Raabe J. Hydro power - the design, use, and function of hydromechanical, hydraulic, and electrical еquipment
Подождите немного. Документ загружается.
Fig.
9.7.1.
4-hole quick response boundary laycr probe on runner
\lane
after
Fur-trror
(9.1
351
with
micro transducers,
2
mm
in
diameter. a) Perspective view and frontal view of the probc.
A.
C
ports
for directional
rncasurement;
R
port for total pressure;
D
port for static pressure. b) Hydraulic device
fdcilitati~ig a stepwise axial adjustment of rotating probe by nicans ofg;isoline fed an~li~lar scrvcnio-
tor, supplied by pipes from a stationary head via shaft and vane.
1
probe;
2
threaded bush;
3
scrcw
for adjustmcnt with grooved ball bearings;
4
annular piston;
5
axial piston;
6
restoring spring;
7
free
wheel device;
8
relieving boring;
9
pressure borc for axial piston;
10
pressure borc for annular piston;
I I
toothed ring for fccdback;
12
contact for feedback;
13
head cover and guide of probe;
14
probe's
guide bearing;
15
adjustment screw for prestressing thc bearing;
16
runner blade. Note the annular
piston in Section
C-D.
After
Furrner
[9.135].
inders (Fig. 9.7.2a). The positioning device, incorporated into the vane permits displace-
ment of the probe head up to 9
Inm
normal to the vane surface.
111.
Time-averaged measurements: Quasi steady investigation of the relative flow in the
proximity of the vane face was made possible by averaging the
synchronoi~s records of
pressure fluctuations from the
4
ports vs elapsed time obtained by a light beam oscil-
lograph over
a
certain period, say one runner revolution (Fig.
9.7.2
b). Hencc the \!aria-
lions of the velocity components in the radial and circumferential directions on both sides
of the vane can be obtained. An essential chnracteristic of this flow pattern is the
secondary flow, directed towards the outside wall on the pressure face and towards the
hub
on the suction face of the vane (Fig. 9.7.2b). This follows from the rcliltive eddy
(Cap.
5.5.5.4).
IV.
Con~parison between time-averaged results
with
predictions by the method of singu-
larities and boundary layer theory: Velocities parallel to the surface as functions of the
Fig.
9.7.2.
Time-aieraged vclocitics
from
dynamic mcasurerncnts
aftcr
iV.
Flr).tllc)r. a) Scheme
of
n?cnsuring
stations.
L
chord;
c
cstcrnnl cylindrical section, radius
r-
=
0,3
r,;
m
nlcati cylindrical
zccrion, radius
r
=
0.77
i-,:
i
internal cylindrical section,
radius
!*
-
0.56
r,,
I.,
runner tip r;tdius.
b)
Schcmatic \.ie\v
<>f
proiilcs
of
the relative velocity on
a
Kaplnn t~~ibillc runlicr vatic close to the
ivu!l.
P
prcbsurc race:
S
suctior? fi~ce. Out\vard; ~ricntcd radial vclocitics on t11c
pressure
LICC,
inwardS
i)rientc':
Oil
the suction f;:cl-..
11
=
0,615
rn;
bcp. c) Compnrison of thc c:~lcul;~tcd anti
measured
l~cloririzs
ir:
:lie
cyli~?dr
ical st.i.iio.ls, c;~lcul:itiol~
by
thc singularity mc[hocl iiftcr .\ILI~.~~/IsL'II
and
Jttcob
[6.7:
6.31.
o
point r:leasi~rc.d.
---
li~c calc~~lstcd. After Fiirti~o
[9.135].
actual
vane zeotnctry
and
the infloiil and outflovi angles mexsured by
Ca?ror.pk
[9.116]
\:;ere
comptited
by
msar:b
of the method of singularities of ,l.I(7rterlserl rind
.Iuc.o!~
[6.2;
6.31.
Rcstllts sho\v
a
icr! good azreemefit with the time-averased velocities ~ne;~sured at the
outer edge of tire
bo!~rlclnry
1:1yer,
Fie.
9.7.2~.
Characterixtics of the boundary layers at the three cylindrical vane sections have been
computed.
~ibitlg
;:a
integratioii mt:ih~d desclibed by Rotta [9.136]. Approxim;ltion of the
timc-averaged
bounclar
laycr
prc~filea by means of e~poncntial functions and their
integrations. yielded
thc values determined experimentally. Only the results obtained
for
the outer c~iindricrll scction ivere very satisfactory
[9.82].
V.
Unsieitd; measurements: Esse~ltinlly the wake areas of
thc
24
guide vanes
in
the
relative
flt~il. ca!l on11
bc
found
by
pressure ~~~t.asuremcnts it1 the unsteady state.
The
resi~ltitig ~ncxiifications
of
relative vzlocity
111
magnitude and direction were
of
the order
of
20
7'0
of
the nican
relati!
.-
velocity and
6
degrees. This gives,
by
the way, an amplitude
of
the hydi-odynrlnlic
lift.
rlt
le~~st in the outer cylindricril section of the order of
40%
of
r
he time-ai sraged lift.
11na~ine that this iluctuatin~i'orce attacks the vane with
a
Icng
nloment arm
with
respect
to the vane's root
at
the h:lb. creating tlizre a11 oscillating bending stress of the same
order. and in connection with fatiple affects a gossible rupture of the vane from the disk.
The usual cause for sucli wake-induced stress I'luctuations
is
on the one hand the endeav-
our to
run with oi7ergare and then to shift the gate's stem towards the shaft so as to have
o~crgate, suidc vanes that e:.:cnd deeply into the
flow
with Ihclr rc;irw;tl-d edytl.s and
.
,llc
;~cij;ice~~l blutli lower front f;~ces ("lowsr"
at
;I
tiirbinc with ;.crticul sll;if~)
~~~nuse
of
thc csccssive thickness of the gatcs (\~lns thickncss
to
clios~l ratios
oi
1
:
5
;~rc
the g:itc slsucture from wcldcd plntcs
;~lso
oil the rc;lr,
;~nd
the impoxsibillty
o!
stlch
n
welded guide
writ:
at its lower front surhcc in ths strcnni\~isc dir-cctior..
Ihc
opened guide vane lcavcs
a
big wake in this direction.
,,
[allows
from tlleorstical considerations that fluctu:ttions of direction
:III~
nl:tgnitudc of \.c.locitv
,*d
hcncc also prcssure will incrciisc
011
appro:~cliing thc vane sul-f:lce. This is
confilmcd
by
,.,,c,surerncnts (Fig.
9.7.2).
According to the findings of
SP.YI
[9.137]
wit!i
rcspcct
to
an
nscilli~~in~
bminur flow (not a had model siricc
any
flow
has
a
laminar sublayer), thc niiisimiim of the oscilla-
t;,lls induced by the gates with
112
cyclcs must be loc;~tcd at
a
\\J;III
dist;t~ice
(i*
of
obviou~ly this range cannot
bc
covcrcd by the probo.
Even
if
fluctuations
of
the
velocity in magnitude and direction
long
tlie leading
etlge
oT
the
vane are
of
the same nature
both
on the prcssvre
and
suction faces
of
tlit. ~~:inc., tlieir
further de~lelopment in the strcamwise direction differs strongly along both the faces.
f'lg.9.7.3.
Vclocity fluctuations aic!ng thc rucner vane.
a)
Fluct~~ations of the rclatite
velocity
at
43
Dagain5t
wall
distance in the strcam\vise direction oi> thc suction fitce. stations
9,Y,
7
of
Fig.
9.7.2.
I~~rcasc
of
the fluctuation in the streamwise dircc~ion and also to\\,urtls thc wall. b) Ilitto
on
the
PRsure
face. Hcre also
an
illcrease of fluctuatio~?~ ton.;lrds the wall, but
a
decrease
5tre:lrn-\vise.
In
Rncral
a
s!ream\vise decrease, or incrcase
of
lluctustions
wilh
stre:tm\\.ise dccrc;~~~.
or
i~icrci~?;c
of
pasllre in tlic main flow, in accord;~ncc
with
turbitlcnce theory. After
Fur.o~er
[9.135].
)(~surernents m:~de
on
the outniost cylindrical vane section (Fig.
9.7.3)
shou-
21
pro-
'Lounced streamwise increasc
oi
iluctuatio!~~ on tile suction face, whcreas fluctui~tions
bd
10
decrease
on
the
pressurc lace. This is in agreement rvith tile da!ilpin- or c~citing
'at of turbl~lcnt disturbilr~ccs under a main
flow
having
n
pressure drop or n pscssule
in
thc streamrvise direction
[5.4].
Fig.
9.7
2.
U>!?arn~:
n:i..izilr.crncItt
of
111~
rcl.l[i\c
flo\\
clc~su
to
n
rlitlllcr vrinc
of
3
Francis
turbink
;it
R:~~lial
:iriJ
p,ltri.rn >ccii~>~!
in
clc\arinn (;ii>o\.e)
and
plan
(bclo\v)
with stritions oi~ncasuring
point
-
Ich
h,tnal
briic.
,In
.i
I~!I..II? irrc.\r;l
>:~~lilcc.
on
rhe suctlon
fact
S
(,ibovc),
ancl
on
thc.
prcssurc
fa(
l'(bclo\s
I.
-r,~t~on
6.
5.
4
,:cc.,>rd~~;g
to
,I!
r\ft~i.
Fu~.tr~cr.
[9.135].
~li~n
i~oundnry layer s~~ggcsts the presence of
a
laminar hound;~ry 1,1!,cr
ns
a
l1:isis
col;li,;~~-i.;o~~ with rhal
on
;I
pli~t~. Quite contrary,
on
the s~iclion Llce of thc
V;IIIC
il
,,
i~lcl-ci\sc
ill
thickness of rhc boundary layer in the direction
of
flow and also fsom
\v;lrds thc shroud is I;->und. Increasing distortions of the velocity profiles in the
tlo\v
)~~,.~~~n (Fig.
Y.7.4b)
is a re:lsnn for ;issuming tllc formation of \fortex layer\ in the
Qrc
J:lry layer.
(har3cte~-istics for this conliguratiotls are the drop of total and static pressure in the area
d,,flte~ axes,
as
deterinincd
ill
the course of investigating pressure variations
011
t!~e
I;ol, face
[9.135].
The steady pressure variations on the pressure face are quite con-
iui
,,,
From the nleaSUrelnentS it is seen how in the region of mnxinlum vane curvature
trjr.
8)
thc static pressI1I.e increases on approaching tllc vane surface.
-
flidClltly
time
inodilicatiens here are attributed to centrifugal forces acting
on
the
P'
Iricles, linkcd to the sharp vane curvature.
unsteady measurements:
Unsteady measurements permit the demonstration
of
the
P-
sence of the wakes of
28
stay vanes and
12
guide vanes at
141
rpln which corresponds
60.4
Hz
in the relative flow. Unequal spacing of the guide elements is respoilsiblt? I'or
;alcrfcrence
frequencies
up to
980
I-lz.
This leads at this measuring point to aver:ige
E~ocit~ fl~~ctuations of
18
%
of the mean velocity with directioilal fluctuations of approx-
m~tcly
12'.
lo
with measurements carried out on the axial turbine, velocity fluctuations in
the
mixed flo~v runner show a reduction of velocity fluctuations
011
the pressure
fact
for
rnclcrated general flow. On the suclion face the fluctuations are increased as the senera1
kw
is deceleratcd.
9.8.
Ji~vestigation on unsteady
flour
in a Francis turbine
by
R.
Gericll
and
G.
Molienkopf
9.8.1.
Introduction
The
bulk of flydroelectric energy demand now is covered with Francis turbines. By virtue
d
their
adjustable gates they are also predestined for the supply of pcak load. Un-
kxtunately this is usually connected with working away from the bep with a restilting
unsteady flow, as was mentioned
in
Cap.
9.2
and
5.5.5.
This can be predicted only
on
the
basis
of model tests.
hfollowing research has been carried out in
a
closcd test rig by
Mollenkopf[9.46]
and
Gericll [9.45]
the
Lab.
f. hydraulische Maschincn und Anlagen,
TU
Miinchen, West Ger~nany. The shell
%Yarn
of
the
turbine
with
its
individual operating points
is
shown
in
Fig.
9.8.1.
For
the
protection
dlhc
hot
filn~ probe used the research work has been carried out under cavitation free operation
4fd
with
a
head of
H
=
2
n.
The tests were made
on
a model with
D
=
272
mm
and
a
bep at
'11
=
768
rpm,
Q,,
=
1,275
m3/s
[9.45].
'4.2.
Unsteady
flow
and
turbulence level
at
runner exit
These
investigntions were carried out
by
means of
a
measuring device shown in
hg
9.8.3a, in which the semi-co~lductor probe shown in the drawing was substituted by
film anemometer. The rnetllod of holding the instrument enabled the probe to be
Fig.
9.8.1.
Efiiciency hiil diagram of 1:r;inc.i~ t\~rSinc invcstig:~:cd \vith
periodical
flt~ctu:ltions of total
;3rc.;;!:r.r
.
I
I-,;i'
at
the bnttui-ri ring beforc thc runner. plant.
V
:!
(scc Fig.
9.8.31,
fro111 thc cork
screw
c!r::fi
tube
\.ortci. (1Vhi:c ;trc;i.) 111
tlis
haichctl
arc;i
non
pcrioctici~l fliici~~;ltions
f~.om
the
draft
tube
vorlcs. rigkt limit
of
a
band wit11 whirl-frcc anti surgo-frcc operation of
an
appro~imate width
:lil
iil
thc hpccd axis direction:
.4t1
11
-
0.2.
Data oi thc
FT
at
Dcp:
Q,,
=
1.275
m3!s,
I],,
-
76,8
rpm,
11,
=
S6,7
rpln; test head
il
=
2
rn.
See
Go-icl~
[0.137].
Strongur
prcssurc
1luctuntic)ns ilppcar under
o\.ersatc (hntc11t.d area;.
Xi
point
32.
and
OII
ths
~vrtli
o!'the ilrart t~~hc inlet, thcj can ;~tt;~in see
(9.1331
lO?,,b
of
thz
head.
and
even
70%
of
Iicad in
rhc
mixing laycr of
111::
drzift tube vortex core.
turned about its
S~ZIII
during its rotation
with
the runncr. The electrical signal
was
transmitted
b!
me:.lns
of
mercury chambers.
Fig.
9.5.3
a
shows
the
distribution of the time-averaged relative velocity at the iildividual
points
of
dperation.
F
ioI11 this
it
is recognized that the
wakes
are not symmetric to
the
runnzr
vane.
They
sho~
a
nearly constant thicknebs past'the boundary lnyer on the
pressure face. The
vslocity distribution to~vards the suction face and especially the wake
past it varies stronglx nith the point of operation and is ~~sually thicker th;,!~ that
frorn
the pressurc face. Becnusc
of
the extrerne large angle
of
incidericc especi:~lly the operation-
rtl points
4
and
31
show periodical stall
of
the flow past th~ runner'".
Tliis is reconfirmed
by
thz
lcvc-1 of turbulence (Fig.
9.5.2
b).
In general this level increases
within the
ith he
to\snrds its cenrreIine except point
4.
As
an
avzraze mngni tudz of these
fluctuations
1
1
YO
of
the
mean
relative velocity were measured. Periodic fluctuations
of
relative velocitx due
t.~
the pitch
of
{he gates could
be
recognized
clearly at points of large
sate
opcfiing,
e.g.
poi11t
!6
with
6,5?6
of
IV.
")
Stu
also
tjotnote
').
392
fig9.8.2.
Relative vc!ocity and Icvcl of' turbule~lce past tl~c runncr vane closc
to
thc shroud
at
Ba2.m
rneasurcd
by
a
hot
film
probc from
Gerich
[9.133]
and
.llolle~ll;o,?f~[9.36].
Numbers
msrk
ktcst point on the hill diagram of
Fig
9.8.1.
:I) Relative vclocity
v>
circumfercncc
across
thc
pitch.
#Level
of isotropic (assumed)
turbulence.
Note probe pobition rcla~ivl: to runncr \.;lnc cxil
ctlgc.
hctuations of the relative ve!ocity due to the prcccssion and
frequency
of the llnstiihic
@fk
Screw part load vortex pil~t tile runner could bc observcd
in
the neigbbourliood
of
shroud
only
in
the extrcnie parl load points
4
and
11,
with small
and
irregular
h~litudcs. Ncnr to thc
hob
thcsc fluctuations have been registralcd; ;is expected.
with
amplitude.
393
d)
qos
0,lO
AY,/~
Q
0,lS
Fig.
9.8.3.
Dynamic mcasurcmcnt ol'total prcssurc in
a
rotating
:~nd
stationary framc of rcfcrcncc,
n)
Elevation
of turbine tcst section ~ith
a
rotating
Kulite scmi-conductor probc of quick response for
dyn;~niic
mcasLil.ciiicni of :otal pics~~~rc: in tlic rclativc, or absolutc systcm. b) Distribution of thc
refcrred time-averagcd total prcssurc minus tlic vapour pressure
in
the rcl:itivc flow at runncr oullct, close to shroud vs circumfercncc, in measuring
plane
VII/A,
sec a) nu~nbcr indicating tcst points acco~din~
to
Fig.
9.8.1.
z
13 vailcs. b) Distribution of the surpl~ls of tlie total pressureg,,, in the rclati\'e
system over the vapour pressure, rcfcrred to head
1:
at runncr outlet near tlic shroud measuring plane
VIIjA
(see
a).
Vane number
z
1,.
Detail
A:
probe orientation against outlet edge of runner vane. c) Measuring plane
V/1
to head
A
Y,/Y
across
the
runner inlet,
plane
V/1.
e) Turbulence level
Tu
acrg
9.8.
3.
j;luctuntions of
total
prcssure
at
runrlcr
exit
nest
fluctuations \\-ere ~neasrired in pl:lne
V!I/A.
For thc measurement thc nrl-angemcn~
rig.
9.8.3a
was crnployed with
a
transducer of
2
mn1 diameter for
thc
djrnamic meo-
'
cmel~t of total
head.
The curves in Fig.
9.8.3
b
show the distribution of tirne-averngcd
rtlr
.
,
aunllable-
A,,loeo~~s to the distl-ibution of tlie relative velocity a rathcr steep change of total
ssill.c in the \\lake past the pressure face is seen here, whercas in the wake due to the
F.
ruct,o~~
face the range of total pressure drop vs circumference becomes more broac! and
more on the individual point of operation. Only at operating point
31
a
rather
dcp
trougll of total pressure extends across the wake, probably duc to a vortex there.
brgc
1hc
fluctuation of total pressure reachcs
a
n~axi~num within the transicn
t
rcpion betweel;
t~
boundary layer or wake and thc main flow. The maximum of totnl pressure fluctua-
tions
due to gate position
16
occurs there at overgate.
9.8.4.
Unsteady abso?ote
flow
between gate
and
runner
This
unsteady flow is generated mai~lly by the two bladed cascades of runner al~d gate
apinst each other. Besides them fluctuations of lower frequency are expected clue
the
action of thc asymmetric and precessing flow field within the draft tube at past load
&o
propagating into the runner. In Fig. 9.8.3 d the fluctuations of total pressure during
&period of a runner \lane and the resulting turbuleilce level, are pl~tted \.ersus the span
d
the gates. They are
generated
by displacement of the runner biade. 'The strong increase
dthese fluctuations towards the shroud and with increased gate opening is caused by the
decreasing
distance bctween gate exit
and
runner. Hence the pulsations originating froin
cutting of the wakes from the gates by the runner vanes become
enhanced.
(Note that the
turbulence level increases towards both the ~valls, Fig. 9.8.3e.)
fhc
precessing cork
screw
vortex under
part
load
within
the
draft
t~rbc.
produces also periodic:tl
nrialions
of
flow both before
rind
past the runner. The conscqucnces are more or less strong
bctuations
of pressure and velocity propagating
also
into
the
spiral
casing
and
ihe
penstock.
A
gowlh
of
the
wake
reglon past the hub and
an
incrcase of the vortex core dialnctcl
nith
dzcrzaslng
bw
and
increasing
moment
of momentum past the runner causes an
increase
of
pressure
fluciua-
*ns
due
to
the
draft tube vortex
with
increasing distance
frnrn
the
bep.
Af
a
certain
gate opening, they belong to a draft tube
vor
~cx
cf
a
certain form and origin,
which
ends
as
well on
the
runner
design. So, c.g.
at
part load
witk
i!b
i:oi-k
screw vortex, the origin
may
kon
the
hub,
sometimes
deep
in
the vaned zonc of thc runner as
was
dcscribcd
by
S~~l~l~r~inier,
Gerich
Koabe
I9.831,
or
it
may
be on the top of the hub cone as was described
by
Ilcnrjl
[10.49].
The
9rat
diversity
of torch-like, rope-like, twin,
axial,
helicoidal, steadily precessing
and
unsteadily
lroccssing
draft
tube vortices as
a
function of speed and discharge of
the
working
point
at
a
certain
has bccn shown
by
Schlenlr~ic~r
[5.18]
for
a
given Francis turbine.
Usually only the fluctuations due to the draft tube vortex could be recognized from the
wsuring signals of the static or the total pressure or the
velocity.
The broad spectrum
4aII
the disturbances ;tppenring could not be determined from the apparently stocllas-
measuring signals. As all the signals have been combined with considerable noise,
a
andysis of frequencies has been carried out for the most important si_enals.
I9.1331
has registered several power spectra of pressure fluctuations at the head
"er
and
within the intcrspace between runner and gate by means of a correlator and
transformer. Besides the expected resonance frequencies
/.
of pressure surges
I
Qliii
(2
-j
-
i;
i'
'>
-)I
/-.Oj,~jA
=0,2i5
;
1
!,@
,%,.",,
./.
i.
,g.
'J.S.4.
~'II~:IIOI~:CI~~
duc to
/.
J,OL
I
_
-1.
,/'.
[:I<
~~CCC'SS~II~
cork
SCCCW
VO~~CX
+----
..
*--
?,.*F.
--
--
-
9.L:s
in
rile
Jr:~ft
tube.
;I)
Distribution
-.y
-
r
4
As,.,
4
A.
'1
1
/,
I-
of \~~rplus
(>f
:ol;!l cricrgy
in
the
954
;
/'
/rM
-
/
rclati\.c syhtc.!?i o\cr thc vapour
/
/./
,
<&
p:.cs';~xciii
!.c'~;I!~oII
to tllc I1c;ld
0,oz
-
----
--.
/
4
1'
/
111
tlic n~c~.i:Ji;:;: :lcroc,j the dr:1<1
./
/
/
-.------*
.
::SS
tai,.:
scction. pl;cr~c
1V.
ill
a
dis-
:
/
/'
/--/-----
/-%-r
;
j/////,
r;l!icc:
1.5
1)
tso~n runner exit.
~-
i
/
0
\\;ill~t~m. ca.!i~;ilion in the
.
--
vk>rtcx
core;
-
--
\virh cnvita-
I,@
I,\
I,
:
1,3
1,4
1,s
1.6
---*
"-1
-
tion thcrc; ;trrow:
instants-
lwt
IIC'OLIS
p~sitic)~:
o!'
vortex axis.
b)
Rccdr~lcd ctirnc~si,~nlcsj tot;~l prcssurc vs tirrlc ulon? circunifcrcncc
Fo.)r
a
ht:~tioli;~ry
probe,
at
test
pci~t
3:
I!c~)
ar,d
37
~ri~iir). (Tcst poir,t
scc
I.'i:.
9.S.1.).
.c
I)I-CCCS\;lti11
perioti 01 tile
VOS:~Y
core,
TL
pc:iotl
of
s
runner rf\-olution.
I1
=
2
rn,
~~icusi~ring pianc
Ill
(172.
9.5.31.
c)
Dis~ril)ution
of
the
~YAS~SC
I?~tcti!aiion~
sf
thc total prcssurt rcfcrrcd
tu
~hc head, sj.nihr~l?~l~~
\\.ill]
;I
pl-eccssing dt-;ift
rut;;.
vvr1c.r 1ie3r thc dr;iSt ::~bc \\;all
in
relation
to ilic differ~:nt v:ilr~cs
of
~!ic
r;~:i,,
of
unit speed
(/I:,)
to
ihr
unit
spc:d
with zcrc>y\\-!)is1
(I],,
,,.,-,),
:.lrcorJing
to
111s
sketch.
:it
cliFkrcnt r::tios
I?S:IIC
33te
il~igl~
1,
to
thc
garc
;tngls
of
the
-
.
--
line crossii~g thc hcp. (See shctcli.) Ai:.cl.r
Gcv-ic-11
[9.133].
ivi:i~
j;
(runner).
.I.-
(runner
blade),
f;
(draft
tube
vorte:tl
and
-f,
(vibration of casing)
a
~nu!:itudc
of
other
resonance
freq~~encies
exists.
Fro111
the
sixctrii
of
t!ie
tut;ll pressure
within
the
pen st^^!;
rlnd
in
thc
nci~hbourhood
of
the
drafc
tubc
wall.
a
stxienlent
could
be
tnade
nbou!
thc
reiction
of
tile
penslock
on
the
pressure fluctuatior~s
oi
[lie
draft tube
compnring
the
"power"
of
the
signals nizasured
on