Development of light-weight steel castings for efficient
aircraft engines - Summary report
Martin Risberg, Anders GotteBox 2033, 550 02 Jönköping, Sweden Tel: +46(0)36 - 30 12 00 swecast@swerea.se http://www.swecast.se c 2014, Swerea SWECAST AB
296585 LEAN - Development of light-weight steel castings for efficient aircraft engines
Author Report number: Date
Martin Risberg, Anders Gotte 2014-003 2014-02-25
Summary
This report summaries the work carried out in the project, LEAN - Development of light-weight steel castings for efficient aircraft engines. The overall object in the project was to develop thinner cast steel components for aircraft engines. The work was conducted in collaboration between research institutes and industry. Swerea SWECAST, the Swedish foundry institute, which has a long tradition in collaboration with the industry in research projects, was the project leader and leader in one of the work packages. FRI, the Polish Foundry Research Institute, which has advance equipment for material science and material data investigations, was the leader of the second work package. The investment foundry TPC Components AB conducted casting trials and contributed with their wide knowledge of investment casting.
Casting trials have been performed in order to investigate the influence of different process pa-rameters governing the fluidity of thin walled investment castings. The alloy used was CbCu7-1, i.e. the cast analogy of the stainless precipitation-hardening steel 17-4PH. Two levels of geometry complexity were used as well as top- and bottom gated casting systems. In the first trial, a blade thicknesses ranging from 0.7-2.0 mm was used. In the second trial, some features were added to the blade as well as a textured surface on one side. It was shown that the top gated casting system showed an overall improved fluidity compared to the bottom gated casting system for the simpler geometry. Blade thickness and pouring temperature were shown to have the greatest impact on fluidity. Adding some geometrical features to the simple geometry drasti-cally decreased the differences between the filling systems. Using a one side textured blade with thickness of 1.3 and 1.5 mm was comparable with 1.5 and 2.0 mm flat castings thus reducing weight of the thinnest sections of a steel casting. Predictions of miss-runs with simulations were shown to be in good agreement with experiments and gave valuable insight to problems in the casting trials. Differences in porosity levels were seen between the top- and bottom gated casting systems for the simpler geometry at high metal temperatures, where the former showed a larger amount of porosity.
Besides the work performed on fluidity of the cast analogy of 17-4PH, a number of other al-loys not commonly used for castings today were evaluated in terms of their fluidity and were compared to 17-4PH. The casting trials were performed with a bar casting. It was shown that JETHETE 152M had the best fluidity followed by Custom 465, L0H12N4M and 17-4PH. CSS 42L and PH13-8M ranked worst in the fluidity comparison and were therefore excluded from further investigations. Mechanical testing at both ambient and elevated temperatures (400 degrees Celsius) was performed. It was concluded that JETHETE 152M was the best per-forming alloy with respect to mechanical properties, followed by L0H12N4M, 17-4PH and Custom 465. However, JETHETE 152M was later excluded due to its poor corrosion prop-erties. In the corrosion test, at 400 degrees Celsius, for 100 hours with salt spray fog, it was determined that the 17-4PH and L0H12N4M showed similar corrosion rates. Wettability test performed on two different shell systems with 17-4PH showed that the shell/alloy system is im-portant to consider during filling of thin sections. The 17-4PH alloy was used in the casting trials of a demonstrator. It was demonstrated during the casting trials that filling of a wide flat section with a thickness below 2 mm is hard to achieve.
Sammanfattning
Denna rapport sammanfattar arbetet som är utfört i projektet, LEAN - Development of light-weight steel castings for efficient aircraft engines. Det övergripande målet för projektet var att gjuta tunna sektioner för komponenter i flygmotorer. För att kunna göra detta behövde befintlig gjutmetod, vaxursmältning, utvecklas. Arbetet har delats mellan forskningsinstitut och industri. Swerea SWECAST, det svenska gjuteriinstitutet som har en lång erfarenhet av forskningspro-jekt i samarbete med industrin har koordinerat arbetet och lett ett av arbetspaketen. Foundry Research Institute, det polska gjuteriinstitutet, som har utrustning för materialvetenskap och materialundersökningar, var ledare för det andra arbetspaket. Ett av Europas ledande precision-sgjuterier, TPC Components AB, genomförde gjutförsök och bidrog med sin breda kunskap om vaxutsmältningsgjutning.
Gjutförsök har utförts för att undersöka olika processparametrars inflytande på flödet vid fyllnad av tunna sektioner. Legeringen som användes i arbetet var CbCu7-1, d.v.s. gjutversionen av den rostfria utfällningshärdade 17-4PH. Två nivåer av komplexitet på försöksgeometri har använts. Även gjutningssystemen har varit olika med inlopp både över- och under ifrån. I den första studien har väggtjocklekar mellan 0,7-2,0 mm studerats. I den andra studien har geometrins komplexitet ökats med bl.a. en strukturerad yta på tunna sektioner. Resultatet för den enkla geometrin visar att inlopp ovanifrån generellt hade bättre fyllnad jämfört de där inloppet kom in underifrån. Tjocklek och gjuttemperatur påverkade fyllnaden. Vid ökad komplexitet minskade inverkan från de olika varianterna av ingjutsystem. Geometrin med strukturerad yta med tjocklek 1,3 och 1,5 mm kunde jämföras med 1,5 och 2,0 mm med jämn yta. Resultatet gav en viktreduk-tion för tunna sekviktreduk-tioner. Simulering av flödesfel visade på god jämförelse med försök och gav viktig kunskap för förståelsen av försöken. De olika ingjutsystemen påverkade porositetsbilden, framförallt för de enkla geometrierna och höga gjuttemperaturer. Fallande gjutsystem gav mest porositet.
Förutom arbetet som utförts på 17-4PH undersöktes ett antal andra legeringar som vanligtvis inte gjuts. Dessa legeringar har utvärderats med avseende på fyllnadsmöjlighet och jämförts med 17-4PH. Gjutförsök med olika tjocka stänger visade att JETHETE 152M hade bäst flyt-barhet följt av Custom 465, L0H12N4M och 17-4PH. CSS 42L och PH13-8M rankades sämst och uteslöts från fortsatt undersökning. Dragprov vid rums- och förhöjdtemperatur (400 grader Celsius) utfördes. Slutsaten var att JETHETE 152M var bäst följt av L0H12N4M, 17-4PH och Custom 465. Korrosionsprov utfördes vid 400 grader Celsius under 100 timmar med saltdimma. JETHETE 152M uteslöts på grund av dåliga korrosionsegenskaperna. Resultatet visade att 17-4PH och L0H12N4M var likvärdiga. Vätningstester utförda på två olika skalsystem tillsam-mans med legeringen 17-4PH och visade att skalsystemet är en viktig faktor när tunna sektioner skall fyllas. För demonstratorgjutningen använde 17-4PH och resultatet visade att det var svårt att fylla breda plana sektioner tunnare än 2 mm.
Acknowledgement
The authors would like to thank the European Commission and Clean Sky Joint Undertaking for financing the project. The authors also would like to thank all participant partners and personnel in the project. The authors specially would like to thank Topic Manager Mr. Anders Sjunnesson and his colleagues at GKN Aerospace System in Trollhättan for support during the project.
Table of Contens
1 Introduction 1
2 Scientific methodology and work plan 1
2.1 Overall strategy of the work plan . . . 1
3 Description of work packages 2 3.1 Partners involved in the work . . . 2
3.2 Project organisation . . . 2
3.3 WP 1 - Project management . . . 3
3.4 WP 2 - Minimizing section thickness of steel alloy 17-4PH . . . 3
3.5 WP 3 - Castability limits for other steel alloys . . . 4
4 Detail description of work WP 2 4 4.1 Task 2.1 - Investment casting of simple geometries . . . 5
4.1.1 Experimental work Task 2.1 . . . 5
4.2 Task 2.2 - Investment casting of more complex geometries . . . 8
4.2.1 Experimental work Task 2.2 . . . 8
4.3 Task 2.3 - Post processing of investment castings . . . 9
4.3.1 Experimental work in Task 2.3 . . . 10
5 Detailed description of work in WP 3 10 5.1 Task 3.1 - Casting of selected alloys with simple geometries . . . 10
5.1.1 Experimental work in Task 3.1 . . . 10
5.2 Task 3.2 - Final casting of selected alloys with complex geometry . . . 13
5.2.1 Basic experimental work in Task 3.2 . . . 13
5.2.2 Complex geometry casting experiments . . . 16
6 Conclusions 22 6.1 Conclusions of WP2 . . . 22
6.1.1 Fluidity . . . 22
6.1.2 Micro-structure and defects in thin sections . . . 23
6.1.3 Numerical simulations . . . 23
6.1.4 Post-processing . . . 23
1 Introduction
Today, cast titanium components is the predominant choice over alternatives like nickel base alloys and stainless steels for critical components in the cold section of aero engine. The main downside of using cast titanium components is the additional cost accompanied with the ma-terial. Using a technically equivalent, but cheaper material, leaves more economical headroom for using higher performing light-weight materials in other part of the air craft. By reducing the total weight of the aircraft it is possible to decrease the emissions during the total life cycle. The focus of the project was to find durable light-weight cast steel materials as alternatives to the materials used today. To successfully accomplish this, the weight penalty of using cast steel compared to titanium components had to be decreased. In order to realise this, materials- and processes development needed to be considered within the project. The scope of the project was to explore and push the minimal castable thickness as well as investigate other possible steel alloys.
The challenge of producing thin-walled castings is related to mould filling and securing an over-all sound component meeting over-all specifications. Casting simulations are frequently used today and are considered a necessary tool to predict mould filling and solidification. However, cast simulation software is only able to reliably predict mould filling for section sizes greater that the relevant wall thicknesses used in this study. Hence, the physical parameters affecting the simu-lation results needed to be determined, studied and adjusted to cover the thin section effect of the casting. This was addressed in a number of trial castings and subsequent material investigations. In addition, a large part of the proposed project was related to exploring the castability limits of a number of steel alloys that are not commonly cast today. The alloys were also evaluated in terms of corrosion and mechanical properties and compared to a reference material.
2 Scientific methodology and work plan
2.1 Overall strategy of the work plan
The overall objective for the LEAN-project was to further develop existing investment casting manufacturing method to enable casting of thinner steel components for aircraft engines. The goal was to facilitate replacement of the lighter and more expensive titanium components with investment cast steel components. Hence, the objective of the proposed project was to develop methods to manufacture high-strength, thin-walled steel castings and thus making stainless steel a viable engineering material for castings used in aerospace engine applications. The overall strategy to reach the goal was to work both with a currently used steel alloy and at the same time explore other steel alloys which are not commonly cast today. The cast analogy of the stainless steel alloy 4PH, CbCu7-1, was chosen as the reference alloy. The work performed with 17-4PH focused on identification of process parameters that had the largest impact on the fluidity of thin walled investment castings. Within the project, different process parameters affecting the minimum wall thickness achievable and final quality and post processing methods in order to achieve desired properties investigated. Heat treatment and Hot Isostatic Pressing (HIP) was used to improve the final properties of the castings. The possibility of using chemical milling for further reduction of wall thicknesses was also considered at the beginning of the project. The work with steel alloys not commonly cast today, focused on adopting promising steel alloys and compare them with the reference alloy in terms of both fluidity and material properties. The goal was to identify higher-strength steel alloys for aero-engine components which are possible
to use at a higher service temperature than 17-4PH. The alloys for this part of the project are used today as sheet metal or as forged materials. The castability limits for the alloys were investigated through lab scale cast trials followed by metallurgical, corrosion and mechanical investigations. Finally a more complex demonstrator was cast to raise the TRL-level of the project findings.
3 Description of work packages
3.1 Partners involved in the work
In order to have an efficient research and development organisation, the work was divided be-tween industry and research institutes. Swerea SWECAST (SSC) was the responsible partner in the project. Swerea SWECAST has a long history of working close to the industry in applied research projects. The Swedish investment casting foundry TPC Components AB (TPC) con-ducted the casting trials and contributed with their wide knowledge of investment casting. The Polish Foundry Research Institute (FRI) had all necessary equipment to conduct state of the art research in terms of materials science and material data investigations. FRI have a long history of international and multidisciplinary projects focusing on foundry technology. FRI also has the facility to conduct controlled investment casting experiments.
3.2 Project organisation
To have an efficient organization the work was divided among the different partners into different work packages and sub tasks which is illustrated inTab. 1andFig. 1.
Work Package/Task Title Activity Lead participant
WP 1 Project management Other SSC
WP 2 Minimize section thickness of steel alloy 17-4PH
RTD SSC Task 2.1 Investment casting of simple geometries RTD
Task 2.2 Investment casting of more complex geometries RTD Task 2.3 Post processing of investment castings RTD
WP 3 Castability limits for other steel alloys RTD FRI Task 3.1 Casting of selected alloys with simple
geome-tries
RTD Task 3.2 Final casting of selected alloys with complex
geometry
RTD
Topic Manager, (GKN)
WP 1, (SSC)
WP 2, (SSC) WP 3 (FRI)
Task 2.2
Task 2.1 Task 2.3 Task 3.1 Task 3.2 Manangement Group
(SSC, FRI, TPC)
Figure 1: Work package description.
3.3 WP 1 - Project management
Project management, technical coordination, information, communication of project and dis-semination of research results were carried out in WP 1. WP 1 was the single point of contact for communication with topic management (TM) according to a communication plan set by TM and SSC. WP 1 was also responsible for economy, milestones, results, and effects on short and long term. The main project leader was responsible for the information and communication strategy. The objectives of WP 1 were to secure that the work progressed and to report the progress and any deviations from the work plan. The results were communicated from WP 1 to TM. The deliverables from WP 1 are shown inTab. 2.
Deliverable no: Name Project leader Nature (Diss. level) D1.1 Detailed project plan with milestones SSC Other (Confidential)
D1.2 Final report SSC Report (Public)
Table 2: Deliverables of WP1.
3.4 WP 2 - Minimizing section thickness of steel alloy 17-4PH
The overall goal in WP2 was to explore, characterise, and push the minimum castable section thickness of 17-4PH steel to its limit. The objectives for WP2 can be summarised as follows:
• Optimise the casting conditions on simple geometries based on casting trials and numeri-cal simulation.
• Develop and validate a reliable simulation methodology for thin walled steel investment castings.
• Validate optimal casting conditions on a more complex model.
• Investigate casting quality requirements, defects, material properties of thin walled steel investment castings.
• Explore the possibilities of post processing operations like HIP and Chemical Milling on thin walled steel castings.
Deliverable no: Name Project leader Nature (Diss. level) Reference D2.1 Test plan for simplified geometry 1 SSC Other (Confidential)
D2.2 Test plan for simplified geometry 2 SSC Other (Confidential)
D2.3 Documentation of manufacturing and process SSC Report (Confidential) [1], [2], [3] simulation of simple geometry 1 and 2
D2.4 Documentation of post cast processing SSC Report (Confidential) [4], [5] experiments
D2.5 Documentation of manufacturing and process SSC Report (Confidential) [6] simulation of more complex geometry
Table 3: Deliverables of WP2.
3.5 WP 3 - Castability limits for other steel alloys
The overall goal of WP3 was to explore the castability limits of steel alloys that are not com-monly cast today. The objectives of WP3 were:
• Evaluate and rank the castability of wrought steel alloys for simple geometries. • Develop optimised casting process parameters for selected alloys.
• Develop optimum heat treatment parameters for selected alloys. • Evaluate the castability of selected alloys in complex geometries.
Deliverable no: Name Project leader Nature (Diss. level) Reference D3.1 Test programme for castability FRI Other (Confidential)
experiments for steel alloy with limited castability
D3.2 Documentation of first batch FRI Report (Confidential) [7], [8] of casting trials
D3.3 Documentation of final cast trial for FRI Report (Confidential) [9] one steel limited castability
Table 4: Deliverables of WP3.
4 Detail description of work WP 2
The main objective in WP 2 was to identify and optimise the governing parameters for fluidity in thin walled investment castings. CbCu7-1, the cast analogy of 17-4PH was chosen as alloy. Fluidity, defects and mechanical properties of the thin sections were investigated. Additional
post processing methods, such as Hot Isostatic Pressing (HIP), were also investigated as a possi-bility to reduce potential porosity. The work was based on a series of carefully planned Design of Experiments (DoE) to facilitate statistical evaluations of the results. Both single effects as well as interaction between effects of different process parameters were studied. The casting trials were performed under production conditions in an investment casting foundry. Two levels of geometry complexity were used as well as top and bottom gated inlet systems. First, a simple slightly curved blade was used to identify and study parameters related to the fluidity. Later, some features were added to the blade to increase its complexity. Also in an attempt to decrease the weight of the thin sections, a non-flat surface was evaluated. As the properties after heat treatment of the castings are of major importance in depth studies were performed, defect levels and mechanical performance of the thin sections were taken into account. The work in WP2 was divided into three different tasks, where Task 2.1 and 2.2 focused on minimising the wall thickness and task 2.3 focused on post processing methods.
4.1 Task 2.1 - Investment casting of simple geometries
A literature survey was fist conducted to give a solid base for the choice of relevant process parameters for the DoE. The result from the survey revealed that thin walled investment castings have been a research interest for a long time. The main conclusions from the survey of thin walled investment castings are summarised below:
• A distinction is made between flowability and fillability depending on which mechanism is reducing the ability of the metal to flow. Fillability is related to surface tension and flowability is controlled by heat transfer.
• Fluidity is improved by increasing super heat, mould temperature or pressure head. • Fluidity is influenced by changes in alloy composition.
• Flow rate and shell permeability are also important parameters that will influence fluidity. • Vacuum-assisted casting and mould vibration during filling are methods that have been
presented in literature to further increase fluidity in thin sections.
An additional literature survey was performed which focused on the properties of precipitation-hardened cast steel. Also calculations of properties were performed in JMatPro. The main conclusions from this survey can be concluded as follows:
• The final properties can be tailored in detail by fine-tuning alloy composition and heat treatment, even though the result is very sensitive to small changes in production environ-ment.
• Copper was found to be the alloying element that has the most pronounced effect on fluidity in terms of viscosity and surface tension.
• Above the upper limit for cast Cb7Cu-1 (Cu>3%) might cause problems with precipitation control, mechanical properties and welding.
4.1.1 Experimental work Task 2.1
In the experimental work in Task 2.1, two different levels of complexity were used, shown in
Fig. 2. The terminology of the different levels from here on will be Simple Geometry 1 and Simple Geometry 2, where the latter had an increased level of complexity. It was decided to use
Figure 2: Two levels of geometrical complexity, (a) Simple Geometry 1 and (b) Simple Geometry 2
both the standard casting procedure, top down pouring, as well as bottom up pouring. All trials were performed under production like conditions. Ceramic moulds were produced, following standard foundry procedures, by continuously dipping and stuccoing the wax assemblies until the desired thickness of the shells were achieved. Due to company secrets no information about the recipe of the slurries used for the shelling can be made public. After the investments had dried they were placed in an industrial autoclave for de-waxing. After subsequent sintering had been accomplished to harden the shells, they were preheated in a push-through furnace for a minimum of two hours before molten steel was poured into the mould.
CbCu7-1, cast analogy of the stainless precipitation-hardened steel 17-4PH was used for the casting trials. Before casting trials, thermal analysis was carried out to establish the liquidus temperature of the alloy as well as the chemical analysis. The chemical composition are shown inTab. 6. Filming the pouring process continuously from start to end for all trials was done by conventional DVD-recorder as well as IR-FLIR instrument.
Fe C Si Mn P S Cr Ni Cu Nb N
Bal. 0.042 0.72 0.60 0.014 0.014 16.36 3.73 3.1 0.042 0.057
Table 5: Composition of cast alloy.
Fe C Si Mn P S Cr Ni Cu Nb N
Bal. 0.042 0.72 0.60 0.014 0.014 16.36 3.73 3.1 0.042 0.057
Table 6: Composition of cast alloy.
For Simple Geometry 1 a slightly curved blade was used, with the dimensions of 100 x 150 mm in width and height respectively. Four different thicknesses between 0.7 and 2.0 mm were used. A standard four armed runner system was used for both filling systems. A 10 ppi filter was placed in the pouring cup for both the bottom and top gated casting system. Process pa-rameters for Simple Geometry 1 were chosen based on the literature survey. Besides thickness,
pouring and mould temperature were chosen as the primary parameters and varied in two levels. The flow rate in these trials was assumed as constant due to the filter position at the cup. The test plan amounted to a total of 128 castings trials. However, since each tree consisted of all four thicknesses, the number of poured trees was reduced to 32. All shells were placed in the same position in all trials i.e. same blade dimension was pointing towards the ladle for all trees poured.Fig. 3is illustrating the casting trials for Simple Geometry 1. To a larger extent resem-ble the GKN geometry from the proposal a boss was added at the centre of the blade for Simple Geometry 2. Based on the casting trials for Simple Geometry 1, the blade with thickness of 0.7 mm was excluded since it was shown to be impossible to fill. The blades were attached to a downsprue via a three armed runner system. The runner system was designed to facilitate inlets to the blade at three different locations. No filter was used in this series of casting trials. As the main goal for Simple Geometry 2 was to further narrow the process window it was decided to vary pouring temperature and pressure head in three different levels. The variation of pressure head was only possible for the bottom gated casting system. In the trials for Simple Geometry 2, the mould temperature was assumed to be constant as the blades were insulated with K-Wool. All shells were pre-heated to 1050◦C. No filters were used in these trials.
Figure 3: Casting trials performed at TPC, Simple Geometry 1.
Both bottom and top gated runner systems were used. As for previous trials, the casting was performed under production like conditions. Each set up was repeated three times. Besides these trials also trials with a non-smooth (textured) surface were evaluated in terms of fluidity. In all trials the response of fluidity were evaluated as the projected area of the partly filled blades by means of digital image processing according toFig. 4.
Nova Flow&Solid was used as the primary simulation code and was used extensively through-out the project. The objective was to develop a simulation methodology of a generic type to be used for miss-run predictions of steel investment castings. Also, equally important, was that the simulations facilitated defect tracking in a realistic way. A lot of effort was therefore devoted in identification and adjustment of the simulation parameters.
Figure 4: Evaluation of fluidity by measuring filled projected area of the blade.
Because one of the major findings in the literature survey was the improved effects on fluidity using vacuum assisted casting, it was decided to evaluate this technique. The purpose is to create a pressure difference which will aid the filling of thin sections and small narrow features. The principle is equivalent to increasing the pressure height. The system consisted of three parts, one casting tank in which the shell is placed, one vacuum tank and a vacuum ejector. The basic idea of the pressure system was that the ejector should be used to decrease the pressure in the vacuum tank before casting. The shell was then placed in the casting tank. The valve between the casting tank and the de-pressurized vacuum tank was then released during filling to get a quick decrease in pressure in the casting chamber while the vacuum ejector was used to maintain the low pressure during the whole casting sequence.
4.2 Task 2.2 - Investment casting of more complex geometries
The main goal of Task 2.2 was to achieve a sound casting with the minimal castable thickness with results from Task 2.1 as input. All necessary information about process parameters to achieve full filling of the complex geometry was provided by Task 2.1. Some main considera-tions were considered important besides the filling of the complex casting. Firstly if needed, a final calibration of the numerical model to have a generic model which is not dependent on a spe-cific geometry. Secondly, geometrical stability which meets the requirements from the end user was also considered as an important task. Thirdly, ensure that the defects are within acceptable limits. Lastly, ensure a sound micro-structure which meets the aerospace standard mechanical requirements after heat treatment.
4.2.1 Experimental work Task 2.2
The results from Task 2.1 were taken as input to perform more complex castings with even thinner wall thickness. For simplicity, a combination of the geometry used for the casting trials for Simple Geometry 2 with a one side texture surface was used. Simulations were performed to decide the optimal parameter set up to be used for the casting trials. In these trials special considerations were taken to the feeding during solidification to reduce the defect level in the casting. The feeding was accomplished by adding wedges at one side of the blade, seeFig. 5. The size and positions were based on numerical simulations. Two castings were mounted on a two armed gating system with the same dimensions as in previous trials. A bottom gated casting system with a sprue height of 300 mm was used. The manufacturing of wax patterns and shells followed the same procedure as in Task 2.1, seeFig. 6. CbCu7-1, cast analogy of the stainless precipitation-hardening steel 17-4PH was used for the casting trials. Insulation of the thin sections was used to decrease the temperature drop during the time from removal from the pre-heating furnace to pouring. Within the frame of the project, a demonstrator of the GKN geometry was cast. The wax patterns were produced at FRI and shipped to TPC for shelling and casting. The inlet and gating system were designed at TPC. The shell manufacturing followed the same procedure as previous and CbCu7-1 was used as alloy.
Figure 5: Illustrating feeding wedges added to the blade.
(a) (b)
(c) (d)
Figure 6: Manufacturing of the shells for the Complex Geometry.
4.3 Task 2.3 - Post processing of investment castings
The possibility to further improving the final quality of the resulting castings by post-processing was investigated in Task 2.3. Among several possible methods, Hot Isostatic Pressing (HIP) was chosen as the most promising technique. The fabrication of investment castings is prone to residual internal defects, resulting in a reduction of fatigue strength and creep resistance, but also in an increased scatter in mechanical data. HIP treatment of the casting, a method known as casting densification, can increase the average level of these properties significantly and reduce the statistical scatter. The method can be used to remove gas porosity as well as shrinkage porosity.
4.3.1 Experimental work in Task 2.3
Investment castings for HIP trials were produced at TPC using a bottom up pouring system and a blade thickness of 2 mm. The HIP was performed at FRI using a pressure of 1500 bar and a temperature of 1180◦C for five hours. Porosity evaluation by X-ray images and 3D X-ray tomography was performed on the castings before and after HIP for comparison. Tensile testing of the blades which were subjected to HIP treatment was compared to the non HIP treated samples. Also the dimensional changes of HIP treated blades were investigated.
5 Detailed description of work in WP 3
CbCu7-1, the cast analogy of the stainless precipitation-hardening steel 17-4PH was fully in-vestigated in WP2. The main objective of WP3 was to evaluate and rank the castability of at least 3 wrought steel alloys not commonly used today as cast material and compare them with CbCu7-1. The work in WP3 was divided into two different tasks. Task 3.1 focused on ranking and optimising the castability of selected alloys.
The alloy selection process was conducted in collaboration with the TM. This work also focused on the final properties of the castings, i.e. the development of heat treatment parameters, tak-ing defects into account and investigattak-ing corrosion and wetttak-ing properties. The goal was to condense the number of alloys by excluding the worst performing alloy after each test. In Task 3.2 these optimised conditions were applied by casting a thin walled demonstrator of the most promising alloy.
5.1 Task 3.1 - Casting of selected alloys with simple geometries
A literature survey was first conducted which was supposed to give enough information about which alloys that was shown to be the most promising ones. However, it was shown that the information in literature was inadequate. Therefore the alloys were selected by TM and the project management. The following alloys were chosen as candidates at the beginning of the project:
• 17-4PH (reference alloy) • Custom 465
• PH13-8M • CSS 42L
However, after discussions two additional alloys were added to the evaluation: • JETHETE 152M
• L0H12N4M (steel alloy commonly used in Poland) 5.1.1 Experimental work in Task 3.1
Fluidity tests were performed for the selected alloys under laboratory conditions. The test geom-etry consisted of bars with different thicknesses (2-10 mm), seeFig. 7The casting was performed with a bottom gated casting system. All bars (2-10mm) were mounted on a circular bottom plate. The fluidity for each alloy was measured as the filled distance of each bar. The whole casting procedure consists of following steps:
• Melting and overheating
– Chemical composition measurement – Chemical composition correction • Slagging and deoxidisation
• Temperature measurement – Final chemical composition • Transfer of the mould to pouring stand • Metal tapping
Figure 7: Test geometry used for fluidity test.
Chemical composition was taken for each batch and the results are shown inTab. 7. Differential Scanning Calorimetry (DSC) was used for establishing liquidus and solidus temperatures for each selected alloy. DSC was also used to measure the specific heat. Dilatometer was used to measure thermal expansion and Laser flash technique was used for thermal conductivity mea-surements. These thermo-physical properties were later used as input to the simulations. Also wettability, corrosion and gasification tests were performed for some of the alloys.
In the casting trials the mould temperature was held constant to a temperature of 930◦C to 950◦C at the moment of tapping for all alloys. However, the tapping temperature differed to some ex-tent depending of the liquidus temperature of the alloy. A summary of the liquidus and solidus temperatures along with tapping temperatures for the alloys are presented inTab. 8. After cast-ing, each alloy was heat treated and subjected to mechanical testing at both ambient and at an elevated temperature of 400◦C. The heat treatment cycle differed among the alloys, see [8]. Because each alloy differed in chemical composition also different heat treatment routes were needed for each alloy.
The results from the fluidity experiments, illustrated inFig. 8, show an almost linear relation between fluidity and thickness for most of the alloys up to 4 mm. All alloys, except PH13-8M
Alloy C Mn P S Si Cr Ni Cu Ni+Ta Mo V Co Al 17-4PH 0.04 0.75 0.012 0.009 0.503 16.39 4.74 3.97 0.173 CSS 42L 0.21 0.80 0.500 18.85 3.00 3.90 0.40 5.50 Jethete 152M 0.14 0.45 0.135 11.87 3.15 2.29 0.18 PH13-8M 0.04 0.46 0.01 0.01 0.200 10.40 8.45 2.05 0.52 Custom 465 0.02 0.35 0.008 0.006 0.255 12.15 10.46 1.26 L0H12N4M 0.06 0.54 0.013 0.007 0.59 12.56 4.35 0.92
Table 7: Chemical composition of the selected alloys.
Alloy TLiquidus(◦C) TSolidus(◦C) TTapping(◦C) TSuperHeat(◦C)
17-4PH 1478 1444 1640 161
PH13-8M 1480 1464 >1700 220
CSS 42L 1477 1362 1640 163
JETHETE 152M N/Aa N/Aa 1580 N/A
CUSTOM 465 1479 1464 1650 171
L0H12N4M N/Aa 1485 1620 N/A
a outside measuring limit
Table 8: Liquids and solidus measurement for alloys.
and L0H12N4M, showed an improved fluidity using a pouring height of 10cm. Even though the alloy PH13-8M was shown to have good fluidity properties it was excluded for further testing due to the need of an extremely high temperature for flowing.
To get the alloy to flow properly a tapping temperature at or above 1700◦C was needed. At this temperature it is well known that the metal dissolves a lot of gases and the risk of defects thereby increases drastically. A degradation of the lining in the furnace (Al2O3) was also ob-served at this temperature. CSS 42L was excluded due to extremely poor fluidity properties and that the alloy was shown to be prone to hot cracking. Taking this into account the ranking from the melting- and castability experiments showed that JETHETE 152M had the best fluidity followed by Custom 465, L0H12N4M and 17-4PH.
The remaining alloys were subjected to micro-structure investigations as well as mechanical testing. After heat treatment, all alloys showed a dendrite martensitic structure with islands of delta-ferrite. The mechanical testing showed that all tested materials have a tensile strength within aero specifications. Based on these tests, JEHTETE 152M was shown to have the best mechanical properties followed by L0H12N4M, 17-4PH and Custom 465. Even if JETHETE 152M was introduced as an alternative alloy at the beginning of the project and was shown to have good fluidity and mechanical properties it was excluded for further testing because of in-sufficient corrosion resistance. This decision was taken in consensus at a coordinate meeting with the TM and all project partners involved.
Taking all results into account and also making a simple estimation of stiffness, weldability, and machinability 17-4PH was considered as the highest ranked alloy, followed by Custom 465 and L0H12N4M.
1 2 3 4 5 6 7 8 9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Bar width (mm) Fluidit y (-) 17-4PH 10cm 17-4PH 30cm CSS42L 10cm CSS42L 30cm Jethete M152 10cm Jethete M152 30cm
Figure 8: Illustrates fluidity results for 17-4PH, CSS42L and Jethete M152.
1 2 3 4 5 6 7 8 9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Bar width (mm) Fluidit y (-) Custom 465 10cm Custom 465 30cm L0H12N4M 10cm L0H12N4M 30cm PH13-8M 10cm PH13-8M 30cm
Figure 9: Illustrates fluidity results for Custom 465, L0H12N4M and PH13-8M.
5.2 Task 3.2 - Final casting of selected alloys with complex geometry
The objectives of Task 3.2 were to evaluate castability of the alloys chosen in Task 3.1 for more complex geometries and find a process for the manufacturing of the model casting (demonstra-tor).
5.2.1 Basic experimental work in Task 3.2
Additional experimental work was performed in Task 3.2 to investigate gasification and wettabil-ity of mould materials and corrosion resistance of selected alloys. The alloys tested were 17-4PH from both FRI and TPC along with L0H12N4M. Custom 465 was excluded from further analysis because of availability. The interaction of the ceramic mould and liquid alloys was investigated
using a unique apparatus for high temperature studies because such phenomena as gas release, chemical reaction and wetting of mould material are responsible for porosity formation as well non-metallic inclusions in castings. Two ceramic shell systems, taken from TPC and FRI, were examined with respect to gasification. Wetting phenomenon is taking place between molten al-loy and ceramic mould when the contact angle formed at the triple point melt/substrate/gas is below 90◦.
InFig. 10the wetting behaviour for 4PH from TPC and FRI are shown. As can be seen 17-4PH from TPC shows a non-wetting behaviour while the alloy from FRI shows wetting. This is an additional factor affecting quality of the final castings since it is responsible for liquid metal penetration inside porous mould accompanied with its fragmentation and detachment, causing introduction of non-metallic inclusions into the melt.
(a) 17-4PH TPC (b) 17-4PH FRI
Figure 10: Illustration of the wettability test.
Moreover, in most metal/ceramic system wetting has reactive character related with the forma-tion of wettable reactive product at the interface. However, this effect may cause strong bonding between the mould and the casting, dimensional changes of the mould as well as local detach-ment of the reaction product and its introduction to solidifying alloy (“washing” effect), all contributing to the formation of casting defects. The wettability tests were performed for every combination between the shell systems used at TPC and FRI in combination with two different chemical compositions of 17-4PH (TPC and FRI) and L0H12N4M. The sessile drop method was applied for measurements of the contact angle values versus time using contact heating procedure and high-speed CCD camera with 100 fps equipped with LabView software for drop image acquisition. The wettability tests lasted for four minutes.
All investigated alloys showed high reactivity in 4-min contact with all mould materials result-ing in the formation of reaction in the region under the drop. The wettresult-ing results presented in
Fig. 11to13showed that the ceramic mould material from TPC was non-wettable for all inves-tigated alloys. Thus one may expect that this material has also lower reactivity with examined alloys. On the contrary, under identical testing conditions, the mould material from FRI showed a larger wetting tendency, compared to the ceramic shell from TPC. More detailed structural characterisation of reaction product region is needed in order to understand the mechanism of interaction between the liquid metal and the mould material.
Using the same apparatus and by the method of large drop, measurements of surface tension and density of liquid metal of 17-4PH steel alloy were performed. From these measurements a surface tension of σ = 1560 mN/m and a density of ρ= 7.00 g/cm3was found.
Based on the preliminary results of wettability studies the following conclusions were made: • The best combination is: 17-4PH TPC alloy with mould from TPC
• The worse combination is: 17-4PH FRI alloy with mould from FRI
(a) FRI Shell (b) TPC shell
Figure 11: Results from the wettability for 17-4PH TPC.
(a) FRI shell (b) TPC shell
Figure 12: Results from the wettability for 17-4PH FRI.
Gasification of the ceramic shells was estimated by continuous registration of the amount of residual gases using a Quadrupole spectrometer QMS 200 during heating of as received mould samples to a temperature of 1200◦C. The following parameters were recorded during the cycle from ambient temperature, heating and cooling:
• Pressure inside the chamber • Composition of residual gases
(a) Shell FRI (b) Shell TPC
Figure 13: Results from the wettability for L0H12N4.
The results from the gasification are shown inFig. 14for the shells from TPC and FRI respec-tively. Both types of ceramic mould materials showed gasification during heating to processing temperature, particularly from the mould delivered by TPC. However, proper preheating helps significantly reduce gas release from the moulds thus special attention should be paid to the mould treatment directly before the contact with molten alloys.
Corrosion tests were performed using the same alloys as for the wettability tests for a time period of 100 hours. The corrosion tests were performed in a new developed thermo-chamber using a saline atmosphere. The corrosion tests continued for 100 hours at a temperature of 400◦C in a chamber with an atmosphere containing a 5% salt fog. The following materials were subjected to the corrosion test.
• 17-4PH from TPC • 17-4PH from FRI • LOHI2 from FRI
It was clear from these tests that the salt fog accelerates corrosion for all studied samples. The alloy with the lowest observed corrosion rate was L0H12N4M, even if the difference was small. The formed oxide surface during the test consisted of iron oxide (FeO) with small quantity of Cr (6-16%).
5.2.2 Complex geometry casting experiments
After laboratory testing the aim was to cast a demonstrator of the GKN geometry given in the proposal. It was decided to carry out the casting in two stages by first casting 1/16 and later 1/4 of the full casting. The geometries are illustrated inFig. 15(a) to15(c). The thin sections of the GKN casting had a nominal thickness of 2.0 mm. Due to a combination between acces-sibility and taking all properties into account it was decided to use 17-4PH as cast alloy. The chemical composition of the alloy used in the casting trials is shown inTab. 9. Before casting, excessive work was carried out designing the inlet and gating system. This was accomplished using MagmaSOFT and FLOW3D. The material data used in the simulations were compiled from both experimental measurements and calculations in JMatPro. From the simulations it was shown that important factors for filling thin walled investment casting are among several; mould temperature, pouring temperature, satisfactory venting of the cavity for good gas evacuation and providing laminar flow of metal during filling.
(a) TPC shell
(b) FRI shell
Figure 14: Results from the gasification tests.
The final inlet system for the 1/16 casting is illustrated inFig. 16which is based on the simula-tion results inFig. 17. The inlet system and ceramic shell for the 1/4 GKN casting is illustrated inFig. 18.
Fe C Si Mn P S Cr Ni Cu Nb+Ta Mo
Bal. 0.02 0.52 0.65 0.025 0.002 15.93 4.28 3.23 0.21 0.14
Table 9: Composition of cast alloy.
Silicon rubber was used as tooling material for the wax patterns. This method is used for proto-type making to prevent high costs of metal tools and allows for manufacturing about a dozen of wax patterns. According to the designed concept, separate wax models were connected to a gat-ing and feedgat-ing system. This part of manufacturgat-ing was performed manually. Ceramic moulds were produced by consecutive dipping and stuccoing the wax assemblies until the desired thick-ness of the shell was achieved.
(a) GKN-Geometry (b) 1/16 part (c) 1/4 part
Figure 15: Illustrating the complex geometries used in Task 3.2.
(a) (b)
Figure 16: Illustrating the final gating and inlet system after optimisation for the 1/16 part geometry.
Due to FRI secrets no information about the recipe of the slurries used can be made public. After the investment the shells were left to dry and later subjected to de-waxing in an industrial autoclave. After the subsequent sintering had been accomplished to harden the shells they were preheated and ready to be poured. The shells were placed in a metal container and the shells were surrounded by sand, seeFig. 19. This served two purposes. Firstly the surrounding sand gave support to the shell and thereby decreased the risk of shell cracking during filling and solid-ification. Secondly, the surrounding sand also served as an extra insulation which minimised the temperature drop of the shell during transport from the furnace to the pouring area. The process parameters used in the casting trials were finally decided to:
• Mould temperature: 950◦C • Tapping temperature: 1650◦C
Miss-runs were present in the 1/16 of the demonstrator as can be seen inFig. 20. None of the castings had a complete fill. Miss-runs were seen at different locations of the thin parts of the casting. It was also obvious from the results that the design of the inlet system resulted in that two melt fronts met at an outer edge, a position with a high potential of miss-runs.
(a) (b)
(c)
Figure 17: Illustrating simulation of filling (Flow3D).
(a) (b)
Figure 18: Illustrating, (a) the final gating and inlet system after optimisation for the 1/4 part geometry and (b) the final shell.
In Fig. 21, a comparison between X-ray computer tomography and defect simulation is pre-sented for the 1/16 of the demonstrator casting. The casting of the 1/4 of the demonstrator was
Figure 19: Illustrating shells embedded in sand.
also unsuccessful, seeFig. 22. During filling effects of gas explosions were observed indicating that a high degree of moisture was left in the shell during filling. A main factor responsible for successful filling of thin-walled investment castings seems to be castability of the alloy. There are several factors which influence alloy castability and not all of them have been investigated during this project.
In investment casting, there are basic factors which should be controlled during the process, most of them are temperature parameters but also such factors as air humidity, chemical composition, ceramic shell parameters and the presence of oxides or impurities should be taken into account. Anyway, casting of thin-walled castings from steel alloys is till now not quite recognized and still requires additional investigations and both practical and simulation based experiments (in the range of gating and feeding system configurations to ensure a laminar flow of melt into cav-ity). Information relating to case studies is still poor. Performed simulations gives only a picture of some parts of the process, but a wide range of input parameters should be confirmed since confrontation with real results often require a lot of practical corrections.
(a) (b)
(a)
(b) (c)
Figure 21: X-ray Computer Tomography and simulated porosity of 1/16 part geometry.
Figure 22: Illustrating the results from the casting trials for 1/4 part geometry.
The cast element made from the certified 17-4PH alloy was examined for the structure homo-geneity after heat treatment according to Tab. 10. The samples collected from different parts of the element (1/16 of the whole demonstrator) were examined by metallographic technique, seeFig. 23. The metallographic observations confirmed the homogeneity of the sample micro-structure, which consisted of a delta ferrite embedded in a martensitic matrix with some non-metallic inclusions.
Temperature (◦C) Time (min) Cooling Homogenization 890 120 Furnace Under critical annealing 700 360 air
Hardening 1050 60 oil
Tempering 480 180 air
Tempering 480 180 air
Table 10: Heat treatment sequence.
(a) 100x (b) 500x
Figure 23: Illustrating micro-structure of the 1/16 GKN casting.
6 Conclusions
6.1 Conclusions of WP2 6.1.1 Fluidity
In the statistical evaluation the projected area of the blades was used as a measure of fluidity. Generally, the top gated casting system showed a larger degree of filling compared to bottom gated system for Simple Geometry 1. The pouring temperature followed by the thickness had the largest impact on fluidity for the top gated system while the opposite behaviour was shown for the bottom gated system. Both the pouring temperature and the thickness had larger impact on fluidity for the top gated system compared to the bottom gated system.
For Simple Geometry 2 the differences between top- and bottom gated casting systems showed only minor differences in filling. However, the top filling system was shown to be much more sensitive to chosen process parameters than the bottom filling casting system. The top filling system also showed a somewhat larger scatter in the fluidity results compared to the bottom fill-ing system. For the top fillfill-ing system the pourfill-ing temperature and the fillfill-ing time showed to be the most important parameters. The bottom filling system showed only minor or no influence on the chosen parameters. By using a one side textured side of the blades it was shown that the thickness could be reduced for the same fluidity which corresponds to a weight reduction for minimum thickness panels of 15-20%.
The filling results of the Complex Geometry showed a complete filling of all castings. X-ray in-vestigation of the blades showed that the defects in the thin sections had been heavily decreased. However, some minor cracks and surface defects could be observed in the castings.
Miss-runs were observed for casting trials of the 1/16 part demonstrator cast at TPC Components AB. This could however be attributed to a deviation in section thickness. The measured thickness at the edge of the miss-run was in the range of 1.0 to 1.3 mm instead of the nominal 2.0 mm. The assisted vacuum casting experiments failed because enough low pressure was not achieved. This could be attributed to an excessive leakage. The problem was identified to the sealing between the casting chamber and the lid.
6.1.2 Micro-structure and defects in thin sections
Penetrant testing showed large differences between top and bottom gated castings with respect to the amount of large pores close to the surface for Simple Geometry 1. Average porosity area of the heat-treated samples varied from 2hfor the bottom gated system to almost 10hfor the top gated system. The larger pores were examined with SEM, which indicated a shrinkage mechanism. CT scanning of the blades showed a substantially larger porosity level for the top gated 2 mm blades at high pouring temperature compared to the bottom gated blades. But for the lower pouring temperature no difference was observed. Heat treatment had a drastic effect on microstructure with respect to morphology, grain size and amount of delta-ferrite. The amount of delta-ferrite in the bulk samples was heavily decreased upon heat treatment. In thin heat-treated sections, the amount of delta-ferrite decreased further due to different temperature response during cooling. There was no difference between top and bottom gated samples with respect to grain size or delta-ferrite. Also the resulting hardness was above the minimum requirements for ASTM A747A standard. The 17-4 PH castings from the trials in Simple Geometry 2 were heat treated following the specifications for condition H1025 in the aero standard ASM 5343E. The expected martensitic micro-structure was obtained.
A total of sixteen principally equivalent samples were tested for tensile properties. All samples had a yield and tensile strength equal to or above the minimum value according to the AMS 5343E standard. Four samples, however, had an area reduction below the minimum value. This could readily be explained by the occurrence of shrinkage porosity within the material.
6.1.3 Numerical simulations
Miss-run predictions and defect tracking was shown to be possible with simulations to a high degree of confidence for all geometries including the demonstrator. Accurate material properties for both the alloy and shell are however needed. Also, the knowledge of the correct temperature profile for the shell at the time of pouring was also shown to be important.
6.1.4 Post-processing
For the HIP treated blades the internal porosity level decreased considerably, as was shown by 3D X-ray tomography. The resulting material therefore shows less scatter in tensile properties compared to material that was not HIP treated. The amount of delta-ferrite in the microstructure was decreased due to the HIP process. This is a result of the extended homogenisation of the material during high temperature cycle and is generally believed to beneficial for the tensile properties of the material. It might be possible to improve the material properties of HIP treated 17-4 PH by optimising the total heat treatment process, increasing the HIP temperature or the temperature of the ageing step, thereby decreasing the yield strength and increasing the ductility of the heat-treated material. Due to the HIP treatment some dimensional changes was observed.
6.2 Conclusions of WP3
• Wall thickness limits which guarantee obtaining of good casting were observed during trials of thin-walled castings made from 17-4 PH alloy, thickness below 2 mm is not fully realisable.
• CT X-ray tomography is a very helpful tool for identification of internal defects (porosity, discontinues), especially in thin-walled castings.
• Best performance was observed for 17-4 PH alloy and ceramic mould made from used in TPC material (company secret).
• Gas release during the casting process may cause the formation of porosity in the castings. • HIP treatment improves mechanical properties and reduces porosity in ready castings, but
also causes slight dimensional deviations.
• Despite that first demonstrator trials did not give positive results is advisable to continuing the investigations to obtain a good thin-walled casting from steel alloys like 17-PH.
References
[1] R. Svenningsson, 2014-002 Thin walled investment castings, Literature survey, Swerea SWECAST (2011).
[2] R. Svenningsson, S. Farre, Å. Lauenstein, D2.3 - Casting trials for simplified 17-4PH in-vestment castings, Techical report, Swerea SWECAST (2013).
[3] Å. Lauenstein, 2014-001 Casting of precipitation-hardening steel, Literature survey, Swerea SWECAST (2012).
[4] Å. Lauenstein, Post-processing of cast 17-4 PH steel, Literature survey, Swerea SWECAST (2012).
[5] Å. Lauenstein, D2.4 - Post-processing trials of 17-4 PH castings, Techical report, Swerea SWECAST (2013).
[6] R. Svenningsson, S. Farre, D2.5 - Casting trials for complex 17-4PH investment castings, Techical report, Swerea SWECAST (2013).
[7] J. Przybylski, View on steel alloys due to their casting properties, Literature survey, FRI (2012).
[8] J. Przybylski, D3.2 - Documentation of first batch of cast trials, Techical report, FRI (2012). [9] J. Przybylski, D3.3 - Documentation of final casting trial for one steel alloy with limited
