EP0184949B1 - Process for the rejuvenation of nickel-based superalloy articles being at the end of their service life - Google Patents

Description

The invention relates to a method of heat treatment for parts arriving at the end of operating potential after having undergone damage by creep in particular; the aim of the method is to make them recover their initial properties in order to prolong their lifespan. It relates to parts made of heat resistant alloy with a nickel base comprising a hardening phase -y 'and applies in particular to the moving blades of a turbomachine.

The blades must be able to resist creep at high temperature because they are mounted on a disc rotating between 5,000 and 20,000 rpm while being exposed to hot gases from 900 ° C to 1300 ° C and oxidants leaving the combustion. We therefore turned to cast alloys, allowing the optimization of their chemical composition and susceptible to significant hardening by precipitation in order to improve the resistance to rupture by creep. The nickel-based superalloys used in aeronautics have a hardening phase -y 'whose volume fraction can reach 70%.

However, during operation, the blades subjected to such mechanical and thermal stresses undergo permanent elongation by creep which inevitably leads to their systematic scrapping after a certain number of hours of use in order to avoid the risks of catastrophic rupture. For example, the high pressure turbine blades of a certain number of engines currently have their operating potential limited to about 800 hours by creep.

This creep deformation process resulting in a degradation of the microcrystalline structure the object of the invention is the production of a heat treatment method allowing the restoration of the initial structure under conditions compatible with the geometrical criteria of the parts.

These alloys designed for use at high temperature exhibit poor corrosion resistance above 900 ° C., in particular in a sulphurous atmosphere; they therefore require surface protection which may be a coating of nickel aluminide obtained thermochemically. The problem posed by this type of protection is that a heat treatment of the part beyond a certain temperature and a certain duration results in intermetallic diffusion modifying its chemical composition and its properties. To avoid this, it is normally sufficient to have a prior treatment for removing this layer. However, this operation appeared to be impossible on turbine blades provided with internal cooling channels because it would prohibitively reduce their thickness of already thin walls.

The invention therefore has the second objective of carrying out a heat treatment which does not require the prior operation of removing the protective layer. In accordance with the invention, the method of regenerating parts of heat-resistant alloy based on nickel NK15 CAT comprising in particular Co 13 to 17%, Cr 8 to 11%, Al 5 to 6%, Ti 4 to 5%, Mo 2 at 4%, Va 0.7 to 1.7%, C 0.1 to 0.2% and forming an eutectic comprising a hardening phase y`, the part having consumed at least part of its operating potential due in particular to creep damage at high temperature, consists in maintaining the part at a temperature between 1160 ° C and 1220 ° C and for a period between 1 h and 4 h to re-dissolve at least 50% of the phase γ 'the method consists in cooling the part at a controlled speed to a temperature below 700 ° C corresponding to the precipitation range of the phase, this speed being chosen as a function of the desired microstructural morphology.

During previous work, regeneration treatments have already been developed. For example, patent FR 2 292 049 describes a method for extending the duration of the secondary creep of certain alloys; it consists of an unconstrained heat treatment, carried out at a temperature lower than that of dissolution of the compounds. This temperature corresponds in practice to the maximum operating temperature of the room; moreover, maintaining the temperature is quite long because, according to the hypothesis put forward, it should allow the annihilation of the lacunar networks by a diffusion process. This treatment, limited in temperature, is certainly ineffective for parts which have operated at high temperatures, such as 1100 ° C., since it does not allow the regeneration of the microcrystalline structure because it excludes the re-solution of the hardening compounds. . In addition, its duration makes it economically uninteresting in an industrial application.

The document "Rejuvenation of turbine blade material by thermal treatment", Final Report for Air Force Wright Patterson AFB, ohio 4533 of July 1979 provides a method of regenerating blades after creep comprising a first heat treatment at a temperature less than or equal to 1149 ° C for 2 to 4 h followed by tempering at a temperature below 927 ° for 10 to 12 h and then air cooling.

Patent FR 2 313 459 relates to a method for improving the service performance of metal parts which have undergone permanent elongation. It consists in subjecting these parts, before the appearance of surface cracks, to hot isostatic compression, at a temperature lower than that where a magnification of the grains occurs, then to apply a treatment of re-solution of the phases followed by 'a hardening income. The major advantage of compaction lies in the fact that it closes the creep decohesions and the unclogged foundry pores. chants.

This technique is however quite cumbersome to implement, it is not justified in all cases. In addition, the following heat treatment does not allow control of the precipitation mechanisms; it also does not take into account a deterioration of the protective layer on the surface; finally it does not allow an economical industrial application.

• - Les figures 1 et 1A sont des microphotographies réalisées au microscope électronique d'une aube après 50 heures de fonctionnement sur moteur.
• - Les figures 2 et 2A sont des microphotographies analogues aux précédentes pour une aube ayant fonctionné 800 heures.
• - Les figures 3 et 4 sont des microphotographies révélant l'aspect des dislocations d'interface γ-γ' après 800 heures de fonctionnement.
• - Les figures 5A à D donnent une représentation schématique du processus d'endommagement par fluage.
• - La figure 6 montre l'évolution microstructu- raie de l'alliage en fonction de la vitesse de refroidissement après un maintien à 1190 °C pendant 1 heure sous vide.
• - Les figures 7, 8 et 9 montrent l'effet microstructural du traitement de régénération : la figure 7 est une microphotographie d'une aube neuve, la figure 8 d'une aube ayant fonctionné 1000 heures et la figure 9 d'une aube régénérée après 1000 heures de fonctionnement.
• - la figure 10 représente dans un repère temps-allongement le comportement en fluage d'une éprouvette respectivement sans régénération et avec régénération à 0,5 % d'allongement.

The following description will better understand the invention and its advantages over the prior art. It refers to the alloy with the trade name IN 100 but it will be understood that the method is more general and its scope is not limited to this alloy.

• - Figures 1 and 1A are photomicrographs taken with an electron microscope of a blade after 50 hours of operation on the engine.
• - Figures 2 and 2A are microphotographs similar to the previous ones for a dawn having operated 800 hours.
• - Figures 3 and 4 are photomicrographs revealing the appearance of γ-γ 'interface dislocations after 800 hours of operation.
• - Figures 5A to D give a schematic representation of the damage process by creep.
• - Figure 6 shows the microstructural evolution of the alloy as a function of the cooling rate after holding at 1190 ° C for 1 hour under vacuum.
• - Figures 7, 8 and 9 show the microstructural effect of the regeneration treatment: Figure 7 is a photomicrograph of a new blade, Figure 8 of a blade having operated for 1000 hours and Figure 9 of a regenerated blade after 1000 hours of operation.
• - Figure 10 shows in a time-elongation mark the creep behavior of a test piece respectively without regeneration and with regeneration at 0.5% elongation.

The IN 100 alloy of formula NK 15 CAT is a nickel-base cast alloy. Its composition is as follows: Cobalt 13 to 17%, Chromium 8 to 11%, aluminum 5 to 6%, titanium 4 to 5%, molybdenum 2 to 4%, vanadium 0.7 to 1.7%, Carbon 0.1 at 0.2% etc ...

Poured under vacuum at 1460 ° C, the IN 100 is designed for long-term use at 1000 ° C and 1100 ° C for short periods. In all cases, its poor resistance to corrosion, in particular in a sulfurous atmosphere, requires protection, obtained for example by the vapor phase aluminization method of patent FR 1 433 497.

From a microstructural point of view, the IN 100 has a γ-γ 'dendritic structure decorated with eutectic aggregates and carbides. The size of the basalt grain dendrites and the morphology of the hardening phase depend on the rate of cooling on casting, therefore on the local thickness of material in the part, and on the content of B and Zr. It varies from a few tenths to several mm for thicknesses ranging from 1 to 10 mm.

The matrix γ, hardened by the effect of a solid solution of Cr and Co in Ni crystallizes in the CFC system. The maximum hardening comes from the precipitation of the -y 'phase, ordered, of type L1 ₂ (Cu ₃ Au) of the same system. crystal clear and consistent with the matrix. Its volume fraction is around 70%. The approximate composition is (Ni, Co) 3 (Ti, AI). The exceptional mechanical resistance when hot conferred by γ 'on nickel-based superalloys comes essentially from the flow stress of this phase, which has the remarkable property of increasing when the temperature increases.

When considering the alloys γ-γ ', the variation of the mechanical resistance as a function of the temperature obviously depends on the volume fraction of -y', but also on the morphology of the precipitates, due to the type of obstacle to the movement of the dislocations they represent.

Furthermore, the alloy is rich in γ-γ 'eutectic islands, located in interdendritic spaces. The temperature of formation of these aggregates is linked to their chemistry during the passage of the solidus, and can vary within wide proportions. The thermal analysis places it between 1210 and 1275 ° C depending in particular on the carbon content.

Two types of carbides are observed in lN 100. Primary carbides of type MC, rich in Ti or Ti-Mo, without orientation relationship with the matrix, appearing well before the end of solidification of the alloy. Secondary carbides, type M 23 C6 rich in Cr and in orientation relationship with the matrix, precipitating at lower temperature between 850 and 1000 ° C.

Experiments were carried out on aluminized blades of a high pressure turbine of an IN 100 alloy aeronautical turbomachine, comprising internal channels for the passage of refrigerant air. It will be recalled that the principle of aluminization is to maintain the part at a temperature above 1000 ° C. in an atmosphere of aluminum fluoride; on contact with the part, the gas dissociates into atomic aluminum on the surface and into gaseous fluorine which maintains the reaction. AL combines with the nickel in the part to form the aluminide which gives it its oxidation resistance properties.

Microstructural observations were made on these blades in new condition and then successively on blades having operated 50 h, 800 h and 1000 h. The operating conditions correspond approximately to a constraint of 130 MPa and a temperature of 1000 ° C.

The new dawn presents at the leading edge as at the trailing edge a γ-γ 'structure rich in eutectics and primary carbides. Two populations of precipitates -y 'co-exist: y' “coarse” of size close to 2 µm precipitating soon after the solidification of the alloy, and -y '“fine”, of size close to 0.2 wm precipitating during cooling following the protective treatment. In the immediate vicinity of the eutectics, only the end γ 'is present. Primary carbides precipitating while the alloy is not fully solidified, are repelled in the interdendritic sites where the grain boundaries are located, which are essentially distinguished by the difference in orientation of the γ 'between 2 contiguous grains.

For blades having operated from 50 to 800 hours, the first microstructural evolution observed consists in the precipitation of secondary intergranular carbides, around the primary carbides and at the γ-γ 'interfaces of the eutectics, after 50 h of operation (Figures 1 and 1A) . For increasing operating times, the precipitation intensifies to become intragranular. At the same time, phenomena of coalescence of the γ 'phase lead to the gradual disappearance of the fine precipitated γ'.

After 800 h of operation, the size of the globules -y 'reaches 3 to 4 μm and can double in the vicinity of eutectics, primary carbides and grain boundaries (Figures 2 and 2A).

The examinations on thin blade show a particular arrangement of the dislocations of interface γ-γ 'and M23 C6 - y: tendency to an arrangement either parallel to the stress of centrifugal origin (figure 3), or in polygonization (figure 4) .

For blades having operated for 1000 hours, the microstructure at the leading edge in the middle of the blade has a dendritic appearance. The interdendritic spaces are rich in eutectics and consist of precipitates γ 'significantly larger than in the heart of the dendrites. The geometry of certain foundry pores reveals a beginning of deformation, as already observed after 800 hours; the coalescence of the phase there causes the disappearance of the precipitated ends.

• - coalescence du γ'
• - orientation des dislocations d'interface γ-γ' parallèlement à la contrainte centrifuge et polygonisation sur certains globules
• - réseau dense et régulier de dislocations d'interface M23 C6-γ' ou M23 C6-y
• - pas d'ancrages des dislocations dans la matrice γ.

Observations in transmission electron micrographs confirm observations made after 800 hours of operation, namely:

• - coalescence of γ '
• - orientation of the γ-γ 'interface dislocations parallel to the centrifugal constraint and polygonization on certain globules
• - dense and regular network of interface dislocations M23 C6-γ 'or M23 C6-y
• - no anchoring of dislocations in the matrix γ.

FIGS. 5A to D give in summary a schematic representation of the process of damage by creep of the alloy subjected to a stress of 130 MPa and a temperature of 1000 ° C., in particular observed on test pieces.

FIG. 5A shows the state of the structure after aluminization, there are 3 populations of γ ': relatively coarse particles of interdendritic γ', fine particles of dendritic γ 'and very fine particles uniformly distributed obtained during cooling after aluminization treatment.

In FIG. 5B after primary creep, the disappearance of the very fine γ 'and the precipitation of secondary carbides are noted.

In FIG. 5C after the start of the secondary creep, we note the oriented coalescence of the dendritic γ '.

In FIG. 5D at the end of secondary creep, the coalescence of the y 'is more marked, it is oriented for the -y' dendritic and not oriented for the interdendritic γ '.

The study of the damage by creep which precedes thus revealed a whole of metallurgical process governing the deformation.

According to the invention, the alloy is subjected to a creep potential regeneration treatment comprising a thermal cycle erasing the microstructural effects of the deformation and leading to a microstructure approaching that of the alloy before stress. The part to be treated, as observed, that is to say after 1000 hours of operation, is placed in an oven, preferably under vacuum in order to overcome oxidation problems. It is heated to a chosen temperature to re-dissolve a sufficient volume fraction of the hardening phase. In the present case of IN 100 alloy vanes protected by aluminization, this temperature is also determined as a function of its compatibility with maintaining the protection; in fact a too high temperature would cause the diffusion of aluminum and the dilution of the layer of nickel aluminide. For the present application, this temperature was chosen at 1190 ° C but may vary depending on the case between 1160 ° C and 1220 ° C. The choice of temperature is also guided by the need for a sufficient margin with the melting temperature of the eutectic for industrial application.

Tests have shown that maintenance of less than 4 hours, preferably of the order of one hour, was sufficient to re-dissolve a volume fraction of γ 'phase of at least 50%, which amounts in particular to destroying the connections between γ 'globules which had developed during damage by creep.

After this maintenance at a temperature of 1190 ° C. for one hour under vacuum, the part was cooled by injecting a flow of inert gas, argon, into the oven. The flow rate was controlled in order to control the cooling rate of the part to a temperature below the precipitation range of the γ 'phase.

It appeared that it was not necessary to control the cooling to room temperature; indeed below 700 ° C, the cooling rate had no influence on the precipitation.

The set of microstructures obtained is represented in FIG. 6. It is observed that the argon coolings lead to the precipitation of two populations of -y ', and that the volume fraction of "large"γ' increases while decreases the content of fines constituents, while decreasing the cooling rate. Microstructural observation reveals a complex phenomenon of "germination-growing" and "growth-coalescence", the respective kinetics of which vary according to the local chemical composition of the matrix giving rise to γ '. There is therefore a compromise between volume fractions of large γ 'and a few allowing the best mechanical behavior to be obtained according to the criteria sought. Indeed, a microstructure consisting only of fine precipitates γ 'is favorable to the creep behavior, but detrimental to the cold and hot ductility of the alloy. In contrast, slow cooling, leading to a microstructure containing only a population of "coarse" would bring no gain to the creep behavior. According to the morphology which one wishes to obtain, one can control the speed between 600 ° C / h and 2500 ° C / h. In the present application, the best choice was between 1085 ° C / h and 1145 ° C / h, the microstructure of which is in Figure 9. Under these conditions, it is no longer possible to differentiate a new blade (Figure 7) d '' a regenerated blade (Figure 9) by only examining their microstructure: distribution of γ-γ 'identical in both cases, absence of secondary carbides, the latter having been dissolved during treatment.

Examination of the effect of the treatment on the protection revealed an increase in its thickness. It is due to the diffusion phenomena involved during the solution treatment. Sulfurizing corrosion tests by sweeping with combustion gases enriched in chlorine and sulfur were carried out in order to compare new aluminized blades with aluminized blades having operated 900 hours and treated according to the method of the invention. After 250 hours, observations make it possible to conclude that the effectiveness of the protection is not altered by the treatment because if the kinetics of corrosion is increased essentially by the diffusion of aluminum in the substrate, it is compensated by an increase of the thickness of the protective deposit.

Tests were also carried out on test pieces in order to characterize them in creep. The IN 100 alloy test pieces underwent: 0.5%, 1% and 3% elongation under a stress of 130 MPa at 1000 ° C; in engine operating equivalent, 1% elongation is equivalent to 800 hours of operation for the above conditions. The test pieces are regenerated and then reassembled in creep. The test results are presented in FIG. 10. It is observed that, under the test conditions, the alloy present after regeneration of the primary and secondary creep stages, the more reduced the greater the pre-deformation.

The maximum gain in treatment is obtained after a pre-deformation of 0.5%. We note that if the time to obtain 1% elongation is 83 ± 10 hours, the time to obtain this same elongation after treatment with 0.5% elongation increases to 103 ± 16 hours, ie a gain of 24% .

The gain is similar on the break time. It is 145 hours normally and goes to 180 hours after regeneration at 0.5% elongation.

These observations make it possible to establish that for the test pieces, the duration of the stationary stage ends shortly before 0.5% elongation and represents the maximum deformation limit for undertaking the regeneration. After 1% elongation, the combined effects of the development of the cavities and the oriented coalescence of the γ 'tend to reduce the effectiveness of the treatment.

The comparison of microstructural observations between test tubes and blades where, for the former, differences in morphology in γ 'dendritic and γ' interdendritic remain after treatment unlike the blades, show that the damage of a blade at the end of potential is less to that of a test piece after 0.5% elongation, which suggests a gain greater than that determined on test piece.

It emerges from the preceding discussion that a blade having consumed its creep potential after 800 hours of operation is regenerated by a heat treatment according to the invention. Comparative examinations on parts and test specimens give hope, taking into account their respective damage processes, of a gain greater than 30% over the service life of the blades.

When the parts have exceeded the secondary creep but they do not show through decohesions, it is possible to combine this treatment with a prior treatment of isostatic hot compaction which is also known per se and which consists in maintaining for 4 hours at 1190 ° C under a pressure at least equal to 1000 bar.

Claims (6)

1. Method of rejuvenating components at the end of their working potential linked in particular to creep damage in gas turbine components in NK15 CAT nickel-based cast alloys notably with the composition of Co 13 to 17 %, Cr 8 to 11 %, AI 5 to 6 %, Ti 4 to 5 %, Mo 2 to 4 %, Va 0.7 to 1.7 %, C 0.1 to 0.2 %, forming a eutectic system γ-γ comprising a hardening phase -y', characterised in the fact that it consists of maintaining the said component at a temperature between 1160 °C and 1220 °C for a period between 1 hr and 4hrs to put back into solution at least 50 % of the volume fraction of the hardening phase V, followed by cooling the component at a controlled rate of temperature drop until the temperature of the component falls to below 700. °C which corresponds to the precipitation zone of the γ phase.

2. A method of rejuvenation in accordance with Claim 1 characterised by a cooling rate of between 600 °C/hr and 2500 °C/hr which is controlled until the temperature of the component is below 700 °C.

3. A method of rejuvenation in accordance with Claim 2 characterised by a cooling rate of between 1085 °C/hr and 1145 °C/hr.

4. A method of rejuvenation in accordance with one of the above Claims for a component which has undergone a corrosion-protection treatment, notably aluminising, characterised in the fact that the solution-restoration temperature is chosen to be lower than the dilution temperature of the said protective coating.

5. A method of rejuvenation in accordance with the preceding Claim characterised in the fact that the solution-restoration temperature lies between 1185 °C and 1195 °C.

6. A method of rejuvenation in accordance with one of the preceding Claims for components displaying sub-surface decohesions characterised in the fact that they are subjected to a preliminary isostatic compaction treatment under heat.

Images