JPRS Report-: Science & Technology

JPRS-JST-91-005
31 JANUARY 1991
FOREIGN
flBBBl BROADCAST
■■■■■I INFORMATION
SERVICE
JPRS Report—
DISTRIBUTION STATEEIMTX"
Approved for pir'ic release;
Distribution Unlimited
Science &
Technology
Japan
BUG QUALITY INSPECTED 3
REPRODUCED BY
U.S. DEPARTMENT OF COMMERCE
NATIONAL TECHNICAL INFORMATION SERVICE —
SPRINGFIELD, VA. 22161 <J1
JPRS-JST-91-005
31 JANUARY 1991
SCIENCE & TECHNOLOGY

JAPAN
CONTENTS
SCIENCE & TECHNOLOGY POLICY
Japanese Government's FY90 Nuclear Power Budget Proposal

[GENSHIRYOKU SANGYO SHIMBUN, 18 Jan 90] .. . 1
ADVANCED MATERIALS
Gas Pressure-Sintered (MoSi2-SiC)/TiAl FGM

[Junzo Fujioka; KEISHA KINO ZAIRYO KENKYUKAI, 22 May 90].. 20
Fabrication Process, Evaluation of FGMs

[Ryuzo Watanabe; SHIN SOZAI 21 SEKI FORUM, 24 Apr 90]..... 32
Thermally Resistant FGMs by CVD Method

[Chihiro Kawai, et al.; SHIN SOZAI 21 SEKI FORUM,
24 Apr 90] 42
AEROSPACE
Developing Independent Manned Space Technologies

[REPORT OF MANNED SPACE ACTIVITIES KEY TECHNOLOGIES
RESEARCH WORKING GROUP, Mar 90] 52
- a -
COMPUTERS
.RISC Processor Array for Artificial Neural Networks

[Atsunobu Hiraiwa, Shigeru Kurose, et al.; HEIRETSU
SHORI SHIMPOJIUMU JSPP '90, May 90] 64
MARINE TECHNOLOGY
Information Processing and Equipment Arrangement in the Pressure

Hull of 'Shinkai 6500'
[Shuichiro Hamaguchi, Itsuro Maeda, et al.; KAIYO
KAGAKU GIJUTSU SENTA SHIKEN KENKYU HOKOKU, Mar 90] 76
Design and Construction of Spherical Pressure Hull of

'Shinkai 6500'
[Shinichi Takagawa, Daisuke Kiuchi, et al.; KAIYO
KAGAKU GIJUTSU SENTA SHIKEN KENKYU HOKOKU, Mar 90]'... 95
b -
SCIENCE & TECHNOLOGY POLICY
Japanese Government's FY90 Nuclear Power Budget Proposal
90CF0359A Tokyo GENSHIRYOKU SANGYO SHIMBUN in Japanese 18 Jan 90 pp 4-5
[Text] Government Nuclear Power Budget Proposal in,Tabular Form

rH CT\ rH CO <f CM m co
00 H CO O r-» oo m m
O ON rH r- CM t-> 00
in co oo CM CO CM rH CO rH in r-< m
A #■ » A *
n N M
* *l
vO CM -* VO 00 CM CO H oo m
rH CM rH rH rH .
* * *
a
to MO N m r-~ co o
o rH O CM
u
ON cn ON o m ON CM CO 00 CM <■
O
CM
O rH
CM oo >N4-I
•H
CM <r CM in CO CM o ON CM ON O ON CO
•* «i * ON •W CJ
•H 3
HNH vO ON o oo 00 vO V4
01
04
* CM * 00!
G
ai
Ti
* rH * * •H
u Ü 0)
c •H O
o oo 10 60 CO C
•H o
■rl
I
o
4J *? u : c •rl 01 C 14H O
•H 01 3
to O
0) VO
H
1-5
X
u
a •rt X rH ■H CJ
rl rl 4-1 CJ Xrl
XI 3 M 4J I
I4J rH U iH 4-1 CO C 0 /*. rl 0) ►> 111
U/^
to 14H 4J H CO CO X CO 4) O 4J psj 0) 01 4J C Ü >, 01 U Fn
to c ^ O
rl 0) 4J ß H T4 CJ H X C OJ •H 4J C CO W
rl a O , U O H 4J 01 tO rl CJ tO co H e *H MH X <-i CD -H OJ CJ
CO o •H 14H ! 01 Irl 1-3 a C 0) C OJ 3 3
Xi
01
t-i
o 4-1 O i x u
00 0) 14H
OJ
4-1
OJ
e>
t-l -H
x ß
4J ß, 14H
01 H
rl *-* I
-
cfl w
00
a to
10 Q) I*H 60 to p
CO' C 01 ""
tO CD rl'
OJ CJ ' -H a o c
•o
4J rH
I PS o
o
CM I
c CO
3 60 l-l
0 0> CO
CJ rH 01
o ^ >N
VO CM
rl oj IV •rl
0) (U u
1-1 l-l >
C ri u u 01 01
01 H c 0!
O M 13
iH
■—• (0
< c 4J
H O to
CO iH CO
•—> u
to
>NZ C CO u
o 01 to 0> 0>
c •• E O > o
 a 00 CM o a
>> c o oo r» CO O to
t>0 0) 60 4-1 oo oo rl VO
o >. 01 o m 0) rH
■-I O 60 #i #>
a i
o c ON "O co r>»
d o >i 3 co ON to U
X -H * oi oi oo
0 iH >.ZH
01 .-I
H
"ä T3
01
c 
o 00 CM m
0) T3 <-t vT
CJ •• 3 OS CO
C 4J X 4J n p.
0) -H c CM <r
•H C ON 3 CO C3>
O D 00 o «
es
>N
01) OI
rl 4J
0> 3
X c 4J
CO w •rH
H 4-1
c u t/l
o
eo a
•H
•H M
4J
tfl 4J X
N < y
•H u
c
CO
e CO
CO 01
00 0. to
u cfl 0)
o i-j c*
4-1
CO COm d X m oo •rl vD oo r~
oo CO rH CO
in 00m o CJ H oo rH •* co «*
r^ r-» CO
es co i-i co CM CO o CM CM M i-~ oo
#■ •> 4-> cd * *
•rl
CJ
CO CM
co o- CO CJ r-~ m cd cu co -a- CU Cd
4-> CM vH co 60<4H
CO cu * ■3 cu * P U
tu N-» cd J-l WS-/ Cd 4-> *s O
.3 bO p rH ,3 4J
m o . OOst r<. ü CO CM
«* CU w m co 3 m O CTi >.T3
oo HH 00 00 MH lH CTl \o 3
.3 co P co •* vO •H <fr -Ö CO ^3 CU CTi
co o o \D CO O rH CT. H TJ
O rH 3 (UHN P H CTi CU vD P CO O •H CM
CU « CU cu 3 «4-1 H • 3
H u *i 4J 4J
cd
* A
in in 3 cu m CM T3
P
CU CO 00 15 CO CO 3
.3
co vo 3 O 3 CU rH O
Cd 3 H
4-1
sCU
0)
CO
<D
u ÖOrH
cu 3 *
CO o 3 H
«H •rl •H *
.3 4-1 60
3
P 3
O
!3
o
Cd Ö 4J
*?£ 60
3 co
CM O
•H
4-1
I*
CJ
O
•rl
4J 4->
.3*. .
;
CJ
U
cu
CO
-% u CJ •H p cd cu 4J cd CJ 3 Ü
6
sO
H !■
CJ
4-1 CJ
3 g cd ■ f-l > o P 4-1
cd
rl cd
cd o_
' &
3
rl
u u
•rl
>>
Hi
•u
M iH MH 00 00 co 3 •rl CO > o CO P
CU CU 3 <D H 3 >
3 CU 3 0 CU 3
•rl P CU ,3 «4-J
rl
PM
Cd
rl 5S B rl
e 3 •rl 3 cu
to CO u
•H O ,3
•3 3 > O
N (0
CM CM co
CM
co m
i I
& a
o o CU I
iH rH *) ft
cu 4» O ^ ^-*
-" co > CN O > «* r^ 00 VO m CM CM CO rH CM O
u o CU . co «a cu co r^ CM 00 m «a- CN 00 U CU 00 CO
E CO T3
a)
i-l CO B
« flj
O o u «
vO CM CO 00 > CO O
60 cu cu *
E H ■<* e C CM CO ß -a- CO 0 > ■o -a
CU a O a O rH cu rH •H
CO o 4-1 * o 6 * B 4-i O 4->
Cfl
cu
o
cfl " r* 0) i>- «a- l~» CN > © CO CM «a- rH -a CO to > B rH CJ CO VO O
rl U CO TJ B co CJ 00 T) s CO
% U cu « cu ■ft CO A
o >% CU CM CO •H f CO T3 HN 3 4J CM T) CM "O •H 4J 4J U o
■a * rH 4J H
«
CO e "■■ a 3 B B * cu
o cu
cu
4-i
to
o
CJ a
o
CU
u
u
4J
01
o %
•H
I
o
60
e
X
60
a JS
•H
ca
CJ
cr
cu
60 >,
cu a)
go oft CO
•H
CU
>
4-1
rH
CO
CO
4J
CO
O
B CJ CO 4J s JJ CO CJ
4J 4J « o •H a) o
(0 c •c o cu e U CO B > > ß J3 ft 4-1
Cfl 0) •H ►. CO •rt 9 CO CU > CJ W X) •H CU
4-1 >-> 4J ft C2j 4J CJ C8 4J a T3 ß
•H •H •rl CU <4H •rl
I I CO
vo -a-
o r-~
0\ u-i
o
(U
ft
CO
CM
^N V_^
6~S 6-S
r-- cjv
1 r! • ^-s U ' CM
B ^-\ cu o CU CU CM
0 4J rH CU > rH rH <-t > O H
o CO /-v CU rH O rH CU ft O rH 0)
Cfl 6~S B ft ^> B O ^ B
0) rH m B O , CU
CO vO 
ft 1
v—' M •* CU
O 4J
ft
-a- VO
in
M I-l Cfl > ^^
CU Cfl CU CU
m H M CO
CM CTv CU cfl r^
>
cu
W 00
CM
U co S
-a cu cfl cu
* Pi ft^RN * ft rH Z CM * H ft rH a
-a- r~ m oo
rH VO CM r~
CM CM co <r in vo
r^ rH r» co -a- -a-
vo CO 00 -a- -a-
* *
B
cfl
rl rH
O Cfl
4J fi •H CO
B O U CU
Cfl Cu CU -H CU -rl
01 ft B
US M ft cfl CO Cfl
rl CU
Cfl O rl
O
ft
rl B
CU r-t CU ft CU o cfl
CJU > 4-1
o 3 Ö CU 4-1 o
ft Z Q U o H
CO
4J 4J
CO 1 co U
CU cu
o ft CU /-v ,—\ CO /»-\ /~\ 0 ^ CJ
/-v O o>
o o VO rH r-~ vo r^. in > ON ,3 O 3 O rH o m O O rH (3 vO
i_l rvl O CU vO sf Ü vO I-» •H ON /-N o sf O r-. rH n O ON O cd vO vO
m ON Mm o CJ ON
Lön development
3,172 ( 2,8:
•H O 00 CO 00 U rH rH O t-\
CM CO I-» > O sf
y compr ehen
t accountin
esea
dica
5,
* 8,
* 0)
ilide separation and
struction processing
4->
co rH TJ rH CM 3 rH 4-1 Sf rH CM >>
3 cd cd
ower ba se
o CU
CJ richmeni * * CU #. u * r< ,3
676 (■
evelopm
3,091 (
researc
166 (
4,711 (
2,843 (
149 (
676 (
1,318 (
2,005 (
1,921 (
irticle
rticle
4-> 60
Sf cu o
vO o c O CO O
•rH m CU
CM
O
sf
CM 3 CO sf
1
r-» sf rH
9H oo m o cd r~
oo m
.ft
O
c 4-1
CD
>, Ü " « •H «\ ,c .
c rH "iH r-- \o 4-> CM CJ
o TO u
15
a *
* H c CO * <•■- co CU ft M C Cd 4J * CO T3 * ; CO * CJ
■CD CU •P •H CU O ,3 cd CJ M ■ •H ft 3 ftrH 3
4J CO ,£5 4-1 ft U rH CJ *Ö cd co cd xi ,fi CU rl cd ft
o 4-1
edun
Heavy
speci
sear
ense
clea
chno
vy p
earc
o: o :.4J
cd .
1
4J
•H CO
ety
U o H
s 3
o
4-j
Ö CU
co
CU
>
0) PS u-t ft cd co
•H
«
cu
SO
O
o
S
r4
ft CU cd Ö ,0 cd CU P CU -Ö
3 M 4J i-4 T3 jz; 4J ri v-' cd X cu cu 4J
X CU
i 1
rj CU w In 1=) a ■- •H
1 1
CO (UK r-l •H
is.
I 1
S3 -o
in so
oo CM
ON CM
CM VO
I »
Sf
r« »
cu m
> rH
O rH
v—'
CU
00 r(
cd cd
CM CM 4-1 CU
O 00 3 >N
00 H cu
n •V CJ 4-1
r^ CM j-i co
rH cu cd
* P-l rH
o o
O vO
rH m
oo o
rH
cd
ü
•H
-3
3
c o
O 4J
•iH cd
4J U
cd o
•H ,Q
T3 cd
Cd r4
Pi
CU /~\ •"N ^-s 43 /-N b0/-> /^ ^^ ^^ ^^
CU Ö O oo CO CJ m O CM 4-1 0\ <r Oi 4-> CO CO
43 4J o 43 a\ oo U CO rH CO d rH m l>> rH CD <}■ 0)
4J cj N CJ CM rH cd r-» O «* CU H cu +J r^
 •H
CU O
4-J «
-H
M
C M-l CU cd CM co H 43 ft •H rH CM 00 4-1
O «4-1 c 03 CU cu T3 cj o co •rl r- d co
CU o O co i-i O CU i-l 13 CJ IT! >H -rl
43 •H -Ö v-^ CU 43 60 MH rl |3 -H
Ü -4J0 CO a O o r^ e «N CJo uo cu 4J «tf CU CM CO CU vO u m vO O O O 0
Cd •rl o co CU o m o CO d o CM e dd-dm Cd U CM CO 4-1 o CO M 6 >> 6 00 * A CU «43 Ö
cu o n cd CU * rH d CJ * H CU 43 00 u o >4-t v-l rH rH 4J CM CU
rl •H •H rl O CO CO CJ O cu cj O H * cd co B
d IW •o o O •H cd •H H •• > bO d C co
o cd co •• 4J 4-> fi iH ri 4-J O co o >s Ü CU cu o u cu
fr-8
CO 4-> cu
M
Ö
CO cd CJ
•H u3
O
•r! u d d d
CU CU 43
•H
43
ft 60
O 43
O >
•H'«*
rl
60
-rl
4J
CU v-l
> O
d CO CO !5 o X rl rH CJ 4-1 4-> 00 43 cd O C
•H
§
•H
CJ
<U
o o
;H 13 rH
•H •P H 4J
CU CO
C CJ Ö
r
cd
H B ft cu
3 3 O 4J
CJ •H rH
Cd O rl cd CJ
0. CU ö cd rl Cd
•H H 43 0) Cd O co
Cd CJ N
U «H
cd (3
00 CU
O
OO
cd
H
CU
U
-H cfl cö
VI
CU m Ö
£
cd
CU
•H CJ o
43 
cu
CU
co
cd
43 CJ CJ CD ft'H
4J
•H
3
gun
CU CU CU Tj
rl Cd
CO
O
rl
CU
CD
CU
CO
00
rl
r-l
O d
TJ
ft > o •H Ed •H 1 I £ 3 T) pq & ft rl O rl O 4J Cd
co CM CO
lO O
o -*
o\ CO
B^ 6~S
m 00
rl * rl •
cu o> cu o
> o > o
O rH O rH
v^ ^-^
cu cu
00 rl 00 U
cd cd cd cd
u-l <Ti 4J CU 4J CU
O O d >. c t^
CTi 0> cu o cu
«« .»i O 4-1 vO CJ 4J
CO U CO o rl CO
cu cd *l cu cd
* ft rH CM ft rH
<7\ CO
vO <f
m o
A
43 Wt
CO Ü CM
43 rl
rH CJ cd
cd rl cu
Ü cd CO
•H CU cu
CO en rl
!* cu
43 Pi 4J
ft rH
CD CO
CU d
U-l cd 4J o
o CJ •H
•H rH 4-1
cu
4J
ecu cd cd
d N
3 43 o •H
4-1 U •H d
•H 4-1 cd
4J Tl cd 00
CO d !3 rl
d cd • o
H m
CO CU
CO
4-1 cfl
rH
•H •2cd £ •
iH
CJ 01 CD -rl CJ to CO 4J
O »-s O /~\ >-,/<"> (U CJ ^-N /~v C /-\ 41/^
CD <7i
0/>
P«sl-
Id
00
n
cd
•rl r-« CJ i-l JJH CO -rl CTi Oi 3 r-l
tu
to r-~ o •H vD Ö IH CO> CT\ O Ö vO tO H -rl
03 iH CU rH u co 0) O ,0 >^ CM O ^> CU 00 -rl rH 4-> >>
•H > o d, u 4J 4-1 PU
x
T3 CO
<y
•rl ,0 X CO CU CU >■. CU rH
s 4-1 4-1 CU 4-> 4J «4-1 M MH cd CU CU > 4-1
o cd 3 Ö cd O ■Cd rl .4-1 C CO
cd
O -' N
00^-' *tj v_> U (1) Tj v 00 v-x CO >»• CO CU CU v-/ CD ^ «H
'rl > H C •rl C > cd rH
>-> co 4-> in >^ oo •rl Cd CO rH >> r-> > rH
O T3
>> OJ •H r-»
00 00 y rl
> r- >-,o
00 4-1 o u
bO 00 CO o 60 f~- 4-14-1 O 60
M H CU H u co CO CD H u d CO < M Cd C\ CU -rl CM (U
•• > U 00 4J CU 0 a) >-, U > > * >
CD
g CO g CU
fi 4-1 ß CD
W H CO -rl -rl
a
w
•rl
CO
C
o
Ö 4-1
4J
W -H CD
«rl
4->
CO >rl ^3
CU 4J CJ
O
W •H •rl
J3 ca
•rl CO 4-1 CO •H rl •rl O *-l O S-l CU
60
u 4J H u u a m c CJ •dg 4-1 CJ O ö cd p cd cd
cd
•H cd •rl CU •H tU <D •rl cd •H ,d •rl O CO O CU
Ö •rl S C TJ u o 0 g •rl 6 4J S >H Cd -rl CD 4-1 6~S
•rl Ü O CU 5 CJo •tf OPT) T3 CU TJ CU Ö r-l
CU •
3 ,0 CU 4-1
<
00 cd C to
H
4J
<: 5
Pi
4-1 <3 CO CO B CO H
Ü o
p.
u o
•H co
is • • • • • • o > o
O rH O rH
\~s s-x
CU ■> O t>-
<u o CO CU CO O
CJ 4-1 CM O CJ 4J
CM M CO ft U CO m
VO <u to CM cu cd
vO pH rH * PH rH * rH
vO CM
CM vO
en CM oo
co cyi
oo st-
vO
sf r-
* rH
•H
rl
O
u
•> CN H O u-i n o CN CN r~
CTi CN cn m m m oo CN m
ft ft ft
■s cu at cu co 1 n #» A «t
CD CO ,0 3 CN -* m CN oo cn <*
cfl co o H H H H H
CU CU rl
co CJ
^ >-l > * * ■■■* '
CJ CU CU
oo d cu > .V
H Td O p.
cu ^
60 u
T3 3
3 CO
PP cfl P. cfl
CU CO
•p ^ S Q
 o cn i^ O VO rH rH
cd cu u cn o O rH H O CT\ cn CN
> ft A A A * #i ft A A A
cu a o r-. in in rH <* m VO CN vO
f*i o 00 4J <t r-. CN CN CN c*q o> r^ <y\
1 •H cu H CN
:*« O 60 *
M CO
M 2! tH p
0)
*
o
•rl
6
o §
4-1
«J
d
CU o
•rl
•S rH
to H
d •H
cu e
I vO CN H rH CN rH Ov r~- cn
■sn a)
d
CU
00
CN vO
CN 00
A ft
H 00
cn1^
A ft
■vl-
vO
*%
rH cn
CO rH
ft
cn <r
m vO
A . m A
P.O 13 oo <t- r^ vO oo f~» oo m H

S P CO O rH cn oo oo 00
o .£> rH CN
o •W
•H CT<
* * *
00
IS
ß d
CU O P CU CU
CN CJ -rl O CJ O 4J
U 4-> CJ u u d
CU P o cj p p p
H O B cfl o o o
■s
H
•P
co o
U H
CD ' to CJ
CJ
d M ft co M M cd
P CU -H CU 4-1 CU
o is -y CJ & d £ ti
Ü o p cu O P O CU
o ft CU' ft ft o ft-H
§ cO ■ 6' co Ü MH
CJ P. CJ ü CJ .H
rH •H O >-, •rl Cfl •rl CO
Cfl M rH CJ M n n CO
rl 4-1 CJ -H 4-1 CU +J CU rH
CU Ü > rH CJ 4J CJ > Cfl
Ö . CU CU O , CJ -rl CU -H 4-1
CU H-O ft rH CO rH TJ o
o W W w H
1 CO 10
•a d CO 0) 4J
CO cu o 4J rl d •-<
4-1 CO 4J -H d ^ 3 cfl O
CO a; -o ^■s CO 4-1 /-> •-s U co r~ 4J ^^. M ro
a) •H a Olrl 4 CM o 0 M 00 cj m BO vD
4J 4J CO rH 0) d m ■a 4J 60^0 oi m •s
•H m M TH 00 QJ CO M-l -* d «N
c H a « ■ T3
■H
* 4J 01 rl >.H
CU 4J ^^
CU s-^ O
l-l ■H
s^
tfl 0 •H 0 . vO 0) M CO CO
•H Ü -H v-'rH 0 CO N—/ r-l •rl CO ^ ß -H CO. CM 4J CO
4J CO 4-1 O O 4J 01 l-l 4J at pi in CJ O
ri
CO CO IH « r\ O >vO CO 00 rl a co r-» Ö0-H rH 3 m cu -a-
E O cj co 00 O 0 o M 0 r-l CJ rH rl Pi *
01 •H 60-H 4«! co T3 CJ rH >, 0) CJ rH >,-H r~ 43 T3 CO CM
Pi M-4 C M-4 CO »r-l C #> 60 4-1 60 > « 4J -H d
•rl •H -H CO r^ o) co 4-1 CM U d 4-1 rl . H 1 CO •H
M CO M 4-1 3 a 01 01 d 01 CO P. 1
a d oi •d d d
o Si Si
(U CO 01 HH fi
> 01 > rl 6 01 ■fi .§ 01 -H 43 iH o
CJ fl> H -H a CJ d 4J CO 1 •H
CO 4J >a
>s 0 5>.4S CO 4J 60 CJ u t>o cj -H -a
4J U 4J 4-1 aj o •H •H CO •H •H H -rl rl 4J CO -H
di a v o
IH 0) 4-1
. Ho eo eCO E
0
OJ
CO
CO
d
E i-i CO m c3 •rl
0 V 1 > CO T3
Cfl
1
5 i-i-> CO rH CJ
I CJ
1 1 1
3
o
u 1
o ^■s 0)
< l-l 1 M
0 s-' P. rl
CO
o> 01 0) 01 a) «3- 00 O
•rl en en XI >> CM m rH
tu rt « 4-1 m m
•rl •H
> H T3 0 >
•rl
O
ca
•H
CJ c cfl
01 01 CO
ex g
w c 2 m o CO
(SI 00 00 \D
01 ai . CO
o > a o CM
u ^ oBO 4J
3 d
O 0) 0> CM
O 60
M d
ai o E ,P
> -H
0 .H
•H
O E
•H 4-1
M 01 01
•u C 60 rH m 00 CM
CJ O -O O CM •tf
01 3 m «* CM
iH •• 43
Ed 4->
<c 'c 00
H D
01
0 >■>
rl 1 M CO
u CJ 3 d 3 4J
OJ 3 •H >-. 0 •H o d
u O H 4-> CO 1 CO CO
tu u CO 4-1 0 QJ
MH >, rl d
CO
4-r I-i
rl
BO
rH 3 CO >> P.
0 rl 0 60 >, CO "G s to Btl «Iri S •rl cfl .»
rH
H |* 01 4J 4J rl CO 0 4J
q 60
rl O cfl
l-l 4-1 b >> d 0) CO 0)
d
d ji P.
E
P. -rl
CJ
9* d P. CJ 01 MH 0 0 3
1* 3 •H CO CJ CU •H CO CJ 0 CO CJ CU
o o u H CJ CO 4J -H u cu •H P.
P. CJ •H 0 -H 4J CJ CO >. rl p. u U CO •a
CJ rl a. s a d •H
R
M CJ 4J 3
cu en
4J
CO o cu CO
•H
ü
•rl
C!) 4J 0 0
CJ 0) 4-> •H
<ve 0
0) -H
d rH
CJ
 CO 00
4J
<
0) 0
BO P.
rH •H
td CO
0) •H rH -H d
E CO td CO •H
CJ ■rl W CO rH u
•H
(11 CO *~s 01 CO
rH • • d d
Ed H H 01 o
P CO 1 p
a3!
O Q» PI 4-1
•H -H -H 4J ^-N ^ P /~\ P ^
4JA(J 1 /N ^-s P /-> CT\ PO\ cu <r O rH
cd o T\ co CT> CO CU CO O o vo a CJ\ •H -st
CD 60 rH 4-1 vO co B r-l O rH oo pm CO O
CU «H -H (3 m p. U - CU •> cu o «i
rl "
•H +J ü Cd >. O 0) -* > CO 4-1 rH rH ^ "3 CO CU CO cu >
■H CI IW 00 u p CU 00 P CU * TJ cd > P
rH > w CU D N-^
N-^ > U O w CU v—' v_> £ CU v-^ O ^
•H C U >. P iH CU CU -rl M >. tj O
CJ -H O CJ CU 4-> -O ß 4J ,o CM OOf"- cu m -*
ca o *j -H -* 3 r-N. 0) 3 co o o vO > >,lO a to
«4H 4J O (0 H N CJ 43 CO C ^5 H 4-1 O- rH rN -H MOO 3 P rH
o CM w o <y> •rl -H O CJ -H co •> O •N 4J O «i
•HO"
to cd »cup. a u •rl -H M cd m P «tf CJ rH 00 P «H rH
3 PrH P O 4J 4-> a +J UH # ,P cd o O 4-1
•H a . W 3 Cd O P CJ op 4J ca
CO TH 00 4J Cd O MO m CU •rl \G 3 U
CO CU CJ cu cd cj o 4-1 TJ CJ H CU
CD >,-3 >>>4H H P , 3 Cd CU P P
CJ 4-1 -H 00 Cd Cd P o p p P T-> 00 rl 4J O
O iH Cd U CO P O o o o o p P -o
M > 1 je ' CO CU P CJ CU
CU 4J 'rl CU CJ 4J Cd cj cd cd O CO > -H Cd «H
M CJ 1 P Cd N •rl N N 3 rl cu CU CO 4J
cd co u a)
01 O
Ö0-H
4J-H
Ö B
b0
M
g-s
cu cd
3 -H -H
OOP
p cd Cd
rl
4-1
CO
O
4-1
CJ
CJ
o
u
rH CO
CU
,3 CJ
CJ i4
•HrH
rl -rl
T3 cd o a) -d
M 4J 00 O 00 00 P cd p 00 O 4-1 O
Cd «Ö rl 4-J a "H c n cj M M o cu cu •H rl •H Cd
i-l ri 00-4 (1) cd
i I
H
1
O MOO
1
O
1
ri
1 Cti P 55 MH
CO
,304
,972
CO ,469
,085
O H 00
CO CM CO VO
co oo
CM 00 O CM -*
H rH I I
* *
CO o CM
m oo o oo CM m VO
o CO o r-» vO
CO
ft
H o CJV
wt n
O r^
f\ A
CO
«i
rH
m rH in CO m
CM Ov CM VO rH
* *
o 1-^ rH H r-- O r-~ m rH <r

■<f 00 oo
*37,31
61,68
10,94
H CO H
«tf H Cjv vO
* CO o
rl
co cd m H
^
•H
cu
> rl
cu
CJ
Cd
rl
O
4-1
a
P.
o P
4J
CJ
cd
cu
CO
P
■4—1 CU
rl
rl 1 CJ CU
CU
>
co
rl •H 4-1 SflU rH O cu CU cd 00 •rl T) 4-1
cu rH cd o CU Cd CU *rl u p CU P rl CO
p o rl CO p •H CU > 4J X r-l •rl p Q
cu p 4J CU <4H Pi CU Cd u CU CJ CO CU CJ
CO CO CJ •H Q rl EU 3 CO o
CJ s^-o •H P •rl CO rl o S -3 P CU s
rH •3
H 4-1 •H P CU rl U o 3 O
£ <-i
cd P rl 4J
CU p co
P J3 cd
4-1
< co •rl
"§
<3•
•H
rH •3
P4 P4 U
• cu cu p cu o
cd CJ r-i
rl CU CU
W rH S3 rl CO rl CJ P •*s
rl
CM
m vO CM CO
10
>> 1 4-1 o
u 4J e ti 4J rH
ß •rl O CU 4J C O ■P ß &-S s^s
cu U CJ 8 H ß cu ß CU O o -*
CO 3 P Cfl CU CU 10 45 •rl -rl •' •
5 4J Ö ,0 O +J 3 8 cu ti CJ MH 4J o CO
CO co o rH CD MH ß rl Ü0 CU •rl Cd rH I-l
00 CU
o CU O 00 CU -rl 4J CO B iH H
■rl CO M '■
CO
CO *H Cj cd > CJ rl 'H CO CO rl H
Wl H S-l H (1) cu ' cd co H ß H 4-1 CU Cd
ci nâu /-"\ cd co /—-. ■O *J /-\ (U (3 /> 51 O /~- cd co /-\ > > /-N
•H o n J-i c r-. 4J cu 00 B CO rH O 00 CO Ü o CD O co •rl CU O U
cd
(-1
cd
rH 1— 0) CU CM ß rl CTi M cu oo cj a oo o CTi O Ü O T)
4J 4-1 O 4-1 e 00 cu m o e ti ti. ti CM P" w i-l 4-1 CU cu
CO
to CD CM
" ß
CU
sM S H
ß cd
4J S
CJ 00
ß CO
4-1
CD O
-rl -H
CO ß
•rl O
CU ß
CJ CO
>^ >.
Cd -H v / M CO v -O 4-1 T3 -rl rl 8 v 4-1 4J
8 s ^ u ti
•rl
CO
O 4-1
v^ cu co ~ o cu -' cd v-x 4J >_• 3 P -' co CO
CO P^-' 4J
•HO ß ö •rl CU }4 ß 4-1 4-> CO 00 CU Q o o cd cd
a) o CD r-i ' ' rH *H
-Ö H > 8 O CJ 4-1 -H 4J 8
CU CM CO 4J -* CO 0 oo cu o
a) H 0 CJ CM B-fl vO >j cj VO Cfl
CJ o cd co r-- CO rl o I-l
}-l > -tf ,0 -fr 0) rl co m o oo cu I-l > vO }-l
o co CM CJ ö <r cu 4J r-- (J SHU AHM sf 3 P.i-1 cu cu . CU CO
4-1 t3 « -H o OJD P. CO cd O cd > J5 "P > >
CJ CM u •rl ß r! o rl iH rl 0 a> ß CU 4-1 O o o
ca >-, d 4J •rl "H > ß P K^
CO 00 0) CO co ß cd 4-1 O -H •rl CU 00 CU CO
4J CU 00 cd >W 4J >•> CJ O 00 00
M o 60 CO CO
CJ P S>i •rl H cd cd
iH 6 •rl cu cu ß 60 -H !S rl CJ H
cj O 3 4J CJ U 5 4-i cu cd cd cd o 13 rl O 4J 4-1
•rl ß Cfl -H CO o P 0 CU CO 4J P. CO O ti CD O rH •rl 4-> ß CO
St
cu fj
S 43 w ß CD M CO 5 00 CO vC 4-> «rl -H 4-1 •rl CU CD CJ 45 4J
o
CO
CJ
o cJ co cd > P cd 60 rl > OCP CD T> 0 CO T3 f> 4} CU CJ CO
u
4-1 CD O U 6 cu cu •rl Cd ß "iH O O Cd -rl O Cd CU P H CU O Vt
<l4J Up •s Pj< 0 co ,J .rl
1
g.n Of<H
1 1
CJ Pi 13 CO W
1
4-1 CJ CU
ex,
CO
1 1
CO <r H <r vo
VO O CM O sf
00 CO 00 CO CO
CM 0t rt n n
CM 00 CM sf
rH H H
* *
m oo o oo r^
CO en o vO O CO
en OOlOH
■ ->i ■ #\ #» *
■»I.
:sf m vo m H
CM CTi CM CM
H
* *
CM CTi rH rH
vO CO CO rH 00
-it CO rH CO r-
vO
I-- 00 Is» vO
CO 00 CO O
rH
* *
CO
CO
.ß. CD
CJ
rl CO ß
cd > CU
cu •H p.
4-1rl
co cd CO
CU
rl CD
rl 4J CD <-\
Cfl CD cd
•rl CO 4-1
cd' ß ß o CO
rl 4-1 H
Q)
•H •H
8 CD 1 cd
"^ P 41 4J
< 43 P o
• w H
CO
•
H *
■
11
4J 1 00
s ß c 1 B 13
CU O •rl 0 O 4-1 3 CU
^ 0 CJ S*. rl 3 u c •H -3 rl rl
ooxi cu 00 3 •H p cj a) e 3 CU o M
O CJ ri O *3 ß •H j-i e 3 O CO 0 IH CU
rH -H rH cd S 6 CU ft-H rH 4-1 o 3 3 14H
O U MH O TJ rl 4J SO CO •H 3 CU 3 1 4-1 •H CO co
3 3 O C 3 3 0 . >H 4J cd 3 cd 3 rH rH CO 3 4-1 3
XI > 3 4-> 4-1 P rH rl o O rH MH
•H •H 4J rt ß •H rH ßCd cd cu g <4H rl o •rl rH 3 >»
0 0) ecu
 «H cd ^>4H 4J •rl •r! o
cu u rl -H 4-1 X5 CU 4-> 4J -H J3 
TJ
CU 4J
4J CO >> 3 4-> _ rlft o
rl Cd O 4J O -H R CO XI p
cu 4-1 4J CU > ß v'H 00 u js ß CO TJ >
X CJ Ü 4J CJ CU •H u cd 3 4-1
3 0) (U 00 cd 3 CJ o >^ H O
o > 4J o ß u •H •H CO > O ,0 3 -d 3 4J 3 4-1 3 3 O 3 a. CJ 00
CO «H Cd i-t O o rl rl 4-1 •rl «H 4J •H -3 . O Q) O c cd 4J CO 4J CU Ö >.
3 O 4-1 •rl CO B
00 CD 4J >1 •rl CO 3. rl
u
cu
cd a) ß +J CD »H cu 4J Ö tJ t« E cd 00 cd 4J 3 4J
o
3 3 3 1 3 00 CD rl M
60 rl x; cd rl ß CU (U 00 rl CU cd •rl % ^ 
g
FM
4-1
CO
CO <U -H o MH 3
TJ 4J M-t •H ß •H
0
cuCJ
XI CU & rl rl cd & rl ß > CJ CO Ü CD H U 3 o > CU 3 O rl X! H CO 0 3 3 (U CU u
•H ß rl
öt
CU CU rl
rl
4J
CJ
H
1
4J
r > O 3 H
1
0) CJ H TJ
1
H H CO rl p CO O o o r~- CPi vD
o u-i CO ro r~- 00 m ro
60 4J
Iff
+J CJ
O Ml
CU
0\ TJ
<N
co o
3 fe xi
cu
0 13 >>
M CU
(3
TJ m O
3 iH •H
CO CO rH
rl H 4-1
<U CU -H (U
00 o -tf <Ti o CN o
CO »H TJ o o en o m CM
„n n a) en fO m Oi CO
H 3
o •°l
CM O
CM H
cd co g
3 4-1 •H
o
•H
•H & "■ ca
3
0)
CD 00 CD
0 CO 4-1 3 Xi
4J 4J co CU 4J 0 4J CD 3 •H 3
CO cu co 4J 4-1 CO x! 3 O 00 ca ca
3 co o CO CO o CJ u p>, H CJ 3 !>, CO
u •rl CJ 4J O >, Ü •rl 0 TJ 3 M-l TJ 3 3
cu rl /-N 3 CJ CO rl P-H 3 _ 4J -H <H CJ 4-1
4-> P. 4-J 3 CU Q) 3 O CD **H 0 3 3 rl >~, CO o co
5 rl 3 O
CU CU -H
4-10 4-1
0 4-1
XI ö
CJ CU
3 >
O
•H
-H
4J
3 rH Xl
CU 3
0 > CO
3 O
■—i -H •rl
rH 3 4-1
3 4-1 00X1
0 3 O 3
3 >H •3 4J 3 0 CO rH •rl 3 3 CJ ca CU 4-1
3 3 0) rH o co rl 3 O cu am O CD •rl rl O XI O
•H XI •H •H 4-1 „ 3 rl >, 3 X O CJ 3 U 00 4J rl rl
CJ 3 ca 3 co 0 -H 3 CO O O MH O -H 3 3 3 P
•H O 3 3 CU 3 CO -H cu •H rl -H 3 3
g rl •H O rl •rl 4J 3 >i 4J TJ >> CD 4-1 3 rl 0 H 3
3 4J P. 3 :>> 3 3 rH 00 3 3 00 3 3 3 3 xl Ü CO
CU 3 CD 00 3 01 O 00 XO 00 4-1 CO 3 -rl
00 CU rH O rl 0 rl rH -rl •H
H rH •rl 0 CJ -H 3 rl
0 ■rl rl 3 rH >i 3 P. P*. 3 O 4-1 O 3 4J 3 3 rH
0 P
cu 3 4-1 CJ O T3 O T3 3 3 CD ■—3 g CO -rl M-l Xl 4-1 rl
H
■3
I •rl
3 a) H
cd > H
CD H •H 3
0 x|
CU CJ
•rl
CO
XI
H rH -H
3 3 CO
4-1 > Xl
rH X! (U
CJ CJ >
3 3 3
XI H 3 3 3 3
X cj 3 > 3 3 4-i
Q 3 X! 3 rl 3 CO
CU
P 3
3 3
4->
H u 3 S xl cu 3 CU CU 3 a 4J -H g +J 4J •rl D 0 3 CO 3
35 •H ^^ CJ 4J CO g -3 CO
CN ro m
12
01
CJ H
<4-l >4-l *o OJ c >
E c tn cfl
rH
■o
o
d OJ -rl Cfl
C0 rH 5^ OJ
■o 0) 0) 4-1 O ■• CO (U 60 cfl M d
c £ > c , c E O 1 o 4J 0 o •a tu o CJ Q.HH
E a CJ
« 4J O
rJ
tu
E tn
tu
I «:
en E
ftrH
o o
cu B MH
O rl
d
CO
r.
>-,
4J
cfl
(3
O
TH
4-1
d 3
CO MH
•H
to >,rH o a 3
tn
to a 3 3 4-1 60 M -H cfl 0 •rl 4J 'P. CJ CO d
1 I*... cu
U OJ
in 0) U-.
TJ •H
c 60
o I"
•H 0)
to o
O r-l 0)
a.M-1
o o
rH
•rl
e
rH
CU
>
Ö
43
CJ
en
CO
43
ft CO 0
•rt rH
CU
C
4-1
CJ MH
rl
CU
4-1
d >> *
to
H cfl
a
« - -
MH CO 60 d Cfl
dcu
0 o
u 43 > 0) CO CU OJ 60 43 to O
e
CU CU O P. 3
o
60
a
IJ rH rj C 4J to > •rl > co 4-1 -a u •a 4-1 •H 60 CU 60 p. 43 O CO CO >>TH 43 MH
■u o 01 •H tu 4J o) tn tn 3 43 Tt rl 43 CO 4J rH en C 43 4-1 ^ 0
c E -a -o O O cfl I cu 43 1 O E. C d c MH O cfl CU tu

M-l O
O 14-1
01
0) CU «4-4 cu to c O 01 4«! 4-1 0) E •H 43 0 >> 4J d E o d
0 i-t 01
0) ►. CO S CO rl CO 4J 43 a 4-1 4J O • 4-1
4J 0) 0 en 4-1 43 t-l
O 0) o to 3 3 CU en 60 o CJ d OJ CO t>% CU 60-O MH 4-1 E
4J n en tn E QJ 4J "d
4-1 4-1 er 4J •H rH 01 CU 4-1 4-1 60 MH d 0 CJ 43
0 a 01 3 4-1 CU 4-1 H iH
3 -rj a E en
4-1 U- fi CJ 4-1 01 IH
O 43 o
CU 4J 4-1 •—• 4-1
-a en cfl
en cfl
CU rl
43
4J
•H
CO
CU 43 0)
> CO Pu"0 tu
O Cfl •H CO
M CO 4«! 4-1 cu a-H
0) C0
c 0 ■H o 0) w O 43 0) CJ
c
•H ^ 4J >i CO a CU cfl U 4-1 0)
a
U r3 MH CU o cu 4J
0 "O •rl rH 4-1
O CJ H to C0 H
d to 0) 3 0 d XI
0) 0) a tn d O tu pe! 4-1 CU CO T3 §
•
5H ao e 43 to P. -rl cfl
c 3 4= c O rl O l-l TH Cfl E r-l
cu 1
0) (3 OJ CO d
o
43
CJ M-l CJ CU U 4-1 4-1
o r-i 4J 60 o u •rl O •HOT) rH (U
■rl 4J 43
cfl U
Ö
4-1 '•» CU > rH
d >> CJ 01 CU •H cu o I ri 3 ^ en d
•iH m d •H 0) O 4-1 MH 4J <4H Cfl 4-1 M-l OJ OJ >
i-i -a cu cu a•a M 4-1 4-1 en a. 4J co cfl
4-1 > •H o 4-1 CO CO _ CO l-i CO O 4-1 60 MH CU 60
s 60 en co cu rH
« to to O 60 E 60 E 60 •H rH cfl CU •rl rl rl E O CO 1
M o
4J tn
i-t •H T> 01
en E &
rJ
cd to
•H
4-1
CU
4-1
•rl
•H
4-1
cu
« _
60 Cfl
en cfl o
>,rH I
•O II J
•rH 3 O
g3
•H -H #■
43 3
4-1
CJ
CU
a o o
a H
CurH (3 • " en
0) en U
4J 4J 0 •rl
a cu d 0 iH CU MH cu P.
60-H U en 43 cfl 43
CO 4-1 CJ p-, U CO 4J t-l
4-1 E 0) ü •H rH CO CO
UH 4J rH c 0 4-1 MH U MH rH cu
71 o 4-1 o MH iH M 01 0) >> c
' CU U 1-4 CD d rH 3 43 01 43
O cfl |J It r «1 rl >
cu a co rJ d -H > tn co > O T3 43 (3 en CU M 60 £> Ü •H 43 o d
4J rl O d Ö 43 0 0
> o iH a O 4J .*& 13 4-1 cfl 3 MH 14H 0 3 3 4-1 •H CU CU cfl 0 co O
E M MH CJ CO p MH Cfl MH 4-1 MH a
c ü
M 01 rJ
01 01
u
U tn
I
O M H to M
!
WOO
1
O 43 MH
1
CO 43 TJ 4J
1
00 © o u-l
en 1T|
CN
oo o O <0- •*
CS
CN
o i-H PI
O
PI r-»
•a cfl
•H 4-1 •H
H d !H
0 OJ 01
01 en fi en 4-1 en
> 4J a. 4J •H 4-1
0
rl
R- E
d
4) e <-\0 en U
3 tu 0
0 CJ
en
0
CJ
to
•rl
E d 4J > H
4-1
01 r^ u O r*.
•H 60 01 tu 4-1 CO A 4-1 cfl rl 4-1
•rl U -a d CO d 4J CU OJ 01 ■rl
r^ Cfl tu 0 01
s MH T-i g Ü
60 d n E E P
H M Ctl CJ g •H
0 0 60 01 d tn O to 3 0 rJ
rH CJ d 4J 60 •H 60 MH d CJ 4-1
0 •H en •H
•H -0 CU CJ
d d en r^ en CO en 43 u l-l 01
43 0 tn en d 01d 4J 4-1 o 0 T-i
CJ •rl a) o
0 4-1 tfl MH MH 01
OJ 4-1 CJ cu o
CJ to 0 MH
4-1 Cfl o CJ CO u o CU -a MH
60 u d d
Cu cflo
* d
0
T) 01 0
60 •rl 4-1 4-1 0 to
d 4-1 cu cu •rl
4-1 •H 3 dCJ
•H CO rJ ft 4J > 4-1
Cfl
d
0)
d
OJ o
to 01 0) cfl •H E en to •H
en > CO CJ 60 4-1 60 s CU •H u 4J
tu d tfl CJ •rl CJ •H en > en 0 CO
o •rl 0> cfl 4J CO 4J •rl o r*> 4J U
o CO tn CO 0 CO en u rH o CU
u 4-1 4-1 M r^ cu
•H CU d ft cfl CO d
a. d CO •a > 0 E d CU 01
01 0> o 0 a cfl d o •H Cfl rl 60
PA E CJ «§ rD •rl PÜ •rl
•
CJ\ o
rH
13
ß
ß o 13 ß
co o •rl CU O
M-l 4-1 M 0) •H 4J 4J •rl u A
o c o CU 4»i 4J >> cd cd >, CO o rl 4-1 O >,

cd >l«H 43 -H cd 43 M rH fl. y n «-I O ß 4-1 43 ß
4-1 ü rH M 4J r-l CJ o 0) cd co cu 4-1 CU •H
ß O -rl a> oo co CJ e 0 CO
SJ
B I
*J rH
ft
4J
«■ri O
ß !5 M
O CO
«
a> Pi
H-l 0)
•H M o
M
>>
4J
co
m ß
o
01 cd ß
•rl 01 O
CU rl Cd
4J O ß
43 CO XI ft rH >i ft m M 43 H M •H U 60 cd o 00 rl XI CO 4J 01
co co a) cd S"-1 o 4-1 «3 01 3 O cu O ß ^ Cd ß CO
•H O 4-1 s P. M >■. 4-1
-o o
>
er 4J rH 43 4-1 & rH cd cd CO rl O rl
rH M cd >, Oft« xi CU o> ß O 4-1 O CU O 43 01 ß o
43 O rH 43 cj cd cu 3 cd ß 43 1—1 M cu ß ß ß cu cd 01 ft 0) 4-1
«d o> i-H 4-> 01 cd 4-> M B 43 U ß 43 T3 ß O 43 CJ
4J M M 01 so e o CO u 0) 0) ft CJ o o cd O •H rl •H ft cd
CO O ß ß 01 3 «V MH 4J 4-1 O 0) IW -H 01 co o 43 CU 01
CD HH
CO O •H ,0 ß ß & M o M CO r-l 4-1 4J 13 4-1 ß 4J 'f. CJ 43 l-l rl
•H X} 4J & o CU O QJ O 0) n cd ß •H O eg 4-i cd
cu a)
03 a) cd •H ß 4-> 4J > CJ > r-l O N Cd U cd B ß rl
X! XI >> CD CU ß 4J CJ ß ß 01 O 4J -H O CO cu ■S.IH to a
4J O H CJ S O MH Cd M-4 cd CU o ß P "4-1 O H rl rl M-l 01 u ß O 3 4-i
o cd ß •H O 3 O 0) Ü CJ O cd cd o O u cd ß cd
M ß cd »4-1 4J rH u a> M •rl M CO 0) CU 4-1 4J CO ß bo B ß 3 &
O CO Cd 4J O CJ 4-1 cd ß H 0) rH 4-1 01 4-1 rl rl CJ CJ 4-1 4J ß 0)
IH >H CO o u CJ ft cd CJ IS CO cd cd CO Ü O cd x> >> 4-i
CO !*.iH IU am CU -H 01 >^ ß ß O 01 |5 01 01 01 cu ß e rl H 43
XJ r>> 4-1 CO co co ß 4J X) CJ o M M Pn 4-1 0) 4J rl u 4-1 rl Cd M-l 3 4>ä ÖO
<D H CU cu O ß 43 cu cd 01 co CU 4J ß cd 4J O 43 CJ TH
M cd »4H l-l ft-H CO > N 01 H 43 CO CJ ß •H 4J 4J ß CO >^ CO •rl H
•d £ cd •H -H CU u 00 4-1 ß -rl
O Cd rl CO CO O M 01 4-1 CU 3
3 cd CO CU 3 ^H CO H 43 3 ß O r-l •rl ft O -H
ft 0) CU •H ß 43 er H
er A! P. 4J ,0 ß cd IM •H ß H Ö0
X) CJ 4-1 4J O O
4J 4J 4J CU 4-1 Cd
0) >^ 60 cd a) cd 01 -rl 4J N 0) cd
CJ cd 4J rl
O 0 B ÖO'H
M 4-1 ß 3 UIHiJ rß CJ CO M •H rl u Fj M ß r-l
CJ M 01 R ft ft ß ß CJ
a) •H o*43 cd co • 01 M cd cd XI p 01 o p ß 43 O
ß O H >1 ft O 01 'H rl
CO «H A! 43 4J CO CU u M 01 M-l 0) •H cd 4-1 W •H r—1 Cd
*ß 4-1 •H •H CJ ß HUH Hl
w cd CJ 4-1 ß & 1 rH co > CJ M-I cu
tu CJ H-4 4J 01 43 0) 43 X? g
CO CO a) M M (U 01 Cd B § CU O 43 x> cd a) ß rH XI CD cd ß 4J 4J CJ > 60 ß g
CU ,ä cd O XJ3H o o 43 ß ß l cd 0) 43 O O ß rl cu O ß CU 01 -H Cd O
H cd CJ (U <4-l 4J 4-1 ft U CJ 4J ß CO —"' M 4J O CO Cd 43 u U ü Cd 4-1 Q rH 43 CJ
i 1 1 1
CO m CT\ O CO CM o
CM l~~ -Jl- rH cn CM
o vD <N •* r-» i~~
oo
CM
r^ CM vD CO O
rH oo r^ r~
o\ -* \£> vD co
oo
CM
4-1 :*.
CO 60 4J
0) O 4J ß co
> MH rH ß CU M
ß O 60 O CD S o
•H ß 4J ß ß ft 4-1
4J -H /-N CO 43 ß •H CJ
CD ß 4J ß CD 4-1 CJ 60 , B >>
>4H ß 43 01 Cd O 4-1 CO 01 -H 43 er 0) m 4-1 cu x>
cd 9i co 4J g H ■•H O 4-1 CO 4-1 0> M o •ri 4-1 -rl
co S 4-1 ft CD 4-1 rl CJ ß 01 cj co co
cu CQ MO ß rl O M O M o O rH 01 O 4J ß O CJ o o cd 4-1 o M CO 3
ß o a IH 01 60 ft CJ O 4-1 m cu •H 4-1 CO
cd u !> cd •H CJ co CO rH M 4J CJ 4-1
4J
CO
&•
B ß
4J CO CU rl M
4-1 XI O H
CU 4-1
rl CJ
Cd
CD
4-1
CO
CO
4-1
4-1 CJ O
CO 3 MH
cd
M
01
rH
M
O
CO
4-1
•H •H cu CO 1HH 3 M a> CO CU ß CU 0) ft CO
co B o , CO ß 4-1 O 4-1 O ß 3 >. & 4-1
CD CJ co Ö M-H CJ 4-1 rl O cd «H ß -H O H 4-1 O ß ß
4«! ß CU O Cd «H 0) CJ M 4J CU 4-1 X} 0) cd -H ft O <ä
cd co o B cu CJ > cu cd B cd ß M •H CJ •rl B
3 •H CJ ß ß rl «H ß M >> 3 S cd o) CJ »H M 4-1 ft
er co 60 O MO O XI 01 H 60 M 43 £ M M Cd Cd O CO
43 !>. co •H -H &4-1 CJ 4J •H 60-H co •H MH Cd CU 4J CU M H rH
4-1 H 4-1 CO 4-1 0>
Ü CJ co m 4-1 CO UH 4J B CJ rH 0) CD cd
rl cd CO ß Cd ßCU ^ cd 43 > ß CQ ß ß M cd B oi y ß > 4J
cd ß CD O 60 rH CU cu 3 0) o o O O O 43 O rH ß CU CD o
W cd 4-1 U -H Cd 01 !3 M co H Ü CJ O CJ m 4-1 CJ CD !25 ÖOXl H
• • • • •
oo CTi o
rH rH rH rH rH CM
14
cu
p M rl >
cu cd •rl
4J a) CO
p co
cu cu ■3
8
p
CU
co
00 is
o
0)
H
4-1
P P
4J
CJ CJ
cd p CJ cd cd >> 4J
I
U
S P •H O
o co 3 rH O 4J CU
a) p. cd Cfl H 43
4J CJ 43 CJ IS 4-> CU 4-1 4-1 P •rl TH
^ 9 &
4-1 rH Cd
P CU
> •H p1* co 3 CO (0 CO "3 rH
CU ^ ■3 CU P CO U p cd P CU H <4-l CU •rl CU W> CU Cd -rl •3
O > P 4-1 U o rH Cd rl CU U M rl P rl rl Ü (U
4J
4-1 CU >"» cd P Ü •H P P CU 3 ri CU P CU •H (U Cti
MH cd
*Ö 43 •3 4J Cd CU 4-1 cd 4J M-l Cd 4-1 CU 4-1 CO 4J
H
co p P H rH CJ u *-> P CU P P co P CU
CU
3 CU •H U rH CO •H •rl CU •rl > CU CU
CO +J cd <u
4-1 n3 CU P CU
U U CU u CJ CU 00 rl
co 3 P cd' P
cd cu
O
CO
CO
U /-N
« IS
o
!U
•rl CU
CU
p co o
|S
CO
o 43
4-1
P 4-1
£
OH
Cd CJ
CU 3
f!
0) &
u O •rl O
o
u
I
rl
60 U o£ CO u 3 1? •H P CO
o co cu p
«4-4 cu p 4J M-l
 O
MH
CO 3 4-1
4-1
CO 4-1 O '
U
O rl
•H CO >> p CU Ö MH •H cd P MH U p p P p
»I
0) o CJ 4-1
U CU 4-1 H >% 3 a> 3 rH O
CO iH
> cd
ri
•rl 8 H CJ 43
cd 3 CJ cd CU
43 P4 B rH CO
4J Cd CO cd
. Ü rH 3 6 rj
cd M 4-1 Cd
CD
CU
bO U CO 4-1 43
•rl 4J
P iH e< 4? CU -H rH rH •H 4J P fU CU •H 4J CU •rl P •H CD
cd co M 4-1 cd !> P U •rl 43 & •rl & > P O rl > p cd P P P CO
cd cd
M o O 3 'rl p cd o 3 3
, Cd MH CU o P cd u cd p Cd rH Cd O
;? P o CJ O H -1 43
CJ M-l P H cj o 43 P-i H o p PH CJ PH CJ CJ H CJ cj <!
I 1 I
CO CO CO CD to CO CD
CO
CU CU CU CU CU cu cu CU
rl u rl rl u rl u U
§ 3 3 3 3 3
3
00 60 8> 00
•rl
00
•rl
bO
•rl
bO
•H
P 3 rj r-i O H
cd cd
c o rt cd Cd
P oo
P
g
4J
(50
o MH p
■w
CO
o
cu o\
CN U
CU
E
CU
rl
cu
fi
CU CU CU
rl
CU
4J 4J 4-1 4-1 4-1 4H
CO Oi CO o 4-1
0) rl
fc •3 p.
(-1 P
M
MH
O
P
H
P
H a H
P P
H H
P
co CD CO CO CD CO
CO /"\ co CO
cu CU 0) CU
•rl P
■8
■3 CU CU CU
ü a) cu u rl rl u rl rl rl
P
CO
4J
CJ
3 T3
00 CU
3
t>0
•H
3
00
•rl
3
60
•rl
3
60
•H
s>
•rl
3
bO
•H
3
bO
•H
P •H rl
•H
CU
•n
D 4-1
MH -3 8>. O
oo
M-l MH MH MH MH
U P H § » cd
■H
cd
rH rH
cd
H
Id
H
cd
rH
cd H
P. 3 cd 43 p cd
P p m
p "—' o rH P 3 u
ON ß ■H rl rl fi rl
CO oo cd OJCOH CU CU cu CU 0) CU
4-> 4-1 4-1 4-1
tf o rH 4J
QJ O
U r-i
E H
P •rl
s H
P p P
5 P
H
H 4J
H "rl .-H
H a cd
4-1
4J 4-1 CU 3
m > O
p P CU
•rl •rl
a) co p M cd cd rH 4-1
p H H CJ 4J rH 4-1
H 4J cd
P 01 CJ cd
43 p p cd P P >". 3 rl
cd cd cd w CJ CU cd
H u i-j
00
MH
4J
p >> is
u
CU
IS 4-1 CU
U
cu H a
tj
MH
rl
o
■rl H
T3 .cd u
u o
O
P
oo o cu M O o P 3 H •H cd bO cd co CU o
P Fi U P 4J p. CU MH (U rl cu P rl O
IS
•rl P. QJ p ti , 3 P rH >rl p
o 4J
o P rl 0) rl P u MH CU U CO rH CO
I
P & 3 CO CU 'rl Pi
„ P H cu cd e cd O cd
B cu fll CU P CU H CU rl s
p
P CU
CJ
> -3
 CJ H •rl rH CU H cd •rl p
co < Cl) •H CJ 3 CJ > u cu •H 4-1 O H 
rH -3 H H
•~\
rH rH
IN CO
15
ÖO
o
X) CD >> u rH
d i > rO CD 4H O
•k CO u 4-1 cd •rl X) ft d
ÖO rl CD o u 4J js d
0 •> (1) 4-J X) CD CO n a CJ 3 >> CJ l
•H C ft d Ö CD > (ü cd cd u i-i V CD
o ... cd cd co d H ,
H O X
CD -H tu
0) 4-1
•H O
•rl
O
d
cd
o •rl
X)
3
CD
CJ
•rl
X)
Q) CD CD
CO 4-1
CD
ÖOX)
I
O
,0 Cd 4J r?S > cd co 3 Cd ß3 0) 3 <J d ft
ÖO d 4-1CD XI 00 CD d u CD 3
rl 4J cd
^•rl CD 4-1 CO J-l cd ■rl «H 4-1 d •H 4H J-l
4-1 4-> ß d 3 «H 4J H x) a) co 4-1 4-J d CD cd
CD CO tu cd X> XI <4-l d ^4-1 co co 4-> CJ CO CD co
co 4H CD ÖOi-l d o . cd 4-1 co o CD d d CD rH CD
A! -cd > cd a •rl 4-1 -rl 4H CD CO CJ (D
n co d c d CO 4-1 H d Ö -H d 3 >,
cd •H cd M o o d O cd o x) X. Ü CD d o
eCD o
1
4H ß CD -rl •rl d CO -rl CD CJ co co rH
o * £ 4J fr.
PH . c co o ca
4-1 o Ö ft M ft -rl ÖO
4-1
cd co
•> ÖO 4J
d Cd
CD -H öo
'rl X) cü
4-J r-l CD ß
cd CJHÄ
O 4-1 4J H CD
,ft cd cd cd Ü H CD O
CO 4-1 H CD rl Cd CD
i"
Ö -H H •H -rl ft •rl ÖO •rl Ü CO •H d CD CO ft-rl X) ÖO
0) 4-1 4-J rl O 4-J -H 4-1 >,T-I X) CD CO u CD d X)
e cj 4J Cd -H CD 4-1 rH CO 4-J CJ H B 3 4J rl -H CD
CD CD cd CD d CD d <D CD CO CD CD •H d 4J
CJ ft U rH «rl > 3 > i^H Cd 4J H 3 -rl Cd
cd co (DOS d O CD CD CD 4-1 CO ft CT1 Cd H
H d ft 3 X) H CJ X) > 3 co cd ß CD rl CD
d O ja cd CD 4-1 W ft 4-1 4-J pj 4J rl
I
d
o 4-1
•rl c
rH
H S 4-1
•H d tu
e i-i 00
CD X)
o > 3 rH vD CTi O C^ CO
o O ,0 cy> CO CN co
ÖO 00
4J
O 4H
CT\ Cd
•H
k X)
5
4J
3o
o xi
tu
4i 4-1
CD
ÖO
cd 13
u 3 vo <y> CM CO CN
CD CO CM oo CO
d 00
CD CTi O
o co ß
co
x) i jä
d ,ö co CJ
CD cd CD rH rl CO
4J d 4-J U cd cd o
0) > cd CD CD :
co -i-i U >-. CD co J-l "> .
> 4-J u B •rl •rl
4J u CO 4-J X) 4-1 4-1
d CD co d CO cd
cd ft
H 3
II
r-l
•3d 3 O
X) -H
CD
4-1
CD
4-1
u
4-J
ft CO •rl d 4J CO CO
co •H O cd 4H •H
'n d u o rl d S 15 O d
CD
&
O
ft
o
-H
4J
cd
ÖO
CD Ti
> 4J
o cd
ft u ft cd
I ÖO
o
•rl
4->
H
CD
4H
3 ft
O
M
4J
CD
>
•rl
4-1 co
4-1
d
9i
B
ß
•rl
X)
cd
U -rl u CO rl -rl u d CJ CD CD rH
Cd 4J cd •rl Cd 4-1 cd CD cd rl o cd CO
4-1 CD co CD CD CO ■CD o 3 d H
H CD •5 CD rH
CJ
B
,d
co
•H co cd
XI cd >
rl
CD
d
'
4-1
cd
d 3 -H CD XJ CD o
•rl jz; cd •rl !3 rH 'S 0 < ü ■ H
CO in vO oo
16
r-l
U 4H •rl 4-1 !>, (50 >>
co cd o )-( O H d 4J 4J
CO -rl d
0) "O CU cd rl •rl CJ
CU 4J rH CU
•H ßrl d cu d d tu cd CO -rl
4-1 cd CJ o X -H o flIHH- cd u ja d -H |
d T) -H 4J T3 •rl 4J o d CU CU 4J 4J cd ja d
I -rl ^ ja ü rH Cd ÖO
U <-i C0 d d 4-) 4-1 rl 4-J 4-1 ÖO
4-> Cd CJ O CU 0 CJ O > CU CU 4J CU 4-4 P. -rl 'rl
d -rl O rH CO
co cj dm d H d rl 4J rl ja i-i
cd gj o d cu d
«*d •d d CU iH 4-1 ÖO 0) . u cu d
>S4-I "O . to o o ja o o o xi d co d o d CU U O
CJ 4-1 •rl >-> 0) M •rl ■& -rl n •H g «H d cd co tu
Ö0 4J 4J CJ >H
d cd CO -H' 4J
°
4-1
sb"^ &>
•rl O 4J
4J :P. d CO
CO •rl CÖ M d iH -rl N d n •H N d > -rl cd J3 Cd Cd 4J
CJ rH •rl r-l •rl o •H OH •d p. CJ V4 rH CD
d CU d •rl CU -H tU S 4-1 co •rl Cd P. CU
o u ft-rl ft O CJ «rl CU d •rl Cü
•rl 4J «H Cj J3 u
tgîj u &0J3 CJ J* rl
•ri a
O O CU M CJ.
•rl rH H
4J
4J 4H H > Cd 4J CU CU 4-1 H 4J tu (U 4J Cd rl •rl
o d CD >. 4H , -d u 0 *d H : 4-1 o du w +J u cj cu d
U CU
O •H & cd cu u O 4-1 rl o &
.4-1 IS er cd d •rl cu. d > o
> d O -rl
ft o ft XI +J O B d o O 0 d Ö 4-j. CU CU rH
P. 4J
d co ft 4J d 4H •rl 0 4-1 •H d o M co e •rl
CU Q) 4-i cd "Ö CO 4J T3 co cd 4-1 4J ja CJ U 4H Cd
«J co d >> 4J H co o co cd O O U CJ
H a -H ÖO O d H t0 d !* 4J 4-4 Cd 'rl /"\
co CU 4-1 d •rl ft CD cd rl d CU CU cd rl d P. CU CU CU -rl 4-1 CU 4H co
> •H rl C0 1-4 4-> CU 4-> u 4J -,
OrlU 4-i co u d co C0 T3 d d cQ to "d U d d «4-|i ja u CO 4J H «rl 4J
ÖO
d -H < «H CO o
fr u CO 4-1
cu cd
d CU o
!a > CJ
•a 4-1 •H rl P- 0 W O CU p. H «V
u 4-1 •H rl CJ
H
O
H <4-l
I l 1 1
e
H
d
O d
4J -rl
6 *->
3 ja - . A A
<3 o o cd MO
■
-m
*
n
o ÖO
4-1
CN
00 CN
fO rH
CU H
+J O 4-1
•rl •' 0\ Cd
>4 H
•rl ft T3
cd v^
&
o
u ÖO
rl 13
0) QJ
cd 4-1
•H <D
in 00 m
CJ ÖO
tu T3 4J CN CM es oo
00 O oo CTN vo
ft 3 CM
ft ft A
CD J3 ro oo rH
o CM r-H
QJ CJ o\ 0 CM
CJ d 00 CÖ
ü g co
C/3 TO 1
co tu
1 4J •H d
tu *n t-i d 1 (U 4J o
cd cd CJ ja •H >, co •rl
%»
o tu
CJ
d
-rl
cd •H u CU 4J rH 4J 4J . 4J
cd i CJ ÖO 4H •rl d CU tu ja
CJ O T3 -rl ft 4-j >. U CJ cd -d cd CJ d -H
ft-H 4H 'H CD Cd 4H 4J
•H cd l CO d 4J
H -H d
rl Ö co cd -H jd •o cd 4-1
(U rH d ■o & tu rl 1 4J d ft rl CU
4J O 0) 4J
CJ -rl 4J d 4J ft-rl 4-1 ft tu d d d
öü /-N
H rl CO rl CO U «rl T3 ÖO •H 0 1
Cd O ÖO Q) > •H d J3 ft CO ft ÖO > >> •H H
H cd o CD 4J •H O 4J co H
ft U & ■ ■
U CD p. TH d H
H CU CO CU •d o co 1 u CJ d
rH o SJ
H -d & cu |* •H ft CU d JS CO •rl Cd
U M
Q) d
& o CJ s-/
6 n b cd •H •H 4J •rl
cd
rl -H
Ja
ft d ft 1 rl 4J 4J
Ü
ÖO 0
ftcu
Ü
cd co co
co d cd -H X) 4J CO rl
CD H -H 4J 4J
CU t>>
I CJ cd
•rl
U B
CU
CJ •rl CU H •H
•H 1 H TH cd
rl CO Ü CJ i
d cu
Cd rl
rl d
CU
15
O
i-\ 4J
<Ü -H
• u tu ü
Cd rl d
^'H
w >
3 0)
Sa e
rH
VO>
"
CM fO
17
u
ÜD cd u CO 4-1
C 111 OJ 00 cu MH CO CO -H 4-1 CO
•rl rH £ ß M o 4-1 O MH . •rl ß rH -rl 4-1 ji ,0 •rl
CJ a) 4-1 -rl CJ >i o M •rl 4-1 rH co ft > •rl CO 4-1 rH
>-.'rl 4-1 CO > •H rH cu 4-1 a M-I & 4-1 ß CJ cd cd 13 MH
4J CO •rl
o n 1
4-1
4J
CO •rl
>> cd cu
ß T3 a ß
CJ CU
CO ft
o
CJ •H
CU > 4-J 4J
>> 5
ftMH
cd o
H rH
MH 4J
co
s
CU
CU 4-1 4-1
•rl CD 4-1 !* ß cd r-l 4J CO 4-1 O ß cd ß o ft ß CU O ft CU CU 4J CJ ß
CJ XI 2 w cd •H CO cu •rl Cd > •rl •rl XI CO cu cd MH u r"> CO J3 co CU cd
•H C U e «H rH ,0 CJ 3 -H i« 60 XI 00 ß ß ,ß rH cd Cd 4-1 u rQ CU 4-1 CU >, ß H
H Id o o !> ft Ü cd fl n o cu >-, cd co CU cd •rl 4-1 ft co CU ß CU •rl J3 CO ß ft
-O. MH rl -H 3 «4-1 ß «rl U H cu rl rH CU IS ß 13 MH 4-1 o
ft 4-1 ri co •H CJ CU ft rl 4-1 4-1 >> ß U 13 CJ x) O 3 O O u CJ rl
ft 4-1 CO CJ 01 M rH CU ft o •rl CU ß 4-1 •rl CU ß ß -H ft rl ,o rl cu S!
•rl 4-1 a) cd & " ß u co cd cj ß MH CO CJ cu •rl & cd ß co 4J o 4-1 CO &
CO •H > CO 5 rl co CO ß 4H CO cu o
•u
£B •H
CU 0] O O ■> cd CU 3 91 13 cd •rl ft
•rl CU H •rl CO •H r-l O 4-1 CU 4-1 o ,w 4-1 •H 4-1 ß cu •rl Q 6 CO CJ
O ft rl 4J cu >►> u CO CO •H cd CO •H Ü 4J u •rl •H CU H 4J ß CU CU •rl ß rl
CU 0) o cd -H CJ X) ft CU CU ■u rl rH cd •rl cd O XI ,3 CJ •H O > rl 13 CU rt
o (-1 Ü -H 01 H o •rl rl ,fi u u •rl o u ,D o H cu •rl U 4J ß H ß o •rl M CU
H r-l u CO CU 4J a) cu CO s> cd •H H H cd ß •H 0 rl 3 cd rl H
1 b
•rl »rl •rl 1
4J 4-1 rQ & > 4J 13 CJ Ü ,0 Ml rl CJ 4J ft CT 9! CJ
ca ,ö O 3 3 cd 4H ß 3
la 4J ft CO P4 ß M-i u w cu -H ft H CO CJ & U MH ß M MH o <4H cd H rl o CU ß
i f ? 1
m o\ o cc en O m 00 CN co i-\ CO
m vo n CO CM VO r-- vO CM o . . CO
-* vO u-i H vD CN vO VO O <r CO
r- <r CN r^ m co
o\ CO CO >d- oo rH o r~ CO O 0\ , vO
rH 00 m CO r~- m CN rH CO r-{ 0\
m oo CM CO o H o> r^ vO o-ft ■
r~-
n A A n «k A §1 •t
VO CM CM m CM m CM rH
CO
ß ß cu
O o rl >.
CU 13 •H •rl 3 4-1
4-1 CU bO 60 CO CU
•rl 4-> CU CU ß cd MH
co ß cd ß U rl !^ O CU cd
o H o >. 13 >M3 -rl B co xi
CU -H cu •rl 4J CU CU -H 4J •rl 4-1 •rl
O 4J f-l 4-1 o cu 60 cd 3 bf >, 3 3 cd ß ft < CU ß ß ß
H O -H CD rH •rl CO CO O ß 13 o co CO •H CO H MH Cd ■> CO CO 4-1 U CO co 4-1 CO «H ^ CO ft rl CO
u cu o CJ CU O ß rl CJ > CJ cd a) 4J ß CU CO cd ß >M3 13 XI rl S ß
> & ß rH ß H >, 3 CO CO £ o rH cd 4-1 -H ß -H a) cu cd
O O vH 4J CU •rl 4-1 ftl3 o o O CJ ft rl -H Cd cd cd & rl
^3 ft ß 3 ß •H ftMH CU ft bO > 1 I O CO M
cd ß cu MH ß CU u CO CJ ß rl •H ß >, ß ft cu
CJ o B o B CU ,40 CJ >> ß o o CU CO 4J -H 4J -H •rl CO
•H ß & 3 •H rl Cd •H -H
5£
.CO -H -rl ß rl 4J a i •rl 1 U 4-1 CU
Cd rl 4J W) cd 4J bO b CO rl 4J 4J rl 4J ß cd co O CO Cd -H M
4J O -H cu 0 'H rl ft 4-1 CO CO 4-1 O U ft CU O 4J •rl 4-1 <U rH ß
cu u S co rH B to cu co CJ ß -rl CJ e CU S TH ß H ß rl 'H 10
e cu o ß Ü O ß j3 rl CU CU 13 CO CU 0 rß rl CU 13 cd ,a cd CJ CJ cd
Cd H rl O 3 rl O 4-1 cd u rH ß CO rH U 4J cd rl U CO 3 rl 3 cd cu
OT W ft CJ 55 ft O o cu 3 W >H Cd W ft O cu 3 W bO PM bO S3 MH B
rH CO rH co
•-N y^s »^N ^^ CJ cd ^™\ ^"\ ^^ Ü cd /"^ •-s /—\
<N CO -* m 3 9i H CN CO 3 CU H CS co
v—' ^^ S^
S3 B *~s
s-/ \s ^-^ (3 B ^ ^^ N^
m vo
18
B B
fl) a)
>. 4J
S*
>1
60 B B B
u O °{ o
m •H o Ö •H
B •d
rH P<rH
w IH g O rH
•H o rH •H
o 0 fi OJ S
•H o >
S sr u (U CO
o • M P •
u r^. H
< N-* 14-1 »—'
o fi
rH •-s ed
CO —/
+J
0)
N
4J
cd 53
3
•H
B
(U
B a) P
CJ
B cd puO
+J a) M o w
ß on rt 0 o
H <! o u v-X
1 1
CM \0
CM 00 00 vO
rH O O
#1 A A
CM CM CO
vD
rH
r- o oo m
vo oo o\ oo
CM <Tv o
n. #i
rH CM
!*
60 CD
rl B
aj o
B -H ' •
CJ 4->
3
CJ ,H
•H vH
6 rl
o -u
JJ B
cd O
CJ
rH
cd B
B O
O -H
rl >H +J
<U -u cd
,S Cd N
+J B -rl
O C B rl
CJ cd CJ cd
/-v 4J 60
.-■
,fi
+J O
4J
<r B u
^ H O o H
oo o\
19
ADVANCED MATERIALS
Gas Pressure-Sintered (MoSi2~SiC)/TiAl FGM
91FE0069A Tokyo KEISHA KINO ZAIRYO KENKYUKAI in Japanese 22 May 90 pp 1-5
[A series of 19 slides used by Junzo Fujioka, Kawasaki Heavy Industries Ltd.,

for his presentation at the Ninth Workshop of the Functionally Gradient
Material Research Association, 22 May 1990]
[Text]
No. 1
The Ninth Workshop, Functionally Gradient Material Research Association

(22 May 1990, Tokyo)
Development of (MoSi2-SiC)/TiAl Functionally Gradient Material

by Gas Pressure Sintering Method
Junzo Fujioka and Yuji Matsuzaki, Akashi Technical Research

Institute, Kawasaki Heavy Industries Ltd.
No. 2
Schematic Diagram of Gas Pressure

Sintering Method (as developed by the
Industrial Research Institute, Osaka
University)
V///////////A
20
No. 3
(5)
(9)
}§B£Ti^<7)Btf>SoI (7) (8) TiB,«aS*
'(10)
Illustration of Mechanisms of Gas Pressure Sintering Method
Key:
1. Combustion wave 6/ Rapid dispersion of B into molten Ti
2. Raw material mixture 7. Nuclei creation
3. Preliminary heating zone 8. Growth
4. Reaction zone 9. Grain bonding
5. Sintering zone 10. TiB2 ceramic material
No. 4
Characteristics of Gas Pressure Sintering Method
1. High-melting point materials (materials not to be readily sintered),

such as carbides, borides, silicides and nitrides, can be sintered
without a sintering aid.
1
2. High-density ceramics can be produced instantaneously.
3. With high energy efficiency due to isotröpic pressure application,

materials with complex shapes can be produced.
4. By placing an ignition agent around a material to be sintered, a high-

melting point material and a metal can be simultaneously sintered,
i.e., a functionally gradient material can be produced by combining
wide-ranging materials.
21
No. 5
Characteristics Required for Thermal Stress-Alleviating

Functionally Gradient Materials (FGMs)
"— i
High-Temperature Side _ „. .... ,, TTM Low-Temperature Side

Material Material
(1) Heat-shielding (1) Thermal stress (1) Thermal conduc-

capability alleviating tivity
(2) Heat stability function (2) Strength
(3) Oxidation (2) Thermodynamic (3) Toughness
resistance stability
22
No. 6
Prospective High-Temperature Side Materials for Thermal Stress-Alleviating

FGMs
Auto-
Heat stability: highest
fillegible exponential]
exo- • «*
coefficient, x 10 /K
Thermal conductivity,
Specific heat, J/g'K

thermic rH CO i
•H
use temperature, °C
Thermal expansion-Ö,
Pd
reac- to O CO
c co
Melting point, °C
tion a) a)
(R.T. to 1100°C)
1 CO
4J
•• Si00 1
s r-t Remarks
reaction AH ,
,C0 00
10 ß
•H o g
4-J !
00 •• *o
Ä C
e <1)<N
ö
Heats of
A
r-t 4-> 4 3
PM
• 70IOI5BBRWB / •} \
o 311 4.5 SI» 1100' 4.1-5.2 M at.) o.«~i.s 0..-7« 52» S—10. • isore«ZT*acufc v -5 /
T* 70 uoirc 274 aiore)
1549 Kft»B**i«5orasMTr //\

«50 5.0 42 OUT.) 0.7-1.3 a. '171 3Ii 2-5
sic 0 74 S.2 IMO-Cftt
is anor:
**»'< 1 a..'310 aT'300-500^
C »4 (1200D '•"IMllJUMÜBU» /r\

A" 751 3.1 1)110 1000—1400 3.4 10 B.T.) 0.1-1.3 o... »»-490 55-213 5-4 saxaucs: ^J,J
SI.». 75—4» (1200T) &T^«o-4ore
• •• Kt (1200-0
*SI. 0 10» (.1 1030 1(50 1.3 4» CLt.)
17 (110ITC
0.5 0..-M4
si aiorc)
421 »
■ i4orcaiT*a=eft
(6)
0.1 a ■ .245—304 ••■•sic toa**s /_.
Hi 4.0 2050 175C—U80 7~» U B.T.) (0.1-1.1) 11 U4.D-C) 343-392 3-5 SSI0,.4il*JC-Ul,M3SIC f / )
Alt«« A" v
t uoooc «..■245-53» B9932S: AT.200-C '
1 ••••TIC t»a**«t /nv

LI 1150 iioo 7-10 2 <LI.) 0.5-0.7 c ..'174-784 147-19» •-15 TI0..Zr>C -Zr0,«TlC ( Ö )
2/0i A"
2.3UC4HTC ?sz s*aaeAT-3oo-40c^ v
'
a. -21»
TIC o 751 4.1 3257 1.0 32 CLT.) 0.5-0.» 255 Q200-C) 451 a***E-w**i« (9)
40 (11001: O..-2B0-I5I
': me» (in:B««*2t *S(l»Sm.lO »}S*:7T4 5'-fc?5»?X (19H.BHX)

■nn. ttm>mSS5mmwm am.*** ere*, ♦»■"■ww «"S'-""5"' " 1 dJ/i-ci-T •411 »/••I
l-UJl OW*: IIJWH* (1HZMUB OTT*: «ST?* 8 (IN» (.4M lul ■4.2 Joil
' etawE: r-»r»«iL««*««3t (lin.Bysimü «.rnur » u.:5tt in. c.^. c.»«. *ur. ins .».sis i itr/u* • I.I »i
SlIlO. *»l:*«B4t*1» Cl»4.Bf!XMriBMt) •E.A.C.tl.r n >L:C<ri>. lu. Ml. free ( QMS) 7/1 .715 lid/«"« • 0.31 JOVi""
Key:
1. Sublimation started at 2200°C
2. 1900 (Sublimation)
3. •Material good for fundamental studies on FGMs
•Rapidly oxidizes at about 1300°C
•Highest use temperature in atmosphere is 900°C
4. Usable up to 1650°C in an oxidizing atmosphere
•Thermal shock temperature gap: AT = 300 to 500°C
5. (*l)Difficult to be simultaneously synthesized and sintered
Thermal shock temperature gap: AT = 500 to 600°C
6. •Stable up to 1650°C in an oxidizing atmosphere
•Rapidly softens above 1400°C
(*23Composite formation with SiC:
3Si02 + 4A1 + 3C -> 2A1203 + 3SiC
Thermal shock temperature gap: AT = 200°C
23
8. (*2)Composite formation with TiC:
ÜO2 + Zr + C -> Zr02 + TiC
Thermal shock temperature gap:.,AT = 300 to 400°C
9. Preliminary experiments were conducted with the electric [illegible
character] ignition method
10. *Data were taken from sources listed below:
(1) Sata and Ikeuchi: Journal of Japan Ceramics Industry Association,
95, 2, 243 (1987)
(2) Mohri and Kawashima: "Fire-Proof Materials and Special Heat
Resistant Materials," Seibundo (1964)
(3) Yoshiki: "Fire-Proof Material Engineering," Gihodo (1982)
(4) Samsonoff et al.: "Handbook on High Melting Point Compounds,"
Nisso Tsushinsha (1977)
(5) Yamaguchi and Magoshi: "Intermetallic Compounds," Nikkan Kogyo
Shimbunsha (1984)
(6) Motoki: "Fine Ceramics," Gihodo (1976)
(7) Miyamoto and Koizumi: Nikkei Material, 13 Jul 86, p 79
(8) Miyamoto: Ceramics, 22, 6, 489 (1987)
No. 7
7 0
Metallic
material
C'/AI (next generation) Intermetallic"
compound
FRM
FRC.CC
jiC/Ai (next generation)
:vn (next generation)
500 1000 1500 2000

Temperature (°C)
Relationship Between Specific Strength and Temperature of Prospective Low

Temperature Side Materials of Thermal Stress Alleviating FGMs
24
No. 8
Production Method and Development Needs
Production method Development needs
1 Raw material powder lamination (1) Optimal gradient distribution

(2) Powder lamination technology
•Powder: Mo, Si, MoSi2» SiC, 'Automatic powder lamination
TiAl •Doctor blade method
2 Molding and hardening (1) Selection of a binder

(when the shape of a material is
complex)
•CIP (2) Binder removal technology
3 Encapsulation (1) Vacuum encapsulation technology

•High-temperature encapsulation
•Glass capsule in a vacuum furnace
4 Gas pressure sintering (1) Temperature control technology

______ 1 •Ignition technology
•Heater-heating: temperature rise
*HIP device 'Dilute powder addition: tempera-
ture lowering
■
(2) Low pressure sintering technology

:
.■
(3) Larger, more complex shape forma-

tion
No. 9
|*-!= >H
Distribution parameter : n = 1
Maximum specific stress: 5.1
Heat flux : 2.6 MW/m"2
MBIERIflL-fl . MoSl2
—.— x-e mTERIBL-8 i Tlfll
, x-i N • I.B
, IMX 1ST ' 4
_._ MW lTTfE - ««*
mres« 2
1FUC ■ I
THICK -
TB - 2293.88689
; TEMP.
» FUÄ
S1RESS-2
- OENSITY
Design Results for (MoSi2-SiC)/TiAl-Base FGM (0-Order Design Based on

Literature Data)
. '25
No. 10
Before vacuum After vacuum

encapsulation encapsulation
Before vacuum After vacuum

encapsulation encapsulation
Encapsulation With Petri Dish
No. 11
MoSl2-TiAl Base
MoSl2 Side TlAl Side

surface surface
MoSi2 (33vdl%SiC)
—TiAl- Base
30mm
MoSia-SiC side TiAl side

surface surface
FGM 30 mm in Diameter (showing SiC-addition effect)
26
No. 12
A Theoretical Density
o Density
3. OH
2.5
0 20 40 60 80 100
molXTiAl
Density of (MoSi2-SiC)-TiAl (0 to 100 weight percent) (NFGM)
No. 13
0%T iAl
MoSia, SiC
^JL^.'/J:.
25mol%TiAI
Mo (Si, Al) 2,
SiC
_WbJL\ : ;;
50mol%TiAl
Mo(Si;AlJV
Ti 5S i 3 •
75mol%TiAl
T i 5S i 3,
JMS I At !''■' JN^JA*.
Al3Ti
100%TiAl
TiAi'.'.TisAl:
; *■&
30 40 50
28 (degree)
Microstructure and X-Ray Composition of NFGM
27
No. 14
Compositional distribution
parameter n = 2
MoSiv-SiC T!Al
Surface Surface
Compositional distribution
parameter n = 0.5
Compositional distribution function
f(x) = (x/t)n,
f(x): Mol percent of TiAl
x: Distance of FGM layer from
MoSi2-SiC TiAl surface
Surface Surface t: FGM layer thickness
FGM 30 mm in Diameter (showing the effect of compositional distribution

function)
No. 15
Fll« *•. ; ll LtMC ANALYSIS

iMOi»
» ii M urn. * •■#«
»I. II.» fcl
■ '
i0
V
Element Concentration Distribution at Cross-Sectional Plane of FGM 30 mm in
Diameter
(Left graph) Compositional distribution parameter n = 2

(Right graph) Compositional distribution parameter n = 0.5
28
No. 16
FGM
Ti-containing Ag solder
o Solder
Cu plate (1 mmt)
Samples for High-Temperature Gap Tests
29
No. 17
* 30
FGM-
OFHC Cu;
LN2 or LH2
FGM Sample Mounted on High-Temperature Gap Test Apparatus

(Akinaga Kumagawa: Sixth Workshop Paper, FGM Research Association, 21 June 1989)
No. 18
16
14 .-**'
12-]
10 ■•+ +
8-
W D 0%TiAI
H 6" •+ 2(hol%TiAl
O 4 O 40«ol%TiAl
A 60mol%TiAl
X 80aol%TiAl
2 V 100%TiAl
0
0. 200 400 600 800 1000
Temperature PC)
Average Thermal Expansion Coefficient of NFGMs
30
No. 19
■—•
us
Rfl
D 0%TiAl
_B. 70 + 20wl«TiAl
0 40nolKTiAl
^ 60 A 60«)l%TiAl
X 80wl«TiAl
50 V 100%TiAl
>
40
o
=3
T3 30
a
o
<_3 20
«a
B
10
05 0 —i 1 1 1 1 r——i—
200 400 600 800 1000
Temperature fC)
Thermal Conductivity of NFGMs (MoSi2: 10W/(m*K) at 1600 to 1.800K)
31
ADVANCED MATERIALS
Fabrication Process, Evaluation of FGMs
90FE0317A Tokyo SHIN SOZAI 21 SEKI FORUM in Japanese 24 Apr 90

pp 29-34
[Article written by Ryuzo Watanabe, Material Processing Department,

Faculty of Engineering, Tohoku University]
[Text] Introduction
Functions most suitable for a material to use environments may be

distributed by creating gradients in the internal composition and
microstructure of the material. If not, whatever material function
one desires may be manifested through gradient formation of the
material's composition and structure.
Functionally gradient materials (FGMs) were born as a product of these

ideas. In other words, various different materials, including metals,
ceramics and plastics, are combined, and gradients are formed in the
mixtures' compositions and structures, in order to gradiently
distribute material properties for functional manifestation. Industrial
material fields that are expected to benefit from the application of
FGMs are wide-ranging, as shown in Figure 1. In particular, applications
of such functions as environmental stability, joint formation,
electromagnetic, opto-acoustic, chemical and biocompatibility, are
anticipated with great interest.
32
t H- p H-
K- °
a n
3 ro *
_&
u1 a
MI a n
ro re
H H-
3 3
o a- rt
H Mi
MI g o B) o
M> a*
rt>
H
H- tn
0 rt o
S3to
ri-
a I e tt H*
a. a
*d MiT3 F- O
, H H n 3 H
H. O O i-l
(o n tt H- i-t tt rt ro
M rt rt (t rt rt ft W rt t-" H> 3
rt H- F- 3 rt W* H- 3 F> O rt H- (D H- tt ft
H' n O rt 0> 3
'•4 O
3 tt D>
OOTJOÖHM î? ^ i n n ro >rJ '» tfl a W

OK tfl ft 2 W K W KW Ä rt rt O H- !«
w rt TJ 3 ro O F* O F- (t> F> O 3^ H- O M
ftH O F- tttt ttC täC tt 3 ro ro ro *ö H. & tt
3* ID 3 trirjq rtoq woo tn (W 3 (U rt d Hi oo ro n to new
rt O * tn 3 00 C OO
rlrttfrtH'rfÜft * g tt - n rt F* I F* rt r) ft O. in O
rt F- rt F- 3 H« rt H- F- r-> rt
H- o H- o tt n H' n Ort H>
rt I <
tt Hi S S rt H- Mi O
HCCOH'SOH
MI e_i Mi [fl M
c o . 3 rt
3 tt <
3
3 H- n en-
F- S O F- 3
O TJ 3 rt 3
3 tt W ^ ro
2
ro ro ro ro
p. en tn
>-t
tn
M
tn
tn 33
31
-JR
T F-
m Cfl to «
rt rt rt
to {U (U H- ro
3 3 3 3
O 3
ro ro ro ro
O O O o CO F-
VJ H- n ro Ä O
o
£3 X 3
ro H- O
(n
tn
rt a ►1
rt
O
ttrt --
tf
tt
H
rt 3 to H* ro
»* O F*
Q
H - 3
Today, the development of FGMs targets primarily on ultra-heat-resistant
materials of a thermal stress alleviation type, which can withstand high
temperature gap environments, for ultimate applications for aerospace
Today, the development of FGMs targets primarily on ultra-heat-resistant
planes' fuselage and engine parts. In these applications, an extremely
complex design is required of a material as follows. A heat-shielding
property and heat resistance need to be imparted to the high-temperature
side of a material, while high toughness and joint-formation function
must be imparted to the low-temperature side, and in addition, the
material as a whole needs to manifest a thermal-stress alleviation function
to reduce thermal stresses caused by the temperature gap. There are
many ways to blend different substances at a desired composition and
to form a gradient of the composition in the composite. Methods range
from the vapor deposition method, that deposits an atomic or molecular
layer on top of another, to the method of laminating thin sheets.
Methods currently under development include PVD, CVD, plasma sputtering
method, electrodeposition method, auto-exothermic method, slurry
lamination method and particle arrangement/sintering methodl. Each of
these method is characterized by the material size it can produce and
by the microscopic structure it can form. Judging from today's level
of technologies, we may be able to say that PVD and CVD are for thinner
materials, the particle arrangement/sintering method is for thicker
materials, and other methods are for intermediate materials, This paper
will deal with the preparation and evaluation of thermal
stress-alleviating-type FGMs prepared by the powder process (particle
arrangement/sintering method), which has the highest degree of freedom
in composition control, structural control and shape formation.
Production Process
POWDER SUSPEKS10»
mmne
Figure 2. Particle Spouting Arrangement Device for Compositional

Gradient Formation
34
First of all, raw material selection and material design must be carried
out in order to achieve targeted functions (which may be heat-shielding,
heat-withstanding, toughness and thermal stress alleviating functions).
For raw materials, ceramic powders are used for the high-temperature side,
while metal or plastic powders are used for the low-temperature side. In
some instances, ceramics and metal whiskers are used to control various
thermomechanical properties. Raw material powders will be arranged in
either a step-wise lamination or a continuously changing distribution
according to a blend ratio as designed in advance. Subsequently, the
powders are molded in a metal mold under compression or molded with
hydraulic pressure, and the green compact is sintered with or without
pressure. The author's group developed a particle spouting arrangement
device, as shown in Figure 2, to arrange particles under precision
composition control. Because generally there are significant differences
in sintering temperatures between'ceramic and metallic powders, it
is technically critical to ensure balanced sintering between different
types of powders, as will be discussed in the next section . Also being
developed is a temperature gradient-added sintering device (Figure 3),
which will give a sintering temperature appropriate for each component
raw material, in order to expand the utility range of an available powder
raw material,'and therefore, to be able to manifest functions with a
larger degree of freedom. In order to solve various technical problems
concerning FGM preparation according to the above-described process, the
author's group pursued fundamental studies primarily using combinations
of zirconia ceramics and metals, including nickel, molybdenum and
stainless steel3»4>5.6.
MHWOR'
JMS» *',Bt*,
(1)
Vacutii
Figure 3. Temperature Gradient-Added Sintering Device
Key:
1. Radiation thermometer 2. YAG laser
3. Computer 4. Gas collection port
5. Gas inlet 6. Temperature adjusting device
7. Servovalve 8. Gas coolant
35
Sintering Characteristics Dependent On Composition
Sintering behaviors of a blend of ceramic and metallic powders are

uniquely dependent on the blend ratio of the two powders. In general,
when sintering shrinkage behaviors are not uniform due to a compositional
difference, various flaws will be formed in a compositional
gradient-arranged material. In Figure 4, typical flaws caused by
sintering imbalance are schematically illustrated. Pressurized sintering
methods, such as hot press and HIP methods, are extremely effective in
reducing the influences of sintering balance, or lack of balance, due to
composition.
Warping
Frustum
Splitting
Crack
Figure 4. Flaws Created by Sintering Imbalance
Size of Composition Control Region
Today, the thickness of a layer of one composition is limited to

approximately 100 microns in the case of the step-wise lamination
arrangement packing. Therefore, through arrangement of 10 layers,
approximately 1-mm-thick gradient layer will be formed. On the other
hand, if a particle spouting arrangement device, as shown in Figure 2,
is used, one can freely create compositional distributions ranging
from ceramics to metals within a thickness of 100 to 200 microns. Shown
in Figure 5 is a photograph of the structure of a zirconia/stainless
steel-base thermal stress alleviating-type FGM which was prepared by
the particle spouting arrangement method.
Material Design to Create Thermal Stress Alleviation Function
There are two key points to material design to create.the thermal stress
alleviation function. One is to predetermine the compositional
distribution of a material so as to keep the thermal stress, which is
generated during rapid cooling from a high temperature to room temperature,
sufficiently smaller than the material's strength'; This point is unique
to any high-temperature preparation process starting from powder raw
36
materials. The other is the design of a functional gradient for
alleviating thermal stress due to the temperature gap in the real
environment^, in order to design such a material, one needs data
concerning relationships between composition and various physical
properties (thermal expansion coefficient, thermal conductivity,
elastic constant and fracture strength). Using these data
and assigning a sample shape and size as well as temperature conditions,
a compositional distribution that minimizes thermal stress will be
obtained by the finite element method. It has been found from
previous studies that an optimal design for alleviating thermal stress
during a preparation process is also effective in alleviating thermal
stress due to a temperature gap5. In addition, a study is underway
concerning optimal designs that simultaneously consider thermal stresses
at the preparation and at the actual use.
Structural Change Accompanying Compositional Change
An FGM made by sintering has unique structural transition, as shown in

Figure 6. The structures at both extreme compositions are dispersions
in which a matrix and a dispersed phase are reversed. The intermediate
structure is a network structure in which the structure continually
changes with the composition. The compositional-phase changes
continually and rapidly at a rate of 0.2 to 0.8 ceramic volume fraction®.
Relationships Between Composition and Properties
c , . PSZ/SOS
ttl *CA)0.W»n/2»M*n
£ 20,
at
O
Thermal Expansion
Ol
^V
o
r . _» 1 : I
0 20 40 60 80 100
Volume Fraction of 3YPS2/•/.
Figure 7. Relationship Between Thermal Expansion Coefficient and
Composition for Zirconia/Stainless Steel-Base Sintered Mixture
The relationships between composition and properties, such as thermal

expansion coefficient, thermal conductivity and electrical conductivity,
are shown by reverse-S curves in Figures 7, 8 and 9 [TN: Captions for
Figs. 8 and 9 are reversed], in correspondence with structural transition.
These curves are considered to represent transitional forms of the
37
Maxwell formula, which concerns the two types of dispersed structure
formed at the two extreme compositions, for the intermediate
compositions . On the other hand, properties, such as Young's modulus
and bending strength, appear to behave in a complex manner depending
on structural characteristics or the degree of bonding/debonding
between ceramic and metallic phases3'10.
all U.1*
PSZ/SUS
6 \ 111*»/ J *■ .
^ 0.12 •
>«0.10
2 ' \
3 aos
"D
J 0.06
■ \
•
gao4
a»
£0.02 •—•
ft t t ■ * 1
0 .20 40. 60. 80 100

Volume Fraction of 3YPZ /%
Figure 8. Relationship Between Thermal Conductivity and Composition for
Zirconia/Stainless Steel-Base Sintered Mixture
1200
susaocrsz
A CO 2~10*m:o.02S*m
1000
800
u
600
C
O
O
400
u
t 200
m
0
J
0 20 40 60 80 100
Volume Fraction of 3YPSZ / %
Figure 9. Relationship Between Electrical Conductivity and

Composition for Zirconia/Stainless Steel-Base Sintered Mixture
38
Pliability and Brittleness Transition and SP Test
As the composition changes continually from a metal to a ceramic

material, the material's pliability and brittleness will also change.
These changes can be found through SP (small punch) tests 10,U
An SP test device is schematically presented in Figure 10. The area
(SP energy) bounded by a load-displacement curve obtained by this
test will be a measure for a sample's pliability. Shown in Figure 11
is the relationship between SP energy and composition. The curve shows
that the property changes from pliable nature to complete brittleness as
the ceramics' volume fraction changes from 20 to 80 percent. It is keenly
important to widen the pliable region in this pliability-brittleness
transition for improving fracture toughness of an FGM and increasing
the degree of freedom for designing it. Efforts are being made to widen
the region through structural micronization and blending of metallic
and ceramic whiskers.
LOAD
FLNCHER
Vpptr^W
CLAMPING
SCREWS
STEEL BAU.'
LOWOT 41«
SPECIMEN
(OSxIOslOinn))
Figure 10. Schematic Diagram of Small Punch Tester for FGM Development
39
A (A) J-10*m:o.eM*m
• (O 3#m:i#ra
I
o
k_
c
0- 1
'0 20 40 60 80 100
V
Volume Fraction of 3YPSZ / •/.
Figure 11. Transition from Pliability to Brittleness With Compositional

Change in Zirconia/Stainless Steel-Base Sintered Mixture
Thermal Shock Test
A laser thermal shock test device, which was developed for testing FGMs
for thermal shock resistance, is schematically illustrated in Figure
12 12 The device uses an ÄE sensor to clearly describe a material's
fracture behavior. By using this device, it was discovered that the
zirconia/stainless steel-base FGM, prepared by the particle spouting
method, was approximately three times more resistant to thermal shocks than
pure zirconia (3YPSZ).
CO, User Uns Specimen
AESensor
Translation Stage
Figure 12. Schematic Presentation of Laser Thermal Shock Test for

Thermal Shock Resistance Evaluation of FGMs
High Temperature Gap Subjection Test
A temperature gap tester, which applies a temperature gap between the

front and the back of an FGM sample, has been developed to evaluate its
heat shielding property and thermal stress alleviation function^. The
device is capable of giving a constant thermal load of up to 5 MW/m2
to a sample while cooling the sample's back with liquid nitrogen. In
addition, the author's group devised a burner-heating test method which
can evaluate more readily an FGM's performance against temperature gaps.
40
Thus far, a surface temperature of 1,400K and a temperature gap of 900K
have been obtained with a 4-mm-thick, disc-shaped zirconia/stainless
steel-base sintered FGM.
Epilogue
Three years have passed since the start of the FGM research project
(sponsored by the Science and Technology Promotion Coordination Fund).
It has been found that the first term's target, a 30-mm-diameter, 1 to
10-mm-thick FGM, can be prepared through several routes. FGMs are much
more diverse than conventional, homogeneous materials in all aspects
ranging from material preparation to application. The currently
prevailing cooperation system between design and evaluation experts
is likely to become increasingly important in the future.
References
1. Third FGM Symposium Lecture Collection, FGM Research Association

(1989).
2. Watanabe: Kino Zairyo 8, 932 (1989).
3. Watanabe, Kawasaki and Murahashi: Sozai Busseigaku Zasshi 1, 36

(1988).
4. Kawasaki and Watanabe: Powders and Powder Metallurgy 37, 73 (1990).
5. Watanabe: Ceramics 24, 932 (1989).
6. R. Watanabe, A. Kawasaki and H. Takahaäi: To be published in Proc.

MECAMAT Intern. Confe. Saint-Etienne,France, Nov. (1989).
7. Kawasaki and Watanabe: Journal of Japan Metal Science Society 51, 525
(1987).
8. Uchiyama, Kawasaki and Watanabe: Third FGM Symposium Lecture

Collection, p. 161.
9. Kawasaki and Watanabe: ibid. p. 35.
10. Kawasaki, Watanabe, Cheng and Takahashi: Powders and Powder

Metallurgy 36, 143 (1989).
11. Saito, Cheng, Takahashi, Kawasaki and Watanabe: ibid. 37, 350
(1990).
12. Hashida, Takahashi and Miyawaki: ibid. 37, 307 (1990).
13. Kumakawa, Sasaki, Maeda and Adachi: ibid. 37, 313 (1990).
41
ADVANCED MATERIALS
Thermally Resistant FGMs by CVD Method
90FE0317B Tokyo SHIN SOZAI 21 SEKI FORUM in Japanese 24 Apr 90

pp 35-40
[Article written by Chihiro Kawai et al., Sumitomo Electric Industries,

Ltd.]
[Text] 1. Introduction
Recently, plans to develop ultrahigh speed airframes, including space

shuttles, have been proposed in several countries. Super-heat resistant
materials that have excellent strength and oxidation resistance at high
temperatures are demanded for fuselage thermal insulators and burners of
these airframes.
Carbon fiber-reinforced carbon composites (C/C composites) are one of

the most promising materials for that purpose because of their excellent
specific strength and specific rigidity at temperatures higher than
1,000°C. However, C/C composites have serious deficiencies in oxidation
resistance and the difficulty of ceramic coating for imparting
antioxidation stability. In other words, because C/C'composites have
an extremely small thermal expansion coefficient, a coating layer (SiC)
will show thermal cracks in a post-coating heating cycle, significantly
lowering oxidation resistance.
The objective of this research is to prepare a coated carbon fiber-

reinforced composite with excellent antioxidation stability through
the compositional gradient formation in the coating layer by the CVD
method and the replacement of carbon, the matrix in the C/C composite,
with a ceramic material, to reduce thermal stress between the matrix and
the coating layer.
42
solas! I ff g ä ft
(l) (5) (3)
C-SIC*
00_ 1' r
SiC ■ sie SIC
t : «HS»«*) t f
(C-SiC) (C-SiC)
(6)
CS1C
I 7) t <H*<t>
C/SIC
ric-sic*
Sa**.
sstm sie sie
tt t t
TiC Tie-sic
CT1C (7'
(9)
t «UHU
C/TiC
(10)(«W)
Figure 1. Conceptual Diagrams for Compositional Gradient Coating and

Functional Gradient Matrix Formation To Produce Super Heat Resistant
Structural Material
Key:
1. Control of Residual Stress in Coating Layer
2. Control of Thermal Expansion Rate
3. Formation of Gradient in Matrix
4. C-SiC-Base
5. TiC-SiC-Base
6. (Tensile Stress Reduction)
7. (Weight Reduction)
8. Compression Stress Introduction
9. Self-Restorativeness
10. (Thermal Insulation Capability)
The coating layer's thermal expansion coefficient was controlled by

compositional gradient formation as follows. In order to reduce tensile
stress generated in the surface layer of SiC, a C-SiC-base ceramic
material, which contains carbon with a small thermal expansion
coefficient and a small Young's modulus, is selected for the interface
between the coating and a C/C composite, and compression stress was
intentionally introduced to the surface layer of SiC to make it
extremely strong. In order to control the occurrence of thermal cracks
in the surface layer, a TiC-SiC-base ceramic material, which contains
TiC with a thermal expansion coefficient larger than SiC's, was
selected for the interface.
43
For controlling the substrate's thermal expansion coefficient,
C/Ceramics composites, in which the matrix C of a C/C composite was
replaced with SiC or TiC, were used to increase the substrate's
thermal expansion coefficient.
Conceptual diagrams for the targeted, coated carbon fiber-reinforced

composites are shown in Figure 1.
2. Preparation of FGMs by CVD Method
(1) C-SiC-base Compositional Gradient Coating!'2
The compositional gradient coating by the CVD method needs to be

examined from the standpoints of composition control and
precipitation rate.
The CVD-precipitation phase diagram with raw materials of SiCl^,

CH4 and H2f as shown in Figure 2, was prepared from thermodynamic
data by computer simulation. The compositions of SiCl and CH,
were changed along the line A in Figure 2, and coating was done
with C-SiC-base materials of various compositions.
Figure 2. Equilibrium Precipitation Phase Diagram for SiCl^-

Cfy-H2 System (at 1,400°C and 60 torr)
The relationship between the Si/C mol ratio in a raw material

gas and the composition of a precipitated layer is shown in Figure 3.
The experimental composition was richer in SiC than the thermodynamic
equilibrium composition. The higher the C content in the
precipitated layer, the total precipitation speed became slower.
44
*Q tbtmodrniale
^0.
20 - «xptrlMntal
eoMposltlon
I JU _L
0.4 0.8 1.2 1.6
Sl/C ratio In gas phase
Figure 3. Relationship Between Raw Material Gas Composition

and Precipitated Layer Composition
It is conjectured that the precipitation speed retardation is

due to the slower precipitation speed of C than that of SiC.
Structures of C-SiC-base composite layers of various compositions are
shown in Figure 4. With increasing C content, the structure
became more porous.
A cross sectional structure of a C-SiC-base compositional gradient

layer, which has been coated over a C/C composite, is shown in Figure 5.
By using a C-SiC-base compositional gradient layer as the coating

layer for a C/C composite, the occurrence of thermal cracks in
the outermost layer of SiC was lessened but not completely
eliminated.
This is presumably because (i) the thermal expansion coefficient

of the coating layer is almost the same as that of SiC due to the
lack of orientation by C in the C/SiC composite structure, and <ii) the
composition distribution in the compositional gradient layer is
not optimal due to the slow precipitation speed of C-rich composition
regions.
From the above discussed data, it was clarified that the matrix
would need a greater thermal expansion coefficient in order to
thoroughly display the thermal stress alleviation function.
(2) TiC-SiC Double-Layer Coating3»^,5
The antioxidation stability of a C/C composite can be significantly

improved by coating first a TiC layer over it and then an SiC
layer over the TiC layer.
The antioxidation stability of a C/C composite after double-layer

coating of TiC and SiC is shown in Figure 6 in terms of the weight
lost due to oxidation against the thickness ratio of TiC/SiC in
45
the coating layer. Also, in Figure 7, the residual stress is plotted
against the same thickness ratio.
i i 1 1
OSIC 70-120»» J
*SIC U0-M0iim"|
fa «.05 0.10 0.15 0.20 0.25

3-*-f >ini (TI C/SI C) «MteCDlfc
Figure 6. Antioxidation Stability of TiC/SiC-Double Layer Coated

C/C Composite
Key:
1. Oxidation weight loss, mg/cm2
2. Thickness ratio in coating layers (TiC/SiC)
(1) (3)
T 1 1—
OSIC 70-120« m-
* SIC 180-240« m
SIC coating
SI-0.41
£-0.6
«-0.8
M
-I.O Ö (4)
-1.2 -
_L J. _l_
0 0.05 0.10 0.15 0.20
( 2) 3-7W Vi>W (T i C/S I C) «»«Ott
Figure 7. Residual Stress vs. TiC/SiC Thickness Ratio in Double

Layer Coating
Key:
1. Residual stress in SiC layer, GPa
2. Thickness ratio in coating layers (TiC/SiC)
3. Tensile stress
4. Compression stress
46
The antioxidation stability increases presumably because (i) thermal
crack formation in the SiC surface layer is controlled through
the introduction of compression residual stress in the surface
layer, and (ii) the composite can restore itself by having Ti02, produced
by oxidation reaction of TiC in the internal layers, fill in thermal
cracks, as shown in Figure 8.
However, as long as some thermal cracks remain on the SiC surface

layer, oxidation will slowly advance after each heating cycle. In order
to effectively solve this problem, it is conjectured necessary to make
the matrix's thermal expansion coefficient as close as possible to that
of TiC and to completely prevent thermal crack formation by creating a
TiC-SiC-base compositional gradient layer for the coating layer.
(3) TiC-SiC-base Compositional Gradient Coating^
SiCl
no-depoal
sic+c
TiCl
no-deposit
Figure 9. Equilibrium Precipitation Phase Diagram for SiCl^-

TiCl4- CH4-H2 System (at 1,350°C and 60 torr)
The CVD-precipitation phase diagram with SiCl4, TiCl^, CH2 and

H as raw materials, as shown in Figure 9, was prepared from thermodynamic
data by computer simulation. The compositions of SiCl^ and TiCl4 in a
raw material gas mixture were changed along the line B in Figure 9 to
vary the composition of TiC-SiC-base coating. After each coating, the
composition, density and hardness of the precipitated layer were
examined.
Shown in Figure 10 is the relationship between the Si/Ti mol ratio in

a raw material gas mixture and the composition of a precipitated
layer.
47
100
80 Thermodynamlc
ü 60
+
u
P «0
p Experimental _
20 - composition
0(>
0 0.2 0.4 0.6 0.8 1.0
T!/(TI+SI) atomic ratio In gas phase
Figure 10. Relationship Between Raw Material Gas Composition and

Precipitated Layer Composition
Shown in Figure 11 is the relationship between the composition

of a precipitated layer and the precipitation speed. Unlike the C-
SiC system, the precipitation speed was found not dependent on the
composition but remaining roughly constant. Thus, the composition
distribution of the compositional gradient layer could be
controlled easily.
200
20 »0 100
TiC/CTIC+SiC) molt
Figure 11. Relationship Between Precipitated Layer Composition and

Precipitation Speed
The structures of TiC-SiC composite layers of different compositions

are shown in Figure 12, and the relationship between the
composition and the microscopic hardness is shown in Figure 13. In
contrast with the C-SiC system with a porous structure, this composite
has a dense and high-hardness structure, composed of oriented, prismatic
crystals, over a wide range of composition.
48
2« 40 «0 «0
TiC/ITIOSIC)' mtt
Figure 13. Relationship Between Precipitated Layer Composition and

Hardness
Based on the above-described composition-control technologies,

various carbon fiber-reinforced composites, including C/C composite,
were coated with a TiC-SiC compositional gradient layer, and the
coated materials were examined for their antioxidation stability.
The cross sectional structure of a TiC-SiC compositional gradient layer

is shown in Figure 14.
The antioxidation stability of these coated composites is shown

in Figure 15.
100
C-SiC coating TiC-SiC coating
1
6 80 re* i
Test condition
R.T. M300*C(keep IhrJ- »•R.T.
« 60 Z.Shr
I
40 I
(SIC)
I
I
20 m
n
FCM
I FCM * FCM
(SIC) I
c/c c/sic C/C C/SIC C/TIC
o«0 2.0 3.9 0 2.0 3.9
Figure 15. Antioxidation Stability of Various Coated Carbon

Fiber-Reinforced Composites
By the same reason as that given for the TiC-SiC double layer
coating, the coating of the compositional gradient layer significantly
improved the antioxidation stability of the matrix. [TN: This sentence
is incomplete in the text.]
49
3. Functional Gradient Formation in C/C Composite8
(1) Preparation of C/SiC and C/TiC Composites
It is effective to create a compositional gradient on the surface of a

matrix, when a compositional gradient layer having the thermal
stress alleviation function is to be formed by CVD-coating.
As a basic study of the technology to form a functional gradient

in a C/C composite matrix, we used the powder sintering method to
prepare various carbon fiber-reinforced composites in which the C-matrix
was replaced with SiC; coated the composites with SiC and examined
the effectiveness of the coating by antioxidation stability tests.
The surface structures of the coated layers are shown in Figure

16.
By controlling the thermal expansion coefficient of the matrix, we

were able to coat the SiC layer without forming thermal cracks.
Added to previously mentioned Figure 15 is the antioxidation

stability of a C-SiC coated carbon fiber-reinforced composite. It was
found that the composite's antioxidation stability would significantly
improve by controlling the matrix's thermal expansion coefficient
through the replacement of its C-matrix with SiC.
We have also prepared a C/TiC composite, in which its C-matrix

was replaced with Tie, and are now investigating its combination
with a TiC-SiC compositional gradient layer. Our goal is to further
increase the matrix's expansion coefficient and improve its heat
shielding capability.
(2) Preparation of Compositional Gradient C/C-SiC Composite
We are also currently studying a C/C-SiC compositional gradient composite

in which the matrix component of a C/C composite varies from total carbon
at the internal layer to SiC at the outer layer. Our objective here is
to prepare a carbon fiber-reinforced composite that can maintain excellent
heat resistance and strength at high temperatures as in a C/C composite
and can be coated with an antioxidation layer without forming thermal cracks.
Using electrophoresis, two kinds of prepreg sheet were prepared by

depositing self-sintering coke powder, for one, and SiC powder, for
another, onto pieces of woven carbon fiber cloth. These two sheets
were laminated with various lamination patterns, followed by molding,
compression and baking to produce compositionally gradient carbon
fiber-reinforced composites.
50
One example of the composites is shown in Figure 17. By using this
method, it is comparatively simple to prepare carbon fiber-reinforced
ceramic composites having the matrix of different types of ceramics,
in addition to C and SiC.
4. Future Studies
Our research has for the present established FGM synthesis technology,
and is now at the stage of completing elementary evaluation of FGMs,
including that of antioxidation stability. In order to further develop
research concerning FGMs, it is extremely important to pursue not only
synthetic technology, but also material design based on the evaluation
of basic properties of non-functionally gradient materials (NFGMs).
At the same time, we are also pursuing research concerning the thermal
stability of compositional gradients under a temperature gradient as
well as a concentration gradient in a high temperature region, because
FGMs are the so-called non-homogeneous materials'»8.
References
1. Chihiro Kawai and Yasushi Igarashi: Third FGM Symposium Lecture

Collection (1989) p 79.
2. Chihiro Kawai and Yasushi Igarashi: Lecture Abstracts, Japan Metal

Science Society Fall Meeting (1989) p 721.
3. Chihiro Kawai, Yasushi Igarashi and Koichi Iwata: Lecture

Abstracts, Japan Metal Science Society Spring Meeting (1989).
4. Chihiro Kawai and Yasushi Igarashi: Lecture Draft Collection,

Annual Meeting, Japan Ceramics Association (1989) p 395.
5. Chihiro Kawai and Yasushi Igarashi: Lecture Draft Collection,

Second Fall Symposium, Japan Ceramics Association (1989) p 222.
6. Tomoyuki Wakamatsu, Masanori Sakagami, Makoto Kawase, Chihiro

Kawai and Yasushi Igarashi: Lecture Abstracts, 1989 Fall Meeting,
Powder and Powder Metallurgical Association (1989) p 123.
7. Tomoyasu Aihara, Mikio Kaji and Yasushi Igarashi: Third FGM

Symposium Lecture Collection (1989) p 111.
8. Tomoyasu Aihara, Mikio Kaji and Yasushi Igarashi: Lecture

Abstracts/Fall Meeting, Japan Metal Science Society (1989) p 729.
51
AEROSPACE
Developing Independent Manned Space Technologies
91FE0125A Tokyo REPORT OF MANNED SPACE ACTIVITIES KEY

TECHNOLOGIES RESEARCH WORKING GROUP in Japanese Mar 90 pp 1-12
Table of Contents
Introduction
1. World Trends in Manned Space Activities {not included)
2. Current Status in Japan and the Role of the National Aerospace

Laboratory
3. Accomplishments for Related Technologies at the National Aerospace

Laboratory
4. Manned Space Activity Key Technology System, and the Area in Which
the National Aerospace Laboratory should be working on immediately
5. Detailed Research Plans
Appendix 1. Space Development Plans in the World (not included)
Appendix 2. Summary of the Sections Related to the General Guidelines

for the Space Development Technology Program (not included)
Appendix 3. Summary of Research Activities for the Manned Space Program

at the National Aerospace Laboratory (not included)
Appendix 4. Technology Items in Various Areas, and the Detailed

Research Plans (not included)
52
Introduction
The International Space Station Program (ISS), in which Japan has been
participating, is expected to start in 1997. Worldwide construction of
lunar bases and the manned exploration of Mars have also been seriously
considered. In this environment, it is necessary for Japanese space
development and utilization activities, which have so far been confined
to the unmanned space technology, to be swept rapidly into the larger
current involving manned space activities. Even in the General
Guidelines for Space Development Policy, which were revised last year,
it is stipulated that key technologies shall be established to develop
independent manned space activities unique to Japan. It is expected that
large scale research and developmental activities in this area will .take ■;
place soon.
Under this condition, it is the natural responsibility of the National

Aerospace Laboratory, the sole national laboratory for aerospace
technology in Japan, to assume the leading and central role for the
formulation of key technologies in the nation's manned space program.
Nevertheless, manned space technology represents a field in which the
Japanese have little experience, and its technological content is so ■-, ■;■-;■
vastly wide that it is not easy to grasp its total picture. This is the
reason why the Laboratory has-been; conducting internal discussions as to :.
what types of activities are expected of it, and as to what types of
activities.are made possible by it. The Research Council of,the -\ !:
Laboratory has established the "Manned Space Activities Key Technology i
Research Working Group" consisting of the researchers from the relevant
fields,-in the Laboratory. The charge of this working group is to develop
detailed plans for key technologies in the manned space activities by .
taking into account previous research accomplishments as well as the
current.technological'capability in the■Laboratory,, and to propose a new
research agenda,; This, report summarizes:..the Working Group's findings.:-
As shown in the text of the report, throughout the long years of its . ;/
research activities, this Laboratory has inadvertently been building up
a considerable amount of accomplishments and capabilities necessary for :'<
manned space technologies. The conclusion of the working Group is that
"The Manned Space; Activity Key Technology Research";is a-/very essentials
research theme of the Laboratory;..- In order- to develop the detailed
research and developmental program consistent with Japan's space
development scenario-.{for .example, the "Manned Spaces Platform Unique to.
Japanese Design'/and others:) „-.the facilities and 'equipment, plans, and
the budget scale, it is necessary to continue more discussion.
Although it is extremely important, the manned space transport system, -
which may be considered the infrastructure, of .the manned space
activities, has not been addressed in this report simply because its
technical content is too large.
53
2. Current Status of Japanese Technology and the Role of the National
Aerospace Laboratory
Japanese aerospace development has been conducted in the past for

unmanned technologies, mainly due to the constraints imposed from the
political and economical standpoints. This is the reason that the only
significant accomplishments for manned space activities have been the
selection and the training of the scientist crew to conduct the First
Material Testing Project (FMPT), and the development of part of the
environmental control system for JEM, which constitutes the Japanese
contribution to the International Space Station Project. There have been
practically no key technologies developed in Japan for the manned space
activities.
Expectations for the Japanese manned space activities have gradually,

yet steadily, been increased under this situation, the process of 'which
is described in the two successive reports published since 1987, such as
the report for Long Term Policy Discussions and for the Special
Committee for the Space Station, as well as in the General Guidelines
for Space Development Policy. Especially in the General Guideline for
Space Development Policy revised last year, it is stated that Japan must
develop key technologies if the country wants to maintain its
independence and freedom in research. At the same time Japan will be
responsible for one of the central tasks in the world space station
development activities which include international cooperation. The
Guidelines point out that one important element in the key technologies
is the manned space activity technology. Because of the current
technological status in Japan, the Guidelines refrain from mentioning
the development of specific and independent Japanese manned space
activities addressing only the formulation of key technologies needed
for long term development (See Appendix 2).
In order to implement this developmental project, it is necessary to

accumulate actual manned space activity experiences in the space
station, as well as to establish the key technologies for manned space
activities and to maintain strong participation in the advanced research
activities. As detailed in the next section, although it is fragmented,
this Laboratory has already been conducting several essential segments
of the manned space activity key technologies, such as the man-machine
system technology in the Fan-Jet STOL project, research on effluent
material processing systems, and for space sickness, under the Science
and Technology Promotion Adjustment Fund Project, expert systems for
support of space experiments in the Special Research Project under the
Space Environment Utilization Experimentation Technology project, and
others. In addition, research on the Closed Environment for Living
Systems in Space (CELSS) project under the new Science and Technology
Promotion Adjustment Fund project, and the research on Manned Flight
Technology under the Special Research Program for the Aerospace
Transport Element Technology Program are starting. It is very possible
that this Laboratory can exert leadership as well as play the central
54
role for the establishment of Japanese-based key technologies by
assembling and promoting, in a consolidated fashion, these past
accomplishments and future research capabilities mentioned above in the :
"Manned Space Activity Key Technology Research." In this manner, this
Laboratory will be able to meet its responsibility as the sole national
laboratory for aerospace technology in Japan.
3. Accomplishments for Related Technologies at the National Aerospace

Laboratory
Although there have been no large scale research projects directed

toward the manned space activities at the National Aerospace Laboratory,
several related technologies have been investigated as summarised below:
(1) Research Activities in the Special Project for Fan Jet STOL Plane
During the research and developmental efforts for the STOL prototype
plane, the Laboratory identified man-machine interface problems between
its pilot and the plane, and conducted research on the development and
evaluation of various display systems, as well as the control motion
analysis of the pilots.
(2) Research Activities for Special Research Project for Innovative

Aerospace Transport Element Technologies -
As a part of research on innovative aerospace transport element

technologies, the Laboratory is expected to conduct research on manned
flight technology to study the safety assurance problem (emergency exit)
and heat rejection system for the manned transport system. The research
on the intelligent cockpit has already been started.
(3) Programs under the Science and Technology Promotion Adjustment Fund
Project
Under this program, research on the treatment of human effluent, and the
integration of human sensation for rotational motions (space sickness)
have been conducted. This research has lead to a three-year program
starting in 1989 for the basic study of the material flow cycle between
plants and atmosphere in the closed environment (research on CELSS) in
the International Basic Study Project for Material Flow.
(4) Research for Special Project for Space Environment Utilization

Experiments
In order to make effective use of precious orbital resources in space,

research on recycling of water and gas, and the expert systems for
supporting space experiments have been conducted to assist the space
crew.
55
(5) Research for the First Generation Material Experiments (FMPT):
The research project to address the relationship between man's sensory

perception and his body motion in the weightless environment has been
conducted under the Standard Budget Program. The experimental research
is expected to start in June 1991 as a FMPT mission.
(6) Standard Budget Program
From the human factor point of view, research on the cockpit display
systems and the operation motion analysis were conducted as a part of
the R&D for the STÖL prototype. In the research for manned support
technology, CELSS and flight clothes, for the super high speed flying
have been studied. In addition, the Laboratory initiated research on
space robotics with a view toward developing the manned space activity
maintenance system. . <^
4. Manned Space Activity Key Technology System and The Area in Which the
National Aerospace Laboratory Should Be Working Immediately
Depicted below is the Manned Space Activity Key Technology System. The
transport technology system and the bus system technology, which are
shown under the dotted line in the figure, are not addressed in detail
by the Working Group. The Working Group, however, fully recognizes that
the transport technology is a very important key technology for manned
space activities, but its technological content is too wide and it is
closely related to other technology research.
56
Life Maintenance Environmental,Con trol (Gas &
w
System Heat '* V?i ^
•WaterS.Effluent Treatment*
.lo&d%ducUon*(3.(3))/
Processing
■Habitability
O
H
O
ü Extra-Vehicular •Space suit * (3. (6))

0)
H Activities •Interface w/Mother Ship
>i .Manned Space Moving
0)
TTirlf CMMU)
>h
^HSH0*'
■P
•H Man-Machine tobotic?
emote
emote ^vnerxiuein.. ,Lcv-""wlogy V
•5
4-1
Interface Science.
Human Factors * (3. (1)J
■3 •Man-Machine Interface * (i- (1). (2),
QJ .(5»
O
cfl Crew Selection
P- Pilots
en Criteria
£S and MS
Crew Training -■—1
a)
Technology .Life Boat* (3.(2)) (Implementation as a
.Abort* (3.(2)) ATAS Special Research
Emergency Rescue J
Technology Project)
Space Medicine & •Space Sickness* (3. (3))

Physiology
-Transport System Technology
" Bus System Technology

: it- 1th * ate the reseateh „hieh have aXready been atarted,
Note
() is the aetivities mentioned in section 3
ate technologies vbicb ate to be ptioritiaed in the futnta

Items xn
breakdown is given in the Appendix 4.)
(More detailed technology
57
5. Detailed Research Plans
This laboratory considers it necessary to make strong advances in

research areas as detailed in Section 5.1 through 5.3 by taking into
account of the Manned Space Activity Key Technology System and our
previous accomplishments. Section 5.4 indicates detailed implementation
plans for these research areas as an Integrated Key Technology Research
Program.
5.1 Life Maintenance System Technology Research
The life maintenance system technology can be grossly divided into the
room environment control technology, the treatment technology for waste
water and effluent, and the production and processing of food (the
details are given in the Appendix 4.1). Research on the component
technology for room environmental control and waste water treatment has
been carried out under the Special Research Program, while that for the
effluent treatment technology has been carried out under the Science and
Technology Promotion Adjustment Fund Project. It is considered
appropriate that the future research is to integrate the results of this
existing research under a Life Maintenance System. Moreover, the
laboratory has been currently conducting research on an algae culture
system and others for the food production and processing technology
research. This activity should be increased to include the effective
culturing system for higher level plants. Based on these considerations,
research on the following items should take the priority:
(1) The Gas Regenerative Cycle Technology (removal and regeneration of

CC>2 and C>2, and fixation and recovery of IU
(2) Detection and Removal Technology of Minute Quantity of Toxic Gases
(3) Treatment and Recycling of Waste water
(4) Collection and Recycling of Effluent
(5) Integration Technology for a System Including Food Production And

Oxygen-Carbon Dioxide Exchange
5.2 Extravehicular Activity Technology
As the manned space activity program develops further in line with the
Space Station Project and others, it is anticipated that the
requirements for extravehicular activities will also drastically
increase. The traditional Apollo/Shuttle space suit required an
adaptation process of a few hours before and after the extravehicular
activities because that the suit's inner pressure is as low as 1/3 of
the atmospheric pressure. It is obvious that this adjustment requirement
precludes its use for the extended extravehicular space activities. Also
the heat rejection design of the Apollo/Shuttle suit is a sublimation
58
type, or the non-recirculation type, which is not suitable for an
extended extravehicular stay. There are several more technical problems
requiring further research and improvement on space suits for extended
extravehicular stay in the areas of protection against radiation/cosmic
dusts, its usability, maintainability/repairability, maneuverability, and
others.
In the technology research at the Laboratory, it is necessary that these

technical problems must be prioritized while the fundamentals of the key
technologies involved in extravehicular activity are studied and
established. As shown in Section 4, extravehicular activities can
largely be classified into space suit technology, extravehicular
activity support technology (interface with the mother ship), and the
space movement technology (MMU technology) (the details are given in the
Appendix 4.2). Our Laboratory has been conducting research on space suit
technology for the past three years by a team of researchers, engineers
and medical doctors in- and outside of the Laboratory as part of the
"Super High Altitude Flying Suit Study Committee." We believe that
future research should be conducted based on the findings of this
committee's work as shown below.
(1) Study and Investigation of the Current Status of Space Activity

Environment and Its Effect on Humans and Its Relevant Technologies
(2) Basic Research on Life Maintenance System for Extravehicular

Activities (Oxygen Supply, Carbon Dioxide Removal, Toxic Gas Removal,
Pressure Controls, and Temperature/Humidity Controls)
(3) Study of the Space Suit Systems to Identify Technical Research Agenda
o Studies of Subsystems for Exterior Envelope/Mechanism System, Thermal

Control System, Gas Recirculation Control System, System Control System,
and Others
o Understanding of Input/Output Relationships Among Various Subsystems
o Identification of Necessary Basic Design Data
(4) Examination and Testing of Basic Data Acquisition Method/

Prototyping of the Parts
(5) Total System Simulation (Development into the Prototyping of the BBM
Model)
o System Simulation by Software Based On the Results of (3) and (4)
o System Simulation by Hardware Based on a Barrack Set Model
59
5.3 Man-Machine System Technology
The man-machine system can be largely divided into human factors,

robotics, remote experiment technology (tele-science), and man-machine
interface technology. The items listed below should be investigated on a
priority basis.
5.3.1 Human Factors
Technology items for human factors are divided into endurance, quality
maintenance, and motion characteristics of human beings in the space
environment. In the past, research has been conducted for the
deterioration of intellectual and motion capabilities under weightless
conditions (FMPT Project), and the research on integration of
rotational sensations (the Science and Technology Promotion Adjustment
Funds Project). It is considered that the items listed below are
priority research subjects:
(1) Implementation of Low Gravity Flight Tests (by Active Use of the
Experimental Plane Dornie Do228)
(2) Research on Hand Control Under the Low Gravity Field
(3) Research on the Coriolis Force Effect Under the Low Gravity Field
5.3.2 Robotics and Tele-Science
The robotics technology to support the manned space activities consists

of sensing and recognition technologies mainly by transducers, the
manipulation technology by robot mechanisms and others, the autonomous
operation technology based mainly on supervisory controls, and movement
technology such as the space motion (see the Appendix 4, and section
3.2). The tele-science related technology consists mainly of remote
experiment planning technology based mainly on experiment controls,
remote experiment operational technology based mainly of the scientific
crew support and tele-robotics, and data analysis capabilities (see the
Appendix 4, section 3.3).
For research related to robotics and tele-science, it is necessary that

a good interaction be maintained with the research activities of
domestic research organizations in the related areas. Among the research
topics mentioned above, the Group considers that the following tasks are
to be advanced as an extension of previous research under the Special
Research Fund Project and the Standard Budget Program at the National
Aerospace Laboratory.
(1) Feasibility Study for Robot Operations For Operations To Be

Automated And Robotized.
60
(2) Research on Robot Mechanisms For Delicate Operations, Flexible
Operations And Operation Dynamics Under yG Environment.
(3) Layered Type Autonomous And Remotely Controlled Operation Technology,,

by Supervisory Controls And Others
(4) Controls and Transmission of Scientific Knowledge, Such As

Experimental Knowledge (Expert System) ....
(5) Identification and Evaluation of Experimental Operation Work Load
(6) Predictive Control Technology by the Orbital Motion Prediction

Simulation and Others
(7) The Remote Operation Technology of Communication Delay Compensation

Type
5.3.3 Man-Machine Interface
The man-machine interface technology is positioned as the integrating

technology for human factors, robotics and tele-science technology. This
technology can be divided into the design of the work station on the
manned space platform and cockpits for space vehicles and others, and
the task distribution technology for manned space work.
In the past, research on the cockpit display systems, and vehicle drive
motion analysis have been conducted for the man-machine interface for
piloted aircraft. Based on this previous research, the following
priority tasks have been considered appropriate (see Appendix 4, section
3.4):
(1) AI type Operational System Research for Cockpit Design of the Space
Vehicles (Round Trip Planes, and Orbit Change Planes)
(2) Study on the Inter-Man and Machine Work Sharing based on Task
Analysis and Workload Analysis of Manned Space Work
5.4 Method for Implementing the Program as the Integrated Key Technology
Research Program ;
61
0) to
JT r-f
>-, CO U-J
oo o
O dJ 0]
4-> 4J
CO CO
0J H
CU iH I
*-» « -a g
=
B *H
O TO
•H
3 TH- 4J >,
Ä H U 0)
03 3 4J
*J 6 h a
» *H O >.
U^K CO W ffl u-i to
/ v
/ \ /
*i r
i i I I
1 O E CO
(U S
L 1 4J 4J TH
to CO p
a> ts to
JJ en
I I
rH
O 01 (0 111
•H -P 6
4-1 CO CO
CO U
3 OC 00 H
H.-Ä.
h>"-( <y B I I E
■H^> B ffl
-> J= m
3 4J £ 1
W TH 3 L_J LJ
i; j-> a L_J
J «-) u
w 0 0
>»* w»A.
3J
jSf@
n c U
J
o !* m
-H MS a T
^
^<»
n 4J o en C to \ XI «J M
y n) IH HI % 4J •• E U
J o <j ■o 6Hr4 1 « CU fi 0 CU CO
■H C 0) c « o » jr E O *H US i-t
-< Ä B O AJ M 4J i-i 4J a tu
a o C HCU 4J 4J U T3
p. 0}
<: u t > i O 0 o
U-H C J) O
 CU rJ CU ,C
cn > F
u « c CO
c c ^ c
WJ ). ■H Xi 4J rH B O
u) R CO 4-> CJ
71 u Pe 5 CU B ■H
1) i-I U t^ =
O H H O MH i-H 4J
>> 60
h B B
tO T-* o
*a
u
ÜJ2 OH J:
Ü 4J 0) 4-i
tu »J o -a
^»
U
rH
(0 0)
•H i-i
B
c a >H to O
SO M O IM o 4J O
cu ffl •H >,u
E 4J to ja
o «w a E c u
4J
rt H «o su
ß c -l-t O 3 T3 r E
4J a
(T rlWH (J
tu o to cu -
O
O 4-
P-
>,
U U > M i E 0) o
U-i a PU a cu j rH 0J ■H rH
c .c I- tO E
4J 0
R <w 3 B
S O
3 E
4J X 2 B
aj r< at u 03 y
4-1 0 4J
01
* u-< a
ai
E *H
> CU E-t
J3
o <a M*-I co -H • AJ
cu o B JJ *o co3>* m WO, r-i E.
c 4-1 »4 •H to atn cd 03 o
o
T-f 3
P-
4J x: e -H3 4J
*.
&
5i
4J r4 U *J CU H (fl 0) E O rH
(0 P aj E ■»3 4J flJCU 4= grH
rH O •H 0 Tl M CO ÜJ3Ö0 4-)
3 CJ U •H O 4J O SS-H4JC CO 4J ■H 4J
U m 4J M Ö tM E CO B O U-4 H 4= •W l-l
E co
4J o
M a) a ttt a ai CU Xi J2UW E O cu 4j h o o a)
43 cn N 3
•■S °
60 4J 3 UOCO 4 a cu a ja -n
CJ •H *o rH *
B co a at B QJ « B
a) r-l
n i-H cu Tt
rH o m fl ■H JS 4-1 Ö0M O E a-H
4J
E
CO
S
o o
•H XI 4J
o
l*-l
IH CJ
CU 4J
M
O
00 Ss S&rl
J3 OJ-HTM 4J CO
Ps CO
M i-(
•H CO
t-t « I
OC S 4J E3 CO ' (0 PL4
E
•H cn 4J
0) t>% .C 1-4 g 4J ß CO U trj E E
cj 3 a» to c -H n
fH to 0) c
X Ä
4J
01
CO 4J H-l 4JO «-<
•H M
Xi u •H 4J s 8
>.
O COT) Q>
CJ tf]
B fc s
cn o
c
4J
B
E
Q)
4-1
(!) ^
to o tnw -H
}-i4J
CJ CO B 0)t0
B
4-> 00
o'*H
U CO
co a
y to
00 N
CU
u r-i
O
^ >. Id 9 (0 CO XI OTJrH CU 03
3 a. 1 o
cu B •H
tu w •H > TJ 4J
3
cu >>
4J
ff)>%
tfl W CO CO 4JC3J-I M T3 0 B M
£ s
CU
XI & 4J 4J VJ
C/l w CO 4->
-B—
HICO
>.
cn
tOjS
=1 cu
E
•H
>
Ot
43 O
y rH
^
i J=
B o
3<c ä^s
62
Research Council for FY 1990
List of the Working Group for Manned Space Activity Key Technology
Research
(In the order of the Japanese alphabet)
Chairman
Hisao Higashi
Research Council Member
Committee Members:
Yoshiaki Ookami
Leader, the 8th Research Group of the Space Research Group
Koji Otsubo
Leader, the 9th Research Group of the Space Research Group
Mitsuo Oguchi
Member, the Space Research Group
Shiro Kibuse
Principal Investigator, Structural Dynamics Section
Takao Suzuki
Chief, Human Factor Laboratory, Control Section
Hiroyuki Terui
Specialist, Flight Section, Space Research Group
Atsushi Nakajima
Leader, the 12th Research Group, Space Research Group
Yoshihiro Nakamura
Leader, the First Research Group, Space Research Group
Keiji Nitta
Leader, the 4th Research Group, Space Research Group
Masataka Maita
Principal Investigator, Thermo-Fluid Dynamics Section
Kotafo Matsumoto
Principal Investigator, Numerical Analysis Section
Nobuyuki Hachiyagi
Member of the Research Council
(Dates of the Meetings)
First Meeting/Dec. 5, 1989 (Tuesday)

Second Meeting/Dec. 15, 1989 (Friday)
Third Meeting/Jan. 12, 1990 (Friday)
Fourth Meeting/Jan. 31, 1990 (Wednesday)
Fifth Meeting/Feb. 16, 1990 (Friday)
Sixth Meeting/March 8, 1990 (Thursday)
63
COMPUTERS
RISC Processor Array for Artificial Neural Networks,
90FE0178A Tsukuba HEIRETSU SHORI SHIMPOJIUMU JSPP '90 in Japanese May 1990
pp 385-392
[Article by Atsunobu Hiraiwa, Shigeru Kurose, Shigeru Arisawa and Kaoru

Inoue, Sony Corporate Research Laboratories]
[Text] ...... Abstracts
Artificial Neural Network (ANN) requires an excessive processing time because

of its large network size to solve difficult problems and because it needs a\-
large volume of learning patterns to be useful as a general purpose machine.
In order to solve this problem, the Net-Data Division Method, which permits
the efficient mapping of ANN algorithms in the mesh connected processor array,
has been developed. Reported herein is this Net-Data Division Method and its
simulation.
1. Introduction:
ANN has been used in many areas in recent years. A variety of models have
been developed as it has become evident that the ANN application, is very ..,,.,
effective in areas such as image processing, voice recognition» controls, etc.
In order to expand the effective application of ANN so as to make it more
practical, it is necessary that the number of neurons must be increased and
that the network size must be expanded. This, however, necessitates the
rapid increase of the,calculation volume, especially for the learning phase .
in which the data must be supplied iteratively to modify the network weights.
In normal workstations, the simulation of ANN usually requires the computation
time ranging from a few hours to a few days as long as the learning is
included. This has been the bottleneck in conducting large scale ANN
research and applications.
In order to improve this situation, it has been proposed that a special

purpose hardware be prepared to increase the computation speed by parallel
processing to take advantage of the parallelism intrinsic to ANN. One way
to do the parallel computation is to assign one neuron to each individual
calculation element, or to connect the processors assigned to multiple
neurons. The former method requires such a large hardware that it is
impractical to develop a large scale network at least with the present day
technology.
64
Most systems proposed at present are based on the.latter principle, and they
consist of those using the DSP processor, those using the general purpose
microprocessor, and those based on special purpose chips, etc. In any case,
all of these methods are based on multiple processors performing parallel
processing by communicating with each other. For the system using DSP, there
is Sandy (Fujitsu) based on TM320C30, which is a 32 bit floating decimal DSP.
This system has the processing speed of 500 M connections/sec (CPS) with
256 processors. [6][7]
We have been developing an ANN simulator called the GCN (Giga CoNnection)
adopting the systoric array in its archtecture with the latest 64 bit
microprocessor, 80860(40MHz) developed by Intel. Currently we are developing
a system with 16 parallel processors.
We have also evaluated the mesh connected 128 processors in which the parallel
processing is done by mapping the Back-Propagation model and the Kokonen
Feature Maps by the net data division method, and have verified that this
system can achieve the processing speed of 1.39 Giga CPS/1.88 Giga CPS.
In addition, a plan to develop a super-chip that integrates this architecture

will be discussed in this paper.
2. Basic Array for GCN
The Neutral Network simulation is comprised mainly of the add/multiple

operations to determine the sum of the weighted values of inputs to the
neurons. Since the computation for each neuron is done independently of each
other, the simulation speed can be increased.if the processors to perform
the high speed add/multiply operations are interconnected and work in
parallel.
This type of system, however, has several shortcomings such as: it requires
many large hardware for intercommunication among different processors; a
processor overhead is required for processing communication; and its controls
become very complicated. These are the reasons that we employed a systoric
array as our GCN architecture in which the same processing elements (PE)
are systematically arranged so that the communication will be limited:to
those between the adjacent PE's. Hence the communication.hardware can be
made small and the control becomes simpler. Moreover,, the communication
can be done asynchronously, and the processor overhead can be controlled to a
small amount. The structure connecting the same PE makes its expansion easy,
and permits, the improvement of its overall performance by simply increasing -:
the number of PE's.
Figure 1 depicts the schematics of the GCN architecture. Each PE in this '
figure has a 80860 microprocessor, a 4M byte localmemory, and two FIFO s
(64bits x 256W). The local memory is used to store the connecting weights
among the neurons as well as the neuron outputs.
65
Each PE in this figure is connected to the adjacent four PE's through FIFO's
to perform the asynchronous communication. In this manner, the data sending
PE's can send the data and begin the next step operation without waiting for
the receiving PE's to assume a receivable state. The data transfer speed of
the FIFO is 160 Mbytes/sec with the 40 MHz clock.
The host EWS (NWS-1850/3800) and each PE are connected through the VME BUS.
JL
PE
±.
PE
80860
80860 -> FIFO -> $» FIFO ->
4Mbyte RAM 4Mbyte RAM
FIFO FIFO
3E PE
PE
80860 FIFO 80860 FIFO
4Mbyte RAM 4Mbyte RAM
FIFO FIFO
~¥ V
Figure 1. The Basic Architecture of GCN
Table 1. Properties of Intel 80860
Type 64 bit RISC Processor
Arithmetic Unit Integer/Control Core Unit

Floating Decimal multiplier
adder
graphic unit
Pipeline 3 Stages
Bus External 64 bits
Internal 128 bits
Cash Instruction 4 Kbytes
Data 8 Kbytes
Clock 33.3/40 MHz
Peak Performance Single Precision Floating Decimal
80 MFLOPS
Double Precision Floating Decimal
60 MFLOPS
Process CMOS 1.0 u rule
66
The 80860 is the 64 bit RISC type general purpose microprocessor developed
by Intel, and careful attention has been given to achieve the high speed
add-multiply operations. It has three independent units such as the core
unit for the integer arithmetics and controls, the multiplier to do the
floating point arithmetics, and the address unit. These three units can be
operable in parallel. The multiplier and the adder have the three stage
pipeline structure so that the continuous multiply/add can be done to obtain
the results for clock by clock. The data loading/storing by the core unit,
and the floating decimal add/multiply arithmetics can be done in parallel to
communicate with the adjacent PE's without its process being interrupted.
[Table 1]
The arithmetic operations in the GCN are performed in two levels by the
pipelines/parallel processing. In the first level, it is done by the
Fine-Grain Pipeline/parallel processing. The floating point arithmetic and
floating point LOAD in the PE are pipelined by utilizing the pipeline
function of the 80860. All arithmetic units can be parallel-processed.
The second level is the micro-pipeline prpcessing.
The arithmetic operations of the entire system is made into a pipeline by the
use of systoric algorithms. The arithmetic operations and data required in
the ANN simulation are divided and separately assigned to PE's, and each
PE transfers its processed result to the adjoining PE's. The process
proceeds in this manner.
This architecture makes it possible to extract the processor array capability

by the effective use of pipeline/parallel processing within the PE as well
as between the PE's.
3. Mapping of BP Algorithms Over GCN
For the parallel processing of more than 100 PE's, a mesh type connection is
used since the linear ring connection tends to increase the communication
time.
This is the reason that we developed the net division method in which the
BP algorithms can be mapped over the mesh connected GCN.
As shown in Figure 3, each PE in the horizontal ring is used in the data

division method [2], and it has the same weight so that the time required
for inputing the different data as well as the communication time can be
shortened. (We call this the copy method.) Every time the entire data are
seen, the derivatives of the connecting weights and revised connecting
weights will go around along this ring.
67
HOST
NWS-18xO/38xO
VME BUS
A(1)
BL
PE
QL
PE
o. PE
G/
PE
H/
PE
ISONY.GCN > SÖNY.GCN >ISONY.GCN >ISÖNY.GCN > SÖNY.GCN
t>
•«3 JL J&. JL _^_ V
I
PE PE PE PE PE
5 c>
Of
ISONY.GCN >ISONY.GCN >ISONY.GCN -> SONY.GCN >ISONY.GCN
^
XL \J/_ jk X- V
a
e PE PE PE PE PE
ISONY.GCN >ISONY.GCN ->SONY.GCN > SONY.GCN > SONY.GCN
^ Ut V V V
^
PE PE PE PE PE
> SÖNY.GCN >ISONY.GCN > SÖNY.GCN >
V £ ISONY.GCN >ISONY.GCN
5
LJ IT u a a
(2) f-âa©ÄftO*f!lyr _ ^
Figure 3. Net Data Division Method Over Mesh Connected GCN
Key:
1. Vertical ring for the network division
2. Horizontal ring for the data division
Moreover, the number of net divisions, (N) and the number of data divisions,
(D) as well as the total number of data (AD) will be explained for the
three layered net data division BP algorithms. Table 2 depicts the
memory volume of each PE.
(1) When the input data 1 is given, the input to the hidden unit j is
determined by
netj =S? wji x
l
i
where W..: connecting weight
68
(2) The output of the hidden unit j will be determined by
Hj 1
1 + e -netj+6j
Where 0. = threshold value

3 ■ .
(3) The Input to the output unit can be obtained by transferring the
intermediate result similar to (1) onto the vertical ring.
(4) The output of the output unit can be determined in the same way as (2)
(5) The output layer error can then be determined by
■-*« = (Tr-WU-fy
where T; : Trainer Signal
0. : Output of the Output unit
(6) The error of the hidden layer can be determined by the following equation
as the errors are transferred onto the vertical ring
*« = H5(l-Hj)X;tf„kWy
k=l
where q: Number of the output unit
(7) The change of the connecting weight for the mth input data can be
determined for each layer by
AWg(m) = Ôj + AWij(m-l)
where (m = 1, - -, AD/D)
Input -hidden layer ... output of the input unit
J
~ Hidden -output layer ... output of the hidden layer
(8) Steps (1) through (7) will be repeated for all divided data for (AD/D)
times.
(9) The total summation will be obtained for the changes of the divided ,
connecting weights by using the horizontal ring.
(10) The connecting weights will then be modified by the following equation
69
D-l
AW..(t) - T, £ «Wa(AD/D)+aAWu(t-l)
pe=0
W„(t) = AW8(t)+Wu(t-l)
where Y\ : learning constant

Cl : stabilization constant
t : number of learnings
(11) With the use of horizontal ring, the revised connecting weights will be
transferred to the other data divided PE's.
(12) Steps (1) through (11) will be repeated until the data reach the target
values.
If the memory volume is to be reduced, the input data will be divided into
N groups and dispersed over the vertical ring PE's. In this case, step (1)
will be determined by transferring the intermediate results of the similar
add/multiply operations to the vertical ring, and step (7) will be
transferring the hidden layer errors on to the other PE's on the vertical
ring.
Table 2. The Memory of Each PE in the Net Division BP Algorithms

(Number of Net Division (N), Data Division (D))
Area Memory Volume
Errors, Trainers, Hiddeness, Number of Output

Units Connecting Weights, Changes of Connecting
weights 1/N times
Number of Input Data 1/D times
Change of Connecting Weight in the Preceding Step
(N divisions) 1/(ND) times
4. Mapping of Kohonen Feature Maps (KFM) onto GCN
The KFM is the two layer learning algorithms without trainers [9]. Hence,
it requires an excessive amount of computations to perform the large amount
of iteration for self-organizing the network with respect to the input data.
In a similar manner to BP, each PE in the vertical ring will be used by the
network division method and will possess the different and divided weight.
70
The intermediate results, the maximum values, and the revised neurons go
around along this ring to perform the inner product, the maximum value
search, and the search of the revised neurons.
Each PE on the horizontal ring is used by the data division method, and has
the same weight so that the', time required for handling different weights and
communication can be reduced. Every time the total data is seen, the
derivatives of the connection weights and revised weights will go around
along this ring.[10]
Next we would like to describe the net data division KFM algorithm by
denoting the number of net divisions by (N), the number of data divisions
by (D), and the number of the total data as (AD).
(1) As depicted in Figure 4, the input vector 1 and the weight Wij will be
divided into N groups. The inner products of these quantities will be
obtained by transposing the intermediate results on the vertical ring.
(2) The local maximum of the inner product is determined for each PE, and
the global maximum of the inner product will be obtained by going around
the vertical ring once.
(3) The global maximum will be informed to each PE for every circulation of
the vertical ring.
(4) Whether the neurons assigned to each PE need to be modified or not will
be determined, and this information will be given to each PE by going around
the vertical ring once.
(5) The change of the connecting weight for each PE will be determined by
4Wij(m) = lj - Wij + AWij(m-l)

where (m = 1, —, AD/D)
(6) Steps (1) through (5) will be repeated for all divided data for (AD/D)
times.
(7) The changes for the connecting weights divided for each data will then
be summed up with the use of the horizontal ring.
(8) The connecting weights are then revised by the following formula:
71
D-l
AWg(t) = 7/5] AW„(AD/D)
pe=0
Wy(t) = AWyCtJ+W^t-l)
where f) : Learning Constant

t : Number of learnings
(9) The revised connecting weights will then be transferred to other data
divided PE's.
(10) Steps (1) through (9) will be repeated for K times.
0= W•I
PEÖ T3 TO | Tl | T2 PEO
PE1 T2 T3 | TO | Tl PE1
a
PE2 Tl T2 | T3 | TO PE2
PE3 TO Tl | T2 | T3 PE3
PEO PE1 PE2 i PE3
Figure 4. Network Division (4 divisions)

(The connecting weights to be used at time Tx)
(1) AJ]-<? K'*I
^>oooooo
-oöoeOQO
KXXXXKX)
Figure 5. Kokonen Feature Maps
wn: winning neuron
Nc: neighborhood
Key:
1. Input vector 1
72
5. The Simulation Results:
The capability of our GCN can be evaluated by performing the algorithms on

the Sim860, which is the software simulator for the 80860.
The three layer network evaluated by this simulation consists of 1600 input
layers, 64 hidden layer and 32 output layers. The total input data is of
5100 patterns, and the PE's are performed by the 32 bit floating decimal
arithmetics. The network is divided into four groups, the input data are
dispersed into 32 PE groups mapped on the 128 mesh connected PE's.
The net-divided BP algorithms developed under these conditions are converted

into a program consisting of 900 line assembler language statements and 600
line C-language statements. The simulation based on this program resulted
in the 4.6 machine cycles per connection for each PE.
This result includes the forward and backward calculations, modification of

the connecting weights and inter-ring communication time.
If each PE was operated on 50MHz (20 ns), the performance of 1.39 Giga CPS
(Connection Per Second) could be obtained.
For the KFM, we mapped the 1600 input layers, 64 output layers, and total
input of 5120 patterns in a similar manner for the simulation. The simulated
result showed that each PE required 3.4 machine cycle per connection. If it
was operated on the 50 MHz, it would have been possible to obtain the
performance of 1.88 Giga CPS.
6. Concluding Remarks:
In the large scale neural network simulation, we have verified with the use
of software simulator that the mesh connected RISC processor array based on
the net data division method can attain the high performance.
In the future, we plan to evaluate the mapping of neural network algorithms

onto the GCN. This architecture can be integrated, with the use of
multichip module technology, into developing superchips similar to the
80860 having CPU core, two FIFO and 4 Mbyte memory. If this is done, it is
possible to realize a parallel neuro-accelerator having the performance of
1 Giga CPS on a (printed circuit) board. It is also expected by the year
2000 the realization of a Tera CPS which would allow the parallel processing
of a few thousands to a few ten thousand units.[11],[12],[13]
73
FIF0.32bit FIF0.32bit SRAM.4Mbit
|»MI|mlfäü-|
^P
(•uM||aUMItWUM|
HH
l*_ll—«—.1
SRAM_4Mbit SRAM.4Mbit
80860
SONY.GCN
SRAM_4Mbit SRAM_4Mbit
SRAM.4Mbit SRAM.4Mbit SRAM.4Mbit
Super-chip for GCN
Acknowledgement:
We would like to express our appreciations to Directors Mlyaoka and

Matsuda, and Section chief Iga who gave us the opportunity to present this
paper during this conference.
Table 3. Comparison Between GCN and Warp
Sys terns Warp GCN
Pipeline 7 stages 3 stages

Local Memory 16 kbytes 4 Mbytes
Activation Learning Data Activation Learning
Data Connecting
Weights
Communication 80Mbytes/sec 160Mbytes/sec

Between PE's
PE's Computation Speed 10MFL0PS 80MFL0PS
Number of PE's 10 128
PE Connection Ring Type Mesh type
Performance 17MCPS 1.1GCPS
References
(1) D.E. Rumehart, G.E. Hinton and R.J. Williams: Learning Internal
Representations by Error Propagation in Parallel Distributed Processing,
Vol. 1, pp 318-363, 1986.
74
(2) D.A. Pomerlean, G.L. Gusciora, D.S. Touretoky and H.T. King: Neural
Network Simulation at Warp Speed: How we get 17 Million Connections Per
Second. ICNN88, Vol 2, pp 143-150, 1988.
(3) S.Y. King, J.H. Hwang: Parallel Architectures for Artificial Neural
Nets, TCNN88, Vol 2, pp 165-172, 1988.
(4) SiY. Kung: VLSI Array Processors., Prentice Hall Inc., 1988.
(5) A. Twata, Y. Yoshida, S. Matsuda, Y. Sato and N. Nakamura: An Artificial

Neural Network Accelerator Using General Purpose 24 bit Floating Point
Digital Signal Processors., IJCNN89, Vol 2, pp 171-175, 1989.
(6) H. Kato, H. Yoshizawa, H. Iciki and K. Asakawa: A Parallel

Neurocomputer Architecture Topward Billion Connection Updata Per Second,
IJCNN90-Wash. DC, Vol 2, pp 47-50, Jan 1990.
(7) A. Hiraiwa, S. Kurosu, S. Arisawa, and M, Inoue: A Two Level Pipelined

RISC Processor Array for ANN, IJCNN90-Wash. DC, Vol 2, pp 137-140, 1990.
(8) N. Fukuda, Y. Fujimoto and T. Akabane: A Transputer Implementation of

Toroidal Lattice Architecture for Parallel Neurocomputing, IJCNN90-Wash. DC,
Vol 2, pp 43-46, 1990.
(9) T. Kohonen: Self-Organization and Associative Memory. (Second Edition),

Springer-Verlag, 1988.
(10) R. Mann and S. Haykin: A Parallel Implementation of Kohonen Feature

Maps on the Warp Systoric Computer, IJCNN90, Wash. DC, Vol 2, pp 47-50,
Jan. 1990.
(11) R. Johnson: Multichip Modules: Next Generation Package, IEEE Spectrum,

Vol 27, No 3, 1990.
(12) "Special Edition; Multichip Module" Nikkei Microdevice, No 54, 1989.
(13) P. Geslinger, P. Gargini, G. Parker and A. Yu: Microprocessors Circa

2000, IEEE Spectrum, Vol 26, No 10, 1989.
(14) Mitsuo Kawato: Neural Network, Spectrum, Vol 13, No 4, 1990.
(15) Intel Corp., i860 ™ 64 bit Microprocessor Programmer's Reference

Manual, 1989.
75
MARINE TECHNOLOGY
Information Processing and Equipment Arrangement in the Pressure Hull öf

'Shinkai 6500'
916C0003A Tokyo KAIYO KAGAKU GIJUTSU SENTA SHIKEN KENKYU H0K0KU in Japanese
Mar 90 pp 303-314
[Article by Shuichiro Hamaguchi, Itsuro Maeda, and Kazuhiko Baba, Deep Sea
Technology Department, Mitsubishi Heavy Industries, Ltd., Kobe Shipyard,
Submarine Department]
[Text] [Abstract]
The 6,500 m deep research submersible Shinkai 6500 is operated by a crew of
three seated in the pressure hull which is a 2 m diameter sphere of titanium
alloy.
Much equipment for maneuvering, navigation, communication and observation,
etc. is installed in the pressure hull.
In order to operate the submersible safely and smoothly, it is very impor-

tant to arrange the equipment effectively in the narrow space so that the
crew can easily operate and monitor it, and also offer necessary information
instantaneously in an easily understandable form.
This report explains the outline of the design procedure and the results of
the actual equipment layout in the pressure hull. The Integrated Informa-
tion Display System (IIDS) uses a CRT to display various information.
Information processing and display modes etc. of the IIDS, and data reproduc-
tion after recovery on the support vessel "Yokosuka" are also discussed.
Key words: Shinkai 6500, Pressure Hull, Arrangement, Information Processing,

Integrated Information Display System, CRT Display.
1. Introduction
The 6,500-meter-deep research submersible Shinkai 6500 was completed at the

end of November 1969, and delivered to the Japan Marine Science and Tech-
nology Center. The submersible, operated by a three-man crew, has the
following features: 1) To ensure that its navigation—steering and survey
and observation activity—will be conducted safely and efficiently,
76
ingenuity from a human engineering viewpoint is employed in the arrangement
of the observation portholes and equipment inside the pressure hull.
2) Various information and data necessary for the submersible's navigation
are centrally processed and displayed on an integrated information display
system using a CRT on board, and are simultaneously recorded in a magnetic
bubble cassette. 3) The recorded information and data may be reproduced
in the integrated information display system's data reproduction division
aboard the support ship Yokosuka, and processed into forms which can be
easily utilized.
This report describes the arrangement of the observation portholes and the
equipment inside the pressure hull, and gives outlines of the integrated
information display system and the information processing and display
methods being adopted in the equipment.
2. The Layout of the Observation Portholes and the Equipment Inside the
Pressure Hull
2.1 Positions of Crewmen Inside the Pressure Hull and Observation Portholes
Inside the pressure hull the three crewmen are positioned with the pilot
seated in the center facing front, the co-pilot seated on his right and the
observer on his left. The pilot can see the outside through the observation
porthole at all times while steering the submersible. This layout was
adopted based on experience obtained from actually operating Shinkai 2000. ,
In the Shinkai 2000, the pilot is designed to steer the submersible while
observing the instruments on the console and a monitor TV, and he is to look
into the observation porthole to see the outside world only when needed. It
has been discovered, however, that in actual navigation the pilot finds it
much easier to steer his vessel if he can see the outside world at all times
through the porthole. As a result, in Shinkai 6500 the console system was
abandoned, and in its place a new layout was adopted'.' The pilot is seated
at"the center to the submersible's front and an observation porthole for
monitoring the submersible's direction of advance was also installed at its
front. In order to improve the side views, the observation portholes for
the co-pilot and observer on Shinkai 6500 are installed farther to the out-
side direction '(sideways direction)' than their counterparts on Shinkäi 2000,
affording a vision of field close on the beam (plus or minus 90 degrees).
Figure 1 gives a comparison of the inside of the pressure hull of Shinkai

2000 and that of Shinkai 6500, Figure 2 is a comparison of the locations of
the view'ports for the two submersibles, and Figure 3 shows a comparison of
the fields of view on the sea floor.
In Figure 3, Shinkai 2000 appears to have a better field of view than

Shinkai 6500 in the area directly to the front of the pressure hull, but in
Shinkai 6500 the shortfall is compensated for by upgrading the capabilities
of the TV cameras (two cameras are provided, and one of them can make a 360-
degree turn or can have its camera angle raised or lowered).
77
Figure 1. Inside of the Pressure Hull of Shinkai 2000 (left) and Shinkai
6500 (right)
Shinkai 2000 Shinkai 6500
(Elevation)
« 120X2
# 80X1
Figure 2. Comparison of the View Port Location
78
Inside diameter o
pressure hull
Figure 3. Comparison of the Field of View on the Sea Floor
Key:
1. Field of view to the left of both eyes (43 degrees)
2. Field of view to the left of both eyes when the angle of depression
is 25 degrees (43 degrees)
3. Center of pressure hull
4. Field of view to the right of both eyes when the angle of depression
5. Field of view to the right of both eyes
6. Field of view to the front of both eyes when the angle of depression
2.2 The Layout of Equipment Inside the Pressure Hull
As explained earlier, in Shinkai 6500 the console system has been abandoned
and various instruments are arrayed as if they are pasted onto the wall (the
boxes containing instruments are embedded in a skeleton called a "bird cage"),
Photograph 1 gives the layout of various instruments inside the pressure hull.
As described before the decision as to where to install what equipment has

been made with due consideration to the positions of the crew so that each
will be able to operate or monitor the instruments with ease. For example,
the integrated information display system and navigation instruments such as
an advanced sonar, and an acoustic position finder are positioned closer to
the direct front of the pilot and co-pilot, while systems that need to be
handled while steering the submersible, such as the weight trim operation
panel, are placed immediately to the front of the pilot. The control box
79
for the main propulsion system, the manipulator master arm and the joy stick
for the grabber, among others, are built portable so that the pilot will be
able to shift his position inside the pressure hull while observing the
outside world through the view port.
Photo 1. Inside of the Pressure Hull
The manipulators for the underwater still camera and the payload instruments
which are used mainly by the observer, however, are concentrated on the left
side. The monitor for an underwater TV is installed near the view port for
the observer on the left side of the pressure hull. He, himself can
manipulate it into making a rotation or raising its camera angle upward or
downward (using a portable controller) so that he will be able to have a
broad field of view.
Equipment and instruments which are not used ordinarily, such as the
emergency breathing equipment and the distribution panel for communications,
are concentrated in the aft section of the pressure hull.
Figure 4 gives the layout of major equipment inside the pressure hull.
In order to determine the shapes and dimensions of the equipment and instru-
ments, and best utilize the small space inside the pressure hull, each
instrument was actually placed in position in a mockup of the pressure hull
manufactured in advance.
3. The Integrated Information Display System
Due to the advance in electronics technology, the technique of CRT display

of information, centered on data loggers for surface ships, was employed as
early at 1975. The adoption of the CRT technology for the information
80
CO
H
O 3
 rH
Q) M-l 3 cd .... ■ A3
T3 ■•■; O rl ^-s Cfl
3 rl CJ tU CO 4J
- 4-1 CU ;
P, rl' rl
rH •H 4J -•: "ö o tu ö o
I-l • 4-1 CO 7 e 4J o p.
3 ■ Cd cd -rl -rl ^-'
m Cd U
■ tu
UHU
<U tU CJ rl
<u ■ bOt) fö s <r cu tu
ri ■- 3 (3 3 TO i-l CO A3
?
CO .■■-:.H
•H 3
■ •■
1
4J
•H
CJ N-' A3
3 cd
CO 0 rl CD 4-1 rH rl O rl
CU 4-1 O -Ö rH H CU -rl M
u •H M-l •rl Cd -rl C 4J
P4 s, '■ CO 4-1 -rl Cd rl
: O H •> CO CO rl O
■P <0 . e tu 4-1 A3 4-> CU M-l
3 A3 3 U 4-1 rl 3 P ,
o 4-1 rl CO O P. O CJ
O CT3 CO S 4-1 (3 MH . 13 cd T) -H
" O -H -H CU ■ •rl A3 (US 3 cd 4J
•rf tu rl CJ ,0 ".« H A3 4-1 Cd 4-1 4-1 tU (U 00 H
Cd -H CO (3 3.-H ^ 13 l>>>.^
OT3 CJ tu os 3 '3 3 X cd o
CO
4-1
O
+J
e CO CO O H 3> O PM r,
p.
3 •rl
O oo 3 -• < • • • • • •
J-i 3 O" OvO H CM en <!■ m
•rl W
CJ
■..■ H r?l rH rH f—1 r—i
cd
ctt
P. rl
•H CU
CJ
3
•H
1 :
•H :■
U MH ./, T3
PH 'CO (U
■■■■ rl ■ ,4-1
tu"' o cd
c
•■_'
.' 0) - cd .
A3 •H rl rl, 4-1 rl
4-1 3 : 4J cd cu tu
CU 00
O ß ST) Q)
M-l •rl •••■■ co cd O -H 4-1
TJ
o
4-1
4-1
CJ
0)
o o co Q) X
p. : rl O
.'. > 3 3 -H
:s
•rl
to S CO
rH
CJ
•H
H O
■H
CD "3
CO
rl
o:
3<uco I rH
O
(U
3
td 3
4-1
CO
rl
CU
4-1
Cd
CU 3
rl O
M-l
OTSri
,Q-H M CJ 3 -. tu A3 O ■IS
T3 3 MH
< CU rl 13 CU CO
3 O 3
1
CU rl a CO
tu CU
bO 60
MH 3 • '-■■■■ t-
CJ :TJ Ct)
3 to 4J rl H
■ «H ■ Ö 4A tU O 'P
0) p. cd 3 CJ to
M . S CU -rl
M 4-1 rl
CU
rl
•H
13
3 Ö0
Ü0 3 '>> rl 3 tu
•rl •H rl 3 CU 4J 'r3 3
N CU P. CJ 4-1 tU 3 O
•H +J so to co o e cd •H
3 4J 3 -H "■-•■■ « P O 4J
o cd •H T) >. a •H rH -Cd
ed oo 3 4J o a
rl A3
•StU CN H cd tu ß MB
p 4J Ö0 a u.o
,3 CJ CO rH >~. 4J 3:
cu •H O X
erf P > o
O O
P O •S
H N CO<* ih lON 00
tu
81
system for Shinkai 6500 was determined from the initial stage of the sub-
mersible' s construction program, and the effort was realized in the form of
the integrated information display system on board. Data from the various
systems and instruments carried by Shinkai 6500 (both those located inside
and outside the pressure hull) are input in the integrated information
display system and comprehensively exhibited on a CRT display. The system
simultaneously records the necessary data in a magnetic bubble cassette.
This enables the crew to see the submersible itself and its surroundings at
a glance (unlike the Shinkai 2000). It eliminates the need for the crew
to check the gauges on the consoles and equipment, thereby contributing
greatly to the ease of steering.
The gauges/displays for the equipment are built in duplicate, which raises
their reliability. An outline of the integrated information display system
is described below.
3.1 Configuration of Hardware and Performance
The integrated information display system has the following hardware con-
figuration and performance.
—Control and record section
CPU (16 bits, 8 MHz)

Two bubble cassette slots (512 kilobytes x 2) (recording time of 10 hours)
Eight serial input channels
Two serial output channels (each having a 10 mA current loop)
Parallel input x 32 bits
24 analog input channels (voltage—±10 V, or current—4 to 20 mA)
—Input/output section
AC 100 V, 60 Hz, 1 $
(Power source) A built-in speaker for voice alarms
—Display section
A 9-inch color CRT display with a touch sensor panel
3.2 Input Signal
The majority of signals input (automatic input) into the integrated informa-
tion display system come from the submersible's various equipment, but some
of the signals are manually input by the operator via the touch sensor
panel.
Table 1 gives a list of input signals.
82
Table 1. List of Input Signals
Input mode
Signal name
Automatic Manual
Electric power system

Bus voltage 0
No. 1 main battery current 0
No. 2 main battery current 0
Amount of discharge from No., 1 main battery 0
Amount of discharge from No. 2 main battery 0
Insulation resistance bus 0
Insulation resistance AC 100 V 0
Insulation resistance DC 28 V 1 .'■ 0
Communications system AC 100 V voltage 0
Communications system AC 100 V current 0
Communications system DC 28 V voltage 0
Communications system DC 28 V current 0
Propulsion and steering system
Main propeller rotation number 0
Main propeller angle of oscillation 0
Rotation number and direction of vertical thruster 0
Rotation number and direction of horizontal thruster 0
Hydraulic system
Oil pressure
Weight trim and adjustment system
Quantity of liquid in trim tank .0
Quantity of liquid in auxiliary trim tank 0
Internal pressure of auxiliary trim tank .: >: 0 ,■..-....'
Retention/release of ballasts for diving 0
Retention/release of ballasts for surfacing . 0
Pressure of air storage device(air pipe) 0
Navigation and communications system
Depth ;.■ .0-. ..:.
Altitude 0
Direction of bow 0
Vertical angle of inclination 0
Horizontal angle of inclination 0.= :
Diving/surfacing speed (descent/ascent speed) 0
Position of submersible 0
Second, minute, hour, day, month, year 6
Air pressure inside ship 0
Temperature inside ship 0
Humidity inside ship 0
Flow direction 0
Flow velocity 0
Environmental control system
No. 1 oxygen partial pressure 0
No. 2 oxygen partial pressure 0
Carbon dioxide partial pressure 0
Oxygen flow 0
Pressure in oxygen container 0
[Continued on following page]
83
Input mode
Signal name Automatic Manual
Survey and observation system

Salt content (electrical conductivity) 0
Water temperature 0
Depth (pressure in the depth)
Sound speed
Emergency and safety
Various alarms 0
3.3 Software
The software for the integrated information display system consists of four
kinds of operation programs which correspond to the operation flow of
Shinkai 6500 and a kind of test program, for a total of five kinds of pro-
grams. Each is recorded in aii individual bubble cassette.
Figure 5 gives the operation flow and corresponding software, while Figure 6
gives the configuration of software.
(Operation flow) (Corresponding software)

Operation check program
Diving preparation
Crew check o Diving preparation check
Check before touching Ship operation program

Hatch closed
water d .
o Check before touching water
Check after touching o Check after touching water
water
Diving o Maneuvering
x)
Diving 0)
CO o Preparation of trim
Descent o
Trim adjustment CJ o Data display
Landing sea floor CJ o Recording of inboard
4-1
CO environment
»
Survey and observation
Floating
Leaving sea floor
Ascent
Check after surfacing Hatch opened o Check after surfacing
After being hoisted above mother ship o Check after being hoisted
Check
Maintenance check program
Maintenance o Maintenance check
Maintenance check
Figure 5. Operation Flow and Corresponding Software
84
Integrated information display system
Ship operation program —Check mode before touching water

—Check mode after touching water
—Maneuvering mode
—Trim preparation mode
—Data display mode
—Recording mode of inboard environment
—Check mode after floating
Operation check program T—-Diving preparation check mode

'—Check mode after being hoisted aboard
mother ship
Maintenance check program
Teat program
Sea trial program -Check mode before touching water

—Check mode after touching water
—Maneuvering mode
—Trim preparation mode
—4)ata display mode
—-Recording mode of inboard environment
—Check mode after being hoisted aboard
mother ship
Figure 6. Configuration of Software
3.4 Information Display Method
As a represenative example of the information display method employed in the

integrated information display system, brief explanations are given on the
maneuvering mode and the data display mode of the operation program when
alarms have been sounded.
85
3.4.1 Maneuvering Modes
The maneuvering mode is displayed at almost all times while Shinkai 6500 is
submerged (the mode selection is effected by selecting the appropriate mode
in the "Menu" at the start of the operation program via the touch sensor
panel). Data are displayed in graphics so that the operator will be able to
see the submersible and its surroundings at a glance.
Photograph 2 gives the display pattern of maneuvering mode, and Figure 7 and
Table 2 give explanations of the display pattern.
Depth
Altitude
Trim
Heel
Trim tank
Auxiliary
tank
Photo 2. Display Pattern of Maneuvering Mode
3.4.2 Data Display Mode
The data display mode displays data on the power source, oil pressure, etc,
as bar graphs, and consists of a three-page screen. For scrolling to the
next page, you need to touch the area marked "NEXT" on the screen.
Each page contains the data listed below.
—1/3 page—source voltage (bus bar, ac system, dc system)

insulation resistance (same as above)
—3/2 page—amount of discharge from main battery (Nos. 1 and 2)

current (main battery, ac system, dc system)
—3/3 page—hydraulic system pressure

air storage tank pressure (air pressure)
amount of liquid in the auxiliary tank and pressure
amount of liquid in trim tank
86
# 1 N N N <-(D @<
\/*
#2 NNN — © * ( Â
® ß
Depth NNNNHD \ \ ■ I NNN | M i-
' ■■ RPM j A'NNN^

Altitude NNNN-© pc?^ 1
descent for ascent Flow
Trim XXNN- I1IL.NNN jflfljîrection
Heel XNN- > <'"•- ~~~ ^ , Flow
'(gf r^ ^ 'X-'~7T& city
yelocit-
Salt
Trim tank XNNN^® content
Auxili; ry tank NNN v Water temp
80 40 40 80
YYYY MM- DD h h : m m" s s maneuvering LOG

/
m N
©
Figure 7. Display Pattern of Maneuvering Mode
Photograph 3 gives the display pattern of the data display mode (of 1/3
page), while Figure 8 and Table 3 give explanations of the display pattern.
3.4.3 Displays When an Alarm Is Sounded
When an alarm has been sounded, the ongoing display is interrupted and a
voice alarm is heard while the contents of the alarm are flashed on the
upper left side of the screen (see Photo 2). Simultaneously, the time when
the alarm was sounded and its contents are recorded in the bubble cassette.
Alarm signals themselves are input from the alarm display panel which has
built-in display lamps and an alarm buzzer.
Since the flashing alarm display area and the information display area are
two different areas, even when an alarm is being flashed, other data are
displayed on the screen.
Table 4 gives a list of alarms.
3.5 Recording and Reproduction of Data
When the operation program is activated, time and date, the position and
height and the direction of the bow of the submersible, the direction and
velocity of (current) flow, as well as CTDV data are automatically recorded
at intervals of two seconds in the bubble cassette, while the majority of
other data are recorded when the "LOG" area on the screen is depressed (LOG
recording).
87
•H
4-1
a r—i
a>
a)
■u W)
c ccl
o a-
o
60
4-1 ö
H •H
CO CO >
co a) o
QJ H
H M iH
a) Cd 00 O
tj o c
o o
60 cd +J
U -H T3
•5u PM Ö
wtD
0)
g
•3
4-1
a) a
o
I o
CM u
cd
Q) cu
i-t
■s
U
o
<D XL
M
cu
4J
g
■u «
n a
 Ü
,H cd cd cu
Ci< (A |3 W
ifl vD IN CO
CN CN CN CN CN CN
88
u
o
I-l
o
o
b
Cd
I-l a) CU
p. •u 4J
to •H = T3 -rl
•rl CU ,C
p Ü Pi &
cd
o o CU
.4J 4J
3 3
O O CU
>.
co
>►>
co
5 CU
IS |5 3 .0
o o •rl 4J
r-f I-I
>. tK i-l H 3
CO co cd cd >H
i-r Ö c o 
tfl 01 Ü o 4-> ,3 CO 3
r-t M CU rl M CO & CU
ft Pirl Oi cd cd o 3
X CU CU cd M f*>
<D ß cd >>
4-) rl -rl H cd
0 cd -d ftr-H
U co ft
CO ' >s O •H CO
•u 
89
Photo 3. Display Pattern of Data Display Mode (1/3)
(R) - Red
•• Green
/__CY) " Yellow
nf-
NEXT MENU
V
(R) (G) (R)
Bus NNN
(R) (G) (Y) (R)
AC NNN
CR) CGT"© (Y) (R)
DC
KG
(11) (Y) (G) -®
Bus' NNN
■"s. (R) (Y) (G)~
m-
AC NNN
(R) (Y) (cr~©
DC NNN A:
Data ]i/o6 LOG
YYYY MM • DD h h : m m s s display /
/ -V
/
©
Figure 8. Display Pattern of Data Display Mode (1/3)
90
-O -d ■d TJ ■a
CU cu CU CU (U
u u rl rl rl
#1*1 A •t , ,#i
IS
o
|3
o
IS
o
13
O
go
: .rH --rH rH H rH
•0' . -rH H rH rH rH
cu cu cu CU CU CU
M , !>v !* >s '.>> ■ r^
co |S «'&' -. V'« ■JJ & « A

S •'••■.'.!' ;
.
5 o cu
&'-■•
ri cu ß O ß ß O O ß o ß O ß 
I-< H E E • 2 2 •rl CU H CU
CU H CU rH H CU H CU H •H . E' ... , E ■d .rl
o CU fl U CU U CU U H O |H et 3
.?- Ü
-
>l O tH
;-■.„..,.,.'■:..._,-
■ rr) '-
m
',''.''■'
. >. cd '
Cd 4-1
co d) CO CO rH td
CU <X) (U CU CV-Ö
> Q > >: CD
>. •H S "H TH . ■-'.■rl ß
...■d -H
cd 60 00 00
t-t CO
& td *s» cd
CU CM, 0)
cd
0)
CU CD
■ : . rl «H
co .
•rl rl . rl U. cd
-d > > td >> cd cd ß
: > C3 C3 « rg ■ .. cd • i .CD (U '
«t-t o o s s S Q) HO)
,ß: cux o) 0) a <u
o m in o Ä ö0 u u
H rH u-i rH rH rH 4-1 CO 4J TJ 4J ß cd u
ß •H O • ■• •H H CD
o CU CU , 0) CU CU CU öo-ß oo a öOTJ cd
•H rH rH rH H rH rH ß ß :• • ß rl 0)
cu 4J cd cd cd cd cd td •H Cd -H 3 -H O Ö0,ß cu
•d tö C? ü a ü CJ Ü CJ CD 4J .01 Ö CD CJ ß 4J "d
O ß > E E AS E £ CO CO CO CO co co CD Cd C0 CU CD 0) -rl O
S 13 > 0- •H rH rH rH rH rH rH A, 3 P4 3 ft O cd O co
cd X e =
3 3 3 3 3 3 (U O E E 1*4 pH p"4 pH [XI PK p >s P S P iJhtflH
CO
•H
P
ß ß
rt cu cu O o
4-1 i-l rH •H •H
to cd cd 4J 4J
P ü ß cj ß td cd
ß CU co o co O ß CU Ü (U Ü
O o •rl •H O H •H ß rH •H
• •rl ß cu w cu 4-1 •H cd
TJ O td T3
co 4-> cd u oo to oo cd 4-1 Ü ß -H CJ ß " , ■
CO 4-1 cd cd u cd CJ ß cd CO •H 4-J CO •H
0) i-l CO 2 CU 4J -rl 4J •rl O rH td
rH 3 •rl >. r-t T3 rH T3 •H 3
cu
rH
cu
,0 co co O ß o ß 4-1 CO 3
rt ß CU fl > >rl > •rl cd ß CJ CD Ü
u o
H
>
•rl
> ß >
CU
td -d
•rl
•H
>
§
4J
ß
•H s
4-1
> > o > o ß CO > co
O O 0 o CO •H O •H •H
o oo O oo o oo O CO bo CD
r-1 CM rH CM «I rH CM r-i CU CM CU
CU (U u rl
O O CJ CJ cj cd o o CJ CJ O
<! p ß . CU >i J3 rH *H î PS CU ■>> >. ,cu CU '
p. CO rl CO cd -d co co M CD co rl rl
co A CJ ß cd cd
•H co ß co cu CO vH CO CO ß CO co
o a •u
>> ■''
T3 ß ß ß O ß ß Ö0
b
o •rl O 3 o o •H o o ß cd ^
m CiJ vH • 4-> >H ß CU TH -H 4-1 •H •H •H r-{ Cd
o 60 4J at 4-i •rl 60 4J 4J cd 4J 4J na Cbi-i
co to iH cd a cd
■u
cd cd rH
3
cd td Cd
cu
Cd
cu
rl
o
0) ft
•rl CD
CO +J CJ 3 Ü CJ CJ CJ Ü
4J i-l -rl CO vH «1 rH -H -H CO •H •rl ri rl O •d -rl
ß O a ß Ö "Ö O ß ß ß ß ß cd cd cu •d
 3E •H 3 2 ß > E 3 E 3 E -H • 3 = aÖ 2 rl a
4-1 o 4-13 u cu
ß CO § cd 1 o CO 0 a CO B § rl ß Ü cd -d
o 3 O 3 O CU 3 O O 3 o O CU 0) o H O
o PQ O pq o CO pq u u pq o O S5 • JE! . r3 «! s
o HMCO «* LO vO r^ COO\OHN n -* m vO r>« oo o\ OH CM co <(•

S3 rH rH rH rH H rH rH rH rH H CM CM CM CM CM
91
6C
3
•rl
U O O
•rl
3 HO«
0 n) ft 3 3
s s
3 U 60 •rl s ft
O u
•H CO CO
4J t) 13
•rl CU 4-1
•Ö 60 3 ■U 0 ft
3 -Ö ft: 3 o co :
O 3 3 O U -H
O r, -H O Wtl
I I Vi
Vl Vi CU ft ft CU
3 60 o O
3 CJ U
cd •■ ft
u H
t) M Vi
4-1
Vi
0> CU CU
>> >■. 43 o cu
vi Vl Ü 60 rl s 3 u VI 60 ÖC Vi M Vi 00 4-1 4J 60
60
o o CO CO 3 3
CU
4-1
CU
4-1
co e
•rl -H •H •rl
CU
4J
CU
4-1
4J o
ö _ ö
3 rl 3 -H
3■ O 3
CU
4-1
co
.53 3
U
3 ■rl
Vi 3
•rl cd
ft
4-1 4-1
I
4-> 4-1 -Ö 3 C 4J
CO 00 «S 60 Vl o cd cd cd cd •H M -H U 3 Vi 43 43 VI
H 43 43 CO cd co cd Vi cd 4-1 cd 60
•P ^ Cfl 3 >. cd •H H 3
13 H 13 43 13
3
ft
•rl Vl
3 's 3 3 CU
4-1
cd CO
3
co 3 3
r-i
3 3 4-1 *vS H ft •rl
•H Vi -H n 4J 3 •H •rl ft ft cd 13
4-1
cd cd cd cd 4J j*, 3 •5 •rl
5P 0
cd
60 0
Cd O
60 U
o
r! cd 3 >
00
3 3 3 ft ft 4-1 Ü 4-1 CJ 4J o 3
•rl
4J
C Ö
•H -rl 'H rH
O •rl 00 CN rl Vi N •rl o
CU H <N CU •rl 3 <4H
CJ 4-> 3
4-» ' 3 O 3 IN3
•rl • 3 • Ö •H
O . •H > •rl 33 o
> Mgtr PQ 15 <! !* P S (3
o -o
cu
4J Vi
CO
•H
o
o
O
o
CU
00
CU
4-1
3
•rl
I
o o M
3 Vi Vi VI 4-1
cd ft
o 3 ft CU cu cd CU 3
■sCO
4-1 3
4-1
gu s
CU
M
T}
O
•rl
CS O
M •rl
4J
C3
AS
rl
CU
rl
CU
3
O
Vi
ft
>
o
vi
cu
r-i U
cd cu
CJ 4-1
4-1
CO
3
4-1
3
o
'I
•rl
O
O
!-l •rl 4J
in cd m O co •rl VI N
3 u H •CJ 3 cd
1
CU •rl 3 3 4-1 43 •rl
0) 3 3 O rH CN rH
CM cd cd
CO •rl Vi VI VI
>. •H 3 3 43 43
■s
4J CJ cj o
CO > CO > r-i cd 43 CU
c M
3 O 3 oo 3 3 3 'S > r-i Vi Vi U 4-1 Vi
60 CU cd rl CO Vi
TH CO cj cd
3 
CO •H CO rl
cd 43 > rl IS 4-1 O EC -H & 4-1 O CO
I Vl
VI VI cu CU
3 3 60 CU CU CU 4-i
CJ CJ VI CJ CJ Ü
Vl • VI rl
cd 3 3 3 CU CU
cd cu cd cd CU
!>> 43
4-1 Vi Vl CU 60 Vl CU Vl 60 4J CU 4J 60
M VI O 60 4-1 H 4J
CU r-i CU 3 CO r-i CO %
CU CU CO 3 CO 43 CO CO cu CO rH 3
•rl •rl •rl 3 •rl 60-H 60 4-1 4J 3 43 •H 4-1 43 4-1 ■rl 3 43 3 •H
4J CO 3 CO 3 U 3 Vi 3
3 W 4J O 3 3
I
T3 3 CO o CO
cd OO cd VI CU VI CU •rl Cd cd •H O Vi 3 O 3 Vi 43 43 Vi
43 CO Vl co cd Vi Vi Vi cd 4J cd
X> G & >, fd VI 4J VI 3 43
rH 4J 43 4-> 43 > 13
•H
3 3 3
H > 3 (3 cd 3 3 3 • 4-1 4J r-i ft
o co o •rl ft ft cd
•rl
. . VI -H
cd cd cd 4J >» •rl 3 •rl ^ä
0 0) 0 cd > <-H ft 4-1 ft 4-1 CJ 4J CJ o 4J r-i
VI Vl Vl N Cd 43
rH CM 4-1 rH VI r-i O rH OO rH 43 CN 43
3 3 3 cu •H & 3
s 4J
• 3 -H
• 3
O CU O cu cd 34
4J 3 3 CU 3 O 3 CN
CO 4-1 co ' "
3 CU 3 CJ 3 CJ O Vl
• O
3
•rl
cd
•rl
Cd
a >
Vl
CU &
Vl
O
cd O
cu Vi
55 vi la vi a a H 0 H <! H P 55 a •H > •H W OO 4-1
co «* m vo t~» oo a\ CN PO in
s r-i
92
6t
3
•H
U o o o o oo oooo
•H
4J H
3 öO CU
P. s 3
3 -H cd
3 •H 3 ft
O M
•H 4J Cd >>
+J O |5 cd
•H 3 H
T) ■P 0 ft
(3 e O
O o M •H
U CJ M-l T3
M M M ÖO 0) H
M cu cu cu s M cd
0) 4J 4J 4J vH 3 «H
4J MUM 3 CD U
U CU CU CU M ÖO CD M
CU ' > > > cd ö cu cd
> 0 ti 3 ÖO & -HM ft ÖO
3 •H ,-H «H t) ß 3 ft 3 0'
•H •H Tt M -H CU
ft ft CD 3 3 Ai Cd rH CU3-U
& 0 ■■' 'S 3 C7- M 3 & cd -ti M CD
+J e 3 3 O -H Cd cd -H -H cd >-.
3 3 ft ft -H rH £ 4-> -U K |3 CD
ft ft •u ;e M O
•u O 0 cd JsS >, CU Cd • -H CJ
3 U •H -H O 3 >•» U 4J ft T3 CU -H
o Q) ÖO iHÖOrHÖO-HÖOCd+JCd CD ÖO M rH ÖO
4-> 3 3 3 3 3 3 3 4J -H -H >, 3 3 3 3 3 3 ■
tu CO -H Cd -H Cd vH 3 -H 4J H CD CU -H O CD Cd -H
o & 3 M 3 U 3 i 3 0 3 -H ' ÖO 3 ,£> CD M 3
•H cd u •o ■M t» U S M-H Cd X U >, M M CU -Ö M
o cu cd >% cd >> cd 0 cd U 3 3 -H X cd cd U >> cd
> w & EÖ & S SO 13 H W cj M cd U U 3
u cd ,3 -H cd cj cd '3 CD
CO CU Xi •H H ^3 -H \C CD CD 0 • -
4J ■P o 303,0 COCU Q)
3 cd n p. cd u 3 M ^<ä cu u u
dl |3 CU 0 M CU 1 CU 3 U P. CD'
4J cd > 3 T3 > 0 ; > cd p. >>
s cu o P. >» O O ■O 4J rH CD
o CD ,3 0 0 rH Cd
o O ^ >i CU Cd -HO
u •H M . M 3 M 4J >r( 4J -H
>, ÖO CU rH ÖO , U- cd 3
i-4 •H M cd >H M -H M H CO cd P. Cd
ft M 3 CU 3 CU 0 -H 'P- M
CO
•H
S'2
cd .3
TJ
'>»'■'
M
cd
>
3
M > -H
cd 3 M ■
X ■
3
M ' ■
-HCMO
CM T3
!>%
Q & -H 33 & -H |3 -rf H <! <J O ■ U ffl
M U U (U M ÖO CU
M CU CU CU H T4 3 M rH
CU 4-J-W+J ^3Cd-H3cd
S
4J U U U 3 -'3C0'H
M CU CU CU O U CO 4J
CU > > > M CU Cd CU M CU
> 3 3 3 öO-UO^M 3öOM
s •H -H -H ■■■' t) C -H P. p, 3 3
•H •H -H > -H CD
P, p, CD SB^CUCUrH CU3C0
&
■"•S
0
3
0
3
3
O
O* U 3 -3 r-i 3
-H cd cd 3 -H
fl M Hl
-H cd M
.3 P, P, -H rH 'S 4J <U W ■ +J X S P.
ft 4J ÖO CD U O
O Ü 3 ^S ^ Cd <U 3 -HO
1-1 •H -H O 3 t^ U U U P. TJ CU vH
CU ÖO rH CU rH ÖO -H CU Cd 4J Cd O P. CU U H Q)
•U 3 3 rH 3 3 3 rH 4J -H vH 4J • 3 rH 3 3 3 rH
cd -H cd ^3 cd -H 3 rQ u H ■ ja/-.'dl 4i o oi d ,o
a § 3 U 3 M 3 S 3 0 3 -H ." , CU ÖO 3 41 CD M 3
CU cd M T) O T3 M S O -H cd X M P. ^ O U . U >, Cd O M M 3 3 -H -H. X U Cd M >, M
H w £ WWW S 'ü u H rji <j <| fto 4J y & 3 iJ
•
O vO r^ "öo oi '■ o rH CM n " -<r m
S3 rH rH rH rl N N N CS CM CM
93
For reproducing these data, they must be fed into the integrated data
display system's data reproduction section aboard the support ship Yokosuka,
after the submersible is hoisted aboard the mother ship. The reproduced
data are processed as needed, and the results are forwarded to the printer
or plotter, while simultaneously being recorded on a floppy disk, to be
output as a daily report or as graphics.
Of the recorded data, those concerning the submersible's position, CTDV, etc.
are delivered to the acoustic navigation system (ANS) aboard the mother ship
for sound speed corrections and are recorded as the corrected diving data.
The conversion of the recording format (MSDOS <-+ IBM format), a process
necessary for effecting an exchange of data with ANS, is also conducted in
the integrated information display system's reproduction section.
4. Conclusion
The foregoing gives brief outlines of the layout of equipment inside the
pressure hull of Shinkai 6500 and information displays by the integrated
information display system aboard the submersible.
Much has been left out about the integrated information display system and
the system's data reproduction section in this technical report, and we
would like to give a more detailed report the next time around if it is
possible.
References
1. Shinichi Takagawa, "Shinkai 6500 Challenges an Unknown World," Marine,

Vol 21, No 10, pp 62-72, 1989.
94
MARINE TECHNOLOGY
Design and Construction of Spherical Pressure Hull of 'Shinkai 6500'
916C0003B Tokyo KAIYO KAGAKU GIJUTSU SENTÄ SHIKEN KENKYU HÖK0KU in Japanese
Mar 90 pp 329-343
[Article by Shinichi Takagawa, Daisuke Kiuchi, and Kenji Takahashi of the

Deep Sea Technology Department and Yutaka Yamauchi and Kazuya Inoue of the
Kobe Shipyard, all of Mitsubishi Heavy Industries Ltd., and Takashi
Nishimura of Kobe Steel Co., Ltd.]
[Text] [Abstract]
The pressure hull of a research submersible is required to have enough

reliability and strength to ensure the safety of the crew against the water
pressure, and at the same time lightness in weight is also required for good
performance of the submersible, it is therefore very important in the
design and construction to select a reliable and high strength material, to
apply an appropriate design method fit this material and to apply minute
machining technologies for required qualities.
JAMSTEC has decided to use a spherical pressure hull made of titanium alloy
(Ti-6A1-4V ELI) for the "Shinkai 6500" submersible after confirming the
characteristics of the material by pressure test of scale models, consider-
ing production capability of the material and the development of design and
machining technologies.
This paper describes the development of the pressure hull from the selection
of material to the final test.
Keywords: Spherical Pressure Hull, Titanium Alloy, Three-Dimensional

Machining, Electron Beam Welding.
1. Introduction
In bodies like research submersibles that move around underwater on their

own, it is very important to build the hull as small and as lightweight as
possible. Since the amount of energy that a research submersible can carry
is limited, it cannot travel on its own to the survey point. Consequently,
it is transported, carried aboard the support ship, to the survey point
where it is lowered to water and hoisted aboard another support ship.
95
The submersible must be small and lightweight if the job of lowering it to
water or hoisting it aboard the mother ship is to be done when the sea is
rough. Furthermore, if the surveying activity underwater or on the sea
floor is to be conducted efficiently, the submersible will have to be able ;
to move around nimbly. Such agility could not be expected of a large
submersible. Therefore, the submersible will have to be built as small as
possible. In downsizing a submersible if the design of any constituent
element took on large dimensions, everything else would have to be built
large; the related supportive elements would have to be built large,
including the propeller, in order to obtain the required thrust. The
battery, the energy source, must be built large and their supportive
elements must be built large. Consequently, when seeking to downsize a
submersible, we must strictly apply the principle of downsizing to the
design of every one of the ship's components.
The largest and heaviest of all components of a manned research submersible

is the pressure hull that houses the crew. A success or failure in down-
sizing it has a great impact on whether or not the whole research submersible
can be built small and lightweight. Therefore, in the development of Shinkai
6500, a research submersible, the key was how to reduce the size and weight
of its pressure hull.
The pressure hull of a large depth research submersible must be small and
lightweight yet it must guarantee the safety of the crew. A spherical
pressure hull is most desired, but there is a limit to how far it can be
downsized because it also must accommodate the crew. As a result of a human
engineering study conducted using a prototype, it was decided that the
pressure hull of Shinkai 6500 should take a spherical form with 2 meters in
inner diameter. With the size of the pressure hull determined, the next
question was focused on how to make it lightweight.
To that end, the requirements were, 1) to selectively use high-reliability

material of high relative strength (0.2 percent proof stress divided by
specific gravity) and 2) to apply an appropriate design method and engineer-
ing technology that would reduce the superfluous flesh as much as possible.
In the course of selecting the material, factors which included the

domestic technology of manufacturing such material, the record of develop-
ment of the technology to manufacture a prototypical pressure hull, and the
progress in the development of the design technology centering on the
collapse strength, were taken into consideration. Repeated pressure and
crush tests were conducted using scale models before finally deciding on the
adoption of a titanium alloy (Ti-6A1-4V ELI) for the pressure hull.
The planned outline of the pressure hull, the procedure pertaining to the
selection of the material used, the design, the manufacture and the results
of tests are described below.
96
2. An Outline of the Pressure Hull
The pressure hull of Shinkai 6500 is a spherical of 2,000 mm in inner

diameter. It has ä 500-mm-diameter hatch at its top which is used as the
entry and exit port for the crew and three observation portholes of 120 mm
in inner diameter on its down side. Two central watertight cable glands
that contain various wires and cables for power or electrical signal
exchange'between the pressure hull and the outside, are installed at the
rear of its top. ■
The observation portholes and integrated wire arid cable conduits were not
ground out of the pressure hull material, but they were installed later on.
That is, the separately manufactured window frames (viewport1 coamings) and
casings for the integrated wire and cable conduits, were fixed in place and
welded.
An outline of the pressure hull is given in Figure 1.
Cable gland connector Conical

sheet hatch
receptacle
Retainei
171 ng
iewport
Viewport
coaming
Figure 1. Structure of Pressure Hull
Material Used
A titanium alloy was used for the pressure hull. Compared with steel,
titanium alloy has a smaller specific gravity and a larger relative
strength (0.2 percent proof stress divided by specific gravity), so the
material can greatly reduce the weight of the pressure hull. Thanks to
titanium alloy's corrosion resistance to seawater, a feature that is not
available with steel, a titanium alloy-based pressure hull does not need
to be built an extra millimeter thicker as a precaution against corrosion.
In the case of Shinkai 2000, the hull was of the plate thickness considered
necessary for the safety design, plus an additional 1 mm as admeasure to
counter corrosion.
97
Table 1 gives a comparison of calculated weights of the pressure hulls, one
made of titanium alloy T1-6A1-4V ELI and the other of ultrahigh tensile
strength steel 10Ni-8Co steel. From the table, one can learn that the
titanium alloy-based pressure hull is lighter than the ultrahigh strength
steel-based pressure hull by some 20 percent.
Table 1. Estimation of Pressure Hull Weight for Material Composition
0.2% Pressure hull

^*-\^^ Item proof Specific Plate Differ-
stress gravity Specific thick- ence in
O-0.2 strength ness* Weight weight
Material -^"^ (kgf/mm 2
) (g/cm2) an ?/Y (mm) (ton) (ton)
Ü-6A1-4 ELI Above 81 4.42 18.3 73.5 4.39

titanium alloy
1.07
10Ni-8Co steel Above 120 7.85 15.3 52.5** 5.46
* Inside diameter 2 m; safety factor 1.55; sphericity 1.00

** To include 1 mm as a margin for corrosion
4. Design of the Pressure Hull
4.1 Design Conditions
The design conditions for the pressure hull are given below.
(1) The pressure at the maximum diving depth was set at 680 kg/cm2, a value
equivalent to the pressure at a depth of about 6,500 meters.
(2) In accordance with the regulation on steel ships set by the Nippon
Kaiji Kyokai (Japan Marine Association), the design crush pressure was set
at a value above 1,058 kg/cm2, a value equivalent to the pressure at a depth
of about the submersible's maximum diving depth, times 1.5, plus 300 meters.
(3) The 0.2 percent proof stress of the titanium alloy used was set at
81 kg/cm2 in accordance with the regulation of the governing standard
AMS4907.
(4) The sphericity, which is the ratio between the standard radius R0 and
the local radius of curvature R10, was set at 1.004. One mm and 0.45 mm
machining tolerances were given for the radius and the plate thickness,
respectively. As described in paragraph 3, no special allowance was given
for corrosion.
The strength of the pressure hull is given by formula (1) or formula (2),
described later, and it is determined by the ratio between the plate thick-
ness and the local radius of .curvature. When manufacturing a pressure hull,
the plate thickness cannot be changed according to the local radius of
curvature, and it is a constant within the limits of a given machining
98
tolerance. Consequently, if a pressure hull, having a strength to withstand
a given pressure is to be made under the condition that the local radius of
curvature fluctuates greatly, the plate thicknesses all over the spherical
hull will have to be in conformity with the largest local radius of curva-
ture. The result is that in areas where the local radius of curvature is
smaller than the largest local radius of curvature, the plate thicknesses
are larger than the required values, that is, those areas are burdened with
excess amounts of flesh.
For reducing the weight of a pressure hull, eliminating the extra amount of
flesh is highly important. To that end, it is necessary to minimize the
fluctuations in the local radius of curvature and the index for the change,
which is the ratio between the standard radius and the local radius of
curvature (i.e., sphericity) and must be brought to as close a value as
possible to 1. The sphericity of the pressure hull for Shinkai 2000 Was
1.07 but the corresponding value for Shinkai 6500 was set at a figure
smaller by more than an order of magnitude, at 1.004.
4.2 Formulae for Calculating Collapse Strength "
For calculating the collapse strength of the pressure hull» we used the
collapse strength formula of spherical shells developed by M. Krenzke et al.
at the David Taylor Research Center (DTRC) in the United States.
The formulae for calculating strength are given below.
Elastic collapse formula :;
Elastic and plastic collapse formula
PE = 1.4V Et, Es/3 (1 - v» ) • CA, //?io)2 .-;. ;\(2) ■■■....■
where R^Q = the local radius of curvature

ha = the local thickness of plate* :
E = Young's modulus M *•'"••:: ■'
v = Poisson ratio
Et = Tangent modulus to the average stress
Es = Secant modulus to the average stress s
Modulus •"•■'■■■ ...<■■.■■<:
a
avg = PBRlo/2ha(Rlo-ha/2) :
p„ = kpE ' ' •' ■ ■ ' . ■ ■'■■'■'— '■■•
k = the correction factor for collapse strength by the
machining method
Since the formulae for calculating collapse strength developed by Krenzke

et al. are empirical ones, in order to confirm their reliability, the
correction factor k was obtained in a series of tests which were undertaken
99
by considering the size of the initial-stage inaccuracy, the material used,
the effect of openings, etc.
The design curve obtained from the results of these tests was used in the
formula (3) by Krenzke et al. It is given in Figure 2. The design curve is
the same as that used for the pressure hull of Shinkai 2000.
1.2
1.0
0.8
OHT80 OHT80(with flatspol)
0.6 (DNS 63 (D'NSG3(wUh halch)

Krenzltu
( rJ,V-
iS
"-'20(l(
'. L X- *■ <• • ' ©NS90 <D'NS63(with vidwpoi
\£) ssoo KM ma / of peiiL'trofon )
0.-1 01S%Nimaraging
GNS90(with hutch)
O10Ni-8Co
0:2 ATi-6AMVELI
3 4
.Pi/P*
Figure 2. Compressive Strength for Spherical Pressure Hull
Key:
1. Experimental value
2. Design curves for Shinkai 2000 and Shinkai 6500
4.3 Local Strength and Allowable Stress
The formulae described above apply to the general area of the spherical
shell. However, a pressure hull is not a simple spherical body. It has
fixtures that differ from the general area of the spherical shell, such as
openings like a hatch and observation windows, as well as central water-
tight cable glands. The aforementioned formulae cannot be applied to these
fixtures. Therefore, these local strengths were designed whereby the
allowable stresses met the following two conditions.
(1) The total of membrane stress and bending stress is below 75 percent of
the 0.2 percent proof stress prescribed for the material used.
(2) The local stress is less than the 0.2 percent proof stress prescribed
for the material used.
Along with the above, in the design of the openings and coamings the follow-
ing considerations were given.
100
4.3.1 Observation Porthole Coamings •..,.
The observation porthole has the shape of a truncated cone having a 90-
degree apex angle. The observation porthole coaming to which the observation
porthole is fixed is of a shape that enables its rigidity to equal that of
the pressure hull, and consideration is given to making the bending moment,
where it joins the pressure hull, as small as possible.
The results of stress calculations by the finite element method of a

spherical shell having an observation porthole coaming are given in
Figures 3 and 4. Figure 3 is a stress contour along the longitudinal
direction (the North-South Pole direction) while Figure 4 gives a stress
contour along the latitudinal direction (the direction parallel to the
Equator). The maximum value of the sum of membrane stress and bending
stress, which was -60.4 kg/mm2, occurred at the root of the observation
porthole coaming, while the maximum value of.the local stress, which was
-77.4 kg/mm2, was generated in the bore of the observation porthole coaming
opening. All the stresses were within the limits of the allowable stresses.
Unit:, kgf/mm Unit: kgf/mm
Figure 3. Stress Contour Along Figure 4. Stress Contour Along

Longitude Section -Latitude Section
(External pressure (External pressure
680 kgf/cm2) 680 kgf/cm2)
4.3.2 Central Watertight Cable Glands
The centralized watertight cable gland is of a multipore plate type in

which a reinforcing insert plate contouring, along the spherical hull, is
provided with openings. These are for the attachment of the receptacles
for the connectors for the penetration pieces. The insert plate has a shape
101
that enables it to have a rigidity on a par with the general area of the
spherical hull and consideration is paid to making its bending moment as
small as possible.
The result of calculations by the finite element method, using a three-

dimensional thick shell element of the stresses that a spherical shell,
provided with a centralized cable gland, receives when subjected to an
external pressure of 680 kg/cm2, is given as a contour of the maximum main
stress in Figure 5. The maximum value of the local stress was generated at
the outer side of the opening on the external circumference of the shell,
and the value at -69.1 kg/mm2 satisfied the allowable stress.
Central opening
ôpening on
S external
circumference
Figure 5. Contour of Maximum Main Stress (External pressure 680 kgf/cm2)
4.3.3 Hatch
The hatch, which is a conical sheet hatch of a truncated cone shape, looks
as if it was made by cutting out a part of the spherical hull. It has a
structure which when subjected to external pressure, retains continuity
with the pressure hull proper. The key to retaining the strength of a hatch
is to make sure that it will not slide when subjected to an external
pressure. The behavior of a hatch is affected by the chamfer angle (the
angle created by the line extending from the external circumference of the
globe to its center and the truncated cone generatrix), so, based on the
results of friction and behavior tests of titanium alloy and the repeated
pressure and collapse tests using spherical shell models having a conical
sheet hatch (described in paragraph 4.3.4), the chamfer angle was set at
3 degrees.
102
The stress analysis by the finite element method was conducted for a spherical
shell having a conical sheet hatch. In the analysis, the existence of a gap
friction element was considered between the conical sheet hatch and the sheet
surface of the pressure hull proper, but the hatch generated no slide when
subjected to external pressures and the stresses were all within the limits
of the allowable stress.
4.3.4 Tests Using Scale Models
In generating aforementioned designs and calculations, we conducted tests

using scale models.
First we produced three titanium alloy models (a model with no opening, a

model with an observation window, and a model with a hatch) for the testing.
Our collapse tests and analyses confirmed the following:
(1) It is possible to estimate the collapse pressure of a titanium alloy

spherical shell by the conventional design methods.
(2) The use of the conventional methods enables the effect of openings,
such as an observation window or a hatch, on the collapse behavior to be
ignored.
(3) A creep at ordinary temperature does not have any effect on the collapse
behavior.
Then, we manufactured a 35-percent scale model, made of titanium alloy, of

the actual pressure hull, using the same design and manufacturing method.
We subjected the model to a repetition of compression tests numbering 1,500
compressive actions under conditions equivalent to the pressure at 6,500
meters below sealevel, and then subjected the model to a crush test. From
the experiment, the following have been confirmed:
(1) There was not observed any deformation in the shape of the pressure
hull nor a crack in it even after it had been subjected to a repetition of
1,500 compressions.
(2) The hatch and observation portholes were well sealed even after
repeated compressions and decompressions.
(3) The repeated application of pressures numbering 1,500 times had no

effect on the collapse behavior of the pressure hull.
(4) The crush pressure was 1,397 kg/cm2.
The crush pressure in (4) is 1.32 times as high as the design collapse
pressure, and a survey has revealed that the large crush pressure was mainly
due to the fact that the proof stress of the material used was larger than
the 81 kg/cm2 described in paragraph 4.1.3. The proof stress of material,
however, varies from manufacturing lot to manufacturing lot, and it is not
practical to confirm the proof stress of the material to be used each time
103
and accordingly change the design, so we decided to adopt the guaranteed
performance in the standards, 81 kg/mm2, ln the design of the real pressure
hull.
5. Manufacture of the Pressure Hull
Manufacture of the pressure hull was undertaken in the order of these ^

described procedures: manufacture of rolled plates; hot forming (heating
the plate in a furnace and pressing it into shape while it is hot); heat
treatment; preliminary machining of hemisphere; putting into place of the
watertight cable gland and welding; three-dimensional machxnxng and _
processing; equator line welding; machining and processing of the equatorial
section; fixing into place of the observation porthole/watertight cable gland
connector receptacle/conical sheet hatch; compressive strength and leakage
tests; and coating. A schematic flow of these procedures is given in
Figure 6, and explanations are given on some of the major steps of these
procedures.
5.1 Manufacture of Raw Material
The raw material for the pressure hull's hemispheres was dissolved in
vacuum (mixing the raw material porous titanium with the additives such as
aluminum and vanadium and dissolving the mixture by an electric arc m
vacuum to obtain titanium alloy); the obtained ingot was forged, beta-
processed (the heat treatment involving high-temperature heating and rapid
quenching in order to obtain finer grains of the titanium alloy) and rolled
to obtain a rolled plate of the titanium alloy 125-mm thick. A disc of
about 3 meters in diameter, cut from the rolled plate, was subjected to
hot forming in a press to obtain the raw material for the hemisphere. #
Further, after the hot forming the raw material for the hemisphere received
a solution treatment and over aging (STOA) (the heat treatment involving
quench-arid-temper to increase the strength and tenacity of titanium alloy).
An external view of the hemisphere after hot forming is given in Photo 1
[not reproduced].
Test pieces were obtained from the extra portions of the hemisphere material
after heat treating and they were subjected to tensile strength.te^s in
order to confirm quality. The results of the analytical test of the wiled
pLte are giveii in Table 2, while the results Of the tensile strength tests
of the extra portions of the hemisphere material are given m Table 3. The
test results were all satisfactory.
Forged material of titanium alloy, manufactured by forging, beta processing,
and billet forging of the ingot obtained by means of vacuum melting, was
used as the starting material for the observation porthole coamings central
watertight cable glands and the conical sheet hatch The material for the
SntralSwatertight cable glands was manufactured by. ?°^ «^^^^
shaping and heat treating the forged material, m this order. The material
forPtht observation porthole coamings and the conical sheet hatch was manu-
factured by forming and forging and then heat treating the forged material.
As with the material for the hemisphere, both materials were subjected to
analytical tests and tensile strength tests to confirm their good quality.
104
D (2) (5) (4) (5) ( 6)
(10)
1 <»*,»»■ 1 „ t
£ — * — f—- — — a
> ft L (11)
■
u J «s KB«
t L_ a «0 I
y a
;«
(12) (15) (14). (15) (16) (17) (18) (19) 120)

«»BITS* »«»«««a imu »We-'»«
BB»*« (PWHT) (HlVUa)
■'©Ä
(22)?(.2$) (24) (25)

Sift jvi/- h
Figure 6. Production Procedure of Pressure Hull
Key:
1. Sponge titanium 15. Edge preparation of openings
2. Melting in vacuum 16. Fixing into place of pene-
3. Ingot tration piece and its
4. Forging electron beam welding
5. Beta processing 17. Attachment of lower structure
6. Slab forging of hatchway and its welding
7. Rolling 18. Post-welding heat treatment
8. Cutting 19- Machining (surface and
9. Hot forming passages)
10. Heat treatment 20. Electron beam welding in
11. Preliminary machining of equator
surface 21. Post-welding heat treatment
12. Attachment of the mount for 22. Machining of the equator
pressure hull and of the 23. Fixing into place of view-
pieces for machining and ports, cable connectors and
their welding conical sheet hatch
13. After preliminary machining 24. Compressive strength and
of inside and outside surfaces leakage tests
14. Drilling of openings 25. Coating
Conical sheet hatch opening
Central watertight cable
gland opening
Viewport coaming opening
105
Ö C Ö a a
•rl •rl •rl •rl
■a CO cd cd 'cd
s §
ei
§
a! a!
s
r-4 O o o o o
CO CO CO r-l
cd o o o 
o o o CD
O
1—t rH rH
o ■ o o "^ *^
:
> ' 8 o
o
o
o
O
O
o
q
U
pQ
o
o V V v V o
co 05 13
CM co CO to o
co <N)
o
o o
a o
o o o o o cu
I8 o o - o o o PQ
CD OO
4-1 CO ■ -a- LO
•rl
CO
o o o o
o
E-l
U
z O
o
o
o
o
o o 1 2
O
14-1
■H
CO
£2
CO 5
•rl o o
o
o
o
o
o o
oc
°.
cu
PQ
CO o
>»
rH
Cd
T
o
OS en a; ,iH
c
o o
o
o
o
o
o 2 co
—.
CU
PQ
o o
•rl
4J
•H i LO CO
O
CO
CX> OS
t—I r-4
rH CO
CO LO a)
T-H
O CS1
o o o o co. CQ
c
! o
Ü LO
O ■
CM
in
CO
INI
i
o
^
CM
> -a* LO
cö
CD
H o
LO
■8 05
co
CO
<
o
«j
< cd CO CO CO LO
6o eo cu
O-,
o o
+->
..H . o .. c cd
CQ CQ >
Ö
o aaaqds aaaqds
-r <"
•rl
CO
uaaq^nos JO
•rl ar^d psiio Harold psiio
- a
o ssiaqdsTuiaq aoj CO
CU
siöTIB ranfuB^Ti V-i
_1
106
Table 3. Tensile Strength of Titanium Alloy of Hemispheres
1
What
direc-
Where tion 0.2% : Contrac-
sample sample proof Tensile Elonga- tion of
was was stress strength tion area
Division taken taken (kgf/mm2) (kgf/mm^) (%) (%)
■
Titanium Alloy for 1/2 t L 85.7 91.7 11 26

alloys northern
for hemi- hemi- 84.5 90.2 14 24
spheres sphere
T 85.5 91.5 12 26
84.4 90.7 14 26
Alloy for 1/2 t L 86.7 93.0 15 25
southern
hemi- 85.7 90.7 15 . 19
sphere
T 84.5 90.4 16 23
84.7 90.7 16 . 12
Prescribed value Above 81 Above 88 Above Report

10 only
5.2 Penetration Piece Fixed Into Place and Welded s
The material for the hemisphere was cut. in a preliminary machining so that
the inside and outside surfaces retained a margin to machine later on; and
then the openings for the observation hole coamings and the penetration
piece for the central watertight cable glands were drilled, followed by the
drilling of the opening for the conical sheet hatch. The penetration piece
was also cut so that the inside and outside surfaces retained a margin to
machine later on, and it, together with the openings on the hemisphere and
the penetration piece, was machined for edge preparation.
Fixing into place of the penetration piece and its welding was conducted by
electron beam welding by attaching a backing strip so that the beam would
not pierce through. The welding work of the penetration piece is given in
Photo 2 [not reproduced]. After the welding was over, the welded part was
subjected to a post-welding heat treatment mainly aimed at removing .the
residual stress.
After the fitting and welding and the post-heat treatment were over, the
welded part was subjected to color contrast penetration testing, ultrasonic
testing and a radiographic inspection to confirm its soundness, and the
results showed that both the surface and the interior were sound.
107
5.3 Three-Dimensional Machining
After the penetration piece was fixed into place and welded, the external
and internal surfaces of the hemisphere were subjected to three-dimensional
machining. In the work the position of the cutting edge of the cutting tool
was precision-controlled three-dimensionally by computer and the bulging
portions of the spherical shell, such as observation porthole coamings and
the penetration piece, were cut and machined at levels of high precision to
remove the superfluous flesh from the pressure hull as much as possible.
The machining work is given in Photo 3 [not reproduced].
In the case of the manufacture of the pressure hull of Shinkai 2000, the
hemisphere was first machined to the finish dimensions and then the penetra-
tion piece was fixed into place and welded by TIG welding, and the result
was that the deformation resulting from the welding stayed put in the
finished pressure hull. In the case of the pressure hull of Shinkai 6500,
the penetration piece was fixed into place in the preliminary machined hemi-
sphere with a margin for machining later on and welded by electron beam
welding, and thereafter the inside and outside surfaces of the hemisphere,
including the bulging portion of the penetration piece, was three-
dimensionally machined and ground to the finish dimensions. This enabled
the finished pressure hull to take the shape of a rough sphere.
In the machining of the hemisphere, the external surface near the equatorial
joint was cut with a margin for machining later on, left intact, and it was
machined to a finish after hemisphere welding in equator was over.
5.4 Hemisphere Welding in Equator and Machining of the Equator Section
After machining of the hemispheres was over, the north and south hemispheres
were welded together in the equator. Electron beam welding, the technique
used for the insertion and welding of the penetration piece, was also used
for the equatorial welding. Photo 4 [not reproduced] gives a view of the
equatorial welding.
After the welding was over, the whole sphere was subjected to a post-welding
heat treatment to remove the residual stress. The work followed the same
procedure as described in Paragraph 5.2. After the post-welding heat
treatment was over, the sphere was subjected to nondestructive testing to
confirm that both the surface and the inside of the weld were sound.
Later, the margin for machining on the external surface near the equatorial
joint was cut and ground to the planned dimensions.
5.5 Testing Welding Work
It is impossible to directly test the mechanical properties of the welds of

the actual pressure hull. We therefore used a testing method in which test
pieces were welded by the welding techniques that were used in the real
pressure hull and subjected to a post-welding heat treatment. These test
pieces were then examined for their mechanical properties.
108
The results are given in Table 4, and all of the measurements satisfied the
prescribed values.
Table 4. Welding Procedure Test Results for Pressure Hull
Joint between
Division Judgment standards Equatorial pressure hull
joint of the and penetra-
pressure hull tion piece
Tensile 0.2% proof Above 81 93.5 87.2
strength stress '
of welded (kgf/mm2)
metal
Tensile Above 88 104.4 97.5
strength
(kgf/mm2)
Elongation Above 5 8.1 6.2
Elongation < Only need to be 21.6 14.8

reported
Tensile test of Tensile strength above j Tensile Tensile

welded joint the base material's ' strength ) strength
standard value I 100.4 | 100.8
(88 kgf/mm2) j kgf/mm2 kgf/mm2
— 1
Side bending test Bending at an angle of j O.K. (no ! O.K. (no
1
105 degrees when the crack | crack)
bending inside radius
R=12t does not give rise
to any crack larger than
3 mm on the bent surface
5.6 Shape Measuring
The plate thickness of: the completed pressure hull and its shape were
measured. An external view of the completely assembled pressure hull is
given in Photo 5 [not reproduced].
The thickness of the plate in the general area of the sphere other than its
equator ranged from 73.24 to 73.93 mm for the northern hemisphere and from
73.40 to 73.95 mm for the southern hemisphere, thereby satisfying the
specifications of plus or minus 0.45 mm.
Figures 7 and 8 give the contour maps of the northern and southern hemi-
spheres of the pressure hull, prepared based on the results of the measure-
ments of their shapes, respectively. The sphericity (local radius of
109
of curvature divided by the standard radius) at 1.004 satisfied the speci-
fications. Results of the measurements of the shape of the pressure hull
are summed up in Table 5.
No. 2 central
watertight
cable gland Conical
sheet
atch
Equator
No. 1 central
watertight
cable gland
Figure 7. Sphericity Deviation (Radius)

Distribution of N-Hemisphere
No. 3 viewport
coaming
No. 1 viewport
coaming
No. 2
viewport
coaming
Figure 8, Sphericity Deviation (Radius)

Distribution of S-Hemisphere
110
Table 5. Results of Sphericity Measurements
Results of measurements
Northern Southern
Item Specifications hemisphere hemisphere
Maximum 1074.03 Maximum 1074.06

Outside radius (mm) 1073.5 ± 1
Minimum 1073.51 Minimum 1073.49
General area plate Maximum 73.93 Maximum 73.95

73.5 ± 0.45
thickness (mm) Minimum 73.24 Minimum 73.40
Sphericity Below 1.004 1.004
6. Compressive Strength and Leakage Tests
The "Submersible Special Standards" of Japan stipulate that pressure shells

must be given compressive strength and leakage tests. The compressive
strength test is conducted to confirm that the pressure shell is free from
any abnormality when subjected to a pressure of 1.1 times the pressure
equivalent to its maximum diving depth. The leakage proof test is a test
that is conducted to confirm that no water has seeped into the pressure
hull under the above pressure condition.
Japan, however, has no facility for testing as large a pressure shell as

ours. So, after the pressure hull in question was completed, we sent it to
the United States and tested it on facilities leased from the DTRC there.
6.1 Procedures of Compressive Strength and Leakage Tests
The conditions of the compressive strength and leakage tests regarding the
composition, maximum pressure, and retention time are given in Table 6 and
the procedures of compression and decompression are given in Figure 9.
Table 6. Condition of Compressive Strength and Leakage Test
Maximum Retention
pressure time
Division (kgf/cm2) (minutes)
Compressive Preliminary pressurization test* 748 15

strength test
Compressive strength test 748 15
Leakage test 748 15
* In the preliminary pressurization test, pure water was poured into the
inside of the pressure hull (in accordance with the DTRC test procedures
and standards).
Ill
Deaeration of test tank
and overflow line
Test tank
for pouring solution

test tank and
Measuring compressing it
instruments
Figure 9. Arrangement of Pressure Test in DTRC
Key:
1. Central watertight cable gland 6. Strain gauge
2. Pipes for pouring water in and 7. Lead wire for a gauge
deaeration 8. Water block
3. Penetration piece for piping 9. Viewport
4. Penetration piece for measuring 10. Lead wire for a gauge
5. Cable for measuring 11. Feedthrough
The maximum pressure was set at 1.1 times the pressure at the pressure
hull's maximum diving depth (748 kg/cm2) and the*retention time at 15
minutes. The prevailing pressure at each of the steps shown in Figured
was retained and the strain of the pressure hull was measured.
Insulating oil was used as a medium of compression in the pressure test, and
strains at various parts of the pressure hull were measured using a strain
gauge. Pure water was used as' the medium of compression in the leakage test
to confirm the pressure hull's water tightness. In the leakage test, strains
were measured at four points on the external surface of the general section
of the pressure hull to confirm its behavior.
6.2 Facilities for Compressive Strength and Leakage' Tests
The compressive strength and leakage tests were conducted using the lO-feet-
diameter spherical test tank installed in the deep-sea simulation experiment
building of the structure division of DTRC. The two-part test tank,
112
consisting of the upper and lower tiers, is a spherical tank of a laminated
structure.
For the leakage test the pressure hull was placed inside the test tank and
attached to the jigs inside the test tank. A schematic arrangement of the
pressure test is given in Figure 10. The observation portholes and water-
tight cable gland connector receptacle used in the test were identical to
the actual products mounted in the pressure hull.
Retention for 15 minutes
.,.., .--. Time '■.■.■.■■■';.,■-
Figure 10. Procedure of Compression and Decompression
6.3 Results of the Compressive Strength and Leakage Tests
Shown in Photo 6 [not reproduced] is the pressure hull being subjected to

compressive strength and leakage testing. The relationships between
stresses and strains, as measured by a strain gauge in the compressive
strength and leakage testing, are given in Figures 11 and 12. The stress-
strain relationships followed a linear line, and there was no residual
strain observed. The results were excellent.
Comparison is made of the actual values of the strains measured in the

vicinity of the viewport coaming, and in the vicinity of the central water-
tight cable gland with their calculated values, and the results are given in
Figures 13 and 14. Although there was observed partial scattering in the
measurements, the two types of values showed good agreement. The stresses
detected by the strain gauge were all within the allowable limits. For
reference, in the leakage test, measurements were taken of the stresses at
the external surface of the general area of the pressure hull and the
measurement results were roughly in agreement with the measurement results
obtained in the pressure test, thereby revealing a good reprbduclbility.
113
1000
1000 1000
Latitudinal Longitudinal
direction 750
direction
750 750 v
Wx External , \\ External
'x surface \ ,'v\( surface \
500 & 500- Inner- \ 500
600 Inner' Q> surface "\
surface \\ U u
3 3
250 CO 260 ■ 250 : (0
CO CO
a) 
Strain <>'<>*>
Figure 11. Stress-Strain Transition Figure 12. Stress-Strain Transition

(General Fart of Hull) , (Bottom of Viewport ,
Coaming)
*/"
.—-*
a A A
•H
cfl
u
+J
QMM
m ftBI
Figure 13. Longitudinal Strain Figure 14. Latitudinal Strain

Distribution Around Distribution Around
Viewport Coaming Central Watertight
(external pressure Cable Gland (external
680 kgf/cm2) pressure 680 kgf/cm2)
Key: Key:
1. Calculated value 1. Calculated Value
2. Measured value 2. Measured value
3. Inner surface 3. Inner surface
4. External surface 4. External surface
114
While the maximum pressure was retained during the compressive strength and
leakage testing, no abnormal occurrences such as drops or fluctuations in
pressure were observed. We examined the existence or nonexistence of leaks
to the inside of the pressure hull, but to our satisfaction, no leakage was
observed.
The results of the compressive strength and leakage tests gave us confidence
that the pressure hull has ample enough compressive strength and water
tightness. After the compressive strength and leakage tests were over,
major weld joints were subjected to dye penetrant tests and the results
revealed that the welds were sound.
7. Conclusion
This paper has described the design, manufacture and testing of the 6500-
class survey submersible Shinkai 6500. A summary of its contents follows.
(1) The pressure hull of Shinkai 6500, the first large-scale thick plate
structure manufactured in Japan by using titanium alloy (Ti-6A1-4V ELI),
was excellent in terms of the material and shape accuracy.
(2) As for the properties of the base material titanium alloy, they were
proved to satisfy—more than enough—the required performance capabilities
in the structural and mechanical tests.
(3) As for welding, the electron beam welding employed enabled us to

obtain sound welds inside and outside of the surface and the joints had
strength that fully satisfied the demanded values.
(4) For geometrical accuracy, the adoption of the technique—first welding

the reinforcements for the openings to the hemisphere by electron beam
welding and then machining the entire hemisphere three-dimensionally—
enabled the sphericity of the pressure hull to obtain 1.004, an accuracy
close to a sphere.
(5) The compressive strength and leakage tests conducted under the condi-
tion of 1.1 times the maximum pressure of the maximum design pressure,
confirmed that the pressure hull fully met the demanded performance
capabilities in terms of compressive strength and water tightness.
References
1. Ryoichiro Sasahara, Tsutomu Toyohara, Katsuhito Fuchigami, and Genta

Takano, "Basic Research on Electron Beam Welding of Thick Plate
Ti-6A1-4V Alloy," MITSUBISHI JUK0 GIHO (MITSUBISHI HEAVY INDUSTRIES
TECHNICAL JOURNAL), Vol 20, No 2, 1983.
2. Norimasa Endo, Kimio Yokota, Ryoichiro Sasahara, Tsutomu Toyohara, Genta

Takano, Takashi Nishimura, and Kenichi Yasui, "Research on Engineering
Methods of Titanium Alloy Pressure Hulls for Deep Sea Research Submersi-
bles (No 1)," JOURNAL OF THE SOCIETY OF NAVAL ARCHITECTS OF JAPAN, —
Vol 156, 1984.
115
3 Norimasa Endo, Kimio Yokota, Ryoichiro Sasahara, Tsutomu Toyohara, Genta
Takano, Takashi Nishimura, and Kenichi Yasui, "Research on Engineering
Methods of Titanium Alloy Pressure Hulls for Deep Sea Research Submersi-
ble (No 2)," JOURNAL OF THE SOCIETY OF NAVAL ARCHITECTS OF JAPAN,
Vol 157, 1985.
4. Kimio Yokota, Hidemasa Morihana, Koichi Uragami, and Yutaka Yamauchi,
"Research on Collapse Strength of Spherical Shell Models of Titanium
Alloy," JOURNAL OF THE KANSAI SHIPBUILDING ASSOCIATION, Vol 198, 1985.
5. Tamotsu Shinohara, Shinichi Takagawa, Kimio Yokota, Hidemasa Morihana,

Koichi Uragami, and Yutaka Yamauchi, "Research on Collapse Strength of
Spherical Shell Models of Titanium Alloy (No 2)," JOURNAL OF THE KANSAI
SHIPBUILDING ASSOCIATION, Vol 207, 1987.
6 Masanobu Sato, Takatoshi Yamauchi, Genta Takano, Hidemasa Morihana,

Katsuhito Fuchigami, and Katsuji Miyamoto, "Properties of Titanium
Alloy for Deepsea Research Submersibles and Their Evaluations,
MITSUBISHI HEAVY INDUSTRIES TECHNICAL REPORT, Vol 15, No 6, 1978.
7. Norimasa Endo, Hidemasa Morihana, Katsuhito Fuchigami, Yutaka Yamauchi,

Tsutomu Toyohara, and Toshihiko Hiraishi, "Pressure Vessels of
Titanium Alloy for Deepsea Research Submersibles," MITSUBISHI HEAVY
INDUSTRIES TECHNICAL REPORT, Vol 17, No 5, 1980.
8. M. A. Krenzke, "The Elastic Buckling Strength of Near-Perfect Deep

Spherical Shells With Ideal Boundaries," DTMB Report 1713, 1963.
9. M. A. Krenzke and T. J. Kiernan, "Test of Stiffened and Unstiffened

Machined Spherical Shells Under External Hydrostatic Pressure, DTMB
Report 1741, 1963.
10. M. A. Krenzke and T.J. Kiernan, "The Effect of Initial Imperfections
on the Collapse Strength of Deep Spherical Shells," DTMB Report 1757,
1965.
11. Japan Marine Machinery Development Association (JMDA, "R&D Report on
6,000-Meter Research Submersible," 1971.
12. JMDA, "R&D Paper on 6,000-Meter Research Submersible," 1972.
16. Kazuhiko Kanai, Hidemasa Morihana, Toshiki Yamazaki, and Kunio Terada,
"Experiment and Research on Collapse Strength of Spherical Shells,
JOURNAL OF THE SOCIETY OF NAVAL ARCHITECTS OF JAPAN, Vol 132, 1972.
- END -
116
'.9\ZZ on 'ai3ijaMiy,::is
ay lüAoa iyod sszs
£01 SS330ad «N..U..W
SUN
.... . /:i:
■.:■:. 19 J2Z. epresent the
,r: iblication may
cite FBIS or JPRS provided they do so in a ma,,. ,*.. cis-;:, j them as the
secondary source.
Foreign Broadcast Information Service (FBIS) and Joint Publications Research Service (JPRS)
oubHcations contain political military, economic, environmental, and sociological news, commentary.
SdrtSSTÄSfcK^ as scientific and technical data and reports. All .nformation has been
obtained ram foreign adio and television broadcasts, news agency transmissions, newspapers boote
andI periodicals Items generally are processed from the first or best available sources. It should not be
kSSSXSS^Yme been disseminated only in the medium in the language, or to the area, indicated.
Items from forekjn language sources are translated; those from English-language sources are> tran-
K^c^JidSi certain diacritics. FBIS renders V«^^***^""1??^
with the romanization systems approved for U.S. Government publications by the U.S. Board of
Geographic Names.
Headlines, editorial reports, and material enclosed in brackets []| are supplied by ™S/JPRS.
Processing ndicators such as [Text] or [Excerpts] in the first line of each ite,m indicate> hovvthe
Xmation was processed from the original. Unfamiliar names rendered phonetically are enclosed in
parentheses. Words or names preceded by a question mark and enclosed in paren heses were: not clear
from the original source but have been supplied as appropriate to the context. O^r "riattributed
parenthetical notes within the body of an item originate with the source. T.mes within items are as given
by the source. Passages in boldface or italics are as published. ^
SUBSCRIPTION/PROCUREMENT INFORMATION
The FBIS DAILY REPORT contains current news provided by NTIS upon request. Subscriptions are
and information and is published Monday through available outside the United States from NTIS or
Friday in eight volumes: China. East Europe, Soviet appointed foreign dealers. New subscribers should
Union, East Asia, Near East & South Asia. Sub- expect a 30-day delay in receipt of the first issue.
Saharan Africa, Latin America, and West Europe.
Supplements to the DAILY REPORTS may also be U.S. Government offices may obtain subscrip-
available periodically and will be distributed to regular tions to the DAILY REPORTS or JPRS publications
DAILY REPORT subscribers. JPRS publications, which (hardcover or microfiche) at no charge through their
include approximately 50 regional, worldwide, and sponsoring organizations. For additional information
topical reports, generally contain less time-sensitive or assistance, call FBIS, (202) 338-6735,or write
information and are published periodically. to P.O. Box 2604, Washington, D.C. 20013.
Department of Defense consumers are required to
Current DAILY REPORTS and JPRS publications are submit requests through appropriate command val-
listed in Government Reports Announcements issued idation channels to DIA, RTS-2C, Washington, D.C.
semimonthly by the National Technical Information 20301. (Telephone: (202) 373-3771, Autovon:
Service (NTIS), 5285 Port Royal Road, Springfield, 243-3771.)
Virginia 22161 and the Monthly Catalog of U.S. Gov-
ernment Publications issued by the Superintendent of Back issues or single copies of the DAILY
Documents, U.S. Government Printing Office, Wash- REPORTS and JPRS publications are not available.
ington, D.C. 20402. Both the DAILY REPORTS and the JPRS publications
are on file for public reference at the Library of
The public may subscribe to either hardcover or Congress and at many Federal Depository Libraries.
microfiche versions of the DAILY REPORTS and JPRS Reference copies may also be seen at many public
publications through NTIS at the above address or by and university libraries throughout the United
calling (703) 487-4630. Subscription rates will be States.

JPRS Report-: Science & Technology

Uploaded by

JPRS Report-: Science & Technology

Uploaded by

JPRS-JST-91-005

BUG QUALITY INSPECTED 3

SCIENCE & TECHNOLOGY

SCIENCE & TECHNOLOGY POLICY

Japanese Government's FY90 Nuclear Power Budget Proposal

Gas Pressure-Sintered (MoSi2-SiC)/TiAl FGM

Fabrication Process, Evaluation of FGMs

Thermally Resistant FGMs by CVD Method

Developing Independent Manned Space Technologies

.RISC Processor Array for Artificial Neural Networks

Information Processing and Equipment Arrangement in the Pressure

Design and Construction of Spherical Pressure Hull of

Japanese Government's FY90 Nuclear Power Budget Proposal

90CF0359A Tokyo GENSHIRYOKU SANGYO SHIMBUN in Japanese 18 Jan 90 pp 4-5

[Text] Government Nuclear Power Budget Proposal in,Tabular Form

P.O 13 oo <t- r^ vO oo f~» oo m H

o 1-^ rH H r-- O r-~ m rH <r

o c o CU 4»i 4J >> cd cd >, CO o rl 4-1 O >,

CU ^ ■3 CU P CO U p cd P CU H <4-l CU •rl CU W> CU Cd -rl •3

60 U o£ CO u 3 1? •H <U H P r-l CJ 3 rH 4-1 rH p O

Gas Pressure-Sintered (MoSi2~SiC)/TiAl FGM

91FE0069A Tokyo KEISHA KINO ZAIRYO KENKYUKAI in Japanese 22 May 90 pp 1-5

[A series of 19 slides used by Junzo Fujioka, Kawasaki Heavy Industries Ltd.,

The Ninth Workshop, Functionally Gradient Material Research Association

Development of (MoSi2-SiC)/TiAl Functionally Gradient Material

Junzo Fujioka and Yuji Matsuzaki, Akashi Technical Research

Schematic Diagram of Gas Pressure

Illustration of Mechanisms of Gas Pressure Sintering Method

1. High-melting point materials (materials not to be readily sintered),

3. With high energy efficiency due to isotröpic pressure application,

4. By placing an ignition agent around a material to be sintered, a high-

Characteristics Required for Thermal Stress-Alleviating

High-Temperature Side _ „. .... ,, TTM Low-Temperature Side

(1) Heat-shielding (1) Thermal stress (1) Thermal conduc-

Prospective High-Temperature Side Materials for Thermal Stress-Alleviating

Heat stability: highest

Specific heat, J/g'K

•H •H 4-1 W/m-K "oo 4-i pi

1549 Kft»B**i«5orasMTr //\

C »4 (1200D '•"IMllJUMÜBU» /r\

1 ••••TIC t»a**«t /nv

': me» (in:B««*2t *S(l»Sm.lO »}S*:7T4 5'-fc?5»?X (19H.BHX)

:vn (next generation)

500 1000 1500 2000

Relationship Between Specific Strength and Temperature of Prospective Low

Production Method and Development Needs

Production method Development needs

1 Raw material powder lamination (1) Optimal gradient distribution

2 Molding and hardening (1) Selection of a binder

3 Encapsulation (1) Vacuum encapsulation technology

4 Gas pressure sintering (1) Temperature control technology

(2) Low pressure sintering technology

(3) Larger, more complex shape forma-

Design Results for (MoSi2-SiC)/TiAl-Base FGM (0-Order Design Based on

Before vacuum After vacuum

Before vacuum After vacuum

Encapsulation With Petri Dish

MoSl2 Side TlAl Side

MoSia-SiC side TiAl side

FGM 30 mm in Diameter (showing SiC-addition effect)

FGM 30 mm in Diameter (showing the effect of compositional distribution

Fll« *•. ; ll LtMC ANALYSIS

(Left graph) Compositional distribution parameter n = 2

Samples for High-Temperature Gap Tests

FGM Sample Mounted on High-Temperature Gap Test Apparatus

Average Thermal Expansion Coefficient of NFGMs

Thermal Conductivity of NFGMs (MoSi2: 10W/(m*K) at 1600 to 1.800K)

Fabrication Process, Evaluation of FGMs

90FE0317A Tokyo SHIN SOZAI 21 SEKI FORUM in Japanese 24 Apr 90

[Article written by Ryuzo Watanabe, Material Processing Department,

Functions most suitable for a material to use environments may be

Functionally gradient materials (FGMs) were born as a product of these

OOTJOÖHM ^i? ^ i n n ro >rJ '» tfl a W

Figure 2. Particle Spouting Arrangement Device for Compositional

Figure 3. Temperature Gradient-Added Sintering Device

Sintering behaviors of a blend of ceramic and metallic powders are

': me» (in:B««2t S(l»Sm.lO »}S*:7T4 5'-fc?5»?X (19H.BHX)