The Technology of Manufacturing Turbine Wheels for Turbomachines Using Selective Laser Melting

Konstantin MikhailovBaltic State Technical University “Voenmeh” named after D. F. Ustinov, St. Petersburg, 190005, Russia

Alyona KirshinaBaltic State Technical University “Voenmeh” named after D. F. Ustinov, St. Petersburg, 190005, Russia

Anton KirshinBaltic State Technical University “Voenmeh” named after D. F. Ustinov, St. Petersburg, 190005, Russia

Kirill PereshiloDepartment of Control Systems and Robotics ITMO University, St. Petersburg, Russia

Abstract — We review methods of manufacturing turbine wheels for small-­sized turbojet engines using Selective Laser Melting technology and analyze causes of defect occurrences.

Keywords: selective laser melting; small turbojet engine; blade machines, turbine wheels

© The Authors, published by CULTURAL-EDUCATIONAL CENTER, LLC, 2020

This work is licensed under Attribution-NonCommercial 4.0 International

I. Introduction

The shift to universal digitalization and automation of production paves the way to the qualitative leap referred to as the Fourth Industrial Revolution or Industry 4.0.

The key technologies of Industry 4.0 include the Industrial Internet of Things, total digitalization and Additive Manufacturing (AM) [1]. In relation to the engine manufacturing industry, AM is capable of replacing a number of traditional manufacturing processes in the long run, but in today’s engines we only see a limited share of AM parts.

In this paper we aim to define the direction of design and process adaptation of parts, units and assemblies that comprise gas turbine engines and gas turbine power plants.

The object of this study is a turbine wheel of a small turboshaft engine with a wheel diameter of 113 mm. The turbine is tested in a small turboshaft engine, as a part of its rotor that rotates at a rate up to 90,000 rpm, which dictates the need for the precise manufacturing of the said part [2].

Selective Laser Melting (SLM) is a very promising technique for manufacturing turbomachine turbine wheels, as it has a number of advantages that the traditional subtractive technologies lack.

The advantages of SLM relevant to manufacturing turbine wheels include [3]:

• The ability to produce external surfaces of parts with highly complex geometry;

• The ability to construct internal cavities and passages;

• The ability to construct internal lattice structures;

• No need for extra equipment and cutting tools that are crucial to produce objects with similar structure using the traditional manufacturing techniques;

• Efficient use of the materials (up to 95%);

• No need to create program instructions for CNC machine.

At the same time, when using SLM to manufacture turbomachine turbine wheels, it is important to keep in mind a number of issues that appear in connection with the features of the technology:

• The necessity of using supporting structures;

• High internal residual stress that may lead to warpage and deformation of parts on the build platform;

• The design of internal cavities or structures should provide the ability to remove residual powder materials;

• It often happens that surface roughness of parts produced by SLM is unsatisfactory for the conditions of the working mass flow.

II. Basic SLM Principles

To deliver power to metal powders it is necessary to focus the laser beam in a spot of minimal diameter, which will align the focal plane of the optical system of the SLM machine and the plane of the build platform. When this condition is met, it creates the necessary surface power density to fuse the powder material. We distinguish several core processes that occur on the heated bed when a laser beam scans a layer of powder material deposited on the substrate. Absorption of laser radiation energy by the powder material, which forms a molten pool, is followed by transfer of the quantity of heat into the mass of the substrate, transfer of the quantity of heat into the powder material, as well as a loss of the quantity of heat through radiation and convection. Simultaneously to the mentioned processes the alloyed material on the heated bed experiences mass loss due to evaporation [4]. The laser beam moves according to the target trajectory, which is predetermined by the choice of scanning strategy and the geometry of the object in production.

Overall, the trajectory and scanning speed determine the evolution of the molten pool on the surface of the layer, along with a partial overlapping of adjacent tracks with subsequent partial remelting, as well as fusion the material into the substrate.

The depth of the fusion determines the degree of track adhesion, which in turn affects the strength and porosity properties of the material. The geometric properties of the track are determined by the energy supply [5] per unit of powder materials:

 (1)

Where E is the power density [J/mm3];

P is laser power, W;

V is laser speed, mm/s;

h is layer thickness, mm;

d is the diameter of laser spot, mm.

The level of energy density ranges from the lower bound of the powder material melting completely and upper bound, when the melt starts to flow over the surface of substrate. Keeping that range in mind, the values of the parameters mentioned above are selected by trial and error until the track forms a stable shape, that is to say, it does not flow over the surface of substrate and goes at a fixed height.

The producing of a molten pool that is followed by its rapid cooling causes internal stresses to build up in the material. Internal stresses can occur through a variety of mechanisms that include uneven heating and cooling of the layer of exposed material, shrinkage and structural deformation of the material, and overall are to a different extent characteristic of all the available metal powders.

Generally, internal stresses are the main reason why deformations occur and the parts get ripped from the build platform.

III. Design and Process Preparation of a Part to be Manufactured by Means of SLM

When manufacturing turbomachine wheels it is important to ensure that the geometry of the produced item is fully equivalent to the original model, which can only be achieved if there are no residual internal stresses. It is possible to relieve residual stresses by heating the part in inert atmosphere before removing it from the build platform.

Thus, the design and process task of preparing the model for manufacturing by SLM consists in rational combining of the following factors:

• The design of the product should support uniform distribution of areas with fused surfaces along the height of the model;

• Part orientation should be chosen in such a way as to require minimum support structures;

• The arrangement of supporting structures should draw heat away from the melt zone to the substrate (build platform);

• The model should have escape holes that allow removing powder from the internal cavities of the product during post-production;

• The model should have structural elements that ensure that thin-walled elements remain rigid during manufacturing.

Further, we comment on the adaptation of the axial turbine runner of a small turbojet engine for manufacturing by SLM.

The axial turbine of a small turbojet engine that we analyze (Figure 1) is a blisk consisting of a massive disk and thin-walled blades attached to its outer surface. Such design of a turbine disk is typical for most variations of small turboshaft engines, due to the fact that this component is compact and for economic reasons it is impractical to produce a turbine disk with removable blades. It is traditionally designed to be manufactured by casting [7, 14, 15], however, when the objective is to produce a construction of long and thin sharp-­edged blades, the application of this technology is limited, since it is impossible to adequately fill thin channels with molten metal.

Figure 1. Axial turbine model.

Computational modeling of AM process [6] shows that when the object is being printed the temperature gradient changes in radial direction; the temperature in the center of the disk is at its maximum and cools down as it approaches its periphery.

Such temperature distribution is typical to all disk-shaped bodies manufactured through SLM, but in this case the temperature difference between the center and the periphery is more pronounced owing to the thin-walled elements of the part with extended surface (blades) that operate as heat transfer surfaces. Manufacturing a product like this without adapting it to the specifics of AM would result in accumulating periphery stresses and inevitably leads to deformations and ripping from the support structures on the periphery of the disk (Figure 2). In this example we have a turbine disk made of a heat-resistant nickel alloy, and it is shown in Figure 2, the blades of turbine are deformed, making it impossible to use.

Figure 2. Turbine wheel building defects.

One way to solve the problem of high levels of residual stress is to design a turbine wheel with reduced span of fused areas. In this case, we optimize the traditional wheel design (Figure 3) by adding a hollow section to it (Figure 4), which results in reduction of the cross-­sectional area of the disk (Figure 7). This design solution allows reducing the temperature difference between the center of the wheel and its periphery through reducing specific energy supply.

Figure 3. Traditional wheel design.

Figure 4. Wheel with a hollow section.

However, this design solution requires introducing additional structural elements to the blisk design, namely, adding internal stiffeners. Stiffeners are directed in radial direction from the bushing to the periphery of the wheel to reinforce the structure, and also serve as elements that conduct heat from the upper edge of the wheel to the construction platform.

Figure 5. The distribution of the cross-­sectional area of the turbine wheels by height.

IV. Arrangements for Withdrawing Heat from the Build Platform

The standard support structures are comprised of thin-walled cubic elements with a square side of 1–1.5 mm [16],[17]. During the process of printing an object, these elements are filled with powder material and their overall thermal conductivity is substantially lower than the thermal conductivity of solid supports. Such supports often do not adequately transfer heat from the melt zone to the substrate (build platform).

Figure 6. Standard support structures, ‘block’ type.

Excessive heat in the build platform leads to local overheating, deformation and warpage of the part.

Adequate heat transfer that does not increase the temperature gradient can be achieved by integrating solid supporting structures into the design of the product. Thus, in manufacturing a turbine blisk it is reasonable to combine such supporting structures as seen on Figure 7 with standard supporting structures, such as ‘block’.

Figure 7. Solid support structures.

Figure 8. Using holes in solid supports.

The solid supports that we are talking about are made in the form of thin concentric rings 0.15 to 0.3 mm thin. Increasing ring thickness over that limit is undesirable, as it also increases the complexity of post processing the part. Such placement of rings makes it possible to remove them through benchwork.

The number of such solid supports is not regulated; however, their placement in areas adjacent to the more massive elements should result in a higher quality result.

In order to reduce the consumption of powder and prevent it from sintering during heat treatment, we made tear-drop holes (angle max 60°) on the lateral surface of the rings. Use of this form (Figure 9) does not need support structures [[11]] in this cross-­section and leaves escape holes for metal powders.

Figure 9. Tear-drop hole.

V. Construction Elements that Help to Remove Powder Materials

It is unacceptable to leave unmelted powder in the cavities of the turbine blisk, as products of this class operate as part of a rotor that rotates at high angular speed. Therefore such products are to be balanced, which is impossible if there is any moving mass in the internal cavities of the product. To be able to completely remove the residue of loose powder, it is necessary to consider a number of construction elements that would not affect the strength and operating characteristics of the product.

Figure 10. Holes in the bushing and the top part of the wheel.

The position of the holes in the center and on the periphery of the wheel makes it possible to remove the metal powders completely; if necessary such holes can be plugged to exclude any effect on the aerodynamic properties of the wheel.

VI. The Rigidity of Thin-Walled Elements

The contacting parts of turbomachine wheels must meet criteria of preciseness of manufacture, roughness of the surface and tolerance of shape deviations [[12],[13]].

It often happens that in the process of manufacturing the turbomachine wheels by SLM, developers take into account possible deformations of the main body of the product, but neglect the deformations of thin-walled elements (blades), which can lead to their deformation. In the mentioned example (Figure 2), the accumulation of internal stresses ripped support structures from the build platform, which subsequently deformed the blades. When the work of turbine wheel is affected by centrifugal and thermal load [10], residual internal stress can lead to off-nominal deformations and, ultimately, to an emergency. An increase in the number of support structures is a reasonable step when facing such risks.

However, extra support structures become impractical at the stage of post processing, as they ask for more labor input and ultimately require that the blade surfaces are finished by a CNC machine.

This problem can be solved by adding a thin-walled cylindrical shell on the periphery of the produced turbine wheel (Figure 11).

Figure 11. Turbine wheel with a thin-walled shell.

The shell is 0.25 mm thick, with a gap of 0.15 mm between its inner wall and the blades. The shell connects to the blades with a series of contact elements

Figure 12. The connection of a blade with the shell.

The shell serves as a construction element and is to be removed after heat treatment.

This solution serves to intensify the process of heat transfer from the melt zone and eliminates the possibility of deformation of the blades.

VII. Results

In this work we successfully confirmed the effectiveness of our design solutions by adapting a turbine wheel for SLM manufacturing and producing it. The blisk is made of heat resistant nickel alloy ПР-O8ХН5ЗБМТЮ, equivalent to Inconel 718. The alloy is widely used in aerospace, nuclear and other industries due to its stable mechanic properties at high temperatures [17–20].

Key process parameters:

• SLM setup: Concept Laser M2 Cusing;

• Max laser power, W: 400;

• Layer height, µm: 25;

• Build plate temperature, °C: 200;

• Scanning strategy: skin-core [9]

The key process parameters of the machine correspond to those used to manufacture the part from Figure 2.

Support structures of ‘block’ type, 1.5 × 1.5 mm, allow reducing labor input to remove them and cut down the number of post processing operations.

Figure 13. Turbine wheel on the build plate.

Figure 14. The produced turbine wheel.

This resulted in creation of a turbine wheel with the given properties (Figure 14). After we removed the remaining metal powder and heat treated the wheel, its geometrical properties did not change, no warpage or deformation appeared. The product was ready to be tested as a part of the turbojet.

Design for manufacturing and assembly cycle included following operations:

• Aerodynamic calculation of the turbine performance and receiving the geometric properties of the wheelspace;

• Structural analysis;

• Adapting the model to AM;

• Creating program instructions for AM machine;

• Printing the turbine wheel;

• Heat treatment;

• Removing the print from the plate and removing supports;

• Post-processing the wheel.

VIII. Discussion

The suggested ways of design and process adaptation of turbine wheels of small turbojet engines can be applied to designing a wide range of products with different application and adapting them to AM. The design and process solutions were verified by producing a number of parts with zero defects.

IX. Conclusion

In the course of this research we did the following:

1) Analyzed the basic principles of SLM and identified the main factors that affect the quality of the product;

2) Determined the primary ways in which to prepare a turbine wheel for AM;

3) Analyzed the causes of defects that occur when manufacturing similar turbine wheels;

4) Based on calculated data and data found by experiment, developed a process that reduces the span of fused surfaces, found a way to intensify heat transfer from the melt zone, described the major technological means to remove powder residue from the internal cavities of the wheel. Additionally, we solved the problem of reinforcing the thin-walled elements of the turbine wheel.

The suggested ways of designing and adapting turbine wheels of small turbojet engines were experimentally verified, and the produced parts are currently tested as a part of small turboshaft engine units.

REFERENCES

[1]   Shvab, K. The Fourth Industrial Revolution. Moskow: Eksmo, 2016

[2]   Kamps, T. Model Jet Engines: Traplet Publications Ltd., 2005

[3]   Zlanenko, M.Nagaitsev, M., and Dovbysh, V. Additive Technologies in Mechanical Engineering, 2015

[4]   Thorsten, H. Synchronized Two-­Beam Strategies for Selective Laser Melting, 2018

[5]   Tsvetkova, EAustenitic Steel 03H17N14M3 Type: The Specifics of its Formation and Qualities in SLM, 2017

[6]   Bruyaka, V. Engineering Analysis with ANSYS Workbench: Textbook, 2010

[7]   Kustarev, Y. and Kostyukov, A. Axis Turbines for Civil Turboshafts, 2006

[8]   Kovalenko, Y.Gruzdev, V.and Vesloguzov, Y. Termal Conductivity and the Structure of Porous Fine Powder Metals, Solidified by Compression, 1995

[9]   Stirling, R. Design Rules For DMLS, 2017

[10]   Gahun, G. Liquid-­Propellant Rocket Construction and Design, 1989

[11]   Kokkonen, P.Salonen, L.Virta, J.and Hemming, B. Design Guide for Additive Manufacturing of Metal Components by SLM Process, 2016

[12]   Balyakin, A.Drobyshkina, E.Vdovin, R.and Alekseev, V. Research on the Use of Rapid Prototyping to Produce Turbojet Blades, 2016

[13]   Agapovichev, A.Balyakin, A.and Smelov, V. Developing a Method to Cast Complex Parts for Aerospace Application Using Additive Technologies, 2014

[14]   Smelov, V.Balyakin, A.Agapovichev, A.and Vdovin, R. Developing A Method To Cast Complex Parts For Small Turbojet Engines Using Additive Technologies, 2014

[15]   Vdovin, R. 3D Virtual Modelling and Process Optimisation for a Part ‘Swirler of the Second Duct’ for Turbojet Using Computing Technologies, 2012

[16]      Autonomous Manufacturing Ltd. 3D Printing Support Structures: A Complete Guide. Available at: https://Amfg.Ai/2018/10/17/3d-­Printing-­Support-­Structures-­Guide

[17]   Waterman, P. 3D Printing Metals: Why Support Designs Matter. Available at: https://Www.Digitalengineering247.Com/Article/3d-­Printing-­Metals-­Why-­Support-­Designs-­Matter

[18]   ASM Metals Handbook, vol. 2. Properties and Selection: Nonferrous Alloys and Special-­Purpose Materials. ASM International, 2002

[19]   Rosenberg, V. The Basics of Thermal Resistance Of Metallic Materials, 1973

[20]   Fridlyander, I.Senatorova, O.Osintsev, O., et al. Mechanical Engineering: Encyclopaedia, 2001. P. II-3: Nonferrous Metals and Alloys. Composite Metallic Materials