PLASTIC ADDITIVE MANUFACTURING AND ITS POTENTIAL …1198354/... · 2018-04-17 · CAD Computer Aided Design CLAD Construction Laser Additive Directe CJP Color Jet ... manufacturing

DEGREE PROJECT IN INDUSTRIAL ENGINEERING AND MANAGEMENT, SECOND CYCLE, 30 CREDITS STOCKHOLM; SWEDEN 2017

PLASTIC ADDITIVE MANUFACTURING AND ITS POTENTIAL EFFECTS ON A SUPPLY CHAIN A case study concerning selection of beneficial parts for additive manufacturing through utilizing a methodological part evaluation framework

VALENTIN KREHL Production Engineering and Management

i. ABSTRACT

Recently additive manufacturing (AM) is rapidly growing and evolving due to the advancements made in speed, quality, resolution and performance. Consequently, AM is starting to become, beyond prototyping, increasingly important for the manufacturing of end-use parts. Several successful case studies are reported and companies are starting to investigate in the opportunities of using AM for production processes and supply chain integrations.

The purpose of this research is to evaluate the potential of using AM for manufacturing plastic end-use parts in a supply chain. Different AM technologies for plastic manufacturing are described and the main advantages and challenges are identified. Based on the company’s plastic part scope, a methodical framework for assessing parts regarding their AM suitability is established, in order to research the potential benefits of an AM implementation. The framework contains a methodical preselection and scoring process utilizing a top-down approach and an analytical hierarchical process (AHP), followed by a technical and economic assessment of the promising parts. In the research, only the existing design is taken into account: the same part, designed for conventional manufacturing technologies, is manufactured by AM without changes in geometry. Both cases, in-house manufacturing and purchasing from a service provider, were investigated with the employment of a cost model for FDM technology and a request for quotation from a general AM service provider. Both cases are compared to each other and the current conventional manufacturing technology.

It was found that, currently, utilizing an AM service provider is more beneficial for the company, due to the low number of parts that could be currently produced with AM. Hereby the lead time for AM profitable parts could be significantly reduced. Using AM, a centralized production with one AM service supplier is clearly seen as the preferable supply chain configuration in the case of the company.

The research provides guidance for the evaluation of part suitability concerning AM production of end-use plastic parts and contributes to the research concerning AM implementation in a supply chain in general and the aim of the company to acquire valuable AM knowhow in particular.

i. SAMMANFATTNING

På sista tiden växer Additive Manufacturing (AM) snabbt och framläggar på grund av de framsteg som gjorts i hastighet, kvalitet, upplösning och performans. Följaktligen börjar AM, förutom prototypning, att blir viktigare för tillverkning av användningsdelar. Flera framgångsrika fallstudier rapporteras och företagen börjar undersöka möjligheterna att använda AM för produktionsprocesser och integrering pa supply chain.

Ändamål med denna forskning är att utvärdera potentialen att använda AM för tillverkning av plastdetaljer för slutanvändning i en supply chain. Olika AM-tekniker för plasttillverkning beskrivs och de viktigaste fördelarna och utmaningarna identifieras. På grundval av företagets plast delar omfång etableras en metodisk ram för bedömning av delar om deras AM-lämplighet för att undersöka de potentiella fördelarna med en AM implementering. Ramverket innehåller en metodisk förhandsval och poängprocess som utnyttjar en top-down-strategi och en analytisk hierarkisk process (AHP), följt av en teknisk och ekonomisk bedömning av de lovande delarna. I undersökningen beaktas endast den ekonomiska synvinkel: samma del, konstruerad för konventionell tillverkningsteknik, tillverkas av AM utan ändringar i geometri. Båda fallen, egen tillverkning och inköp från en tjänsteleverantör, undersöktes med anställning av en kostnadsmodell för FDM-teknik och en offertförfrågan från en allmän AM-tjänsteleverantör. Båda fallen jämförs med varandra och den nuvarande konventionella tillverkningstekniken.

Det konstaterades att det för närvarande är en AM-tjänsteleverantör som är mer fördelaktig för företaget på grund av det låga antal delar som kan för närvarande produceras med AM. Härav kan ledtiden för AM-lönsamma delar minskas betydligt. Med AM används en centraliserad produktion med en AM-tjänsteleverantör tydligt som den föredragna leveranskedjekonfigurationen när det gäller företaget.

Forskningen ger vägledning för utvärdering av delkompatibilitet avseende AM-produktion av plastdetaljer för slutanvändning och bidrar till forskningen om AM-genomförande i en supply chain i allmänhet och syftet med företaget att förvärva värdefullt AM-knowhow i synnerhet.

ACKNOWLEDGEMENTS

The undertaking of this thesis has been a pleasure and I am very grateful that I had the opportunity to research in the field of additive manufacturing at the company.

First and foremost, I would like to express my sincere gratitude to my company’s supervisor for the support and advice during my thesis and my time at the company as well as for the confidence and the independency I got in developing my thesis.

I would also like to thank Mr. Amir Rashid for the academic support during my thesis research.

Furthermore I would like to thank all other people in the company, especially in the sourcing and logistic planning department, who made this thesis possible by providing valuable information through interviews, supporting me with the data gathering or giving feedback that helped to improve my work. This thesis would not have been possible without their collaboration.

Finally I would like to thank my family for always supporting me during my studies, Becky, Tiina and Peter for their time and mental support during the thesis and last but not least Gustav for all the countless lifts.

ii. ABBREVIATIONS

Abbreviation Full term 2D Two Dimensional 3D Three Dimensional 3DP Three Dimensional Printing AHP Analytical Hierarchical Process AM Additive Manufacturing BJ Binder Jetting CAD Computer Aided Design CLAD Construction Laser Additive Directe CJP Color Jet Printing DC Distribution Center DfAM Design for Additive Manufacturing DMD Direct Metal Deposition DMLS Direct Metal Laser Sintering DOD Drop on Demand EBDM Electron Beam Direct Manufacturing EBM Electron Beam Melting ERP Enterprise Resource Planning FDM Fused Deposition Modeling FFF Fused Filament Fabrication FLM Fused Layer Manufacturing GMU Global Manufacturing Unit ITR Inventory Turnover Rate LENS Laser Engineered Net Shaping LOM Laminated Object Manufacturing MML Material Master List MOQ Minimum Order Quantity PDM Product Data Management RFQ Request for Quote RM Rapid Manufacturing RP Rapid Prototyping SL Stereolithography SLA Stereolithography SLM Selective Laser Melting SLS Selective Laser Sintering UAM Ultrasonic Additive Manufacturing

Table of Contents

i. Abstract ..................................................................................................................................... 7

i. Sammanfattning .................................................................................................................... 8

Acknowledgements ......................................................................................................................... 9

ii. Abbreviations ...................................................................................................................... 10

1 Introduction ............................................................................................................................ 7

1.1 Purpose and motivation .......................................................................................... 7

1.2 Delimitations ................................................................................................................ 8

1.3 The company’s plastic part supply chain ........................................................ 9

2 Theoretical framework AM ........................................................................................... 11

2.1 AM technology overview and current state ................................................ 11

2.2 AM processes and their suitability for end-use plastic parts .............. 14

2.3 AM advantages and challenges ......................................................................... 21

3 Possible impact of AM on the company ................................................................... 24

3.1 Production alternative for low volume parts ............................................. 24

3.2 Part consolidation ................................................................................................... 24

3.3 Decentralization ...................................................................................................... 24

4 AM part assessment frameworks ............................................................................... 25

5 Methodology application part assessment for AM .............................................. 27

5.1 Preliminary Assessment ...................................................................................... 27

5.2 Technical Assessment ........................................................................................... 44

5.3 Economical assessment ........................................................................................ 47

6 Effects on the company’s supply chain ..................................................................... 52

6.1 Make scenario using FDM technology ........................................................... 52

6.2 Buy scenario using an AM service supplier ................................................. 52

7 Conclusion and discussion ............................................................................................. 55

8 Reference list ....................................................................................................................... 58

9 Appendix ................................................................................................................................ 64

9.1 Appendix 1 ................................................................................................................. 64

9.2 Appendix 2 ................................................................................................................. 65

9.3 Appendix 3 ................................................................................................................. 67

9.4 Appendix 4 ................................................................................................................. 68

9.5 Appendix 5 ................................................................................................................. 71

9.6 Appendix 6 ................................................................................................................. 72

9.7 Appendix 7 ................................................................................................................. 73

Introduction

7

1 INTRODUCTION

1.1 PURPOSE AND MOTIVATION

According to a recent study (researchandmarkets.com, 2016), the additive manufacturing (AM) market is projected to reach $6.2 billion by 2021, from $2.5 billion in 2015. Besides AM technologies were already emphasized by for industrial application in the NCMS (National Center for Manufacturing Sciences) roadmap study on AM in 1998, it was only in the recent years, that AM is rapidly growing and evolving, due to advancements made in speed, quality, resolution and performance. Consequently, AM is starting to become, beyond protoyping, increasingly important for the manufacturing of end-use parts (Khajavi, Partanen and Holmström, 2014; Lindemann et al., 2015).

This development has led to an increasingly number of successful case studies, particularly in the automotive, medical and aerospace industry (Gibson, 2017), and more and more companies are starting to investigate in the opportunities of using AM for production processes and supply chain integration (Wienken and Kilger, 2016). According to the world’s leading research and advisory company Gartner, the use of AM both for supply chain operations and the production of end-use parts is expected to have a high benefit. Depending on the industry, it will take an estimated 5 to 10 years more for mainstream adoption (Michael Shanler and Pete Basiliere, 2016). Another research of the consulting company E&Y highlights that companies have a shortage of in-house knowledge related to AM and still lack of information resulting in limited awareness of different AM technologies (Wienken and Kilger, 2016).

The company is successfully using AM in their prototyping workshop, but has not yet investigated to use AM beyond. As it is expected that the importance of AM in production and supply chain processes accelerates, the company wants to acquire deeper understanding of AM technologies, its capabilities and opportunities to research a possible potential implementation for end-use parts and the resulting benefits and effects on the supply chain. Based on the purpose, following research question was obtained:

Introduction

8

How and for which parts in the company could the implementation of AM be beneficial and what would be the effects on the supply chain?

From this question following sub-questions were defined:

• What kind of AM technologies are currently available? • What are the advantages and challenges of AM?

• What are the opportunities of applying AM in a supply chain in

general? • What are the relevant opportunities of applying AM in the supply

chain of the company?

• What is the most suitable methodology to be applied in order to obtain the most promising parts in the company concerning the application of AM?

• What is the potential of AM for the considered part scope in the company?

• What are the technical requirements of the selected potential parts with regards of compliance and material?

• What are the cost benefits for specific identified parts if produced by AM?

• What are the knowledge needs and effects on the supply chain?

1.2 DELIMITATIONS

The research is conducted in the sourcing department at the company and it considers the part scope contained within the company’s supply chain in Europe. It is expected to cover the biggest share of the company’s plastic part scope via Europe. Due to the nature of the company’s parts the researched part scope is focused on plastic parts only. The scope of metal parts used in the company are seen as not complex enough in order to gain benefits from metal AM. The research timeframe is limited to 20 weeks. The research is explicitly focused on the existing part design and does not consider any assemblies and the resulting AM opportunity of part integration nor the redesigning of parts.

Current AM technologies and their suitability for plastic end-use parts, and the related advantages and challenges are researched. In order to obtain the most suitable part candidates for the use of AM, a literature study investigating in a suitable part assessment methodology is performed, and the methodology thereafter executed. Potential cost benefits using AM are researched for the identified parts and the effects on the supply chain and the knowledge needs in the company are highlighted.

Introduction

9

1.3 THE COMPANY’S PLASTIC PART SUPPLY CHAIN

With the aim to identify possible relevant opportunities of applying AM in the company’s supply chain, the supply chain of the investigated part scope has to be understood.

Figure 1 shows exemplary the European supply chain of the company for plastic parts (only plastic injection manufactured and plastic machined). Information regarding the company’s plastic part supply chain was gathered by interviewing employees from the departments project sourcing and logistic planning, as well as with the help of enterprise resource planning (ERP) data.

The company does not manufacture plastic machined or injection molded parts itself and hence parts are purchased from suppliers. The company uses one main distribution center (DC) in Germany acting as a hub for its supply chain, as well as a smaller DC in Russia. The European global manufacturing centers (GMUs) are located in Sweden and Poland. Relevant parts in the DC in Germany are classified into two categories. LOCA parts are stored locally in the DC, in contrast to NORM parts, which are stored at the supplier or the GMU. The NORM parts are, with the call-off order, shipped via the DC Germany to its destination point (e.g next stage). The Next stage can be a so called platform, a dealer, service technician or end customer. Usually the parts are consolidated in the DC Germany before they are shipped to the next stage. In very rare cases, the parts are shipped from the GMU Sweden directly to the customer (BANS parts).

The finding is, that by only considering the parts scope of the DC in Germany nearly all injection molding and plastic machined parts in Europe can be taken into account. The DC in Russia receives all parts from DC in Germany and the part scope of GMU Sweden is either stored in DC Germany or routed via the DC. GMU Poland’s use of relevant plastic parts from suppliers is negligible and can therefore be disregarded (see Appendix 1).

Introduction

10

Figure 1 Supply chain Europe for injection molding and plastic machined parts

Theoretical framework AM

11

2 THEORETICAL FRAMEWORK AM

By conducting a literature review, the current state of AM technology, its processes and relevant advantages and challenges with regard to the supply chain are presented. It is important to know when investigating in AM, that the state of the art with its limitation and possibilities is rapidly evolving and therefor a continuous reviewing is necessary.

2.1 AM TECHNOLOGY OVERVIEW AND CURRENT STATE

AM is defined by the F42 Committee (2012) of the standardization organization ASTM International as:

“A process of joining materials to make objects from 3D model data, usually layer upon layer, as opposed to subtractive manufacturing methodologies.”

Khajavi, Partanen and Holmström (2014) also describe AM as a digital technology for manufacturing physical objects layer by layer from a three-dimensional (3D) computer aided design (CAD) file. The production process time is dependent on the objects’ size and required part precision and usually varies from a few hours to a few days.

The generic AM process contains, according to Gibson, Rosen and Stucker (2015), eight steps:

(1) Generation of a 3D CAD model of the to-be-manufactured part including its details and dimensions

(2) Conversion of the 3D CAD model into a so called .STL file format and slicing into two-dimensional (2D) cross sections or layers, using a computer program

(3) Transference of the .STL file to the AM Machine (4) Adjusting machine setup according to the produced object (5) Building of the object (6) Removal of the object (7) Post-processing, for instance removing support features (8) Use in an application

AM Technology was invented in the 1980s, when it was originally introduced as rapid prototyping (RP) or 3D printing (3DP) (Bourell et al., 2009) and initially used as a method to produce rough physical prototypes of the final products (Li et al., 2016). Since then, the technology continued to evolve and major AM techniques have been developed. Nowadays objects from different materials, such as plastic, metal or glass, are able to be produced. The input material hereby can be provided in the form of powders, filaments, liquids or


12

sheets (Gibson, Rosen and Stucker, 2015). To summarize, throughout the years, various abbreviations were used for AM and its technologies. The most commonly used are RP, rapid manufacturing (RM) and 3DP. The abbreviation 3DP is used both interchangeable with the abbreviation AM, as well as an abbreviation for the binder jetting technology.

In July 2016, the company Gartner published an annual article in which the current AM market is analyzed using the Gartner Hype Cycle. As shown in Figure 2, by July 2016, 3D Printing in Supply Chain was seen on the peak of inflated expectations, whereas 3D printing in Manufacturing Operations was already seen as entering the trough of disillusionment. 3D Printing in Supply Chain means to use AM to produce a finished item or subassembly, whereas 3D printing in Manufacturing Operations stands for the use of AM to print tools, fixtures or dies and molds. Michael Shanler and Pete Basiliere (2016) define the peak of the hype curve as:

“Overenthusiasm and unrealistic projections, [where] a flurry of well-publicized activity by technology leaders results in some successes, but more failures, as the technology is pushed to its limits.”

The trough of disillusionment is described as loss of media interest due to overinflated expectations and the becoming unfashionable of the technology.

Figure 2 Gartner hype cycle additive manufacturing 2016

(Michael Shanler and Pete Basiliere, 2016)


13

It can be stated, that due to the made advancements, e.g. precision, speed, affordability and range of materials, the AM technology is currently transforming from the use for prototyping to a future use for production (Wohlers and Caffrez, 2016) and it potentially will fundamentally revolutionize manufacturing operations and supply chains (economist.com, 2012; Ott, 2015).

This conclusion can also be confirmed by a large survey conducted by E&Y in 900 companies in April 2016. The advances in AM already enables the production of high quality end-use components or products, which are predominantly used in plastic and medical applications, the aerospace and the automotive industries. According to E&Y, around 30% of plastic, automotive and medical companies that already utilize AM, use it for printing their own end-use components or products. By 2021, 38% of all the surveyed companies expect that AM will become part of their production processes for end-use parts (Wienken and Kilger, 2016).

Currently, huge investments are made in new AM facilities, which are a sign of change within the manufacturing operations and their supply chain. For example, GE has recently opened a $32 million AM R&D center in Pittsburgh, Pa., and Siemens has announced that they will invest €21.4 million to open a metal AM facility in Sweden (Wohlers and Caffrez, 2016).

To summarize, AM is currently rapidly transforming from a manufacturing technology dominantly used for prototyping to manufacturing technology for end-use parts. High investments and more and more successful case studies indicate the disruptive potential of AM, both for manufacturing and the supply chains.


14

2.2 AM PROCESSES AND THEIR SUITABILITY FOR END-USE PLASTIC PARTS

Various AM processes have been introduced to the commercial market by industrial companies such as Stratasys, EOS, Arcam or 3D Systems. The AM processes can be classified according to the ASTM F42 Committee (2012). In this chapter, a short description and advantages and challenges of each technology classification will be explained briefly. A detailed overview of selected AM technologies can be found in Appendix 2.

2.2.1 Material jetting

Figure 3 Material Jetting (additively.com, 2017)

Commercially used nomenclature: DOD (Drop on Demand), Inkjet printing, Thermojet

Droplets of building material are selectively deposited using an ink-jet printing process similar to 2D printing technologies. Advantages include a lower cost of equipment compared to other AM technologies, the building material can be changed during the process and high building speed and scalability is possible. However, material jetting is limited by its choice of materials as only waxes and photopolymers are commercially available (Gibson, Rosen and Stucker, 2015, pp. 192–216). The technology is commonly used for prototypes and the production of wax models. Wax models are used for investment casting and its lead time can be significantly reduced using material jetting. As photopolymers are aging, the mechanical properties are degrading and thus the technology is not suitable for end-use plastic parts. Compared to other AM technologies only a few providers are available worldwide, which are 3D Systems, Keyence, Stratasys and HP (Basiliere, 2016b).


15

2.2.2 Binder jetting

Figure 4 Binder Jetting (additively.com, 2017)

Commercially used nomenclature: 3DP (3 Dimensional Printing), BJ (Binder Jetting), CJP (Color Jet Printing)

In contrast to material jetting, a liquid binder is printed onto a powder bed which then forms the bulk of the part. The advantages, as with material jetting, are the lower cost of equipment compared to other AM technologies, as well as the high building speed and scalability. Binder jetting was originally developed under the name 3DP at MIT University. A wide range of polymer composites, metals and ceramics can be used. It is possible to manufacture parts in different colors. As particles are glued together, post-processing infiltration steps are needed if good mechanical properties or a high density part is required (Gibson, Rosen and Stucker, 2015, pp. 205–217). Binder jetting is commonly used for producing sand casting applications, and is not very suitable for end-use plastic parts, due to the needed excessive infiltration post processing steps (Basiliere, 2016a).


16

2.2.3 Material extrusion

Figure 5 Material Extrusion (additively.com, 2017)

Commercially used nomenclature: FDM (Fused Deposition Modeling), FFF (Fused Filament Fabrication), FLM (Fused Layer Manufacturing)

Originally developed by the company Stratasys as FDM technology, there are now many variations of the original technology available, due to patent expiry in 2010. Material is selectively extruded through a nozzle, often two nozzles; one for building material and one for support structures. The material extrusion process is relatively inexpensive, but it only processes plastics and needs substantial post-processing, such as sanding and polishing, compared to other AM technologies. Nevertheless, material extrusion is seen as suitable for the manufacturing of end-use plastic parts. The main advantage compared to the powder bed fusion process is the easiness of use, as material can be changed very quickly instead of the laborious powder changing in machine that uses a powder bed fusion process (Basiliere, 2017).


17

2.2.4 Powder bed fusion

Figure 6 Powder bed fusion in form of SLS (additively.com, 2017)

Commercially used nomenclature: SLM (Selective Laser Melting), SLS (Selective Laser Sintering), DMLS (Direct metal laser sintering), EBM (Electron Beam Melting)

Powder bed fusion processes use a thermal energy source to selectively process a powder bed. There are various patented technologies available, using the powder bed fusion process. A wide variety of materials, for example plastic, metal and binder-coated sand, can be processed. When printing plastic material no support structure is usually needed, as the surrounding powder gives enough support. A relatively high level of detail and accuracy can be achieved and only one single material per built is possible (Gibson, Rosen and Stucker, 2015, pp. 107–144; Zalak Shah and Pete Basiliere, 2016b). Based on the available materials and its characteristic, powder bed fusion technology is seen as very suitable for end-use plastic parts.


18

2.2.5 Directed energy deposition

Figure 7 Directed Energy Deposition (Loughborough University, 2017a)

Commercially used nomenclature: CLAD (Construction Laser Additive Directe), DMD (Direct Metal Deposition), LENS (Laser Engineered Net Shaping), EBDM (Electron Beam Direct Manufacturing)

Directed energy deposition processes melt metal material, which are being deposited onto a built platform. Commercial processes utilize material in the form of powders or wires, which is melted via an electron beam. Therefore high density parts with controllable microstructural features are able to manufacture. The process can be utilized for repairing defect parts such as high-technology components. The main drawback of this AM process is the poor resolution and surface finish, as well as the slow build speed (Gibson, Rosen and Stucker, 2015, pp. 245–267). Multiple materials can be deposited simultaneously which enables the production of metal gradients. Another unique advantage of the process comes with the possibility of multi-axis motion (Zalak Shah and Pete Basiliere, 2016a). Due to the use of metal material only, this process is not suitable for the manufacturing of end-use plastic parts.


19

2.2.6 Sheet lamination

Figure 8 Sheet Lamination (Loughborough University, 2017b)

Commercially used nomenclature: UAM (Ultrasonic Additive Manufacturing), LOM (Laminated Object Manufacturing)

Layer-by-layer lamination processes utilize various built materials, including plastics, metals, and ceramics. Lamination is realized with different bonding mechanisms, such as ultrasonic welding, adhesive or thermal bonding (Gibson, Rosen and Stucker, 2015, pp. 219–242). Laminated objects, are not suitable for structural use and mostly used for aesthetics and visual models (Loughborough University, 2017b). Hence, sheet lamination is not suitable for manufacturing of end-use plastic parts.


20

2.2.7 Vat photopolymerization

Figure 9 Vat Photopolimerization (additively.com, 2017)

Commercially used nomenclature: SL (Stereolithography), SLA (Stereolithography)

Ultraviolet light selectively cures a part cross-section using liquid photopolymer that is contained in a vat. Unlike powder based methods, support structures will often need to be added. The main advantage over other AM processes are the part accuracy and surface finish, as well as high building speed. The main drawback is the exclusive use of photopolymers, which do not have the impact strength and durability of injection molded thermoplastics. Additionally photopolymers are aging and mechanical properties are degrading within relatively short durations of exposure to sunlight, therefore it is mainly used for functional prototypes and not suitable for end-use plastic parts (Gibson, Rosen and Stucker, 2015, pp. 63–102).


21

2.3 AM ADVANTAGES AND CHALLENGES

The company’s analyzed part scope can be seen as most similar to the characteristics of an after-sales service supply chain as both represent mainly high-demand variability, low-demand rates, high part variety and remote service locations. Therefore, a literature research was conducted to successfully identify potential advantages of AM in the supply chain, focusing on the after-sales service supply chain.

According to Khajavi, Partanen and Holmström (2014), the characteristics of AM makes it a potentially disruptive technology for supply chain management. Especially the after-sales service part characteristics appear to be tailor made for the use of AM technology (Cohen, Agrawal and Agrawal, 2006).

All findings concerning advantages and challenges are described in the following paragraphs and if not stated differently, literature references are summarized in Appendix 3. Appendix 3 also presents an overview of the main findings and the most relevant literature.

2.3.1 Advantages of AM Reduction of manufacturing or order costs for small batch sizes While conventional manufacturing processes, such as injection molding, are very time and cost effective for big batch sizes, AM is better suited for small batch sizes. Setup costs with AM are comparably lower as no start-up costs for tooling occur. Therefore costs are nearly independent of the size of the batch and small batches of parts can be produced cost effectively without having to achieve economies of scale (see Figure 10).

Figure 10 Cost comparison (Cotteleer and Joyce, 2017)

Mass customization AM is very suitable for mass customization as every product can be manufactured slightly different with no significant increase in manufacturing costs due to the needlessness of tooling. For example the health industry is currently heavily transforming due to the AM enabled possible mass customizations of products, like hearing aids (Knofius, Van der Heijden and Zijm, 2016).


22

Complex design Almost any shape without typical manufacturing restrictions, that are due to the casting or subtractive manufacturing process, are achievable. Furthermore, there is no direct connection between complexity and manufacturing costs (Hopkinson, Hague and Dickens, 2006), which means the complexity of the part does not affect manufacturing cost. AM is commonly phrased as an manufacturing technique where “complexity is free” (cf. Cotteleer, 2014). As a result, the product can be optimized for its function through for example topology optimization (see Figure 11) or featuring cellular mesostructures.

Figure 11 Design optimized bracket (3dprintingbusiness.directory, 2016)

Part consolidation As it is possible to manufacture more complex parts with AM, former assemblies can be consolidated into a single part or reduced to a smaller number of parts, which leads to a reduction in production and assembly costs. Perhaps the most mentioned example in recent times is the functional integration of 20 components for a fuel nozzle injector into a one piece part at the company GE (geadditive.com, 2016).

Shorter lead time No tooling is needed for AM and therefore the lead time from idea to product is significantly shorter. AM can also be beneficial for the production of molds for conventional manufacturing technologies, where the lead time can be significantly reduced and complexity, such as cooling channels, can be added (Leandri, 2015). If AM is used in an after-sales service supply chain, shorter lead times are achievable as AM technology can be implemented closer to the point of consumption, inducing an overall higher supply chain responsiveness.

Decentralization If AM is applied in a suitable case, literature states that AM technology can help to decentralize production. Decentralized production can bring various benefits such as achieving higher potential for customer satisfaction, higher flexibility or decreased need for 3PLs (third-party logistic providers).


23

Inventory holding cost reduction AM can be used for on-demand production and therefore achieves safety stock reduction, which leads to a reduction of inventory holding costs in the entire supply chain.

Sustainability In comparison to a subtractive process, AM produces nearly no waste and is therefore considered a green manufacturing technology. Hence, the carbon emissions of a supply chain can be significantly reduced with AM technology. Furthermore, worn out parts can be repaired, rather than replaced. An example is the company Siemens AG (2016), which is utilizing AM to repair gas turbines in Sweden.

2.3.2 Current challenges of AM System price Even though prices have fallen the last years due to patent expiries, the price for an AM system that can offer a reliable production system for end-use parts is currently still very high. According to Wienken and Kilger (2016), the system price is the most commonly cited barrier for adopting AM technology in a company.

Building speed With a building speed of up to 10,5l/h (SLS, plastic), the cycle times are still significantly larger in comparison to injection molding manufacturing techniques and are therefore inferior in achieving economies of scale via batch fabrication of standardized part geometries. Nevertheless, major progressions in reducing the building speed where recently achieved by HP. In May 2016, the company introduced an AM printer based on the powder bed technology, which they claim is up to ten times faster compared to existing FDM or SLS printer solutions (HP Development Company, 2017).

Material Although different ranges of materials are available that can be used for AM, it does not cover the wide variety of materials that can be utilized with conventional manufacturing methods. Due to deficiencies of interlayer bonding, current AM parts suffer from anisotropic mechanical properties. Furthermore, most of the commercially available AM technologies can only process a single material at a time (Ahn et al., 2002). Extensive research is performed to create more materials, however, in some cases the requirements for end-use products cannot yet be achieved with AM when it comes to surface quality or accuracy and post-processing is required (Mellor, 2014).

Possible impact of AM on the company

24

3 POSSIBLE IMPACT OF AM ON THE COMPANY

Resulting from the company’s business model and its current situation, and based upon the previously researched advantages of AM, following can be concluded on possible impacts on the company:

3.1 PRODUCTION ALTERNATIVE FOR LOW VOLUME PARTS

The company's supply chain sources, manufactures and distributes a high mix of low volume parts. AM can be a production alternative, as manufacturing costs are nearly independent from the manufactured quantity and no tooling costs occur, which leads to a manufacturing cost reduction for low volume parts. Furthermore, it is stated in the company that extensive transport and holding costs occur more likely for low volume parts, and these costs could sometimes exceed the manufacturing costs. Using AM could decrease transport and holding costs, due to possible on-demand manufacturing and the manufacturing directly at the needed location.

3.2 PART CONSOLIDATION

As more complex designs are able to manufacture without increase in manufacturing costs, AM could be used for part consolidation through redesign. This would minimize assembly costs as well as sourcing actions within the company. This exceeds the scope of this project but should be considered in the future if the company wants to adopt AM technology.

3.3 DECENTRALIZATION

With the continuous development of the AM market, it is expected that AM system prices drop in the long term future and therefore decentralization of manufacturing closer to the customer could be considered as a business model. By using AM directly on the site of the after-sales service provider of the company, customer needs could be fulfilled more precisely. Transportation and holding costs could be decreased in addition to the lead time.

AM part assessment frameworks

25

4 AM PART ASSESSMENT FRAMEWORKS

At the present time, limited scientific literature is available and only a few methodologies are developed to assess and identify feasible part candidates for AM production from a given part scope. Usually, a practitioner realizes that AM technology might improve the characteristics, such as cost or design for functionality, of a specific part. This then leads to a part assessment regarding the technological feasibility and the benefits if the part is printed. Such an approach is called bottom-up approach and relevant frameworks are, among others, developed by Simkin and Wang (2014) and Lindemann et al. (2015).

Simkin and Wang's approach consists of a two-stage method that examines the parts to determine whether it falls into one of seven proposed categories of supply chain scenarios, such as improved functionality or lower import/export costs. Simkin and Wang's approach is used commercially by the company Senvol LCC. If a part falls into one of the proposed categories, according to Simkin and Wang, the parts may be more cost effective to produce via AM. In the second stage it is examined which AM technology can be used for manufacturing the identified part candidates, based on a non-public algorithm.

Lindemann et al. approach includes a workshop concept for the identification of suitable part candidates. After part candidates are identified in the first workshop, a second workshop is used to evaluate identified parts by AM experts and company representatives. The evaluation is based on a scoring methodology which assesses different part characteristics, concerning primarily technological aspects. In the last step, economic aspects are taken into consideration as well.

Knofius, Van der Heijden and Zijm (2016) proposed a framework, in contrast to the above presented methodologies, using a top-down approach, which can be initiated with a large scope of parts. The advantages of the framework are, according to the authors, minimization of risk to disregard promising parts as well as decreasing dependency on the expertise of practitioners, which subsequently minimizes the chance of underestimating logistical improvements. The framework is based on the analytical hierarchy process (AHP) and relies on suitable part information that is retrievable from the company's database. The result is a ranking of parts according to their potential value when produced with AM, which then leads to a manual evaluation of the most promising parts.

Another assessment framework is described by Zanardini et al. (2015), providing a three-steps evaluation guideline, verifying technical and economic feasibility of the implementation of AM for specific parts. The framework is

AM part assessment frameworks

26

based on Simkin and Wang (2014) including additional quantitative drivers. According to Zanardini et al. (2015), quantitative drivers can be used to assess which parts are more promising for AM, nevertheless it is not stated what methodology is used to assess the parts, which is named preliminary assessment. The proposed evaluation framework contains furthermore a technical assessment, followed up by an economical assessment.

Besides the presented scientific literature, several AM printer companies are offering their consultant service to assist with the identification of potential AM suitable parts. Unfortunately it was not possible to obtain more information about their methodology approach, as this is their key business knowledge.

Methodology application part assessment for AM

27

5 METHODOLOGY APPLICATION PART ASSESSMENT FOR AM

The proposed AM evaluation framework of Zanardini et al. (2015) is utilized as a skeleton of the part assessment methodology. Based on the findings, the proposed assessment framework by Knofius, Van der Heijden and Zijm (2016) is seen as the most suitable approach for the company and utilized for the preliminary assessment. The methodology is continued with a technical assessment and followed up with an economical assessment. Figure 12 presents the structure of the methodology including the employed research.

Figure 12 Adapted methodology structure for AM part assessment

5.1 PRELIMINARY ASSESSMENT

In the following, Knofius, Van der Heijden and Zijm (2016) framework, adapted to the company's characteristic, is used for the preliminary assessment in order to identify suitable part candidates on the basis of the company’s empirical data.

The bottom-up approach of Lindemann et al. is not suitable for the company’s environment and the nature of the case study for two reasons. Firstly, the bottom-up approach relies, according to Knofius, Van der Heijden and Zijm, on the expertise of practitioners, which the author has limited access to in the sourcing department of the company. Secondly for the identification process, a bottom-up approach can only take a limited number of parts into consideration due to its design and the parts must therefore be preselected. In this project no selections have been made previously and the entire plastic parts scope, which contains a high number of parts, must be considered in order to avoid the overlooking of promising parts. This is conducted in three steps:


28

Step 1: Scoping and data-cleaning: The scope of the considered parts is defined and the provided part data is filtered and cleaned, based on Go and No-Go criteria. This provides only potentially suitable part candidates, which are the base for Step 2.

Step 2: Part scoring: All qualified parts from Step 1 are scored based on available part attributes (criteria) that are indicators for AM suitability. As a result, Step 2 provides a ranking of the most suitable parts for AM based on the particular criterion.

Step 3: Criteria Weighting: The criteria are related to the company’s AM supply chain goals and weighted according to their importance regarding the goals. For the evaluation of the supply chain goals and the determination of the company specific weighting of each criteria, an AHP is used. As a result, an overall ranking according to the AM part suitability can be computed.

5.1.1 Scoping and data-cleaning (Step 1) As mentioned in chapter “Delimitations”, only plastic parts are considered, as part requirements for beneficial utilization of metal AM are absent in the company. Metal AM is more advantageous if parts are highly complex and consists of advanced materials, such as titanium alloys. Therefore it is mostly used in aviation and aerospace, medical and automotive industries, where it can help to reduce weight through topology optimization or used to manufacture complex parts with integrated cooling channels made from advanced materials (Guo and Leu, 2013).

Since the company focuses on only specific key manufacturing technologies, plastic manufacturing is outsourced to external suppliers. Within the field of sourcing, the company uses so-called “Material Master Lists (MML)”, which are created for every DC and GMU on a monthly basis. The MMLs contain necessary information about the parts, with respect to sourcing and logistics. The majority of MMLs are generated via the same ERP system and therefore their provided information is consistent. Furthermore, information about material properties and assembly structures are available through a product data management (PDM) system.

Through the evaluation of multiple MMLs and discussions with employees from both sourcing and logistic planning, the European market and hence the MML of the DC in Germany was chosen as the focus of this research. The majority of the company’s plastic suppliers are based in Europe and furthermore all plastic suppliers for the important European market are situated on the continent with the majority routed via the DC in Germany. An overview over the plastic supply chain flow within Europe can be seen in chapter 1.3 and Appendix 1.


29

In order to clean the data, the MML of the DC in Germany, which contained 32735 parts, was cleaned and filtered using Go and No-Go criteria. The Go and No-Go criteria were defined by reviewing literature and AM printer datasheets and discussions with employees within sourcing and logistic planning at the company. As a result, a remaining number of 464 parts were left to consider in the proceedings. Figure 13 provides an overview and explanation of the data cleaning phase.

It is crucial to note, that due to the constant advancement of AM and the evolving of parts in their life cycle stages (see Figure 24), parts that were disregarded at the time, could become interesting in the future. For example, at the time high demand parts could transform to low demand parts when reaching a certain life cycle stage and therefore regular analysis is recommended.


30

Figure 13 Data cleaning phase

(2) Parts are only considered if they are made from metal or plastic through constraining the part attribute “Category”

(3) It was decided together with the responsible sourcing managers that only parts that are manufactured via injection molding or plastic machining are potentially interesting (e.g. all parts with the value Plastic Injection Moulding and Plastic Machined Parts in part attribute “Subcategory”). Other manufacturing techniques such as rotational moulding, blow moulding, vacuum forming and extrusion are disregarded. In general, parts with a too low demand for injection moulding are plastic machined, hence part costs are comparably high. Injection moulding parts can be generally considered as candidates for AM.

(4) All parts that have a deletion date of 2017 or earlier are disregarded, as their remaining usage time is too short

(5) The part attribute “Part Volume” is manually created and generated using part attribute “Length”, “Width” and “Height”. According to the maximum building envelope of commercially available plastic AM printers, the part volume for considered parts is restricted.

(7) The manually created part attribute “Expected Demand” is used as a Go/No-Go criterion. “Expected Demand” uses figures from the part attribute “APO-Saleforecast_12M”, which is a forecast created by the advanced planning and optimization (APO) application of the used ERP system. If no forecast is available, the sum of consumption and invoice quantity of the last year is used (part attribute “Consumption/Invqt 2016”). If no consumption or invoice quantity for the last year is available, a generated “Average Demand”, containing the years 2012 to 2016, is used as “Expected Demand”. Since the break-even point of AM compared to conventional manufacturing technologies depends on the specific part, a high threshold of 1000 was chosen in order to lower the amount of data, but to make sure that any potentially suitable parts were not excluded from the beginning.

(8) If more stock is available than is needed for the next 5 years, assuming that the “Expected Demand” does not change throughout the years, the part will not be considered.

Steps Go /No-go Criteria Constraint Explanation Remaining parts

1 MML DC Germany start point with all parts that are listed in the MML for the DC in Germany

32735

2 Category Metal_plastics all parts that are made from metal or plastic 6660

3 Subcategory Plastic Injection Moulding, Plastic Machined Parts

manufacturing process 1000

4 Deletion Date blank or ≥ 2017 if deletion date was before or in 2017, parts are not considered

930

5 Part Volume < 226800 cm3 if bigger, no SLS AM machine available 929

6 Consumption or Invqt

no consumption or invqt since 2012

if no consumption or sales since 2012, demand too low to consider parts

889

7 Expected Demand <1000

Expected Demand:IF APO_Saleforecast_12M <>0 THEN

APO_Saleforecast_12MELSEIF Consumption/Invqt 2016 <>0 then

Consumption/Invqt 2016ELSE

Average demand 2012-2016END IF

602

8Stock / Expected

Demand <5if there is more stock than 5 times the expected demand for the next 12 months, parts are not

considered464


31

5.1.2 Part scoring (Step 2) All remaining parts are used in a scoring process. Part attributes, which are available from the ERP and PDM systems, are examined together with the functions manufacturing, sourcing and logistic planning to determine whether their characteristics could indicate part suitability regarding AM. Six different part attributes, in the following named criteria, are considered to be suitable for the following scoring process (see Figure 14). The criteria are either calculated with the help of other available part attributes, or simply adopted.

Figure 14 Criteria used for the scoring process

As the remaining 464 parts that are left after Step 1 have different values for the different criteria, the criteria values are normalized by applying a Min-Max scaling procedure. Hereby, the criteria values are scaled to a fixed range from “0” to “1” with scores according to their AM suitability. Subsequently, criteria values get comparable and assessable. The score “0” represents the least favorable outcome (e.g. the worst value), and “1” the most favorable outcome (e.g. the best value). Values in-between receive a proportional linear or logarithmic score. The following equations were used for the assignment of scores:

𝑧𝑧𝑖𝑖 =𝑥𝑥𝑖𝑖 − 𝑥𝑥𝑚𝑚𝑖𝑖𝑚𝑚

𝑥𝑥𝑚𝑚𝑚𝑚𝑚𝑚 − 𝑥𝑥𝑚𝑚𝑖𝑖𝑚𝑚 𝑧𝑧𝑖𝑖 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑧𝑧𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑣𝑣𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛

𝑥𝑥𝑖𝑖 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑜𝑜𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑣𝑣𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛 (1)

𝑦𝑦𝑖𝑖 =log 𝑥𝑥𝑖𝑖 − log 𝑥𝑥𝑚𝑚𝑖𝑖𝑚𝑚

log𝑥𝑥𝑚𝑚𝑚𝑚𝑚𝑚 − log 𝑥𝑥𝑚𝑚𝑖𝑖𝑚𝑚 𝑦𝑦𝑖𝑖 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑧𝑧𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛𝑜𝑜𝑛𝑛𝑛𝑛𝑛𝑛𝑙𝑙ℎ𝑛𝑛𝑛𝑛𝑚𝑚 𝑣𝑣𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛

𝑥𝑥𝑖𝑖 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑜𝑜𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑣𝑣𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛 (2)

The assignment of scores is a crucial step as it heavily influences the outcomes of the part assessment methodology. Therefore, resulting score distributions for the different criteria have been analyzed and discussed together with employees. If required, the scaling procedure was adjusted both using thresholds and/or logarithmic score distribution instead of a linear one. Whether to normalize the values in a linear or logarithmic way is fully

Criteria Explanation Category

Expected Demand Expected Demand for the next 12 months Adopted

Lead TimeLead Time worst case: leadtime_contracted

or leadtime_operational Adopted

Lead Time Difference Leadtime_contracted - leadtime_operational Generated

ITR Inventory over Turnover rate Adopted

Standard Price Standard Price Adopted

MOQ/Demand Minimum Order Quantity/Expected Demand Generated


32

dependent on the resulting distribution of the criteria values and therefore dependent on the company’s analyzed part scope.

In the following section both a detailed justification of the chosen and considered criteria, as well as an analysis of the outcome, are provided.

Expected Demand The previously generated criterion Expected Demand is taken into consideration as a very relevant criterion, as literature (e.g. Holmström et al., 2010 or Mellor, 2014) indicates that slow movers are the most interesting parts for AM and illative low demand parts receive a higher score.

For computing the scores of the criterion Expected Demand, a linear Min-Max scaling was applied, using equation (1). In addition, the logarithmic Min-Max scaling has been considered, but it will, as seen in Figure 15, only give a small number of very low demand parts a high score and is therefore not suitable.

As the majority of the analyzed parts have an Expected Demand below 250 pcs per year, over 60% of the parts receive a score between 0,8 and 1 (see Figure 15). It was discussed that the criterion range is set too wide, causing the majority of parts to receive a high score. Nevertheless, it can be argued that it is valid to consider such a wide Expected Demand range as the feasible yearly production volume for AM is very depending on the part volume. A greater score separation will be achieved through other criteria ranking.

Figure 15 Comparison linear and logarithmic score distribution of Expected Demand

0

0,2

0,4

0,6

0,8

1

1,2

1 57 142 246 382 601 965

Scor

e

Expected Demand

Linear Min-Max Scaling

Logarithmic Min-Max Scaling


33

Figure 16 Score distribution of Expected Demand

Minimum-Order-Quantity over Demand The MOQ over Demand is a relevant criterion, as it shows potential cost saving possibilities in case of a low demand part with a comparably high MOQ. If the MOQ is higher than the Expected Demand, comparably high holding costs occur or the actual purchase price increases for ordering less than the agreed MOQ. As there is no need for setup or tooling change, cost effective MOQs of 1 can be achieved using AM. Hence, low values of MOQ over Demand receive a lower score than high values.

A logarithmic Min-Max scaling is applied to spread the scores more evenly (see Figure 17). Since MOQ over Demand values smaller than 1 do not show any saving possibilities, a lower threshold of 1 is used. Additionally, due to a few very high values, an upper threshold of 5 is applied, and the scaling is computed as following:

50% of the parts MOQ value is either of smaller or equal size compared to the Expected Demand. No cost savings can be made and hence these parts receive a score of 0. On the other hand, 50% of the parts MOQ value is higher than the Expected Demand value and therefore these parts are more interesting for AM, which is why they receive a higher score (see Figure 18)

0

200

400

600

800

1000

1200

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

# Parts 7 12 12 15 28 20 41 44 62 102 121

Percentage 2% 3% 3% 3% 6% 4% 9% 9% 13% 22% 26%

Max Value 1000 917 845 744 648 545 450 350 250 150 50

0

20

40

60

80

100

120

140

Max

Exp

. Dem

and

per S

core

Part

s pe

r Sco

re

𝑦𝑦𝑖𝑖 =log 𝑥𝑥𝑖𝑖 − log 1log 5 − log 1

log 1 = 𝑛𝑛𝑛𝑛𝑙𝑙𝑛𝑛𝑛𝑛 𝑙𝑙ℎ𝑛𝑛𝑛𝑛𝑟𝑟ℎ𝑛𝑛𝑛𝑛𝑛𝑛log 5 = 𝑣𝑣𝑢𝑢𝑢𝑢𝑛𝑛𝑛𝑛 𝑙𝑙ℎ𝑛𝑛𝑛𝑛𝑟𝑟ℎ𝑛𝑛𝑛𝑛𝑛𝑛 (3)


34

Figure 17 Comparison linear and logarithmic score distribution of MOQ over Demand

Figure 18 Score distribution of MOQ over Demand

Lead Time The biggest lead time improvements when utilizing AM can be made if the current lead time is comparably long. Hence, parts with longer lead times receive a higher score compared to parts with shorter lead times.

At the company, two different lead times are documented. Firstly, the contracted lead time, which is defined as the agreed period between purchase order and receiving the purchase at the DC or GMU. Secondly, the operational lead time, which is the real measured lead time.

For the scoring, the worst-case scenario is considered, which means the longest lead time, either contracted or operational, is taken into consideration.

0,0

0,2

0,4

0,6

0,8

1,0

1,2

Scor

e

MOQ over Demand

Linear Min-Max Scaling

Logarithmic Min-Max Scaling

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

4,0

4,5

5,0

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

# Parts 234 25 25 21 24 21 18 18 11 13 54

Percentage 50% 5% 5% 5% 5% 5% 4% 4% 2% 3% 12%

Min Value 0,0 1,1 1,3 1,5 1,8 2,1 2,4 2,9 3,4 4,0 4,8

0

50

100

150

200

250

Aver

age

of M

OQ

/ De

man

d

Par

ts p

er S

core


35

Values were normalized using linear Min-Max Scaling. To level out high peaks a threshold on the top 1% of the values was applied (<65 days). Thus, following equation was used:

𝑧𝑧𝑖𝑖=𝑥𝑥𝑖𝑖 − 𝑥𝑥𝑝𝑝𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑚𝑚𝑒𝑒𝑖𝑖𝑒𝑒𝑒𝑒(0,01)

𝑥𝑥𝑚𝑚𝑚𝑚𝑚𝑚 − 𝑥𝑥𝑝𝑝𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑚𝑚𝑒𝑒𝑖𝑖𝑒𝑒𝑒𝑒(0,01) (4)

Only 14% of the parts receive a score of 0,5 or higher as their lead time is 34 days or more. As seen in Figure 19, the majority of the parts (61%) receive a score between 0,3 and 0,5 which is related to a lead time from 22 to 33 days.

Figure 19 Score Distribution of Lead Time

Standard Price The criteria Standard Price is regarded to be a good indicator concerning AM suitability. The Standard Price includes the purchase price and the costs that occur until the delivery point (e.g. for transportation and custom duties). The MML provided Standard Prices always in Euro, comparing to the purchase price which was given in the local currency. The Standard Prices can be used for comparison as the company only adds a certain percentage to every purchase price, instead of using more advanced models. AM is more advantageous for higher Standard Prices and therefore higher Standard Prices receive higher scores. To receive an evenly spread distribution the logarithmic Min-Max scaling is preferred (see Figure 20). For values lower than 1€ a threshold is applied, as well as for the top 5% values (<27€).

𝑦𝑦𝑖𝑖 =log 𝑥𝑥𝑖𝑖 − log 1

log 𝑥𝑥𝑝𝑝𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑚𝑚𝑒𝑒𝑖𝑖𝑒𝑒𝑒𝑒(0,95) − log 1 (5)

0

10

20

30

40

50

60

70

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

# Parts 3 75 40 158 123 15 8 8 2 26 6

Percentage 1% 16% 9% 34% 27% 3% 2% 2% 0% 6% 1%

Min Value 7 10 16 22 28 34 39 45 52 59 64

0

20

40

60

80

100

120

140

160

180

Aver

age

Lead

TIm

e

Part

s pe

r sco

re


36

As shown in Figure 21, 35% of the plastic parts receive a score of 0 as their Standard Price is under 1€. This indicates a high fraction of simple standard plastic parts, such as washers, plugs or caps.

Figure 20 Comparison logarithmic and linear distribution of Standard Price

Figure 21 Score distribution for Standard Price

0

0,2

0,4

0,6

0,8

1

1,2

Price

Scor

e

Linear Min-Max ScalingLogarithmic Min-Max Scaling

0,0

5,0

10,0

15,0

20,0

25,0

30,0

0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

# Parts 163 38 40 42 36 36 19 22 17 22 29

Percentage 35% 8% 9% 9% 8% 8% 4% 5% 4% 5% 6%

Min Value 0,0 1,2 1,7 2,4 3,4 4,8 7,0 9,5 13,5 19,0 26,7

0

20

40

60

80

100

120

140

160

180

Aver

age

Pric

e

Part

s pe

r Sco

re


37

Lead Time Difference Differences between the agreed lead time and the operational lead time can cause downtimes at the customer site or increase the cost for safety stock. It is argued that, in the right case, AM can minimize lead time differences as the part can be manufactured in-house which simplifies the process control and hence control over the lead time delays. For the scoring, a linear Min-Max scaling is chosen. A lower threshold of 1 is used, as the maximum allowed delay is 1 day. The upper 5% of the values (<7 days) receive the maximum score of 1. The upper threshold is used, to level out the few very high values. The equation used, can be seen below:

𝑧𝑧𝑖𝑖 =𝑥𝑥𝑖𝑖 − 1

𝑥𝑥𝑝𝑝𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑚𝑚𝑒𝑒𝑖𝑖𝑒𝑒𝑒𝑒(0,95) − 1 (6)

The Lead Time Difference for over 75% of the parts is 3 or more days. Only 16% of the parts are delivered within the agreed lead time (see Figure 22).

Figure 22 Score distribution of Lead Time Difference

0

1

2

3

4

5

6

7

8

0,0 0,2 0,3 0,5 0,7 1,0

# Parts 72 43 221 98 3 27

Percentage 16% 9% 48% 21% 1% 6%

Min Values 0 2 3 4 5 7

0

50

100

150

200

250

Aver

age

Lead

Tim

e Di

ffere

nce

Part

s pe

r Sco

re


38

Inventory Turnover Rate (ITR) The ITR is a figure that, in the company, reflects the speed with which the parts flow through the stock. This means the time a part stays in stock on average. In the company the ITR is calculated as follows:

𝐼𝐼𝐼𝐼𝐼𝐼 =𝐶𝐶𝑛𝑛𝑛𝑛𝑟𝑟𝑣𝑣𝑛𝑛𝑛𝑛𝑛𝑛 𝑉𝑉𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛 𝑙𝑙𝑛𝑛 𝐼𝐼𝑛𝑛𝑣𝑣𝑛𝑛𝑛𝑛𝑙𝑙𝑛𝑛𝑛𝑛𝑦𝑦 𝑣𝑣𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑜𝑜 12 𝑛𝑛𝑛𝑛𝑛𝑛𝑙𝑙ℎ𝑟𝑟𝐴𝐴𝑣𝑣𝑛𝑛𝑛𝑛𝑛𝑛𝑜𝑜𝑛𝑛 𝐼𝐼𝑛𝑛𝑣𝑣𝑛𝑛𝑛𝑛𝑙𝑙𝑛𝑛𝑛𝑛𝑦𝑦 𝑙𝑙𝑛𝑛 𝐼𝐼𝑛𝑛𝑣𝑣𝑛𝑛𝑛𝑛𝑙𝑙𝑛𝑛𝑛𝑛𝑦𝑦 𝑉𝑉𝑛𝑛𝑛𝑛𝑣𝑣𝑛𝑛 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑜𝑜 12 𝑛𝑛𝑛𝑛𝑛𝑛𝑙𝑙ℎ𝑟𝑟

(7)

An ITR smaller than 1 indicates that a part stays in stock for one year or more. A linear Min-Max scaling is applied, where values below 1 are scored with the maximum score of 1. In consultation with the function logistic planning, a threshold on the upper 5% (e.g. values higher than 5.2) is used, as a change of stock of 5.2 times in a year is seen as sufficient and no improvements need to be made. For missing ITR values (10% of all parts) the overall mean ITR for the investigated part scope, 1.85, is utilized.

Over 50% of the parts receive a score >0,9 as their ITR is lower than 1,21 (see Figure 23). This means these parts stay in stock on an average of 9,9 months or more.

Figure 23 Score distribution for ITR

0

1

2

3

4

5

6

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Parts per Score 26 10 5 9 11 21 26 28 86 60 182

Percentage 6% 2% 1% 2% 2% 5% 6% 6% 19% 13% 39%

Min of ITR 4,9 4,49 4,14 3,7 3,28 2,88 2,45 2,03 1,62 1,21 0,05

0

20

40

60

80

100

120

140

160

180

200

Min

imum

ITR

Part

s pe

r Sco

re


39

Considered criteria: Part life cycle status It was considered to take the part life cycle status into account for the scoring process. The life cycle status could provide information about demand developments. AM can be beneficial in the situation where a part is previously manufactured using injection molding, but due to the progression in its life cycle (i.e. coming into its phasing out stage), a steep regression in demand can be expected (see Figure 24). Currently, it is not possible to extract this information from the ERP or PDM system. Furthermore there are no established rules or guidelines regarding the provided availability of spare parts for the customer. Therefore the part life cycle status could not be utilized for the scoring process.

Figure 24 Product life cycle stages according to Levitt (1965)

Tooling status Information about the current status of the manufacturing tool, e.g. the injection mold, would reveal if AM could be beneficial. The tools are owned by the company, even if the part manufacturing is outsourced. If a part tool is close to refurbishment or repair and a part is well progressed in its life cycle, i.e. near to its phasing out, AM could be beneficial. However, if the tool is still working or the tool is expected to last the parts life cycle time, AM would not bring as many cost benefits. Tooling status as such is currently not proactively monitored by the company and therefore not possible to utilize in the scoring process.

Transportation and holding cost It was not possible to retrieve the holding cost per part with the provided MMLs. Therefore holding costs have been disregarded. Transportation costs are currently not calculated individually for each part. Transportation costs are estimated as a percentage of the purchasing cost. Hence, an increase of the purchase price (included in standard price) leads to an increase in transportation cost as well. Therefore it is sufficient to include the standard price as a criterion.


40

5.1.3 Criteria weighting (Step 3) The part criteria specified in Step 2 are allocated to the company’s supply chain goals, in the following named goals (see Figure 25). The goals represent the benefits that the company hopes to achieve with an implementation of AM for end-use parts and were defined by the sourcing department as follows:

• Reduce cost through the use of AM • Reduce downtime at the customer site through the use of AM

Figure 25 Criteria linked to supply chain goals regarding implementation of AM

(1) A low demand of parts at the company often involves high demand variability. Therefore, either safety stock costs for low demand parts or the downtime at the customer site are comparably high. By using AM for low demand parts both issues can be improved as the parts can be, in the best case, printed on-demand.

(2) Through AM, the MOQ can be significantly reduced as no tooling is needed. Hence, cost savings can be achieved when utilizing AM on parts with a high MOQ over Demand ratio.

(3) Print on-demand reduces the lead time and hence, safety stock costs or downtime can be reduced.

(4) If the Standard Price for a part is high, the use of AM may offer a cheaper option to manufacture the part.

(5) Using AM, no tooling change is necessary and lead time delays become more controllable, especially if the AM process is held in-house. This can lead to a reduction of safety stock or a reduction of downtime.

(6) As AM, in the best case scenario, can print on-demand, the ITR can be increased, which in return theoretically lowers stock holding costs and tied up capital.

The goals are not equally important and therefore a company specific weighting regarding the importance of both goals is evaluated using the AHP method and its pairwise comparison matrix. For all following AHPs, the Excel template of Goepel (2013), version 2016-05-04, is utilized. The AHP is a decision making tool that helps in different scenarios, such as prioritization, evaluation or benchmarking (Bhushan and Rai, 2003). For further information about the AHP method, see Saaty (2008). All following outcomes of the AHPs can be seen more detailed in Appendix 4.

GoalsCriteria

Reduce Cost Reduce Downtime

Expected Demand x xMOQ/Demand xStandard Price xLead Time x xLead Time Difference x xITR x


41

Eight employees within sourcing, project sourcing, logistic planning and manufacturing were asked to score both goals to obtain the pairwise comparison. Through interviewing employees working in different fields, the subjectivity of the AHP gets more leveled and a more homogeneous overall result can be obtained. Figure 26 illustrates the outcome for the evaluation of the goal weighting.

Figure 26 AHP Weighting of the supply chain goals

Additionally, the importance of each criterion for its allocated goal(s) is obtained by the author through the same AHP procedure. All criteria allocated to the same goal are pairwise compared by the author and one employee from logistic planning. According to the framework of Knofius, Van der Heijden and Zijm (2016), the following question was asked when conducting the pairwise comparison:

“If we improve both [criteria] values for the entire [part scope], which [criterion] does support the achievement of the company goal X better?”

The resulting criterion weighting regarding the importance for the goals can be seen in Figure 27 and Figure 28

Weight

Reduce Cost

- 4 1/5 80.8%

Reduce Downtime

1/4 - 19.2%

Reduce C

ost

Reduce

Dow

ntime

Figure 27 AHP result for criterion

weighting regarding goal "Reduce Cost"

Figure 28 AHP result for criterion weighting regarding goal "Reduce

Downtime"

Weight

Expected Demand

- 4 1/2 3 4 5 29,3%

MOQ / Demand

1/4 - 1/2 1 2 2 11,3%

S_Price 2 2 - 3 5 4 32,9%

ITR 1/4 1/4 1/3 - 7 3 15,8%

Lead Time 1/3 1/2 1/5 1/7 - 1 5,1%

Lead Time Difference

1/5 1/2 1/4 1/3 1 - 5,6%

S_P

rice

ITR

Lead Time

Lead Time

Difference

Expected

Dem

and

MO

Q /

Dem

and

Weight

Expected Demand

- 1 1 32,7%

Lead Time 1 - 1/2 26,0%

Lead Time Difference

1 2 - 41,3%

Expected

Dem

and

Lead Time

Lead Time

Difference


42

By multiplying the criterion importance weighting with the goal importance weighting, the particular criterion weighting per goal is obtained. The obtained weightings, both importance weighting and particular criterion weighting, are summarized in Figure 29. The overall criterion weighting is the sum of its particular weights per goal. It can be noted that the criterion Expected Demand and Standard Price received the highest weighting and are therefore more important than the other criteria.

Figure 29 Retrieving overall criterion weighting (rounded values)

In order to calculate the overall score of a part regarding its AM suitability, the obtained overall criterion weight is multiplied with its criterion score (see Step 2). All six obtained weighted criterion scores are summed up and an overall part score is retrieved. Figure 30 provides an exemplary result for a specific part. To summarize, following procedure is applied beginning with Step 2:

i. Scoring of part criteria for every part ii. Determining goal weighting and their related criterion weightings

with the AHP iii. Compute overall criteria weight iv. Multiply the scores of (i) with weightings of (iii), which results in

weighted criterion scores v. Sum weighted criterion scores in order to obtain a final (rounded)

part score


43

Figure 30 Exemplary part scoring

ArtNr Criterion Criterion Value

AHP Weight Score Weighted Score

xxxxxxx Expected Demand 53 0,30 0,95 0,29MOQ/Demand 0,19 0,09 0,00 0,00Lead time 31 0,09 0,41 0,04Standard Price 181,08 0,27 1,00 0,27Lead Time Difference 0 0,13 0,00 0,00ITR 1,85 0,13 0,79 0,10

0,7Overall rounded score


44

5.1.4 Results of preliminary assessment As it can be seen in Figure 31, the part score distribution shows that the majority of parts were assigned a score between 0,4 and 0,6. Therefore, only parts that scored higher than 0,6, which equals 83 parts or 18% of the overall scored parts, are considered as interesting for AM. These parts are taken into account for the following technical assessment.

Figure 31 Part score distribution

The preliminary assessment provides a good prioritization of parts where the chance to gain opportunities and benefits through AM are higher. Nevertheless, a technical assessment is necessary, as the scoring relies only on economical and logistical criteria. In order to assert if the part is actual feasible to print, PDM data, such as technical drawings or required properties of the specific plastic material must be analyzed manually.

5.2 TECHNICAL ASSESSMENT

By reviewing the part structure of each considered part with the help of the PDM system, it was discovered, that even though a part is not seen as an assembly in the MML, it may consists of different subparts in different materials, if the assembly took place at the supplier’s site. As assemblies are not part of the research scope, these parts are disregarded (see Figure 32). Additionally, all parts which were either not findable in the PDM system or did not have any CAD file linked, were ignored.

Next, material information, provided by the PDM system, was taken into consideration. Currently, transparent end-use plastic parts that withstand the necessary chemical requirements (mainly food contact certified, high chemical resistance and low water absorption) cannot be manufactured additively.

0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

Parts per Score 5 24 43 97 115 97 54 25 4

Percentage 1% 5% 9% 21% 25% 21% 12% 5% 1%

0

20

40

60

80

100

120

140

Part

s pe

r Sco

re


45

There is no suitable substitution material for Polymethylmethacrylate (PMMA) or any other transparent material available. Furthermore, even if material would be available, according to an AM supplier, transparent parts have to be polished after the AM process, which is estimated to be more cost-intensive for the company’s’ eligible parts in comparison to today’s utilized manufacturing method. For the future, the company could look into these parts and assess if it is really necessary to use transparent material or if the benefit of using AM would outweighs the benefits of transparent material. For further manual assessment the transparent parts are disregarded (see Figure 32). Finally, parts that demand special requirements from the material, such as stabilized against dry heat or very high stability against chemicals due to their use in the cleaning process, are disregarded. These parts need to be excluded as currently only a limited range of “basic” plastic material is available for AM.

Figure 32 Manual assement steps

In general, the material which is used for the remaining 16 parts, hereafter named original material, can be substituted as stated in Figure 33. This is a general conclusion and works, according to a material specialist of an AM printer supplier as well as the company’s material specialist, in most cases. None of the 16 parts is in contact with food, which makes it easier to choose substitution material, as from a compliance point of view no food graded certificates for the material are needed. PA 12 is in general comparably expensive, which is the reason why it is very seldom used as original material. Even though none licensed propylene material powder for the SLS process can be found, it is officially not yet supported by the AM printer supplier and therefore not taken into consideration as a SLS material. It is expected that in near future by AM printer supplier licensed propylene material will be available (3dprintingindustry.com, 2017).

Since all materials have different properties, for further more detailed investigations, each part has to be analyzed separately in what kind of environment, in terms of temperature, chemicals or UV exposure it is operating in. Furthermore, the specific requirements on e.g. tensile strength or toughness have to be investigated. To summarize, at the current state of

Steps Filter Explanation Remaining parts

1 Score of > 0,6 all parts with score >0,6 are taken into account for manual assesment

83

2 assemblies parts where assembly took part at the suppliers site are disregarded

55

3 deficient PDM dataall parts that are not findable or don't have CAD drawings in the PDM system are disregarded 37

4 transparent material all parts that uses transparent material are disregarded

23

4 part material original material with special requirements are disregarded

16


46

research it is seen as sufficient to make the assumption that the AM materials, PA12 respectively PC are fully capable of replacing the original material.

Figure 33 Substitution material according to AM Technology

Together with an external AM expert, the remaining parts where assessed based on the technical drawings. It became clear that none of the parts, at the current state with the current specifications and requirements, can be additively manufactured. As the parts are not intended to be manufactured additively, they are not specifically designed for AM. All 16 parts require too tight tolerances, in general ISO2768-m or single specific tolerances, which are currently not able to meet with AM. Figure 34 provides a comparison between the ISO 2768 m tolerances and achievable tolerances using a SLS process.

Figure 34 SLS Tolerances for specific feature length [mm], provided by an AM service

bureau and ISO 2768-m tolerances (ISO 2768-1, 1989)

One reason for the existing tight tolerances could be, that the current manufacturing process for the 16 parts, plastic machining, is capable to deliver the tolerances without any significant additional expenses. Perhaps, if choosing AM, looser tolerances would still deliver the necessary function of the part. Also current requirements on surface finish of Ra 3,2 respectively Ra 6,3 cannot be met with AM and the surface finishing in general is very depending on part orientation in the printer, angled surfaces and AM method. If a part surface is slightly angled visible steps will occur when using AM.

It can be concluded that it is not possible to just replace the conventional manufacturing process without changing the design specifications and adapting them to AM. It is important to emphasize, that in order to make the parts suitable for AM, the design engineer, together with the product owner, has to evaluate and reassess the part design. Tolerances and surface roughness need to be reassessed and certain AM restrictions, such as angel steepness, wall thickness or surface warping must be considered in the part design (c.f. stratasysdirect.com, 2017; EOS GmbH, 2017). It could be sufficient for certain parts to apply a transparent coat on the outer surface to minimize the surface roughness in order to increase the parts dirt repellence. Furthermore, in order

Original Material FDM Material SLS MaterialPA66-GF30, Plasticamide reinforced PC, Polycarbonate PA 12, PolyamidePEHD, Polyethylene high density PA 12, Polyamide PA 12, PolyamidePOM, Polyoxymethylene PA 12, Polyamide PA 12, PolyamidePP-H, Propylene, homopolymer PA 12, Polyamide PA 12, PolyamidePP-H-GF20, Propylene reinforced, homopolymer PC, Polycarbonate PA 12, Polyamide

Feature Length 10 30 50 75 100 125 150 200 250 300 350 400Tolerances AM ±0.20 ±0.28 ±0.38 ±0.39 ±0.45 ± 0.51 ±0.58 ±0.70 ±0.83 ±0.95 ±1.08 ±1.20Feature Length 6-30Tolerances ISO ±0.20

30-120±0.28

120-400± 0.5


47

to gain more AM specific benefits, it would be even better, if the entire part is redesigned according to Design for AM (DfAM) rules.

Another finding is, that all remaining parts are currently plastic machined. The explanation is, that tool costs for injection molded parts were not regarded in the part assessment. As in the company, purchasing costs for tools are development costs, they are not allocate to the purchase price, hence, purchase costs for injection molding parts are in general very low. Few more costly parts that are injection molded contain subassemblies and were therefore disregarded during technical assessment. Injection molding parts should be assessed again, as soon as major costs occur through for example tool breakdown or tool modifications.

5.3 ECONOMICAL ASSESSMENT

To define the potential of AM for the company, the costs utilizing AM for the identified parts were investigated in an economical assessment. For the following course of action the assumption is made that tolerances and surface roughness can be adjusted to feasible values for AM and only minor design adjustments are made. In reality adjustments of design and requirements are a comprehensive and extensive process which requires the expertise of different functions inside the company. For the current state of research it is seen as sufficient to continue with the above mentioned assumptions.

Two different methods for the cost investigation were used. Firstly, the potential in-house manufacturing of the parts were considered and the AM part costs were calculated using proposed equations from previous researches. Secondly a request for quote (RFQ) was sent to a service provider of AM for the potential case of buying AM manufactured parts. The manufacturing costs respectively purchasing costs of both methods are compared with the existing purchasing cost.

For the potential case of in-house manufacturing, only the FDM technology is considered as suitable, as SLS technology requires specific knowhow, which is at the moment not available in the company. According to Schmid and Levy (2014), knowhow regarding the SLS process and powder reprocessing must exists, in order to ensure the necessary quality of the printed end-use parts. Furthermore, the SLS technology requires higher capital investments, as at least two separate rooms with exhaust ventilation and specialized equipment for parts finishing and powder mixing are needed. FDM technology, on the other hand, is more user friendly, clean and safe, even for an office environment and compared to SLS technology, less process knowhow is necessary (stratasysdirect.com, 2017). According to several AM printer suppliers (both providing FDM and SLS technology), SLS technology is mainly


48

used by organizations that specialize into AM. These organizations have the ability to utilize the SLS printer on a higher degree of capacity utilization, which means the building chamber of the printer is packed with as many parts as possible. High utilization of the chamber means less powder waste, as only max. 70% of the unused powder can be recycled (3D Systems, 2016). Hence, the costs for the chamber cleaning and preparation for the next production become a smaller fraction of the total part cost. Using the FDM technology, the chamber does not need to be filled with as many parts as possible as no material waste occurs. Additionally, no cleaning of the chamber is necessary and the cost for preparation of a production run, mainly costs of initial heating of the chamber, are negligible.

5.3.1 Cost Model for in-house manufacturing using FDM Technology

The in-house manufacturing cost calculation is based on the research of Hopkinson and Dickens (2003) and Ruffo, Tuck and Hague (2006), who identified the different types of cost associated with AM. Both studies propose a way how to sum up AM related costs and breaking it down to an actual part. According to Hopkinson and Dickens (2003), the costs can be broken down into:

• Machine costs • Labor costs • Material costs

Overhead factors such as energy, space rental costs or consumables and administrative overhead costs are not considered. It is argued by Hopkinson and Dickens (2003) that these factors contribute with less than 1% to the final costs. Furthermore, when comparing manufacturing cost with purchasing cost of a part, it is argued by the author, that the neglected overhead costs are in the same range as potential transportation costs if the part will be purchased. Also potential costs for designing the part according to DfAM rules and testing are not included, as they are not included in the purchasing part price of the current manufacturing setup. Costs for support material as well as additional print time costs for the support structure are disregarded, as these costs are heavily depending on part design and printing part orientation and therefore not able to determine in general. The FDM technology does not require any post-processing time or manual finishing work as the support structure can be dissolved in water.

It was chosen by the author and an external AM expert to base the calculations on the printer Stratasys Fortus 380. The printer is seen as the most suitable for the parts in investigation from the build envelope and accuracy perspective.


49

Figure 36 shows a scheme of the applied costing model with regarded and disregarded costs. An important assumption was made about the productivity of the printer, which was estimated to have an utilization of 90% (based on Hopkinson and Dickens, 2003). As seen in Figure 36, only material cost is considered as a direct cost. All other costs are allocated to a single part by the volume or the footprint of the part. For the cost calculations, it was assumed that the printers load platform is always utilized to the maximum and per platform built, a set-up time of 0,5h is needed (see Figure 37). No more labor is needed once the platform is set-up. In comparison to a proposed set-up time of 10 min per platform (Hopkinson and Dickens, 2003), the utilized 30 min as set-up time are very conservative. The labor costs can be distributed to the parts considering their footprint (dimension length and width). Figure 37 provides details of the cost calculations and all calculated part costs can be found in Appendix 5.

To assure the consistency of the established cost model, an AM service bureau was asked to provide a cost quote for the case that the parts are printed in-house with the FDM technology. The provided manufacturing costs are based on the assumption that no depreciation costs and maintenance costs occur (e.g. no overhead costs). As shown in Appendix 6, the computed manufacturing costs are very similar to the provided quoted costs if the same material price in the calculation and the quote is used (0,22 €/cm3). This shows that the established cost model is valid and it could be used as a foundation for further investigations.

The comparison between manufacturing cost with FDM technology and the purchasing cost using a conventional technology highlights only 3 out of 16 parts that could potentially be manufactured more cost effective with a FDM printer in-house. As the labor cost is only a very small fraction of the manufacturing cost, differences in manufacturing costs between Germany and Sweden would be very small. In Figure 35, a comparison between FDM manufacturing in-house cost and current purchase price is made. With a calculated estimated annual saving of 674€, in-house production is not interesting for the company, as the return on investment is too low.

Figure 35 Comparison in-house FDM technology and conventional manufacturing

17% 17% 30%

0%

50%

100%

Part 8 Part 9 Part 15Manufacturing costs, AM Purchase price, conventional


50

Figure 36 Cost model for AM manufacturing cost calculation based on Ruffo, Tuck and

Hague (2006)

Figure 37 Cost Details for Manufacturing Cost Calculation

Unit Variable Obtained by

Machine Capacity h/year 5464,8 C 235*24*UMachine Utilisation % 90% UProcess Time h/cm3 0,08 STechnician annual working hours h 2024 WH 235*8Platformsize cm2 1022 PSLabour per full platform h 0,50 f

Material Costs Material Costs €/cm3 0,33 MC

Labour Costs Labour Costs per hour €/h 21,68 LCh Company figure

Labour Costs footprint fraction €/cm2 0,0106 LCcm (f*LCh)/PS

Machine Costs Machine Purchase* € 97300 EPurchase Cost per year € 12162,5 D E/8Maintenance Cost per year € 5838 Ma E*6%

Machine Cost per year € 18000,5 aM D+Ma

Decepriation Cost per cm3 €/cm3 0,18 Ecm (D/C)*SMaintenance Cost per cm3 €/cm3 0,09 Mcm (Ma/C)*S*Prices in general include additional, obligatory devices but exclude PC or material licenses

Calculation of Manufacturing Costs FDM


51

5.3.2 Cost for AM service provider using SLS technology For the potential external production, a RFQ was sent to the AM service provider “Materialise”. The RFQ contained following information:

• CAD files in .STL format • Technical drawings • Request for price calculation with MOQ steps based on the expected

demand • To be utilized material: PA12

It was clarified that the manufactured parts are not used as prototypes but as end-use parts. The provided AM costs using SLS technology can be found in Appendix 7, the costs are hereby purchasing costs. From the provided costs, it can be concluded that the SLS technology is, given a high utilization, more cost effective compared to FDM technology. The provided purchase costs for 6 parts were lower than the actual purchasing price using plastic machining. It appears that Figure 38 is confirming the finding of Baldinger et al. (2016), that a relationship exists between costs per cm3 and volume of the part when using a AM service provider: cost per cm3 decreases with the increased volume of a part until a certain threshold, where it changes to a linear relationship. This could be the reason why the FDM manufacturing costs are comparably smaller for part 9, 10 and 15. The volumes of these parts are comparably small and resulting, the costs per cm3 when using a service provider are high.

Figure 38 Cost comparison FDM, SLS and conventional manufacturing

It must be emphasized that no design changes have been made and potentially more cost savings occur if the parts would be redesigned for AM. Especially parts with high volume of material are seen by the author as good candidates for potential savings using DfAM. Based on the expected demand and the provided purchasing costs for AM, potentially total annual savings for the 6 parts amount to 2667€.

0%

50%

100%

150%

200%

250%

Part 3;388 cm3

Part 4;923 cm3

Part 9;7 cm3

Part 10;7 cm3

Part 14;30 cm3

Part 1517cm3

Purchase price, SLS Purchase price, conventional Manufacturing Cost, FDM

Effects on the company’s supply chain

52

6 EFFECTS ON THE COMPANY’S SUPPLY CHAIN

If the manufacturing technology is changed from a conventional to an AM technology, the supply chain will also change. In this chapter potential effects on the supply chain of the company are presented, both for the case of in-house AM manufacturing and the buy scenario using an AM service provider.

6.1 MAKE SCENARIO USING FDM TECHNOLOGY

When switching process to in-house AM manufacturing, the company’s supply chain has to integrate new suppliers in their network: the AM machine vendor and an AM material supplier. In most cases the AM machine vendor incurs also the AM material supplier function. As labor costs are only a very small fraction of the manufacturing costs for FDM printing, the decision where to base the printer depends mostly on other aspects, such as location of demand or import taxes. As in-house manufacturing is not yet seen as cost efficient for the company, it should rather concentrate on utilizing an AM service provider.

6.2 BUY SCENARIO USING AN AM SERVICE SUPPLIER

According to several researches (see chapter 2.3), stock can be reduced with the use of AM due to the potential of on-demand production. This leads to a reduction of safety stock, which in theory lowers holding costs. Nevertheless, in the case of the company and the researched part scope, AM would not necessarily lead to holding cost reductions. A usual manufacturing time of 4 to 5 days, using an AM service provider within closer distance, is seen as feasible and therefore, most likely a small safety stock must remain in the warehouse. Due to the fact that the great majority of the researched part scope is stored in bin locations in a small parts warehouse, fixed holding costs per bin location are applied. The holding costs are independent on the number of the stocked parts in the bin. Hence, holding costs are applied as soon as the part is allocated to a bin location. AM could only achieve holding costs savings if the part would be produced on-demand with no safety stock at all. Nevertheless, the use of AM would lower capital costs for the stock as it is possible to order parts with an effective MOQ of 1 and the lead times for the feasible parts could be reduced significantly from current 26-31 days to 4-5 days.

Parts that could be produced more cost effective using a SLS technology service provider, based on the findings from chapter 5.3, are currently sourced from two suppliers based in Sweden. All AM can be produced at a single service supplier near Germany. 29% of the total annual volume value sourced from supplier A could be additively manufactured, which would have a relatively big impact on the supply chain setup of supplier A. When deciding to use AM, further investigations regarding the change of logistic costs or


53

negation power due to changed volume value for supplier A are necessary. The fraction of total volume value from supplier B is with 1% neglectable and an impact on the supply chain setup of supplier B is not expected. If the entire supply chain is observed the volume value that is affected by the use of AM is very low and hence no major effects on the supply chain can be seen.

Using conventional manufacturing techniques and its required batch production, a distributed production is not beneficial for the researched parts due to their low volume. Whereas Hopkinson and Dickens (2003) have shown that manufacturing costs for AM parts are not dependent on the quantity produced and ordering with an effective MOQ of 1 is potentially possible. Hence, a supply chain setup for AM parts using a distributed production in national locations close to major markets instead of concentrating production in a centralized location could be an alternative (cf. Figure 39).

Figure 39 Centralized vs. decentralized production


54

A distributed production supply chain setup may bring, according to Khajavi, Partanen and Holmström (2014), essential benefits, such as lead time reduction, and consequently enables service level improvements (Treville, Shapiro and Hameri, 2004). Another possible resulting benefit that has been discussed together with the logistic planning function could be the saving of import taxes and shipping costs. Nevertheless, with the current state of AM adoption in the company, it is seen as more beneficial to only consider a centralized production, preferably near the DC in Germany. At the current state, the costs that occur with the certification of every AM service supplier are estimated to be much higher than the actual benefits of a distributed production. The AM producible part scope in the company is not sufficiently high to justify a distributed production. It is important to have in mind that a distributed production could become interesting as soon as parts are designed or redesigned for AM and the AM producible part scope grows as this is accompanied by a greater costs saving potential.

To summarize, it is found that using the current researched AM producible part scope, centralized production near the DC in Germany is clearly the preferable supply chain setup for the company. The supply chain setup will not change drastically, but AM production will contribute to a shortened lead time in comparison with conventional methods (for current lead time of conventional methods see Appendix 7), reduce capital costs. An additional finding is that AM will not contribute to the reduction of holding costs in the company`s supply chain.

Conclusion and discussion

55

7 CONCLUSION AND DISCUSSION

Following overall research question was defined in the beginning of the research:

How and for which parts in the company could the implementation of AM be beneficial and what would be the effects on the supply chain?

As AM becomes more and more suitable for plastic end-use parts, an incremental implementation of AM in the company is proposed:

1. Production alternative: As elaborated, AM could provide a beneficial manufacturing alternative compared to conventional technologies due to its suitability for low demand production.

2. Redesign and part consolidation: After gaining understanding of AM and its challenges through using it as a production alternative, it is proposed to redesign and consolidate parts. Assemblies can be reduced significantly through redesigning and consolidation, which leads to further cost reductions.

3. Decentralization: In the long term future it is estimated that AM system prices will significantly fall, which could lead to disruption in the supply chain as, for instance, it might be possible to use a decentralize production located at the service places.

The research concentrated on the use of AM as a production alternative and therefore a part evaluation framework was established using a preliminary assessment, a technical and an economic assessment. A top-down approach was adopted for the preliminary assessment, initiated with a large part scope, in order to score and identify parts that could be additively manufactured in a beneficial way. Out of initial 1000 plastic machined and injection molded parts, 83 parts received a score of 0,6 or higher and were seen as promising. Hence, these parts were used for the successive assessments.

It can be concluded, after the technical assessment, that the company could, at the current time, in 6 cases utilize an AM service supplier to beneficially substitute plastic machining manufacturing. The use of AM could help to reduce the lead time and purchasing costs for the parts and hence would support the fulfilling of the supply chain goals “reduce cost” and “reduce downtime at the customer”. Annual cost savings of 2667€ are estimated by using an AM service provider. Additionally the lead time could be significantly reduced to around 4-5 days from current 26-31 days. Further research needs to be done, in order to quantify the benefits of a shorter lead time for the specific parts. Due to the high investment costs for AM in-house manufacturing and the currently insufficient amount of printable parts, it is not seen as


56

beneficial to invest in an AM printer. Furthermore, there is no economic benefit to consider AM for injection molded parts as long as no major tooling breakdown occurs, as the tools are already available.

The research has shown that AM can be used as an economically beneficial production alternative regarding the assumption that tolerances can be adjusted to AM feasible values. Besides that, the research provides a methodical way of assessing parts regarding AM, as well as a cost model for manufacturing parts utilizing FDM technology. Both findings can be used for future research and contribute acquiring AM related knowledge at the company.

Costs of AM are heavily depending on the print time, which is related to the part material volume, and therefore it must be emphasized how important it is to redesign parts for AM. Through DfAM, print time and material volume could be reduced and more benefits would be gained. Furthermore, the parts can be designed for their function as the design restriction are less compared to conventional manufacturing. The cost comparison between AM and conventional manufacturing did not account the costs of moulding tools and therefore it is proposed to investigate if AM could be beneficial for a part specifically, at the time when major tooling costs occur.

The actuality of the research topic is highlighted by the still limited information that was available during the literature review on how to assess parts for AM in a methodical way. Only the existing part design was taken into account: the same part, designed for conventional manufacturing, is manufactured by AM without changes in geometry. However as many researches has shown (e.g. Atzeni et al., 2010), the geometric feasibility of additive processes allows a redesign of the part, which has an impact on the cost. Hence, it is strongly recommended to investigate in AM already in the early phase of part development. Therefore, it is important to sensitize all functions in the company about AM that are involved in the design process. For the short term future, it is recommended to redesign and print one of the economical feasible parts, as this would help to understand how much benefits can be gained when using DfAM. Furthermore, short term future investigations could be made regarding part consolidation of already used parts. Part consolidation provides the possibility to eliminate assembly and consequently, it would lead to a reduction of labor, handling and logistic costs (Zanardini et al., 2015).

In regards to choosing the appropriate manufacturing technology for new parts in the company, a business case should also include AM, as the company’s part scope contains many parts with a demand volume where AM could be an


57

alternative to injection molding. Costs for tooling and the tooling lead time, which is about 6 months, could likely make AM comparably more beneficial.

During the research, it became clear that the media is often too enthusiastic about AM and the simplicity of it. However, speaking to AM service providers and AM printer suppliers, AM is still not mature and a lot of research is ongoing concerning standardization, usable materials, throughput rate or the surface and dimensional quality. It is important to emphasize that due to the rapid growth of AM technologies, an AM market review in constant intervals can help the company to understand and use AM in a better way and will help to decide the right point of time of using AM as a production alternative for end-use parts.


58

8 REFERENCE LIST

3D Printing Media Network (2016) Capabilities Expand With Next Generation Fortus Production 3D Printer From Stratasys - 3D Printing Media Network. Available at: https://www.3dprintingbusiness.directory/news/capabilities-expand-next-generation-fortus-production-3d-printer-stratasys/ (Accessed: 13 June 2017).

3D Systems (2016) Tips and info for building with DuraForm ProX PA. Available at: http://infocenter.3dsystems.com/product-library/system/files/legacy/3DSCentral%20-%20Production%20Printers/ProX%20500/Customer/Manuals/2484_CIB%201601%20-%20Building%20with%20DuraForm%20ProX%20PA%20Engineered%20Thermoplastic.pdf (Accessed: 9 June 2017).

Additively (2017) Overview over 3D printing technologies - Additively - your access to 3D printing. Available at: https://www.additively.com/en/learn-about/3d-printing-technologies (Accessed: 12 May 2017).

Ahn, S. et al. (2002) ‘Anisotropic material properties of fused deposition modeling ABS’, Rapid Prototyping Journal, 8(4), pp. 248–257. doi: 10.1108/13552540210441166

Atzeni, E. et al. (2010) ‘Redesign and cost estimation of rapid manufactured plastic parts’, Rapid Prototyping Journal, 16(5), pp. 308–317. doi: 10.1108/13552541011065704

Baldinger, M. et al. (2016) ‘Additive manufacturing cost estimation for buy scenarios’, Rapid Prototyping Journal, 22(6), pp. 871–877. doi: 10.1108/RPJ-02-2015-0023

Basiliere, P. (2016a) Technology Overview for Binder Jet 3D Printing. Available at: https://www.gartner.com/doc/2981218/technology-overview-binder-jet-d (Accessed: 12 May 2017).

Basiliere, P. (2016b) Technology Overview for Material Jetting. Available at: https://www.gartner.com/doc/2962018/technology-overview-material-jetting (Accessed: 12 May 2017).

Basiliere, P. (2017) Innovation Insight for Material Extrusion 3D Printing. Available at: https://www.gartner.com/doc/2668817/technology-overview-material-extrusion-d (Accessed: 2 May 2017).

Bhushan, N. and Rai, K. (2003) Strategic decision making: Applying the analytic hierarchy process. (Decision engineering, 1619-5736). New York: Springer.


59

Bourell, D.L. et al. (2009) A Brief History of Additive Manufacturing and the 2009 Roadmap for Additive Manufacturing: Looking back and looking ahead. 24 September.

Cohen, M.A., Agrawal, N. and Agrawal, V. (2006) ‘Winning in the Aftermarket’, Harward Business Review (May). Available at: https://hbr.org/2006/05/winning-in-the-aftermarket.

Cotteleer, M. (2014) 3D printing: “Complexity is free” may be costly for some, 2 June. Available at: https://dupress.deloitte.com/dup-us-en/focus/3d-opportunity/3d-printing-complexity-is-free-may-be-costly-for-some.html (Accessed: 2 June 2017).

Cotteleer, M. and Joyce, J. (2017) 3D opportunity: Additive manufacturing paths to performance, innovation, and growth. Available at: https://dupress.deloitte.com/dup-us-en/deloitte-review/issue-14/dr14-3d-opportunity.html (Accessed: 13 June 2017).

Cotteleer, M., Holdowsky, J. and Mahto, M. (2014) The 3D opportunity primer: The basics of additive manufacturing. Available at: https://dupress.deloitte.com/content/dam/dup-us-en/articles/the-3d-opportunity-primer-the-basics-of-additive-manufacturing/DUP_718-Additive-Manufacturing-Overview_MASTER1.pdf (Accessed: 15 May 2017).

Economist (2012) Print me a jet engine. Available at: http://www.economist.com/blogs/schumpeter/2012/11/additive-manufacturing (Accessed: 13 April 2017).

EOS GmbH (2017) Basic Design Rules for Additive Manufacturing. Available at: https://cdn1.scrvt.com/eos/public/ab4f0542d66453fc/5f889ab7e3f72bd3d44b22205ba8b68b/EOS-Basic-Design-Rules_Additive-Manufacturing_EN.pd (Accessed: 17 May 2017).

F42 Committee (2012) Standard Terminology for Additive Manufacturing Technologies. West Conshohocken, PA: ASTM International.

Gao, W. et al. (2015) ‘The status, challenges, and future of additive manufacturing in engineering’, Computer-Aided Design, 69, pp. 65–89. doi: 10.1016/j.cad.2015.04.001

GE Additive (2016) GE Additive Manufacturing in Alabama: The Future Is Now: Press Release. Available at: http://www.geadditive.com/press-releases/ge-additive-manufacturing-in-alabama-future-is-now (Accessed: 13 April 2017).

Gibson, I. (2017) ‘The changing face of additive manufacturing’, Journal of Manufacturing Technology Management, 28(1), pp. 10–17. doi: 10.1108/JMTM-12-2016-0182


60

Gibson, I., Rosen, D. and Stucker, B. (2015) ‘Additive Manufacturing Technologies: 3D Printing, Rapid Prototyping, and Direct Digital Manufacturing’ (509pp). doi: 10.1007/978-1-4939-2113-3

Goepel, K.D. (2013) Implementing the Analytic Hierarchy Process as a Standard Method for Multi-Criteria Decision Making In Corporate Enterprises: A New AHP Excel Template with Multiple Inputs. Singapur (Proceedings of the International Symposium on the Analytic Hierarchy Process 2013). Available at: http://bpmsg.com/wordpress/wp-content/uploads/2014/01/AHPcalc-v2013-12-24a.pdf (Accessed: 5 May 2017).

Guo, N. and Leu, M.C. (2013) ‘Additive manufacturing: Technology, applications and research needs’, Frontiers of Mechanical Engineering, 8(3), pp. 215–243. doi: 10.1007/s11465-013-0248-8

Holmström, J. et al. (2010) ‘Rapid manufacturing in the spare parts supply chain’, Journal of Manufacturing Technology Management, 21(6), pp. 687–697. doi: 10.1108/17410381011063996

Hopkinson, N. and Dickens, P. (2003) ‘Analysis of rapid manufacturing—using layer manufacturing processes for production’, Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 217(1), pp. 31–39. doi: 10.1243/095440603762554596

Hopkinson, N., Hague, R. and Dickens, P. (2006) ‘Introduction to Rapid Manufacturing’, in Dickens, P.M., Hague, R.J.M. and Hopkinson, N. (eds.) Rapid manufacturing: An industrial revolution for the digital age. Chichester: Wiley, pp. 1–4.

HP Development Company (2017) HP Jet Fusion 3D Printing Solution: Reinventing Prototyping and Manufacturing. Available at: http://h20195.www2.hp.com/v2/getpdf.aspx/4AA6-4894ENA.pdf?ver=3 (Accessed: 2 May 2017).

http://www.researchandmarkets.com/research/m5mq6g/additive (2016), 3 May (Accessed: 3 May 2017).

https://3dprintingindustry.com/news/diamond-plastics-polypropylene-laser-sintering-3d-printing-34174/ (2017) (Accessed: 17 May 2017).

https://www.stratasysdirect.com/resources/fused-deposition-modeling/ (2017) (Accessed: 17 May 2017).

ISO 2768-1 (1989) General tolerances-Part 1: Tolerances for linear and angular dimensions without individual tolerance indications: Technical Committeee ISO/TC 3.


61

Khajavi, S.H., Partanen, J. and Holmström, J. (2014) ‘Additive manufacturing in the spare parts supply chain’, Computers in Industry, 65(1), pp. 50–63. doi: 10.1016/j.compind.2013.07.008

Knofius, N., Van der Heijden, M. and Zijm, W. (2016) ‘Selecting parts for additive manufacturing in service logistics’, Journal of Manufacturing Technology Management, 27(7), pp. 915–931. doi: 10.1108/JMTM-02-2016-0025

Leandri, A. (2015) What Can Additive Manufacturing Do for Tooling?, 3 May. Available at: https://3dprint.com/55676/additive-manufacturing-tooling/ (Accessed: 3 May 2017).

Levitt, T. (1965) ‘Exploit the Product Life Cycle’, Harward Business Review, 43, pp. 81–94. Available at: https://hbr.org/1965/11/exploit-the-product-life-cycle (Accessed: 29 May 2017).

Li, Y. et al. (2016) ‘Additive manufacturing technology in spare parts supply chain: A comparative study’, International Journal of Production Research, 55(5), pp. 1498–1515. doi: 10.1080/00207543.2016.1231433

Lindemann, C. et al. (2015) ‘Towards a sustainable and economic selection of part candidates for additive manufacturing’, Rapid Prototyping Journal, 21(2), pp. 216–227. doi: 10.1108/RPJ-12-2014-0179

Liu, P. et al. (2014) ‘The impact of additive manufacturing in the aircraft spare parts supply chain: Supply chain operation reference (scor) model based analysis’, Production Planning & Control, 25(13-14), pp. 1169–1181. doi: 10.1080/09537287.2013.808835

Loughborough University (2017a) Directed Energy Deposition, 5 May. Available at: http://www.lboro.ac.uk/research/amrg/about/the7categoriesofadditivemanufacturing/directedenergydeposition (Accessed: 15 May 2017).

Loughborough University (2017b) Sheet Lamination. Available at: http://www.lboro.ac.uk/research/amrg/about/the7categoriesofadditivemanufacturing/sheetlamination/ (Accessed: 2 May 2017).

Mellor, S. (2014) An Implementation Framework for Additive Manufacturing. Doctor Thesis. University of Exeter.

Michael Shanler and Pete Basiliere (2016) Hype Cycle for 3D Printing, 2016 (Accessed: 16 May 2017).


62

National Center for Manufacturing Sciences (1998) The Road to Manufacturing: 1998 Industrial Roadmap for the Rapid Prototyping Industry. NCMS Report 0199RE98. Ann Arbor.

Ott, T. (2015) Podcast 3D opportunity for the supply chain: Additive manufacturing delivers: Driving supply chain transformation. Available at: https://dupress.deloitte.com/dup-us-en/multimedia/podcasts/additive-manufacturing-3d-printing-supply-chain-transformation-business.html (Accessed: 13 April 2017).

Ruffo, M., Tuck, C. and Hague, R. (2006) ‘Cost estimation for rapid manufacturing - laser sintering production for low to medium volumes’, Journal of Engineering Manufacture, 220(9), pp. 1417–1427. doi: 10.1243/09544054JEM517

Saaty, T.L. (2008) ‘Decision making with the analytic hierarchy process’, International Journal of Services Sciences, 1(1), p. 83. doi: 10.1504/IJSSCI.2008.017590

Schmid, M. and Levy, G. (2014) Quality management and estimation of quality costs for Additive Manufacturing with SLS. Zürich. Available at: https://www.iwf.mavt.ethz.ch/ConfiguratorJM/publications/Quality_Ma_133345735583221/p107.pdf (Accessed: 24 May 2017).

Siemens AG (2016) Press Release: Siemens strengthens footprint in Additive Manufacturing. Available at: https://www.siemens.com/press/pool/de/pressemitteilungen/2016/power-gas/PR2016080358PGEN.pdf (Accessed: 3 May 2017).

Simkin, Z. and Wang, A. (2014) Cost-Benefit Analyses for Final Production Parts.

Statasys Strategic Consulting (2017) Implementing 3D Printing: In-house or Outsource? Available at: https://www.stratasysdirect.com/blog/3d-printing-service-vs-buying-3d-printer/ (Accessed: 24 May 2017).

Treville, S. de, Shapiro, R.D. and Hameri, A.-P. (2004) ‘From supply chain to demand chain: The role of lead time reduction in improving demand chain performance’, Journal of Operations Management, 21(6), pp. 613–627. doi: 10.1016/j.jom.2003.10.001

Wienken, R. and Kilger, C. (2016) How will 3D printing make your company the strongest link in the value chain? EY’s Global 3D printing Report 2016.

Wohlers, T. and Caffrez, T. (2016) ‘Additive Manufacturing: The State of the Industry’, pp. 45–52. Available at: https://www.sme.org/uploadedFiles/Publications/ME_Magazine/2016/May/May%202016%20f1%20Wholers.pdf (Accessed: 12 April 2017).


63

Zalak Shah and Pete Basiliere (2016a) Technology Overview for Directed-Energy Deposition 3D Printing (Accessed: 15 May 2017).

Zalak Shah and Pete Basiliere (2016b) Technology Overview for Powder Bed Fusion 3D Printing (Accessed: 15 May 2017).

Zanardini, M. et al. (2015) ‘Benefits and costs of additive manufacturing applications: an holistic evaluation guideline’, in Proceedings of 20th Summer School Francesco Turco. Available at: https://www.researchgate.net/publication/306100171_Benefits_and_Costs_of_Additive_Manufacturing_Applications_An_Evaluation_Guideline.

Appendix

64

9 APPENDIX

9.1 APPENDIX 1

Figure 40 Company’s plastic part supply chain in Europe

Appendix

65

9.2 APPENDIX 2

Figure 41 AM technologies, corresponding base materials, and advantages and

disadvantages (Cotteleer, Holdowsky and Mahto, 2014)

Appendix

66

Figure 42 Technologies and materials matrix (Cotteleer, Holdowsky and Mahto, 2014)

Appendix

67

9.3 APPENDIX 3

(1) (Gibson, Rosen and Stucker, 2015) (2) (Liu et al., 2014) (3) (Holmström et al., 2010) (4) (Khajavi, Partanen and Holmström, 2014) (5) (Li et al., 2016) (6) (Gao et al., 2015)

Reduc

e man

ufactu

re

costs

for s

mall lo

t size

s

Mass c

ustom

isatio

nCom

plex d

esign Part

cons

olida

tion

Shorte

r lead

time

Decen

tralisa

tion

Inven

tory h

olding

cost

reduc

tion Sus

taina

bility Sys

tem co

sts Limite

d mate

rials Lim

ited s

izeBuil

t spe

ed

(1)

xx

xx

xx

xx

x(2

)x

xx

xx

(3)

xx

xx

xx

(4)

x(5

)x

x(6

)x

xx

xx

xA

dvan

tage

sC

halle

nges

Figure 43 Advantages and Challenges of AM

Appendix

68

9.4 APPENDIX 4

Appendix

69

AHP Criteria related to "Reduce Cost" (EVM multiple inputs)K. D. Goepel Version 04.05.2016 Free web based AHP software on: http://bpmsg.com

Only input data in the light green fields and worksheets!

n= Number of criteria (2 to 10) Scale: 1 AHP 1-9

N= Number of Participants (1 to 20) α : 0,1 Consensus: n/a

p= selected Participant (0=consol.) 2 7

Objective

Author

Date Thresh: 1E-07 Iterations: 5 EVM check: 1,5E-08

Table Comment Weights Rk1 29,3% 22 11,3% 43 32,9% 14 15,8% 35 5,1% 66 5,6% 57 0,0%8 0,0%9 0,0%

10 0,0%

Result Eigenvalue lambda:

Consistency Ratio 0,37 GCI: 0,22 CR: 5,9%

Matrix

Expe

cted

D

eman

d

MO

Q/D

eman

d

S_Pr

ice

ITR

Lead

Tim

e

Lead

Tim

e D

iffer

ence

Crit

erio

n 7

Crit

erio

n 8

0 0

1 2 3 4 5 6 7 8 9 10Expected Demand

1 - 4 1/2 3 4 5 - - - - 29,32%MOQ/Deman

d2 1/4 - 1/2 1 2 2 - - - - 11,34%

S_Price 3 2 2 - 3 5 4 - - - - 32,87%

ITR 4 1/3 1 1/3 - 7 3 - - - - 15,75%

Lead Time 5 1/4 1/2 1/5 1/7 - 1 - - - - 5,11%Lead Time Difference

6 1/5 1/2 1/4 1/3 1 - - - - - 5,61%

Criterion 7 7 - - - - - - - - - - 0,00%

Criterion 8 8 - - - - - - - - - - 0,00%

0 9 - - - - - - - - - - 0,00%

0 10 - - - - - - - - - - 0,00%

normalized principal

Eigenvector

6

0

1

Determining the weight of part criteria that is related to supply chain goal "Reduce Cost"

4-Apr-17

Valentin

6,372

Criterion

Consolidated

Expected DemandMOQ/DemandS_PriceITRLead TimeLead Time DifferenceCriterion 7Criterion 8

for 9&10 unprotect the input sheets and expand the question section ("+" in row 66)

Appendix

70

AHP Criteria related to "Reduce Downtime" (EVM multiple inputs)K. D. Goepel Version 04.05.2016 Free web based AHP software on: http://bpmsg.com

Only input data in the light green fields and worksheets!

n= Number of criteria (2 to 10) Scale: 1 AHP 1-9

N= Number of Participants (1 to 20) α : 0,1 Consensus: n/a

p= selected Participant (0=consol.) 2 7

Objective

Author

Date Thresh: 1E-07 Iterations: 8 EVM check: 2,0E-08

Table Comment Weights Rk1 32,7% 22 26,0% 33 41,3% 14 0,0%5 0,0%6 0,0%7 0,0%8 0,0%9 0,0%

10 0,0%

Result Eigenvalue lambda:

Consistency Ratio 0,37 GCI: 0,16 CR: 5,6%

Matrix

Expe

cted

D

eman

d

Lead

Tim

e

Lead

Tim

e D

iffer

ence

Crit

erio

n 4

Crit

erio

n 5

Crit

erio

n 6

Crit

erio

n 7

Crit

erio

n 8

0 0

1 2 3 4 5 6 7 8 9 10Expected Demand

1 - 1 1 - - - - - - - 32,75%

Lead Time 2 1 - 1/2 - - - - - - - 25,99%Lead Time Difference

3 1 2 - - - - - - - - 41,26%

Criterion 4 4 - - - - - - - - - - 0,00%

Criterion 5 5 - - - - - - - - - - 0,00%

Criterion 6 6 - - - - - - - - - - 0,00%

Criterion 7 7 - - - - - - - - - - 0,00%

Criterion 8 8 - - - - - - - - - - 0,00%

0 9 - - - - - - - - - - 0,00%

0 10 - - - - - - - - - - 0,00%

Criterion 7Criterion 8

for 9&10 unprotect the input sheets and expand the question section ("+" in row 66)

normalized principal

Eigenvector

3

0

1

Determining the weight of part criteria that is related to supply chain goal "Reduce Downtime"

4-Apr-17

Valentin

3,054

Criterion

Consolidated

Expected DemandLead TimeLead Time DifferenceCriterion 4Criterion 5Criterion 6

Appendix

71

9.5 APPENDIX 5

Figure 44: Comparison AM Cost for FDM Technology in-house and current purchasing

costs

Pa

rtE

xpe

cte

d

De

ma

nd

Cu

rre

nt

MO

Q#

Fu

ll p

latf

orm

Pa

rt

Vo

lum

e

[cm

3]

Fo

otp

rin

t [c

m2

]M

an

ufa

c

Tim

e [

h]

Ma

teri

al

Co

st [

€]

La

bo

ur

Co

st

[€]

Ov

erh

ea

d C

ost

s (D

ep

rec

iati

on

+

Ma

inte

na

nc

e)

[€]

Ca

lcu

cla

ted

A

M

Ma

nu

fac

C

ost

[€

]

Cu

rre

nt

Pu

rch

Co

st

[€]

An

nu

al

Lo

ss/W

in

152

734

90,9

139,

757,

330

,01,

524

,055

,518

,219

37,7

24

1212

79,4

816,

426

,20,

920

,948

,023

,099

,9

353

33

387,

729

4,98

31,0

127,

93,

110

2,2

233,

217

5,6

3053

,6

410

33

923,

028

0,85

73,8

304,

63,

024

3,2

550,

824

1,4

3093

,4

513

78

324,

413

1,67

26,0

107,

01,

485

,519

3,9

74,6

1550

,7

66

89

364,

712

2,5

29,2

120,

31,

396

,121

7,7

65,8

911,

3

711

1212

229,

981

18,4

75,9

0,9

60,6

137,

331

,911

59,2

82

109

72,6

102,

085,

823

,91,

119

,144

,110

,168

,1

920

8271

7,0

12,3

750,

62,

30,

11,

84,

325

,5-4

23,6

108

8271

7,0

12,3

750,

62,

30,

11,

84,

325

,5-1

69,5

111

56

406,

017

6,9

32,5

134,

01,

910

7,0

242,

855

,318

7,5

125

54

398,

017

431

,813

1,3

1,8

104,

923

8,1

30,2

1039

,6

1341

2830

53,2

364,

317

,60,

414

,032

,019

,650

5,6

1441

6356

30,3

162,

410

,00,

28,

018

,117

,622

,2

1518

4435

17,0

22,7

51,

45,

60,

24,

510

,314

,8-8

0,5

169

2042

58,6

504,

719

,30,

515

,435

,315

,817

5,2

67

3,6

Po

ten

tia

l S

av

ing

s

Appendix

72

9.6 APPENDIX 6

Figure 45 Manufacturing cost comparison between computed AM cost and provided

quote costs for FDM technology, using same material costs

Pa

rtE

xpe

cte

d

De

ma

nd

MO

Q#

Fu

ll p

latf

orm

Pa

rt

Vo

lum

e

[cm

3]

Fo

otp

rin

t [c

m2

]M

an

ufa

c

Tim

e [

h]

Ma

teri

al

Co

st [

€]

La

bo

ur

Co

st

[€]

Ov

erh

ea

d C

ost

s (D

ep

rec

iati

on

+M

ain

ten

an

ce

) [€

]

Ma

nu

fac

turi

ng

C

ost

, p

rov

ide

d

Qu

ote

[€

]

Ca

lcu

late

d

Ma

nu

fac

turi

ng

C

ost

wit

ho

ut

Ov

erh

ea

d C

ost

s [€

]

152

734

90,9

139,

757,

320

,01,

524

,022

,221

,5

24

1212

79,4

816,

417

,50,

920

,924

,618

,3

353

33

387,

729

4,98

31,0

85,3

3,1

102,

210

2,3

88,4

410

33

923,

028

0,85

73,8

203,

13,

024

3,2

226,

720

6,0

513

78

324,

413

1,67

26,0

71,4

1,4

85,5

77,8

72,8

66

89

364,

712

2,5

29,2

80,2

1,3

96,1

86,4

81,5

711

1212

229,

981

18,4

50,6

0,9

60,6

58,3

51,4

82

109

72,6

102,

085,

816

,01,

119

,122

,417

,0

920

8271

7,0

12,3

750,

61,

50,

11,

81,

61,

7

108

8271

7,0

12,3

750,

61,

50,

11,

81,

81,

7

111

56

406,

017

6,9

32,5

89,3

1,9

107,

095

,091

,2

125

54

398,

017

431

,887

,61,

810

4,9

94,5

89,4

1341

2830

53,2

364,

311

,70,

414

,013

,512

,1

1441

6356

30,3

162,

46,

70,

28,

07,

26,

8

1518

4435

17,0

22,7

51,

43,

70,

24,

54,

94,

0

169

2042

58,6

504,

712

,90,

515

,414

,713

,4

Appendix

73

9.7 APPENDIX 7

Figure 46 Purchase cost comparison between SLS AM and conventional technology

Part Expec ted Demand

DE

MOQ Part Volume [c m3]

Height [mm]

Purc h Cost

(Quote Materialize

Current Purc h

Cost [€]

Loss/ W in Lead time [days]

1 52 15 90,9 65 58,1 18,2 2075,6 322 4 13 79,4 75 51,3 23,0 113,2 263 53 22 387,7 45 142,8 175,6 -1739,9 314 10 10 923,0 100 212,1 241,4 -293,7 315 13 22 324,4 45 75,0 74,6 4,8 316 6 22 364,7 45 69,2 65,8 20,1 317 11 9 229,9 110 87,8 31,9 614,2 318 2 22 72,6 45 42,8 10,1 65,4 319 20 85 7,0 12 11,6 25,5 -278,0 26

10 8 85 7,0 12 11,6 25,5 -111,2 2611 1 28 406,0 36 77,5 55,3 22,2 2312 5 34 398,0 30 77,2 30,2 235,2 2313 41 17 53,2 60 20,7 19,6 42,4 2314 41 22 30,3 45 12,7 17,6 -200,5 2315 18 53 17,0 19 12,4 14,8 -43,4 2316 9 20 58,6 50 16,7 15,8 8,0 23

2666,7Po te ntia l Sa v ing s

PLASTIC ADDITIVE MANUFACTURING AND ITS POTENTIAL …1198354/... · 2018-04-17 · CAD Computer Aided Design CLAD Construction Laser Additive Directe CJP Color Jet ... manufacturing

Documents