Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Qualification of Nd:YAG laser direct metal deposition techniques for repair of nickel superalloy components (October 2006)

   

A A Lugan1 , P A Hilton1 , G B Melton1 and C Rinaldi 2

1 TWI Ltd, Cambridge, UK
2 CESI RICERCA SpA, Milano, Italy

Paper presented at 25th International Congress on Applications of Lasers & Electro-Optics (ICALEO 2006), 30 October - 2 November 2006, Scottsdale, AZ, USA

Abstract

The high temperature operating components of power utility gas turbines are made with gamma prime reinforced nickel based superalloys. As a result, there is a very high cost associated with the replacement of any parts damaged during service. The economic benefits to be gained by using repair procedures as opposed to replacement are considerable, with potential savings amounting to several million dollars per turbine. In this paper, the development of procedures for Nd:YAG laser direct metal deposition of nickel superalloys are presented. The paper describes the process by which high density and imperfection free deposition of the superalloy powder was achieved and goes on to describe production of a series of test specimens with geometries designed to qualify the quality of the deposition in terms of tensile, low cycle fatigue, and creep properties. The results of the tests show that the target of achieving high temperature mechanical properties in excess of 80% of the parent material properties was achieved.

Introduction

The high cost of replacing components suffering from wear and fatigue damage in the high pressure sections of industrial derivative and power utility gas turbines has led increasingly to end-user interest in the use of repair techniques. The most arduous conditions in gas turbines and aeroengines are experienced in the high pressure sections. Components which experience the highest temperatures are generally made from nickel superalloys, which have highcreep strength at elevated temperature. Superalloys are normally used for rotating turbine blades, nozzle guide vanes and also for some static components in the combustion and turbine areas. Nickel superalloy development began over 50years ago and during the past decade a range of nickel alloys has been introduced with compositions and mechanical properties tailored to suit the demanding requirements of the gas turbine. Unfortunately, the higher performance alloysare considered extremely difficult to weld because of their susceptibility to heat affected zone (HAZ) liquidation cracking, weld metal cracking and cracking during post weld heat treatment (PWHT). Even more pronounced problems are encountered with the introduction of directionally solidified and single crystal alloys.

It is widely accepted that during repair, a low heat input to the nickel alloy component is necessary to minimise HAZ and weld metal cracking. As a consequence the adoption of mechanised processes, such as laser welding or laserdirect metal deposition (laser DMD) or mechanical TIG welding and deposition, are key techniques for repair of nickel based superalloys.

The results reported in this paper relate to alloy IN738, which is a vacuum melted, vacuum cast, precipitation hardenable nickel alloy, possessing excellent high temperature creep-rupture strength and hot corrosion resistance. Itwas designed to provide the gas turbine industry with an alloy with creep strength up to 980°C, combined with the ability to withstand long time exposure to the hot corrosion environment associated with gas turbine engines. Thisalloy has 15 - 18% chromium content and a combined Al and Ti concentration of about 7.0% in weight. This yields a volume fraction of γ ' phase of about 47% after solution ageing heat treatment. Two versions of alloy IN738 are produced: a high carbon version designated IN738C and a low carbon version designated IN738LC. The low carbon (and lowzirconium) concentrations in IN738LC are needed for improved castability, particularly in large sections. The work reported here was performed using alloy IN738LC.

Brazing has been widely used for repairing corrosion and erosion damage in gas turbine components in this and other super alloys, but the concern regarding this repair process is the reduction of elevated temperature mechanical properties, as observed in creep-rupture tests. Weld repair of gas turbine alloys is limited principally by the tendency to form cracks, either during welding or PWHT. The precipitate-forming elements, mainly aluminium and titanium,that are added for high temperature strength are primarily responsible for this problem. IN738 is generally considered to have 'poor weldability' purely from a compositional viewpoint. Gregori [1] has recently surveyed the welding and repairing of nickel superalloys for gas turbines.

TIG and laser weld repair have been carried out on alloy IN738LC using various filler metals such as Inconel 625, but this has only about 25% of the high temperature tensile strength of IN738LC, and creep-rupture tests on refurbished blades show that TIG joints possess less than 60% of the strength of the base material. [2,3] It has been shown that controlling the cooling rate of the weld bead is fundamental to obtaining a crack-free HAZ when using a matching consumable. [4] This technique however, employs a high preheat to eliminate cracking and presents several practical problems associated with the use of temperatures up to about 1000°C.

A laser processing technique has been applied to IN738LC with Inconel 625 filler powder, [3] but the mechanical characterisation of the joints showed that the weld metal had properties such that it was not applicable to the hot sections of a turbine, The use of precipitation hardenable weld filler that more closely matches the component alloy composition offers the possibility of extending blade repair to high stress areas of the turbine. Good mechanical properties have been achieved in IN738 with multipass laser repair welding using an IN939filler powder. [5] The use of this filler material and a careful control of heat input, preheat and PWHT, produced properties higher than those achieved employing Inconel 625 consumables.

Alloy IN738LC powder filler material has also been used to restore the tip of an IN738 directionally solidified first stage turbine bucket, using a CO 2 laser powder fusion process. [6] The results showed that a low heat input laser process could produce porosity and crack-free welds, with narrow HAZ.

Published results therefore indicate that weld repair can be performed on IN738 using laser welding employing IN939 or matching filler, although the results show stress rupture properties about 20% lower than those for the parent material, and the techniques employed require a high (900 - 1000°C) preheat. It would be beneficial to exclude the need to preheat before treatment and the results presented in this paper describe work using an Nd:YAG lasersource, to develop welding procedures with very low heat input, for repair of IN738 using a matching powder filler, without the need for high temperature pre heating. The target was to produce mechanical properties in the deposited material in excess of 80% of those in the parent material. To validate the process, high temperature tensile, low cycle fatigue (LCF) and creep testing were employed. To the authors knowledge there is no previously reported LCF datafor such repairs.

Specimen preparation

Test specimens were manufactured from cylindrical blanks of Alloy 738LC parent material supplied in the form of cast bars 0.5 in (12.5mm) diameter ´ 6 in (150mm) long. The material was in the solution heat treated (1120°Cfor 2 hr) and hot isostatic pressed (HIP) condition.

A standard blank test specimen geometry was designed which would allow specimens to be clad and then machined to size for any of the required tests, see Fig.1 and 2. This required the initial diameter of the clad length to be just 3mm, over a nominal length of 15mm. The machined blank was then clad with a deposited layer approximately 1.5mm thick to provide a specimen of approximately6mm in diameter. Final test specimens were then machined back to a diameter of 5mm from these clad samples.

Fig.1. Drawing of test blank
Fig.1. Drawing of test blank
Fig.2. Test blank, tensile, LCF and creep specimens
Fig.2. Test blank, tensile, LCF and creep specimens

An unusual creep test specimen with an axial hole drilled through the bar to remove parent material inside the deposit was chosen, so that only deposited material and the HAZ would be tested without the influence of parent material.This presented problems in machining but electro-spark proved to be successful. The diameter of the hole was established during the procedure trials. A two layer sample was sectioned and from this section it was concluded that a 1.6mmdiameter hole would remove the parent material leaving only a cylinder of clad material and the HAZ, supported at each end by parent material. The total thickness of material tested was therefore 1.7mm.

The matching alloy 738 consumable powder was supplied by Praxair as NI-284-9. Following cladding, a post weld heat treatment (PWHT) was performed in a vacuum furnace and consisted of:

  • Annealing at 1120°C for 2 hours followed by forced gas cooling to room temperature.
  • Ageing at 845°C for 24 hours followed by forced gas cooling to room temperature.

Experimental details

The laser deposition process was carried out using a Sulzer Metco powder feed unit and a TWI built single jet powder delivery nozzle. The particular powder feeder unit mentioned above operates well at the low powder mass flow rates used for laser deposition. The powder delivery nozzle consisted of a 120mm long, 3mm outside diameter, 1.2mm inside diameter, copper tube. The length to internal diameter ratio was chosen to promote smooth, consistent powder delivery.The nozzle was angled at 60° from the horizontal, based on previous optimisation studies. The tip of the nozzle provided the capability to water cool the device during operation. The laser beam used was generated by a TrumpfHL3000D, 3kW CW lamp-pumped Nd:YAG laser, fed to the workpiece by a 0.6mm diameter optical fibre. 1:1 imaging was used in the process head, so that a lens to minimum beam waist working distance of about 250mm was available. The focusing laser beam was surrounded by a coaxial gas shield. The focusing lens was protected by a glass cover slide. For the work reported here, the process head was held at the end of an articulated arm robot and the samples were held in a CNC controlled rotary fixture. All the specimens produced for mechanical tests with this system, were manufactured in a pure argon atmosphere, created by shrouding the sample in a metal sided enclosure and linking this to the coaxial gas delivery nozzle on the process head described earlier. The powder delivery nozzle and laser beam delivery system can be seen in Fig.3. With this optical system, the position of the minimum beam waist could be moved vertically up and down by moving the beam focusing lens. As the powder delivery nozzle was attached to the beam delivery system, this meant that once the impingement position of the powder stream was set with respect to the sample surface, the laser spot size to be used could be chosen without moving the powder delivery nozzle.

Fig.3. Photograph showing the powder delivery nozzle attached to the laser beam focusing optics incorporating a coaxial gas shield
Fig.3. Photograph showing the powder delivery nozzle attached to the laser beam focusing optics incorporating a coaxial gas shield

Powder deposition trials

Early deposition trials on bars of IN738LC, 12mm in diameter indicated the potential problems of liquation cracking which can be found with welding many nickel superalloys, as can be seen in Fig.4, which shows cracking in the HAZ between the deposit and the substrate material. Surface breaking cracks were also evident in this early work A photograph of a typical deposit at this stage of the work can be seen in Fig.5, showing a singe layer spiral deposit. In all the work reported in this paper, the samples were rotated and translated under a fixed laser beam. Travel speed along the length of the sample was adjusted to give a onethird overlap between each deposit. Following this initial work a second series of trials was made using significantly lower heat to the sample, with better results. In this part of the work a design set of experiments was used withthe variables, laser power, P, (range 300 - 500W), powder feed rate, M, (range 30 - 45g/min) and travel speed, V, (range 0.9 - 1.2m/min). Parameter combinations were chosen with small ranges of P/V and M/V, of 24 - 30J/mm and 30 -37g/m, respectively. This resulted in 17 sets of combinations in the experiment. For the 12mm diameter bar, only one of the set of combinations produced porosity and liquation free deposits. A transverse section from this sample can beseen in Fig.6. A cosmetic 're-melt' pass of this sample removed the spatter on the top of the sample as can also be seen in Fig 6. The samples shown in Fig.6 were produced with P/V = 27.5J/mm and M/V = 30.0g/m.

Fig.4. Liquation cracking at the HAZ between deposit and parent material
Fig.4. Liquation cracking at the HAZ between deposit and parent material
Fig.5. Typical spiral deposit from early work
Fig.5. Typical spiral deposit from early work
Fig.6a) Transverse section showing elimination of porosity and cracking
Fig.6a) Transverse section showing elimination of porosity and cracking
Fig.6b) Similar section but with additional removal of surface spatter
Fig.6b) Similar section but with additional removal of surface spatter

These conditions were then transferred to samples machined to the required geometry of the test specimens, this being 3mm diameter in the region of the deposit. Keeping the P/V and M/V values as recorded, for the samples shown in Fig.6, a range of deposits were made based on a laser power of 125W, with a travel speed and a powder feed rate calculated for the new substrate diameter of 3.0mm. A track thickness of about 0.8mm could be achieved, whilestill providing a crack and porosity free deposit, as can be seen in Fig.7.

Fig.7. Crack and porosity free deposit ~0.8mm thick on the 3mm diameter test specimen
Fig.7. Crack and porosity free deposit ~0.8mm thick on the 3mm diameter test specimen

In order to produce the required test specimens a deposit layer thickness of 1.5mm was required. Further work was thus directed at optimisation of a second deposited layer on top of the first layer. To compensate for the increase in diameter of the second layer and keep P/V and M/V constant, the process parameters for the second layer were changed slightly from those used for the first layer. Using two sets of parameters, one for each layer, 25 samples weresubsequently produced for mechanical testing. Following deposition, the samples were heat treated and then machined to final dimensions. After machining, each sample was radiographed and subjected to fluorescent dye-penetrant inspection (FPI). No significant indications of imperfections were seen, other than those attributable to machining marks. Nine additional specimens were also machined from parent material for comparative tests.

Testing

The following high temperature tests were performed:

  • Tensile testing at 850°C
  • Creep tests at 850°C in air on both parent material and weld repair specimens. Specimens with a central hole of 1.6mm diameter were used to eliminate the base material at the core of the sample and to test a true cross weld situation. Stress varied from 150-300MPa and the duration of the tests was up to 5000h for the parent material and 3570h for the weld repairs.
  • The LCF tests were performed at 850°C on both parent material and weld repair specimens. including continuous cycle (LCF-CC) tests and LCF tests with tensile hold time (LCF+HT), to evaluate the material behaviour in creep-fatigue interaction conditions. LCF tests were carried out in accordance with ASTM E606-92.

Results and discussion

The results from the high temperature tensile testing at 850°C showed good repeatability. Proof strength and tensile strength varied from 479-498MPa and 639-662MPa, respectively. Elongation and area reduction varied from 3-14%and 8-26%, respectively. Comparison with literature values for Alloy 738 parent material, Fig.8, showed that the tensile properties of the laser repair welds at 850°C are in the range 80-85% of the values for the base material. Examination of the tensile samples indicated that failure occurred in a ductile mode in the gauge length. Most failures occurred through the weld and HAZ, although one sample failed in the parent material, as shown in Fig.9, and not in the weld/HAZ.

Fig.8. Summary of the high temperature tensile testing results and comparison with literature data for parent material and weld repair. Filled symbols are the average values at 850°C for the weld repairs and the empty symbols and dashed lines represent the typical data for Alloy 738 parent material (PM) from Ref. [1,10]
Fig.8. Summary of the high temperature tensile testing results and comparison with literature data for parent material and weld repair. Filled symbols are the average values at 850°C for the weld repairs and the empty symbols and dashed lines represent the typical data for Alloy 738 parent material (PM) from Ref. [1,10]
Fig.9. Metallographic examination of the tensile sample showing failure in the parent material
Fig.9. Metallographic examination of the tensile sample showing failure in the parent material

A summary of the results of the creep tests for both parent material and weld repair at 850°C and data from the literature for various weld repair processes are reported in Fig.10, on a Larson-Miller stress-rupture parameter plot. The creep rupture properties of the laser weld repairs produced in this work are compared with typical values for TIG repair using Alloy 625 consumable [7,8] and some results obtained on laser clad specimens produced using type 625 filler. [3,9] The presence of the hole in the creep specimens used in this work, seems to reduce slightly the creep strength values. Fig.11 compares the results obtained for parent material specimens with a hole, with standard test data for the parent material obtained from the literature for Alloy 738. [1,10] The results show that laser weld repairs produced using a matching powder consumable are very good with respect to both previous laser repairs and standard TIG repairs obtained using a type 625 filler. Comparison with the stress-rupture data for the parent material ( Fig.10) showed that the reduction of stress-rupture strength of weld repair samples is about 20% for a test duration of up to 3500 hours.

Fig.10. Larson-Miller stress-rupture parameter plot comparing weld repair with literature data
Fig.10. Larson-Miller stress-rupture parameter plot comparing weld repair with literature data
Fig.11. Larson-Miller stress-rupture parameter plot comparing center-hole creep samples for the parent material with literature data
Fig.11. Larson-Miller stress-rupture parameter plot comparing center-hole creep samples for the parent material with literature data

The results of the LCF tests are summarized and compared to data reported in the literature in Fig.12. The results obtained with and without a tensile-hold cycle gave very similar results. Weld repairs showed a five-times reduction of endurance compared with the parent material. However, comparison with LCF literaturedata for Alloy 738LC at 850°C [9,11] shows that the results from the weld repair samples fall within the scatter band for the parent material data.

Fig.12. LCF results comparing weld repair with parent material and literature data
Fig.12. LCF results comparing weld repair with parent material and literature data

A review of the available literature has shown that the use of the repair technique developed in this work, gives weld metal properties that match adequately the requirements of the parent material. The most frequently used superalloy filler metal, Inconel 625, has only about 25% of the high temperature tensile strength of Alloy IN738LC and the properties of conventional TIG weld repairs on Alloy IN738LC possess less than 60% of the strength of the base material.

Conclusions

This project has demonstrated that Alloy 738LC can be repaired using a matching consumable at room temperature, without the need of preheat. The objective of achieving high temperature mechanical properties in excess of 80% of the parent material was achieved.

  • The high temperature tensile properties at 850°C were 80-85% of the parent material properties.
  • The creep resistance showed a stress reduction of less than 20%.
  • LCF test results on weld repair specimens fell around the lower limit of the scatter band of the parent material data. Tests with tensile hold time gave a very similar response to tests without.

Acknowledgement

This work was funded by a TWI Group Sponsored project. The sponsors were Ansaldo Energia SpA, CESI Ricerca SpA and Portland General Electric Company. The sponsors are thanked for their permission to publish this paper.

References

  1. Gregori, A. 'A survey of welding and repairing of nickel superalloys for gas turbines'. TWI Members Report from its Core Research Programme, no. 774 August 2003.
  2. Keienburg, K., Esser Wand Deblon, B. 'Refurbishing procedures for blades of large stationary gas turbine'. Materials Science and Technology, Vol. 1, 1985 pp. 620-628.
  3. Rinaldi, C., Gavito, L., Gerelli, I., Singleton, P., Walker, P., Guardamagna, C., Moscotti, L., Bayard, P., Bertoli, A. 'Automatic refurbishment of gas turbine components by CO2 robot laser'. Proc. Conf. 'Materials Solutions 1997 on joining and repair of gas turbine components'. Indianapolis, USA, 15-18 September 1997 109-117. Publ. ASM, Materials Park, USA, 1997.
  4. Foster, M., Updegrave, K. 'Welding superalloy articles'. US Patent No. US6333484B1, Chromalloy Gas Turbine Corp.
  5. Sandy, D., Frederick, G., Peterson, A., Stover, J., Viswanathan, R. 'Development of a laser-based/high strength weld filler process to extend repair limits on IN738 gas turbine blades'. Proc. Conf. 'Welding and Repair Technology for Power Plants - EPRI 2000', 7-9 June 2000, Marco Island, USA. Publ: Electric Power Generation Institute, Charlotte, USA, 2000. Paper P05-GI.
  6. Hu, Y. P., Reynal, V., Ramundo, F., Dainiak, G. A. 'Laser powder fusion repair of industrial gas turbine blades using IN738LC nickel based superalloy'. Proc. Conf. 'Welding and Repair Technology for Power Plants - EPRI 2000' 2000. Marco Island, USA. Publ: Electric Power Generation Institute, Charlotte, USA, 2000. Paper P02-GI
  7. Stover J T and Gandy D V: 'Status of weld repair technology for Ni-based superalloy Gas Turbine Blading'. EPRI Report TR-108272, 1998.
  8. Esser W, McLean M, Schneider K: 'Effect of fabrication and repair procedures on the performance of Alloy IN738LC and IN939'. Proc. Conf. 'High temperature alloys for gas turbines and other applications 1986', Liege, 6-9 October 1986, Part1, pp.593-622. Commission of the European Communities, 1986, EUR 10567. Publ: D Reidel Publishing Company, Dordrecht, The Netherlands, 1986. ISBN 9027723044.
  9. BRITE EURAM Project BE- 4314, 1994-1997.
  10. 'Alloy IN-738 alloy - Technical data'. INCO Report 497. Publ.: The International Nickel Company Inc, One New York Plaza, New York, USA, 1981.
  11. Strang A: 'High temperature properties of coated superalloys'. Proc. Conf. 'Behaviour of high-temperature alloys in aggressive environments', London, 15-18 October 1979, pp.595-611. Publ: The Metals Society, London, 1980. ISBN 0904357309.

For more information please email:


contactus@twi.co.uk