Additive Manufacturing Applications
for Wind Turbines
NIKLAS WAHLSTRÖM
OSCAR GABRIELSSON
Additive Manufacturing Applications
for Wind Turbines
Niklas Wahlström
Oscar Gabrielsson
Master of Science Thesis MMK 2017:98 MKN 191 KTH Industrial Engineering and Management
Examensarbete MMK 2017:98 MKN 191 Additiva Tillverkningsapplikationer för Vindkraftsturbiner Niklas Wahlström Oscar Gabrielsson Godkänt 2017-06-13 Examinator Ulf Sellgren Handledare Ulf Sellgren Uppdragsgivare Vattenfall AB Kontaktperson Marcus Klingevik
Sammanfattning
Additiv tillverkning, "additive manufacturing" (AM) eller 3D-printing är en automatiserad tillverkningsmetod där komponenten byggs lager för lager från en fördefinierad 3D datormodell. Till skillnad från konventionella tillverkningsmetoder där en stor mängd material ofta bearbetas bort, använder AM nästintill endast det material som komponenten består utav. Förutom materialbesparingar, har metoden ett flertal andra potentiella fördelar. Två av dessa är (1) en stor designfrihet vilket möjliggör produktion av komplexa geometrier och (2) en möjlighet till en förenklad logistikkedja eftersom komponenter kan tillverkas vid behov istället för att lagerföras. Detta examensarbete har utförts på Vattenfall Vindkraft och har till syfte att undersöka om det är möjligt att tillverka och/eller reparera en eller två reservdelar genom AM och om det i så fall kan införa några praktiska fördelar. En kartläggning av komponenter med hög felfrekvens och/eller som kan vara lämpade för AM har genomförts. Av dessa har en roterande oljekoppling även kallad roterskarv valts ut för vidare analys.
En omfattande bakgrundsstudie har utförts. En nulägesorientering inom området AM för metaller redogörs, här redovisas även en generell jämförelse mellan konventionella och additiva tillverkningsmetoder. Vidare behandlas aktuella och framtida användningsområden för AM inom vindkraftsindustrin. I bakgrundsstudien behandlas också arbetssättet ”reverse engineering”, huvudkomponenter i ett vindkraftsverk inklusive roterskarven samt flödesdynamik.
Under arbetets gång har en roterskarv med sämre driftshistorik undersökts. I syfte att finna andra konstruktionslösningar som bidrar till en säkrare drift har en bättre presenterande enhet från en annan tillverkare granskats. Då viss detaljteknisk data och konstruktionsunderlag saknas för de undersökta enheterna har "reverse engineering" tillämpats. Ett koncept har sedan utvecklats för den första enheten där förbättrade konstruktionslösningar har introducerats samtidigt som en rad konstruktionsförändringar har gjorts i syfte att minimera materialåtgången och samtidigt anpassa enheten för AM. Konceptet har sedan evaluerats med hjälp av numeriska beräkningsmetoder. För det givna konceptet har även kostnad och byggtid uppskattats.
Arbetet visar på att det är möjligt att ta fram reservdelar till vindkraftverk med hjälp av AM. Det framtagna konceptet visar på ett flertal förbättringar som inte kan uppnås med konventionella tillverkningsmetoder. Emellertid finns det en rad begränsningar såsom otillräcklig byggvolym, kostnader och tidskrävande ingenjörsmässigt arbete och efterbehandlingsmetoder. Dessa förbehåll i kombination med avsaknad av 3D-modeller begränsar möjligheterna att nyttja tekniken i dagsläget. Framtiden ser dock ljus ut, om tekniken fortsätter att utvecklas samtidigt som underleverantörer är villiga att nyttja denna teknik kan AM få ett stort genombrott i vindkraftsindustrin.
Master of Science Thesis MMK 2017:98 MKN 191 Additive Manufacturing Applications
for Wind Turbines
Niklas Wahlström Oscar Gabrielsson Approved 2017-06-13 Examiner Ulf Sellgren Supervisor Ulf Sellgren Commissioner Vattenfall AB Contact person Marcus Klingevik
Abstract
Additive manufacturing (AM), also known as 3D-printing is an automated manufacturing process in which the component is built layer upon layer from a predefined 3D computer model. In contrast to conventional manufacturing processes where a vast volume of material is wasted due to machining, AM only uses the material that the component consists of. In addition to material savings, the method has a number of potential benefits. Two of these are (1) a large design freedom which enables the production of complex geometries and (2) a reduced compexity in supply chain as parts can be printed on-demand rather than be kept in stock.
This master thesis has been performed at Vattenfall Wind Power and aims to investigate the feasibility to reproduce and/or to refurbish one or two spare parts on a wind turbine by AM and if it can introduce any practical benefits. Components with a high failure rate and/or with an suitible design for AM have been investigated. A rotating union or fluid rotary joint (FRJ) was selected for further analysis.
A comprehensive background study has been conducted. A current status of metal AM is described as well as a comparison between conventional and additive processes. Furthermore, current and future applications for AM witihin the wind turbine industry are presented. The mehodology "reverse engineering", main components in a wind power plant including the fluid rotary joint as well as fluid dynamics are also treated in the background study.
As a part of the process, a fluid rotary joint with worse historical failure data was disassembled and examined. In order to find other design solutions that contributes to a better and more reliable operation, another better performing fluid roraty joint was investigated. Since detail drawings and material information are missing for the examined units, reverse engineering has been carried out to gather details of the designs. A concept for the first unit has been developed, in which improved design solutions has been introduced and a number of changes have been implemented in order to minimize material consumption and to adapt the design for AM. The concept has been evaluated by the use of numerical methods. Costs and build time have also been estimated for the developed concept.
This project has illustated that it is feasable to manufacture spare parts by the use of AM. The developed concept demonstrates several improvements that are not possible to achieve with conventional manufacturing processes. Nevertheless, a number of limitations such as insufficient build volume, costs as well as time cosuming engineering effort and post-proccessing methods are present for AM. These restrictions, in combination with lack of 3D-models, limits the possibility to make use of the technology. However, the future looks bright, if the technology continues to develop and if subcontractors are willing to adapt to AM it will probably have a major breakthrough within the windpower industry.
FOREWORD
This master thesis was performed at Vattenfall Wind Power during the spring semester of 2017 and concludes the studies for a master's degree in machine design at the Royal Institute of Technology (KTH) in Stockholm.First, we would like to thank our supervisor Marcus Klingevik who announced this thesis and introduced us to Vattenfall, he has supported us along the way and provided us with valuable input and contacts. We would also like to thank all the employees at Vattenfall that have been involved in this project and supported us with expertise or assistance in one way or another. We are grateful to our supervisor Ulf Sellgren at KTH for guidance and contribution into the project process as well as to Jan Stamer who has provided us with essential expertise in manufacturing processes.
Finally, our sincere thanks goes to Torbjörn Holmstedt at Lasertech LSH AB for the help with the manufacturing of a mockup for our developed concept.
NOMENCLATURE
Here are the Notations and Abbreviations that are used in this Master thesis.Notations
Symbol
Description
τ Shear Stress [N/m2]
u Fluid velocity [m/s]
y Displacement perpendicular to the velocity vector [m]
µ Dynamic viscosity [Ns/m2]
ν Kinematic viscosity [m2/s]
ρ Density [kg/m3]
Re Reynolds number
V Mean fluid velocity [m/s]
λ Fluid friction coefficient ΔPF Frictional pressure loss [Pa]
ΔPM Minor pressure loss [Pa]
ΔPtot Total pressure loss [Pa]
ξ Local loss coefficient
Q Flow rate [m3/s] hr Hours $ U.S dollar € Euro
Abbreviations
AM Additive ManufacturingCAD Computer Aided Design
CFD Computational Fluid Dynamics
CNC Computer Numerical Control
CSS Cross-Section Scanning
DED Direct Energy Deposition
DfAM Design for Additive Manufacturing DMLS Direct Metal Laser Sintering
EBM Electron Beam Melting
FRJ Fluid Rotary Joint
HPDC High Pressure Die Casting
LM Laser Melting
LS Laser Sintering
MS Maraging Steel
OEM Original Equipment Manufacturer
PBF Powder Bed Fusion
RE Reverse Engineering
RP Rapid Prototyping
SLM Selective Laser Melting
SLS Selective Laser Sintering
WPP Wind Power Plant
TABLE OF CONTENTS
1.
INTRODUCTION ... 1
1.1 BACKGROUND ... 1 1.2 PURPOSE... 1 1.3 DELIMITATIONS ... 2 1.4 METHOD ... 22.
FRAME OF REFERENCE ... 5
2.1 ADDITIVE MANUFACTURING ... 52.1.1 Process Steps in Additive Manufacturing ... 6
2.1.2 Additive Manufacturing Process Categories ... 7
2.1.3 Powder Bed Fusion ... 7
2.1.4 Directed Energy Deposition ... 13
2.2 ANALYSIS OF THE POTENTIALS WITH ADDITIVE MANUFACTURING ... 16
2.2.1 Material Capabilities and Geometric Limitations ... 16
2.2.2 Material Characteristics and Mechanical Properties ... 17
2.2.3 Surface Roughness and Hardness ... 20
2.2.4 Conventional Manufacturing versus Additive Manufacturing ... 20
2.2.5 Additive Manufacturing Applications within the Wind Power Industry ... 25
2.3 REVERSE ENGINEERING ... 29
2.3.1 Definition ... 29
2.3.2 Geometric Measurement ... 29
2.3.3 Material Identification ... 29
2.4 WIND POWER PLANT COMPONENTS ... 29
2.4.1 Major Components ... 30
2.4.2 Rotating Union ... 31
2.5 FLUID DYNAMICS ... 34
2.5.1 Basic Theories ... 34
2.5.2 Hydraulic Losses in Pipes ... 36
3.
PROCESS ... 39
3.1 SELECTION OF ADDITIVE MANUFACTURING APPLICATION ... 39
3.1.1 Qualitative Study ... 39
3.1.2 Quantitative Study ... 39
3.1.3 Weighted Evaluation Matrix ... 41
3.2 REVERSE ENGINEERING ... 43
3.2.2 Creation of CAD-models ... 43
3.2.3 Surface Roughness ... 46
3.2.4 Material Estimation ... 48
3.2.5 Original Manufacturing Method ... 48
3.3 SELECTION OF AMTECHNIQUE AND MATERIAL ... 49
3.4 DEVELOPMENT OF A CONCEPT ... 50
3.5 CFDANALYSIS AND VALIDATION ... 56
3.5.1 Define Operating Conditions and Fluid Properties ... 58
3.5.2 CFD simulation on the Original Design ... 59
3.5.3 CFD Simulation on the Concept Design ... 61
3.5.4 Analytical Pressure Loss Calculation ... 63
3.6 STATIC STRUCTURAL ANALYSIS ... 64
3.6.1 Stator ... 64
3.6.2 Rotor ... 66
3.7 COST AND BUILD TIME EVALUATION ... 68
3.7.1 Production Cost Estimation ... 68
3.7.2 Build Time Estimation ... 69
4.
RESULTS ... 73
4.1 DESIGN IMPROVEMENTS ... 73
4.2 BUILD TIME AND PRODUCTION COST ... 75
4.3 MANUFACTURING OF A MOCKUP ... 76
5.
DISCUSSION AND CONCLUSION ... 77
5.1 DISCUSSION ... 77
5.1.1 Stress Analysis and Optimized Material Layout ... 77
5.1.2 Surface Roughness ... 77
5.1.3 Build Time and Cost ... 77
5.1.4 Material selection ... 78
5.2 CONCLUSIONS ... 78
5.2.1 Advantages and Disadvantages for the Developed Concept ... 78
6.
RECOMMENDATION AND FUTURE WORK ... 81
7.
REFERENCES ... 83
APPENDIX A: QUESTIONNAIRE ... 87
APPENDIX B: SURFACE ROUGHNESS ... 89
1. INTRODUCTION
This chapter describes the background, purpose, delimitations and the methods used in the presented project.1.1 Background
Additive manufacturing (AM), commonly known as 3D printing is a group of manufacturing processes used to add material layer upon layer in order to build components. Depending on the process and settings is it possible to process a wide range of materials including metals. AM, as a novel technique, has several potential benefits. Two of these are (1) a greater design freedom compared to conventional manufacturing which enables the production of complex geometries and (2) a reduction in supply chain as parts can be printed on-demand rather than be kept in stock (Ford & Mélanie, 2016).
Vattenfall is one of Europe’s largest generators of electricity and owned by the Swedish government. Today, the company has almost twenty thousand employees in several countries across Europe and operates within various energy sources (Vattenfall, 2016). The wind department, in which this master thesis is performed, is responsible for the operation and maintenance for over one thousand installed turbines and more are about to be built (Vattenfall, 2017).
Vattenfall does not develop turbines. Instead, these are provided from external subcontractors. Usually there is a two year service contract agreement between Vattenfall and the subcontractor for the wind turbine. Vattenfall is less dependent of the original supplier once the service contract and the warranties have expired. At this stage there are opportunities to find other suppliers or to perform in-house repairs. Are there any suppliers, who can offer innovative and cost effective solutions through AM, it would be of Vattenfall’s interest. Further, there are no formal restrictions for Vattenfall to manufacture spare parts.
1.2 Purpose
Vattenfall experiences a lack of knowledge about AM and how this technology can impact on the wind power industry in the future. To bridge this gap, the task of this project is to find suitable applications where AM can be used and provide an insight of how these manufacturing processes can affect the spare part supply chain.
In order to fulfill the purpose, the assignments of this project are defined as:
• Investigate the feasibility to reproduce and/or to refurbish one or two spare parts on a wind turbine by AM.
• Find a suitable spare part(s) for AM and discover if AM can introduce any practical benefits.
1.3 Delimitations
This project is a pilot study of the introduction of AM at Vattenfall and will not include implementation of the technique into the existing business. The study of different AM techniques and materials will be limited to these that are currently commercially available and will only treat relevant metal based processes. Primarily, one component will be investigated during this project. However, this number can be extended to two components depending on its complexity.
1.4 Method
A background study will take place during the first phase of the project. The purpose of this study is to gather information about different AM processes, design of wind power plants and their related maintenance as well as the spare part supply chain. The information will be collected through published work, related internet sources, technical literature and descriptions. Qualitative surveys will be conducted in order to find specific refurbishment and/or design cases that can be relevant for the objectives of this project.
In the next phase of the project some specific refurbishment and/or design cases will be evaluated, using the survey and background study as a foundation. The identified cases will be examined in terms of how AM might be useful or not. During this phase, the qualitative data obtained from the surveys will be compared to quantitative data taken from Vattenfall's enterprise software. The quantitative data can be reports such as the amount of failures for a specific spare part or stockholding of spare parts.
Usually, Vattenfall does not have access to detail drawings/3D-models and material specifications belonging to components installed in their wind turbines. Thus, a need of reverse engineering will be required to bridge the gap of technical data. Geometric data will be collected using either 3D capture technique and/or handheld measurement tools. Surface roughness will be measured by using appropriate surface roughness measuring equipment.
Computer-aided design (CAD) software, Solid Edge ST8, will be used to create 3D-models of the investigated parts as well as for new concepts. ANSYS Workbench 17.1 will be used as Computer-aided engineering platform for design analysis and validation. The module ANSYS Fluent will be used for computer fluid dynamics (CFD) analysis. ANSYS Static Structural will be used to analyse the mechanical stresses and to ensure sufficient dimensioning. Additionally, for design and visual purposes a mockup will be manufactured using an appropriate additive manufacturing process.
Figure 1: Schematic flowchart of the project methodology.
2. FRAME OF REFERENCE
The reference frame is a summary of the existing knowledge and former performed research on the subject. This chapter presents the theoretical reference frame that is necessary for the performed project.2.1 Additive Manufacturing
Additive manufacturing (AM) is the formalized term for what used to be called rapid prototyping (RP) and what is commonly called 3D Printing. 3D Printing however is particularly associated with machines that are low end in price and/or overall capability (ASTM F2732, 2013). In a product development context, RP describes the process for rapidly creating a prototype or basis model directly from digital model data (Gibson, Rosen, & Stucker, 2015). Eventually a final product will be derived from the prototype. More recently, technologies previously used for RP have been developed to the level that the output is suitable for end use. This is one reason why the terminology has essentially evolved from rapid prototyping in the early days to additive manufacturing today.
Wohlers Associates, is a 30-years old company that provides technical, market, and strategic advice on the new developments and trends in AM. Every year they publish a report which conveys the state of the AM industry. This year’s annual report reveals that the entire 3D printing industry grew by 17.4 % in 2016 and is now an industry worth over $6 billion, consisting of all 3D printing products and services worldwide. By 2017, Wohlers forecasts the sale of 3D printing products and services to reach nearly $8.8 billion. By 2021, the market is expected to grow to more than $26 billion. In 2016, 97 manufacturers produced and sold industrial AM systems. This is up from 62 companies in year 2015 and 49 in year 2014 (Wohlers Associates, 2017). Sales of metal printers have accelerated in the past few years. Wohlers states that worldwide, an estimated 808 metal AM machines were sold in 2015, representing a growth of 46.9 % over 2014. Sales of metal AM systems from year 2000 to 2015 are seen in Figure 2 (Wohlers, 2016).
Figure 2: Sales of metal AM systems.
drilling, grinding, carving, etc.) from a bulk solid to leave a desired shape, as opposed to additive manufacturing.” These two manufacturing methods are not necessarily separated from each other. A hybrid system can add material at one point but also utilize the benefits of the subtractive processes at later stages.
2.1.1 Process Steps in Additive Manufacturing
Additive manufacturing is not a “push button” technology. It often requires pre- and post build effort, particularly for parts produced in metal. AM is not inherently superior to subtractive or formative methods of manufacturing, although it depends on the types of parts being built and the design optimization that goes into them. The process of AM has been summarized as a procedure of eight steps (Gibson, Rosen, & Stucker, 2015). The steps involved in the AM process are dependent on the products complexity and engineering content. The process sequence will also vary depending on the used AM technology. To some degree, most AM processes involves the following steps:
1. Conceptualization and CAD: All AM processes begins with a digital 3D solid or surface
representation. This 3D model can be created in a CAD software. It is also possible to create a 3D surface representation from an existing part using a 3D scanner, a process often referred to as reverse engineering.
2. Conversion to STL: Next step is to convert the file into a file format accepted by the AM
machine. STL file format is the most widely used file format nowadays.
3. Transfer to AM Machine and STL File Manipulation: The STL file is transferred to the
AM machine. Additionally, it may be required to manipulate the file to achieve the correct size, position and orientation for building.
4. Machine Setup: The proper build parameters are set up for the build process. Such
settings can include thermal energy, layer thickness, speed, resolution etc.
5. Build: The building step is mainly an autonomous process. Superficial monitoring can be
necessary to ensure no errors have taken place.
6. Removal and Cleanup: The part is removed once the AM machine has completed the
build.
7. Post-processing: Post-processing can be required in some situations. The extent of this
step is very application specific. This step can include polishing, removal of support features, hardening processes, surface preparation, painting etc. In some situations, post-processing may involve chemical or thermal treatment of the part to achieve the final part properties.
8. Application: When required additional treatment has been executed, the part is ready to be
2.1.2 Additive Manufacturing Process Categories
There is a wide variety of different AM processes available. The range of possibilities can be overwhelming for newcomers to the field. In addition, AM system manufacturers have created unique process names to differentiate themselves from their competitors which has contributed to the confusion. In January 2012, ASTM International Committee on Additive Manufacturing Technologies voted on a list of AM process category terms and definitions (Wohlers, Industry Briefing, 2013). The ASTM-approved AM process terms are listed below with the precise standardized definition (ASTM F2732, 2013).
1. Binder Jetting: An additive manufacturing process in which a liquid bonding agent is
selectively deposited to join powder materials.
2. Directed Energy Deposition: An additive manufacturing process in which focused
thermal energy is used to fuse materials by melting as they are being deposited.
3. Material Extrusion: An additive manufacturing process in which material is selectively
dispensed through a nozzle or orifice.
4. Material Jetting: An additive manufacturing process in which droplets of build material
are selectively deposited.
5. Powder Bed Fusion: An additive manufacturing process in which thermal energy
selectively fuses regions of a powder bed.
6. Sheet Lamination: An additive manufacturing process in which sheets of material are
bonded to form an object.
7. Vat Photo Polymerization: An additive manufacturing process in which liquid
photopolymer in a vat is selectively cured by light-activated polymerization.
All these process categories will not be studied during this work. Direct Energy Deposition (DED) and Powder Bed Fusion (PBF) are the AM processes considered as most relevant in this thesis. These are the two most relevant technologies in AM for metal-based production. PBF is characterized as the most convenient manufacturing process for prototyping and direct part production, because of that, the focus will be positioned on it in this thesis. Below is a description of PBF followed by DED.
2.1.3 Powder Bed Fusion
Figure 3: Schematic setup of a PBF system.
PBF processes are widely used with a broad range of available materials including polymers, metals, ceramics and composites with material properties often comparable to many other engineering-grade materials. In principle, all materials that can be melted and re-solidified can be used for PBF processes and any metal that can be welded is considered to be a good candidate. A wide range of metals are commercially available for PBF processing and the number continues to grow as more materials are being developed. Material which are commercially available in some form are stainless and tool steels, titanium alloys, nickel-based alloys, some aluminum and cobalt-chrome alloys. Some companies can even offer PBF of precious metals, such as silver and gold (Gibson, Rosen, & Stucker, 2015).
Benefits and Drawbacks with PBF
PBF processes are most competitive for geometrically complex low-to-medium sized parts. The technology is especially used today for aerospace and biomedical applications in which geometric complexity and lightweight structures are highly sought. Emerging technologies which will further decrease the cost and built time will make PBF processing even more competitive. It is likely that PBF processes will remain one of the most common types of AM technologies for the foreseeable future. Some of the key process benefits are the wide variety of materials. In the PBF build process, loose powder works to some extent as a support structure which enables advanced geometries to be built (Gibson, Rosen, & Stucker, 2015). There are some design guidelines for metal PBF processing to stick to when designing for AM (DfAM). One essential guideline when DfAM is to consider where support structures are needed. As a rule of thumb, overhanging structures greater than 35° between the part and the build platform can be built without requiring a support structure, see Figure 4.
Figure 4: Minimum overhang angle between platform and part to avoid support structure.
PBF will also involve some drawbacks. For example, when processing metals, parts will experience high residual stresses. Because of this, support structures are required to keep the part from excessive warping. Expensive and time consuming post-processing may therefore become unavoidable. Part orientation and locations of support structures are key factors when setting up a build. Accuracy and surface finish are not superior for powder-based AM processes. These properties are strongly related to operating conditions and the powder particle size. Finer particle size will produce smoother parts with better accuracy but are more difficult to handle and spread. Larger particle sizes allow easier powder processing and delivery but with inferior surface finish. Parts can exhibit shape distortion due to thermal deformation. Materials with low thermal conductivity leads to better accuracy as melt pools and solidification are more controllable and part growth is minimized when heat conduction is kept low (Gibson, Rosen, & Stucker, 2015).
PBF Processes
Similar to other AM processes, PBF based technologies have a wide range of different terms when referring to different processes. The classification of these technologies is mainly based on what type of thermal sources they use and the binding mechanism used to fuse the powder particles. Powder particles can be bounded by four primarily binding mechanisms, these are called; solid state sintering, chemically induced binding, liquid phase sintering (partial melting) and full melting. Chemically induced binding is used in the AM process category called binder jetting and is thus not used in PBF processes. Sintering means that the powder creates metallurgical bindings without reaching melting temperatures (Gibson, Rosen, & Stucker, 2015). PBF processes which induce partial melting from a laser beam are often referred to as laser sintering (LS) processes while processes which induce full melting are referred to as laser melting (LM) processes. Additionally, AM system manufacturers have created unique process names over the years although they are based on common principles. Some of the most popular are called Selective Laser Sintering (SLS) or Direct Metal Laser Sintering (DMLS) from EOS, Selective Laser Melting (SLM) from Renishaw and SLM solution and Laser cusing (a combination of the words concept and fusion) from Concept Laser (Valmik, Prakash, & Shreyans, 2014). Electron beam melting (EBM) is easier to distinguish from the other PBF processes since this process uses an electron beam as thermal source. Figure 5 shows how all these processes are linked together based on the principles of how they operate.
Figure 5: PBF technologies and terms.
techniques can involve several binding mechanisms. The standard says that the word “sintering” is a historical term and a misnomer (ASTM F2732, 2013).
Laser Sintering
Selective Laser Sintering (SLS) was the first commercially developed LS process commercialized by DTM Corporation in 1992. In the SLS process, a laser is used to fuse the powder by partial melting the powder in to a body in an enclosed chamber. However, this machine was used for sintering polymer materials, whereas the term SLS is typically used when referring to polymer PBF processes nowadays. The first commercial metal sintering machine was introduced in 1995 by the German company EOS. To distinguish this technology from the previous one, they use the term Direct Metal Laser Sintering (DMLS) (Valmik, Prakash, & Shreyans, 2014). In the beginning, the DMLS process was based on a liquid phase sintering mechanism involving partial melting of the metallic powder (Dongdong, 2015). Since sintering typically requires more time consuming post-processing compared to fusion by melting, few AM processes use sintering as a primary fusion mechanism today (Gibson, Rosen, & Stucker, 2015). Even though the process term DMLS is used by the company EOS today, the process induces full laser melting. Therefore, the word “sintering” is a historical term and a misnomer.
Laser Melting
The demand to produce fully dense components with mechanical properties comparable to those of bulk materials has driven the PBF technologies towards LM processes. The desire to avoid time-consuming post-processing cycles involved in LS has also encouraged the development of LM. It is based on the same processing apparatus and procedures as LS. The main difference is that LM induces full melting as its primarily binding mechanism. Continuously improved laser processing conditions (e.g. higher laser power, smaller focused spot size, smaller layer thickness, etc.) in recent years has made full melting of metal powder achievable. The development of LM processes has significantly improved microstructural and mechanical properties as relative to those of early time LS-processed parts. Parts approaching 99.9 % density can now be produced in a direct way, without post infiltration, sintering or hot isostatic pressing (HIP). Nevertheless, LM is not superior to LS in all aspects. LM requires a higher-energy level and thinner powder layer thickness which results in slower build rates. LM-processed parts will accumulate considerable stresses due to more thermal-shrinkage (Dongdong, 2015). Due to full melting, support structures can also be inherently required to avoid collapse of molten material for overhanging structural members (Valmik, Prakash, & Shreyans, 2014).
SLM is an advanced form of the SLS process where, full melting of the powder bed particles takes place by using one or more lasers to fuse each layer of the powder bed in to a complete part in an enclosed chamber. The SLM process was commercialized in 2004 by a company which today is known as SLM Solutions. Laser cusing is another process name but similar to SLM processing. The difference is that laser cusing utilize a different thermal exposure strategy to ensure thermal equilibrium on each powder bed layer to minimize the residual stresses. This technology was commercialized by the German company Concept Laser in 2004. Most of the PBF machines use one fiber laser of 200 W to 1 kW capacity to selectively fuse the powder. The build chamber is usually provided with inert atmosphere of argon gas for reactive materials and nitrogen gas for non-reactive materials (Valmik, Prakash, & Shreyans, 2014).
reciprocally, thus guaranteeing constant production with minimal downtimes (Concept Laser, 2017). Concept Laser has a large portfolio of metal powders ready for processing such as Stainless steel (316L), Aluminum alloys (e.g. AlSi12 and AlSi10Mg), Titanium alloy (TiAl6V4), pure Titanium, Stainless hot-work steel, Nickel-based alloys, Cobalt-chromium alloy and precious metal alloys (e.g. Silver alloy, yellow/rose/red gold and platinum alloy) (Concept Laser, 2017).
The company SLM Solutions has three LM systems in their product portfolio and they have trademarked their laser based PBF process as Selective Laser Melting (SLM). Their largest system is SLM 500 offering a build envelope measuring 500x280x360 mm. The unit can be equipped with up to four 700 W fiber lasers operating simultaneously to increase the build-up rate up to 105 cm3/h (SLM Solutions GmbH, 2017). They have a variety of metal powders ready for processing such as Aluminum-, Cobalt-, Nickel-, and Titanium alloys as well as tool steels and stainless steels (SLM Solutions GmbH, 2017).
The company EOS has six different DMLS systems in their product portfolio. The machines which offers largest built envelope are called M400 and M400-4. The model M400-4 offers a building volume of 400x400x400 mm. It fuses the powder with four 400 W fiber lasers operating independently in a 250x250 mm square each which enables a built rate up to 100 cm3/hr. The
system is capable to process metals such as light metals, stainless and tool steels to super alloys. EOS M400 offers the same build volume but instead of having four 400 W fibers lasers this system has one single 1 kW powerful fiber laser (EOS, 2017).
Electron Beam Melting
Figure 6: Schematic view and process steps in the EBM process.
Currently Arcam has three EBM machines in their product portfolio commercialized for different AM applications. These machines are Q10plus - for orthopedic implant manufacturing, Q20plus - for production of aerospace components and A2X - for aerospace production and research and development of materials. The Arcam A2X system is designed for production of functional parts within the aerospace, as well as general industry. This system is suitable for demanding applications where parts that must meet the highest material standards. A2X offers a maximum built envelope of 200x200x380 mm. The machines strongest feature is the built chamber that can withstand process temperatures up to 1100° C. High process temperatures offers a wider range of materials available for processing, such as titanium aluminide and Inconel but it is also ideal for development of new materials. The Arcam Q20plus system offers a built envelope of 350x380 mm (diameter x height) and is specifically designed for cost-effective production of aerospace components. This system allows easy powder handling for fast turn-around times and offers a large built envelope (Arcam, 2017).
Arcam has their own portfolio of standard materials. They provide metal powders, process settings and support for the following standard materials: Titanium alloys (Ti6Al4V and Ti6Al4V-ELI), Titanium Grade 2, Cobalt-Chrome and Nickel alloy (Inconel 718). In addition to these standard materials, Arcam allows and supports their customers to independently develop a process for other materials (Arcam, 2017).
Comparison between Laser and Electron Beam PBF
Table 1: Characteristic features of LM and EBM.
LM EBM
Power source One or more fiber lasers of
200 to 1000 W
High power Electron beam of 3000 W
Build chamber environment Argon or Nitrogen Vacuum / Helium bleed
Method of powder preheating Platform heating Preheat scanning
Powder preheating
temperature (°C) 100-200 700-900
Maximum available build volume (mm)
800x400x500 (x,y,z)
(Concept Laser, 2017)
350x380 (ØxH) (Arcam, 2017)
Maximum build rate (cm3/hr) 120 (Concept Laser, 2017) 80 (Arcam, 2017) Layer thickness (μm) 20-100 50-200
Surface finish (Ra) 4-11 25-35
Geometric tolerance (mm) ± 0,05-0,1 ± 0,2
Minimum feature size (μm) 4-200 100
EBM has superior powder preheating capability in comparison to the laser based systems. Powder preheating up to 200 °C is achieved by platform heating for laser based systems while in EBM, a build chamber temperature up to 900 °C is maintained by scanning each layer with the electron beam. Higher build chamber temperatures will reduce the thermal gradient and thus the residual stresses in the parts which further eliminate required heat treatments. Preheating holds loose powder together which enables EBM to eliminate structural supports and allows manufacturing of more complex geometries. In addition, entire EBM process takes place under vacuum which reduces thermal convection, thermal gradients and contamination and oxidation of parts. Laser based manufacturing take place in a nitrogen or argon gas environment. Despite this, laser based systems are still more popular due to lower machine costs, higher accuracy, more available materials and availability of large build up volumes (Valmik, Prakash, & Shreyans, 2014). Figure 7 shows the market share amongst the various PBF AM systems from 2015. The companies EOS and Concept Laser represented more than half of the market share, making them the industry leaders. Arcam is the only company selling EBM systems and represented 10,5 % of the market share of PBF systems (Seifi, Salem, & Beuth, 2016).
Figure 7: Market share among metal PBF AM systems, (Except Trumpf which uses DED method).
2.1.4 Directed Energy Deposition
Directed energy deposition (DED) is an additive manufacturing category in which a various number of materials including polymers, ceramics and metal compositions can be processed. However, the approach is most commonly used for metal powder and is occasionally termed to as “Metal Deposition Technology” (Gibson, Rosen, & Stucker, 2015).
AM process can be traced to the technique used when welding in which a wire is melted by an arc (Sames, List, Pannala, Dehoff, & Babu, 2016). See Figure 8 for a schematic illustration of a typical DED system (Sciaky, 2017).
Figure 8: Schematic illustration of a wire-fed DED process.
The melt pool is generated by focused energy that can either be a laser beam, electron beam or an arc of plasma. However, the most common commercialized process uses laser in combination with a powder deposition system (Gibson, Rosen, & Stucker, 2015). See Figure 9 for an overview of the different techniques.
Figure 9: Overview of different DED technologies.
Laser-based Powder Deposition
The process for a laser-based powder system can be described as the following. The metal powder is fed by pressurized gas into a deposition head which normally consists of sensors, laser optics and one or more powder nozzle(s). When the deposit powder gets in contact with the heat source it will liquefy and thus create a melt pool. Initially, the process requires some sort of substrate on which it can act. After a certain amount of powder has been deposit will the substrate and deposition head move relatively towards each other. This will generate a pre-defined path upon the substrate in which the material will solidify as the deposition head with its laser beam moves on. The process continues to generate tracks of solidified metal powder which eventually will form a layer. The process is repeated until several layers has formed the desired part (Thompson, Bian, Shamsaei, & Yadollahi, 2015).
enclosed chamber in which the building process takes place or by being deposit out of the deposition head and thus act as a shield gas (Sames, List, Pannala, Dehoff, & Babu, 2016). Excess of powder is to some extent unavoidable since a portion of the powder will not get in contact with the melt pool and thus not melt during the process. However, this is not necessary a bad trait since the powder flow can be tuned in by several parameters and therefore improve the process. Furthermore, the excess powder can be collected and reused (Thompson, Bian, Shamsaei, & Yadollahi, 2015).
Several companies develop and market this type of DED process. However, the first company that commercialized this particular technique is the American company Optomec. The brand name of their technique is termed “LENS” and is referred to as “Laser Engineered Net Shaping”. Optomec markets their technology in both standalone machines and as modular print engines that can be integrated into existing CNC machines which allows for hybrid manufacturing systems (Optomec, 2017). Besides hybrid manufacturing of parts, Optomec profiles their technology as especially relevant for component repair where material can be added onto existing parts that have been worn out. Today, the company has three standalone machines for metal additive manufacturing: the LENS 450, LENS MR-7 and their largest machine LENS 850-R in which parts up to 900x1500x1000 mm can be built. Several metal alloys can be processed including titanium, nickel-base super alloys, stainless steels and tool steels (Optomec, 2017) Another company that is specialized within AM and repairs is the American company DM3D Technology who market their technology as "Direct Metal Deposition" (DMD). DM3D has three standalone machines and one deposition system that is integrated with a 6-axis industrial robot. The company claims that their machines can be used for restoration of precision components such as shafts, turbine blades and diaphragms (DM3D, 2017).
Insstek is a Korean company that currently has four machines which all utilize the laser-based powder deposition system. The company also markets the technique as "Direct Metal Technology" (DMT) and their largest machine (MX-GRANDE) is capable of building and repairing parts in sizes up to 4000x1000x1000 mm (InssTek, 2017). Insstek does not have their own material portfolio, instead they refer to industrial supplier of these. However, their machines are capable of process several metal alloys such as steel, copper, titanium, nickel and cobolt (InssTek, 2017).
Electron-based Wire Deposition
The process for an electron-based wire system is to some extent similar to the described laser-based system. Instead of powder and laser, a metal wire is fed into the deposition head and melted utilizing an electron beam. Electron beams function efficiently in vacuum rather than utilizing inert gas, making it more suitable to process materials that have a high reactive rate to oxygen (Gibson, Rosen, & Stucker, 2015).
By utilizing wire as a feedstock, a higher capture capacity rate can be achieved compared to powder. Almost all of the wire will be capture by the melt pool and used in the building process. However, the drawback with such feeding system is the limited complexity that can be fabricated, partly because of the difficulties related to the control of the feeding system. For this reason, wire-fed machines are more suitable for building more simple geometries (Gibson, Rosen, & Stucker, 2015).
deposition process. Sciaky refers to a list of 380 different welding wire products that can be used within their applications (Sciaky, 2017).
Other DED Processes
Even though a few other DED processes exist, these have not yet reached the same level of commercialization. A Norwegian company, Norsk Titanium, has developed a technique named “Rapid Plasma Deposition" (RPD). The technique is similar to the earlier described DED processes but utilize a plasma arc in combination with a titanium wire feedstock to build parts to a near-net shape (Norsk Titanium, 2016).
Benefits and Drawbacks with DED
Unlike PBF processes, DED enables for a larger build volume since there is no need for a pre-laid powder bed. The main limiting build size factor for these machines is either the size of the gas/vacuum filled chamber or the working envelope of the deposition head. The latter case applies to the machines in which the shield gas is delivered out of the deposition head. Another unique benefit for DED processes is its capability of being integrated to existing subtractive methods and its utility within the field of repairs.
DED processes require support structures and/or special arrangement to create complex geometries such as cavities or undercuts since the machine only has the last laid layer as support when producing the next coming layer of material (Gibson, Rosen, & Stucker, 2015).
As for many other AM methods, DED usually requires some sort of post-processing such as removal of substrate/support structures and/or thermal treatment in order to relief stresses. However, the main limitation for DED is its poor resolution and surface finish which means that the parts normally requires further machining in order to meet its requirements (Gibson, Rosen, & Stucker, 2015).
2.2 Analysis of the Potentials with Additive Manufacturing
2.2.1 Material Capabilities and Geometric Limitations
Table 2: Material capability and build volume for selected PBF machines.
Method Company/
Machine
Build Volume
(x,y,z) [mm] Power [W] Materials Source
DMLS EOS/ M 400-4 400x400x400 4 x 400 Aluminium, Cobalt, Nickel, Steel, Titanium (EOS, 2017) EOS/ M280/290 250x250x325 1 x 200/ 400 Aluminium, Cobalt, Nickel, Steel, Titanium Laser cusing Concept Laser/ X Line 2000R 800x400x500 2 x 1000 Aluminium, Nickel, Titanium (Concept Laser, 2017) Concept Laser/ M Line Factory PRD 400x400x425 4 x 1000 Nickel, Steel, Titanium Concept Laser/ M2 Cusing 250x250x280 2 x 400 Aluminium, Cobalt, Nickel, Steel, Titanium EBM Arcam/
Q20Plus Ø350×380 1 x 3000 Cobalt, Titanium (Arcam, 2017) Arcam/ A2X 200x200x380 1 x 3000 Cobalt, Nickel,
Titanium
High strength materials such as Titanium and Nickel-based super alloys (e.g. Inconel) are often available for processing in PBF systems. One reason for this is that there is a greater market demand for these materials. High strength materials are harder to machine by conventional manufacturing which then makes AM a better option regarding economic benefits and time savings. The primary focus has been to develop attractive materials which the market demands. Because AM is used most frequently within the medical and aerospace industry, the vast range of available lightweight metals is no bigger surprise.
2.2.2 Material Characteristics and Mechanical Properties
Table 3: Mechanical properties for selected PBF materials.
Material Grade Mechanical Properties
Method Machine Condition
Ultimate Tensile Strength, Rm [MPa] Yield Strength, Rp0.2 [MPa] Elongation at Break [A%] Source
Direction Direction Direction
Z XY Z XY Z XY
AlSi10Mg – Aluminium alloy
DIN EN 1706 (Die cast) HPDC ≥240 ≥140 ≥1
(EN 1706, 2010) Laser-cusing Concept Laser 240°C/ 7hr 345±11 345±8 214±19 218±7 3±1 6±1 (Concept Laser, 2017)
DMLS EOS As built 395 244 3,2 (EOS, 2017)
300°C/ 2hr 290 165 7,3
Inconel 718 – Nickel alloy
ASTM B637 (Bars, Forgings) Solution + precipitation ≥1275 ≥1034 ≥20 (ASTM B637, 2006)
EBM Arcam As built 1068±139 1060±26 706,5±38 822±25 21-22 22 (Arcam, 2017)
HIP+STA 1232±16 1238±22 1187±27 1154±46 1,1 7 Laser-cusing Concept Laser 980°C/ 1hr 720°C/ 8hr 620°C/ 8hr 1300±50 1050±50 10±2 (Concept Laser, 2017) DMLS EOS As built 1040 710 26 (EOS, 2017) 954°C/ 1hr 718°C/ 8hr 621°C/ 18hr 1470 1200 15
X3NiCoMoTi - Maraging Steel
DIN 1.2709 (Plates, Bars) Annealed ≥712 ≥426 ≥13
(DIN 1.2709,
2017)
DMLS EOS As built 1100±100 1100±100 1000±100 1050±100 10±4 10±4 (EOS, 2017)
490°C/ 6hr 2050±100 1990±100 4±2 Laser-cusing Concept Laser 540°C/ 6hr 1650 1550 2,5±0,5 (Concept Laser, 2017) 316L - Stainless steel
ASTM A276 (Bars, Shapes) Annealed ≥485 ≥170 ≥40
(ASTM A276, 2017) Laser-cusing Concept Laser 550°C/ 9hr 570 470 15 (Concept Laser, 2017)
DMLS EOS As built 540±55 640±50 470±90 530±60 50±20 40±15 (EOS, 2017) Ti6Al4V – Titanium alloy
ASTM F1108 (Castings) cold-worked Annealed/ ≥860 ≥758 ≥8 (ASTM F1108, 2014) EBM Arcam HIP, 920°C/ 2hr/ 1000Bar 1020 950 14 (Arcam, 2017)
DMLS EOS As built 1240±50 1290±50 1120±80 1140±50 10 7 (EOS, 2017)
800°C/ 4hr 1100±40 1100±40 1000±60 1000±50 14,5 13,5
components. If the material is brittle, the component will fail catastrophically and abruptly in case of loads which exceed the materials yield limit. If the material is ductile it will deform, showing evident traces of overloading which is critical for safety. Therefore, post-processing heat treatment is often required to achieve a more ductile material behavior. Besides the mechanical properties presented above, there are a lot of other material properties and factors to consider when choosing the right material for a certain component. Below is a brief description of the PBF-processed metals that have been examined.
Aluminium Alloy (AlSi10Mg)
AlSi10Mg is typically used for casted aluminium parts with thin walls and complex geometry. It characterized by good strength, hardness and dynamic load bearing capacity which makes it suitable for parts subjected by high loads and is ideal for applications which require a combination of good thermal properties and low weight. A big advantage when it comes to AM is the high build rate capability. The build rate can reach up to 100 cm3/hr in a EOS PBF machine. Material density when processed is 2,64 g/cm3. The materials ductility is increased
when heat treated at 300 °C for two hours (EOS, 2017).
Nickel Alloy (Inconel 718)
Nickel alloy (IN718) is an iron-nickel-based hardenable super-alloy characterized by good tensile, fatigue, creep and rupture strength at temperatures up to 700°C. Ideal use is in corrosive environments due to its outstanding corrosion resistance. Due to great heat resistance, it is also ideal for high temperature applications such as gas turbine and process industry parts. The build rate is 15 cm3/hr in a EOS PBF machine. Material density when processed is 8,15 g/cm3 (EOS, 2017).
Maraging Steel (MS 1.2709)
Maraging steel (MS 1.2709) is martensitic-hardenable steel known for combining high strength, toughness and dimensional stability during aging. It is suitable for many tooling applications and various high performance industrial and engineering parts due to high hardenability and wear resistance. Built parts are easily machinable and can be post-hardened to achieve higher surface hardness. This material has excellent surface polish ability. The build rate can reach up to 20 cm3/hr in a EOS PBF machine. Material density when processed is 8 g/cm3 (EOS, 2017).
Stainless Steel (316L)
Stainless steel (316L) is characterized by a high strength, ductility and good corrosion resistance. It is used for many engineering applications such as aerospace, oil & gas, food processing and medical. The greatest disadvantage when it comes to AM is the slow build rate of nearly 7 cm3/s when processed in a EOS PBF machine. Material density when processed is 7,9 g/cm3 (EOS,
2017).
Titanium Alloy (Ti6Al4V Grade 5)
2.2.3 Surface Roughness and Hardness
Surface roughness and hardness are investigated for parts produced by laser based PBF machines. Laser based PBF machines are capable of producing parts with better part resolution and surface finish than EBM machines. The effective melt pool size is larger for EBM creating a larger heat-affected zone. Finer metal powder is often desired to achieve better surface finish. However smaller particles can form dust clouds that interfere with the electron beams efficiency. Consequently, the minimum feature size, median powder particle size, layer thickness, resolution, and surface finish of EBM processed metals are typically larger which degrades the surface finish (Gibson, Rosen, & Stucker, 2015). The achievable surface roughness, hardness and build accuracy for the selected metal alloys processed in EOS laser based PBF machine are shown in Table 4 (EOS, 2017).
Table 4: Surface roughness, hardness and achievable part accuracy for selected PBF metals.
Material Machine Post- Processing Layer Thickness [μm] Surface Roughness Ra [μm] Hardness [HB] Build Rate [cm3/hr] Build Accuracy [μm] AlSi10Mg EOS M 400 As built 90 11 119 100,3 ±100 Inconel 718 EOS M 400 As built 40 6,5 446 15,2 ±40-60 MS 1.2709 EOS M280 1) As built 2) Age- hardening 3) Polishing 20 4 1) < 0,1 3) 322 1) 513 2) 5.8 ±20-50 50 9 1) < 0,1 3) 322 1) 513 2) 19,8 ±20-50 316L EOS M280/290 1) As built 3) Polishing 20 13 1) < 0,2 3) 180 7,2 ±20-50 Ti6Al4V EOS M280/290 As built 60 8 314 32,4 ±50
2.2.4 Conventional Manufacturing versus Additive Manufacturing
Digital Storage and Transportation
production. Inventory occupies physical space, buildings, and land while requiring rent, utility costs, insurance, and taxes. The cost of holding inventory is approximately 25 % of the value of the inventory. Reducing inventory will reduce expenses and frees up capital (Stewart & Gilbert, 2014).
Impact on the Tradition Supply Chain
The impact of AM on supply chains takes many forms, including reduced material waste for leaner manufacturing, simplified production processes, reduced costs, increased flexibility, faster reactions to demand and the ability to decentralize production. There are great potential cost savings in overall management of supply chain including transportation, and inventory carrying. Costs related to the supply chain often account for company's largest expenses. Approximately 5% of the total value of goods consists of costs in the supply chain. AM has several potential ways to reduce complexity in the supply chain. Some of the benefits are (Attaran, 2017):
• Consolidation of assemblies into one part will reduce inventory complexity. This will reduce several labor intensive and time consuming assembly and pre-assembly steps. It will also reduce the supplier base of a company.
• AM enables individualized offers to each customer and the supply chain can quickly react to changes in the marketplace due to increased production flexibility. The need for large bulk inventories will not be required.
• AM will allow smaller manufacturing runs and by that reduce the economic lot size. • It will reduce transportation cost by eliminating the need for both remote high-volume
production facilities and low-level assembly workers. Manufacturing can practically take place anyplace at the same cost.
• It enables more localized production that will reduce logistics costs and environmental impacts.
• Shorter manufacturing lead time enables manufacturing on-demand.
Structural Improvement Potential
In most conventional manufacturing processes, increased geometric complexity increase time and effort to fabricate parts, thus affecting the cost. In AM, the complexity level is not resulting in higher part prices which is often described in terms of providing "complexity for free" (Gibson, Rosen, & Stucker, 2015). A comparison between AM and conventional manufacturing regarding the impact of complexity on costs is seen in Figure 10 (Roland Berger, 2016).
In AM, costs are constant for higher complexity levels (blue line), while in conventional manufacturing, costs rise over increasing complexity levels (black line). Hence, a cost intersection point is described as the starting point for the “complexity for free”.
A useful methodology when designing complex part is so-called topology optimization. Several software’s are available in which the user can set constraints such as loads and supports to a volume of material, in the next stage of the process will the software remove material that is not necessary in order to fulfill the constraints. The outcome or design solution of such process is an optimized, most likely, organic-shape part that has just enough material to handle the initially set constraints (ANSYS, 2017). The methodology is not new. However, additive manufacturing enables to make use of the design solution due to its ability to manufacture complex geometries (Zegard & Paulino, 2016).
Improved Material Utilization
Today, many industrial subtractive manufacturing methods consume vast quantities of raw materials that never end up as part of the final product. In contrast, AM starts with nothing and only add material that the final product requires. Reducing material wastage will reduce cost and make manufacturing more sustainable. For example, a typical one ton Rolls-Royce civil aircraft engine is produced from about 6.5 tons of metal where large parts consist of high-priced metals like titanium. Engineers at Rolls-Royce have estimated that material savings in order up to 80 % may be achieved through AM. Another benefit using AM is the potential to create less dense parts that weight less (Barnatt, 2013).
Costs
Current research reveals that AM is cost effective for manufacturing small batches with continued centralized manufacturing. But distributed production may be cost effective with increased automation. The costs in AM are difficult to measure and current studies are limited in their scope. Most studies examine the production of single parts and those which examine production of assemblies do not examine supply chain effects such as decreased inventory, transportation and assembly costs. Studies has shown that material costs constitute a major proportion of the cost of a product produced using AM. However, increasing adoption of AM may lead to a reduction in raw material cost. The cost of AM machines is also rather high but this cost has continually decreased over the recent years. Between 2001 and 2011 the average machine price decreased 51 % after adjusting for inflation (Stewart & Gilbert, 2014). The machine prices for the PBF systems shown in Table 2 are given by the supplier only after request of quotation. Thus, the machine prices are given in Wohlers 2014 report. In year 2014, the biggest PBF system was X Line 1000R with a build envelope measuring 600x400x500 mm and came with a price tag of about €1.5 million. The price for Arcam’s model Q20 were about €800 thousand and EOS model M280 were priced at €415 thousand (Wohlers, 2014). The price differs greatly depending on the size and laser power capacity. Generally, laser-based PBF machines are cheaper than Arcam’s EBM machines.
Figure 11: Illustration of cost and gains in AM.
In AM, there is no need of investment in designing and fabricating the necessary tooling and fixtures before production begins. This can significantly reduce time and cost from the design phase to production for small scale production. In fact, a remarkable cost reduction can be obtained if the component is designed to exploit the potentialities of AM. A good example of this is a cost comparison analysis on a 1:5 scaled landing gear assembly on an Italian P180 II aircraft. The landing gear was originally made of an aluminum alloy (AlSi10Mg) and manufactured by high-pressure die casting (HPDC). The mould cost and processing cost per assembly for HPDC were €21.29 + €21 000/N, where N is the number of parts produced and the cost of the mould is €21 000. The landing gear was redesigned for AM and a EOS M270 PBF machine was chosen to produce the parts in aluminum (AlSi10Mg). The total cost per assembly for DMLS was €526.31, regardless of production volume. However, this applies to the assumption that only one assembly is produced for each batch. How the cost per assembly would be affected if several parts were built simultaneously (with full utilization of the machine's build envelope) has not been examined. For a production volume of less than 42 assemblies, DMLS was more cost effective than the traditional process of HPDC (Atzeni & Salmi, 2012). The cost per assembly as function of production volume is seen in Figure 12 (Stewart & Gilbert, 2014).
Figure 12: Breakeven point for HPDC and DMLS.
tool allows users to find which machine and process settings are the best choice for the user’s point of interests.
Skin-core is a novel process method which enables faster build rates without compromising the surface quality. Skin-core works in such a way that the inner “core” area of parts is built with bigger beam diameters, thicker layers and higher beam power. The outer “skin” area is built by using smaller beam diameters, thinner layers and lower beam power. This method allows faster build rates without compromising on surface quality and accuracy (Kretzschmar, 2015).
The developed simulation tool for parts costs estimations in PBF calculates the net part cost with regard to material cost (with a waste factor of 2 %), expenses on worker (with a total labor cost of 46116 €/year) and expenses on machine (i.e. machine costs, maintenance-, software- and hardware cost). In the cost evaluation, the machine is assumed to be used 6120 hr/year which corresponds to a utilization rate of 70 %. In a published report, part cost and build time for varying input parameters was evaluated on three different components to find the main effects and influences on costs and build time. The three different components implemented in this comparison is shown in Table 5(Kretzschmar, 2015).
Table 5: Components implemented in the cost comparison.
Bearing block (BB) Turbine wheel (TW) Venturi (V)
Dimensions: 127x76x52 mm Dimensions: 54x54x28 mm Dimensions: 9x9x30 mm Volume: 96,6 cm3 Volume: 20,6 cm3 Volume: 0,96 cm3
The resulting parts cost and build time for variable production volume, accuracy levels and machines when processing MS 1.2709 is shown in Table 6. In this cost analysis, the material cost for MS 1.2709 was set to 89 €/kg or 0,69 €/cm3. The three different machines implemented in
this study are EOS M280 (EOS), SLM Solution 500HL (SLM) and Concept laser X Line 1000R (CL).
Table 6: Costs for MS 1.2709 for variable production parameters (Kretzschmar, 2015).
Machine Accuracy Production volume
Part Part Vol. [cm3] Cost per part [€] Cost per cm3 [€/cm3] Time per part [h] EOS high 100 V 0,96 4,6 4,79 0,2 EOS low 1 BB 96,6 320 3,31 12,2 EOS skin-core 15 TW 20,6 74 3,59 3,3 CL high 100 TW 20,6 135 6,55 3,2 CL low 1 V 0,96 113 117,71 2,1 CL skin-core 15 BB 96,6 635 6,57 15,0 SLM high 15 V 0,96 9 9,38 9,0 SLM low 100 BB 96,6 122 1,26 2,1 SLM skin-core 1 TW 20,6 176 8,54 6,3
dramatically when using CL's machine. CL's machine is intended for production of large components and is therefore not the ideal choice for these components and production volume. Cost per cubic centimetre is given for six of nine different combinations regarding production volume and accuracy levels if data from CL is excluded in Table 6. As a rough estimation of cost per cubic centimetre for all combinations, the missing values have been generated by interpolation. The impact of cost on what part (i.e. V, BB or TB) is built has been excluded. The cost per cubic centimetre for MS 1.2709 for variable accuracy levels and production volumes is seen in Figure 13. Note that the production volume effects the cost mainly if it is possible to manufacture several parts simultaneously in one build. So, the cost depends on how effective the build envelope utilization is rather than the production quantity itself (Kretzschmar, 2015).
Figure 13: Cost per cubic centimetre for varying production parameter (MS 1.2709).
Lowest cost per cubic centimetre is 1.26 €/cm3 when accuracy level is low which gives faster
build rates and when the sufficient utilization of the build camber by larger production volume. Highest cost per cubic centimetre is 20.17 €/cm3 when the build rate is slower by higher
accuracy level and when utilization of the build camber is poor due to small production volume. If a single part would be large enough to fully utilize the build chamber alone, the production volume would be less important.
2.2.5 Additive Manufacturing Applications within the Wind Power Industry
AM has not yet reached as large use in the wind industry today as in the aerospace- and medical industry. Since AM has matured and great benefits has attracted an attention from other industries the wind power industry has begun to explore this technology (Dodd, Jan, 2017).
In a Product Development Context
Vestas Wind Systems is a Danish manufacturer, seller, installer and servicer of wind turbines. They are the only global energy company dedicated exclusively to wind energy with more than 71 GW installed to date (Vestas, 2017). Vestas have their own 3D printing facilities but they have so far mainly used the technology for rapid prototyping in product development. It has allowed Vestas to judge value-chain impacts in manufacturing, transport, installation and service while reducing their time to market (Dodd, Jan, 2017).
