Raabe J. Hydro power - the design, use, and function of hydromechanical, hydraulic, and electrical еquipment
Подождите немного. Документ загружается.
ps~o~k
~IKI
draft tuhc. The ratio of both the Inst figures was ~.cco@nized to
~IIC~C:ISC
\vitl~
inrrs;l~i~lg distance of the olrcra~ing point from the stripe of 7ero \vliirl past the run1;er
d
hcncC'
quiet opcration.
an
9.8.5
Unsteady
flow
in tlie draft tube of the turbine
nis isr~estigation is limited to part load operation. In a systcm rotating with the vortex
core. the tir~c-alleraged distlibution of total pressure across the draft tube section shows
nnerallg n parabolic drop of the pressure towards the eccentric axis of the vortex core
certain instant, (Fig. 9.8.4a). At a certain load this distribution is
independent
of the
rt
a
bold
provided
the core docs not cavitate (as usually happens). Then the vortex core is
c~iiracterized
by
constant critical pressure (Fig. 9.8.4a).
The origin of the vortex core is usual!y in the runner and due to
a
rotating stall by
flow
Lsyll~metry Greeter stall in the runner happens with growing distance from the line of
operation without whirl. Hence the periodicity and
the
form of the draft tube vortex
are
subjected to stronger changes.
The variation of moment of mpmentu~n past the runner and the throughflow distribution
and the variation of the vortex core combined herewith cause pressure fluctuations on the
draft
tube wall with varying amplitudes.
~~~roxirnately it can be said, that the ainplitudes increase with growing moment of
momentum at runner exit and with increasing distance from the bep, Fig. 9.8.4~. (Note
that also the frequency of the surge increases with its amplitude.)
9.8.6.
Remedies
against draft tube surge
and
power swing
I.
Guide ribs at the hub: The liydsa~~lic fundamentals for this remedy consists in the
following.
Any llo\v guided by vanes or baffles arranged in a pitch to chord ratio
below
'one' follo\vs Inore or less the direction of these devices. Rut this is at the expensc of
so-called inlet shock loss under a flow not along this device at its inlet, Cap.
5.3
(8.951.
Ilence the whirl
c,,
in the draft tube as the origin of the vortex formation may be
suppressed by such guide ribs attached to the hub as the local origin of the effect.
But
this
functions only at a certain desired gate position and shifts the bep irito this position. This
occurs in connection with eventual shock loss at
the inlet of the guide ribs and hence
at
the
expense of optimum efliciency.
Moreover the loss due to inlet shock in connection with possible stall there may worsen
!he
efficiency over a Inrger operating range. In an imaginary design the lowest part
of
the
hub's end may be arranged axially adjustable by means of
a
servonlotor which comes into
action only at the desired point.
On
the basis of his thesis work also dealing with this phenomenon. Mr.
Schlei~rrner
at Lehrstuhl und
Labor
iiir hydrat~lischc Maschinen und Anlagcn
dcr
TU
Miinchen, madc such a proposal about
1972,
but
in vain since the author,
hcading
the lab did not believe in it.
Afterwards
thc author learncd
'.b
from other inves~igators, Mr.
Grein
from Sulzcr Eschcr Wyss, and Mr.
A1esri.q
from KaMeWa,
such
it
remcd
jl
would help in the case of cnlergcncy (by the way the adjustable hub was never
Udc)
[8.95].
'I-
Guide ribs
at
draft tube wall: I-lere the ribs may have the same effect. But they are not
@effective, since they
are
now more distant from the hub region
as
the origin
of
the
397
.-
-
I
V.
Air injection: This can he ckcted by s~liftiiig v:llvcs ccl~incsting the vortex
core
(pro~~idcc!
tl~cy
founcl
it!
;vit!l the atmosphcrc. But
i1
can
be
cionc
also
by
press!~rizr.d
~consun?ptio~i
I
IJ,:,
ol
ratcri
ilurv
n;c:l\i~red
at
nornial niJjs) to different slelions. They
nlay
~IC
bcfcl-I.
ti:^
ru1iiit:r.
011
the
\Y:III
~f
tlie throat rins, at tllc
hub
(t111.0ligll the h~llow
shaft)
and
in
the centre
of
d~;~fi
rnhc inlet
19.13X;
9.1391.
in
tl?c
I;l!lcr
cas:
tllc
sir
injection Inax bc through
;I
pcrfor:~tcc!
pi11
~.;l~-ricd
by
tripod
on
thc
draft
tllbc
,>r
[!lro!!gh
a
pig:,
pcnctrating
thc bclicl of
[lie
clr;~i[
iui>c.
?'hc
I;~ttcf
resuit5
ill
;I
ratller
.:ifj-c!L,,igil.
..\il
t]i:.;c.
dcviccs
in;t>.
s~~!fcr
from \\/car.
tc;~r
:~rltl
vibration.
,\i~- inicction along
the
runner
\
211~
f:tCC
m;1>
o<cajjonall;
rcc.lucc so~-~ic\:'!l;lt tlic clr;~? and hence
i;iit-~.:\sc
the
crficicncy.
/\
pos;iblc
2fiici2nc5 tlr-op
-jLlch
dc\iccs ir~crc;~.;~.;
with
thc sl>ccil'ic spcccl
of
the
turbine,
\~IICI::IS
the
weor
(?:?cl-<:l>cs.
V.
'The rc;ll rc~rreciy:
:\
pr~pcj-ly plan~icd plant. especially
in
dc\;c.lopir!2 coi~ntrics shonld
;~ii\.e
a
first
scr
tiimensiont:cl
a.s
tc
I-ricet the first ~.)o\\~t'r dcm;lnil
by
opzr-nti!l,y ;~rul;nd its
hcp.
If
a
~:t
is
forced
to
\\.ark
f:~r
i~n-ay
from its
'ncp
tluring lor~ycr pcr.iods t1ii.n
:i
ti~rbine
\\-it11
ad!l~,?;tblc run!lc.r :.ant:
(K;ip:an
-f.
Deriaz
F)
shou!cf be prt.ferl-c:l.
if
thc head range
;ii!u\i.s
tili.;.
TIic!- alone frlciii:a~c
a
\.vhi~.l-i'rce opc~.ation p:~st 1I1c riinnct- at
any
load
and
',icncc ab(j!i~ii tile milin source
of
pressure srlrze
in
thc draft tube.
From
tociit!-'s
point
of
view. Laser. Dopplcr Anernometry
(LDA)
surely is the
most
rnodern and as
i\.2
think best
Lvay
of
tnensuring velocities in
turbulent
flows. Especially
i!?
the ;licinitj
of
n;lils. LD.4-technique is supcl-ior
to
all othcr anemonleters, because
no
probe
is
pl:~ce~l in
the
flc~v
passage. The only disadvantages
of
I,DA's
are the rather high
price
and
r
he
complicated handling.
---
\i
the Lcl~rl~uhl fiir h~draulischc hlaschiucn
und
Anlngcn
of
~he
'Teclinic;~l
University
Illunlch,
LLIA's
huvc
L~c:l
urcd
in
no\\
nicnsurcnlcnt
ior
ten years (1953). Ecforc describing assignments,
\\!:irationcll I~IOJCS. ~igil~l pr~ccssing units, let
us
have
a
sliort
view
uithe physical pri~iciplcs
of
Laser
Doppler
Ar;~.momslr>.
i8.1361,
(9.661.
.-\s
expressed
by
its name,
the
bnsic principle of
I,DA
measurement is the optical
Doppler-
el'kect
[9.140].
Although
it
is not precise in all crlscs, the basic theory is more easily
described
by
the so-cslled inrtrfurenct: model.
Two
focussed laser beams produce
in
the
39s
,r~,shing-scction a stalion:lry fringe p:!ltern. Tl~c dist:~licc
A.v
of the fringes is give11 by the
rsl;ltin~l:
1
(see Fig.
9.9.1.1
)
*.hel.c:
=
wavelength of laser beam
O
=
cross-angle between the two beams.
Fig.
3.9.1.
Illustrating
the
me;lsuring princi-
plcs
in
Laser ancn~on~ctr)t.
I)
Interference
rnodcl
with
sintionary
fringe
pattern, \relocity
nmponent
I>,,
~ne;isured
by
thc Laser, ~najor
,xis
I;,
111inor
a\is
1:
of
the
prohe.
2)
The
probe's
\loliimc close to
the
wall,
u
glass ~vall.
b
liquid
with
the
same
refractive index as glass.
Nearly every kind of
flonl,
carries n~icroscopicallp snlall particles with
it.
These scattcring
piarticles, cause
a
poslion
of
scattered light, passing through the probe volume
(=
crosb-
section of the laser beams).
If
the flow dces not carry cnough scattering particles, one must
inject synthetic ones. The size of scattering particles 11as to be smaller than
A.u!2.
The
frcquency
A
1,
of the scattel-ed light is directly proportional to the particle-velocity, accord-
ing to the
equation:
In
a
way, we do not measure flow velocity, but particle velocity. Rut
if
the scattering
particles
sa
tisfy c~rtain conditions, they represent the flow velocity, because they are
carried by the flow nearly without inertia. For Inore infor~nation about this see: Dlrl-st,
\!cliiilg,
\Z'hifc~la\c.
[9.140].
Dificultics in applying LDA-technique often appear by diffraction of light beams
at
optical col~tilct surfaces, complicating the determination of the probe volume. Redress
an
be n~ade by tuning thc se!rnctivc index.
In
doing so, a fluid is used, having at a certain temperature the same refractive index as
'st:
observation window of the test section.
A:
our laboratory decisive successes were
nade with this technique by measuring
Re),nolds
stresses
in
an
u~steady turbulent pipe
low.
One of the main advantages by tuning the refractive index is beside the siinple
jetermination of the probe volulne's coordinates, that
we
are able to approach to the wall
*ith
the narrow edge of the probe volumc. Probe volumes usually sre ten times larger
'long the major axis thcn along the minor axis. This means that we may approach the
with
the rlnrrow cdgc ien times closer without toucl~ing the wall
(Fig.
9.9.1.2).
Probc
'~lumes hitting the wi~ll. are
;I
critical point
in
LDA
measurement, as they produce "wall
'gnals", caused by particles adhering to the wall.
~0~11~'
i)/U~r(lllO~tf//
!llO(/''.S
(f
I.D.*IS
Rclk~cncc bca~.rl
\)\lc'~ii
for I-di~ncnsioni~l ~ncosurcmcnt.
Doublc rcfcrcncc
I>c':~ni
\ystcn~ (?-dim. ~ncasurcrncnt).
1-or~nrd
an(!
I,,~ch
sc,i~tzri~ig fringe syste111 for
1,
2,
and
3
bclocity componcIits.
blcnsi~rcnit.nt of flow vcIocitics
in
all kind of Iriinsprlrcnt media at Jifkrent tcrnpcral
h.lc.;~sutc'~iicnts
i:i
thc viciriity of walls without any disturbancc of the flow.
Using
frequency
sliifting, forward and backward turbulent
velocities
of
high frcqucncy
nated bctwcen.
Sizntrl
procvssit~g
Frequency
tracking processors.
Counting
processors.
Transient
recorders.
9.9.2.
Example
of
a
flow
measurement, carried out
by
lncans
of
a
two
dimensioni11
l,D,4
with
tracking processors
III the lab of Lehrstuhl fiir hydraulischr: k.1aschincn und i1nlagt.n of
?'[Jbl,
Munich in
a
periodically
r~nstcady,
fully
developed turbulent pipe
flow,
the ~urbulerrt kitlctic energy
--
q.'iZ
and the
Rc.!trtol~!s
stress
-
_om
were measured. Becau:ie all velocities and even the
mean ones
were
ti111e-dependent:
an
electronic data h:tndling system
was
indispensable.
A
period of this
iiow
nlay look as shown in
Fig.
9.9.2.
The er~semble mean values of
the
vc!ocitics were ca;cul;~ted within
a
small time-space
At,
in which
ihe
flow was considered
io
bs
steady.
Thc
flow
period
T
was
divided into up
to
100
time-sptices according to
.
Fis.
3.9.2.
In
cach
mean
values and
Keyrtol(!s
stresses were cornputod and stored. This
was
done
by
repeating
LY
times the salrle mean flow condition. The pulsation period also
was
measured and controlled by the computer.
Fig.
9.9.3
is
a
scheme of the electronic equip-
mcnt, used for the measurements. The oncoming analogtie data froin the tracking
pro-
cessors were digitalized
by
a fast 12-bit AD-unit and
transmitted
to the computcr, doing
the calculations and
stori~lgs of the mean values for evcry time-space. So in two series
~f
msasurcments the radial and temporal distributions of
U,
~i',
v',
w'
aild thereby the
Fig.
9.9.2.
Graph of
an
axisymmetric
pulsating
fldw
in a straight circular pipe,
showing the
\relocity
11
in the direction of
pipe axis
:
vcrsus time elapsed
t,
a certain
cross section given by
2,
and at a radius
r,.
:It
ti~nc increment at
:he
instant
to.
T
period. After
llartn<r.
Fig.
9.9.3.
Scheme of electronic equipment used for ~neasurement and data acquisition. 1 Laser-
beam,
2
optical hend.
3
measuring section (pipe), measurement of the instantaneous velocity compo-
~nts in the
3
coordinate directions 11,
5,
w.
4
detecting unit;
5.1
photo multiplier for
-I-
VV,
1,;
5.2
ditto for
+
V
u,
\r;
6
Doppler signal;
7.1
Doppler signal processor type Brow11 Bovsri and Cie for
";7.2
ditto for
v
and
it.;
8
loop closed signal;
9
signal noise ratio ok; 10
L..,
\v:
1
1
11;
12
computer
PSI
80
F=
,-===7--
Kontron, analog digital conversion of
(I,
V.
U:
LIP,
T,
p,
J17'2,
,1
\vP2,
11
11
,
11'
wf,
tl'
\v';
13
trigger signal;
14
inductive pressure transducers to measure the instantaneous pressures pi and
I2
~irnultaneously at stations
1
and
2
a distance apart
Ar
in the streamwise direction:
15
p,,
p,;
16
arrier frequency amplifier
KWSI 3s-5
type Hottinger Baldwin:
17
pressure gradient
L3P;.I::
18
terminal;
19
data logger, discs; 20 to subsequent treatment in the computer centre. After
Ifarrner
19.1
331.
dynamic pressure
q2/2
and the most interesting value
u'
were obtained. Logical signals
from
the
Dopplel.
processor told the conlputer whether the oncoming velocity signals
wcre
good or not. These signals were "SNR-ok"
(=
signal noise
ratio
ok) and "loop
1
dosed"
(=
signal processor has found a
Doppler
signal).
In
this way oniy good
Dopplcr
signals were
taken
by
the computer.
An
example of
the
measurement results is given below in the
Fig.
9.9.4.
At
a
constant
point
of time, radial distributions
of
tl'",
vl*,
wl*,
q2/2,
and
tl'v'
were plotted.
N
=
counting variable
T
L-
period of pulsation
J
I
=
time-space
ti,
V
=
mean velocity component in the direction of the pipe axis, normal to it
8).
d,
10'
=
fluct~~ating vclocity compnncnts in the
z,
r,
cp-directions
*
-
*
-
root-mean-square valucs of
11'.
u',
w'
dd
=
Re).r~ol(l.s-shcnr-stress
pcr unit inass
Ax
=
fringe distance
b
=
mens~~red velocity component by
LDA
kr.9
I
=
cylindrical coordinates in axial. radial and azilnuthal direction, respectively
8
=
wavclenglh of light
=
cross-angle between illumination beams
1
v
=
Doppler
frequency
I
=
density
Fig. 9.9.4. Exainplcs for
measurement
of a pulsating pipe flow at a certain cross section versus
dimer?sionlcss pipe radius: Reynolds ~li~mber of thc timc-averaged flow Re
=
20,l .
lo3; period
T
=
1
s; ratio of stroke of pulsation generator to pipe radius
OS
=
2,03; ratio of amplitude of thc
sinusoidally oscillating pnrt af
the
flow to the timc-averaged flow rate
P
=
Q,/Q
=
0,197; Womersley
P;irameter
R
=
K
=
53.5. The root-mean-square values of the fluctuating vclocity compo-
nents
u'".
c'".
\v"
are made dim~:nsionle~s by dividing by tile wall shcar velocity 11,; the ki~?ctic energy
of the turbulent motion is rnadc
dinlensionless by dividing by the squared wall shcar velcicity
uf;
phase relationship: full I~ne: 0,075
T;
hatched line; 0,563
T.
a) Root-mean-sqaurc values
ti'"
of
velocity fluctuation in
thc pipe axis direction; b) ditto
a'"
in the radial direction; c) ditto
\v'"
in
the
tan~ential direction; d) kinetic encrgy of turbulent motion; e) correlation
rr'
corresponding
to
apparcnt (Reynolds) shear stress on a coaxial cylinder and in direction of the pipe axis. After
Hcrrtner
[S.
1361.
Concluding remarks
on
measuring techniques:
The conversion of any mcasurcment into a signal as
a
computcr input by analogue or digital
technique
iuld its computerized e~nluation for the dcsired output as described above, may
be
considered to bc the present trend in any up to-date experimental
techniques.
Tlle computerized
processing
c.f
~neasuring signals is of special value for dynamic measurements of unsteady phenome-
na. Quickly
repcatcd measurements at
a
certain measuring point in connection with the high
processing
speed of a computer also yielas the timc-averaged mean value of any measurement and
its
time-aver:~gtd errors during a ccrtain observation period. Hence human errors in observation
will
be excludcd more and more.
10.
The
turbomachine, its design
and
construction
10.1.
Introduction
hjany types of hydro turbo machines exist. This is a consequence of their different duties
as
a)
turbining alone,
b)
pumping and turbining in a pump-turbine,
c) direct and reversed turbining and pumping in tidal power plants, and
d)
pumping in a pumped starage plant with a ternary (tandem) set.
~ut
it is also a consequence of the varying water requirements and topography, the hydro
turbo machine
has to be adapted to.
~~arnessing of low head requires axial turbines
(AT)
with high percentage reaction and
susceptibility to cavitation.
Thcir appearance is characterized on the one hand by the
classical design of
I!
Kaj~lait,
nanlcly the Kaplan turbine (KT) with vertical shaft and a
spiral or semispiral
czsing as the generator of angular momentum required, an elbow
draft tube and
a
runner usually with adjustable vanes. The version with full spiral casing
overlaps the range dominated earlier by the
F-rancis turbine (FT).
In
the lower head range the AT has the variant of the tubular turbine (TT). This is
characterized by a straighter water duct than
the original KT. Here the generation of
wl~irl is left to the stay and gate vanes. The variants of the TT are manifold. With the
generator in a submersible bulb, they.appear as bulb turbines (BT).
Tiley have also the generator outside the water duct directly coupled or connected by
a
bevel gear. Then they may have the straightest possible water duct in the form of
a
rim
generator
type (RGT) or Straflo type (ST).
In
the lowest head range, the
IT
may enable the harnessing of the kinetic energy
in
river
rapids. To
date, the
TT,
especially in its straflo version, seems to be the effective tool for
harnessing tidal power. But there, in the usual case
of
a
one basin plant, the function of
a
reversible turbine and pump requires special consideration and compromises of the
blading, when the runner vane position is nearly retained.
The conical and axial distributor of a
TT
is a rather poor generator of whirl and needs
careful consideration of the flow in the gates both with respect to geometry and loss.
The design of a
KT
has to start from the runner tip diameter either with respect to
lntenlal loss or suction requirements. Emphasis has to be given also to the proper draft
lube
design, since the kinetic energy at runner exit may amount to
80%
at overgate.
Therefore TTs, with two bends in the draft tube to allow the shaft passing through the
bller, may have additional losses.
fie
layout of
a
turbine has to account for the stresses in the rotor vanes, the hub and the
:;l~r.,,~icl
1111cIcs
[hc
~:iiergcncy
C;~SC
of
run;rw;ly.
Yllc:
sl:.i~ft clcsig~i should shift the
1
CI-i~ic:il spcccl
;!tlo\c
tlljs tlircshold.
=
-
%
-
I-11,-
ticsign
of
;iccc\sorles like be:it-ings, s11:lft se:\ls ~tc.. slloulrl f~llo\v tlic clircciioll
or
'
stiflctling tlic sh.lR
aitll
respect to
lilt
critic:il
speed.
Murcover tllcre elcn~ents [nust
bve
c;!sy acccsslbilil
y
for assembly and clisrnnntling.
Tlic circi~lation of the lubricant ,hould be cffectecl by thc shaft itself.
1
Icad cover, gakq
stay vanes
and
tlicir rings and the spiral casing have
to
bc clcsisned for low values of1
0%
dzfnlm;ilion, and ndmissible stress even under cn~wgency loads, like
water
hammer,
and
runaway. ant1 fc>r
:I
limited radial extension, p:irticularly of the outrnost pasts like the
spiral casing, stay vane and gate ring.
In vertical
scts, thc accessibility of the lov~er parts such as thc runner
and
the bottom
cover of gates slic~ll(l be fxilitatctl by auxiliary cranes, inspection doors in the draft tube
or intermediate
shaft
pieces to be removed for dismantling of the runner.
Fliirnessing higher heads requires Frallcis turbincs
(FT)
with mcdiiim percentage reac-
tion. For
highcst
heads, sm;~llest discharges,
acd
mean heads, in~pulzc t~-lrbines of the
Yeiton type
(PT)
witl~out any reaction are applied. Wide zoncs
c;f
overlapping exist
between
FT
and
KT
on the one hand
and
FT
aild
PT
on the other.
Thc
cross flow turbine, a cross-breed between reaction and sction turbines, but more an
action turbinc. orisinally designed by Michell
arid
Ranki, and now built by Ossbergcr,
appears mcstly in micro power plants.
It
nlay be interposed bctween
FT
and
KT
especially
ill
cases of highly varying discharge, head, and tail water level.
The
!latr?ess of the erticiency vs flow and the bep efficiency ch:lrncteristics of
a
FT
depend
m::ir;ly
on
the
hsacl and hence
on
the given specific speed. This figure determines
deci-
si~cly :he _geometry in the sectional view of
a
FT.
-
I
~ansportr~tion. fibsicarion on site, submergence
of
the set, mode of working and overall
economy, sand
zrosion and simplicity of disnantlin~, and maintenance may further
decide the
final type of
a
turbine.
Flow field and strength of
a
FT
runner are related to geometry of the runner vane and
slirouds.
The
limitation of output for
a
give^!
runpcr dianeter results mainly from its
relative span of vanzs, and the permissible stress of runner
vane
material under emergency
zorlditior,~ such as runaway.
The desisu of
a
pump-turbine
has
to start from that of
a
mere pump. It is based on suction
requiiements, optimum efficizncy, different hill diagrams during pumping and turbining
and
speecl clianse.
The broac! 5eld ofhydrcturbomachinery is reflected in a long reference list.
i10.1
to
10.51, [1.50], [6.Q],
[6.35],
i8.1321
give
2
survey on machinery. Kaplan turbines are treated in general in
[10.6],
special
co~~structio~s
=f
them.
nrith emphasis
on
high-head design, in
[10.7
to
lC.191,
their layout, based on
modern
tendencieb
1:1
[10.1
I],
special design problems, like geometry
of
the runner chankber,
in
[10.12].
thc torque
on
the runner vane spind!e in
[1U.l3],
pteclicted and calc~~lated flow in
[10.14],
its
losses,
gencrall?
in
j10.15]. and
on t'nc runner bladcs in
[10.10; lG.171.
The tubular turbine, as the
cthcr important variant of tlie axial turbine, is presented in
n
survey of bulb turbines
[13.18],
in
general
treatises
[lo.
i
9
to
10.211
by
studies of individual constr~~ctions
[i0.22; 10.231,
with
its
prob1er.l~ a1
tl:e
intakc
[10.24].
\iith
experiences on its operation
(10.25; 10.261.
by a comparison of
predicted and
measbred stresses
(10.271.
A
spccial diagonal turbine emerged from
the
bulb turbine
[10.28],
then the Straflo design
[10.29;
10.301,
also used :o cxploit tidal power
l10.31;
10.321.
hlodr.rn trends in Francis turbine dcsign arc given in (10.331, somc\vhat older
it]
[10.34]. the
const;
ti~.fion:il limits of
I'~;IIIC~S
turbines in [10.35], its
computer-aidcd
design in snnle nct~~al con-
Str,tctic)ns In [10.37 to 10.3!]. its his!] head dosigns in [10.42
111
10.441, 11s ~elativc
11~)~
in [10.35 to
]0.47], its draft tube surgcs in 110.48 to 10.501. the welding of onc of its 1;lrgcst runners in [lO.jl],
be
slrcsses in runners in [10.52]. Operating cxpericncc on Francis ti~rbincs and problems due to thc'
of large units arc rcportcd in [10.53 to 10.581, the features of its single gate scrvolnotors in
llfi59],
ils start up in [10.60]. ilnd the design of blildi~lg in [IO.hl].
~j~dcrn trends
of
Pclton turbinc design arc found
in
[l0.62], the alternative of Fr;incis or Pelton
tUrbil~~ is discussed in (10.631. details of i~ctu;ll l'cltor, turbines are found in [10.64; 10.651, the stresses
ln
their important parts in [10.66; 10.671, and the qucstionablc potential reduction of thcir runalvay
,peed ratio in (10.681.
Storage
pumps of ternary scts (tandem scts) for puinpcd storagc are treated in [10.69
to
10.721,
problcllIS
of
pumpcd storagc plants in (10.731. The pump-tul-binc in general is described in [!0.7!,
110.75 to 10.791, niodern design trends in [10.80; lO.Sl], some actuiil constructions in (10.82
to
lO.S9],
operating expcricnces in I10.90 to 10.921, its oprirnizntion. with rcspcct to loss and cavitation
features, in [10.93].
~hcf~liowing other topics, related to pump-turbines, are treated: limit of head per stagc (10.941, start
upillto puinpin~ Inode [13.35]. dynamic load [10.96],
transients
[10.97], axial thrust [10.98], runaiiay
110.991, part load opcration [10.100], the strcshcs on its components [10.101], a possible hysteresis of
lfi
ch;lractcristics (10.1 011 g~tc vibrations [10.103 to 10.105], gatc spiildlc torque under misaligned
gates [10.106], start up through boostcr J:rancis turbinc [lo. 1971. the sloehastic radial forccs on its
rotor [10.108 to 10.1 101.
A
no\~clt~ is the gale-rcgulatcd two-stasc piimp turbine [10.1
1
I]. Unregulated multistage pump
turbines (10.1 12 to 10.1 141 and the lsogyrc dcsign [lo.
1
151 arc sporndicallp used.
With rising
six i111d spcttd of the set, its critic:~l shaft vibration, its rcsponse to r:itlial forces and the
sddcd mass
dus
to flcxiiral kibrntions bcco~ne 1111cresting [lo.] 17 to 10.121], also the bchaviour of
ciastic adjustable gatcs
[
10.1221.
Losscs
ir~
the spiral casiny [!0.123], its stress calculation (10.124: 10.135], reduced to
a
conical shell
jIO.T26]. i?rusc! IifundTFcydcs of pumpcd storagc [10.127],
it>
dcsign with
a
stay ring of parallel
plates [10.128j, the distribution of the dynaniic pressure on
it
[10.129] are treated; further the fatigi~c
problcn~s of stay vanes [IQ. 1301 and thcir dynamic bchr;viour [IO.l-3l
j.
Thc
thrust bearing
of
the tiltcd pad typc [10.13:! to 10.134].
and
of
ths newly introd~~ced hydrostatic
type (10.135] Ivcrc dcscribcd.
also
sclf lubrication of trunnions ivithout any lubric21nt (10.1 361, sh:ifc
vals j10.1371, the condcnscr operation of pumped storagc sets [10.136], and colnputcr aided dcsign
l10.36].[I0.122], [10.139]. supported by the fi~litc clement mcthod (10.1-40; 10.141]. I'lle standardiza-
tion of
srnall turbines makcs progress [10.142]. Spccial features
of
stnall turbinc types have been
invcsligated, e.g. thc influcncc of the jet and wl~cel sizc on the perforlnnnce of inclined jet ilnpulse
turbines [10.143].
10.2.
Project
and
construction
of
axinl
turbines
10.2.1.
Geseral
survey
of
types
Axial
turbines exist actoally in ttvo variants.
One
usually with a vertical shaft, radial
hlributor, adjusiablc
runner
blades wit
11
t'ulcrt~nl in
the
hub
is
an i~lvention
of
V
Kr~plun
'nd
hcnce callcd
KT.
This is also valid for
the
K1'
with
fixed
runner blades the so-called
ProPcller turbine.
nexcond axi;il turbillc type
used,
has
a
more
or
less horizo~lt;il shaft,
an
axial or conical
d~slributor as the pencrilror
or
the whirl component
c,,
before the runner, and an axial
ruilncr
of
t11c
KT
type with tisunlly adjiist:~blc ninncr vii!lcs.
It
is aclincd llrre as
a
t~~hulpr
t\lrbinc
(-TT).
711e
:1x1:11
turbine
wit11
hori,:ui~t;~l shalt
;111t1
its ycncr.~tur s!l~-unh
on
its
shroud, ;iccorcling to
;in
US
patc~it of
ilarza
from
I9
19,
rcnlizcd
:lt
first hy
11.
Fi.scllpr
Escher Wyss, bclo~~gs
;IISO
to this type of KT.
It
will bc considcrt.d more in detail.
Adjiistzbility of runner vanes originillly was rnrc in this desig:~. With a hydrostatic
watcr-lubria~tcrl. self-iidjusting radial ar;d axial bcarii~g for thc
il~iilI
thrust arranged
on
thc shroud, to facilitate runner vane adjustn~ent, this typc now
has
besn revived under
the name
of
Str:iflo turbine. This was a specin! development of the Swiss firm
&her
Wyssi
h
joint vcntgre of the Sulzcr Concern, hencc named Sulzcr Eschcr Wyss.
10.2.2.
The true
Xaplnn turbine
This has usually (Figs. 10.2.1 to
7)
a vertical axis and a generator outside the watcr
duct.
The flow admission is usually by
a
radial gate precetlcd by
a
spiral casing when
tile
turbine has
a
larger size. This tends to be
a
semi-spiral casing (Fig.
3.4.12;
14) with
a
trapezoidal cross section of concrete in the lower head range. For higher heads, the full
spiral casing is used of circular cross sectioil, either lined with steel, or welded from steel
(Figs.
3.4.15
and 10.2.5). In conjunction with radial gates, the generation of a constant
angular
monlentum bcfore the runner is achieved easily.
In
a
few cases, small high head KTs have also a horizontal shaft and a rnetal spiral casing
(Fig. 3.4.17). In
the
usual design with
a
vertical axis, the machine is equipped with an
clbow draft tube. Its bend increases the diffuser loss which becomes important for the low
head KT
with its relative high kinetic energy (up to 80
%)
at
the runner exit. Therefore
this
elenlen
t
(Figs.
3.4.12,
13,
14) has
to
be
carefully
di~mensioned.
The large cavitction coefficient
0
of
s
KT may require subtnergcnce cven at modest heads.
Thereby the head of
KTs is limited [10.7] up to 70(80) m.
Bzcailse
of
its hi~h degree of rcaction up to
90%
and the impossibility of an axial double
flow cdesisn, rlie
KT
has a high axial thrust. This rnay lead with high heads to
a
diameter
of thrust bearing even larger than that of the runner. Hence the axial thrust also limits
the head of
KTs.
The width of a power house with vertical sets results from the lateral distance of
3
or
4
runner diameters for semi or full spiral casings (Fig.
3.4.45).
10.2.3.
The
tubular turbine
(TT)
The
previously mentioned features of performance of design also occiir in
TT
with the
difference that the straizht draft txbe used here has
a
low difTuser loss and requircs
a
smaller excavztion. Further the distance between sets can be reduced to
2
runner diame-
ters.
h4oreover the
'TT
saves the
flow
deflection between gate and runner (see Fig. 10.2.9).
For
its smaller drzft tube loss the
TT
is more susceptible to cavitation than a usual
KT.
Therefore its head is limited to about
20
m.
A
weak point of the TT is its nearly 1101.izontal
shaft in the case of the bulb turbine (Figs. 10.2.10 to
16).
I-?ere the assembly is complicated
by
the shaft's bending line in order to retain the small
gaps of both the rotors within their stationary part (throat ring, generator stator).
In
the
L
usual case
of
a directly coupled ~~lternator and small admission hatch to the bulb through
a vertical hollow rib, the
asse~nhly and disnlantling of the generator through
rr
hatch
at
the downstream end of the bulb becomes diflicult.