Final Report Summary - DATACAST (Development of a low cost Advanced gamma Titanium Aluminide Casting Technology)
Executive Summary:
Novel gamma-TiAl based intermetallic light-weight high temperature structural materials are recently being used as Low Pressure Turbine blades in aircraft engines.
Depending on the LPT performance (low or high speed), the technical requirements on the materials properties are different. Anyhow, there is a basic requirement for chemical and microstructural homogeneity which exceeds corresponding standards of steel or Ni-based superalloys remarkably. Thus, the production of gamma-TiAl parts is very challenging.
The Project leader, MTU Aero Engines GmbH, has decided to introduce forged TiAl TNM blades in the last stage of the Low Pressure Turbine of Pratt & Whitney`s novel PurePower PW1100G aircraft engine family in order to gain efficiency and reduce specific fuel consumption. The novel TNM alloy has been developed in order to improve the poor wrought processing capability of gamma-TiAl by the addition of the bcc structured beta-Ti phase into the well known alpha2/gamma microstructure of Titanium Aluminides. As a side effect, the microstructure remains fine grained due to a different solidification path of the alloy which is characterized by a complete β-solidification mechanism. It has been demonstrated that the as-cast TNM alloy can be subsequent wrought processed without primary ingot conversion via extrusion technology.
The major objective of the project was to develop an industrial production technology for TNM feed stock parts for forging (forging stocks) offering highest product quality and lowest production costs. Different production routes have been evaluated for the metallurgical alloying (Plasma Arc Melting, Vacuum Arc Melting), the homogenization (Induction Skull Melting, Vacuum Arc Skull Melting) and casting technologies.
Based on experimental work and simulations of the casting and solidification mechanisms the following production technology has been identified to meet the requirements with the best compromise:
• gamma-TiAl ingot manufacturing from Ti-sponge, Al and master alloys via VAR processing
• Ingot conversion and homogenization in a VAR Skull Melter
• Centrifugal casting in permanent moulds
• HIP
• Mechanical machining
The technology has been completely developed and industrialized for the production of
gamma-TiAl TNM forging stocks.
Project Context and Objectives:
Novel gamma-TiAl based TNM alloys have been developed in order to increase wrought processing capability by the addition of the bcc structured beta-Ti phase into the well known alpha2/gamma microstructure of Titanium Aluminides. As a side effect, the microstructure remains fine grained due to a different solidification path of the alloy which is characterized by a complete β-solidification mechanism. It has been demonstrated that the as-cast TNM alloy can be subsequent wrought processed without primary ingot conversion via extrusion technology. Unfortunately, this novel TNM alloy exhibits disadvantages in the metallurgical processing compared to State-of-the-Art intermetallic TiAl alloys auch as TiAl48-2-2.
Thus, the objective of the materials manufacturing technology covers both the development of an industrial stable alloying technology to ingots and the subsequent conversion technology to homogeneous small size parts.
For ingot alloying, PAM and VAR should be evaluated as the most common technologies for Ti alloy ingots containing substantial amounts of low melting elements such as Aluminum. EBM is not applicable due to the severe evaporation of Al during the melting operation.
MM and ACCESS have demonstrated recently that the following two technologies for the conversion of VAR ingots into forging stocks are basically feasible.
a) Remelting of gamma-TiAl VAR ingots in a VAR skull melter combined with gravity die or centrifugal casting in permanent moulds, subsequent HIP and mechanical machining
b) Cutting of gamma-TiAl VAR ingots into segments, remelting segments in a VIM skull melter combined with gravity die or centrifugal casting in permanent moulds, subsequent HIP and mechanical machining
None of both technologies have ever been used in industry before. Both technologies exhibit technical Pro`s and Con`s. Both technologies allow basically the use of non-contaminated recycling materials and prevent substantial materials losses during the entire manufacturing chain except final machining. Both technologies can be basically industrialized.
Neither the melting/casting process nor the solidification mechanism are sufficiently investigated and understood up to now. Additionally, ingot manufacturing via VAR is based on a novel technology which has been developed at MM particularly for TNM alloys in order to overcome the “cracking phenomenon” during the VAR processing. This technology is not yet approved and may require amendments to existing industrial standard procedures. A patent application has been filed on the corresponding procedure (DE 10 2009 050 603.9 and PCT/EP 2010/064 306) by MM. The procedure itself results in absolutely stable VAR processing behaviour.
An additional alternative which needs to be evaluated is the production of remelt stocks using “plasma-arc-melting” (PAM) instead of VAR. Because Plasma Arc Melting will be probably the most promising recycling technology, PAM ingots have to be extensively investigated with regard to their chemical composition, microstructure and homogeneity. Due to the cracking phenomenon, PAM ingots have to be cut in remelt stocks in order to be further processed via VIM skull melting combined with gravity die or centrifugal casting in permanent moulds.
In any case, HIP has to be applied for the closure of any casting porosity.
Subsequent machining of the billets is one of the most important cost drivers. Thus, the casting process should preferably result in a net shape part which does not require surface machining before further processing. The casting process has to be optimized with this regard.
The design of the casting tool (permanent mould, casting wheel) is of remarkable importance of the entire cost structure as well and must address in particular
- assembling and disassembling issues
- repair and maintenance issues
- entire lifetime of the casting tool
The objective of the project is the development of a low cost casting process for γ-TiAl based TNM alloys which guarantees to meet all technical specifications of the products.
Two basic requirements determine the scientific and technology methodology of the project:
→ uniform product properties such as homogeneous microstructure, reproducibility
→ application of robust low cost mass production technologies exhibiting high specific materials yields
As a precondition, all single steps in the production chain must run absolutely stable and repeatable without remarkable materials losses. Therefore, an initial ingot manufacturing technology has to be used which overcomes the cracking behaviour of TNM consumable electrodes during conventional VAR processing and VAR Skull Melter processing as well.
Alternative 1
is based on a novel VAR technology patented by GfE. Remelting in a VAR skull melter offers excellent homogenization due to large batch sizes.
Alternative 2
does replace one initial VAR step by an additional VIM step. Homogenization in the VAR skull melter is comparable to alternative 1. The application of gravity die or centrifugal casting depends on the remelt stock size required in the subsequent VIM process.
Alternative 3
is based on a PAM ingot which cannot be further processed in the VAR skull melter due to the cracking issue. Thus, the separation into feed stock materials for the VIM skull melter must be done by mechanical sawing or similar technologies which causes additional materials losses and may limit homogenization potential.
Project Results:
3. Description of the main S&T results
• Evaluation of TNM ingot manufacturing routes
a) VAR ingots
Since at least one Vacuum Arc Remelting (VAR) step is still required for the production of Ti-alloys used in rotating aircraft engine parts, the VAR process for ingot manufacturing has been evaluated in very detail.
In contrast to conventional TiAl alloys, novel beta-stabilized TiAl alloys such as TNM show a remarkable tendency to crack during VAR processing. It could be demonstrated that this behaviour originates from abnormal thermal expansion processes at the γ-solvus temperature which is much lower in beta-stabilized TiAl alloys compared to the conventional ones. Thus, the capability to remove stresses induced by abnormal thermal expansion by plastic flow is remarkably decreased due to the lower temperature and stress relaxation occurs via cracking.
A novel VAR technology has been developed by GfE MM in order to overcome this “cracking phenomenon”. This technology runs absolutely stable and has been approved meanwhile. A corresponding patent has been filed in 2009 and was issued (EP 10 765 988.0-2122) in the EU recently. All the VAR processing in the Project is based on this patented technology.
The ingots are being used either as consumable electrodes in the VAR skull melter or are being cut (separated) into lumpy pieces as remelt stocks for the VIM skull melter.
In the first option, approximately one third of the ingot (about 40 kg) is homogenized in the VAR skull melter and converted to cylindrical slugs via centrifugal casting in permanent moulds. In the second option, the homogeneization in the VIM skull melter is applied to much smaller parts of about 8 kg mass. In order to get an idea about the homogeneity of a double VAR TNM ingot, one ingot of approximately 170 kg mass has been separated into 85 pieces of about 2 kg mass in average. Every piece was chemically analysed.
The deviation of Al is most important to assess the chemical homogeneity of TiAl materials. Double VAR melted ingots exhibit an Al deviation of ± 0,75 wt.-% which means that this deviation is too large in order to meet the technical specification requirements. Details are given in Figure 2 of the attached document.
Several TNM ingots via double VAR processing have been produced for both as consumable electrodes for the into forging stocks in the VAR Skull Melter and mechanically separated remelt stocks of approximately 8 kg weight for the conversion into forging stocks in the ISM (Induction Skull Melting). It is evident that the homogeneity is not that sufficient for the manufacturing of forging stocks if the further homogeneization is applied to relatively small quantities such as some kilograms.
Investigations with regard to the homogeneity of single VAR melted ingots show that the local element deviations exceed remarkably the ranges given in the corresponding materials specifications. Corresponding investigations in the past demonstrated a local Al deviation within a range of about ± 5 wt.-%. On the other hand, a major part of about 40 kg of such an ingot used as consumable VAR skull melter electrode is being further homogenized in the skull melter and results in products with specified chemical compositions. The homogenization of locally inhomogeneous consumable electrodes during VAR SM processing is very effective. The chemical analyses of two sets of centrifugally cast slugs resulting from single or double VAR processed consumable electrodes are undistinguishable and, thus, cannot be adjusted to the type of the consumable electrodes (Figure 3 in the attachement). From this point of view, the use of single VAR melted consumable electrode for VAR skull melter processing is a realistic option for cost reduction.
b) PAM ingots
A third additional alternative which has been evaluated is the production of remelt stocks using “plasma-arc-melting” (PAM) instead of VAR. Because Plasma Arc Melting will be probably the most promising recycling technology, PAM ingots were investigated with regard to their chemical composition, microstructure and homogeneity. Unfortunately, due to the cracking phenomenon, PAM ingots cannot be used as consumable electrodes in the VAR SM furnace. Thus, PAM ingots have to be cut into remelt stocks in order to be further processed via ISM.
In the Project, non-contaminated scrap from castings (feeders, tundish etc.) has been used for PAM remelting. As expected, cracking occurred three times during the ingot withdrawing procedure despite the processing was adjusted to beta-TiAl alloys. Furthermore, non-melted inclusions were found. There are serious doubts if a reliable PAM processing of TNM can be established. The patented VAR technology is not applicable to PAM due to technical reasons.
If compare between the chemical composition of the scrap material before PAM conversion and the chemical composition of the PAM ingot it becomes evident that there is a remarkable impurity pick up during PAM processing. In particular Iron, Copper and Oxygen are being picked up wheras Aluminum seems to be partly evaporated. The Iron content increased from 300 ppm to 700 ppm, Copper from 60 ppm to 300 ppm and oxygen from 400 ppm to 1000 ppm. The Al content was reduced by 0,4 wt.-%. Thus, the Al evaporation has to be compensated during PAM processing which should not be an issue (see Figure 3 in the attachement).
In summary, PAM processing seems to be applicable in order to convert TNM casting scrap materials into ingots. Because such ingots cannot be used as consumable VAR SM electrodes for further VAR SM homogenization due to the cracking phenomenon, a PAM technology has to be developed which results in very small diameter ingots to be cut and used directly as forging stocks. For this technology, a sufficient chemical homogeneity has just to be demonstrated.
• Development of casting wheels and permanent moulds
The casting wheel with permanent moulds is a complex thermo-mechanical system which has to meet the following requirements:
▪ stability of shape and size during thermo-cyclic processing of both the wheel and the product as well
▪ simple assembling and dis-assembling
▪ no mould and wheel erosion
▪ reasonable maintenance efforts
The major problem is to control the interaction between the thermal expansion of the different parts after the pour followed by negative thermal expansion during the cooling phase.
In the past, several casting wheel designs have been evaluated (e.g. casting wheels for automotive valves) more or less with negative results. Based on this experience, a complete novel casting wheel design was developed.
The casting wheels work sufficiently. There is no hint for materials erosion up to now. The functional parts of the casting wheel as well as the moulds remain stable after more than 100 pours. It is expected that the number of pours can be further increased to much larger figures since there is no hint for any thermal deformations up to now. The system runs stable.
In summary, PAM processing seems to be applicable in order to convert TNM casting scrap materials into ingots. Because the ingots cannot be used as consumable VAR SM electrodes for further homogeneisation, the use as direct source for remelt stocks will require double melting in case of virgin materials.
• Development of VAR skull melting
The principle processing is shown as a sketch in Figure 4 and in reality in Figure 5 of the attachement. A consumable electrode is partly remelted into a water cooled copper crucible by applying a high arc current which keeps the material liquid except the skull. An induction crucible coil supports materials homogeneization during the melting procedure. The total amount of liquid material is defined through the entire energy consumption of the arc.
After the defined total energy has been consumed, the arc is switched off. The electrode removes very fast from the crucible and the crucible tilting starts initiating the pouring operation. The liquid material is poured through a tundish off-center into the rotating casting wheel.
The following melting and casting parameters were evaluated and have been fixed meanwhile:
• Arc current and arc voltage
• Melting atmosphere (vacuum)
• pouring time → crucible tilting curve
• rotational speed of the casting wheel
• melt flow control via tundish into the casting wheel (off-center, radial flow)
• Development of Induction Skull Melting
The tests of induction skull melting and subsequent centrifugal casting have been performed at Access in a Leicomelt casting facility of ALD (see figure 6 in the attachement).
The process characteristics are as follows:
● VISM in water-cooled copper crucible (preventing unwanted reactions between melt and crucible)
● 2-chamber-system (separation between melting and casting chamber)
● melting capacity 8 kg
● Optionally gravity or centrifugal casting (0-400 Rpm)
The melt is poured via a funnel into a rotating permanent mould.
• Simulation of the casting process
To support the optimization of the VAR skull melting process as well as the induction skull melting process into permanent moulds with different shapes, several casting process simulations have been performed at Access.
In the beginning the principle casting set-up of the VAR skull melter has been adopted to the casting simulation software.
With this set-up simulations on filling, solidification and evolution of casting porosity have been performed. As a result a first optimization step concerning the geometry has been established. Casting porosity can be effectively suppressed by extensive feeding (material consuming) and pouring into tapered moulds with increasing casting cross section. Both elements are being used in the existing casting process. Mould filling and solidification mechanisms for plenty of different mould geometries have been simulated which gave valuable information with regard to an optimum mould geometry. Figure 7 in the attachement illustrates the casting simulation work.
• Evaluation of cast forging stocks (slugs) resulting from VAR SM processing
Within one VAR SM pour, the homogeneity of the cast cylindrical parts is excellent. The deviation of Al within the slugs of one pour determined at the bottom and top of each slug is less than ± 0,15 wt.-% which meets the accuracy of XRF analysing procedur (see following table shows the chemical composition of 18 parts produced in one pour at the bottom and the head.
It can be further stated that there is no difference in the chemical homogeneity of VAR SM processed slugs poured from single VAR or double VAR melted consumable electrodes.
As one may expect with respect to equivalent chemistry, the results of the casting and pouring process are not related to the type of the consumable electrode since the homogenization effect of the skull melter compensates completely local inhomogeneities of the consumable electrodes.
• Forging stocks originating from ISM processing
Slugs have been manufactured by centrifugal casting in tapered moulds. Investigations were performed in the as-cast condition. Thus, in any case, remaining casting porosity was detected particularly in the upper part of the slug along the centreline region. The different alternatives with regard to VIM feed stock materials show uniform results after VIM SM processing. The homogeneization effect of the VIM SM furnace is effective as well and results in homogeneous cast slugs but relatively small batch sizes. Both macro- and microstructure seem to be not related directly to the type of the feed stock material, if the feedstock in summary represents the nominal alloy composition. Evaporation losses of Al can be easily pre-compensated by feed stock alloy production.
• Manufacturing of feed stock materials for VAR SM and ISM
In case of the use of virgin material, the production of feed stocks for VAR SM processing is much less expensive compared to the production of ISM feed stock materials.
If revert material is available, there is no conversion technology to VAR SM feed stocks (except the use of maximum 15 % as starter material) available up to now, whereas revert can be used almost unlimited as feed stock in the VIM SM. The adjustment of the Al content is technically feasible without remarkable additional efforts.
If starting with virgin materials, the most important argument for VAR SM processing of the initial feed stock is the much larger size since VAR skull melters are available up to some tons of pouring weight whereas the Induction skull melters are limited to maximum 100 kg. Furthermore, ISM processing is remarkably more expensive compared to VAR SM processing.
As a consequence, VAR SM processing is preferred for the “virgin route” against ISM processing.
If starting with qualified revert, direct ISM conversion into semi-finished products is the most effective technology. PAM conversion into ingots requires a second conversion step and is not applicable to beta stabilized Titanium Aluminides. More details on the Pro`s and Con`s are given in Table 2 of the attachement.
• Production and characterization of different sets of forging stocks
Two production sets of forging stocks of different geometries have been manufactured and characterized in detail. Two sets consisted of each 5 pours and 45 resulting forging stocks.
Results of the chemical analyses, the evaluation of the macro- and the microstructure give evidence that the applicable specifications have been completely met.
For more details, please see Figures 8 and 9 of the attachment.
• Demonstration of the stability and reproducibility of the production
process of target forging stocks
The optimized (preferred) production process has been extensively evaluated in the work packages 1-4 and subsequently defined in WP 5 of the DATACAST project.
Based on this fixed process, in total much more than 100 batches of forging stocks for the low pressure turbine of the PW GTF aircraft engine have been produced and characterized up to the end of the project. There is evidence on the stability of the entire production process. The forging stocks meet the requirements of all applicable materials and processing specifications.
The technology has been finally approved by the Task Leader MTU. Any commercial production is based on this technology.
• Evaluation of expected series production costs and identification of
processing improvements for cost reductions
Future serial production costs depend mainly on technology improvements whereas the influence of the mass production factor is comparably small. Based on the current production costs appropriate cost saving potentials have been detected, evaluated and presented in detail to the Task Leader.
a) effects of mass production
b) effects of future technology improvements
Since the production technology is a fixed process, all technology improvements are subject to Process Change Requests and corresponding approval procedures.
Potential Impact:
All novel major civil aircraft engine programs (GEnx, LEAP, PW GTF) contain light weigth LPT blades made up of Titanium Aluminides. The precondition for this is the availability of high quality feed stocks / semi-finished products to reasonable costs.
VAR skull melter based homogenization and casting technology is the unique basis for the production of all TiAl components which are in industrial use up to now. The demand for TiAl LPT blades is dramatically growing. It is expected that the demand exceeds 100.000 blades in 2016.
As the market and technology leader, GfE Metalle und Materialien GmbH has developed and industrialized an entire production chain starting with raw materials down to feed stocks and semi-finished products for different parts / TiAl alloys.
GfE has started a huge investment program with two objectives:
• Increase of the production capacity by a factor of 4 to 5
• Decrease of production costs by implementing an effective recycling technology
According to this, GfE will double the production capacity in 2015 based on the virgin route and double the production capacity again in 2016 by further improvement of the technology and introduction of the recycling route.
All these efforts will have a substantial impact on green air transportation and the development of new business opportunities in the field of high performance materials in Europe.
Further R&D project are necessary in order to exploit the full potential of TiAl as a light weight high temperature high strength material in aircraft engine, land turbine and automotive applications.
GfE has reported about the foreground results in conferences and scientific papers.
List of Websites:
none
Novel gamma-TiAl based intermetallic light-weight high temperature structural materials are recently being used as Low Pressure Turbine blades in aircraft engines.
Depending on the LPT performance (low or high speed), the technical requirements on the materials properties are different. Anyhow, there is a basic requirement for chemical and microstructural homogeneity which exceeds corresponding standards of steel or Ni-based superalloys remarkably. Thus, the production of gamma-TiAl parts is very challenging.
The Project leader, MTU Aero Engines GmbH, has decided to introduce forged TiAl TNM blades in the last stage of the Low Pressure Turbine of Pratt & Whitney`s novel PurePower PW1100G aircraft engine family in order to gain efficiency and reduce specific fuel consumption. The novel TNM alloy has been developed in order to improve the poor wrought processing capability of gamma-TiAl by the addition of the bcc structured beta-Ti phase into the well known alpha2/gamma microstructure of Titanium Aluminides. As a side effect, the microstructure remains fine grained due to a different solidification path of the alloy which is characterized by a complete β-solidification mechanism. It has been demonstrated that the as-cast TNM alloy can be subsequent wrought processed without primary ingot conversion via extrusion technology.
The major objective of the project was to develop an industrial production technology for TNM feed stock parts for forging (forging stocks) offering highest product quality and lowest production costs. Different production routes have been evaluated for the metallurgical alloying (Plasma Arc Melting, Vacuum Arc Melting), the homogenization (Induction Skull Melting, Vacuum Arc Skull Melting) and casting technologies.
Based on experimental work and simulations of the casting and solidification mechanisms the following production technology has been identified to meet the requirements with the best compromise:
• gamma-TiAl ingot manufacturing from Ti-sponge, Al and master alloys via VAR processing
• Ingot conversion and homogenization in a VAR Skull Melter
• Centrifugal casting in permanent moulds
• HIP
• Mechanical machining
The technology has been completely developed and industrialized for the production of
gamma-TiAl TNM forging stocks.
Project Context and Objectives:
Novel gamma-TiAl based TNM alloys have been developed in order to increase wrought processing capability by the addition of the bcc structured beta-Ti phase into the well known alpha2/gamma microstructure of Titanium Aluminides. As a side effect, the microstructure remains fine grained due to a different solidification path of the alloy which is characterized by a complete β-solidification mechanism. It has been demonstrated that the as-cast TNM alloy can be subsequent wrought processed without primary ingot conversion via extrusion technology. Unfortunately, this novel TNM alloy exhibits disadvantages in the metallurgical processing compared to State-of-the-Art intermetallic TiAl alloys auch as TiAl48-2-2.
Thus, the objective of the materials manufacturing technology covers both the development of an industrial stable alloying technology to ingots and the subsequent conversion technology to homogeneous small size parts.
For ingot alloying, PAM and VAR should be evaluated as the most common technologies for Ti alloy ingots containing substantial amounts of low melting elements such as Aluminum. EBM is not applicable due to the severe evaporation of Al during the melting operation.
MM and ACCESS have demonstrated recently that the following two technologies for the conversion of VAR ingots into forging stocks are basically feasible.
a) Remelting of gamma-TiAl VAR ingots in a VAR skull melter combined with gravity die or centrifugal casting in permanent moulds, subsequent HIP and mechanical machining
b) Cutting of gamma-TiAl VAR ingots into segments, remelting segments in a VIM skull melter combined with gravity die or centrifugal casting in permanent moulds, subsequent HIP and mechanical machining
None of both technologies have ever been used in industry before. Both technologies exhibit technical Pro`s and Con`s. Both technologies allow basically the use of non-contaminated recycling materials and prevent substantial materials losses during the entire manufacturing chain except final machining. Both technologies can be basically industrialized.
Neither the melting/casting process nor the solidification mechanism are sufficiently investigated and understood up to now. Additionally, ingot manufacturing via VAR is based on a novel technology which has been developed at MM particularly for TNM alloys in order to overcome the “cracking phenomenon” during the VAR processing. This technology is not yet approved and may require amendments to existing industrial standard procedures. A patent application has been filed on the corresponding procedure (DE 10 2009 050 603.9 and PCT/EP 2010/064 306) by MM. The procedure itself results in absolutely stable VAR processing behaviour.
An additional alternative which needs to be evaluated is the production of remelt stocks using “plasma-arc-melting” (PAM) instead of VAR. Because Plasma Arc Melting will be probably the most promising recycling technology, PAM ingots have to be extensively investigated with regard to their chemical composition, microstructure and homogeneity. Due to the cracking phenomenon, PAM ingots have to be cut in remelt stocks in order to be further processed via VIM skull melting combined with gravity die or centrifugal casting in permanent moulds.
In any case, HIP has to be applied for the closure of any casting porosity.
Subsequent machining of the billets is one of the most important cost drivers. Thus, the casting process should preferably result in a net shape part which does not require surface machining before further processing. The casting process has to be optimized with this regard.
The design of the casting tool (permanent mould, casting wheel) is of remarkable importance of the entire cost structure as well and must address in particular
- assembling and disassembling issues
- repair and maintenance issues
- entire lifetime of the casting tool
The objective of the project is the development of a low cost casting process for γ-TiAl based TNM alloys which guarantees to meet all technical specifications of the products.
Two basic requirements determine the scientific and technology methodology of the project:
→ uniform product properties such as homogeneous microstructure, reproducibility
→ application of robust low cost mass production technologies exhibiting high specific materials yields
As a precondition, all single steps in the production chain must run absolutely stable and repeatable without remarkable materials losses. Therefore, an initial ingot manufacturing technology has to be used which overcomes the cracking behaviour of TNM consumable electrodes during conventional VAR processing and VAR Skull Melter processing as well.
Alternative 1
is based on a novel VAR technology patented by GfE. Remelting in a VAR skull melter offers excellent homogenization due to large batch sizes.
Alternative 2
does replace one initial VAR step by an additional VIM step. Homogenization in the VAR skull melter is comparable to alternative 1. The application of gravity die or centrifugal casting depends on the remelt stock size required in the subsequent VIM process.
Alternative 3
is based on a PAM ingot which cannot be further processed in the VAR skull melter due to the cracking issue. Thus, the separation into feed stock materials for the VIM skull melter must be done by mechanical sawing or similar technologies which causes additional materials losses and may limit homogenization potential.
Project Results:
3. Description of the main S&T results
• Evaluation of TNM ingot manufacturing routes
a) VAR ingots
Since at least one Vacuum Arc Remelting (VAR) step is still required for the production of Ti-alloys used in rotating aircraft engine parts, the VAR process for ingot manufacturing has been evaluated in very detail.
In contrast to conventional TiAl alloys, novel beta-stabilized TiAl alloys such as TNM show a remarkable tendency to crack during VAR processing. It could be demonstrated that this behaviour originates from abnormal thermal expansion processes at the γ-solvus temperature which is much lower in beta-stabilized TiAl alloys compared to the conventional ones. Thus, the capability to remove stresses induced by abnormal thermal expansion by plastic flow is remarkably decreased due to the lower temperature and stress relaxation occurs via cracking.
A novel VAR technology has been developed by GfE MM in order to overcome this “cracking phenomenon”. This technology runs absolutely stable and has been approved meanwhile. A corresponding patent has been filed in 2009 and was issued (EP 10 765 988.0-2122) in the EU recently. All the VAR processing in the Project is based on this patented technology.
The ingots are being used either as consumable electrodes in the VAR skull melter or are being cut (separated) into lumpy pieces as remelt stocks for the VIM skull melter.
In the first option, approximately one third of the ingot (about 40 kg) is homogenized in the VAR skull melter and converted to cylindrical slugs via centrifugal casting in permanent moulds. In the second option, the homogeneization in the VIM skull melter is applied to much smaller parts of about 8 kg mass. In order to get an idea about the homogeneity of a double VAR TNM ingot, one ingot of approximately 170 kg mass has been separated into 85 pieces of about 2 kg mass in average. Every piece was chemically analysed.
The deviation of Al is most important to assess the chemical homogeneity of TiAl materials. Double VAR melted ingots exhibit an Al deviation of ± 0,75 wt.-% which means that this deviation is too large in order to meet the technical specification requirements. Details are given in Figure 2 of the attached document.
Several TNM ingots via double VAR processing have been produced for both as consumable electrodes for the into forging stocks in the VAR Skull Melter and mechanically separated remelt stocks of approximately 8 kg weight for the conversion into forging stocks in the ISM (Induction Skull Melting). It is evident that the homogeneity is not that sufficient for the manufacturing of forging stocks if the further homogeneization is applied to relatively small quantities such as some kilograms.
Investigations with regard to the homogeneity of single VAR melted ingots show that the local element deviations exceed remarkably the ranges given in the corresponding materials specifications. Corresponding investigations in the past demonstrated a local Al deviation within a range of about ± 5 wt.-%. On the other hand, a major part of about 40 kg of such an ingot used as consumable VAR skull melter electrode is being further homogenized in the skull melter and results in products with specified chemical compositions. The homogenization of locally inhomogeneous consumable electrodes during VAR SM processing is very effective. The chemical analyses of two sets of centrifugally cast slugs resulting from single or double VAR processed consumable electrodes are undistinguishable and, thus, cannot be adjusted to the type of the consumable electrodes (Figure 3 in the attachement). From this point of view, the use of single VAR melted consumable electrode for VAR skull melter processing is a realistic option for cost reduction.
b) PAM ingots
A third additional alternative which has been evaluated is the production of remelt stocks using “plasma-arc-melting” (PAM) instead of VAR. Because Plasma Arc Melting will be probably the most promising recycling technology, PAM ingots were investigated with regard to their chemical composition, microstructure and homogeneity. Unfortunately, due to the cracking phenomenon, PAM ingots cannot be used as consumable electrodes in the VAR SM furnace. Thus, PAM ingots have to be cut into remelt stocks in order to be further processed via ISM.
In the Project, non-contaminated scrap from castings (feeders, tundish etc.) has been used for PAM remelting. As expected, cracking occurred three times during the ingot withdrawing procedure despite the processing was adjusted to beta-TiAl alloys. Furthermore, non-melted inclusions were found. There are serious doubts if a reliable PAM processing of TNM can be established. The patented VAR technology is not applicable to PAM due to technical reasons.
If compare between the chemical composition of the scrap material before PAM conversion and the chemical composition of the PAM ingot it becomes evident that there is a remarkable impurity pick up during PAM processing. In particular Iron, Copper and Oxygen are being picked up wheras Aluminum seems to be partly evaporated. The Iron content increased from 300 ppm to 700 ppm, Copper from 60 ppm to 300 ppm and oxygen from 400 ppm to 1000 ppm. The Al content was reduced by 0,4 wt.-%. Thus, the Al evaporation has to be compensated during PAM processing which should not be an issue (see Figure 3 in the attachement).
In summary, PAM processing seems to be applicable in order to convert TNM casting scrap materials into ingots. Because such ingots cannot be used as consumable VAR SM electrodes for further VAR SM homogenization due to the cracking phenomenon, a PAM technology has to be developed which results in very small diameter ingots to be cut and used directly as forging stocks. For this technology, a sufficient chemical homogeneity has just to be demonstrated.
• Development of casting wheels and permanent moulds
The casting wheel with permanent moulds is a complex thermo-mechanical system which has to meet the following requirements:
▪ stability of shape and size during thermo-cyclic processing of both the wheel and the product as well
▪ simple assembling and dis-assembling
▪ no mould and wheel erosion
▪ reasonable maintenance efforts
The major problem is to control the interaction between the thermal expansion of the different parts after the pour followed by negative thermal expansion during the cooling phase.
In the past, several casting wheel designs have been evaluated (e.g. casting wheels for automotive valves) more or less with negative results. Based on this experience, a complete novel casting wheel design was developed.
The casting wheels work sufficiently. There is no hint for materials erosion up to now. The functional parts of the casting wheel as well as the moulds remain stable after more than 100 pours. It is expected that the number of pours can be further increased to much larger figures since there is no hint for any thermal deformations up to now. The system runs stable.
In summary, PAM processing seems to be applicable in order to convert TNM casting scrap materials into ingots. Because the ingots cannot be used as consumable VAR SM electrodes for further homogeneisation, the use as direct source for remelt stocks will require double melting in case of virgin materials.
• Development of VAR skull melting
The principle processing is shown as a sketch in Figure 4 and in reality in Figure 5 of the attachement. A consumable electrode is partly remelted into a water cooled copper crucible by applying a high arc current which keeps the material liquid except the skull. An induction crucible coil supports materials homogeneization during the melting procedure. The total amount of liquid material is defined through the entire energy consumption of the arc.
After the defined total energy has been consumed, the arc is switched off. The electrode removes very fast from the crucible and the crucible tilting starts initiating the pouring operation. The liquid material is poured through a tundish off-center into the rotating casting wheel.
The following melting and casting parameters were evaluated and have been fixed meanwhile:
• Arc current and arc voltage
• Melting atmosphere (vacuum)
• pouring time → crucible tilting curve
• rotational speed of the casting wheel
• melt flow control via tundish into the casting wheel (off-center, radial flow)
• Development of Induction Skull Melting
The tests of induction skull melting and subsequent centrifugal casting have been performed at Access in a Leicomelt casting facility of ALD (see figure 6 in the attachement).
The process characteristics are as follows:
● VISM in water-cooled copper crucible (preventing unwanted reactions between melt and crucible)
● 2-chamber-system (separation between melting and casting chamber)
● melting capacity 8 kg
● Optionally gravity or centrifugal casting (0-400 Rpm)
The melt is poured via a funnel into a rotating permanent mould.
• Simulation of the casting process
To support the optimization of the VAR skull melting process as well as the induction skull melting process into permanent moulds with different shapes, several casting process simulations have been performed at Access.
In the beginning the principle casting set-up of the VAR skull melter has been adopted to the casting simulation software.
With this set-up simulations on filling, solidification and evolution of casting porosity have been performed. As a result a first optimization step concerning the geometry has been established. Casting porosity can be effectively suppressed by extensive feeding (material consuming) and pouring into tapered moulds with increasing casting cross section. Both elements are being used in the existing casting process. Mould filling and solidification mechanisms for plenty of different mould geometries have been simulated which gave valuable information with regard to an optimum mould geometry. Figure 7 in the attachement illustrates the casting simulation work.
• Evaluation of cast forging stocks (slugs) resulting from VAR SM processing
Within one VAR SM pour, the homogeneity of the cast cylindrical parts is excellent. The deviation of Al within the slugs of one pour determined at the bottom and top of each slug is less than ± 0,15 wt.-% which meets the accuracy of XRF analysing procedur (see following table shows the chemical composition of 18 parts produced in one pour at the bottom and the head.
It can be further stated that there is no difference in the chemical homogeneity of VAR SM processed slugs poured from single VAR or double VAR melted consumable electrodes.
As one may expect with respect to equivalent chemistry, the results of the casting and pouring process are not related to the type of the consumable electrode since the homogenization effect of the skull melter compensates completely local inhomogeneities of the consumable electrodes.
• Forging stocks originating from ISM processing
Slugs have been manufactured by centrifugal casting in tapered moulds. Investigations were performed in the as-cast condition. Thus, in any case, remaining casting porosity was detected particularly in the upper part of the slug along the centreline region. The different alternatives with regard to VIM feed stock materials show uniform results after VIM SM processing. The homogeneization effect of the VIM SM furnace is effective as well and results in homogeneous cast slugs but relatively small batch sizes. Both macro- and microstructure seem to be not related directly to the type of the feed stock material, if the feedstock in summary represents the nominal alloy composition. Evaporation losses of Al can be easily pre-compensated by feed stock alloy production.
• Manufacturing of feed stock materials for VAR SM and ISM
In case of the use of virgin material, the production of feed stocks for VAR SM processing is much less expensive compared to the production of ISM feed stock materials.
If revert material is available, there is no conversion technology to VAR SM feed stocks (except the use of maximum 15 % as starter material) available up to now, whereas revert can be used almost unlimited as feed stock in the VIM SM. The adjustment of the Al content is technically feasible without remarkable additional efforts.
If starting with virgin materials, the most important argument for VAR SM processing of the initial feed stock is the much larger size since VAR skull melters are available up to some tons of pouring weight whereas the Induction skull melters are limited to maximum 100 kg. Furthermore, ISM processing is remarkably more expensive compared to VAR SM processing.
As a consequence, VAR SM processing is preferred for the “virgin route” against ISM processing.
If starting with qualified revert, direct ISM conversion into semi-finished products is the most effective technology. PAM conversion into ingots requires a second conversion step and is not applicable to beta stabilized Titanium Aluminides. More details on the Pro`s and Con`s are given in Table 2 of the attachement.
• Production and characterization of different sets of forging stocks
Two production sets of forging stocks of different geometries have been manufactured and characterized in detail. Two sets consisted of each 5 pours and 45 resulting forging stocks.
Results of the chemical analyses, the evaluation of the macro- and the microstructure give evidence that the applicable specifications have been completely met.
For more details, please see Figures 8 and 9 of the attachment.
• Demonstration of the stability and reproducibility of the production
process of target forging stocks
The optimized (preferred) production process has been extensively evaluated in the work packages 1-4 and subsequently defined in WP 5 of the DATACAST project.
Based on this fixed process, in total much more than 100 batches of forging stocks for the low pressure turbine of the PW GTF aircraft engine have been produced and characterized up to the end of the project. There is evidence on the stability of the entire production process. The forging stocks meet the requirements of all applicable materials and processing specifications.
The technology has been finally approved by the Task Leader MTU. Any commercial production is based on this technology.
• Evaluation of expected series production costs and identification of
processing improvements for cost reductions
Future serial production costs depend mainly on technology improvements whereas the influence of the mass production factor is comparably small. Based on the current production costs appropriate cost saving potentials have been detected, evaluated and presented in detail to the Task Leader.
a) effects of mass production
b) effects of future technology improvements
Since the production technology is a fixed process, all technology improvements are subject to Process Change Requests and corresponding approval procedures.
Potential Impact:
All novel major civil aircraft engine programs (GEnx, LEAP, PW GTF) contain light weigth LPT blades made up of Titanium Aluminides. The precondition for this is the availability of high quality feed stocks / semi-finished products to reasonable costs.
VAR skull melter based homogenization and casting technology is the unique basis for the production of all TiAl components which are in industrial use up to now. The demand for TiAl LPT blades is dramatically growing. It is expected that the demand exceeds 100.000 blades in 2016.
As the market and technology leader, GfE Metalle und Materialien GmbH has developed and industrialized an entire production chain starting with raw materials down to feed stocks and semi-finished products for different parts / TiAl alloys.
GfE has started a huge investment program with two objectives:
• Increase of the production capacity by a factor of 4 to 5
• Decrease of production costs by implementing an effective recycling technology
According to this, GfE will double the production capacity in 2015 based on the virgin route and double the production capacity again in 2016 by further improvement of the technology and introduction of the recycling route.
All these efforts will have a substantial impact on green air transportation and the development of new business opportunities in the field of high performance materials in Europe.
Further R&D project are necessary in order to exploit the full potential of TiAl as a light weight high temperature high strength material in aircraft engine, land turbine and automotive applications.
GfE has reported about the foreground results in conferences and scientific papers.
List of Websites:
none
