the erosion of steam turbine blades - illinois: ideals home
TRANSCRIPT
4- ' ^
f4
UNIVERSITY OF ILLINOIS
LIBRARY
Class Book Volume
M rlO-20M
4-
f -f- f 4- 4 4
-4
4
r ^4 ^
'7- ^.
X.
% -f
4 4-
1^ f
t -h
'^ tI- -v* ^j-
1^ -f- f < f. * 4
- > * ^ f
•^^ f -^^ ^ ^ ^
+- sf- 4 ^ '
^ ^ * ^ -f- -i
^ '
'
f -t- -f- -4 «*•
^ *- 4 4-
h i^J'
r 4 >f- ^ -y.
f t -f
'
-4^- ^- ^-
^ - - fv 1^ >^ ^ji^^
r -'k =Jf* 7- f
, ,^ -f *-
^ ^ -fc ^ ^ ^ ^ ^ ^-
^ ^- ^ t- ^ ^ 4^ ^ ^- ^
-f. „t
--#^ 4 -i^ 4- ^-
^ - - - - f-
>* + ^ V +
-+ 'i- * 4 ^ ^i.-
TIIK I:R<)S1<)N of STILV.M TURI5IM: IJI.ADKS
nv
CIIARLKS IIKXRV .lACOiiHENROY IvENNKTll MURDUCKl
THESIS FOR THE DEGREE OF BACHELOR OF SCIENCE
liV MECHANICAL ENGINEERING
IIV THE
COLLEGE OF ENGINEERING
OF THE
UNIVERSITY OF ILLINOIS
JUNE, 1910
UNIVERSITY OF ILLINOIS
May. 3.1 i9ao
THIS IS TO CERTIFY TH.AT THE THESIS PREPARED UNDER MY SUPERVISION BY
Charl.es Henry Jacobsen and Roy Kann.e.tli....Mur.duck
ENTITLED Tlie Erosion of. Steam T.ur.bine. Blades.
IS APPROVED BY ME AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE
DEGREE OF Bachelor of Science in....Iy;;..ecM.n^^ Engin.e.e.r.lng
Instructor in Charge.
APPROVED:
HEAD OF DEPARTMENT OF Mechanical En^.i.n.e..e.r..i..n4..
167534
TABLE OP COTTTl'iNTS
Page
.
INTRODUCTION 1
THEORY OF THE FLOW OF STEAM. .
Determination of the steam velocity 2
Design of the nozzle 4
DESCRIPTION OF APPARATUS.
Description of the blades. 7
Description of the testing apparatus. ... 7
METHOD
.
Method of conducting tests 12
DISCUSSION.
Discussion of results 14
RESULTS
.
Data sheets. 15 to 30
Result sheet. , 31
Ciirves. 32 and 53
LIST OP ILLUSTRATIONS.
Fig, 1 Diverging^ nozzle. , 6
" 2 Converging nozzle 6
" 3 Cross-secion of the apparatus. . . 8
" 4 Photograph of the apparatus. ... 9
" 5 Diaphragm at the exit side 10
" 6 Westinghouse Parsons turhine "blades. 11
" 7 Cross-section of a blade 12
1.
THi^ OP riTEAM TUKBIN.. BLAD^sS.
Tntroduc tior
.
Unlike the steam engine the steam turbine is not subject
to much wear. In the steam turbine, practically the only we:-r is
caused by the cutting action of the steam upon the blades. In most
cases Lhis wear is not serious, but it may cause an increase in
the steam consumption.
In 1904 Mr. Francis Hodgkinson presented a paper before
The American Society of Mechanical Engineers, in which he gave the
results of some experiments concerning the erosion of steam turbine
blades. In these experiments hard drawn Delta metal blades were
exposed to the action of steam for 128 hours. Two nozzles were used
with velocities of approximately 2900 feet per second and 600 feet
per second, the wear being much greater in the case of the high
velocity than with the low velocity.
In 1909 Messers Ekblaw and Wolf performed a series of ex-
periments on Keir Turbine buckets comparing the action of wet and
dry steam at different velocities. It was found that the wear was
greater in case of the wet steam than with the dry steam. These
tests also confirmed Mr. Hodgkinson' s statement concerning the
effect of different velocities.
Aside from these there is an absence of data concerning
the erosive action of steam upon turbine blades.
To determine, then, the action of steam at practically
commercial quality, at different velocities upon different blades
is the object of this thesis.
Theory of the Flow of Steam
Steam in dropping from a region of pressure to a
region of lower pressure {^through a channel or no^izle, and, if the
expansion is adiahatic , the drop in heat contents is used in in-
creasing the kinetic energy of the jet. This heat drop is repie-
sented "by (H;^ - H^), ( where = the total heat in one pound of
steam at pressure p^^, and Ho = the heat in one pound of steam at
pressure p^.). It may be assumed that the velocity of the steam at
entrance is so small as to he neglected.
If Cq represents the theoretical velocity with which the
steam leaves the nozzle, then the kinetic energy of one pound of
steam is Gq/^& foot pounds. This ir the work equivalent of the
heat given up during expansion. One B. T. U. is equivalent to
778 foot pounds of work , or
Co/-g = 778(H3_ - Hg)
Letting g represent 32.16 this equation reduces to
Go = V 2g X 778(Hi -
or Go = 223.7 V(Hl - K^) — — — — (l)
The heat drop for a certain adiahatic expansion may he readily
obtained from a Heat Chart, or by getting the total heat in one
pound of steam at pressure p^, correcting for quality;
^1 " ^1*
Then subtract from this the total heat in one pound of steam at
pressure p,^, correcting for quality;
The quality of the steam at the initial pressure was obtained by
use of a throttling calorimeter. The formula used was
XI = (H^ - q^4. 0.48(Tq - Tg^))/r^— — (2)
6.
\Vhero H^^ ~ Total heat in one pound of steam at atmospheric pressan
q-j^- Heat of the liquid at pressure p^.
r^ = Heat of vaporization at pressure p^.
- Temperature of the steam in the calorimeter in
degrees Pahr.
= Temperature of the steam at atmospheric pressure in
degrees Pahr.
0»48 = Specific heat of the steam.
The quality of the steam after expansion was obtained by using the
relation;
e *{rjX^)/Ti = e +(rixi)/Ti - (3)
or , with adiabatic expansion there is no change of entropy,
where 6 = Entropy of the liquid .
(x-j^ ^l)Al = Entropyof vaporization, corrected for quality.
All terms are known except X2 which can be obtained as above
explained from equation (3) .
Having the total heat drop due to the adiabatic expansion
the theoretical velocity can be obtained as explained above, from
equation (1). But there a certain amount of the heat drop lost
due to the friction of the steam in the nozzle*
Let G-j^ = The actual velocity.
ra = The fraction of the heat drop lost due to the friction in
the nozzle.
Then from the energy formula -
G^/2g = (l-m)(Cy2g)
or = C^V 1 - in
Taking the friction loss m. as 0.1 , which is an average value
( from the I. C. S. pamphlet "Steam Turbines" ), then
4.
\/ I - m = 0.95 approximately, oi a lofes of velocity
of 5^ , or
= 0.95 Co
Ol = 0.95 . 223.7 " )
= 212.5 V(H]L " ^^2^"
- — — — —Design of the Nozzle.- The factors entering into the
design of the proper cross-section of a nozzle at any point of
its length, are , the pressure, the velocity of the jet, the
weight of the steamflowing per second, and the specific volume of
the steam.
Let G = The numher of pounds of steam flowing per second.
G]L = The velocity of thesteam in feet per second.
V = Thevolume of one pound of steam in cuhic feet.
F = The area of cross-section in sq. ft,
f = The area of cross-section in sq.in.
Then F = G v
or p = Gv/C;j^
or f =r l44Gv/ C^^ — — — — — — (f)
It is found that the cross-section is smallest where the
pressure is 0.57 of the initial pressure. Thus if the initial
pressure is 100 pounds^ the pressure will be 57 pounds at the
smallest cross-section or the throat. In the preceding statement
it is assumed that the final pressure is less than ^7f of the
initial pressure. If this were not the case— that is, if the drop
in pressure is small, or if the final pressure is greater than
57^1 of the initial pressure the walls of the nozzle should
converge
.
In "brief, if the final pressure pg is less than 0.57 p^^
5.
the nozzle should have a throat, then diverge, ns shov/r^ ir fig. (1)^
but if Lhu final prcissuro po is greater than 0,57 uho nozzle
should convergu, as shown in fig. (2).
The nozzle '.vas horijd out with a uniform taper for the
portion in which expansion takes place. The taper was approximately
1 in 10 as shown in fig.(l). The area of cross-section of the
throat was determined from formula (5), "by using pressure p' equal
to 0.57 p;j^, and the area of the outlet end by using the pressure
P2« If the nozzle is too short the taper will be too great and
the jet will spread. If the nozzle is too long the loss due to
friction will be increased. The usual taper is about a 10 degree
angle
.
Let = The diameter in inches at the large end .
b = The diameter in inches at the throat,
c = The angle between the sides of the nozzle .
L = The length of the nozzle in inches.
L = (0 - b)/2 tan(c/2).
And if c = 10 degrees,
L = 5.715(0 -b).
Description of Apparatus.
Apparatus Tested,- The blades wure of the type used in the
Westinghouse Parsons pressure turbine. A sketch showing a cross
section is shown by Fig. 7, pagefE. There were four blades,
each of different composition, hnov.'u oaly to the manufacturers,
the Westinghouse Machine Go.
Apparatus for Testing,- Two pressure chambers of 6" pipe were
bolted together with a blank flange between them. The steam nozzl
extended through this flange and was held in place by a flange
and strap over it bolted to the blank flange. One chamber was
connected through a reduction to a 1" steam pipe from the steam
main. The other chamber connected with a Worthington surface
condenser through a 4" pipe.
The blade holder was bolted to the exhaust side of the
blank flange, to one side of the nozzle. The blade fitted into a
split cylinder, which, in turn, fitted into a hole in the holder,
where it was secured in the proper position by three set screws.
The nozzle for the first set of tests- at the lo7/est vel-
ocity of steam was a straight nozzle. The nozzlo used in the re-
mainder of the tests was tapered one in ten, being a diverging
nozzle.
Two Crosby steam gages were used to obtain the steam
pressure. They were connected to the two chambers, the one on the
exhaust chamber reading' both vacuum and pressure.
Each chamber was provided with a drip cock.
The throttling calorliaQtor was connected to the high
pressui'e chamber through a l/r" sampling nipple.
Method Ox Conducting Test.
Four ten hour tests were run on each of the four blades
The steam velocity was varied, while all other conditions were
maintained as near constant as possible. The velocities used,
were approximately 1400, 2100, £800 and 3400 feet per second,
the latter being the maximum attainable under the conditions.
The pressures, initial and exhaust, used to obtain the desired
steam velocities were taken from Prof. 0. G. Thomas' "Steam
Turbines".
The quality of the steam was maintained as high and as
constant as possible. Steam was taken from the top of the main
and the piping was lagged.
The blade on which the test was to be conducted, was
thoroughly cleaned, dried in a desiccator and weighed on a sensi
tive balance. It was then fastened on the blank flange close to
the end of the nozzle, in such a position, that the steam flowin
through the nozzle would strike it on a tangent to the inside
short curve, as shown in figure 7.
13.
After connecting up the apparatus, ctearn war; turned on and
the valves on the high and low pressure sides adjusted to give the
proper initial and exhaust pressures to give the desired steam
velocity
.
A condenser vvas used to condense the steam and to obtain
a vacuUiTi on the exhaust side when needed. A throttling calorimeter
was used to obtain the quality of the steam.
The following readings were taken every thirty minutes;
Initial steam pressure.
Exhaust steam pressure.
Temperature of steam in calorimeter.
Temperature of steam in exhaust chamber.
Three barometer readings were taken during each test.
All instruments used were calibrated and readings corrected
At the end of the test, the blade was removed, allowed
to cool, cleaned, dried in a desiccator and weighed. The amount
of erosion was obtained by taking the difference between the
initial and final weights.
14,
Discassion of Results.
The results of the series of experiments given on pages
15, to 31, show rather clearly the action of steam at different
velocities upon turbine blades of different materials. Enough data
was obtained to make a fair comparison between the action of the
steam at high and low velocities. In all cases there was an increase
of the wear with an increase of velocity. In the two higher vel-
ocities, in all cases, the quality of the steam ran up slightly,
which would cause the wear to be less than if the quality had been
more constant
.
With blade No.(l) subject to the action of steam at a vel-
ocity of 1400 feet per second for 10 hours the wear was 0.002 grams,
and at a velocity of 3350 feet per second the wear was 0.0041 grams,
or an increase of 0.0021 grams , or 105/^ with an increase of 1950
feet per second in the velocity.
Taking the highest velocity with blade No. (4), the wear
was 0.0035 grams, but in these tests the velocity of the steam is
absolute, while in the turbine the effective velocity is the veloc-
ity relative to the blades. The blades tested belong to the Parsons
type of pressure turbine. The relative velocity in a turbine of
this type is much less than that used in these tests, being less
than 1000 feet per second. The curves on pages 32 and 33 show the
relative hardness of the different blades.
1 Pehr Z6
S /'^ ZSI
Z/O1
1-92,3 Z^S
36. <S" ^ 23 J
/-s:<9 23a
35" .S" 23 / S-/3 / fc>, /^ r C-.cz.
22.3-4
//TO' 3'<?. ^ 23/
a, /s
3/TS /6 6 2. -52.
/^.^ a? pCry
5
^.'/ t5"36.^ / 5: 5 S3a.
^8.S 2-S8
6>/5-. / zsS MOO
3 9. 6~ S.B'f-
7.'/^'
3<?. ^ 2.5'^
IF. 3 23/
Z3/. 7- ZS-o. 1
Mar. I
/ o : 3o 37. sr 2. 'O 2^0
II IOC 2.
1Ji : 3o
/R : cc2.^0
:3c /<?./ ^/<?
J : (Po2.^ /
/Y./^ S'Q
/:3c 33. 3 2./S'0. 001
£:oe? 3/. / 2/7- 2^4-2-3.^ 1-^
z:3o2^ /
3: 30 /r. ^ / p
floo
2^, T
5:CO
£>': ffc3c?. J-
2.-4- 3- S^O
6 - ^ c _ I3&T
7: CO >^.^
7: 3 c 3r. 4'
6:€c /^.^
8:30
-zti A *
^ n o
^ : O G / 3. 7' i-^ ^3
S 3 <:> r- / , o7
/<D : / 57 i)~
/C ,3 <=/A. / ^ ~^ KT 23. G:,-?"^^
// . O c:>^ 1 ,
\~f f t csJ ' • o
23. ^^T'P
// ,'^ fS" <^ Qj / T r/ O / o /O-OOIT"
/^ .' £> <i> /2 3 9_ 29. ^2-
/a /^<=> /iT. P
' ' / o /(/ ' ' ^
^ a /cP. 5
n J ' f > o ^ 5> 9
^ . :3'=> "<b / J / V) • ^ ^::> ^
- O cS>^ / (
—
'
5'.'3' <=> •^CL. O. . 9<?. 6
•9-: o<=>/ <J/ /T- ^ ^ y "o - ^3, ^
^ o / <o ' /
-5^.
' <o o
-5". C
3R.^ 2cr/
'o : 3o 39. S' 3
39.1
2:?T
Mar. 3
3P.i~ /G.o 2.^ 7=
S>:/^, o
& : 6^' Z3 TI
£> SS' Sr. s'
^ . -ys' 3 7-. J' 23^ 20
.
932/
//p .20, 5S<3<?
37^5'C, OO / ^
23 /
3^.r / 5'.6
" 37. /Y.O
'/<$.o
y
:
37.S- /6.9
y<P.o
2 3-3
-4/.?'
/^./
/^.^
/r-.o
/^o
i'^'^G -
.
1 1 1
3r.s. ?1
s-3 2-^4 22<?
g .
S2. 23 /
1
SI 2
^ . /^'
9 : ^•b' So
^8 57// (d , ^ to lZ-
^8 37/0. 00 22
4y S.3
S'.S' £^ Y 22.^
^3 230
/^/^^' S2L s.s- 22.'^
/ , / ^' ^a. 2^9 230
^.9 23o
e.:/ a' ss £.5-PU
S'o ^/ 22
^4. /
^//^ ^9 -2 S'o 22 8
^3 2So 230
s.^
3-3 2-^^ 2B.^
1 1 1
20.
Mar r
t:4S ^9 22 £^
2.ss 2 30
£> S-Q SO
/o , z^" S-Q S.a
^6
// .^/ SG
///^^' ST (^,S
,' /-i' ^9 So
.'^6 2.2. G
/ ./y
s:o
So 2.S0
3;/y
^:/^' y3 S3^S^."^
So
1
So,
^S/,(^^
1
Univkrsitv liJJNOJS
T
so
So ^•3
^3 5: 3
5"0 ^- T ^S'S
^:^-> SO ^'S2 3. ^T'S,?
/ £>, S3 23. ^7^/6
/I' z^"^' sz ^,0 2.sr(i?, 002.'^
/y
:
^/ ^~ o-.23.24^
/y . J>\5' 5a S £^2 /
.'£.5" ^9 ^-9 /
^^ B^o
^^ ^SS3 s:a 2^3
^/ s^ s /Cc^'pute:d Results
S3
90.7SJ52
^7
So ^'7 ^-=?^
S6 ^.^
S/.2.1
s-.o,
23^.81
s
S.8
///^ 5^ 239
S 5^ .S^3
2S6' 232. 20.^306
^8 S.-3So. 93.00
^.^ ^-^-^0. oa / ^
3sr S.7
///-^^ S3 S.S'--- /^'^
/^//£> s/ S/
S3 ^.3 2 S'G 23?-
So
^:/<^ ^'<?
^. P 2 S~0 23o9s. a
S.'/o 2./C>
z^a <?
S3 23 3
ss-
-5: /if ^3 ^
' 1 ! i t
S3-?-
3
I
y
83.^ //.<? /3o
/ 8(o,o //.cP1
78. o
So.
3
2^e
/ /.a
/£},/ //.<f/6. 2 (£>27'
//.' /<P J2./ 2T7-
///^ <S/.s
: /o
/ ^ .
/./£> /^. ^
1 : ^e> /2. S-
^ : / o //.<? ^^(^
//•cP
3/^o S3,^
//c
^:^c
// 3 ^^(^
<PB.3 //
24.
UNIVr,RSIT>
I
10
/o. 8 a3
IS 77./ / /.3 273 2^, 2_
S>/. s /o.s 263
2Z^t 262. 76. 2.5-6 9
2 7^-O. OO 2-=? .
//./s- 2 7^- 23. 30 3
//.'^^ ^ 7
/Z: 2 7-3
12.: ^s' <SS~, / 27^
/ : IS' //. c^> ^75-
/
^^O
3: rr. / 2.73 2^<?
3: ^s' TT. / 27<z 2^S> 27Vj^
4: /s-c
2" 73 2 6-<f
77- / 3-7^
S'. / 6~~ aV3> 2"5"z
/ o.(^
//-/,
2 73,
5
t
UNIVt:RSITV OF llJJNOIS
n Data try DA) Aor. 7 l^^]o
i^. Foh.r
7 • 30 S /. 5
3//, 6)
"7'/'
/3. ^9 2. YT
y. 3o 2.r^ ^ / «o^ cj cr
/o:oo1
'^3.^^ST
/o:5o 7<o, o /3. o /9cf
/ / :oo
//. 3o TT. O /O.^ 2 7(S
<r/, 6 S.O
/^:3o (S3. S ^ 3.0
/ oo
/:3o
£:oo rs.o
£ :3o
3 :oo a^o /9o
3:30 VG.o / /.o 2 77 2Z6o
^: oo S/,s ^.^ 2 7c^ / S"^
2-7S'
s: oo 2 r9 / 33
s : 3o T/. O ^73
27T.S' ^^7: 5":
72.f
cf ; OS re /o,$
c- 3S & 3 /A c? / 9
3: OS' S3 /0,§' 5"3
^3 //. c? 2 7-9
/ O ' os~ ^ 7 —7-^ 2 6/
/ 2 7-6t—' y/
id. ^//, <7»5 // 3
0.
/ / ' 5 z—
// ^5 2: 7-6
3'. T 273 237
/ .' 0:5^ /A^ 2 76 ^:36
^7^ 2.3 d
: ^§ 22l^
^^ 2TY ^6/
3:05- /Ac? 2YJS
3: 5S 27-6 /6 7 9
: oS ^3 ^7k5- 27^0
6~: o5~ 27Z
^~'3s
^ .' OS' /A§ ZYT2 7-5" 2 3cP. /
Umvi.rsi7'v of Illinois
GREEL OF B,->
r
TEST NO. ;3 OATA SHETT t^A /^/or-. 6. 1910
No. T, me
---
T: 3o 2/ 2 2-=9'^~
1 3^, 2-^3Blade: r^fo 2Series ^
TniTi".--;. vv:^r.,t ases o'
Jifterenc. O.OO^/^
<S':3o ^37. 22 ZL^'7
/ 3T, 2 / 26<?
5 20.
26Zo / 0:oo y 3,5-8 /9 2^7-
/0.-3O 20 269
23 <P 2.7- le Baroruer ^9. /32.
.e.Afrnos. Pre55 3 Los/^;,2 7"/
^37.0 Z.T/
/ 32.0 '2LJ £Ti
/'oo /2 7.0 a/
?3 /'3o y^-^. 2 2 Po
a; voComputed Results
2T/QuaVrty Steom ... 59,-??
3:00
^:3o ^30.0 2";^/. IOC . , 3rsom 33 5"0
^' 00 2 "73
19 -4-: 3o ^ /
29
S: 3 a 137.0
2/.-^ 29/
28.
UNivr.Rsn >
I
M Data Smft V
— .
2 / 2<Pt 2.5-7
-
1^1.
o
20 260
1 3T.8 2/ 3
2/ ^SB.
/^I.O 2/ ^ ^<f>
/37c? ^/ / . <£r 17
/ 37 O 2/ T 2.5T1 / > . r' /37 c?
i / t ^ .//.^^ /37cP05. Press /-^f.-f
/ 2 ,' / / 37c:P ^/
/ -3 , .
/.o Z/
/;/ :r .2/ ^^ T
2a
/3S-.0Cor
. , ..asuLT5
zo
___ 2379
/^^.^ 20 32cfo
2o1
1
1
t
1
^^^^^ a/ j
/34.7,
i 1 •
20.
;5 /V^A 3/ »
/ o : a 5"
/
3J / ' C3 .— C 'f-j ^<r 1. —
—
/ / c- / nY ^ .Q /
2.3. o ^6 <?
/ :p> \~ c~/ ^- . o / Z5n1 ^—^ l_/
7* >^ .?= JP "7-f 1
- 2 3. 6 2."
/ 2- f '^7 f\ -—
«
a6cp- -2 9.62^ Trs.Hy
z* 22 .o
''3 7
' *^— o
^'Z • «— I— "> ^
^'t"'
_ 2^ z 3
3:3^0
22. s-
^-.'^^ ^^^^
a2. 5"
1
1
1
Jo!
///^/-. /,
23 .o 23«f
2-6 3
8: 3Q /a/ 22. (3
^:ao 22. 2-6 7-
_ 2^. S»23-2.
/o :oo 22. o 2^5, 52./ 7
/o :3o /2 3 2 3, ^
/ / oo ^3. O 23 3 £± C 1
- Z3. S-^^
: 3 o 2-73£-0 /^^
/Z.oo Z 3. O 2S'cP 27--<t
3o 3o 27--^
/ - oo ^73 1
j
//3o ^3. O 27^2.i
21 oo 2:3.
o
27/
2,3.
a
2.33_ 9^.6
3 :oo 2*7/
\
23. <3 2©3 \
// z\
2ZzS
1 :
263./\
1
1
31.
Result Shuet
Kxp. No. Blade No. Avge
.
mFt • per
1 1 1400
2 2 1387
3 3 1405
4 4 1375
5 1 1975
6 2 1990
7 3
8 4 \j \j \j
9 1 2760
10 2 2745
11 5 2760
12 4 2750
13 1 3350
14 2 3280
15 3 3340
16 4 3360
1. Avge, Qual. Wear.of i n
steam Gram
90.4 0.002
98.1 0.001
98.6 0,0017
98.5 0.0013
98.9 0. 0022
99.0 0.0012
99.1 0.0022
98.8 0.0018
99.5 0.0027
99.5 0.0024
99.6 0.0031
99.4 0.0022
99.4 0.0041
99.4 0.0035
99.5 0.0042
99.6 0.0035
4^. ^ 4. ^
^N i^ # ^ 4^ ^ ^
1^ f
^ ^ % ^ ^ ^ ^ 4-- ^ r ^ 4 ^^'4^" ^ 4 ^ ^-Hfe^ . ^ 4 ^ 4^
^^s^' ''-^s^ •^ifc' '-4-
^ ^N^ ^ ^ ^ t 4. 4, ^ ^-^^ ^
^ ^ % .^fe ^ ^ ^ #f ^
-^4
4^
4
^ ^ 4^ ^
^ ^ ^ -it ^ ^ *
+-4^' ^
f
-f
4 4
^ f
4
4^
4-
4 4 ^ ^ 4- 4 i)^ -f-