Full text of DOI:10.1051/metal:2002114

Development of burden distribution
control with an advanced bell-less
top at Chiba No.6 blast furnace*
T. Sato, T. Nouchi, K. Takeda, T. Kawai,
H. Kamano, N. Takashima
(Kawasaki Steel Corporation Japan)
■
INTRODUCTION
For the improvement of blast furnace operation, more steel
companies have installed the burden distribution control (1-6)
with a bell-less top which has an advantage in controllabi-
lity compared with a bell+MA top. At Kawasaki Steel
Corporation, the bell-less top for a large blast furnace was
installed at Chiba No.6 BF in 1977 (7) for the first time in
Japan. Then, the more advanced three parallel bunker bell-
less top was developed at Mizushima No.3 BF in 1990 (8).
These efforts resulted in a prolonged campaign life (9) or
in the use of a large amount of a small size sinter (10).
Recently, demands for cost reduction of hot metal have
increased and resulted in low cost operations of blast
furnaces with high PC rate, low coke rate and high-ratio
charging of low quality, low cost materials. At tough condi-
tions such as low coke rate and low agglomerate ratio
operation, a small fluctuation of ore layer thickness may
cause local widening of the cohesive zone and its dropping
to the deadman zone. Therefore, high accuracy and resis-
tance to disturbance of burden distribution control are
required more than ever. Authors have proposed ideas for a
desirable burden layer profile and required functions for the
charging system (11) in order to correspond to hard opera-
tions such as mentioned above. These ideas were realized
first with reduced scale model experiments and installed at
the revamping of Chiba No.6 BF for the 2nd campaign in
A new bell-less top was installed at Chiba n° 6
blast furnace in 1998. It was designed in view of an
advanced burden distribution control method. Its
main features are a rotating chute with a stabilizer
and the possibility of reverse and forward tilting.
Reduced scale model experiments have been used
to develop this system. An accurate simulator was
also developed for burden distribution. The results
of simulation have been used to implement
operational charging conditions making use of
reverse tilting. The test operation confirmed that
burden profile and gas distribution could be
controlled satisfactorily with this method.
1
998. The new bell-less top is named SMART (Strate-
gically Multi-functional and Accurate bell-less Top). In this
report, details of the development of a new bell-less top
with scale model experiments and results of the filling tests
are described as a confirmation of several new functions,
then operational charging trials are presented.
■
DEVELOPMENT OF THE NEW
BELL-LESS TOP
Concepts of the new bell-less top
The concept of innovative burden distribution control
which provides high controllability and stability at hard
operations such as low coke rate or high-ratio charging of
small size sinter is shown in figure 1. Required functions
are :
*
Subject of a presentation at the 4th European Cokemaking and Iron-
making Conference, Paris 2000 (June, 2000). Also published in the
conference proceedings, Éditions de la Revue de Métallurgie, vol. 1,
p. 308-314.
– multi-batch charging with more than three parallel
bunkers,
©
La Revue de Métallurgie 2002
La Revue de Métallurgie-CIT
Mars 2002
231

Contrôle de la répartition des
charges au haut-fourneau n° 6 de
Chiba équipé d'un nouveau type
de gueulard sans cloche
T. Sato, T. Nouchi, K. Takeda, T. Kawai,
H. Kamano, N. Takashima
(Kawasaki Steel Corporation Japan)
Depuis 1997, Kawasaki Steel équipe ses hauts-fourneaux
de gueulards sans cloche pour améliorer la répartition
des charges. Divers perfectionnements ont été apportés à
cette technologie, notamment à l'occasion de la marche à
fort taux d'injection de charbon et de l'utilisation
d'agglomérés plus fins et plus économiques.
déposer les matières en quantité suffisante à la périphérie
du fourneau grâce à l'augmentation de la longueur de la
goulotte (de 0,2 m seulement), de son angle d'inclinaison
maximal et à la maîtrise de la trajectoire de chute.
Compte tenu de la modification du système de
chargement, il a fallu améliorer le simulateur existant de
répartition de la charge : prise en compte des nouveaux
types de trajectoires de chute des matières et du
stabilisateur de dépôt. La trajectoire de chute unique du
gueulard classique est remplacée par la superposition de
trois trajectoires ayant des composantes horizontales de
vitesse différentes et interagissant entre elles. Les essais
sur maquette ont permis de caler le modèle.
Mise au point d'un nouveau type
de gueulard sans cloche
Avec les conditions de marche très dures du HF6 de
Chiba, le système de chargement des matières au
gueulard doit disposer d'un nombre suffisant de silos en
parallèle, pouvoir s'adapter aux changements de
granulométrie des matières, réaliser des conditions
stables de dépôt à l'extrémité de la goulotte, et permettre
d'obtenir sans risque d'éboulement un profil plat des
couches déposées. Cette dernière condition impose de
pouvoir faire varier l'inclinaison de la goulotte à la fois de
manière classique (de la périphérie vers le centre) et de
manière inverse. La méthode de chargement que cela
nécessite a été mise au point à l'aide d'une maquette de
gueulard à l'échelle 1/18.
Confirmation des résultats de simulation
au moyen de mesures lors du remplissage
du HF6 après réfection
Lors de la réfection du HF6 en 1998, ce haut-fourneau a
été équipé d'un dispositif de chargement s'inspirant de
celui des essais sur maquette. Lors du remplissage du
haut-fourneau à son redémarrage, on a prélevé des
échantillons à chaque rotation de la goulotte et vérifié les
deux types de ségrégation granulométrique réalisables à
la sortie du silo de chargement. La relation entre
l'inclinaison de la goulotte et la position radiale du point
de chute des matières est conforme aux prévisions de la
simulation. Le profil des couches de coke et de minerai a
été suivi et comparé au profil calculé. Quel que soit le
mode de chargement (classique ou inversé) les prédictions
du modèle sont vérifiées. On constate de plus que les
conditions d'éboulement de la couche de coke sous
l'impact du chargement de minerai sont différentes suivant
le mode de gestion de la goulotte : avec le chargement
inversé, la couche de coke sous-jacente est beaucoup
moins affectée. La ségrégation granulométrique radiale
des couches déposées est conforme aux prédictions du
simulateur.
Pour obtenir un profil de couche peu incliné de la
périphérie vers le centre, il faut déposer plus de matière
au centre qu'à la périphérie à chaque rotation de la
goulotte. Le chargement avec modification réversible de
l'inclinaison de la goulotte permet d'obtenir le profil de
couche désiré, même en cas de fluctuations de la quantité
et de la qualité de la charge ou de la durée totale de
chargement. Les essais sur maquette confirment la
validité de cette approche et la possibilité de prévoir avec
précision les résultats obtenus au moyen du modèle de
simulation développé parallèlement. Pour obtenir une
marche centrale du haut-fourneau, les conditions de
ségrégation granulométrique des matières à la sortie du
silo de chargement doivent être différentes de celles du
mode de chargement classique : les particules les plus
grosses doivent être déchargées avant les particules les
plus fines.
Essai sur haut-fourneau en marche
Afin de mieux contrôler la régularité de chute des
matières, l'extrémité de la goulotte a été équipée d'un
dispositif de stabilisation. Ceci permet d'éviter de
rallonger la goulotte, ce qui conduirait à des trajectoires
de chute trop proches de l'horizontale. Il s'agit également
d'éviter de réduire la largeur de la zone de dépôt, afin de
ne pas compromettre la stabilité de la couche déposée. La
nouvelle méthode de chargement permet effectivement de
Après un début de marche avec chargement de couches
alternées coke-mine, la séquence définitive CCMM a été
adoptée. Le simulateur a permis de définir la manière de
passer du mode de chargement classique au mode inversé.
Pour conserver la même répartition gazeuse, des
changements importants dans le schéma de chargement
sont nécessaires. Les résultats de la simulation ont été
utilisés pour faire un essai dans les conditions ainsi
232
La Revue de Métallurgie-CIT
Mars 2002

déterminées. Le profil radial d'utilisation du pouvoir
réducteur du gaz dans la partie supérieure de la cuve a
été trouvé inchangé par rapport à la marche dans les
conditions classiques.
seconde phase, l'addition en deux étapes du minerai est
faite de manière à favoriser une marche périphérique : le
minerai est d'abord déposé vers la périphérie, puis vers le
milieu. Lors de ces modifications du profil de chargement,
la marche du haut-fourneau est restée stable et le profil
des couches a pu être bien maîtrisé avec une faible pente.
La nouvelle méthode de gestion de l'inclinaison de la
goulotte permet donc d'agir de manière efficace sur la
répartition gazeuse et le profil des couches. Elle est à la
base de la marche économique du haut-fourneau à fort
taux d'injection de charbon, à faible mise au mille de coke
et avec une forte proportion d'aggloméré de seconde
qualité dans le chargement.
D'autres essais avec la nouvelle méthode de gestion de
l'inclinaison de la goulotte ont été conduits afin de
déterminer dans quelle mesure on pourrait agir sur le
profil des couches et la répartition gazeuse. L'objectif
était de passer d’un profil de coke à courtes terrasses à un
profil à plus faibles inclinaisons. Une méthode de
chargement en deux phases a été utilisée pour cela, en
accord avec les résultats de la simulation. Dans une
première phase on diminue la pente du lit de coke, ce qui
favorise une répartition gazeuse centrale stable. Dans la
–
–
two modes of tilting direction of the rotating chute, either
peripheral (conventional tilting) or central (reverse
tilting), that provide more flexibility about the formation
of layer profile compared with conventional charging,
formation of very stable conditions against a collapse
phenomenon of the burden layer with low surface angle
or long terrace profile.
Experimental apparatus
An outline of the experimental apparatus is shown in figure 2.
To reproduce the burden transport process of an actual
plant, ore bin surge hopper, charging system including top
bunker and belt conveyer linked together are installed. The
reduced scale of the apparatus is 1/18 and crushed coke and
sinter with a size adjusted to the scale are used as charging
materials.
Fig. 1 – Concept of the new charging system.
Fig. 1 – Principe du nouveau système de chargement.
–
–
control of particle size changes during discharging from
the top bunker by controlling the segregation conditions
in the off-centre type top bunker,
installation of a stabilizer on the end of the rotating chute
that provides adequate width and trajectory of burden
from the chute,
Fig. 2 – Experimental apparatus.
Fig. 2 – Maquette de chargement.
La Revue de Métallurgie-CIT
Mars 2002
233

In addition to this, adequate central gas flow is necessary
for the stable operation of the blast furnace and fine
particles should not be charged at the central part. To
charge coarse particles at the central region in spite of the
tilting direction of the rotating chute, a specific control
technique of particle size changes during charging is requi-
red. An off-centre type top bunker with the type of segre-
gation control plate installed at Mizushima No. 3 BF (8)
has as characteristic that fine particles are discharged at the
early stage and coarse particles at the later. This has an
advantage for the stable central gas flow at conventional
tilting. On the basis of the control concept of segregation in
the top bunker, a new top bunker was designed to provide
two modes of particle size change, one is the discharge of
fine particles at first and the other of coarse particles at first.
Results of particle size changes during discharge from the
top bunker at the scale model are shown in figure 5.
Reverse tilting charging method
To form a low angle surface profile which provides
maximum stability of the layer, the amounts of charging at
the central to middle region have to be larger than with
conventional charging. An example of low angle surface
profile is shown in figure 3 with each layer profile dumped
at one rotation of chute calculated by the burden distribu-
tion simulator (l2). The layer thickness of burden dumped
in the central part at one rotation is larger than that in the
peripheral part, since the quantity of burden dumped at
each rotation is nearly equal during charging. If the forma-
tion of a low angle profile as shown in figure 3 is carried out
with conventional tilting, strict end point control of the
charging time is required to avoid a huge fluctuation of
layer profile due to the changes of burden quantity and
quality. On the other hand, reverse tilting charging guaran-
tees a designated quantity of burden at the central part in
spite of the fluctuation of total charging time, and provides
good reproducibility of low angle layer profile. The result
of a scale model trial to form a low angle layer profile with
reverse tilting is shown in figure 4. The identical profile
calculated by the simulator is materialized, so the advan-
tage of reverse tilting for the formation of a low angle
profile is confirmed.
Fig. 5 – Changes in particle size
discharged from the bunker.
Fig. 5 – Évolution de la granulométrie des matières
en cours de chargement.
Comparaison de deux schémas de chargement.
Development of a stabilizer
Fig. 3 – Flat profile (calculation).
Fig. 3 – Obtention d'un profil de couche plat (simulation).
At this revamping, the throat diameter was enlarged by
0
.9 m, and the ability of a sufficient charge at the wall
region is a required function for the new charging system.
Extension of the rotating chute length enables the burden to
fall further, but it turns the falling trajectory in a horizontal
direction because of the increased horizontal velocity of the
falling burden. A typical disturbance for burden distribution
control is the fluctuation of filling level due to the slip or
non-uniformity of burden descent. As the falling trajectory
turns in a horizontal direction, a fluctuation of the filling
level affects the radial falling position of the burden and the
resultant gas distribution changes remarkably. Conse-
quently, the new charging system must provide sufficient
charge at the wall region without turning the falling trajec-
tory in a horizontal direction. An excessive decrease in the
falling width of the burden reduces the stability of the depo-
sited burden (11) because of an increase in the falling
energy per unit area. Therefore falling width should not
Fig. 4 – Flat profile (model experiment).
Fig. 4 – Vérification sur maquette des conditions d'obtention
d'un profil de couche plat.
234
La Revue de Métallurgie-CIT
Mars 2002

HAUT-FOURNEAU
Development of the burden
distribution simulator
Corresponding to the renewal of the charging system, a new
burden distribution simulator was developed based on the
existing one (l2, 13). The main improvements of the simu-
lator were the treatment of the falling trajectory of burden
from the rotating chute associated with the installation of
the stabilizer. As shown in figure 8, while the falling beha-
viour of burden is treated as one main flow for the conven-
tional simulator, the new one treats the three falling
trajectories from the chute with consideration of their inter-
actions at the exit of the stabilizer. These flows combine
into one in the main flow. The interaction between each
flow is considered by introducing an interaction coefficient.
The coefficient applies to the horizontal component of the
falling velocity of the flow located at the lowest position
Fig. 6 – Falling trajectory of the main flow.
Fig. 6 – Trajectoire de chute des matières pour différentes
configurations de goulotte.
Fig. 7 – Radial distribution of burden
from the rotating chute.
Fig. 8 – Treatment of burden flow on simulator.
Fig. 7 – Distribution radiale de la charge déposée
par la goulotte rotative.
Fig. 8 – Étude de l'écoulement des matières en sortie de goulotte
à l'aide du simulateur.
differ from that of the conventional charging system. To
meet the demands mentioned above, a special device for
stabilization of burden flow (stabilizer) at the end of the
rotating chute was developed, and its effects were verified
by scale model experiments. Falling trajectories of burden
(
locus of the most concentrated point of particles in the
falling width at each level) from the rotating chute for the
conventional and new charging systems are shown in
figure 6. The new charging system provides full charge at
the wall region because of an increase in the maximum
tilting angle of chute and in the length of chute by 0.2 m,
and invariable falling trajectory compared with the conven-
tional system owing to the stabilizer. For comparison, the
case with an extension of the length of the rotating chute by
1
.0 m without stabilizer is shown: the extension of the chute
length only turns the falling trajectory in a horizontal direc-
tion. As shown in figure 7, the falling width of burden with
the new charging system is identical to that of the conven-
tional system.
FIG. 9 – Falling trajectories of burden.
Fig. 9 – Exemples de trajectoires calculées et observées.
La Revue de Métallurgie-CIT
Mars 2002
235

(flow 3 in Fig. 8), and is adjusted to fit the measured falling
trajectories at the scale model (Fig. 9). Calculations of layer
profile and percolation of particles are treated with the
same methods as in the conventional model. Changes in
particle size during charging are modeled based on experi-
mental results shown in figure 5. The accuracy of the simu-
lator was confirmed during the filling test.
■
CONFIRMATION OF THE NEW
FUNCTIONS (FILLING TEST)
A new charging device developed on the basis of the scale
model experiment was installed during the revamping of
Chiba No.6 BF in 1998 (14). Performance of the new func-
tions and accuracy of the simulator were confirmed at the
filling test just before blow-in.
Fig. 11 – Relation between chute position
and falling position.
Fig. 11 – Relation entre l’orientation de la goulotte et
la position de la zone d'impact du jet de matières.
Change in particle size during discharge
from top bunker
Layer profile and particle size
distribution
Changes in particle sizes during discharge from the top
bunker were measured at the filling test. Samples of burden
were taken at each turn during the charge by use of a collec-
tor installed at the level of a gas temperature probe at the
throat. Results are shown in figure 10 and the ability of the
function which provides two ways of particle size changes
during discharge is confirmed.
Surface profile of coke and ore after charging are measu-
red and compared with calculated results by simulator. As
shown in figure 12, the surface profile of burden is predic-
ted accurately by the simulator not only for conventional
tilting charging but also for reverse tilting. On the other
hand, the coke layer collapse phenomenon (15) as a result
of the impact of charged ore shows a different behaviour
between conventional tilting charging and reverse tilting.
In addition to the surface profile of coke and ore, the
subsurface profile of coke after ore charging measured by
the electric sensor is shown in figure 13. While typical
coke collapse occurs for conventional tilting charging,
reverse tilting charging provides very small coke collapse.
Fig. 10 – Changes in size of ore discharged
from the top bunker.
Fig. 10 – Évolution au cours du chargement de la granulométrie
du minerai extrait du silo de gueulard.
Falling position of burden
A relationship between the tilting position (tilting angle) of
the rotating chute and the radial falling position of burden
was investigated by use of a measuring device installed at
the level of a gas temperature probe. The main falling
points with the obtained burden flow are determined from
the most frequently impacted position on the measuring
rod, and show good agreement with calculated results by
simulator as shown in figure 11.
Fig. 12 – Comparison of measured and
predicted burden profile.
Fig. 12 – Comparaison des profils de couches
mesurés et calculés.
236
La Revue de Métallurgie-CIT
Mars 2002

HAUT-FOURNEAU
Operational charging test
After the ultra short revamping period (16), Chiba No.6 BF
was blown-in for its 2nd campaign on May 26, 1998. At the
rating-up operation after the blow-in (17), the conventional
charging method of two batch (CO) charging was practiced.
Since the 17th day after blow-in, four batch (CCOO) char-
ging making use of three parallel bunkers was started, and
this is the basis for burden distribution control ever since.
Furthermore, operational trial with reverse tilting was
performed.
Trial for the change to reverse tilting (18)
For a smooth transition from conventional charging, a
suitable charging pattern of reverse tilting, which provides
identical gas flow distribution in radial direction as conven-
tional tilting, was planned by the simulator. The predicted
results of the heat flow ratio distribution, corresponding to
the gas flow distribution and calculated from the layer
thickness and particle size distribution, for conventional
and reverse tilting are shown in figure 15 with their char-
ging patterns. Drastic changes of the charging pattern are
required to maintain the heat flow ratio, and the tilting
direction change was tested on the B.F. under the condi-
tions determined with the simulation. Figure 16 shows
typical actual measurements of the gas utilization ratio
distribution in the radial direction at the upper shaft level
for the conventional tilting and during the trial. Variation of
gas distribution with changes of charging patterns hardly
Fig. 13 – Subsurface coke profile
after ore charging.
Fig. 13 – Profil de la sous-couche de coke
après dépôt d'une couche de minerai.
A reason for the no-collapse phenomenon for the reverse
tilting charging is estimated as follows:
Coke collapse generally occurs in the deposited coke
layer at the middle or peripheral region. The coke layer
remains stable during the next ore charging because the
central region is already covered by deposited ore at the
first stage of charging and there is no chance of coke layer
collapse.
As shown in figure 14, measured particle size distribution
in radial direction corresponds to the prediction by simu-
lator.
Fig. 15 – Reverse tilting test
planned by simulator.
Fig. 14 – Radial distribution of particle diameter.
Fig. 14 – Distribution radiale de la granulométrie
des matières chargées.
Fig. 15 – Mise au point sur simulateur
du schéma de chargement inversé.
La Revue de Métallurgie-CIT
Mars 2002
237

was controlled to promote peripheral gas flow. An initial
state where O1 is charged at the middle region and O2 at
the peripheral region is changed to a final state where O1 is
charged from the middle to the peripheral region broadly
and O2 at the middle region. As a result of this control, a
change of ore layer profile and an increase of peripheral gas
flow were achieved (period B in Fig.17). During these
trials, stable operation was maintained and controllability
and stability at low angle profile with reverse tilting were
confirmed.
■
CONCLUSIONS
Fig. 16 – Result of the reverse tilting charging test.
Fig. 16 – Résultats de l’essai de marche
avec schéma de chargement inversé.
At revamping of Chiba No.6 BF for the 2nd campaign, a
new bell-less top was installed to provide high stability and
controllability of burden distribution control. The new
device has the following features:
appears, as predicted by the simulation. This confirms with
reverse tilting the efficiences of the simulation to describe
the operation.
–
–
three parallel top bunkers for multi-batch charging,
control of particle size changes during charging with a
control of segregation conditions in the off-centre type
bunker,
Gas distribution control with
low angle profile (19)
–
installation of a stabilizer at the end of the rotating chute
that provides adequate width and trajectory of burden
from the chute,
Further trials of reverse tilting in view of the confirmation
of controllability of layer profile and gas distribution were
performed. Targets of the trial were a change of coke layer
profile from short terrace profile to lower angle and promo-
tion of central or peripheral gas flow. Two stages of char-
ging pattern change were implemented in accordance with
the plans by simulator. At first, the coke layer profile was
controlled to decrease its angle; as a result, stable and
promoted central gas flow is achieved (period A in Fig. 17).
Next, the charging pattern of ore (divided into O1 and O2)
– two modes of tilting direction of the rotating chute,
conventional tilting and reverse tilting.
At the filling test, the functions of the new device and the
accuracy of the simulator developed are confirmed. The
charging trial has proved high controllability for gas distri-
bution and layer profile with reverse tilting. The new func-
tions of the bell-less top will be utilized for low cost
operations with high PC rate, low coke rate and high-ratio
charging of low quality material.
Fig. 15 – Reverse tilting test planned by simulator.
Fig. 15 – Mise au point sur simulateur du schéma de chargement inversé.
238
La Revue de Métallurgie-CIT
Mars 2002

HAUT-FOURNEAU
■
REFERENCES
(
(
(
1) OCHIFUJI (Y.), MATSUI (Y.), SIMIZU (M.), KITA-
YAMA (S.), ISIWAKI (S.) – CAMP-ISIJ, 9 (1996), p. 826.
2) IWATSUKI (K.), SIODA, (H.), KAWAOKA (K.), ABE (T.)
and YANO (M.) – CAMP-ISIJ, 2 (1989), p. 996.
3) ISHIHARA (I.), KAMADA (H.), KAKIUCHI (K.),
UENO (H.), MATSUAKI (S.), YADOMARU (N.) –
CAMP-ISIJ, 10 (1997), p. 882.
(4) ICHIDA (M.), TANAKA (T.), KOIZUMI (F.) – CAMP-
ISIJ, 12 (1999), p. 624.
(5) SHIMOMURA (A.), HATTORI (M.), LINO (B.), KIMURA
(K.), WAKIKA (S.), TUKIJI (H.) – Tetsu-to Hagane, 78
(1992), T 41.
(
6) YAMAMOTO (K.), KISHIMOTO (S.), MAKI (A.), INOUE
H.), TOMIOKA (K.) – CAMP-ISIJ, 6 (1993), p. 107.
(
(7) OKUMURA (K.), KAWAI (T.), MARUSHIMA (H.),
TAKAHASHI (H.), KURIHARA (J.) - Tetsu-to Hagane, 66
(1980), p. 1956.
(
8) SAWADA (T.), UETANI (T.), TANIYOSHI (S.), MIYA-
GAWA (S.), SUGAWARA (H.), YAMAZAKI (S.) – Tetsu
to Hagane, 78 (1992), p. 1337.
(
(
(
9) HIROSE (S.), MATSUO (N.), KOKUBU (H.) – CAMP-
ISIJ, 7 (1994), p. 157
10) KOBAYASHI (K.), MATSUMOTO (T.), YANAGISAWA
(K.) – Kawasaki Steel Technical Report, 32 (1995) 4, p. 9.
11) NOUCHI (T.), TAKEDA (K.), ITAYA (H.) – The 54th
committee (ironmaking), the Japan Society for the
Promotion of Science (JSPS), Report No. 2121.
(
(
(
12) SATO (T.), NOUCHI (T.), KIGUCHI (M.) – Kawasaki Steel
Technical Report, 38 (1998), p. 24.
13) NOUCHI (T.), SATO (T.), MIYAGAWA (S.), TAKEDA (K.)
and ITAYA (H.) – CAMP-ISIJ, 7 (1994), p. 1004.
14) MORIKAWA (Y.), KAMANO (H.), KAWAI (T.), MARU-
SHIMA (H.), MATSUMOTO (T.) – CAMP-ISIJ, 11 (1998),
p. 896.
(15) OKUNO (Y.), ISOYAMA (T.), IRIDA (T.), MATSUOKA
(H.) – Tetsu-to-Hagane, 68 (1982), p. 37
(16) FUJITA (M.), KOJIMA (H.), MARUSHIMA (H.), MAT-
SUMOTO (Y.), YOKOKAWA (A.) – CAMP- ISIJ, 11
(l998), p. 898.
(
17) SHIOZAWA (T.), NOMURA (M.), MATSUMOTO (T.),
GOTO (S.), TAKASHIMA (H.), MORIKAWA (Y.),
KAMANO (H.), KAWAI (T.) – CAMP-ISIJ, 11 (1998), p.
8
99.
(18) SATO (T.), NOUCHI (T.), TAKEDA (K.), KAWAI (T.),
NOMURA (M.), NISHIMURA (N.) – CAMP-ISIJ, 11
(1998), p. 897.
(
19) SATO (T.), NOUCHI (T.), TAKEDA (K.), KAMANO (H.),
GOTO (S.), TAKASHIMA (N.), NOMURA (M.) – CAMP-
ISIJ, 12 (1989), p. 128.
La Revue de Métallurgie-CIT
Mars 2002
239