = 0.37.
The bends on the branches are compound mitres.
The upper branch bend is 90, and Figure 10-16
'
---~---
'
I'
i
.
--1---v--1.
! ~ '
--
4 3 2 0
e,~e
x, X
Fig. 1Q--23: Forces acting on a turbine housing
156
10.8 Baseframe anchor bolt loads that the turbine is rigid (i.e. there is no deflection or
bending in it), the amount if stretch ei at a point xi
This section derives and explains the background
from the pivot line will be proportional to the dis-
of the formulae used in Section 4.13.2.
tance xi. The extra force induced in the bolts is

In Figure 10-23, a turbine and manifold is
proportional to the stretch, and can be written with
considered as a free body. The manifold is joined to
a constant 'k' as:
the penstock by an expansion joint. If the friction in
the joint is ignored, the only force acting along the
penstock is the pressure force, P, equal to the Eq. 10-94
pressure multiplied by the pipe cross-sectional
area. The turbine is bolted to the floor of the The force P has a moment about the pivot line of M
powerhouse with anchor bolts around the = P x a. This must be equal to the sum of the
base frame. moments due to the bolt forces, and this allows us
The force P is taken by the bolts in shear. As- to find the value of each bolt force.
suming that the housing and frame are completely
rigid, if there are N bolts, each bolt will take a shear
force of PIN.
Working out the tensile force in the bolts is no
easy matter. The force P induces a moment on the
turbine, and this will tend to pivot it around its right
hand edge, pulling upwards on the anchor bolts. Eq.J0-95
How the load is distributed in the bolts depends on
the relative stiffness of different parts of the casing. X; - horizontal distance from pivot
Also, when an external load is applied to a preten- n; - number of bolts at position x;
sioned bolt, not all of the force goes into the bolt. F; - force in each bolt at position X;
The amount that the bolt carries depends on the
relative stiffnesses of the bolt and the assembly it is An example of this calculation is given in Section
clamping. The analysis in this text assumes a 4.13.2, which also contains a discussion on how to
completely rigid turbine frame, that all the load adjust the result for a real turbine situation.
goes into the bolts, and that the bolts are all the Note that this formula works if the penstock, and
same. Because these assumptions are never true, force P, is at an angle instead of horizontal. Only
the results calculated are indicative only. With a bit the moment induced by P needs to be considered.
of experience, the figures can be modified to A downwards vertical components of P goes dir-
estimate the real loads, and allow the size of the ectly into the floor, not through the bolts. An up-
anchor bolts to be calculated. A more detailed wards vertical component only affects the
analysis is given in Shigley & Mischke (1989). calculations if a gap appears between the floor and
Assume that the turbine pivots around the line of the turbine - which should never happen (if it did,
the centres of the bolts on the right side. The bolts the bolting arrangement would be shown to be to-
will stretch slightly, and because of the assumption tally inadequate).
157
11
APPENDIX: FULL SYSTEM CALCULATION

SPREADSHEET
The calculations required to choose a Pelton tur- Table 11-1 except for the values in the penstock
bine are quite laborious if done manually, but are thickness calculations. Test the spreadsheet with both
easily handled on a personal computer. Once a values of z and t_close (z = 65%, t_close = 0 or
spreadsheet has been set up, numerous options can z = 100%, t....:.close = 0.8s) to check it.
be tested very quickly. This section shows a The formulae used are the same as those used in
spreadsheet set up for Microsoft Excel, though it other places in this book, with the exception of the
can easily be adapted for other spreadsheet slow valve closure surge formula; this latter is taken
software. from lnversin (1986), though it is easily derived and
Consider the process of starting a Pelton turbine. can be found in most standard hydraulics text-
Initially, when the nozzle valves are shut, the press- books. (Note that there is a slight error in lnversin,
ure at the turbine is the full static head. When a in that the formula omits the term for the gross
valve is first opened, this head accelerates the water head.) The basic turbine equations are given in Sec-
in the penstock, and pushes it out of the nozzle as a tion 3.1, the penstock loss is calculated using the
jet. As the speed of the water in the penstock in- same equations as for manifold pipe losses given in
creases, so do the losses. This has the effect of re- Section 10.7.1, and the thickness comes from the
ducing the head at the nozzle, and consequently equations in Section 4.10.3.
reducing the flow in the jet. There is an equilibrium
flow at which the losses in the penstock give a head
at the nozzle which gives exactly that flow in the jet. 11.1 Using the spreadsheet
This is the operating flow, and it is this flow that the The procedure for using the spreadsheet is to fill
spreadsheet is calculating. in the inputs on the left hand side, and then check
The spreadsheet (Table 11-1) is divided into in- the outputs on the right hand side. The
puts - on the left hand side, and outputs - on the spreadsheet has three parts which can be solved
right. The outputs are calculated from the inputs, separately. It is easiest to sort out Part 1,
and the formulae used are shown in Table 11-3. All 'Penstock losses, Turbine jet size and speed
the input and output value cells in the spreadsheet calculations' first, then move on to Part 2, 'Power
have been given names (shown in the 'Symbol' col- calculations', and finally do Part 3, 'Penstock
umns) so that these range names can be used in the thickness calculations'. The bottom part,
formulae. It would be possible to use the cell ad- 'Constants', needs to be filled in for all
dresses instead, but using the symbol names makes calculations.
the formulae much easier to read and understand.
The formulae are correct for the units given in 11.1.1 Part 1, Penstock losses, turbine jet size
Tables 11-1 and 11-2. Remember that if the units and speed calculations
for any value are changed, all the formulae that The 'Constants' g K_w ro and nu can be filled in
I I
refer to it must be adjusted. with the values given in the example. Provided
When first setting up this spreadsheet on a fresh water is being used, and the scheme is located
computer, it is recommended that the values on earth, these values should suffice. In Part 1, the
given in the sample spreadsheets are tried out to values for flow Q, gross head, H_g, and penstock
check that everything is correct. Two different length L, come from site survey data, and can be
examples are given because the penstock thick- filled in straight away. Values for the penstock ma-
ness calculations have an 'IF' statement, and two terial roughness coefficient k, depend on the type
different sets of equations are used depending on of penstock. Values of k are given in in Table 10-3
whether the closure time is greater than, or less in Section 10.7.1. A figure of 0.1 is used here as a
than, the critical time. The full spreadsheet general value for rough-painted fabricated mild
shown in Table 11-1 has values which test for steel pipes.
instantaneous flow blockage, the part The inlet and penstock minor loss coefficients
spreadsheet in Table 11-2 tests for slow valve K_m, are the sum of the loss coefficients for all the
closure. Table 11.2 uses the same variables as bends, valves, and the inlet. If the penstock has an
158
Table 11-1: Spreadsheet values for instantaneous valve closure
A B c D E F G H
1 PELTON TURBINE FULL SYSTEM CALCULATION SPREADSHEET

2 INPUT DATA Jl CALCULATED OUTPUTS
3 SITE NAME Sample
4 1. Penstock Losses, Turbine Jet Size & Speed Calculations
5 Quantity Symbol Unit Quantity Symbol Unit
Velocity in
6 Flow Q 32 1/s penstock v 1.811 m/s
Friction head loss
7 Gross head H g 60 m in penstock H f 2.676 m
Minor head losses
8 Penstock length L 120 m in penstock H c 0.502 m
Net head at end of
9 Penstock diameter d 150 mm penstock H n 56.82 m
Penstock Penstock
10 roughness k 0.1 mm efficiency e_pen 94.7 o;o
coefficient
Inlet & penstock Nett head at
11 minor coefficient K_m 3 turbine H_tur 55.69 m
12 Manifold efficiency e_man 98 % Jet velocity vjet 32.05 m/s
13 Pelton PCD PCD 250 mm Jet size (water) djet 25.21 mm
Max. allowable jet
14 Number of jets njet 2 size d_max 27.5 mm
Turbine optimum
15 Max. jet dia./PCD beta 11 % speed N 1126 rpm
Nozzle velocity Turbine runaway
16 coefficient c_v 0.97 speed N_r 202"1 rpm
17
18 2. Power Calculations
Total system
19 Drive efficiency e drv 95 o;o efficiency e_tot 56.41 %
Generator Turbine mechanical
20 efficiency e gen 85 o;o power P_tur 13.14 kW
21 Turbine efficiency e_tur 80 o;o Electrical power P_elec 10.61 kW
22
23 3. Penstock Thickness Calculations

Penstock thickness Wave velocity in
24 t 3 mm penstock V_wave 1183 m/s
Penstock critical
25 % of flow stopped z 63 % time T _crit 0.203 s
26 Valve closure time T_close 0 s Surge head H_sur 137.7 m
Corrosion Total head at
27 allowance t_cor 1.5 mm surge H_tot 197.7 m
28 Penstock UTS s_ult 410 N/mm2
Penstock Young's Required penstock
29 Modulus E_p 210 kN/mm2 thickness t_req 2.904 mm
30 Overall safety factor SF _tot 3.96
31
32 Constants
33 Gravity g 9.8 m/s
Bulk density of
34 water Kw 2.1 kN/mm2
35 Density of water ro 1000 kg/m3
Kinematic viscosity
36 of water (5C) nu 1.57 eSt
159
Table 11-2 Part Spreadsheet for slow valve closure
A B c D E F G H
3. Penstock Thickness Calculations 23
24 Wave velocity in
Penstock thickness t 2.2 mm penstock V_wave 1117 m/s
% of flow stopped Penstock critical

25 z 100 % time T _crit 0.215 s
26 Valve closure time T_close 0.8 s Surge head H_sur 34.85 m
Corrosion Total head at

27 allowance t_cor 1.5 mm surge H_tot 94.85 m
28 Penstock UTS s_ult 410 N/mm2
Penstock Young's Required penstock

29 modulus E_p 210 kN/mm2 thickness t_req 2.173 mm
Overall safety
30 factor SF_tot 3.96
31
Table 11-3: Formulae used in Full System Calculation Spreadsheet
Cell Cell name Formula
$G$6 v =4*{Q/1000)/(PI()*(d/1000)A2)
$G$7 H_f =VA2/(2*g)*L/(d/1000)*1/(4*(LOG10((k/1000)/(3.7*d/1000)
+5.74/(V*(d/1000)/(nu/1000000)A0.9))A2)
$G$8 H_c =K_m*VA2/ (2*g)
$G$9 H_n =H_g-H_f-H_c
$G$10 e_pen =H_n/H_g*100
$G$11 H_tur =H_n*(e_man/100)
$G$12 v_jet =c_v*SQRT(2*g*H_tur)
$G$13 d_jet =SQRT(4*(Q/1000)/{n_jet*PI()*v_jet))*1000
$G$14 d_max =PCD*beta/100
$G$15 N =0.46*v_jet*60/(PI()*PCD/1000)
$G$16 N_r =N*l . 8
$G$19 e_tot =e_pen/100*e_man/100*c_vA2*e_tur/100*e_drv/100*e_gen/100*100
$G$20 P_tur =H_tur*ro*g*Q/1000*c_vA2*e_tur/100*1/1000
$G$21 P_elec =H_g*ro*g*Q/1000*e_tot/100*1/1000
$G$24 V_wave =1/SQRT (ro* ( 1/(K_w*1000000000)
+(d/1000)/(E_p*1000000000*t/1000)))
$G$25 T_crit =2*L/V_wave
$G$26 H_sur =IF(T_close<T_crit,V_wave*z/100*Q/1000/(PI()4*(d/1000)A2*g),
+H_g*( (L*z/100*V/(g*H_g*T_close) )A2/2
+SQRT((L*z/100*V/(g*H_g*T_close))A2
+(L*z/100*V/(g*H_g*T_close))A4/4)))
$G$27 H_tot =H_g+H_sur
$G$29 t_req =H_tot*ro*g*(d/1000)/(2*s_ult*1000000/SF_tot)*1000+t_cor
inlet loss coefficient of Ki, and three bends with detailed values of Kb can be looked up in Figure
coefficients of Kb 1 , Kb 2 and Kb 3 , then 10-15 or Figure 10-16. Inlet losses are shown in
Figure 11-1. Generally speaking, the minor losses
K_m = Ki + Kb 1 + Kb 2 + Kb 3 . do not greatly affect the performance, and a rough
As a rule-of-thumb for initial calculations, a value guess for K_m will be sufficient for most
of Kb = 0.5 can be used for each bend. More calculations.
160
~ ~ F: ~
'l
'l ~ 'l ~ ~
'l 'l
'l ~ 'l ~ ~
'l 'l
'Q 'l 'l 'l ... 'l 'l 'l 'l ... 'l 'l 'l '
Ki 1.0 0.8 0.5 0.2 0.04
hooded/ projecting flush, slightly bell-mouth

re-entrant sharp-cornered rounded
Fig. 11-1: Loss coefficients for various penstock inlets (after lnversin, 1986; and Harvey et al, 1993)
The manifold efficiency will typically be 93-97%. changes in diameter, though the latter requires a
A straight forward, smooth reduction for a single good understanding of the equations. Similarly, it is
jet will have high efficiency, a complicated manifold possible to include the manifold as a series of pipe
with many branches will be somewhat lower. pieces plus bend and junction loss coefficients, but
The nozzle coefficient c_v, depends on the type of this is quite complicated to do, and changes for each
nozzle being used. A value of 0.97 is acceptable for manifold layout. Losses in the manifold are usually
preliminary calculations. Better values can be ob- not too significant, and making allowances for them
tained from Section 2.2.1 or Section 2.2.2. as an efficiency is a reasonable approximation.
Beta is the ratio of the maximum jet diameter to
the PCD, and is determined by the bucket design. 11.1.2 Part 2, Power calculations
The bucket used in this manual is designed for a This part of the spreadsheet processes the informa-
beta of 11% nominally, and can be pushed to 12%. tion from Part 1. The input values are the drive
See the discussion in Section 3.1.4 for more detail. efficiency e_drv, the generator efficiency e_gen,
Having put in the basic data, it is now time to and the turbine efficiency e_tur.
make some guesses for the penstock diameter d, The drive efficiency will be 100% for direct-
the PCD, and the number of jets n_j et. For those coupled arrangements, and perhaps 94-97% for
not familiar with Pelton calculations, it is advisable belt drives. The generator efficiency should be
to go through the selection procedure in Chapter 3, available from generator manufacturers. Figures of
and use the values obtained from there in this 80-85% are reasonable for small induction motors,
section. The alternative approach is to start with a up to 90% for larger (50kW+) synchronous
large value of d (say 20-30% more than would be generators.
expected, and the adjust the PCD until approx- The turbine efficiency depends on the size and
imately the required speed, N, is achieved. Next, the accuracy of manufacture. A general figure of
adjust the values of n_j et until d_j et is less than 75-80% is reasonable for micro-hydro, though bet-
the maximum allowable jet size d_max. If this gives ter efficiencies may be obtained for large units
too many jets, the PCD will have to be increased, (50kW+), worse for very small ones (<5kW).
and the speed will have to be accepted. Next, re- Once these efficiency figures are put in, the
duce d until the penstock efficiency e_pen, comes spreadsheet returns the power outputs, which can
to about 95%. Finally, refine all the values until a be checked against the expected values.
reasonable looking design is achieved.
The spreadsheet is set to determine the value of 11.1.3 Part 3, Penstock thickness calculations
d_j et for a given value of Q. Once a turbine has Part three calculates the required thickness of the
been designed, it is often useful to calculate the penstock for surge induced by the turbine. It is not
other way round. For example, a 2-jet turbine may necessary to use this part of the spreadsheet unless
be designed for a maximum flow, but one may wish one is choosing a penstock thickness.
to calculate the flow and efficiency with only one The Young's Modulus E_p, and Ultimate Tensile
jet. This can be done using the 'goal-seek' option in Strength s_ul t, for the penstock material can be
Excel, giving a value for jet diameter and changing put in as soon as the material is chosen. Typical
Q to achieve this. Goal-seeking can obviously be values for the most common materials are given in
used to try other options too, though one has to Table 11-4.
choose variables that are related through the The corrosion allowance t_cor, is the amount of
equations. the penstock wall that can be lost to corrosion over
Note that the spreadsheet assumes a single the design life of the penstock -which might be 15
penstock pipe of the same diameter for its whole or 20 years for micro-hydro plants (large schemes
length. It can be modified for multiple pipes or for work on 50 years for the civil works). The corrosion
161
Table 11-4: Typical properties of common penstock The safety factor used in the spreadsheet SF_tot,
materials is the product of the basic safety factor and the
hand-welding factor. The basic safety factor should
Young's Ultimate Tensile Density

Modulus Strength (kg/m 3 ) be applied to any type of penstock. The hand-
E_p s_ult welding factor is only used for steel pipes if the
(kN/mm2) (N/mm2) longitudinal welds are hand-welded. A penstock
Mild steel 210 350 7850 made from rolled mild steel plate in a local work-
HOPE 0.2--0.8 6--9 900 shop for a 15kW scheme would have a safety factor
uPVC 2.8 28 1400 SF _tot = 3.5 x 1.2 = 4.2. The same scheme using
bought-in, machine welded mild steel tube would
use SF_tot = 3.5. A 5kW scheme using HDPE
allowance depends on the corrosion protection ap-
pipe would use SF_tot =3.0.
plied to the penstock. Typical values are given in
Table 4-7. For plastic penstocks, no corrosion al- Table 11-5: Penstock thickness safety factors (Wal-
lowance is necessary. tham, 1994)
The valve closure time T_close, and the per-
centage of the flow stopped z, are discussed in Sec- Size of scheme Very Small Medium Large
small (8- (20- (>50kW)
tion 4.3.5. If the flow is blocked instantaneously, (<8kW) 20kW) 50kW)
say by the spear valve end falling into the nozzle,
T_close is zero. If it can be guaranteed that such a Basic safety factor 3.0 3.5 4.0 4.0
failure can never occur, then the most rapid poss- Hand-welding 1.2 1.2 1.1 1.1
factor
ible valve closure time can be put in. The amount of
flow stopped depends primarily on how many jets
there are: for a 4-jet Pelton, the value of z is likely Once the values of z, T_close, t_cor 1
to be 25%, but see Section 4.3.5 for more detail. s_ul t E_p and SF_tot have been put in, a trial
I
The spreadsheet calculates the stress in the value of thickness, t, can be inserted. The
penstock from the expected surge or water hammer spreadsheet will calculate the required thickness
head. The allowable head depends on the safety t_req, and t should be adjusted until it is a stand-
factor used. Choosing a safety factor is something ard thickness that is greater than t_req. Note that
of an art, relying mainly on experience. Safety fac- the penstock thickness calculations are done for the
tors for penstocks tend to be large, as there is usu- bottom of the penstock only. Thickness can be re-
ally quite a degree of uncertainty about the loads. duced because of the lower static head higher up
Suggested safety factors are given in Table 11-5. the penstock.
162
12
APPENDIX: 0-RING SEALS

0-rings are cheap, reliable seals that are very useful relative to the 0-ring. A flange-type seal is assumed
for micro-hydro. They form good, leak-free pipe to be static, but a rod seal can be static or dynamic.
flange seals, and can also be used to seal spear The seal on the shaft of a spear valve is dynamic,
valves. A standard rubber 0-ring will readily seal because the shaft moves, and the 0-ring rubs along
up to a lOOOm head. To perform well and give good the sealing surface. If a governor is used to control
life, 0-rings need to be installed correctly. This the spear, the application is truly dynamic, because
means choosing the correct size of 0-ring and the spear valve can be moving continuously. If the
groove, and making sure the machining is accurate. spear is controlled manually, the application is
semi-dynamic, and slightly more squeeze can be
used if problems are experienced with leakage.
12.1 Sealing arrangements
The most common ways of using 0-rings are shown
in Figure 12-1. The first diagram shows a flange 12.2 0-ring section and groove details
seal, and the other two show internal and external Check which 0-ring cross-sections (dimension 'a')
rod seals. First decide which type of sealing method are available, and see which can be fitted. Within
is to be used. Static sealing is where the sealing reason, use the largest section possible. In order to
surfaces do not move relative to the 0-ring. Dy- maintain a pressure inside the rubber to achieve
namic sealing is where one of the surfaces moves sealing, the 0-ring has to be squeezed in the
a 0A
- - - --+--- - -
0-ring
(b) Rod sealing, external
(a) Flange sealing
R
=0.5 for a < 2.5 mm
= 1.0 for a >= 2.5 mm
Groove Detail
(c) Rod sealing, internal
Fig. 12-1: Methods of installing 0-rings
163
groove. This means that the dimension F must be larger than 0.5mm for 0-rings up to 2.5mm section,
less than the 0-ring section, a. The amount of and l.Omm for larger sizes.
squeeze depends on the 0-ring section, and the The surface finish depends on the application.
application. Small 0-rings need more squeeze to Published recommended values are shown in Table
make sure they seal. Static seals can tolerate a high 12-1. Machining needs to be done carefully to
amount of squeeze, but dynamic applications use achieve these values. A dynamic rod surface should
less squeeze to reduce the friction and wear on the ideally be ground, though for a manually adjusted
0-ring. If possible, followed the suppliers recom- spear valve, a turned finish is acceptable. Also,
mendations for selecting the size. If no recom- there is no point having a perfect surface finish if
mended groove sizes are available, then determine the surface is going to corrode in service. If the
the amount of squeeze from Figure 12-2. This surface is exposed to water, the material needs to
graph is plotted from manufacturers recommended be corrosion resistant, ideally stainless steel.
groove sizes for a number of different sizes (Dowty,
1978). Table 12-1: 0-ring housing surface finish for dif-
ferent applications
Static Dynamic
Eq. 12-1
Ra (~m)
The next stage is to determine the width of Housing 1.6 0.8

groove required, W. Surprisingly, perhaps, rubber Mating surface (rod) 0.8 0.4
is incompressible. If it is squeezed in one direction,
it needs space to expand in another. Rubber also
For rod seals, the gap between the rod and the
has a relatively high coefficient of volumetric ex-
housing must be checked. If this gap is too big the
pansion with temperature. The groove needs to
0-ring may be extruded into it. Table 12-2 gives
have space both for the 0-ring to be squeezed side-
recommendations for the maximum values of X.
ways and for thermal expansion. The maximum
width for static applications is not too critical, but Table 12-2: Clearance between shaft and housing
for dynamic applications the width needs to be kept
to a minimum, to stop the 0-ring shuffling about in 0-ring section Xmax
the groove and wearing. The recommended width (mm) (mm)
of the groove can be read off from Figure 12-3 - if Over Incl.
better information is not available. The graph gives
values of Wla for various values of a. 2.50 0.12
2.50 3.00 0.15
The groove has now been sized, but the detail of 3.00 4.00 0.16
the groove is important too. The corners must not 4.00 5.00 0.17
have too large a radius, R, or it will reduce the 5.00 8.00 0.18
8.00 10.00 0.20
volume available for the 0-ring. R should be no
1 .
35%r----------------,----------------~------~
30% .. ;
--------,---- - --- --~-- - - -- -- __.__ ___ _ _ '--------l
I
25%t---~-----;~--+i
1
+----~
A + ___________
l 20% f .. - --- -- - J _________

l
r11Fia~98 ------
= A I Rod, Dynamic
g~
15% ------ -------- ~ ---.---;- -t----
I .. - .___--- -------- !A Rod, Static

10% --------------r-----------~.---~
5% . . . J__ ___ _
I
I
~k ~---------------r----------------~----~
0.0 2.0 4.0 6.0 8.0 10.0
0-rlng cross-section (mm)
Fig. 12-2: Squeeze required in groove for different 0-ring sealing arrangements (Dowty, 1978)
164
1.ro ~--~~------------------------------~
'
1.50 --
~ 1:
1.40

'

L
1
;; ..2
'i 1i ...

Flange
I
8r
II 11 1 30 +--------------~------- Rod, Dynamic
Rod, Static I
~6
1.20
1.10
1.00 L _____________________ ~
0.0 2.0 4.0 6.0 8.0 10.0

Ortng cross-section (mm)
Fig. 12-3: Groove width/section for different 0-ring sealing arrangements
12.3 0-ring diameter When designing 0-ring housings, the effect of tol-
erancing on the 0-ring should be considered. The
The final parameter to choose is the diameter of
need to have reasonable machining tolerances on
the 0-ring itself. For a flange seal (with internal
the various components may make it impossible to
pressure, as assumed here) the 0-ring outside di-
keep the 0-ring within the above limits, especially
ameter needs to be the same size or slightly larger
for small 0-rings. If this is the case, the maximum
than the housing diameter B. For external rod seal-
compression or stretch can be increased to 8~/o ~o
ing, the 0-ring outside diameter should be the
ensure that some compression or stretch remams m
same size or slightly larger than the groove outside
the worst tolerance condition.
diameter. For internal rod sealing, the 0-ring inter-
nal diameter should be the same size or slightly
smaller than the groove internal diameter. These 12.4 Material
relationships are expressed in Equation 12-2, and
For use in water, oil or grease, most standard stock,
W should be chosen so that it satisfies the appropri-
nitrile-rubber 0-rings will work. The recom-
ate equation.
mended rubber for use with water has a nominal
hardness of 70 IHRD (International Rubber Hard-
Flange Seal: ness test Shore type -D), and this should be re-
2 quested for critical applications.
W+ a = 1 00 to 1 01 Large-diameter 0-rings for pipe flanges can be
B made from continuous lengths of material. Calcu-
late the length of rubber required as per the equa-
External Rod Seal: tions above, cut the ends at an angle, and glue
together with cyanoacrylate adhesive ('superglue ').
2 Smaller diameter 0-rings should be bought
W+ a = 1 00 to 0 97 moulded as one piece, without a join.
c
( J.()() to 092 in special circumstances - see below.)
12.5 Installation
Note in Figure 12-1 that 15 chamfers are included
Internal Rod Seal: on the rod or housing for the rod-type seals. This is
to ensure that the 0-rings are not damaged during
D = I 00 to 1 05 assembly. Care should be taken not to catch the
A
0-rings on sharp corners or threads during
( 100 to 108 in special circumstances - see below.)
assembly, and not to use sharp tools to put them in
Eq. 12-2 place.
165
13
APPENDIX: KEYS AND KEYWAYS

This appendix gives recommended dimensions for applications. The tolerance classes given for the
both parallel and tapered metric keyways. Figure key width in Table 13-1 (N9, J 5 9 and DlO) are ISO
13-1 shows the general layout, and Table 13-1 tolerances, which are discussed more fully in Ap-
gives the dimensions and tolerances for standard pendix 14 in the section on 'Limits and Fits'. For a
keyways for various shaft diameters from 012mm- more general discussion on the use of keys, see
110mm, which should cover most micro-hydro Section 4.8.3.
X Taper 1 in 100 X
-.,--
h*
xl X
Section X-X
Enlarged detail of Section X-X
5 X 45~1-
.t_p=====9
h L ========~ h D
~b~
I.
Fig. 13-1 : Parallel and taper metric keys and keyways to BS 4235
166
Table 13-1: Selected parallel and taper metric key and keyway dimensions to BS 42345: Part 1: 1972. Reproduced with permission
SHAFT KEY KEYWAY
Nom. dia. Size Width, b Depth Radius
d Width Height Chamfer Length To/. on To/. on To/. on Shaft, Parallel Taper
b h s I parallel parallel taper shaft t1 key hub t2 key hub t2
shaft hub &hub (parallel) (taper)
Over Incl. Nom To/ Nom To/ Min Max Min Max Nom (N9) (Js9) (010) Nom To/ Nom To/ Nom To/ Max Min
(mm) (mm) (mm) (JJ.m) (mm) (JJ.m) (mm) (mm) (mm) (mm) (mm) (JJ.m) (JJ.m) (JJ.m) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
12 17 5 0 ___2_ 0 0.25 0.40 10 56 5 0 +0.015 +0.078 3 +1.0 +0.1 1.7 +0.1/0 0.25 0.16
- - 2.3
....... 17 22 6 -o.030 6 -o.030 0.25 0.40 14 70 6 -0.030 -o.015 +0.030 3.5 0 2.8 *0 2.2 +0.2 0.25 0.16
0\
-..1
22 30 8 0 7 0 0.25 0.40 18 90 8 0 +0.018 +0.098 __i_ +0.2 ~ +0.2 ~ 0 0.25 0.16
-
30 38 10 -o.036 8 -o.090 0.40 0.60 22 110 10 -Q.036 -o.018 +0.040 5 0 0.40 0.25
12
- 3.3 0 - 2.4
38 44 12 0 8 0.40 0.60 28 140 0 +0.021 +0.120 ~ ~ 0.40 0.25
- ~
44 50 14 -o.043 ~ 0.40 0.60 36 160 14 -Q.043 -o.021 +0.050 5.5 3.8
f..--- ~ 0.40 0.25
-
50 58 16 10 0.40 0.60 45 180 16 ______ ~ ~ 0.40 0.25
58 65 18 11 0 0.40 0.60 50 200 18 0 +0.026 7 4.4 ~ 0.60 0.25
-
65 75 20 0 __g_ -Q.110 0.60 0.80 56 220 20 -Q.052 -o.026 +0.149 7.5 4.9 3.9 0.60 0.40
- - -
75 85 22 -o.052 0.60 0.80 63 250 22 +0.065 9 ~ ~ 0.60 0.40
- 14 -
85 95 25 0.60 0.80 70 280 25 9 5.4 4.4 0.60 0.40
~ - - -
95 110 28 16 0.60 0.80 80 320 28 10 6.4 5.4 0.60 0.40
(* +0.2/0 for taper keys)
14
APPENDIX: LIMITS AND FITS

When a bush has to be fitted in a housing, the size nominal diameter (i.e. the fundamental deviation is
and tolerance of both components affects how they zero). For a shaft-basis fit, the maximum size of the
go together. A heavy press fit can be produced, shaft is always equal to the nominal diameter
where the bush has to hammered or pressed into (again, the fundamental deviation is zero). In the
the housing, and it is almost impossible to get it out ISO system, hole-basis fits are preferred, but shaft-
again. Alternatively, a loose clearance fit means basis fits may be needed if a piece of equipment
that the bush can be easily slipped into the housing, comes with an h-deviation shaft on it.
and can fall out again. By putting the correct toler-
ances on component drawings, a designer can de-
termine exactly what sort of fit results. All too
14.1 Selected fits
often, lack of knowledge of how to do tolerancing Table 14-1 contains a list of recommended fits, with
stops people from attempting to design machinery. a brief description of the type of fit and the appro-
However, it is not magic, and by following a few priate places that they may be used for Pelton tur-
simple guidelines components can be made to fit bines or micro-hydro generally.
together as required. Having chosen a fit, the tolerances for hole-basis
Metric tolerances are usually bases on the ISO fits can be read from Table 14-1, and for shaft-basis
system of tolerances, deviations, limits and fits fits from Table 14-3.
(see, for example, BS 4500: 1969). Each compo-
nent has a tolerance band, which is the difference Example
between the largest and smallest allowable com- A phosphor bronze bush is to be used for the front
ponent sizes. The ISO system assigns different tol- bearing of a spear valve in a Pelton turbine. The
erance grades, which control the size of the bush is to be fitted into a mild steel housing, and
tolerance band. A fundamental deviation is then will be held in place with a circlip. The outside
assigned, which is the distance from the nominal diameter of the bush is a nominal 50mm.
size to the start of the tolerance band. The funda-
mental deviations are given capital letters for Solution
holes, A to ZC, and lower-case letters for shafts, a There is no need to use shaft-basis fits, so use the
to zc. The tolerance grades are given numbers hole-basis scheme. The bearing could be put in
from 1 to 16. A hole might be H7, and a shaft k6. place with an interference press fit (H7-p6), but
Giving a shaft as 030 mm, h6, completely specifies since it is retained by a circlip this is not necessary.
its size for manufacture. The forces on it are not high, and there has to be a
Suppose shaft needs to be put into a hole. The certain amount of clearance between the bearing
nominal size is chosen, and the hole is made to this and the shaft anyway, so any location fit will be
size with a tolerance of H7, the shaft with a toler- good enough. Use a H7-k6 fit, which is a transition
ance of g6. The shaft will fit into the hole with a fit which will normally require pushing in.
good, close sliding fit. This will be the case what- From Table 14-2, for a 50mm nominal diameter,
ever the nominal size, whether is it 30mm or the tolerances for the hole, H7, are +0.025/-0.000,
300mm. The tolerances H7 and g6 can be looked up so the housing should be made 50.025mm to
in tables for the nominal size chosen. 50.000mm. The k6 bush is +0.018/-0.002, so should
Tables 14-2 and 14-3 give the tolerances for be made 50.018mm to 49.998mm. The loosest fit
hole-basis and shaft-basis fits. For a hole-basis fit, that can result is 0.027mm clearance. The tightest
the minimum size of the hole is always equal to the fit is 0.018mm interference.
168
Table 14-1:
Hole basis Shaft basis

Hole- Shaft Shaft- Hole
H9- d10 h9- 010 Loose running fit.

H9-e9 h9- E9 Easy running fit.
Widely space bearings.
Spear valve front bearing.
H8-f7 h7- F8 Close running fit.
Easily produced running fit for bearings.
H7-g6 h6-G7 Sliding fit.
Not intended to run free, but a small clearance fit which can be moved
and turned, yet which locates accurately.
Precision fit for spigots and pins.
Taper locking sleeves on shafts.
H7-h6 h6-H7 Clearance location fit.
Minimum clearance is zero, but usually gives a small clearance. For
accurate location. A compromise between ease of assembly and accurate
location.
Flange spigots.
H7-k6 h6-K7 Transition location fit.
A transition fit that usually gives a small clearance. Not always easy to
assemble or disassemble.
Hubs and pulleys on keyed shafts.
Spear valve plain bearings into housing.
H7-n6 h6- N7 Transition location fit.
May give clearance but is usually a tight fit.
H7-p6 h6-P7 Interference location fit.
A true interference fit. Can be pressed into place, and should be possible
to pull apart.
Bushes and hubs that need to be well located.
Collars on spear shaft for bearings.
169
Table 14-2: Selected ISO fits- hole basis. Reproduced with permission from BS Data Sheet 4500A, 1970
Clearance Transition Interference
Loose running Easy running Close running Sliding fit Clearance Transition Transition Interference
fit fit fit location fit location fit location fit location fit
Nominal size Tolerance Tolerance Tolerance Tolerance Tolerance Tolerance Tolerance Tolerance Nominal size
Over Incl. H9 d10 H9 e9 HB f7 Hl g6 Hl h6 Hl k6 Hl n6 Hl p6 Over Incl.
(mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
- 3 +0.025 -Q.020 +0.025 -0.014 +0.014 -0.006 +0.010 -0.002 +0.010 0.000 +0.010 +0.006 +0.010 +0.010 +0.010 +0.012 - 3
0.000 -Q.060 0.000 -Q.039 0.000 -Q.016 0.000 -o.oo8 0.000 -0.006 0.000 0.000 0.000 +0.004 0.000 +0.006
3 6 +0.030 -Q.030 +0.030 -Q.020 +0.018 -Q.010 +0.012 -Q.004 +0.012 0.000 +0.012 +0.009 +0.012 +0.016 +0.012 +0.020 3 6
0.000 -Q.078 0.000 -o.o5o 0.000 -0.022 0.000 -0.012 0.000 -o.oo8 0.000 +0.001 0.000 +0.008 0.000 +0.012
6 10 +0.036 -Q.040 +0.036 -Q.025 +0.022 -0.013 +0.015 -o.oo5 +0.015 0.000 +0.015 +0.010 +0.015 +0.019 +0.015 +0.024 6 10
0.000 -Q.098 0.000 -0.061 0.000 -0.028 0.000 -0.014 0.000 -0.009 0.000 +0.001 0.000 +0.010 0.000 +0.015
10 18 +0.043 -0.050 +0.043 -0.032 +0.027 -O.Q16 +0.018 -0.006 +0.018 0.000 +0.018 +0.012 +0.018 +0.023 +0.018 +0.029 10 18
....... 0.000 -Q.120 0.000 -0.075 0.000 -Q.034 0.000 -0.017 0.000 -Q.011 0.000 +0.001 0.000 +0.012 0.000 +0.018
-....)
0 18 30 +0.021 +0.015 +0.021 +0.028 +0.021 +0.035 18 30
+0.052 -0.065 +0.052 -Q.040 +0.033 -Q.020 +0.021 -Q.007 +0.021 0.000
0.000 -Q.149 0.000 -Q.092 0.000 -Q.041 0.000 -Q.020 0.000 -O.Q13 0.000 +0.002 0.000 +0.015 0.000 +0.022
30 50 +0.062 -o.o8o +0.062 -o.o5o +0.039 -Q.025 +0.025 -Q.009 +0.025 0.000 +0.025 +0.018 +0.025 +0.033 +0.025 +0.042 30 50
0.000 -Q.180 0.000 -Q.112 0.000 -o.o5o 0.000 -0.025 0.000 -0.016 0.000 +0.002 0.000 +0.017 0.000 +0.026
50 80 +0.074 -Q.100 +0.074 -Q.060 +0.046 -Q.030 +0.030 -Q.010 +0.030 0.000 +0.030 +0.021 +0.030 +0.039 +0.030 +0.051 50 80
0.000 -Q.220 0.000 -0.134 0.000 -0.060 0.000 -Q.029 0.000 -Q.Q19 0.000 +0.002 0.000 +0.020 0.000 +0.032
80 120 +0.087 -Q.120 +0.087 -Q.072 +0.054 -Q.036 +0.035 -Q.012 +0.035 0.000 +0.035 +0.025 +0.035 +0.045 +0.035 +0.059 80 120
0.000 -0.260 0.000 -Q.159 0.000 -Q.071 0.000 -Q.034 0.000 -Q.022 0.000 +0.003 0.000 +0.023 0.000 +0.037
120 180 +0.100 -Q.145 +0.100 -Q.084 +0.063 -Q.043 +0.040 -Q.014 +0.040 0.000 +0.040 +0.028 +0.040 +0.052 +0.040 +0.068 120 180
0.000 -Q.305 0.000 -Q.185 0.000 -Q.083 0.000 -0.039 0.000 -Q.025 0.000 +0.003 0.000 +0.027 0.000 +0.043
180 250 +0.115 -Q.170 +0.115 -Q.100 +0.072 -o.o5o +0.046 -Q.015 +0.046 0.000 +0.046 +0.033 +0.046 +0.060 +0.046 +0.079 180 250
0.000 -Q.355 0.000 -Q.215 0.000 -Q.096 0.000 -Q.044 0.000 -Q.029 0.000 +0.004 0.000 +0.031 0.000 +0.050
250 315 +0.130 -Q.190 +0.130 -Q.110 +0.081 -Q.056 +0.052 -Q.017 +0.052 0.000 +0.052 +0.036 +0.052 +0.066 +0.052 +0.088 250 315
0.000 -Q.400 0.000 -Q.240 0.000 -Q.108 0.000 -Q.049 0.000 -Q.032 0.000 +0.004 0.000 +0.034 0.000 +0.056
315 400 +0.140 -Q.210 +0.140 -Q.125 +0.089 -Q.062 +0.057 -Q.018 +0.057 0.000 +0.057 +0.040 +0.057 +0.073 +0.057 +0.098 315 400
0.000 -0.440 0.000 -Q.265 0.000 -Q.119 0.000 -Q.054 0.000 -Q.036 0.000 +0.004 0.000 +0.037 0.000 +0.062
400 500 +0.155 -Q.230 +0.155 -Q.135 +0.097 -Q.068 +0.063 -Q.020 +0.063 0.000 +0.063 +0.045 +0.063 +0.080 +0.063 +0.108 400 500
---
0.000 -Q.480 0.000 -Q.290 0.000 -Q.131 0.000 -Q.060 0.000 -Q.040 0.000 ~.005 0.000 +0.040 0.000 +0.068
Table 14-3: Selected ISO fits - shaft basis. Reproduced with permission from BS Data Sheet 45008, 1970
Clearance Transition Interference
Loose running Easy running Close running Sliding fit Clearance Transition Transition Interference
fit fit fit location fit location fit location fit location fit
Nominal size Tolerance Tolerance Tolerance Tolerance Tolerance Tolerance Tolerance Tolerance Nominal size
Over Incl. h9 010 h9 E9 hl FB h6 Gl h6 Hl h6 Kl h6 Nl h6 Pl Over Incl.
(mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
- 3 0.000 +0.060 0.000 +0.039 0.000 +0.020 0.000 +0.012 0.000 +0.010 0.000 0.000 0.000 -{).004 0.000 -{).006 - 3
-{).025 +0.020 -{).025 +0.014 -{).010 +0.006 -{).006 +0.002 -Q.006 0.000 -{).006 -{).010 -{).006 -{).014 -{).006 -{).016
3 6 0.000 +0.078 0.000 +0.050 0.000 +0.028 0.000 +0.016 0.000 +0.012 0.000 +0.003 0.000 -{).004 0.000 -o.oo8 3 6
-{).030 +0.030 -{).030 +0.020 -{).012 +0.010 -{).008 +0.004 -0.008 0.000 -0.008 -{).009 -{).008 -{).016 -{).008 -{).020
6 10 0.000 +0.098 0.000 +0.061 0.000 +0.035 0.000 +0.020 0.000 +0.015 0.000 +0.005 0.000 -{).004 0.000 -{).009 6 10
-{).036 +0.040 -{).036 +0.025 -{).015 +0.013 -{).009 +0.005 -{).009 0.000 -{).009 -{).010 -{).009 -{).019 -{).009 -{).024
10 18 0.000 +0.120 0.000 +0.075 0.000 +0.043 0.000 +0.024 0.000 +0.018 0.000 +0.006 0.000 -{).005 0.000 -{).011 10 18
...... -{).043 +0.050 -{).043 +0.032 -o.018 +0.016 -o.011 +0.006 -{).011 0.000 -{).011 -{).012 -{).011 -{).023 -0.011 -{).029
-....!
.......
18 30 0.000 +0.149 0.000 +0.092 0.000 +0.053 0.000 +0.028 0.000 +0.021 0.000 +0.006 0.000 -{).007 0.000 -{).014 18 30
-{).052 +0.065 -{).052 +0.040 -{).021 +0.020 -{).013 +0.007 -Q.013 0.000 -{).013 -{).015 -{).013 -0.028 -{).013 -{).035
30 50 0.000 +0.180 0.000 +0.112 0.000 +0.064 0.000 +0.034 0.000 +0.025 0.000 +0.007 0.000 -{).008 0.000 -{).017 30 50
-{).062 +0.080 -{).062 +0.050 -{).025 +0.025 -{).016 +0.009 -{).016 0.000 -{).016 -{).018 -0.016 -0.033 -{).016 -{).042
50 80 0.000 +0.220 0.000 +0.134 0.000 +0.076 0.000 +0.040 0.000 +0.030 0.000 +0.009 0.000 -{).009 0.000 -{).021 50 80
-{).074 +0.100 -{).074 +0.060 -{).030 +0.030 -o.019 +0.010 -0.019 0.000 -{).019 -{).021 -{).019 -{).039 -{).019 -{).051
80 120 0.000 +0.260 0.000 +0.159 0.000 +0.090 0.000 +0.047 0.000 +0.035 0.000 +0.010 0.000 -{).010 0.000 -{).024 80 120
-{).087 +0.120 -{).087 +0.072 -{).035 +0.036 -{).022 +0.012 -{).022 0.000 -{).022 -{).025 -{).022 -{).045 -{).022 -{).059
120 180 0.000 +0.305 0.000 +0.185 0.000 +0.106 0.000 +0.054 0.000 +0.040 0.000 +0.012 0.000 -{).Q12 0.000 -{).028 120 180
-{).100 +0.145 -{).100 +0.085 -{).040 +0.043 -{).025 +0.014 -{).025 0.000 -{).025 -{).028 -{).025 -{),052 -{).025 -{).068
180 250 0.000 +0.355 0.000 +0.215 0.000 +0.122 0.000 +0.061 0.000 +0.046 0.000 +0.013 0.000 -{).014 0.000 -0.033 180 250
-{).115 +0.170 -{).115 +0.100 -{).046 +0.050 -{).029 +0.015 -{).029 0.000 -{).029 -{).033 -{).029 -{).060 -{).029 -{).079
250 315 0.000 +0.400 0.000 +0.240 0.000 +0.137 0.000 +0.062 0.000 +0.052 0.000 +0.016 0.000 -{).014 0.000 -{).036 250 315
-{).130 +0.190 -{).130 +0.110 -{).052 +0.056 -{).032 +0.017 -0.032 0.000 -0.032 -0.036 -0.032 -{).066 -{).032 -{).088
315 400 0.000 +0.440 0.000 +0.265 0.000 +0.151 0.000 +0.075 0.000 +0.057 0.000 +0.017 0.000 -{).Q16 0.000 -{).041 315 400
-{).140 +0.210 -{).140 +0.125 -{).057 +0.062 -{).036 +0.018 -{).036 0.000 -{).036 -{).040 -{).036 -{).073 -0.036 -{).098
400 500 0.000 +0.480 0.000 +0.290 0.000 +0.165 0.000 +0.083 0.000 +0.063 0.000 +0.018 0.000 -{).017 0.000 -{).045 400 500
-{).155 +0.230 -{).155 +0.135 -{).063 +0.068 -{).040 _Q,O~O L_-{).040 0.000 -{).040 -{).045 -{).040 -o.o8o -{).040 -{).10_ ~-
15
APPENDIX: SPECIFIC SPEEDS

15.1 Definition of specific speed defined as 75kgf.m/s =735.5W. The following form
of the equation is required to use kW for power:
It is quite possible to select, design and install tur-
bines without ever considering specific speed. It is a
slightly awkward concept to grasp, and is mainly of N . = l-166NJP
s-metr1c H t
use for large hydropower stations where it is im- n
Eq. 15-2
portant to know exactly which type of turbine is
best. However, books on turbines often start with a Hn - (m)
chapter on specific speed, so for completeness, this P - (kW)
brief summary is included. N - (rpm)
For a given head, a turbine will have one speed at
which it gives peak efficiency. This is the speed at Note also that specific speed is not actually a speed,
which the water flows most smoothly around the because its units are not simply [rpm). This means
turbine. If the head is increased, there will be a new that when the units are changed, the value of N 5
optimum speed, higher than the first. At this speed, changes. In the literature N, is often quoted in im-
the water will be flowing faster than before, but the perial, or British, units. The formula is the same as
paths followed by the flow will be the same. It is Equation 15-1, but the units used are:
possible to imagine the turbine run under a head of
only one metre, at a speed which gives the same N.JP
flow pattern. Ns-imperial =~ Eq. 15-3
n
When operating under a lm head, the turbine
would produce a certain amount of power. By scal-
Ns-impcrial - Imperial specific speed
ing the size of the turbine, making it bigger or ( rpm.hp0.75fft t.25)
smaller as required, a turbine could be made that
Hn - (ft)
produced lkW of power at lm head. It would be p - (hp)
operating at optimum speed, and the flow patterns N - (rpm)
would be the same as the original turbine at op-
timum speed. This optimum speed for a lkW ma-
Before comparing values of N 5 , note which systems
chine operating under a lm head is the specific
of units are being used. To convert between metric
speed of that turbine design.
and imperial specific speeds:
For historical reasons, it is more common to use
one metric horsepower as the unit of power, in-
stead of one kilowatt. The specific speed for a par- Ns-metric = Ns-imperial X 4.446 Eq. 15-4
ticular design of turbine is then defined as the
The reason for doing this exercise is to compare
speed, in rpm, at which the turbine, with its valves
different turbine designs. By mathematically scaling
fully open, would give best efficiency with a lm
down turbines to a standard head and power, the
head and with its size scaled down to give an output
different speeds at which they operate can be seen.
of 1 metric horsepower. In terms of an equation,
For Pelton turbines, the specific speed is propor-
this can be expressed as:
tional to the bucket width divided by the PCD.
N.JP
Ns- metric = - -1-
15.2 Choosing a turbine
H,: Eq. 15-1
Turbines with higher specific speeds run faster, as
Ns-metric - metric specific speed (rpm.mhp0.75f would be expected, and are usually lighter and
m 1.25) smaller that lower N 5 turbines. Thus it makes sense to
Hn - net head at turbine (m) use a turbine with as high a specific speed as possible.
P - turbine output power (metric hp) Advances in the design and materials of turbines usu-
N - turbine speed (rpm) ally increase the specific speed of that type of turbine.
However, specific speed is by no means the only
This is the conventional form. Note carefully the criterion for selecting a turbine. The other major
strange units for power. One metric horsepower is factor is head, and high-N5 turbines generally only
172
work with low heads. Figure 15-1 shows the specific range of powers and speeds deduced from its Ns
speed and head ranges for a variety of turbines. and and head limitations.
it can be seen that Pelton turbines have low specific o Site-specific features: it may not be possible to
speeds but can work under very high heads. Pro- excavate deep enough to fit a particular type of
peller turbines have very high Ns but can only work turbine, or the generator design may limit the
under limited heads. allowable turbine speed.
Other factors that limit the use of turbines are:
In micro-hydro, where there are often only one
o Efficiency at partial flow: if the turbine is re- or two types of turbine available, and then in only a
quired to run for substantial periods at partial few sizes, the choice of turbine is often quite
flow, its efficiency away from its optimum point straightforward, and can safely be made without
is important. reference to specific speed. But it is a good general
o Strength and hydraulic stability: the turbine may principle to use faster-running turbines to keep the
not actually be able to function over the full size and cost down.
HEAD
1000
[m]
500
400
I
300
200
I
PELTON
100 1 JET
... I FRANCIS
I
50 2JET
40 4 JET
I
I
30 6JET
Propeller
20
I
II Metric
specific
speed
10
10 20 40 60 80 100 200 400 600 800 1000
Fig. 15-1: Specific speed/head ranges for various types of turbines
173
REFERENCES
British and International Standards referred to in the text Daugherty, R.L. (1920) Hydraulic Turbines, McGraw-
(e.g. BS 1400:1985 or ISO 1191 Part 1) can be obtained Hill.
by post from BSI Customer Services, 389, Chiswick High Doble, W.A. (1899) Paper presented to the ASME Cali-
Road, London W4 4AL, U.K. fornia Meeting, September.
Douglas, J.F. (1986) Solving Problems in Fluid Me-
ASCE, (1993) American Society of Civil Engineers, chanics, Volume 1, Longman. ISBN 0 582 28641 7.
'Steel Penstocks', ASCE Manuals and Reports on En- Dowty (1978): '0'-Ring Catalogue, Part I - Technical
gineering Practice No. 79, ASCE, New York. ISBN 0 Data; also Guide for '0'-Ring Selection, Dowty Seals
87262 951 1. Limited, Ashchurch, Gloucestershire, GL20 8JS, UK.
Arter, A. & Meier, A. (1990) Hydraulics Engineering Durand, W.F. (1939) 'The Pelton Water Wheel', Me-
Manual, Vol. 2 and supplement, Selected Nomograms chanical Engineering, Volume 6, pp. 447-454.
and Diagrams of series Harnessing Water Power on a Eisenring, M. (1991) Micro Pelton Turbines, Vol. 9 of
Small Scale, SKAT, St. Gallen, Switzerland. series Harnessing Water Power on a Small Scale,
Bachmann. P., Scharer, Ch., Staubli, T. & Vullioud, G. SKAT, St. Gallen, Switzerland. ISBN 3-908001-34 X.
(1990) Experimental Flow Studies on a 1-jet Model Feinberg, W. (1983) Lost-wax Casting, A Practitioner's
Pelton Turbine. Paper presented to International Asso- Manual, Intermediate Technology Publications,
ciation for Hydraulic Research Symposium, Belgrade, London, UK. ISBN 0903031884.
September 1990. Fischer, G., Arter, A., Meier, U. & Chapallaz, J.M.
Baines, J. & Williams. A. (2000) Website and various (1990) Governor Production Information, SKAT-
downloadable documents: eee/ntu.ac.uk/research/ GATE, Switzerland.
microhydro. Nottingham Trent University. Fraenkel, P., Paish, 0., Bokalders, V., Harvey, A.,
Brown, A. & Edwards, R. (1991) Micro-Hydro Power,
Bier, P.J. (1986) Welded Steel Penstocks, Water Resources
A Guide for Development Workers, Intermediate
Publication, Engineering Monograph No. 3: Bureau of
Technology Publications, London, UK.
Reclamation, Denver, CO, United States Department
Fravit (1999) Fravit sri, Forged Runners for Pelton Units,
of the Interior, USA.
Article in Hydropower & Dams Magazine, Issue 3,
Bjerke, M., Brekke, H. & Berg, K.l. (1990) Modern De-
1999, pp. 82-83.
sign and Manufacturing of Multijet Pelton Turbines.
Gilbert, G.N.J. (1968) Engineering Data on Grey Cast
Paper presented to International Association for
Iron, British Cast Iron Research Association.
Hydraulic Research Symposium, Belgrade, 1990.
Grein, H.L. (1988) Efficiency Prediction for Pelton Ma-
Bodmer, G. R. (1900) Hydraulic Motors and Turbines,
chines. Paper presented to Turboinstitut Conference on
3rd ed. D. van Nostrand Co. New York. Hydraulic Machinery, Ljubljana, September 13-15.
Borciani, G. (1991) Construction of Impulse Turbines, Grein, H. & Keck, H. (1988) Advanced Technology in
(ed. Zu-Yan Mei) Mechanical Design and Manufactur- Layout, Design and Manufacturing of Pelton Turbines.
ing of Hydraulic Machinery, Avebury Technical, Al- Paper presented to International Association for
dershot, UK. Hydraulics Research Symposium, Trondhein.
Brekke, H. (1984) A General Study on the Design of Guthrie Brown, J. (ed.) (1984) Hydro-Electric Engineer-
Vertical Pelton Turbines. Paper presented to Tur- ing Practice, Volume 11, Mechanical and Electrical En-
boinstitut Conference on Hydraulic Machinery, Ljubl- gineering, 2nd Edition, CBS, Delhi, and Blackie & Son,
jana, 1984. Ltd. London.
Brekke, H. (1987) Recent Trends in the Design and Harris, J. (1983) A True Standard of Gravity for Malawi.
Layout of Pelton Turbines, International Water Power Paper presented to IMEKOI/nstitute of Measurement
& Dam Construction Magazine, November 1987, pp. and Control International Conference, London.
13-17. Harvey, A., Brown, A., Hettiarachi, P. & Inversin, A.
Brekke, H. (1994) State of the Art in Pelton Turbine (1993) Micro-Hydro Design Manual, A Guide to
Design, Hydropower & Dams Magazine, March 1994, Small-Scale Water Power Schemes, London, UK. ISBN
pp. 21-28. 1 85339 103 4.
Buckner, B. (1989) Pelton Wheel Using Lost Wax Cast- Hill, D. (1984) A History of Engineering in Classical and
ing, Unpublished notes from course run by ITDG and Medieval Times, Routledge. ISBN 0 415 15291 7.
the Agricultural Development Bank of Nepal; 23 July Hodge, T. (1990) A Roman factory, Scientific American,
1989. November.
Castro! (1984) Ball and Roller Bearing Lubrication, Hulscher, W. & Fraenkel, P. (1994) The Power Guide,
Booklet IND/60c/2/84, Industrial Lubricants Division, An International Catalogue of Small-Scale Energy
Burmah-Castrol Industrial Ltd, UK. Equipment, 2nd Edition, Intermediate Technology
Chapuis, L. & Frosch!, K. (1998) Optimized Fabrication Publications, London, UK.
of Pelton Turbine Runners, Hydropower & Dams Hurst, S. (1996) Metal Casting: Appropriate Technology
Magazine, Volume 5, Issue 2, pp. 30-32. in the Small Foundry, Intermediate Technology Pub-
Cunningham, P. (2000) Energy Systems and Design, lications, London, UK. ISBN 1 85339 197 2.
Canada, Pica Hydro Magazine, April 2000. Inversin, A.R. (1986) Micro-Hydropower Sourcebook, A
Dansie, J. & Bonifay, J. (1996) High Head, Record Out- Practical Guide to Design and Implementation in De-
put, Publicity leaflet on the Cleuson-Dixense scheme, veloping Countries, NRECA International Founda-
Sulzer Hydro. tion, Washington DC.
174
Jordan Jnr., T.D. (1996) A Handbook of Gravity-Flow Roark. W.C. (1989) Roark's Formulas for Stress & Strain.
Water Systems For Small Communities, Intermediate 6th International Edition, McGraw Hill. ISBN 0 07
Technology Publications, London. UK. ISBN 0 946688 100373 8.
50 8. Saubolle. B.R. & Bachmann. A. (1978) Mini Technology
Kempe (1989) Kempe's Engineers Handbook, Edited by I, Sahayogi Press, Kathmandu.

Carill Sharpe. 94th Edition, Morgan-Grampian. Schneebeli, F.. Baltis, E. & Keck, H. (1996) New Tech-
Kisioka, E. & Osawa. K. (1972) Investigation into the nology Earns Acceptance. Sulzer Technical Review,
Problem of Losses of the Pelton Wheel. Paper pre- 1196, Sulzer Hydro, ZUrich.
sented to Second International JSME Symposium on Shaeffer. J. (1991): Alternative Energy Sourcebook, 9th
Fluid Machinery and Fluidics, Tokyo, September. Edition. Real Goods Trading Corporation. Ukiah. CA
Lazzaro, B. & Rossi, G. (1990) Pelton Turbine: Model 95482. USA.
Investigation on Some Peculiar Geometrical Features. Shigley. J.F. (1986) Mechanical Engineering Design, First
Paper presented to International Association for Metric Edition. McGraw-Hill. Singapore. ISBN 0 07
Hydraulics Research Symposium, Belgrade. 100292 8.
Yugoslavia. Shigley, J.E. & Mischke. C.R. (1989) Mechanical Engin-
Macken, P.J. (1970) Copper Alloy Casting Design, Refer- eering Design. International Fifth Edition. McGraw-
ence Data and Foundry Design, Copper Development Hill. ISBN 0 07 100607 9.
Association, London. Siervo. F. de & Lugaresi. A. ( 1978) 'Modern Trends in
Maguire, D. ( 1998) Engineering Drawing from First Selecting and Designing Pelton Turbines. Water
Principles, Arnold. ISBN 0 340 69198 0. Power & Dam Construction Magazine. December.
Miller, D.S. (1990) Internal Flow Systems, 2nd Edition, pp. 40--48.
BHRA (Information Services). Smith, N .A.F. (1984) The origins of water power: a prob-
Moritz, L.A. (1958) Grain-Mills and Flour in Classical lem of evidence and expectations, Transactions of the
Antiquity, Clarendon Press, Oxford. ISBN for set of Newcomen Society, Volume 55. 1983-4.
volumes 0-405-12345--D. SKF ( 1994) SKF General Catalogue 4000/IV E.
Nechleba, M. (1957) Hydraulic Turbines, Their Design SKF ( 1991) SKF Bearing Maintenance Handbook
and Equipment, Artia, Prague. 4100E.
Nixon, N. & Hill, J. (1987) Water Power, Quarry Bank Smith. N. (1994) Motors as Generators for Micro-Hydro
Mill Publishing. ISBN 0 946414 06 8. Power, Intermediate Technology Publications.
Oberg, E., Jones, F.D., Holbrook L., Horton, H.L.. London. ISBN 1 85339 286 3.
Henry, H. and Ryffel, H.H. (1992) Robert E. Green Vivier, L. (1966) Turbines Hydrauliques et Leur Regula-
(ed.) Machinery's Handbook, 24th Edition, Industrial tion, Editions Albin Michel, Paris, France.
Press Inc., New York. ISBN 0 8311 2492 X. Waltham. M . ( 1991) The Manufacture of Pelton Wheel
Perera, L. (1995) 'Making Micro-hydro Turbines in Sri Runners, unpublished booklet for Agricultural De-
Lanka', Hydronet magazine, 1195, p. 10. velopment Bank of Nepal!ITDG, July.
Raabe. I.J. (1985) Hydro Power, The Design, Use, Func- Waltham. M. (1994) Mechanical Guidelines for Micro-
tion of Hydromechanical, Hydraulic, and Electrical Hydro-Electric Installations, unpublished booklet,
Equipment, VDI-Verlag Gmbh, Dusseldorf. ITDG Nepal, January.
175
17
GLOSSARY
y, gamma monobloc runners

bucket exit angle runners made as a single-piece casting
~.zeta surge pressure
efficiency of flow in bucket pressure rise in a penstock due to a decrease in
bore the flow
internal diameter of a pipe synchronous generator
cavitation generator rotor has coils or permanent magnets
air/vapour bubbles which can appear in low to create a magnetic field; speed constant for a
pressure fluid, often causing damage as they given frequency
collapse tangential turbine
direct-coupled turbine turbine in which a jet or jets of water act
running the turbine at the same speed as the tangentially on a runner; a Pelton is one form of
driven machinery a tangential turbine
head turbine setting/set
pressure, measured as height of column of water the vertical position of the turbine, used for
induction generator large installations where the powerhouse is
generator where magnetic field in rotor induced underground; a deep set turbine is lower, giving
by rotating currents in stator; speed may vary higher head, but giving a lower drop for the
even at constant frequency tailrace
manifold water hammer
pipework connecting penstock to the turbine extreme surge pressure due to sudden blocking
of flow
176
18
INDEX
y, gamma 15-16. 136 stresses 37-9

~. zeta 15. 136 theory 15-18
action turbine welded attachment to hub 43-4. 113-14
see impulse turbine butterfly valves 44, 77
alternator
see synchronous generator casing
aluminium 35. 91 see turbine housing
anchor bolts 83-5, 121, 156-7 casting, chill 110
Antipater of Thessalonia 5 casting, die 104
Atkins, Jearam 8 casting, investment
casting runner 110
balancing, runner 110--11 feeder (runner) system 107-9
ball valve 44, 77-13 investing pattern 109-10
Banki turbine process 106
see crossflow turbine wax patterns 106-9
baseframe, turbine 81-3 casting, sand
bearings, deflector 54 feeder (runner) system 98--102
bearings, generator 52, 87 finishing 104
bearings, shaft inspection & testing 102-3, 105
alignment 61 materials for 34-6
arrangement in turbine 55-7 nozzles 45
arrangement on shaft 61-3 parting compound 98-9, 102
assembly 117-20 patterns for 91-7, 100-1
commissioning 122 process 97-9
forces on 58--60 sand, moulding 101-2
friction losses in 18, 20--1, 143 single buckets 99-100
housing tolerances for 64-5 single-piece runners 39-40, 103-4, 105-6
lubrication 65 temperature 101
maintenance 123-4 terminology 98
problems 128, 143 cast iron
removal 124 buckets/runners made from 34-6
seals 66 investment casting, use for 109
shaft tolerances for 64 limits for runners made of 38
sizing 63-4 nozzles 45
types 60 properties 35-6, 145
bearings, spear valve 47, 50, 116 sand casting with 100-1
belt drives 2, 27-9, 31, 56-13, 81-2, 128, 161 shrinkage allowance 91-2, 103
bend, in pipe 152-3 testing 105
bifurcation, in manifolds 153-4 cavitation 12, 32-3, 129
Borda 6 channels
Bracca 6 How calculations 85-7
brass 34-5,38,91,100-1,109,147 chill casting 110
bronze 35,47,91,100-1,109, 147 coefficient of velocity, Cv 11-12, 24-5, 134, 146
bucket coefficient of contradiction, Cc 11-12, 45, 133
attachment to hub 39-44 coefficient of discharge, C 0 11-12, 134
bending moment on 37-13 coefficient, loss
bolted attachment to hub 40--2, 111-13, 148-9 see loss coefficients
definition of front, top, etc 15 commissioning 121-2
design 32-4 concrete 121
drawing 33, 44 contractions in pipes
elliptical section 9, 34-5 losses due to 154-5
exit angle, y 15-16, 136 convention
How efficiency/friction factor, ~ 15, 136 for bending moments 59
How patterns 15-16 for describing Pelton buckets 15
forces/efficiency 136-51 conversion factors
function 2- 3 units of measurement 131-2
head limits for 38-9 cope, casting box 98--104
invention 9 corrosion
machining for bolting 111-13 bucket/runner 35-7
manufacture/casting 91-116 penstock/manifold 74-5, 116, 161-2
materials 34-7 shaft/runner 71
problems 124, 127 crossHow turbine 1, 25, 173
177
deflectors 37,44,52-4,56, 81 , 89,122-3 0-rings for 165
desilting basin 87 properties 65
die-casting 104 temperature limits 122
direct-coupled turbine 2, 27-9, 55-7, 73, 82 Greek alphabet 132

discharge loss 15, 19
distributor head
see manifold bucket limits 38
DCS, Butwal, Nepal xi effect on bucket design 33
Doble 6, 8, 9 gross 27
drag, casting box 98-104 losses, penstock/manifold 27, 74, 124-6, 133,
151-6
efficiency net 27
bucket 15, 136-7 range for Peltons 1-2, 4-5, 10, 24, 26
nozzle/spear valve 12, 134 test, for manifold/penstock 116
optimum 21-2, 28, 122 units 132
overall system, components 19,29-30, 130-1 , headrace 2,86-7,121 , 126
159-61 Hesse, F. G. 8
penstock/manifold 27, 161 housing
runner 17-22, 137-9, 142-3 see turbine housing
scaling effects 20 Hurdy-Gurdy wheel 7-8
turbine 26, 136 hydraulic efficiency 142
typical 5, 8-10, 128-9, 161
ELC impulse turbine 5-6, 8, 11
see Electronic Load Controller impact turbine
Electronic Load Controller 90, 122 see impulse turbine
end-uses 2, 29, 31 , 90 imperial units of measurement 130-2
entrance (to bucket) induction generator 2, 29, 52, 56, 161
see notch, bucket see also generator
Euler 6 injector
see spear valve
falaise 22, 51 iron
fatigue see cast iron
bucket 32,37-9,43,114-15,145-7
limit 147 jet, water
shaft 67-8, 72-3, 121, 150-1 cover for 50,127
feeder systems diameter 11 , 25-7, 45-6, 134-6
for casting 98-102 divergence 14-15, 45
fits flow in 11, 13, 25, 134
see limits & fits force 58, 146
flat belts 28-9 function 2-4
flinger seals 72-3 number of per turbine 27, 28
flour grinding mill 5-6, 29 passing through runner 137-40
flywheel 87-90 problems 126-7
forces spacing 50-1
bearing/shaft 58-64 velocity 11, 24, 133-4
bolts, anchor 84-5 vena contracta 11
bolts, baseframe 157
bolts, bucket clamping 41-3, 148-9 Kaplan turbine 11, 25, 129, 173
deflector 53-4 keys/keyways 68-70, 166-7
housing/manifold 80-1
jet/buckeVrunner 18, 32,37-8,136-7,144-6 limits & fits 64-5, 168-71
forebay tank 2, 123 locking adaptors 68, 70-1
foundations 83, 121 locking sleeves
Francis turbine 2, 11, 25, 129, 173 for bearings 62, 119-20
loss coefficients 77-8, 151-6, 158-61
gate valves 44, 77 losses
generator mechanical 18
drive arrangements 27-8, 55-7, 60, 67-8, 73, 81-2 runner 19-21
efficiency 161 total system 19
fitting runner to 118
load controllers, with 90 machinery, driven by turbine 29
overspeed 22,52-5,87 machining allowance 91-2, 111
short circuit 67 manifold 73-6, 151-6
speed, operating 21 , 28-9, 55 Manning equation 85
see also induction generator Manning roughness coefficient, n 85
see also synchronous generator marking out
ghatta xi turbine housing plates 116
Girard, Madame de 6 Microcast process
gold rush 6 for making runners 115
governor 45,50,53,55,87-90,163 micro-hydro ix, 2, 9
gravity, acceleration due to 132 mitred bend, in pipe
grease fabrication 152
filling bearing housings 117, 123-4 losses due to 152- 3
inspection 123 Mohr's circle 150
178
monobloc runners 39-40 penstock
Moore, Joseph 8 commissioning 122
moulding sand erection 121
for casting 97- 9, 101 flow & losses 30, 47, 73-4, 133, 151-2, 158--62
multi-jet turbines/multiple jets 9--10, 22-3, 28, 45, 50--2, 60, 129 forces 84, 157
general 2-3, 15
Naiver 6 governors effect on 89-90
needle valve inlet 126, 161
see spear valve problems 124--6
Nepal xi, 1, 5, 9, 54, 89,104, 106,110,113,114, 118 protection 75
notch, bucket 2-3, 8-9,21 , 32, 34,39,95, 97,113,127 surge/thickness 51-4, 75, 161-2
Norse Wheel/Mill 5 valves 44, 54--5, 77
nozzle pipes
abrasion 45-7, 87, 124 flow calculations 85-7, 151-2, 158--61
alignment 117-18 Pitch Circle Diameter 3-4
blockage 51-2, 87,162 dimensions as % of 32
casing support 78, 80--1 Poncelot, J. V. 5-7
choice of spear valve or 44--5 power
design 45--6 equations, general turbine 26, 58, 130--1
distance from runner 14, 45-7, 127 Pelton turbine 141-2
fabrication 46 units 132
function 2-4 powerhouse 2, 28, 50, 73, 81 , 83, 85, 121
materials 45 propeller turbine I, 11, 24--5, 173
on precursors to Peltons 7--8
problems 126--7 Ra
sizing 127 surface roughness 64--6
spear valve 3, 12, 46, 136 Re
theory 11-15, 24, 130--6, 161 see Reynold number
types 12 reaction turbine 11
valves used with 44--5, 76 regulator
see spear valve
oil expeller 29 Reynold number 143, 152
0-rings 47, 73, 163-5 rice hulling mill 29
overspeed, turbine roughness coefficients
6--jet Peltons 28 channels 85--6
buckets, effect on 37, 115 pipes 151--6, 158-9
generators, effect on 52, 56, 87 roughness, machined surfaces 64
housing. effect on 81 runaway 22, 28, 37--8, 52,81, 144. 147,159
protection mechanisms 52-5 runner
sensing 54 attachment to shaft 68-71
bolted buckets 40--3. 111-13
paint function 3
housing 81, 121. 124 machined from solid 115
penstock 75, 152 monobloc/single-piece manufacture 39-40,
shaft/runner 72. 121 105--6, 115
parting compound, casting 98-9, 102 number of buckets 39
pattern making optimum speed 21
fabrication 92- 7 welded buckets 43-4. 113-16
feeder (runner) system 98-102
ingate 98, 100, 102 sand, for casting
materials 91 see moulding sand
shrinkage allowance 91-2 safety factor
PCD bolted joints 42-3, 71
see Pitch Circle Diameter bucket/runner 146--7
Pelton, Lester 8-9 key/keyway stresses 70
Pelton turbine materials 146--7
application ranges, head/flow 26 penstock stresses 75, 162
bearing and shaft arrangements 55--8 shaft stresses 67. 150
commissioning 121-2 seals
design 32--87 fitting 118
highest head 4 types 65--6
installation 121 see also 0-rings
invention 8-9 selection
largest possible 10 see turbine selection
largest, world's 4--5, 10 set
maintenance 123-4 see turbine setting/set
manufacture 91-120 shafts, runner 58--63, 67-73, 149-51
problems solving for 124--8 shrinkage allowance
selection/sizing 24--31 casting 91
shaft seals 72-3 shut-down mechanisms
specific speed 172 brakes 54
speed, operating point 21 swirl (English Electric) 14, 55
theory 11-22, 133-57 valves 55
Peltric sets 1-2, 55-7, 73, 78, 81-2, 116 see also deflectors
179
single-piece runners see limits & fits
see monobloc runners torque
Sl units of measurement 130 bolted joint capacity 71

sluice valves bolt tightening 42
see gate valves deflector 53-4
spear valve friction in turbine 18, 143
alignment 117-18 key capacity 68-70
body diameter 47 matching turbine to load 21
choice of nozzles or 44-5 power, relationship to 58, 131
Cv 12 Torricelli, Evangelista 134
design 46-50 Torricelli's theorum 134
efficiency 12 traditional mill 5-6
flow from 12-13, 135-6 trashrack 51, 87, 126
function 3-4 turbine
governors, control by 50, 53 see crossftow turbine, Francis turbine, Kaplan
invention 9 turbine, Pelton turbine, propeller turbine, turgo
jet from 14 turbine
maintenance 124 turbine casing
manufacture 116 see turbine housing
old designs 9 turbine housing
penstock, effect on 51-2 baseframe/foundation 81-3
problems 126-8 bearing support/alignment 61-2, 73, 117-18
shut-off valves, use with 77 construction 80-1
surge pressure due to 51-2, 161-2 dimensions 78-80
trashrack for 87 flow within 22-3, 78-80
specific speed 172-3 forces on 156-7
speed ratio, x installation 121
definition 15-16, 136 maintenance 124
optimum 21-2, 137 manufacture 116-18
splitter ridge 2-3, 8, 32-3,92-6, 117, 124, 127 problems 128
spoon wheels 8 variants 1, 2, 10, 23, 54-7, 73, 80, 82
spreadsheet turbine selection 1, 24-31, 172
for Pelton system calculation 158-62 turbine setting/set 24, 173
sprue, casting 98-103, 108-10 turgo turbine 11, 25
stainless steel
see steel units of measurement 130
steel
cast 34,36 valves
casting buckets/runners 101-2 function 4, 44, 77
friction at bolted joints 71 loss coefficients, Kv 77-8
investment casting 109 losses in 73,77-8,121-5,151,158
keys 70 penstock inlet 54
limits for runners made of 38 problems 128
penstocks 52, 162 surge pressure due to 162
properties of cast 36-7, 146-7 types 44-5, 76-8
shafts 67 see ball valve, butterfly valves, gate valves, spear
spear valve shafts 47 valve
stainless 34, 36-7, 101, 114-15, 147 V-belts 28-9
testing cast 105 vena contracta 11-14, 45, 133
welding 36, 40, 43 vertical mill 5-6, 77, 126, 134, 151, 156
surface finish velocity head 73-4
bearing seatings 64 von Mises stress 150
buckets & runners 97, 100-6, 127 V-ring seals 66
fatigue effects 146
nozzles 13, 45 water hammer 51-2, 162
sealing surfaces 66, 164 water wheels 5-7
surge pressure wedge belts 28-9
avoiding 53, 89-90 welding
calculation of 51-2, 158-62 bearing alignment with 61
test pressure for 116 buckets onto a runner 43-4,
thickness of manifold to withstand 74-5, 113-14
158-62 crack detection 114
valves and 55, 77-8 buckets built by weld deposit 115
synchronous generator 27, 29, 52, 56, 161 bucket construction 37, 115
cast iron buckets 36
tailrace 2-3, 79-86, 121, 124, 128 housings 117
tangential wheels 6-8 penstock safety factors 162
taper locking sleeves steel 36-7
see locking sleeves testing 114-15
see also locking adaptors windage 18-19,21-2,58,78,80,142-3
temperature wiper blade 78, 90
colours of hot metals 101
drop due to turbine 11 x, bucket speed ratio
tolerances see speed ratio, x
180

009

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

009

Diunggah oleh

Hak Cipta:

Format Tersedia

9

APPENDIX: STANDARD UNITS, FORMULAE

Quantity Unit Symbol Equivalent

Base units 9.3 Other Sl units

p = P Q . 1Jcotal equation to use rpm:

From To Multiply by Notes to gravity

APPENDIX: DERIVATION OF FORMULAE

Forebay tank Penstock

Fig. 1D-1: Simplified forebay tank, penstock and nozzle

This result is known as Torricelli's theorem, from

Comparing Equation 10-9 with Equation 10-10 it Also:

Eq.10-11 Eq. 10-17

10.3.1 Simple Theory

Then Equation 10--23 becomes:

~ /, F = mb. vi. (1 - x) (1 +~.cosy) Eq. 10.25

Bucket velocity v b (v r vb) mb. vJ .x.(!- x ).( 1 +~.cosy)

This is maximum when the runner is stationary, x =

Fig. 1Q-4: Angles at which buckets catch the jet

Eq.10-35 x.( vi.[~]- 2R. tanOA)

0a + 8A- 8 =X. (tan 0a + tan8A) 28A.Rl -2R.tan8A )

(20A- 8) So, multiplying Equation 10-28 by this additional

Fig. 1Q-5: Water missing a bucket for x > x1;m

line. Above this speed the splitter goes faster than

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Jet velocity into bucket

Eq. 10-49 v, =Rp.ro

w2 = ~-Wl Eq. 10-52 Eq. 10-56

2 by parts, and rearranging to make the sec48 integral

= f K0 .ArPc .(x.vJ 11w = 0.05 X x3 Eq. 10-67

Eq. 10-64 This gives a Joss of 0.5% at x = 0.46, which is realis-

10.4.1 Runaway load If the turbine is being operated at its optimum

casting shown in Figure 10-9. The buckets are inte- = 30 2 X0 0195. D 2

(b) Welded attachment

the maximum bending moment is going to occur at

Material Ultimate Proof or Endurance Density Tensile design Fatigue design

tensile stress yield stress limit stress stress range

Brass, 60/40 280 90 8300 30 20

No stem 0.0038 0.48

portional to its area multiplied by its moment arm

10.5.2 Analysis of a bolted bucket

sidered here, the von Mises stress is:

avm - von Mises equivalent stress

For torsion and bending only, the Mohr's circle can

Fig. 1Q-13: Mohr circle for torsion and bending only in

A word of caution, though. The estimates for ae

10.7 Manifold loss calculations f = [

The following sections give data for calculating the

Type of pipe Roughness

Fig. 10-14: Compound mitred eo bend with n joints

0.30 I _....-::: portantly, studying the graphs shows which sorts of

Fig. 10--15: Loss coefficients for various bends (Miller,

Figure 10-17 shows the side branch coefficient,

0 0.1 0.2 0.3 0.4 0.5 0.6 0. 7 0.8 0.9

Figure 10-19 shows K 31 for a good junction, that Eq. 10-93

10.7.4 Diameter transition losses 10.7.5 Worked example

/ ... )./ v = 0.02 mm:

penstock, and enters the powerhouse through a 35

795 229 55 200

Fig. 10-22: Manifold system

coefficient velocity loss

Fig. 1Q--23: Forces acting on a turbine housing

tance xi. The extra force induced in the bolts is

APPENDIX: FULL SYSTEM CALCULATION