low cost refractory castable die development for …

7º CONGRESSO BRASILEIRO DE ENGENHARIA DE FABRICAÇÃO 7th BRAZILIAN CONGRESS ON MANUFACTURING ENGINEERING

20 a 24 de maio de 2013 – Penedo, Itatiaia – RJ - B rasil May 20th to 24th, 2013 – Penedo, Itatiaia – RJ – Brazil

© Associação Brasileira de Engenharia e Ciências Mecânicas 2013

LOW COST REFRACTORY CASTABLE DIE DEVELOPMENT FOR TITANIUM SHEET SUPERPLASTIC FORMING

Gerard Bernhart, [email protected] Fabien Nazaret, [email protected] Thierry Cutard, [email protected] 1

1Universite de Toulouse, ICA(Institut Clemet Ader), Mines Albi, F-81013 Albi cedex 09 France 2Aurock, 54 Rue Gustave Eiffel, 81000 Albi

Abstract. Fiber Reinforced Refractory Castable (FRRC) is an innovative material solution to manufacture SPF tools for forming titanium alloy sheets. Indeed, FRRC are able to conserve good mechanical properties up to 900°C. Fiber reinforcement allows to improve mechanical properties and to avoid a catastrophic failure. Moreover, constitutive laws can be considered to reproduce the FRRC’s mechanical behaviour when using FEM codes. As a consequence, it is now possible to proceed to a numerical design of SPF tools that are based on FRRC. Following this route, large SPF tools have been designed, manufactured and tested under industrial conditions. Successful results have been obtained in the field of forming TA6V sheets.

Keywords: Refractory castable, superplastic forming, ceramic dies, fiber reinforcement, microstructure, material modeling)

1. INTRODUCTION

When performing the superplastic forming process (SPF), tools are subjected to severe thermomechanical loadings.

In the case of the SPF forming of titanium based sheets, tools made of heat resistant nickel-chromium cast steels are necessary. But, such materials are quite expensive and their lead-time to manufacture is very long. This drawback is a limiting factor for an expansion of the SPF process. Since several years, Fiber Reinforced Refractory Castables (FRRC) are developed as an emerging material solution to manufacture SPF tools (Nazaret et all, 2004a)). Benefits in using FRRC are related to the short lead-time and to the use of low cost raw materials. Non-reinforced refractory castables keep good mechanical properties until 1000°C and are characterized by a good thermal shock resistance because of their low thermal expansion coefficient and low elasticity modulus. Unfortunately, their mechanical behaviour remains quasi-brittle. Their reinforcement with stainless steel fibers enables to give some ductility to the resulting FRRC and avoid a catastrophic failure after crack initiation (Bernhart et all, 2007).

Nevertheless refractory concretes are heterogeneous materials. Their microstructures are indeed mainly constituted of a cement-based matrix, of aggregates, of porosity and microcracks. In some cases, adding short metallic fibres create a new heterogeneity degree in these materials. Such heterogeneities confer specific behaviours to refractory concretes both from the microstructural and from the mechanical points of view. The temperature level and the thermal history have a main influence on their behaviours. As an example, large microstructure evolutions take place in refractory concretes during their first heating. They are due to multiple and complex mechanisms that are strongly related to the material formulation and they often have large consequences on the mechanical behaviour.

This paper gives first an overview of several studies that have been conducted over the years on several oxide based refractory ceramics and concretes. Main topics are related to the characterisation both of the microstructural evolutions and of the mechanical behaviours as a function of the thermal history of the material. A particular attention is paid to the relationships between these topics particularly concerning damage processes by considering a large panel of scales: from the microscopic to the macroscopic ones.

Than macroscopic engineering approaches are described in order to be able to design and justify SPF forming tools. Based on constitutive laws that allow modeling of the particular damage behaviour of FRRC’s numerical simulations by finite elements are shown. The capability of FRRC’s to sustain high thermal gradient levels was validated in industrial shop production environment.

7 º C O N G R E S S O B R A S I L E I R O D E E N G E N H A R IA D E F A B R IC A Ç Ã O 2 0 a 2 4 d e M a i o d e 2 0 1 3 . P e n e d o , I t a t ia ia – R J


2. MATERIALS Three refractory concretes were successively investigated and are essentially considered in the present paper. One is

a geopolymer based refractory concrete that is reinforced with 1.5 vol.% of metallic fibres (FRRC: Fibre Reinforced Refractory Concrete). It is made of a geopolymer-based matrix (in the SiO2-Al2O3-K2O ternary system) and of cordierite aggregates (Cailleux et all, 2005). The matrix is obtained by mixing a major alumino-silicate oxide, thermal silica fume and an aqueous solution of potassium polysilicate. The maximum aggregate size close to 3mm. Metallic fibres are made of an AISI 310 stainless steel. They have been processed by cold drawing and are characterized by a 0.38mm diameter and by a 12.5mm length. The FRRC refractory concrete is shaped by mixing these components in a planetary mixer and by casting the mix under vibrations. Complete polymerisation of the geopolymer matrix is obtained after an isothermal heat treatment of 12 hours at 80°C.

The second material (And-LCC) is a commercial grade of an andalusite based refractory concrete (Marzagui and Cutard, 2005 , Marzagui et all, 2004). It is made of andalusite aggregates with a 5mm maximum size, of silica fume, of alpha-alumina and of a low content of a calcium aluminate cement. Two fiber shapes are considered : on one hand straight fibers characterized by a 12.5mm length and by a 0,38mm diameter, on the other hand hook-end fibers characterized by a 25mm length and by a 0,38mm diameter too. Samples are prepared by mixing the raw materials in a planetary mixer with a 5wt% water addition. The mix is then cast in the moulds. Moulds and samples are then immediately wrapped in plastic. They are cured at room temperature during 24h and then extracted from the moulds before a 110°C-24h drying step.

The third material (Bau-ULCC) is a commercial grade of a bauxite based refractory concrete (Marzagui and Cutard, 2005 , Marzagui et all, 2004). It is made of bauxite aggregates with a 5mm maximum size, of silica fume, of alpha-alumina and of an ultra-low content of a calcium aluminate cement. The same processing route was retained both for the Bau-ULCC and for the And-LCC refractory concretes

When fired before performing microstructural or mechanical tests, the heating and cooling ramp rates of the firing cycles were of 60°C/min for the three materials. Isothermal dwells at the maximum temperature of the firing cycle were of a 5 hours duration.

Generic steps for FRCC processing and microstructure is shown in figure 1.

Fig 1a: Processing route of geopolymer based FRRC

(Fiber Reinforced Refractory Concrete) Fig 1b: FRRC microstructure

Characterization of the concretes was performed at the material level using X-ray diffraction (XRD), conventional

(SEM) and high temperature environmental scanning electron microscopy (HT in-situ ESEM) and at the mechanical level mainly with High Temperature four point bending tests.

In the later case the nominal strength is calculated at the peak load value (P) according to the material resistance theory :

( )22

3

bh

lLP −=σ (1)

3. MICROSTRUCTURAL BEHAVIOUR

Refractory concretes are based on the association of various components and phases. At high temperature, the

resulting materials are unstable. Heating such materials can result in solid and/or liquid state phase changes, in crystallisation phenomena, in the activation of diffusion processes and/or of sintering mechanisms. Furthermore, because of the large contrasts that exist between the dilatometric, mechanical and physical behaviours of each component, heating or cooling the refractory concrete induces internal stress fields. They often play an important role in microstructural evolutions too.

Mixing

Aggregatescordierite

grog

Powder precursoralumino-silicate

Liquid precursorpotassium

silicate

Fibrescold drawn AISI310

stainless steel

Castingunder vibration

80°C drying

GEOPOLYMER binder phase

Powder precursor :metakaolinite

Liquid precursor :potassium silicate

alkali- reactionat 80°C

Amorphous phaseK(AlSi3)O8

Orthose feldspar

Mixing

Aggregatescordierite

grog

Powder precursoralumino-silicate

Liquid precursorpotassium

silicate

Fibrescold drawn AISI310

stainless steel

Castingunder vibration

80°C drying

GEOPOLYMER binder phase

Powder precursor :metakaolinite

Liquid precursor :potassium silicate

alkali- reactionat 80°C

Amorphous phaseK(AlSi3)O8

Orthose feldspar

Fibre

Refractory castable

Fibre

Refractory castable



During the first thermal cycle, and particularly during the first heating, microstructural transformations take place in the matrix of refractory concretes.

Figure 2 gives some results in the case of the matrix of the geopolymer based refractory concrete. The dilatometric curve is highly non-linear. This indicates that microstructural changes take place in this matrix during the first heating. Four temperature domains can be defined to describe this dilatometric behaviour. After the 80°C polymerisation step, the matrix is made of amorphous K(AlSi2)O8 nanoparticles. The XRD pattern exhibits on one hand the diffraction peaks of the fine cordierite particles and a highly broadened peak that is centred on the theoretical position of the maximum intensity peak of the orthoclase feldspar. When increasing the firing temperature, evolutions of the diffraction patterns confirm that microstructural changes take place in the matrix. They mainly deal with dehydration processes in domains 1 and 2. In domain 3, solid state sintering occurs in the nanoparticle network. From 800°C, vitreous bridges are clearly observed on SEM micrographs and liquid phase sintering processes are active. Diffraction peaks of leucite, of quartz and then of tridymite progressively appear and indicate the progressive crystallisation of nanoparticles.

a)

Fig. 2 : Effects of the first firing on the microstructure of the matrix of

a geopolymer based refractory concrete:

a - dilatometric behaviour during the first heating,

b - SEM micrographs, c - XRD patterns.

110°C dried

800°C fired

b) c)

The dilatometric behaviours of the And-LCC and Bau-ULCC materials, appear also highly non-linear and four temperature domains could be defined too (figure 3a) (Cutard, 2007). For these materials, the matrix chemical formulations are in the CaO-Al2O3- SiO2 (CAS) ternary system. Hydration processes control the development of matrix/aggregate bondings. They mainly consist in the formation of AH3 and C3AH6. XRD tests allow to characterise and to identify the physico-chemical transformations that occur in these matrices as a function of the firing temperature [(Marzagui and Cutard, 2005 , Marzagui et all, 2004). They mainly deal with dehydration phenomena, hydrate conversion, crystallisation of CA and CA2, formation of anorthite (CAS2), of mullite (A3S2) and of cristobalite (SiO2).

a) b) c) Fig. 1 : Examples of microstructural evolutions of the high alumina cement based refractory concretes :

a – matrix dilatometric behaviours both for the andalusite based and for the bauxite based materials during the first heating,

b - high magnification SEM micrographs in areas characterised by high silica fume contents in the bauxite based material (room temperature observations).



4. MICROMECHANICAL BEHAVIOUR As detailed before, the large contrasts that exist between the physical and mechanical properties of the various

components of a refractory concrete often lead to the development of internal stress fields. This is particularly true when considering the consequences of the dilatometric behaviour differences.

Firstly, considering unreinforced refractory ceramics, a generic set of problems deal with the aggregate/matrix association. During the first heating of a refractory concrete, it can be considered that aggregates are mainly characterised by an expansion behaviour whereas the matrix is characterised by large shrinkage domains. The induced internal stress fields often result in damage processes in the refractory concrete microstructure. Only subjecting the microstructure to a thermal cycle leads to damage.

Secondly, adding metallic fibres creates another heterogeneity degree in these materials. A metallic component, characterised by a high expansion coefficient compared to ceramic ones, is so included in the microstructure. When subjecting FRRC’s to thermal cycles, specific internal stress fields are induced. They can be analytically estimated. They result in specific damage mechanisms at fibre/concrete interfaces. As shown in figure 6a, they deal with radial matrix microcracking and with fibre/concrete debonding and opening. The presented micrographs have been obtained after subjecting the FRRC to a first thermal cycle. Heating again the FRRC will lead to a closing mechanism of fibre/concrete (F/C) debondings and to the development of a fretting pressure at F/C interfaces. The opening/closing mechanisms that occur at these interfaces could clearly be observed during high temperature in-situ ESEM observations (figure 4b) (Nazaret et all, 2004b). Fig. 4 : Evolutions of the

fibre/refractory concrete interfaces of a geopolymer based refractory concrete reinforced with short metallic fibres:

a- after a 500°C firing (room

temperature observations), b- during the cooling step of a

800°C firing cycle (high temperature observations).

a) b)

5. MACROSCOPIC THERMOMECHANICAL BEHAVIOUR Main characteristics of the FRRC thermomechanical behaviour have already been described in previous papers

concerning tensile and compression tests (Bernhart et all, 2007, Nazaret 2005). The behaviour is quite similar for four points bending tests. Firstly, the behaviour is linear elastic. When the stress level increases, the behaviour moves then to a non-linear one due to damage processes by microcracking. At higher stress levels, damage localizes and one or several macrocrack appear. After the stress peak, the behaviour curve exhibits an extended softening part. Indeed, stainless steel fibers bridge macrocracks, allowing the FRRC to sustain both high stress and strain levels. Fiber reinforcement induces a quite linear , perfectly plastic behaviour whatever the testing temperature between room temperature and 700°C, i.e. the material is no more brittle.

In the case of the FRRC reinforced by straight fibers, the thermomechanical behaviour evolutions can be observed on fig 5 for various testing temperatures and for two grades of FRRC (Geopolymer based ((5a) and Alumina based ((5b)). From 20°C to 700°C, the behaviour remains quite stable with the three previously described domains. The peak strength is approximately 100% higher for the later grade FRRC. At 900°C, the peak strain and the peak strength both increase. A viscoplastic component characterizes the behaviour at this temperature level.

200°C

fiberoxide layer

concrete

800°C

High temperature in-situ ESEM

200°C200°C

fiberoxide layer

concrete

800°C

fiberoxide layer

concrete

800°C

High temperature in-situ ESEM

fibre/ concrete debonding

20µm

fibre

concrete

radial microcraking

fibre

100µm

concrete


20µm

fibre

concrete


20µm

fibre

concrete

radial microcraking

fibre

100µm

concrete

radial microcraking

fibre

100µm

concrete



Fig 5a : High temperature four point bending stress for Geopolymer based FRRC

Fig 5b : High temperature four point bending stress for High alumina content based FRRC

Figure 6 shows the comparison between the thermomechanical behaviour curves of the high alumina And-LCC

castable reinforced with hook-end fibers compared to the straight fibers. The general behaviour characteristics and evolutions are quite the same than for the case of straight fibers. Nevertheless, two differences can be noticed. On one hand, an average 20% increase of the peak strength is observed when considering the hook-end reinforced material. On the other hand, the post-peak softening branch is characterized by higher stress levels in the case of hook-end fibers too. This results clearly indicate that the fiber shape has a strong influence on the FRRC thermomechanical properties. At 900°C, the peak strain and the peak strength both increase, nevertheless a viscoplastic component characterizes the behaviour at this temperature level.

Fig 6a : Four point bending test behaviour of the 900°C

fired FRRC And-LCC castable with straight fibers. Fig 6b : Four point bending test behaviour of the

900°C fired FRRC And-LCC castable with hook-end fibers.

Thermal thermal history related to the firing has also an impact on mechanical behaviour of FRRC. As an example,

figure 7 shows the difference between the damage behaviour of two bending beams at 20°C. The first one is tested in the non-fired state (figure 7a) and the second one is tested after a 500°C firing. The strain fields that are presented have been determined by using a 3D stereocorrelation method (Bernhart et all, 2007). For the non-fired FRRC, only one strain localisation area is observed and results in the initiation of a single macrocrack. For the 500°C fired FRRC, because of the damage that has been created during the firing cycle, the resulting non-linear behaviour allows multiple crack initiation sites and a crack redistribution phenomenon. As a consequence, several macrocracks initiate and propagate in the beam. Complementary results have confirmed this behaviour type in the case of more complete loaded structures (Bernhart et all (2007), Nazaret ,2005)

Because of their heterogeneity and of their quasi-brittle behaviour, refractory concretes may be characterised by a scale effect. To investigate the effect, three points bending tests have been performed at room temperature and on five sizes of unnotched homothetic beams (cf. figure8a) made of the geopolymer based FRRC fired at 500°C. The five beam heights (from D1 to D5) are respectively equal to 12.5, 25, 50, 100, 200mm and the distance between the lower loading rollers corresponds to 4.5×Di. For the five beams, the thickness has been kept constant to a value of 100mm.

0

5

10

15

20

0 0,2 0,4 0,6 0,8 1

Deflection (mm)

Str

ess

(MP

a)

Geopolymer FRRC material

20°C 300°C 500°C 700°C

0

5

10

15

20

0 0,2 0,4 0,6 0,8 1

Deflection (mm)

Str

ess

(MP

a)


0

5

10

15

20

0 0,2 0,4 0,6 0,8 1

Deflection (mm)

Str

ess

(MP

a)


20°C 300°C 500°C 700°C

0

5

10

15

20

0 0,2 0,4 0,6 0,8 1

Deflection (mm)

Str

ess

(MP

a)

High alumina FRRC material

20°C 300°C 500°C 700°C

0

5

10

15

20

0 0,2 0,4 0,6 0,8 1

Deflection (mm)

Str

ess

(MP

a)


0

5

10

15

20

0 0,2 0,4 0,6 0,8 1

Deflection (mm)

Str

ess

(MP

a)


20°C 300°C 500°C 700°C

0

5

10

15

20

25

0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6

Deflection (mm)

Str

ess

(MP

a)

700°C

500°C

300°C

20°C900°C

0

5

10

15

20

25

0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6

Deflection (mm)

Str

ess

(MP

a)

700°C

500°C

300°C

20°C 900°C



Fig. 7 : Effect of a firing cycle on the damage behaviour of the

geopolymer based FRRC in the case of a four points bending beam

(strain fields determined by using a digital image stereocorrelation

method): a- non-fired FRRC,

b- 500°C fired FRRC.

a) b) Figure 8b shows that the considered FRRC is highly sensitive to the scale effect. The peak stress level decreases of

a factor close to 2 when increasing the beam size from D1 to D5.

Fig. 8 : FRRC size effect characterised in three points

bending : a- the five considered sizes of FRRC

bending beams, b- rupture stress at room

temperature as a function of the beam size and in the case of the

FRRC fired at 500°C (three points bending tests).

a) b)

Various types of scale effect laws have been considered to model the case of the geopolymer based FRRC. Most of

them have been developed to model the scale effects of civil engineering concretes. They deal with the deterministic and deterministic/statistic laws developed on one hand for unnotched beams and on the other hand for notched beams. To model the scale effect behaviour of the considered FRRC, a two transitions law has been proposed (Nazaret, 2005). It links the deterministic laws of notched and unnotched beams. Coupling these two types of law has been suggested by the observation of the damage modes that characterise the studied FRRC. For the smallest beam sizes, multicracking occurs in a band located in the lower part of the beam, mainly subjected to tensile stresses. The damage process could be identified when considering the strain fields obtained from the stereocorrelation method in the case of the 500°C fired FRRC. In that case, the deterministic laws developed to describe the damage behaviour of unnocthed beams are well suited. For the largest sizes of the FRRC beams, when the height of the multicracked band is low compared to the beam height, even if the beam behaviour is not brittle, when a macrocrack initiates this one propagates like in the case of a notched beam. As a consequence, the deterministic laws developed for notched beams allow a better description of the scale effect of the geopolymer based FRRC

6. NUMERICAL SIMULATION

Numerical simulations by finite elements have been performed to evaluate the capability of the FRRC to sustain

SPF processing constrains (Nazaret, 2005, Bernhart et all, 2008). A prototype SPF tool with a 950mm length and a 650mm width was considered. It corresponds to an FRRC insert that is placed in a metallic tank. A meshing with tetrahedral elements supporting quadratic interpolation has been performed. Simulations shown here after were performed using a Mazars (Mazars J. and Pijaudier-Cabot G., 1989) damage model with a commercial FEM code. The material is assumed to behave elastically and to remain isotropic. If Λ0 is the initial stiffness matrix of the material and

D the scalar damage, that ranges from 0 for the virgin material to 1 for the failed material, the following behaviour law can be considered :

σ = Λ0 1− D( ) : ε e ( 2 )

Dt and Dc are respectively the tensile and compressive damage parameters. They allow to take into account the non-symmetric behaviour of the FRRC. The total damage D is a weighted sum of Dt and Dc . Their evolution is given by functions Fi depending on material parameters and equivalent strain.

D = α tDt + αc Dc ; Dt = Ft˜ ε ( ) ; Dc = Fc

˜ ε ( ) ( 3 )

The use of a continuum damage model with softening leads to strain localization. To avoid such a numerical problem, a regularization of equivalent strain by a non local method has been used introducing an internal length Lc.

all experimental datamean value

stre

ss (

MP

a)



This internal length defines an area on which the local equivalent strain will be averaged with equivalent strain of the neighboring weighted by a function. The non local equivalent strain is calculated by considering a gradient formulation

In a first time, considering a purely elastic behaviour, a thermal simulation has been performed to determine the

temperature field after an extraction of the tool from the SPF press. As initial conditions, temperature is fixed to 900°C in all points of the tool. Air convection and radiative transfer are imposed as boundary conditions on the upper face of the tool. There are no boundary conditions on the other faces that are protected by the walls of the metallic chamber. The temperature evolution is used to calculate the stress levels that are induced by thermal gradients. A result example is shown in Fig 9 at a time of 466s after the extraction from the SPF press. The maximum principal stress reaches a level of 44 MPa. On all the upper face, quite homogeneous stress levels close to 25 MPa are obtained. Such stress levels are very high when compared to the FRRC mechanical properties.

Fig. 9 : Maximum principal stress field in the SPF tool

at a time of 466s after the extraction from the press.

Fig. 10 : Maximum principal stress field with a 15 bars

pressure applied on the upper face

Then a simulation of stresses generated by the sheet forming step has been performed too. An homogeneous 900°C

temperature field has been taking into account for the SPF tool. A typical industrial pressure of 15 bars (1,5 Mpa) has been applied on the upper face. Tool clamping is not considered since the FRRC tool is placed in a metallic tank that supports clamping loads. Under such conditions, the maximal principal stress reaches 17.6 MPa in the FFRC tool (Fig 10).

These results that have been obtained by considering an elastic behaviour have shown that stresses generated by thermal gradient are more critical than stresses induced by the sheet forming step. This result is in agreement with investigations performed on SPF steel dies (Gao et all, 2005).

Fig. 11 : Damage field generated by thermal gradients at a time of 30s after the tool extraction from the SPF

press

Fig. 12 : Damage field generated by thermal gradients at a time of 300s after the tool extraction from the SPF

press

In a second time, temperature fields calculated during the step of the tool extraction from the press have been used to evaluate damage and stress levels in the tool. For such simulations, the Mazars model with a non-local regularization has been chosen as the constitutive law for the FRRC. The damage criteria is attained at a time of 30 s after the extraction of the tool from the SPF press. Damage appears first at two corners of the FFRC tool (Fig 11). After an



extraction time of 300s, damage has propagated on all the upper face (Fig 12). The maximum damage value is close to 0.7 and an average value of 0.4 is obtained. On the upper face, the average of the maximal principal stress is equal to 14MPa and is quite homogeneous. This simulation shows that the FRRC sustains thermal shocks induced by the SPF process because of developing damage. Fiber reinforcement enables a smeared damage on all the upper face of the tool and avoids the damage localization and/or crack propagation phenomena.

7. CONCLUSION The thermomechanical behaviour of Fiber Reinforced Refractory Concretes (FRRC) have been discussed in the

20°C to 900°C temperature range. Because of their heterogeneity, refractory concretes are characterised by complex microstructural and mechanical behaviours, particularly at high temperature. Three oxide based refractory concretes have been considered in the present paper, and short metallic and aluminia fibre reinforcements. Microstructural evolutions have been considered both from the physico-chemical and from the micromechanical points of view. It has been shown that the firing duration has a great influence on the 900°C behaviour. The benefits of the reinforcement was clearly shown and the non brittle behaviour was demonstrated at all temperatures. Such material solution are now mature for industrial die production.

Moreover material models have been developed for in service simulation purposes. These simulations show that surface damages appear without major crack. This result has been verified by industrial forming tests in the case of TA6V sheets. The FRRC tool remained intact without any macrocracking and without any sticking between the sheet and the tool. FRRC appears as a promising innovative material solution to manufacture SPF tools. Raw materials are cheap and the manufacturing process is an economic and rapid one. It’s possible to mold directly the FRRC tool to its final geometry without any machining step and with a good surface quality.

8. ACKNOWLEDGEMENTS

Results considered in this paper have mainly been obtained by PhD students (E.Cailleux, H.Marzagui and

F.Nazaret): all of them must be greatly acknowledged. The author, wishes also to thank the Ministry of Industry and various companies (Airbus, Aircelle, ACB, Pyromeral Systems, Snecma, TRB, Vesuvius, Aurock) for their technical and financial support in the different programs where the considered materials were studied.

9. REFERENCES Bernhart G.,Nazaret F. and Cutard T., 2007, Fibre reinforced refractory castables: an alternative solution for SPF die

manufacturing, Materials Science Forum Vol.551-552 (2007), p.37-42 Bernhart, G., Nazaret, F.and Cutard T., 2008, T., Fiber reinforced refractory castables for SPF toolings,

Materialwissenschaft Und Werkstofftechnik, vol. 39, n°4-5, p. 317-321, 2008 Cutard, T.,2007, Study of refractory concretes at high temperature: microstructural evolutions and mechanical

behaviour, 10th International Conference of the European Ceramic Society (ECERS’2007), Berlin (Germany), 17-21 June.

Cailleux E.,Cutard T.,Bernhart G., 2005, Study of a ceramic refractory reinforced with metallic fibers : from the microstructure to the mechanical behaiviours. Industrial ceramics, Vol.25, n°1, pp21-26.

Gao C.Y., Lours P. and Bernhart G., 2005, Journal of materials processing technology Vol.169, p.281-291. Marzagui H,, .Cutard T. 2004, Characterisation of microstructural evolutions in refractory castables by in situ high

temperature ESEM, Journal of Materials and Processing Technology, vol.155-156, pp1474-1481. Marzagui H., Cutard T., Roosefid M., Ouedraogo E., Prompt N., Deteuf C.,2004, Room temperature mechanical

behaviour of two refractory castables, Proceedings of the International Conference of Metallurgists COM2004, Hamilton (Canada), pp 645-659.

Nazaret F., Cutard T. and G. Bernhart G., 2004a, Euro SPF 04 Albi (France),p.139-144 Nazaret F., Cutard T., Bernhart G.,2004b , Thermomechanical behaviour of a fibre-reinforced refractory concrete: tests

and analysis, Sixth RILEM Symposium on Fibre-Reinforced Concretes, Varenna (Italie). Nazaret F., 2005, Caractérisation et modélisation du comportement thermomécanique d'un béton réfractaire renforcé de

fibres métalliques, PhD thesis (in French), Ecole des Mines de Paris (France). Mazars J. and Pijaudier-Cabot G., 1989, Journal of Engineering Mechanics Vol.115, p.345-365. 10. RESPONSIBILITY NOTICE

The authors are the only responsible for the printed material included in this paper.

low cost refractory castable die development for …

Documents