Acciaio Memorie Microalloyed steels for high

Transcription

Memorie
Acciaio
Microalloyed steels for high-strength forgings
A. J. DeArdo, C. I. Garcia, M. Hua
In the past thirty-five years, two families of microalloyed (MA) steels have been developed for high strength bar
and forging applications. The first family was introduced in 1974 and represented the medium carbon steels
to which were added small amounts of niobium or vanadium. These early medium carbon contents steels
exhibited pearlite-ferrite microstructures and showed good strength and high-cycle fatigue resistance.
About 15 years later, microalloyed multiphase steels were introduced, which had microstructures comprised
of mixtures of ferrite, bainite, martensite, and retained austenite, depending on the composition
and processing. These steels were capable of reaching very high strengths, with good fatigue resistance
and high fracture resistance. Prior to the early 1970s, high strength forgings could be obtained only by final
heat treatment, involving reheating, quenching and tempering (QT). It has been shown repeatedly that the air
cooled forgings made from MA pearlite-ferrite steels can exhibit strengths and fatigue resistances similar
to those of the more expensive heat treated forgings. This paper will follow the development
of the microalloyed pearlite-ferrite steels over the past 35 years.
KEYWORDS:
microalloyed forging steels, high strength, microstructures, pearlite-ferrite steels, heat treatment
INTRODUCTION
Today, there are essentially three methods of manufacturing
high strength forgings, i.e., forgings with yield strengths in excess of ~ 600MPa and UTS levels above ~ 850 MPa.: (i) heat
treated low alloy steels, (ii) microalloyed medium carbon steels
and (iii) microalloyed multi-phase steels. This paper will focus
on the second group.
As it is well-known that producing forgings using the QT heat
treatment process is inefficient, expensive and deleterious to the
environment, alternative routes to high strength forgings have
been studied for decades. From an engineering perspective, having high strengths in forgings is very attractive, since this enables the possibilities of higher static and dynamic loads, smaller
components, and particularly in rotating parts, improved high
cycle fatigue resistance. From about the mid-1950s in the USA,
the flat rolled steel industry has shown that the small addition
of elements such as niobium and vanadium, which together with
Ti define the microalloying elelments, to simple C-Mn-Si steels
could impressively increase the strength when the steels were
rolled and cooled correctly [1]. In 1974, this same approach was
taken in Germany, where Nb and V were added in small amounts
to the traditional steel Ck 45, a medium carbon C-Mn-Si steel
[2], again, with impressive increases in strength.
In the medium carbon steels intended for forging applications, V
is normally preferred over Nb because of the solubility behavior
which permits the dissolution of VCN particles at lower reheat
temperatures. The strengthening effect of V can be further improved when used with higher N levels. In today’s technology, N
in BOF steels is in the range of 40-60ppm, while in EAF steels it
Anthony J. DeArdo,* C. Isaac Garcia, and M. Hua
Basic Metals Processing Research Institute
Department of Mechanical Engineering and Materials Science
University of Pittsburgh, Pittsburgh, PA 15261-2284, USA
*also, Finland Distinguished Professor,
Department of Mechanical Engineering
University of Oulu, Finland
La Metallurgia Italiana - n. 9/2010
is normally in excess of 90 ppm. Even higher strengths can be
achieved by using higher N levels in the range of 150-200ppm.
As mentioned above, the choice of V is mainly based on the ease
with which the addition can be dissolved into the austenite during reheating. The lower billet reheating temperatures means
lower production costs on the forging shop floor, more uniform
properties and lower straightening costs. This paper will review
some of the physical and mechanical metallurgy of these V-bearing medium carbon steels [1].
The early development of medium carbon, microalloyed forging
steels was essentially summarized in three papers by Frodl, et al.
in 1974 [2], Von den Steinen et al., in 1975 [3], and Brandis and Engineer in 1980 [4]. These authors presented the attributes and
benefits of adding small additions of Nb, V, Nb + V and V-N to steels containing approximately 0.3-0.5 wt% C, with Ck45 being the
usual German reference steel. They showed that the yield and
tensile strengths could be substantially increased through these
additions in air cooled forgings. They also showed that this strengthening could be further enhanced by slightly increasing the
cooling rate by using compressed air. From the work of Von den
Steinen, et al., Figure 1 shows the influence of V and Nb additions
to steel containing 0.51% C - 0.27Mn – 0.010N [3]. These steels
were first reheated to 1250°C, forged, then air cooled. They were
then reheated a second time to various temperatures, then cooled to RT either in still air or by compressed air. It is clear from Fig.
1 that the 0.1%V addition resulted in an increase in YS of ~33%
(from 450-600MPa) and in UTS of ~10% (from 820-900MPa) after
reheating at 1100°C and air cooling. These increments improved
to 44% (from 480 to 690MPa) and 17% (from 860 to 1010MPa), respectively, after compressed air cooling. It should be noted in Figure 1, that the increment in YS did not vary very much for the V
steel with changes in reheat temperature, again, a benefit to consistency. This early work showed that YS levels near 600MPa and
UTS levels near 900MPa were easily attainable in air cooled, MA
medium carbon steels containing 0.1% V, with 49MnVS3 being
the early successful candidate steel. These strengths compare
well with those found in QT steels containing 0.45 %C, i.e., YS values near ~ 600 and UTS levels ~ 850MPa.
5
Memorie
FIG. 1
Strength characteristics as
a function of annealing
temperature. Initial state:
forged (S) at 1250°C to 20
mm diameter bars.[3]
Resistenza a rottura e a
snervamento in funzione della
temperatura di ricottura. Stato
iniziale: forgiato (S) a 1250°C
in barre con diametro di
20mm.[3]
ALLOY DESIGN, PROCESSING AND STRENGTH
In the early days of the development of these steels, the major
objective was obtaining high strength levels. In medium carbon
steels containing V or V-N, the overall strength is controlled by
the amount and strength of both the pearlite and ferrite, as described by Gladman in Equation 1 for the yield strength [5].
Equation 1 shows the main contributors to YS.
(1)
Where fα is volume fraction ferrite, d is the ferrite grain size and
Sp is the pearlite interlamellar spacing. A typical microstructure of a medium carbon MA steel (10V45), air cooled from near
1200°C, is shown in Figure 2 [6]. It can be observed that the
proeutectoid ferrite decorates the prior austenite grain boundaries, and this is followed by the pearlite reaction that completes
the transformation during air cooling. Notice that the structure
is comprised of about 20% ferrite and 80% pearlite, in this case.
Hence, these steels are often referred to as pearlite-ferrite steels.
FIG. 2
Ferrite-pearlite optical microstructure exhibited by
the as-received 1045Vsteel. Transverse section.[6]
Microstruttura ferritico-perlitica di un acciaio 1045V.
Sezione trasersale.[6]
6
Since the ferrite is known to nucleate on the prior austenite
boundaries, both the austenite grain size and the cooling rate
will control the amount of ferrite and, therefore, also the amount
of pearlite. Low ferrite contents are associated with large austenite grain sizes and higher cooling rates. Lower ferrite contents also generally mean having higher strength levels, as well.
Factors such as bulk carbon and Mn contents obviously also contribute to the final microstructure and properties.
These early papers clearly pointed out the importance of billet
reheating temperature. An example is shown in Figure 3 [2].
In Figure 3 it can be seen that high reheat temperatures generally led to high UTS levels for all but the steel containing V
alone. In the case of the V steel, there was little change in the
amount of ferrite, pearlite or the UTS with reheating temperatures from 1150 to 1250°C. This fact is of course very important in producing forgings with uniform and consistent
properties.
As discussed earlier, the transformation temperature in these
steels is important, hence, the cooling rate in the temperature
range 800-500°C is also significant. This is shown in Fig. 4 [2],
which are CCT diagrams for Ck45 with and without 0.1%V. Figure 4 indicates how the amount of ferrite and the overall hardness varies with cooling rate; the overall higher hardness of
the V-steel compared to Ck45, on the order of 20%, is substantial
at a cooling rate of 1°C/sec, typical of air cooling.,
As shown by Van den Steinen et al., in Figure 1, air cooled V steel
shows a YS near 600MPa while a similar carbon steel shows
450MPa. The cause of this extra 150 MPa in YS is very interesting and the debate continues through the present. Clearly some
of this extra hardening can be attributed to precipitation hardening of both the proeutectoid and pearlitic ferrite by VCN in the
V steel. This view is supported by extraction replica TEM data of
Figs 5 and 6 [7] for V steels. Later work showed this effect was
even stronger in the presence of V and N, Figure 7 [6]. It should
be noted that data of the kind shown in Figures 5 - 7 must be interpreted with caution since the strength does not correlate with
the extracted precipitation, but rather to the elements found in
the extracted liquid. This means that the V and N found in the liquid may have been very small precipitates (causing precipitation hardening) but dissolved during extraction, or solute (causing
substructure strengthening), or as a combination of both.
Acciaio
FIG. 3
% Ferrite UTS and Ferrite Grain Size –
vs. Reheating Temperature. [2]
Percentuale di ferrite e dimensione dei
grani ferritici in acciai F-P vs. Rm.[2]
FIG. 4
% Ferrite in F-P Steels and VHN –vsdT/dt. [2]
Percentuale di ferrite in acciai F-P e
durezza Vickers vs. dT/dt. [2]
FIG. 5
Increment in the 0.2%-proof stress,
dissolved and precipitated amount of V
and Nb. [7]
Incremento di Rp0,2 in funzione della
quantità di V e Nb disciolti e precipitati.[7]
7
Memorie
Fig. 7
FIG. 6
Precipitation hardening effect of V and Nb. Austenizing: 8001300°C, 0.5hr air cool (* 0.11%V ~ Compressed air cool). [7]
Effetto di V e Nb sull’indurimento per precipitazione.
Austenizzazione: 800-1300°C, 0.5h raffreddamento ad aria
(* 0.11%V ~ Raffreddamento ad aria compressa). [7]
On the other hand, Von den Steinen et al. [3], clearly showed the
presence of fine VCN in all of the ferrite in air cooled V steels.
It is important to note that in a steel exhibiting 25% ferrite and
75% pearlite, nearly 90% of the final structure is ferrite, either
proeutectoid or pearlitic. An example of this wide spread fine
precipitation of VCN precipitates in virtually all of the ferrite is
shown in Figure 8. With a reheating temperature of 1200°C in
a cooled 20 mm diameter bar, the V addition had little effect on
the amount of pearlite, but still increased the YS by about 150
FIG. 8
The effect of precipitation hardening by
V and N on Yield strength [4,1]. All data
have been normalized, using Eq. (1), to
0.3Si, fp=0.8 and a section size of 20
mm Φ. All Steels contain 0.7Mn and
were austenized at 1200-1250°C [6]
Effetto dell’indurimento per precipitazione
da V e Nb su Rm [4,1]. Tutti i dati sono stati
normalizzati, usando l’ Eq. (1), a 0.3Si,
fp=0.8 e con una seconda dimensione di
sezione di 20 mm Φ. Tutti gli acciai
contenevano 0.7Mn ed erano austenizzati a
1200-1250°C [6]
MPa for still air cooling and over 300 MPA for compressed air
cooling. The amount of these increases that are due to precipitation hardening or to substructure strengthening is unknown.
In some later work, Brandis and Engineer showed the importance of nitrogen to the effectiveness of V [4], an observation repeated several times over the past 30 years, Figure 7 [3].
FATIGUE RESISTANCE
It is well-known that the high cycle fatigue resistance scales
TEM of extraction replica showing VCN precipitates in both proeutectoid and pearlitic ferrite. Steel Ck45 + 0.1%V, air
cooled from forging temperature 1250˚C, 50mm. [3]
Immagini TEM di repliche di estrazione che mostrano precipitati VCN sia in ferrite proeutettoide che in ferrite perlitica. Acciaio
Ck45 + 0.1% V, raffreddamento in aria a partire dalla temperatura di forgiatura di 1250°C, dimensioni 50 mm.
8
Acciaio
FIG. 9
Rotating bend test of notched samples from
steels with 0.45%C. (Hot rolled to 50mm round
bar from 1250°C. [7]
Prova di fatica a flessione rotante di provini intagliati
di acciaio con 0.45%C. (Laminato a caldo in barra
tonda da 50 mm da 1250°C). [7]
FIG. 10
Tensile strength and impact toughness for the 13
experimental steels. [6]
Resistenza a trazione e resilienza per i 13 acciai
sperimentali. [6]
with the UTS, where the fatigue limit or strength is often found
to be about 50% of the UTS [8]. An example is shown in Figure
9 for Ck45 with and without a 0.1% V addition. [7]
The relatively high carbon content in these steels means that the
final forgings can be induction hardened, especially in the journal radius regions that are especially prone to high cycle fatigue
in crankshafts. Furthermore, the presence of the V in these steels allows them to be strong candidates for gaseous or ion nitriding. Since the V lowers the activity of N, the ease of nitriding
of a V-bearing steel is much easier and faster. The very hard nitrided surface would increase the fatigue limit even further.
FRACTURE RESISTANCE
Microalloyed medium carbon steel forgings were not designed to
have high impact toughness or low ductile-brittle transition temperatures. The two main reasons for the lack of low temperature Charpy impact toughness are: (i) a large prior austenite
grain size due to the absence of any real thermomechanical processing with the subsequent achieving of a high Sv value for the
prior austenite at the transformation temperature, and (ii) the
relatively high carbon content. Both of these factors are known
to contribute to inferior toughness values. In many cases the forging is operating as a rotating part in hot oil where toughness,
especially low temperature toughness, is not a concern.
Although the lack of toughness can be attributed to fairly large
austenite grain sizes and the high carbon contents, nevertheless, much research has been conducted over the past two decades to seek ways of improving the Charpy impact toughness. An
example of this work is shown in Figure 10, which shows the
strength-toughness balance for several experimental steels. [6]
Figure 10 shows that lower carbon content with higher Mn, Si
and N levels microalloyed with V can show good U-notch toughness in air cooled 30mm bars. Of course, transforming from
a much finer austenite would also materially affect the notch
toughness. This work has been encouraging since it does show
that the strength-toughness balance can be improved through
an understanding of the underlying physical and mechanical
metallurgy of these microalloyed medium carbon steels.
ECONOMICS
It is important to recognize the commercial impact of this development, where microalloyed medium carbon steels were deve9
Memorie
FIG. 11
Advantages in manufacturing
crankshafts from microalloyed medium
carbon steels. [7]
Vantaggi relativi alla produzione di alberi a
gomito da acciai microlegati a tenore
medio di carbonio. [7]
loped to replace heat treated carbon and low alloy steels. As
shown in Figure 11, [7] the processing of these microalloyed forgings is quite streamlined, when compared to the normal heat
treatment route. At least four processing steps can be avoided,
plus the cost of handling and transporting the forgings. Furthermore, the problem of mixed steel is much less likely in such
a streamlined process. The economic benefits of this simple process are clearly evident.
CLOSURE
The addition of V to medium carbon forging steels is an old and
successful technology. The benefits to strength and fatigue resistance are quite impressive.
Its broad use around the world speaks to its usefulness. The benefits to the economy(less steel needed for functionality) and environment(less steel produced and less energy consumed) are
unfortunately probably underappreciated.
REFERENCES
1) A. J. DeArdo, “Niobium in Modern Steels,” International Materials Reviews, Vol. 48, No. 6, (2003), p371-402.
2) D. Frodl, A. Randak, K. Vetter, Harterei Techn Mitt, Vol.29, No. 9, (1974),
p169-175.
3) Albert von den Steinen, Serosh Engineer, Elisabeth Horn and Gunther
Preis, Stah u. Eisen, Vol. 95, No.6, (1975), p.209-214.
4) H. Brandis and S. Engineer, Fachberichte-Huttenpraxis Metallweiterverarbeitung, Sept. 1981, p.666.
5) T.Gladman, I. D. McIvor and F. B. Pickering, J. Iron Steel Inst., Vol. 210,
(1972), p.916.
6) Fundamentals and Applications of Microalloying Forging Steels, Conf.
Proceedings, Edited by C. J. Van Tyne, G. Krauss and D. K. Matlock, The
Minerals, Metals & Materials Society, Warrendale, Pennsylvania, (1996).
7) Fundamentals of Microalloying Forging Steels, Conf. Proceedings, Edited by G. Krauss and S.K. Banerji, The Met. Soc. of AIME, Warrendale,
PA, (1987).
8) Mechanical Metallurgy, George E. Dieter, Third Edition, McGraw-Hills,
1986, p375.
Abstract
Acciai microlegati per prodotti forgiati ad alta resistenza
Parole chiave: acciaio, forgiatura
Negli ultimi trentacinque anni, sono state sviluppate due famiglie di acciai microlegati (MA) destinate agli acciai per barre ad
elevata resistenza o ad applicazioni in forgiatura. La prima famiglia è stata introdotta nel 1974 ed era costituita da acciai a medio
tenore di carbonio con aggiunte di piccole quantità di niobio o di vanadio. Questi primi acciai con contenuti medi di carbonio avevano microstrutture perlitico-ferritiche e presentavano una buona tenacità e resistenza alla fatica ad alto numero di cicli. Dopo
circa 15 anni, sono state introdotti gli acciai multifasici microlegati, che avevano microstrutture composte da miscele di ferrite,
bainite, martensite e austenite residua, a seconda della composizione e del processo. Questi acciai erano in grado di raggiungere
una tenacità molto elevata, con buona resistenza alla fatica e alta resistenza alla frattura. Nei primi anni ‘70, i prodotti forgiati
ad alta resistenza potevano essere ottenuti soltanto mediante trattamento termico finale, che implicava riscaldo e bonifica (QT).
È stato dimostrato più volte che i forgiati raffreddati ad aria realizzati con acciai perlitico-ferritici MA possiedono proprietà di tenacità e resistenza alla fatica simili a quelle di forgiati più costosi sottoposti a trattamento termico. Questo studio fa una rassegna dello sviluppo degli acciai microlegati perlitico-ferritici negli ultimi 35 anni.
10
Acciaio
Caratterizzazione meccanica
delle fasi dell’acciaio Duplex 2205
mediante nanoindentazione
M. El Mehtedi, S. Spigarelli, P. Ricci, C. Paternoster, E. Quadrini
Gli acciai Duplex sono caratterizzati dall’avere una struttura bifasica costituita da austenite e ferrite;
le loro proprietà meccaniche dipendono dall’interazione di molti fattori, quali la composizione chimica,
la morfologia delle fasi ed il rapporto austenite/ferrite. Quest’ultimo diminuisce all’aumentare della
temperatura di deformazione. La caratterizzazione meccanica delle due fasi risulta estremamente difficile
con le prove meccaniche tradizionali, le quali forniscono un valor medio e quindi il loro contributo congiunto.
Allo scopo di determinare il contributo e le proprietà di ogni singola fase, è stato quindi adoperato un nuovo
metodo basato sulla tecnica della nanoindentazione. I campioni in Duplex 2205 analizzati, sono stati
precedentemente deformati a caldo a 950, 1000, 1100, 1150 e 1200 °C allo scopo di determinare l’effetto
della temperatura di deformazione sulle proprietà meccaniche di ogni singola fase.
PAROLE CHIAVE:
acciaio inox, deformazioni plastiche, lav. plastiche a caldo, caratterizzazione materiali, tecnologie
INTRODUZIONE
Gli acciai inossidabili Duplex devono il loro nome alla loro particolare microstruttura formata da austenite e ferrite. Dalla combinazione di queste due fasi dalle differenti caratteristiche
deriva una struttura con ottime proprietà meccaniche e di resistenza alla corrosione che combina gli aspetti migliori degli acciai austenitici e ferritici senza raggiungere l’elevato costo delle
superleghe.
Se d’altra parte, determinare le proprietà meccaniche delle singole fasi con le tecniche tradizionali non risulta fattibile date le
dimensioni della struttura, una possibile soluzione potrebbe essere il ricorso alla nanoindentazione. La nanoindentazione è una
tecnica sviluppata di recente, che differisce dai normali test di
durezza, dove le impronte sono dapprima generate imponendo
un dato carico e poi analizzate con tecniche di microscopia. Nella
nanoindentazione, invece, il carico e la profondità di penetrazione sono registrati di continuo dalla fase di carico a quella di
scarico, dando luogo così ad un diagramma carico-profondità di
penetrazione. Tale diagramma fornisce maggiori informazioni
rispetto ad un’immagine di microscopia dell’impronta residua, in
quanto ci rivela la “storia” della deformazione elastica e plastica
al variare del carico e permette la determinazione della durezza
e del modulo di Young in funzione della profondità di penetrazione.
Questa tecnica è stata utilizzata, per ovvi motivi, soprattutto per
lo studio dei nanocompositi e di film sottili [1]; infatti la misura
delle proprietà meccaniche di una superficie su scala micro e
nano metrica sta assumendo sempre maggiore importanza sia
nel settore della ricerca che in quello industriale dal momento
M. El Mehtedi, S. Spigarelli, P. Ricci,
C. Paternoster, E. Quadrini
Dipartimento di Meccanica,
Università Politecnica delle Marche,
Via Brecce Bianche,
I-60131 Ancona, Italia
C
Mn
0.018
1.815
TAB. 1
che i sistemi nelle tecnologie odierne tendono sempre più a decrescere nelle dimensioni. Con tale tecnica si riescono a definire
le proprietà di un materiale in tutti quei casi in cui le tecniche
comuni non sono applicabili, ossia per la misura delle proprietà
meccaniche su scale molto piccole (ad esempio nei rivestimenti
sottili), o ancora quando è necessario investigare le caratteristiche meccaniche dei componenti strutturali di un materiale (durezza e modulo di Young di fasi differenti, di precipitati; effetto
di seconde fasi disperse di una matrice metallica, etc.), nonché
dove siano disponibili piccoli volumi del materiale di prova oppure sia necessario avere una bassa profondità di penetrazione.
La nanoindentazione si adatta bene anche allo studio dei materiali massivi (bulk).
MATERIALE E PROCEDURE SPERIMENTALI
Nella tabella 1 è riportata la composizione chimica dell’acciaio
Duplex 2205 oggetto di studio.
I campioni per l’esecuzione dei test di nanoindentazione sono
stati precedentemente sottoposti a prove di torsione alle temperature di 950, 1000, 1100, 1150 e 1200°C e velocità di deformazione pari a 0,5 s-1. La prova consisteva in: un riscaldamento fino
alla temperatura di deformazione seguito da una permanenza a
tale temperatura per 300 secondi ed in una deformazione fino a
rottura del campione con conseguente tempra immediata con un
getto d’acqua per congelare la microstruttura. I risultati ottenuti
dall’analisi di tali condizioni sono poi stati confrontati con il provino tal quale.
Al fine di migliorare la finitura superficiale e mettere in evidenza la microstruttura per essere in grado di stabilire quale
P
S
Si
0.027 0.0006 0.345
Ni
Cr
Mo
N
5.138 22.057 2.609 0.1493
Composizione chimica dell’acciaio Duplex 2205 (% in peso).
Duplex 2205 chemical composition (wt%).
11
Memorie
Il metodo di Oliver-Pharr [2,3] è stato utilizzato per analizzare le
curve carico-profondità di penetrazione ottenute per dedurre la
durezza (H) e il modulo di Young ridotto (Er) di ciascuna fase.
Un preciso posizionamento dell’indentatore su una determinata
fase e dimensioni dell’impronta considerevolmente minori in
rapporto alle dimensioni del grano in esame, sono stati importanti presupposti per ottenere informazioni attendibili circa le
proprietà meccaniche della struttura. Tutte le posizioni di indentazione sono state infatti determinate tramite scansioni SPM
(Scanning Probe Microscopy), per evitare di includere nelle misure i bordi di grano e calcolare quindi l’effetto di rafforzamento
ad essi dovuto.
FIG. 1
Micrografia ottica del campione Duplex 2205 tal
quale.
As-received Duplex 2205 microstructure.
FIG. 2
Immagine
SPM del
campione
Duplex 2205
tal quale.
As-received
Duplex 2205
SPM image.
delle due fasi fosse sottoposta ad indentazione, la superficie di
ciascun campione è stata dapprima preparata con le comuni tecniche metallografiche e poi attaccata elettroliticamente in una
soluzione al 10% di acido ossalico in acqua distillata (6V di voltaggio per circa 20 s), cercando di raggiungere il miglior compromesso tra rugosità della fase austenitica e distinzione delle
fasi.
Il materiale allo stato non deformato presentava la tipica struttura austenitica in matrice ferritica con frazioni volumetriche
circa uguali (Fig. 1). Con l’aumentare della temperatura di deformazione la struttura diventava in prevalenza ferritica, fino a
raggiungere una frazione volumetrica attorno al 70% alla temperatura di 1200°C.
I test di nanoindentazione sono stati effettuati mediante Hysitron© UBI® 1 equipaggiato di una punta di diamante con modulo
di Young pari a 1171 GPa e coefficiente di Poisson pari a 0.07
[2]. Le prove sono state realizzate con una punta Berkovich, per
la quale la proiezione dell’area di contatto ideale risulta essere
AC=24,5 (hc)2 dove hc è la profondità di indentazione.
In particolare, i test di nanoindentazione eseguiti in questo studio si articolavano nelle seguenti fasi: avvicinamento della punta
al campione, fase di carico con velocità pari a 200 µN/s fino al
valore massimo imposto, mantenimento del carico per un tempo
ad esso proporzionale e successiva fase finale di scarico sempre
ad una velocità pari a 200 µN/s. Il carico massimo imposto variava tra 2000, 5000 e 10000 µN. Un numero variabile tra 10 e
15 indentazioni sono state eseguite in ciascuna fase per ciascuna
modalità di prova.
12
RISULTATI
In Fig. 2 è mostrata la struttura del campione tal quale così come
si presenta dalle immagini SPM.
Nella struttura ottenuta dalla scansione SPM la matrice ferritica
è rappresentata dalla parte chiara, mentre le zone più scure rappresentano la fase austenitica.
In Fig. 3 sono riportate le curve carico-profondità di penetrazione per la fase austenitica e ferritica rilevate nel campione non
deformato con un carico massimo di 5000 µN, le corrispettive
immagini SPM delle impronte residue per ambedue le fasi (Fig.
3b e Fig. 3c) e le sezioni trasversali (Fig. 3c e Fig. 3d).
Dall’analisi delle curve carico-profondità di penetrazione si evidenzia come nella fase austenitica si raggiungano, a parità di
carico, profondità maggiori, arrivando quasi a 200 nm sotto carico massimo della ferrite. Anche la profondità residua risulta
essere maggiore nella fase austenitica, come confermato dalle
immagini della sezione trasversale, ben correlate ai dati ottenuti
dalle curve di carico-penetrazione.
Le immagini SPM mettono inoltre ben in evidenza la differente
morfologia tra le due fasi, dovuta all’attacco chimico necessario
a distinguerle; la soluzione acida impiegata attacca la fase austenitica, causando una rugosità maggiore rispetto all’altra fase,
ma sempre senza superare valori tali da incidere sull’attendibilità dei risultati ottenuti.
In Fig. 4 sono riportate le curve, sempre relative al campione tal
quale, sia per la ferrite che l’austenite nelle altre condizioni di
carico imposte. La profondità di contatto aumenta con il carico
massimo imposto e risulta, anche in questo caso, essere sempre
maggiore nell’austenite a parità di Pmax.
I valori della nanodurezza ottenuti dalle prove sono riportati
negli istogrammi di Fig. 5 in funzione del carico massimo applicato per ciascuna delle due fasi. Le barre di errore rappresentano le deviazioni standard.
Tali valori mostrano una più alta variabilità per la fase austenitica rispetto agli stessi valori relativi alla fase ferritica. La causa
di tale comportamento è da imputare alla diversa morfologia
delle due fasi dovuta all’attacco chimico. L’analisi delle immagini SPM ha infatti rilevato che i valori della rugosità superficiale RMS della fase austenitica sono maggiori rispetto a quelli
della fase ferritica. Questo fatto, come già accennato, può essere
collegato alla preparazione del campione mediante l’attacco chimico che mette in evidenza le due fasi (il problema della rugosità assunta da grani di austenite in seguito a trattamenti di
pulitura chimica o elettrochimica è stato analizzato, per esempio,
in [4]). E’ noto che la rugosità superficiale di un materiale è uno
dei parametri più critici che condiziona l’accuratezza dei risultati delle nanoindentazioni [5]; infatti, se la superficie di prova
presenta delle asperità, solo una piccola parte della punta del
penetratore sarà a contatto con la superficie, portando di conseguenza a valori errati di profondità di contatto e quindi delle
proprietà meccaniche. Ne risulta quindi che le curve di caricoscarico risultano diverse le une dalle altre e di conseguenza
Acciaio
a
b
d
FIG. 3
c
e
Curve carico-penetrazione (Pmax=5000 µN) per la fase ferritica e austenitica (a) del campione tal quale e rispettive
immagini SPM delle impronte residue: (b) per la ferrite e (c) per l’austenite. In (d) ed (e) sono riportate le
corrispondenti sezioni trasversali.
Load-displacement curves (Pmax = 5000 µN) for ferritic and austenitic phase (a) for Duplex 2205 as-received sample and the
respectively imprint image: (b) for ferrite and (c) for austenite. In (d) and (e) are also presented the corresponding crosssections.
a
FIG. 4
b
Curve carico-penetrazione per la fase ferritica e austenitica con Pmax=10000 µN (a) e Pmax=2000 µN (b) per il
campione tal quale.
Load-displacement curves for ferritic and austenitic phases in Duplex 2205 as-received sample with Pmax = 10000 µN (a) and
Pmax = 2000 µN (b).
anche i valori di durezza e modulo di Young ridotto presenteranno una certa dispersione.
In generale, i risultati del presente studio mostrano che i valori
delle durezze diminuiscono all’aumentare del carico massimo
imposto per ambedue le fasi, ma in maniera più accentuata per
l’austenite. Per quanto riguarda invece l’andamento del modulo
di Young (Fig. 6), tenendo conto delle deviazioni standard, questo non risulta essere influenzato dal carico applicato per nessuna delle due fasi. I valori per la fase ferritica sono del resto
FIG. 5
Valori della durezza in funzione del carico massimo applicato nel
campione tal quale per la fase austenitica e ferritica.
Hardness values as a function of the maximum load applied for
austenitic and ferritic phase in Duplex 2205 as-received sample.
13
Memorie
FIG. 6
Valori del modulo di Young in funzione del carico
massimo applicato nel campione tal quale per la
fase austenitica e ferritica.
Young module values as a function of the maximum load
applied for austenitic and ferritic phase in Duplex 2205
as-received sample.
sempre maggiori rispetto a quelli dell’austenite, indice di una
maggiore rigidezza.
La durezza dei campioni sottoposti a prove di torsione varia con
la temperatura di deformazione, nell’intervallo tra 950°C e
1200°C. Come è infatti possibile vedere dalle curve per i tre differenti carichi considerati (Fig. 8), la durezza H è soggetta ad
una lieve diminuzione per le temperature più basse dell’intervallo considerato, fino a circa 1000 – 1100°C, per poi arrivare ad
un picco locale in corrispondenza della temperatura di circa
1150°C. Questo andamento, che è comune per tutti i carichi considerati (2, 5 e 10 mN), è anche comune alle due fasi, austenite
e ferrite. Come è possibile vedere nei grafici di Fig. 8, le due fasi
rispondono in maniera simile alla deformazione plastica a caldo
per una data temperatura, poiché entrambe mostrano andamenti
simili di durezza.
DISCUSSIONE
La nanodurezza, calcolata secondo il metodo precedentemente
accennato [2], è una valutazione delle proprietà plastiche del
materiale. Infatti esiste una correlazione tra il valore della pressione media esercitata nella zona della deformazione plastica di
indentazione e la tensione di snervamento: le due grandezze
sono generalmente correlate dalla seguente formula
H ≈ KY,
a
14
(1)
B
dove H è la durezza e Y la tensione di snervamento, entrambe
espresse nella stessa unità di misura [6], e K è una costante che
in genere viene presa pari a 3. La relazione (1), applicata acriticamente ai dati ottenuti nel presente studio, porterebbe comunque ad una forte sovrastima del valore della resistenza
meccanica delle due fasi; questo fatto deriva dalla sostanziale
complessità delle problematiche legate alla misurazione della
durezza tramite nanoindentazione. La durezza della ferrite è in
genere maggiore di quella dell’austenite, sia nei campioni in condizione tal quale, che dopo torsione. La diminuzione di durezza
per carichi maggiori è dovuta alla presenza di un evidente effetto di scala di indentazione (indentation size effect, ISE [7-9]),
riscontrabile in effetti in tutti i solidi con struttura cristallina. Il
fenomeno di ISE dipende dalle caratteristiche geometriche dell’indentatore, rappresentate dall’angolo tra la normale alla superficie e le facce della piramide del Berkovich, dalla profondità
di penetrazione h, dalla densità di dislocazioni S, dal modulo di
resistenza a taglio µ e dal vettore di Burgers b, queste ultime
grandezze caratteristiche del materiale.
L’ISE è evidente in questo caso per entrambe le fasi considerate,
indipendentemente dalla temperatura. Volendo tentare un’analisi quantitativa del fenomeno, si può partire dal modello proposto da Nix e Gao [8], per cui la durezza può essere espressa in
funzione della profondità di penetrazione secondo la seguente
relazione:
(2)
essendo H durezza ad una certa profondità di penetrazione, H0
la durezza per il limite di profondità infinita di penetrazione, h*
una lunghezza caratteristica e h la profondità di penetrazione. Il
grafico in Fig. 9 riporta lo studio dell’ISE per il materiale nella
condizione tal quale, secondo l’equazione (2); il valore di H0 può
essere dunque stimato, in linea di principio, calcolando l’intercetta dell’asse delle ordinate (punto corrispondente a profondità
di penetrazione infinita). Si ottengono così valori di H0 pari a
circa 3.1 e 3.7 GPa per l’austenite e la ferrite rispettivamente,
con rapporto fra le due grandezze (1.19) sostanzialmente in linea
con i risultati di Cho e Gurland, in base ai quali la resistenza
meccanica a temperatura ambiente della ferrite è circa 1.18 volte
quella dell’austenite [10].
H0 può inoltre essere espresso in funzione di altri termini, e più
in particolare da
(3)
essendo α costante di valore pari a 0.5. Anche h* può essere
espressa in funzione di altri parametri, e diviene
FIG. 7
Immagine SPM dopo
indentazione dei campioni
torsionati: (a) T=1200°C e
Pmax=2000 µN, (b) T=1150°C
e Pmax=10000 µN.
SPM image of the samples
after deformation at:
(a) T = 1200 ° C and
P max = 2000 µN,
(b) T = 1150 ° C and
Pmax = 10000 µN.
Acciaio
FIG. 8
Durezza dell’austenite (a) e della ferrite (b) dopo deformazione per torsione alle diverse temperature.
Hardness of austenite (a) and ferrite (b) phase after deformation in torsion at different temperatures.
lografiche non hanno però evidenziato la presenza di altre fasi
oltre a quelle esaminate, o di precipitati di altro genere, almeno
di dimensioni grossolane. Rimane quindi da chiarire questa anomalia della variazione della durezza in funzione della temperatura di deformazione a caldo.
Una ulteriore precisazione riguarda il fenomeno di pile-up presente ai bordi delle impronte, sia nella fase ferritica che nella
fase austenitica, che è evidente dalla Figura 3. Tale effetto è stato
trascurato nel presente studio, e quindi nel calcolo della durezza
non è stata effettuata alcuna correzione per la presenza del pileup, come presentato da Kese et al. [13].
FIG. 9
Indentation Size Effect sui due diversi componenti
del Duplex 2205 (materiale nella condizione tal
quale).
Indentation Size Effect on both different phases of
Duplex 2205 (as-received condition).
(4)
E’ evidente che la durezza dipende dalla radice quadrata della
densità di dislocazioni immagazzinate, ρS; di conseguenza,
quando per temperature superiori a 0,5 volte la temperatura assoluta di fusione del materiale (= TM), avvengono fenomeni di
addolcimento diversi per le due fasi e cambia anche la quantità
di dislocazioni immagazzinate. In particolare, per la ferrite sono
rilevanti i fenomeni di rinvenimento dinamico (DRV), a causa
dell’elevato valore dell’energia dei difetti di impilaggio dovuta
alla struttura cubica a corpo centrato [11,12], mentre nell’austenite sono più importanti i fenomeni di ricristallizzazione dinamica (DRX).
Per quanto riguarda la durezza, nel presente studio si è visto che
nell’austenite, si raggiungono valori quasi stabili quando si prendono in considerazione temperature di deformazione plastica
superiori a 1050°C, mentre per la ferrite c’è un picco pronunciato in corrispondenza di 1150°C. In linea di principio, nell’intervallo di temperatura considerato si può avere anche la
formazione o dissoluzione di precipitati che possono essere responsabili di una tale variazione di durezza. Le indagini metalLa Metallurgia Italiana - n. 9/2010
CONCLUSIONI
La caratterizzazione meccanica delle fasi austenitica e ferritica
dell’acciaio Duplex 2205 è stata effettuata mediante la tecnica
della nano indentazione in diverse modalità di prova. Si è dimostrato come, combinando le scansioni SPM con la nanoindentazione, si è in grado di determinare informazioni circa le
proprietà meccaniche delle singole fasi, cosa non possibile con
le normali tecniche di durezza.
Nel caso del campione non deformato, la durezza risulta decrescere all’aumentare del carico applicato per ambedue le fasi, in
accordo con la teoria dell’indentation size effect. Il modulo di
Young ridotto, inoltre, risulta essere sempre maggiore nella ferrite, segno di una maggiore rigidezza.
RINGRAZIAMENTI
Gli autori desiderano ringraziare l’ing. Paolo Vivani ed il Sig.
Daniele Ciccarelli per l’aiuto prestato nello svolgimento delle
prove sperimentali.
RIFERIMENTI BIBLIOGRAFICI
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
L. XIAODONG, B. BHARAT, Materials Characterization 48 (2002) 11
W. C. OLIVER and G. M. PHARR, J. Mater. Res., 7 (1992) 1564.
W. C. OLIVER and G. M. PHARR, J. Mater. Res., 19 (1992) 3.
A. SCHREIBER, C. ROSENKRANTZ and M. M. LOHRENGEL, Electroc. Acta 52 (2007) 7738.
K. L. JOHNSON, Contact Mechanics, Cambridge University Press,
Cambridge, (1985).
D. TABOR, The Hardness of Metals, Clarendon Press, Oxford, 1951,
37.
Y. LIU and A. H. W. NGAN, Scripta Mater. 44 (2001) 237.
W. D. NIX and H. GAO, J. Mech. Phys. Sol. 46 (1998) 411.
J. Y. SHU and N. A. FLECK, Int. J. Sol. Str. 35 (1998) 1363.
15
Memorie
[10] K. CHO and J. GURLAND, Metall. Trans. 19A (1988) 2027.
[11] E. INOUE and T. SAKAI, J. Japan Inst. Met. 55 (1991) 286.
[12] H.J. MCQUEEN, E. EVANGELISTA and M.E. KASSNER, Z. Metal-
lkud. 82 (1991) 336.
[13] K.O. KESE, Z.C. LI and B. BERGMAN, Mat. Sci. and Eng. A 404
(2005) 1.
Abstract
Mechanical characterization of phases in Duplex 2205 stainless steel
by nanoindentation technique
Keywords: duplex stainless steel, plastic deformation, hot working, materials characterization , technologies
Duplex stainless steels constitute a class of appreciated materials for their good combination of high strength, resistance to corrosion in presence of chloride-containing fluids and stress corrosion cracking. The microstructure of these alloys is composed by
almost equivalent volume fractions of ferrite and austenite; as a result, the mechanical properties of these materials measured
by the conventional techniques are a balanced mixture of the response of each of the constituent phases. This observation is particularly important in all those cases in which the plastic deformation of the alloy is characterized by large differences of stress
and strain partition between the constituent phases. Even micro-hardness is of limited use in this situation, except when the microstructure is extremely coarse. Nano-indentation is a recent technique developed to measure the mechanical properties of materials on nano-scale; although largely used in nano-science to characterize nanostructured materials, this technique can be
successfully applied also in bulk multi-phase alloys to investigate the mechanical properties of the different constituents. This
study aimed at investigating the mechanical properties of a duplex stainless steel in as received and hot-deformed conditions. The
measure of nano-hardness was found to be affected by the Indentation Size Effect; the analysis of the dependence of the hardness
on the indentation depth resulted in an estimation of the hardness for infinite penetration that could be considered an index of
the yield strength of the alloy. The conclusion is that the ratio between ferrite and austenite hardness is close to 1.19, in excellent agreement with literature data dealing with the tensile response of the two phases. The elastic modulus of ferrite was always higher than that of austenite. The nano-indentation measurements on the alloy deformed at high temperature showed a
decrease in strength of both phases, even though a peak in hardness, more evident in ferrite than in austenite, in the material
deformed at 1150°C was observed.
16
Acciaio
Thixoforming M2 tool steel:
a study of different feedstock routes
P. Kapranos, D. H. Kirkwood
Different aspects of thixoforming M2 tool steel feedstock produced through deformation recrystallization
and partial melting (RAP) and through Sprayforming routes. The spheroidal microstructures obtained
are compared as are the resulting properties after thixoforming. Experiences on die materials are described
as are the various challenges of thixoforming high melting point alloys.
KEYWORDS:
thixoforming, steels, high melting point alloys, sprayforming, GFM, mechanical properties
INTRODUCTION
Since the discovery of Semi-Solid Metal Processing or Thixoforming in the 1970’s at MIT [1], there have been significant developments in this technology. However, although thixoforming
is now a commercial process, with components being produced
each year for the automotive and for the consumer
products/electronics markets mainly using aluminium and magnesium alloys, there is still an apparent lack of visible acceptance in manufacturing.
This apparent paradox can be explained if we separate the thixomolding variant from the thixoforming umbrella of processes.
Thixomolding has achieved undisputed commercial success because it has concentrated on
specific markets using only magnesium alloys and utilizing a
compact process that alloys recycling. The rest of the thixoforming processes have had a long struggle in establishing credible material feedstock routes and recyclability of material
feedstock.
Despite that, because of the fine microstructures and the high integrity associated with thixoformed products, thixoformed aluminium products have replaced steels and some forged,
machined or cast aluminium parts, with the consequent savings
in manufacturing time and weight. Thixoforming indeed produces complex near-net-shaped components of high integrity,
with mechanical properties better than cast components but it
needs to provide repeatability at the right cost in relation to competing processes.
Thixoforming has the potential to be a successful commercially
viable manufacturing process but in order to achieve this it has
to find its own niche in a number of market applications that
will benefit from the advantages it has to offer.
In addition, thixoforming has to seriously establish itself as an
effective ‘hybrid’ manufacturing alternative for shaping high
melting point alloys.
Current research in the later field is concentrating on the deveP. Kapranos, D. H. Kirkwood
The University of Sheffield,
Department of Engineering Materials,
Sir Robert Hadfield Bldg, Mappin Street, Sheffield, S1 3JD, UK
[email protected]
Keynote lecture presented at the International Conference
“Hot Forming of Steels And Products Properties”
Grado, 13-16 Settembre 2009, organised by AIM
FIG.1
Schematic of thixoforming Press at Sheffield.
Schema di una pressa di tixoformatura a Sheffield.
lopment of high melting point alloys such as steels, iron-alloys,
copper-alloys, superalloys and other exotic materials, to further
exploit the potential benefits of this under-utilised metal forming
technique.
However, although thixoforming of high melting point alloys offers exciting possibilities and tremendous potential, and has already been part of the original work of over thirty years ago, it
is currently still in the research stage of development [2].
This paper will provide some insight of possibilities as well as
the challenges involved when shaping such high melting point
alloys by using M2 tool steel as a ‘model’ alloy.
EXPERIMENTAL SET-UP
The thixoforming of the chosen model alloy was carried out
using a Servotest hydraulic press at Sheffield that had been designed specifically for research in semi-solid processing of alloys and this press has been fully described in the literature [3]
and graphically shown in Figure 1. The procedure of thixoforming is schematically shown in Figure 2 and consisted of a number of steps: placing the billet in the induction coil within the
vacuum chamber, evacuate and back fill with N2+5% H2, heat
the billet to the semi-solid state, use a ‘softness indicator’ to
sense the condition of the billet and then inject into the die. Figure 3 shows a photographic series of images of the thixoforming process.
The ‘softness indicator’ consisted of electronic circuitry with a
17
Memorie
FIG. 2
Steps used for the
thixoforming
operation at the
Sheffield set-up.
Passaggi produttivi
utilizzati per le
operazioni di
tixoformatura presso
l’impianto di
Sheffield.
laser beam being introduced below the billet which was standing on ceramic pins.
Once the billet achieved semi-solid status it would slump onto
the pins, disrupt the sensing of the laser beam and thus trigger
the thixoforming operation. Once experience was gained this
system was superseded by a pre-set heating program. In addition, in order to achieve reasonably uniform heating throughout
the volume of the billet, ceramic wool pads were placed on the
top and bottom surfaces of the billets to reduce heat losses.
The modeling of this process has also been described in the literature [4, 5].
As for the feedstock material, two production routes were investigated at the time; deformation by GFM and Sprayforming by
FIG. 3
Thixoforming sequence for M2 tool
steel.
Sequenza per la tixoformatura dell’acciaio
da utensile M2.
FIG. 4
Sprayformed and GFM material feedstock microstructures exhibiting the familiar spheroidal microstructures
obtained through isothermal heat treatment to the semi-solid state.
Microstrutture dei materiali di base formati a spruzzo e con GFM che mostrano le tipiche microstrutture sferoidali ottenute
tramite trattamenti termici isotermici allo stato semisolido.
18
Acciaio
Ospray Metals Ltd. Figure 4 shows examples of the microstructures for these two routes.
RESULTS & DISCUSSIONS
Forgings of M2 tool steel of both feedstock production routes
were thixoformed and their mechanical properties tested and
compared with the as received materials. Figures 5, 6 and 7
show such results. Similar work has been also carried out using
Stellite 21 alloys. Thxoforming of other grades of tool steels such
as T15 and H13 has also been carried out by the Sheffield group
as well as that of cast iron, ductile iron, stainless steel and other
superalloys [2].
The main challenges in thixoforming these high temperature alloys have been: Billet uniformity, oxidation, injection delivery
materials and die materials.
All these had great influence on the final properties of the thixoformed products. As has been described above, the billet uniformity was established by the use of ceramic insulation at the
top and bottom surfaces of the billets.
This was not a fully satisfactory approach as on occasion it resulted in the inclusion of ceramic insulation into the final product adversely thus affecting properties (See Figure 10). Current
technologies have overcome this challenge as can be seen in recent research by the research group based in Liege [6].
Oxidation was reduced by using a vacuum and gas protective atmosphere and once again any discrepancies there resulted in
oxide inclusions in the final product. The injection materials
challenge was solved by the use of ceramic pedestals (Syndanyo
– cement based composite).
Finally the most problematic of the above challenges was the
use of die materials. A number of die materials were used. Graphite of different grades and strengths was a convenient and
‘cheap’ material to use but it had the adverse effect of high conductivity resulting in chilling of the semi-solid slurry being injected and therefore leading to a number of solidification related
problems discussed below. The use of ceramic and even ‘sand’
dies alleviated the problems of using graphite but introduced
their own deficiencies; fracture, erosion, and cost.
As mentioned above the use of graphite dies created the problem of premature chilling of the injected slurry. This in effect
created a solid ‘box’ with slurry still flowing through it as can be
seen in Figure 9. Of course when the slurry filled the remaining part of the cavity by backing up on itself it created a lap
which on testing was the source of cracking and therefore reduction in mechanical performance. The use of ceramic dies eliminated this problem and the mechanical properties obtained
using ceramic dies were consistently less scattered than the
ones obtained using the graphite dies.
FIG. 6
FIG. 5
Fracture strength of M2 as a function of injection
velocity (GFM route feedstock). Values of as
received material in the longitudinal and transverse
directions are also shown.
Strength values of Sprayformed M2 tool steel: as
sprayed and as thixoformed.
Valori di resistenza dell’acciaio da utensile M2 allo stato
di prodotto per formatura a spruzzo e di prodotto
tixoformato.
Resistenza alla frattura dell’acciaio M2 in funzione di
velocità di iniezione (GFM route feedstock). Sono
mostrati anche i valori del materiale come ricevuto nelle
direzioni longitudinale e trasversale.
FIG. 7
Strength values of GFM
M2 tool steel: as received
(Longitudinal and
transverse directions) and
as thixoformed.
Valori di resistenza di
acciaio da utensile M2 GFM
allo stato di fornitura
(direzione longitudinale e
trasversale) e tixoformati.
19
Memorie
FIG. 8
Thixoformed M2 tool steel
cog wheels and resulting
typical microstructures:
GFM feedstock (top) and
Sprayformed feedstock
(bottom). Other alloys are
also shown in the picture.
Ruote dentate in acciaio da
utensile M2 tixoformato e
microstrutture tipiche
risultanti: materiale di base
GFM (in alto) e materiale di
base prodotto mediante
formatura a spruzzo (in
basso). Nella figura vengono
mostrate anche altre leghe.
FIG. 9
Generation of laps in
thixoformed ‘fingers’.
Generazione di difetti in ‘dita’
tixoformate.
20
Acciaio
FIG. 10
Defects in thixoformed
parts originating from
embedded ceramic blanket
material.
Difetti in pezzi tixoformati
generati da materiale
ceramico incorporato.
During these studies the results were statistically analyzed and
the corresponding microstructures and possible defects were
looked at in relation to the corresponding mechanical properties in order to draw reasonable conclusions. It was clear from
the above that when the billets were thixoformed under reducing atmosphere, therefore containing no inclusions, within ceramic dies they consistently had good mechanical properties, on
occasion better than those of the starting feedstock. Any discrepancies introduced through uneven heating, ineffective protective atmosphere and rapid chilling due to the die material all
resulted in reduced mechanical properties.
Thixoforming of high temperature alloys is challenging but feasible. If sufficient care is taken in devising appropriate procedures and using appropriate materials for handling and
delivering the semi-solid slurries intricate products can be obtained that exhibit good mechanical properties.
Since the work described in this paper was completed, a number
of researchers have moved the dream of forming high melting
point alloys in the semi-solid state closer to a commercial reality
[7] but although promising, these efforts are far from complete.
A European consortium under the support of the COST Office is
currently working towards the possibility of bringing the efforts
of past and present researchers in this field to fruition [8].
REFERENCES
1. D. P. Spencer, R. Mehrabian and M. C. Flemings, Rheological behavior of Sn-15 pct Pb in the crystallization range, Metallurgical Transactions, 1972, 3, pp1925-1932.
2. Kapranos P, ‘Thixoforming of high melting point alloys’, Revue / Journal Title Chuzo kogaku ISSN 1342-0429, 2005, vol. 77, no 8, pp. 518525, Nihon Chuzo Kogakukai, Tokyo, JAPON
3. Kapranos P, 'Thixoforming /SSM facilities at Sheffield', Proc. of the
4th Int. Conf. on Semi-solid Processing of Alloys and Composites,
June 18-21,1996, Sheffield, England, pp 360-363 (Edit. Kirkwood DH
& Kapranos P)
4. Kapranos P, Gibson RC, Kirkwood DH, Hayes PJ, and Sellars CM 'Induction heating and partial melting of high melting point thixoformable alloys' Proc. of the 4th Int. Conf. on Semi-solid Processing of
Alloys and Composites, June 18-21,1996, Sheffield, England, pp 148152 (Edit. Kirkwood DH & Kapranos P)
5. Kapranos P, Gibson RC, Kirkwood DH, Hayes PJ, and Sellars CM, 'Modelling the Induction Heating of High Melting Point Alloy Slugs for
High Temperature Mechanical Processing', Vol 12., J. of Material
Science & Technology, March 1996, pp 274-278
6. A. Rassili, J.C. Pierret, G. Vaneetveld, P. Cezard, R. Bigot, M. Robelet,
‘Influence of the Pre and Post Treatment Operations on the Properties of the Thixoformed Steel Parts’, Solid State Phenomena Vols. 141143 (2008) pp 689-694
7. Cezard P and Sourmail T. ‘Thixoforming Of Steel : A State Of The Art
From An Industrial Point Of View’, Solid State Phenomena, Vols.
141-143, (2008), pp 25-35
8. COST Action 541 ‘Semi-Solid Processing of Steels’:
http://www.cost541.ulg.ac.be/index.html
Abstract
Tixoformatura di acciai
da utensile M2:
studio dei risultati ottenuti
con materiali provenienti
da diversi processi
Parole chiave:
acciaio, tixoformatura, processi
Nel presente lavoro sono stati analizzati diversi aspetti della
tixoformatura di acciai da utensile M2 ottenuti attraverso diversi processi: deformazione, ricristallizzazione e fusione
parziale (RAP) e Sprayforming.
Le microstrutture sferoidali ottenute sono state confrontate
insieme alle proprietà rilevate dopo tixoformatura. Si descrivono esperienze su materiali da stampo e anche le varie
sfide poste dalla tixoformatura di leghe ad elevato punto di
fusione.
21
Memorie
Forgiatura
Vacuum treatments for hydrogen removal
in 140 ton ladle for big ingots casting
M. Paura, M. De Santis, M. Calderini, S. Neri, R. Palomba, L. Sartini
Hydrogen removal can be accomplished via different steelmaking routes (VOD, ASEA, RH). Focus was given
on the first two technologies. As main differences between the systems based on a (multi)-plug equipped ladle,
in VOD the vacuum chamber is obtained by coupling a roof with a tank, in ASEA plant
–where electromagnetic melt stirring is also exploited - coupling occurs directly with the ladle, leading often
to a non perfect sealing. Moreover, VOD plant is able to perform under vacuum steel degassing treatment
at pressure values lower than usually reached in ASEA plant, and is also equipped with a oxygen lance
allowing to produce stainless steels with very low carbon and nitrogen content. The industrial need
of achieving very low hydrogen contents for big ingots casting with an acceptable costs/benefits ratio called
for a comparison between performances of different vacuum treatments strategies. A CSM numerical
degassing model was applied to 140 ton ladle conditions either after ASEA or VOD treatment.
KEYWORDS:
acciaio, degasaggio sotto vuoto, affinazione acciaio, deidrogenazione, lingotti
INTRODUCTION
Hydrogen removal can be accomplished under vacuum via different steelmaking routes , e.g., VOD, ASEA, RH. Within the VOD technology, the vacuum chamber is obtained by coupling a roof with
a tank. On the other hand, in ASEA plant –where electromagnetic melt stirring is also exploited – such a coupling occurs directly with the ladle, leading often to a non perfect sealing.
The most relevant capability of a VOD plant for dehydrogenation
treatments is the achievement of a stronger vacuum level than
usually reached in ASEA plant. Another is the lower risk
to have damages due to splashing phenomena respect to
ASEA plant. Due to these features, VOD (VD) plant technology is used by many ingot producers to point toward very low
hydrogen contents in melt, with shorter vacuum treatment than
ASEA plant.
The industrial need of achieving in TKAST very low hydrogen
contents (lower than 1 ppm) for big ingots casting with an acceptable overall costs/benefits ratio called for a comparison
between performances of different vacuum treatments strategies. A CSM numerical degassing model was applied to
140 ton ladle conditions either after ASEA or VOD treatment.
The benefits of a VOD treatment are highlighted, together with
indications on the best working conditions, supported by encouraging experimental results.
ASEA AND VOD DEGASSING
The hydrogen reaction is controlled by liquid phase mass
transfer of hydrogen. For non recirculating systems the reactions increase with argon stirring in a complex manner (scheM. Paura, M. De Santis
Centro Sviluppo Materiali, Roma, Italy
M. Calderini, S. Neri
Società delle Fucine, Terni, Italy
R. Palomba, L. Sartini
ThyssenKrupp Acciai Speciali Terni, Italy
FIG. 1
Schematic of melt degassing phenomenon driven
by inert gas bubbling [1].
Schema del fenomeno di degasaggio di un bagno
metallico grazie al soffiaggio di bolle di gas inerte [1].
matic in figure 1 [1]). Due to the bubbles expansion, which occurs in function of pressure and liquid height, rate equations
are complex [2].
Degassing reaction follows the relationship:
(1)
Here X is the concentration of the element undergoing on degassing up to the equilibrium value XE, V the steel volume, kx
the mass transfer coefficient and A the area exchange (=effective
reaction surface) which under vacuum depends on pressure and
bath height.
23
Memorie
As a result, after integrating (1), one obtains for hydrogen:
(2)
where [H] is the hydrogen concentration, the labels f, i, e
refers respectively to final, initial and ‘ideal’ (as state by
equilibrium conditions) concentration. For ‘strong’ degassing
(below 10 torr) [H]e can be neglected with respect to the other
terms, so the rate equation becomes:
(3)
The CSM model was based on such a relationship, integrated by a bubble size estimation in function of flow rate
[3,4]. The model was applied to ladle treatment in TKAST ASEA
and VOD conditions. The single-straight type ASEA electromagnetic stirrer allows clockwise melt flow, from wall to surface
and again toward the bottom, and the melt velocity imposed decays rapidly from wall to center [3].
The 140 ton VOD ladle has a diameter/height ratio equal to 1,
and two eccentric plugs, both in a half side of the bottom plane,
one at half radius and the second at 800 mm from the first.
FIG. 2
RESULTS AND DISCUSSION
A CSM numerical model, based on equations (1÷3) were applied to TKAST ASEA and VOD conditions, accounting for the
ladle steel volume, the bubble area arising from the flow rate injected and the mass transfer coefficient k dependent on
steel velocity [2]. In this way, the overall rate equation can
be written [1]:
Esempio di risultati con modello fluidodinamico del
moto dell’acciaio nella siviera ASEA sotto l’azione
dell’induzione elettromagnetica e del gas di stirring.
Frazione in volume di gas (mappa) e velocità (in m/s)
nel piano passante per un setto.
(4)
where QM is the melt recirculation rate, related to steel velocity
[5].
The recirculation rates needed were taken from modelling calculations. To achieve the recirculation rate term:
- numerical simulations, based on a Computational Fluid-dynamics model accounting for electromagnetic forces, were
performed for ASEA conditions (see figure 2),
- physical modelling simulation were performed to simulate
VOD conditions. Here, a 1:4 water model scale was built at
CSM labs (figure 3). Flow rates (plugs, lance) were scaled according to the Similarity theory (keeping the same Froude’s
number between model and plant [4]).
The reason of using different modelling strategies for the different technologies is the following.
Recirculation rate (expressed in m3/s), relevant parameter in (4),
represents the liquid flow rate in the ladle ‘batch’ volume. Moreover, recirculation rate is equal to ladle volume divided
by the ‘homogenisation time’ (defined as the time needed to achieve perfect homogenisation throughout all the liquid volume with
respect to a well defined quantity – e.g., temperature, alloying elements concentration). Therefore, a way to achieve recirculation rate values is to measure melt homogenisation
times. In the VOD frame, it is more reliable to achieve this
information from physical model, therefore, the reduced scale
VOD water model was set up.
On the other hand, electromagnetic forces cannot be simulated
in a water modelling frame, and for ASEA the recirculation rate
values were derived directly from the liquid velocity values
after numerical calculations.
To assess mixing times and in turn take information on the
recirculation rates for the VOD dehydrogenation model, a
conductivity technique was used. A salt solution is injected
24
Example of result with CSM fluid-dynamics
modeling of flow in ASEA ladle with
electromagnetic forces and gas stirring. Coloured
map in the plane highlighted passing per a plug
represents gas volume fraction. Velocity vectors
(unit in m/s).
FIG. 3
Image of the CSM water scale model of TKAST VOD
ladle system.
Modello ad acqua in scala della siviera VOD di TKAST,
presso i laboratori CSM.
into the liquid and a probe connected to a conducimeter allows
the visualisation of the conductivity signal.
When an asymptotic value is achieved, complete mixing was attained in the liquid.
For ASEA calculations, different flow rates were simulated,
together with cases where induction stirring effect was in the
same direction or countercurrent with respect to the ascending
gas. The results are summarised in figure 4. Start hydrogen concentration was 3.5 ppm. As a result, on one hand the increase
of flow rate proved to favour more efficient degassing, on
Forgiatura
FIG. 4
Results of the CSM dehydrogenation numerical
model to ASEA TKAST conditions. Hydrogen content
in melt vs treatment time for different stirring
conditions (see legenda).
Risultati dell’applicazione del modello numerico di
deidrogenazione CSM nelle condizioni operative ASEA
in TKAST. Contenuto di idrogeno nell’acciaio in funzione
del tempo di trattamento per differenti condizioni di
agitazione (vedi legenda).
Fig. 6
model to VOD TKAST conditions. Hydrogen content
in melt vs treatment time for different vacuum
levels. Gas flow rate 200 l/min per plug.
deidrogenazione CSM nelle condizioni operative VOD
del tempo di trattamento per diverse condizioni di vuoto
(vedi legenda). Portata gassosa per setto: 200 l/min.
the other hand the adequate managing of induction stirring
compensated for a lower flow rate, useful to reduce risks of splashing in ASEA layouts. The better condition occurs when
induction and gas stirring are counter-current (the residence
time for bubbles in melt is longer, and the higher turbulence can further break gas flow into bubbles enhancing the exchange area for dehydrogenation).
A constant pressure value of 2 mbar was used in the calculations.
For VOD calculations, the dehydrogenation trend in function of
the plug gas flow rates (100 e 200 Nl/min) is shown in figure 5
and 6. The stronger the vacuum, the lower the equilibrium hydrogen (asymptotic value). The higher the flow rate, the more
efficient is the treatment, due to the enhanced exchange area
between bubbles and dissolved hydrogen.
The model sensitivity to flow rate proved to be in agreeLa Metallurgia Italiana - n. 9/2010
FIG. 5
model to VOD TKAST conditions. Hydrogen content
in melt vs treatment time for different vacuum
levels. Gas flow rate 100 l/min per plug.
deidrogenazione CSM nelle condizioni operative VOD
del tempo di trattamento per diverse condizioni di vuoto
(vedi legenda). Portata gassosa per setto: 100 l/min.
ment with other similar validated tank degassing models results [2]. The effect of a lower pressure is much more significant
than that of the gas flow rate, even though as soon very low
pressure values are approached (<10 mbar), the dehydrogenation rate increase is lower.
By comparing figures 4, 5 and 6, the stronger VOD efficiency
with respect to ASEA operations is shown. Taking as a reference
a hydrogen target content of 1 ppm, one can see that the difference between the most efficient ASEA and VOD performances
can also exceed 5 min.
This result shows how equilibrium aspects (effect of pressure,
number of bubbles) are of prevailing importance in dehydrogenation process with respect to mixing phenomena in the melt.
As a matter of fact, ASEA dehydrogenation performances are
less efficient in spite of the better mixing behaviour of ASEA
the electromagnetic stirred melt.
This can be shown in figure 7, where in a literature [7]
diagram correlating power transferred to melt by different
stirring system, the working point corresponding to our two
cases were derived [8, 9] and evidenced.
Besides, the fact that dehydrogenation does not increase automatically with gas stirring is explained by the fact that in gas
stirred systems under vacuum, the gas expansion is mainly concentrated close to the surface. It was estimated that 50% of the
expansion occurs within a distance from the ladle surface equal
to 20% of the total bath height [10].
As a result, when exceeding the flow rate values needed for a
reasonable rapid mixing (around 100 Nl/min) most part of the
further flow rate is spent for energy dissipation at the surface,
with risks of splashing due to the close coupling roof-ladle,
and is not fully available to enhance melt recirculating rate
and bath mixing.
Summarising, the most powerful driving force for efficient hydrogenation process proved to be the dissolved hydrogen equilibrium shift towards stronger vacuum levels.
The further advantage of VOD layout (higher roof makes harmless splashing occurrence) allows to enhance the process by
injecting higher gas flow rate in the melt. Gas stirring policy can
be further enhanced also with suitable flow rate management
in case of multi-plug ladle.
25
Memorie
ction of the metallurgical targets in degassing operations for
the steel being produced.
FIG. 7
Diagram correlating specific power transferred to a
melt with stirring systems (electromagnetic, gas)
and bath homogenisation time [7]. The working
points (calculated from [8] and [9]) corresponding
to the conditions of figures 5, 6 and 7 are
highlighted.
Diagramma che correla la potenza specifica ceduta al
bagno metallico mediante diversi sistemi di agitazione
(elettrromagnetica, gas) e il tempo di omogeneizzazione
del bagno metallico [7]. I punti di lavoro sono calcolati
da [8] e [9] e corrispondono alle condizioni delle figure
5, 6 e 7.
Based on these aspects, VOD-based steelmaking route was
used in TKAST, for forging steel fabrication, to obtain hydrogen values lower than 1 ppm for bottom pouring ingots,
avoiding expensive anti-flakes treatments after forging.
Table 1 shows the results obtained after 30 minutes of vacuum treatment in 140 ton ladles performed in both TKAST
VOD and ASEA plants. The average lower content (-25%)
achieved after VOD treatment, supports the model indications. Here, hydrogen measurements were performed in both
cases by the same Hydris device just at the end of vacuum treatment. This means that results are comparable, and the differences are not dependent on instrumental accuracy. Hydris
devices are widely used for hydrogen measurement inside
ladles, and their accuracy is usually acceptable for this kind
of production.
Moreover, the verified model suitability allowed more in general to get a tool for driving the dehydrogenation strategy,
by optimising treatments times (and in turn, costs) in fun-
Heat n.
1
2
3
4
5
6
TAB. 1
Hydrogen content (ppm)
after treatment
VOD
ASEA
0.8
0.9
0.8
0.7
0.9
0.8
1.1
1.2
1.0
1.2
1.0
1.1
Hydrogen content in the steel after 30 minutes of
vacuum treatment performed in 140 ton ladle
(TKAST VOD and ASEA plants).
Contenuto di idrogeno nell’acciaio dopo 30 minuti di
trattamento sotto vuoto attraverso VOD e ASEA nella
siviera TKAST da 140 ton.
26
CONCLUSIONS
With the aim of achieving very low hydrogen contents for
big ingots casting with an acceptable costs/benefits ratio, a
CSM validated model was used to compare the dehydrogenation
performance at the TKAST 140 ton ladle conditions either after
ASEA or VOD treatment.
The most relevant results were the following:
- ASEA treatment is efficient when coupling purging to
electromagnetic stirring with induced flow counter current
with respect to the gas plume from the plug. Increasing the
gas flow rate per plug is risky, due to the gas expansion,
mainly concentrated close to the surface, inducing risks of
splashing due to the close coupling roof-ladle;
- VOD operations are more effective because of the strongest vacuum levels achievable. As a result, the equilibrium
gas partial pressure is lower than in ASEA operation, and dehydrogenation can continue up to very low level;
- a further advantage of dehydrogenation through VOD operations is related to the VOD layout. The higher roof makes harmless splashing occurrence, so an increased flow rate can be
used to enhance in the melt presence of bubbles where dissolved hydrogen diffusion occurs. Gas stirring policy can be
further enhanced with suitable flow rate management in case
of multi-plug ladle performed by the aid of CSM validated
model indications.
Based on these aspects, even if ASEA based route should be preferable for logistic reasons, VOD-based steelmaking route was
used in TKAST, for forging steel fabrication, to obtain hydrogen values lower than 1 ppm for bottom pouring ingots, avoiding expensive anti-flakes treatments after forging.
Moreover, the verified model suitability allowed more in general to get a tool for driving the dehydrogenation strategy,
by optimising treatments times (and in turn, costs) in function of the metallurgical targets in degassing operations for
the steel being produced.
REFERENCES
[1] B. Kleimt, S. Kohle, K. P. Johann, A. Jungreithmeier and J. Molinero,
‘Dynamic process model for denitrogenation and dehydrogenation by vacuum degassing’, Scandinavian Journal of Metallurgy, 2000; vol. 29, pp. 194 – 205.
[2] The Making, Shaping and Treating of Steel, 11th Edition - Steelmaking and Refining Volume, AISI Steel Foundation.
[3] F. Oeters, Metallurgy of steelmaking, Springer-Verlag, Berlin, 1997.
[4] J. Szekely, Fluid flow phenomena in metal processing, Academic
Press, London, 1979.
[5] F. Oeters, W. Pluschkell, E. Steinmetz and H. Wilhelmi, ‘Fluid flow
and mixing in secondary metallurgy’, Steel Research 59 (1988), n. 5,
pp. 192 – 200.
[6] N. Bannenberg, B. Bergmann and H. Gaye, ‘Combined decrease of
sulphur, nitrogen, hydrogen and total oxygen in only one secondary
steelmaking operation’, Steel Research 63 (1992), n. 10, p. 431-437.
[7] K. Nakanishi, T. Fujii and J. Szekely, ‘Possible relationship between energy dissipation and agitation in steel processing operations’, Ironmaking & Steelmaking, 1975, vol. 3, p. 193–197.
[8] Y. Sundberg, ‘Mechanical Stirring Power in Molten Metal in Ladles
Obtained by induction Stirring and Gas Blowing’, Scandinavian Journal of Metallurgy, 1978, vol. 7, pp. 81-87.
[9] Z. Miaoyong, C. Nailiang and Y. Hongliang, ‘Electromagnetic
stirring in ladle refining processes’, Millennium Steel, 2005, pp.
101-104.
[10]K. Krishnakumar and B. Ballal, ‘Effect of vacuum on mixing
behaviour in a ladle – a watermodel Study’, ISIJ International’,
1999, vol. 39, n. 11, pp. 1120-1124.
Forgiatura
Abstract
Deidrogenazione sotto vuoto in siviera da 140 ton
per il colaggio di grandi lingotti
Parole chiave: steel, steelmaking, dehydrogenation, ingots, vacuum steel degassin
La deidrogenazione può essere effettuata secondo differenti cicli d’acciaieria (che prevedono VOD, ASEA, RH). Questo studio è
incentrato sui primi due. La differenza sostanziale fra questi sistemi è che nel VOD la camera di vuoto è ottenuta con un
reattore chiuso da una volta, nell’impianto ASEA, dove il metallo liquido è anche agitato elettromagneticamente, è accoppiata direttamente con la siviera, non sempre a perfetta tenuta. Inoltre, l’impianto VOD permette di effettuare trattamenti sotto vuoto più spinto rispetto all’ASEA, ed è anche dotato di una lancia ad ossigeno che permette la produzione di acciaio
inossidabile con contenuto bassissimo di carbonio e azoto.
La necessità di ottenere una deidrogenazione spinta nel colaggio di lingotti da forgia in quadro di accettabile rapporto costi/benefici ha portato a confrontare l’efficacia di differenti trattamenti sottovuoto nelle condizioni di TKAST. Per l’occorrenza, è
stato applicato un modello di degasaggio CSM alle condizioni di lavoro negli impianti ASEA e VOD.
I risultati più rilevanti sono stati i seguenti:
- il trattamento ASEA è efficiente quando si effettuano nello stesso tempo stirring con gas ed elettromagnetico, imponendo all’acciaio un moto in senso opposto a quello del flusso di gas. Un aumento della portata gassosa è rischioso a causa
dell’espansione del gas sotto vuoto, che avviene principalmente presso l’interfaccia metallo-scoria, con pericolo di proiezioni
di fluido (acciaio e scoria) verso la volta;
- il fatto che il trattamento VOD dia luogo a deidrogenazione più efficiente rispetto a quella che si ha in ASEA, nonostante l’agitazione elettromagnetica sia più efficace, indica come il fenomeno sia legato più alla chimica del processo che alla fluidodinamica dell’acciaio liquido;
- il trattamento in VOD è più efficace per i livelli più spinti di vuoto ottenibili. La pressione parziale del gas in condizioni di equilibrio è minore che nel trattamento ASEA, e la deidrogenazione può andare avanti sino a livelli molto spinti.
Nel nostro caso, l’obiettivo di un contenuto di idrogeno di 1 ppm è ottenuto circa 5 minuti prima rispetto al processo ASEA;
- un ulteriore vantaggio della deidrogenazione in VOD piuttosto che in ASEA è di tipo impiantistico. La maggiore altezza
della volta in VOD di fatto neutralizza i rischi di proiezioni di fluido verso l’alto. Quindi possono essere usate maggiori portate
di argon, il che vuol dire un maggior numero di bolle nel liquido disponibili per la diffusione dell’idrogeno. La strategia di soffiaggio di gas inerte può essere ulteriormente migliorata con la gestione della portata in una siviera a più setti porosi basata
sulle indicazioni derivanti dal modello CSM.
Sulla base di queste considerazioni, malgrado il ciclo ASEA sia preferibile per motivi logistici, è stato seguito un ciclo di fabbricazione basato sulla deidrogenazione in VOD per il colaggio in TKAST di grandi lingotti. In questo modo, sono stati ottenuti bassi valori di idrogeno, inferiori ad 1 ppm, evitando successivi costosi trattamenti ‘anti-fiocco’ dopo la forgiatura. In più,
il modello si è dimostrato un strumento affidabile ed utile a calibrare, più in generale, tempi e prestazioni legate ai trattamenti
di degasaggio fornendo un contributo importante al miglioramento della produttività.
27
Memorie
Forgiatura
Grain size prediction
during open die forging processes
D. Recker, M. Franzke, G. Hirt, R. Rech, K. Steingießer
One of the most important target parameters during open die forging is the microstructure, respectively
the grain size. This paper details different semi-empiric models that ultimately help to predict the microstructure
properties of a forged block. As a first step, trials in industrial scale were performed by Buderus Edelstahl
GmbH and attended by SMS Meer GmbH. The collected process data was used by the Institute of Metal Forming
(IBF) for the numerical analysis of the open die forging process and to validate the microstructure prediction
module STRUCSIM. The numerical prediction of the grain size shows a good agreement with the results
obtained from the metallography. In a second step models for the core fibre of a forged block were developed
at the IBF. The models use data from the online process measurement and simplified plastomechanical
interrelations for the calculation of equivalent strain and the temperature in the core of the forged part during
the process. With their results the microstructure in the core fibre of the workpiece can be predicted online.
The models are still in development and the most recent results will be presented in this paper.
KEYWORDS: open die forging, strain distribution, temperature distribution,
simulation, microstructure, optimisation, model
INTRODUCTION
Modern open die forged products with high quality properties and
material require a smooth and accurate forging process. An important aspect during the open die forging of large workpieces
is to pursue the right forging strategy so that the desired microstructure can be achieved and possible casting defects can be
removed from the core of the workpiece. To guarantee the desired microstructure and mechanical properties within the workpiece it is of high interest to be able to predict the microstructure within the workpiece during the forging process [1].
Generally, a minimum true strain of the workpiece is required
to accomplish a workpiece with the desired quality. As the local
strain is highly inhomogeneous within the workpiece, the accomplishment of a minimum true strain can not always avoid defects caused during forging. For large workpieces, this can lead
to extensive waste and economical loss, since large workpieces
usually represent a high value. Present trials to avoid these problems with programmed forging through pre-calculated forging
plans failed due to differences between reality and calculation
which were adding up during the process. Aside from that, the
pre-calculated plans do not consider unpredicted interventions
by the forging press operator. This leads to a less reproducible
process and is therefore not satisfactory.
Since there is the desire for reproducible processes and guaD. Recker, M. Franzke, G. Hirt
RWTH Aachen University, Germany
R. Rech
Buderus Edelstahl GmbH, Wetzlar, Germany
K. Steingießer
SMS Meer GmbH, Mönchengladbach, Germany
Keynote lecture presented at the International Conference
“Hot Forming of Steels And Products Properties”
Grado, 13-16 Settembre 2009, organised by AIM
ranteeing the expected microstructure inside of the workpiece,
the project “Entwicklung eines Prozessmodells zur Online-Optimierung von Freiformschmiedungen großer Blöcke” (“Development of a process model for online-optimisation of open die forging of large workpieces”) which is supported by the DFG through
SPP1204 “Algorithmen zur schnellen, werkstoffgerechten Prozesskettengestaltung und -analyse in der Umformtechnik” (“Algorithms for a fast, material-suitable design of a process chain
in metal forming”) [2], aims at developing fast simulation models.
These models shall be the basis for realising a numerical assistant
system, which is able to suggest the best continuation of the forging sequence at any time during the forging process. A fundamental, semi-empiric model was developed, which is able to calculate the equivalent strain and temperature for the core fibre of
a forged block. By connecting its results to the microstructure model STRUCSIM [3, 4], prediction of the expected grain size can
be made.
The models for the equivalent strain and the temperature are derived and verified with a reference solution. As a reference solution the Finite Element Analysis (FEA) simulation is chosen since more cognition can be gathered easier and faster from the FEA
than from real processes, at the moment. As evaluation of the FEA
and STRUCSIM a forging process was carried out at Buderus Edelstahl GmbH and the metallographic examination of the workpiece
was compared with the FEA results.
OPEN DIE FORGING EXPERIMENTS
Open die forging experiments were carried out at Buderus Edelstahl GmbH and were attended by SMS Meer GmbH to evaluate
FEA simulations and the integrated microstructure calculation
(STRUCSIM). The material of the workpiece was 26NiCrMoV. The
ingot had an initial diameter of around 700 mm and was forged
to a diameter of around 500 mm using a total of four passes. The
forging temperature was 1170 °C and a total of 65 strokes were
carried out. The initial average grain size of the ingot was 1000
µm (Fig. 1, left) [5].
For metallographic examination discs were cut out of the forged
29
Memorie
a
FIG. 1
b
Open die forging at Buderus Edelstahl GmbH (a) and positions for metallography (b).
Forgiatura a stampo aperto presso la Buderus Edelstahl GmbH (a) e posizioni degli esami metallografici (b).
Position
=
Average Grain Size
µm
M-12-1
M-12-2
M-12-3
M-12-4
F-12-1
F-12-2
F-12-3
F-12-4
150 - 250
100 - 200
100 - 300
150 - 250
150 - 250
150 - 300
150 - 300
100 - 200
TAB. 1
Results of metallography at some selected
locations.
Risultati della metallografia in alcuni dei punti
selezionati.
ingot at three different positions after the forging process (Fig.
1, right). From each disc nine samples were taken from defined
positions. The average grain size for each sample was determined and some results are shown in Table 1.
THE MEASURING SYSTEM LACAM FORGE®
During the whole forging process all mechanical data was recorded
and the process was recorded per video as well. Aside from that,
the measuring system LACAM® [6] which is installed at Buderus
Edelstahl GmbH delivers relevant process information online [7].
The LACAM® system which is developed by Ferrotron Technologies GmbH provides information about
• length of the workpiece
• surface temperature of the workpiece
• stroke protocol and saddle position
in a uniform coordinate system through laser scanning the forged workpiece during the forging process (Fig. 2). The distance
measurement is based on a time-of-flight measurement using a
pulsed semiconductor laser deflected in two directions. The heat
radiation is measured and provides the possibility to measure the
surface temperature distribution of the workpiece [6]. With this
data the process can be simulated in detail. Fig. 2 shows the surface temperature distribution recorded by LACAM® for the first
stroke. The temperature of the contact surface of the tool and the
higher temperatures of the ingot side are clearly visible.
30
FIG. 2
Temperature measurement with LACAM® after the
first stroke.
Misurazioni della temperatura mediante LACAM® dopo
il primo passaggio.
NUMERICAL SIMULATIONS
The measured and recorded information was used to create the models for the numerical simulations. The models were created with
the software PEP [8] and the simulations were calculated using the
FEA system LARSTRAN [9]. The microstructure was calculated using
the module STRUCSIM which is directly coupled to LARSTRAN.
The four passes of the reference open die forging process were
simulated. The initial average grain size was set to 1000 µm, according to the original process. The calculated distribution of the
average grain size in the longitudinal sections for all four passes
is shown in Fig. 3. At the head and foot of the ingot the results
show a coarse microstructure (between 500 µm and 1000 µm).
In the middle of the forged part the microstructure is finer (between 60 µm and 300 µm) while the grain size increases from the
core towards the surface. The calculated microstructure results
were compared with the microstructure determined from the experiment. This comparison is shown in Table 2 exemplary for the
position 3 (see Fig. 1) of the middle and foot disc.
As Table 2 shows, the experiment and the simulation show a good
correlation. Regarding the comparison between the experiment
and simulation, the experiment validates the FEA, so that in the
next steps the different models could be derived and verified from
FEA reference solutions.
Forgiatura
FIG. 3
Simulated grain size distribution in the
core plane for the four forging passes.
Simulazione della distribuzione delle
dimensioni dei grani nel piano centrale per i
quattro passaggi di forgiatura.
Position
=
M-3-3
M-12-3
F-3-3
F-12-3
TAB. 2
Average Grain Size
Experiment
Simulation
µm
µm
150 - 300
100 - 300
150 - 300
150 - 300
173
318
289
295
Comparison of experiment and simulation for the
grain size.
Confronto tra dati sperimentali e simulati, relativamente
alla dimensione dei grani.
CONCEPT OF AN ASSISTANT SYSTEM
FOR OPEN DIE FORGING PROCESSES
The LACAM® system is a first step to a more reproducible forging
process. The vision of the Institute of Metal Forming is to set up
an intelligent assistant system that uses the provided data of LACAM® to predict the microstructure distribution within the workpiece and finally to assist the forge during the process.
According to the present works at the Institute, a concept for the
assistant system could comprise the modules as shown in Fig. 4.
The process provides current data about the change in length, position of the tools and the surface temperature of the workpiece.
With the change in length of the workpiece a strain model calculates the equivalent strain in the workpiece. Respectively, the
surface temperature is used by a temperature model to calculate the temperature in the workpiece. The results are processed
subsequently by the microstructure model STRUCSIM to predict
the microstructure in the workpiece.
In long-term work the assistant system is to be developed. Aside from the already mentioned models, a visualisation-tool has
to be derived. This tool shall display the current distribution of
microstructure in the workpiece. In combination with this and
different optimisation methods [10], the numerical assistant system shall suggest and display the optimal continuation of the current forging pass.
STRAIN MODELS
The main criterion for the models is a fast working algorithm so
that e.g. the strain can be calculated during the process. As shown
above, FEA simulations can be used to calculate several technical values rather precisely. In the case of open die forging the disadvantage of the FEA is that the numerical calculation of the process takes more time than the actual real process itself. Therefore the FEA is inappropriate to compute e.g. the strain (online)
FIG. 4
Possible concept of the intended assistant system
for open die forging processes.
Possibile impostazione del sistema di monitoraggio nel
processo di forgiatura a stampo aperto.
during the forging process. Thus, a fast and online capable model has to be derived.
In a first step a one dimensional model for calculating the equivalent strain was developed for the core fibre of the forged block.
The basic idea is that the equivalent strain can be calculated on
the basis of global values, such as the change in length of the workpiece. The strain model calculates the equivalent strain through
the change in length of the workpiece during one stroke of a forging pass. The functionality of this approach was described previously in different publications [11, 12].
So far the model for the core fibre showed good results for forging process consisting of only one forging pass. Thereby the main
problem was to superimpose the model results for the first pass
and following passes [12]. In recent trials a method for superimposing the strain distribution for several forging passes was
developed. In this method the stretching of the strain distribution due to the stretching of the forming zone is considered. Hence, the strain distribution for several following forging passes can
be simply added to the total strain distribution:
(1)
As evaluation of the model a two-dimensional FEA reference process was created. The main simulation parameters can be found
in Table 3. After every simulated stroke the change in length of
the core fibre is measured and put into the strain model and it
calculates the strain distribution for the core fibre in the forming
zone for this stroke. Finally the calculated strain distributions for
each pass are added up and so the total strain distribution in the
core fibre of the workpiece is determined.
31
Memorie
For an initial bite ratio of sB0/h0 = 0.8 Fig. 5 shows the results for
the equivalent strain in the core fibre of the reference simulation
at the end of the process. The calculated total equivalent strain is
shown as well as the strain distribution for each forging pass. As
the results show, there is a good correlation between the calculated total equivalent strain and the FEA results. Only small differences between FEA-solution and model-solution can be observed.
The differences may be explained by the half-empirical character
of the model. In the FEA-calculation material flow varies in each
stroke and thus the distribution of equivalent strain in the core
fibre also varies in each stroke. Since the model uses the change
in length of the workpiece to calculate the equivalent strain it does
not consider different material flow behaviours within the workpiece. However, the results show that the model can be used for
a fast estimation of the equivalent strain distribution in the core
fibre of the workpiece during the open die forging process.
The same model was applied to a bite ratio of sB0/h0 = 0.5. Fig. 6
shows the results for the FEA and the model. In this case the model does not deliver as accurate results as for a bite ratio of 0.8.
Especially at positions where a forging pass ends (x = 65 % and
x = 75 %), the model shows unusual peaks. Methods to smooth
the distribution in these zones have to be developed in the further progress of the project. Furthermore, in the areas were single strokes overlap, a better method to superimpose the strain distributions has to be developed as well.
Nevertheless, considering that the model is based on very simple assumptions [11, 12] and calculates the equivalent strain distribution of the core fibre by only using the change in length between to passes, it delivers good results for the examined bite ratios. For the small bite ratio the accuracy might be reduced but
the principle of the distribution is similar to the one of the FEA.
The comparisons show that the principle of the model works. The
results are in good agreement with the FEA reference solution.
Aside from that, the model works fast. In the next steps the model has to be evaluated with three dimensional reference solutions.
Furthermore it has to be implemented in a module so that interfaces to the LACAM® system can be derived.
A different approach is to accelerate the FEA simulation so that
the results can be used during the process. In general, one forging pass consists of several similar strokes. The idea of the IBF
is to simulate one forging stroke and then isolate the computed
strain distribution. Thereby all elements of the FEA results with
an equivalent strain of εV ≈ 0 are deleted. The isolated strain distribution is then transferred to the final shape of the workpiece as many times as strokes were performed so that the strain
results of one stroke are superposed resulting in the total equivalent strain distribution after one forging pass (Fig. 7).
Combined with the LACAM® system the principle of the superposition could work as follows: The system measures the workpiece. With the measured point cloud a net of the final shape
of the workpiece is generated. Combined with the input of required
parameters (bite ratio, tool speed, material, etc.) a database delivers the “template” for one stroke. The database can be generated from FEA simulation results which were performed for different conditions prior the forging process. This template is then
superposed to the total strain distribution within the workpiece.
The basic idea of the superposition had to be evaluated. Therefore the superposition was used on a FEA reference process. For
a certain bite ratio one pass of a forging progress was completely
simulated. Aside from that, one stroke with the same bite ratio
was simulated as well. The results were isolated and transferred
to the final shape of the workpiece of the complete simulation.
For a bite ratio of sB0/h0 = 0.8 the results are shown in Fig. 8. The
left part shows the graphic results of the FEA simulation and the
superposition. Both pictures show a similar distribution of the equivalent strain throughout the whole workpiece. The core fibre shows
slight differences between the FEA results and the results of the
FIG. 5
FIG. 6
Parameter
Value
Unit
l0
h0
saddle width
saddle radius
µ
ϑStart
ϑEnvironment
ϑTool
650
100
200
20
mm
mm
mm
mm
0.3
1200
25
300
°C
°C
°C
TAB. 3
Simulation parameters for the two-dimensional FEA
model.
Parametri di simulazione per il modello FEA
bidimensionale.
Comparison of the equivalent strain in the core
fibre calculated by the strain model and by FEA
simulation after three forging passes. The bite ratio
is sB0/h0 = 0.8.
Confronto della deformazione equivalente nella fibra
centrale calcolata mediante modello di deformazione e
mediante simulazione FEA dopo tre passaggi di
forgiatura. Il “bite ratio” è sB0/h0 = 0.8
32
Comparison of the equivalent strain in the core
fibre calculated by the strain model and by FEA
simulation after three forging passes. The bite ratio
is sB0/h0 = 0.5.
Confronto della deformazione equivalente nella fibra
centrale calcolata mediante modello di deformazione e
mediante simulazione FEA dopo tre passaggi di
forgiatura. Il “bite ratio” è sB0/h0 = 0.5
Forgiatura
FIG. 7
Principle of the superposition method for an open
die forging process.
Principio del metodo di sovrapposizione per un
processo di forgiatura a stampo aperto.
superposition which increase towards the surface. A more precise comparison for the core fibre and the surface fibre of FEA
simulation and superposition is shown in the right part of Fig.
8. In addition, the difference between the FEA and superposition
– related to the FEA distribution – is plotted in both graphs. For
the core fibre both strain distributions match each other quite accurately for the chosen bite ratio. For the forged part of the core
fibre the biggest differences can be observed at the edges of the
different forming zones (25 % up to around 75 %). For the centre
of the forming zones the difference is around 0 % to 15 %. In the
lower right part of Fig. 8 the comparison of the surface nodes is
shown. While the relative difference between superposition and
FEA is rather small for the core fibre, the comparison for the surface shows much bigger differences. For the whole forged part the
relative difference is around 25 % up to 75 %.
Overall, it can be seen, that the superposition delivers good results for the core fibre for the examined bite ratio. However the
differences for the surface can not be neglected. For a two dimensional application of this method an appropriate superposition algorithm has to be developed so that the surface and the areas between the core fibre and surface deliver more accurate results. Nevertheless, the results show that the principle of this method works.
TEMPERATURE MODEL
As in every other hot forming process the temperature of the workpiece plays a very important role during the open die forging pro-
cess. The temperature field within the workpiece occurring during the process influences directly or indirectly many important
parameters such as local flow stress, local forming capacity, strain
and stress condition, force and work requirements. Another important aspect is the microstructure. Especially during thermomechanical forming, the thermal conditions determine the microstructure which directly influences the mechanical properties
and the quality of the final product [11, 13]. Knowing the temperature distribution, in particular in the core of the workpiece is
therefore of high significance to the forging process. Thus, the target is to develop a fast model which calculates the temperature distribution within the workpiece on basis of online measurements
of the surface temperature of the forged block. This model is being
developed at the Institute of Metal Forming. In the first stage the
model calculates the temperature distribution in the core fibre of
the workpiece. As the functionality of this model is described precisely elsewhere [11], only a short overview of the model is given.
The workpiece is divided in three sections, in which different thermal conditions exist (Fig. 9). Since the model is derived from a
two-dimensional FEA-reference-simulation, there can be only heat
flow in x- or in y-direction. It is assumed that only one-dimensional
heat flow in the y-direction occurs in section 1 (forming zone).
For the duration of one stroke, heat transfer occurs from the workpiece surface into the tool. Aside from that, an increase in temperature (dissipation) due to the deformation work may occur in
the core fibre. For section 2, heat loss by radiation to the surroundings causes a one dimensional heat transfer in y-direction.
In section 3, it is assumed that heat flows only in x-direction.
According to the direction of heat transfer, one dimensional finite Difference (FD) models are placed in each section. Based on
a known temperature of the surface node, each FD-model calculates the temperature of the respective node in the core fibre after the current stroke. The dissipation in the forming zone is thereby approximated through the equivalent strain given by the
strain model (see above).
The temperature model was applied to the same two dimensional
FEA simulation as it was used for the strain model. For the second
pass the temperature distribution was calculated for the core fibre and is shown in Fig. 10 (left part) as well as the temperature
distribution calculated by the FEA. For a more detailed view of the
block edges, the temperature distribution of model and FEA is
shown in Fig. 10 (right part) exemplary for the right block edge.
Overall the temperature distribution of the model shows a good
agreement with the results of the FEA. In the formed area (between 40 % and 100 % of the related length of the core fibre) the
maximum difference between model and FEA is around 5 °C. Re-
FIG. 8
The left part shows the graphic results
of FEA and superposition method. The
right part shows the results of both
methods for the core fibre and the
surface, as well as the related
difference between both (FEA and
superposition).
Il grafico della parte sinistra mostra i
risultati dei metodi FEA e di
sovrapposizione. La parte destra mostra i
risultati di entrambi i metodi per la fibra
centrale e per la superficie, e le ralative
differenze fra i due (FEA e
sovrapposizione).
33
Memorie
workpiece. However, more FEA trials have to be run so that material with e.g. a higher dissipation can be tested. This would give
the opportunity to test if the differences between model and FEA
are still as low as described above. Furthermore trials of real forging processes have to be run to validate the model in real life.
FIG. 9
Subdivision of the workpiece in three different
sections with different thermal conditions in each
area.
Suddivisione del pezzo in tre diverse sezioni con
differenti condizioni termiche in ogni area.
lated to the overall start temperature of 1200 °C, this difference
is rather small. Similar results can be observed for the block edges (figure). The differences between model and FEA are not higher than 10 °C in this area.
The results show that the presented model approximates the FEA
results for the temperature distribution of the core fibre of a forged block quite well. Considering that the FEA represents a real
life forging process, the model delivers the possibility to give a
fast estimation of the basic temperature distribution inside the
MICROSTRUCTURE MODEL
Once the data for the temperature and the equivalent strain inside of the workpiece is available, the microstructure model STRUCSIM can calculate the microstructure properties of the workpiece. The exact functionality of STRUCSIM is described in detail elsewhere in different literature [3, 4]. Therefore, only the results
will be presented in the following paragraphs.
The calculated strain and temperature distribution in the core fibre of both, model and FEA, is used as input for STRUCSIM. In
this specific case the temperature and strain distribution after
one forging pass is used. As initial grain size the fictional value
of 1000 µm is used. The STRUCSIM results for the dynamically
recrystallised fraction of the core fibre after one stroke are shown
for model and FEA values in Fig. 11.
Fig. 11 shows that there are some discrepancies between the FEA
and the model values for the dynamically recrystallised fraction.
The differences are around 10 % reaching up to around 20 % in
the area of the right edge of the block (90 % - 100 % of the core
fibre). These differences might be explained by the differences
between the input data. As described above, there are differences between the strain distribution calculated by model and by
FEA. The same applies to the temperature distribution.
Fig. 12 shows the calculated grain size in the core fibre after one
FIG. 10
Comparison of the temperature
distribution in the core fibre
calculated by the temperature model
and by FEA simulation after two
forging passes. The bite ratio is sB0/h0
= 0.8.
Confronto della distribuzione della
temperatura nella fibra centrale calcolata
con il modello di temperatura e mediante
simulazione FEA dopo due passaggi di
forgiatura. Il “bite ratio” è sB0/h0 = 0.8.
FIG. 11 Dynamically recrystallised fraction of the core fibre
calculated by STRUCSIM for model-values and FEAvalues.
Frazione ricristallizzata dinamicamente nella fibra
centrale calcolata mediante STRUCSIM per i valori del
modello e i valori FEA.
34
FIG. 12 Grain size distribution in the core fibre calculated
by STRUCSIM for model-values and FEA-values.
Distribuzione della dimensione dei grani nella fibra
centrale calcolata mediante STRUCSIM per i valori del
modello e i valori FEA.
Forgiatura
forging pass for model and FEA. As well as for the dynamically
recrystallised fraction, the model values do not match the FEA
values exactly. The differences are around 100 µm throughout the
forged part. Nevertheless, the grain size distribution calculated
using the model values shows a similar character to the FEA distribution. This means that the model values can be used by
STRUCSIM to give a fast estimation of the microstructure of the
core fibre during the forging process. Improving the models for
strain and temperature will lead to a more accurate calculation
of the microstructure properties of the workpiece.
CONCLUSION
Summarizing the presented work, the following aspects can be
highlighted:
• The trials carried out show that the finite element analysis can
be used to predetermine the microstructure inside the workpiece. Additionally, the trials show that the FEA can be used
as a reference instead of a real forging process if appropriate
material data is available.
• Two different models for calculating the strain in the core fibre
of the workpiece during the forging process were presented. Both
models work fast and deliver quite accurate results. The different inaccuracies in predetermining the equivalent strain
shall be avoided in the future so that the results can be improved.
• The presented temperature model is able to predict the temperature distribution in the core fibre in principle using measured surface temperatures. Some inaccuracies can be observed. Nevertheless, the model can be used to approximate the
temperature distribution of the core of the forged block during
the process.
• Connecting the model for temperature and equivalent strain distribution to the microstructure model (STRUCSIM) shows that
the grain size of the core fibre can be predicted during the forging process. The estimation of the grain size can be improved
by improving the accuracy of the strain and temperature model.
In future the models have to be implemented in a compiler language so that a fast and automatic work of the models is secured.
Aside from that, algorithms which process the measured data for
the models will be developed. Another important step is to implement the different models in the assistant system. Furthermore,
optimisation methods and algorithms have to be implemented in
the system. These will deliver an online support for the forging
press operator and will suggest the optimal continuation of the
current forging pass.
ACKNOWLEDGEMENTS
The authors would like to thank the “Deutschen Forschungsgemeinschaft” (DFG) for the financial support of these works within the SPP1204 “Algortihmen zur schnellen, werkstoffgerechten
Prozesskettengestaltung und –analyse in der Umformtechnik“.
REFERENCES
[1] Dürr, O., Beitrag zur Qualitätsverbesserung beim Freiformschmieden von Langprodukten und gekrümmten Formteilen, in Institute
of Metal Forming; RWTH Aachen University. 2007, Shaker Verlag:
Aachen.
[2] SPP1204. Algorithmen zur schnellen, werkstoffgerechten Prozesskettengestaltung und -analyse in der Umformtechnik. [cited;
Available from: www.spp-schnelle-algorithmen.de.
[3] Karhausen, K. and R. Kopp, Model for Integrated Process and Microstructure Simulation in Hot Forming. Steel Research, 1992.
63(6): p. 247-256.
[4] Karhausen, K., Integrierte Prozeß- und Gefügesimulation bei der
Warmumformung, in Institut für Bildsame Formgebung, RWTH
Aachen. 1994, Stahleisen: Düsseldorf.
[5] Franzke, M., et al., Online analysis of grain size evolution for open
die forging processes, in International Forgemasters Meeting. 2008:
Santander. p. 1-6.
[6] FERROTRON-Technologies-GmbH, Lacam FORGE, Measuring System. Company Flyer, 2008.
[7] Rech, R., et al., Einsatz von Lasermesstechnik (LaCam Forge) an
Freiformschmiedepressen, in 22. Aachener Stahlkolloquium, G.
Hirt, Editor. 2007: Aachen. p. 53-60.
[8] Franzke, M., Zielgrößenadaptierte Netzdiagnose und -generierung
zur Anwendung der Finite Elemente Methode in der Umformtechnik, in Institut für Bildsame Formgebung, RWTH Aachen. 1999,
Shaker Verlag: Aachen.
[9] LASSO-Ingenieurgesellschaft-mbH, LARSTRAN User's Manual.
2007, Leinfelde-Echterdingen.
[10] Posielek, S., Einsatz kombinatorischer Optimierungsmethoden bei
automatischer Optimierung von Umformprozessen, in Institut für
Bildsame Formgebung, RWTH Aachen. 2005, Shaker Verlag: Aachen.
[11] Franzke, M., D. Recker, and G. Hirt, Development of a process model for online-optimization of open die forging of large workpieces.
Steel
Research
International,
2008.
79(10).
DOI:
10.2374/FRI08SP067.
[12] Recker, D., M. Franzke, and G. Hirt. Online calculation of equivalent
strain, temperature and microstructure within the workpiece during open die forging. in SENAFOR - Conferência Internacional der
Forjamento. 2008. Porto Alegre, Brazil.
[13] Braun-Angott, P. and B. Berger, Berechnung der Schmiedeguttemperaturen beim Reckschmieden. Archiv für das Eisenhüttenwesen, 1981. 52(12): p. 465-468.
Abstract
Previsione delle dimensioni dei grani
durante i processi di forgiatura a stampo aperto
Parole Chiave: acciaio, forgiatura, metallografia, simulazione numerica
Uno dei parametri più importanti da seguire durante la forgiatura a stampo aperto riguarda la microstruttura e, più propriamente,
la dimensione del grano. Questo studio fornisce una descrizione dettagliata dei diversi modelli semi-empirici che possono contribuire a predire le proprietà microstrutturali di un blocco forgiato. Come primo passo, sono state eseguite prove su scala industriale presso la società Buderus Edelstahl GmbH e presenziate dalla SMS Meer GmbH. I dati di processo raccolti sono stati
utilizzati dall'Institute of Metal Forming (IBF) per condurre l'analisi numerica del processo di forgiatura a stampo aperto e per
convalidare il modulo STRUCSIM di previsione della microstruttura. La previsione numerica della granulometria mostra una
buona concordanza con i risultati ottenuti con la metallografia. In una seconda fase sono stati sviluppati, presso la IBF, modelli
per la fibra centrale di un blocco forgiato. I modelli utilizzano i dati delle misurazioni rilevati durante il processo e le interrelazioni plasto-meccaniche semplificate per il calcolo della deformazione equivalente e della temperatura durante il processo, a
cuore della parte forgiata. Con questi risultati è possibile prevedere in linea la microstruttura della fibra centrale del pezzo. I modelli previsionali sono ancora in fase di sviluppo e in questo documento vengono presentati i risultati più recenti.
35

Acciaio Memorie Microalloyed steels for high

Transcription

Similar documents

I CALCHI - Martini

three rolls

Kussmaul`s sign in effusive constrictive pericarditis

to Komoda Modular Stair Installation

3600 SERIES VIDEOPHONES VIDEOCITOFONI SERIE 3600

made in italy

SuperJaws XXL Accessories

Implant JOURNAL

Documento PDF - AMS Tesi di Laurea

Tesi dott. Andreolli Marco - Università degli Studi di Verona