1
Digital Twins and the Maturing of New
Manufacturing Technology
A Case Study in the Aerospace Industry
Felix Lebbink S-2605279
January 2020 Word count: 11.124
Supervisor / University of Groningen (RUG) Prof. dr. ir. J.C. Wortmann / dr. J.A.C Bokhorst
dr. N.B. Szirbik (second supervisor)
Co-assessor / University of Groningen (RUG) S. Waschull (PhD candidate)
Supervisor / GKN Fokker Aerostructures S. Hengeveld (Project Leader Fokker 4.0)
2
Abstract
The integration of the physical and digital world in industry is recognized to be the essence of the fourth industrial revolution, more commonly referred to as Industry 4.0. Digital Twins have been identified as a key component of Industry 4.0 and can be explained as the real-time digital replica of a physical entity. Synchronizing the physical world with a digital replica, allows for different purposes, such as analysis of the physical entity or simulations of its behaviour. Current literature on Digital Twins extensively describe the potential contributions of Digital Twins to the manufacturing industry, however lack in describing the relationship to new manufacturing technology, which is still in development. This paper adds to the existing body of knowledge that Digital Twins promise opportunities to enhance the research and development of new manufacturing technology and speed up the maturing process. The results of this single-case research, conducted at a company active in the aerospace industry show that the potential contribution of Digital Twins increases as the technology matures, measured according to the Technology Readiness Levels (TRL).
3
Table of content
Abstract ... 2 Table of content ... 3 Preface ... 5 I. Introduction ... 6 II. Theory ... 8 Digital Twin ... 8 Digital Model: ... 9 Digital Shadow: ... 9 Digital Twin: ... 10Level 1 Pre-Digital Twin ... 12
Level 2 Digital Twin ... 12
Level 3 Adaptive Digital Twin ... 12
Level 4 Intelligent Digital Twin ... 12
Applications ... 13
Digital Twin model ... 14
Development of a Digital Twin ... 14
Maturity of new manufacturing technology ... 16
Tools to measure the maturity of new manufacturing technologies ... 17
Expected relation between Digital Twins and the maturing of new manufacturing technology ... 19
III Method ... 21
Research design and case selection ... 21
Case Description ... 21
Data collection ... 22
Data analysis ... 23
IV Results ... 24
Situation as-is ... 24
Current working methods and procedures ... 26
Problems within the current working methods and procedures ... 27
Digital Twin ... 27
Digital Twins and the Technology Readiness Levels... 28
TRL 1 and 2 ... 28
TRL 3 and 4 ... 28
4
TRL 7 – 9 ... 28
V Discussion ... 29
Answering the research question... 29
Limitations ... 30 Theoretical implications ... 30 Practical implications ... 31 VI. Conclusion ... 32 VII References ... 33 VIII Appendices ... 36
Appendix 1 Conducting a systematic literature review of Digital Twins ... 36
Step 1 Planning and formulating the problem (definition of the scope) ... 38
Step 2 Searching the literature ... 38
Step 3 and 4 Data gathering and Quality evaluation ... 39
Step 5 and 6 Data analysis and synthesis and Interpretation ... 39
Step 7 Presenting results ... 39
Appendix B Conducting Case Study ... 40
5
Preface
6
I.
Introduction
For the first time in history, we are aware that we are in the middle of a new industrial revolution, more commonly referred to as ‘Industry 4.0’. Contrary to the previous three industrial revolutions, this one is predicted a-priori, rather than observed ex-post (Drath & Horch, 2014), making it an extraordinarily interesting phenomenon. Industry 4.0 can be characterized by the introduction of the Internet of Things and Services into the manufacturing environment (Kagermann, Wahlster, & Helbig, 2013). Amongst other aspects, it requires modelling and simulation of manufacturing systems and the use of advanced artificial intelligence for process control, which includes autonomous adjustment to the operation system. This is a new simulation modelling paradigm and can best be described by the concept ‘Digital Twin’ (Rodič, 2017).
A Digital Twin is a real-time digital replica of a physical entity and forms the knowledge base for various applications. This digital replica of a physical system (product, process or a subset of this) can be used for different purposes, such as analysis of the physical entity or simulations of its behaviour. This synchronization between the virtual and the physical is possible due to the underlying enabling technologies of Industry 4.0 (Negri, Fumagalli, & Macchi, 2017). The purposes associated with Digital Twins include, but are not limited to, reflecting the structure, performance and health status of the physical counterpart, as well as determining when to schedule preventive maintenance, promoting traceability and optimizing service and manufacturing processes (Madni, Madni, & Lucero, 2019). For example, Goossens (2017) describes how a Digital Twin can be used for training staff on virtual machines before operating them or that a Digital Twin can optimize its own performance for given duty cycles (Goossens, 2017).
However, contrary to the purposes of Digital Twins described above, which cover already existing manufacturing processes, there might be a contribution to new manufacturing technology which is still in development as well. Digital Twin technology has the potential to reduce the cost of system verification and testing while providing early insights into system behaviour (Madni et al., 2019), implying that there is a potential contribution to the maturing of a new manufacturing technology. This is an underlit purpose of Digital Twins in the current literature and to the best of our knowledge, no prior research has investigated this relationship.
To begin, one must first familiarize oneself with the concepts of ‘Digital Twins’ and ‘maturing’ of new manufacturing technology. Digital Twins have been identified as a key component of Industry 4.0. The technological basis of Industry 4.0 roots back in the Internet of Things (Negri et al., 2017), and includes the application of Cyber-Physical Systems (CPSs) to industrial production systems (Drath & Horch, 2014). To clarify; through the Internet of Things, all components in a production system can be connected which allows for collection and exchange of data (Negri et al., 2017). Such a system is referred to as a CPS. As the name suggests a CPS is a systems in which the software (cyber) and the physical (physical) components are deeply intertwined (Xiong et al., 2015). A Digital Twin is the digital replica of the physical system.
7
Innovation driven entrepreneurship leads to the development of new manufacturing technology. Innovations in the field of manufacturing technology or “premature manufacturing technologies” (e.g. a new production machine) may promise high potential in terms of product design or cost reduction (S. Peters, 2015). Nevertheless, the process of new manufacturing technology development, starting from idea and ending in the implementation and commissioning of a stable process, can be time-consuming and costly. The ‘maturity’ of new manufacturing technology provides an indication of how far up this process the new manufacturing technology currently is.
To assess the maturity of new manufacturing technology, different readiness level scales have been developed. This research uses the Technology Readiness Levels (TRL) scale, which was developed by the National Aeronautics and Space Administration (NASA) to support the technology planning of space programmes (Flinn, 2019). The TRL consists of nine levels which new technology progresses through as it matures. The primary purpose of the TRL is to aid management in their decision-making process concerning the new manufacturing technology by providing knowledge on its maturity, since it can be extremely costly to implement a new technology too soon or too late (S. Peters, 2015).
Intuitively speaking, manufacturers would prefer to bring new manufacturing technology, which potentially enhances the manufacturing process, to the highest readiness level in the shortest amount of time. As Madni et al. (2019) argue, Digital Twin technology has the potential to reduce the cost of system verification and testing, which leads to the expectation that the use of Digital Twins speeds up the maturing of new manufacturing technology. Therefore, this research aims to find out the relation between these two concepts. The research question can be formulated as:
HOW CAN DIGITAL TWINS OF MANUFACTURING PROCESSES CONTRIBUTE TO THE MATURING OF NEW MANUFACTURING TECHNOLOGY?
To answer this question, a case research has been conducted of a company operating in the aerospace industry. The case research begins with an extensive literature review of all current knowledge in the field, which is described in the theory section. A thorough, sophisticated literature review is the foundation and inspiration for substantial, useful research (Boote & Beile, 2005). Thereafter, the methodology adhered to during the case research is described, subsequently followed by the results, the discussion and the conclusion.
This research is conducted in a case setting of the aerospace industry since the aerospace industry is one of the forerunners when it comes to experimenting with Digital Twins (Tuegel, Ingraffea, Eason, & Spottswood, 2011). The company in the case is a Dutch settled manufacturer of aerostructures, which is in the process of creating a Digital Twin of a new manufacturing technology, making it the ideal environment to conduct this study.
The paper is structured in the same sequential order as described above, meaning that the next chapter elaborates on the theoretical underpinning of this research. Chapter three provides a description of the methodology adhered to during the research. The fourth chapter elaborates on the results obtained during the research and in the fifth chapter the results are discussed. Finally, the conclusions are provided in chapter six as well as both the theoretical and managerial implications.
8
II.
Theory
This chapter elaborates on the theoretical underpinning of the key aspects of this research. An extensive literature review of the current status quo is paramount for performing research. Knowledge is built upon the foundation of prior knowledge and conflicting use of terms and definitions should therefore be avoided, and newly attained knowledge should be complementary to predecessor knowledge.
There are two key aspects concerning this research that are elaborated in more detail in this section1. The first is the Digital Twin. Although the concept is relatively novel, and current literature is
somewhat contradictory and ambiguous in the use of the term, this research endeavours to give a description and place it in the correct context. Secondly, the ‘maturity’ of new manufacturing technology is discussed, as the focus of the research is how Digital Twins may contribute to it.
Digital Twin
A Digital Twin is a real-time digital replica of a physical entity. The term was first introduced and later elaborated by Professor Michael Grieves in Product Lifecycle Management (PLM) classes (Guo, Zhao, Sun, & Zhang, 2019). As mentioned before, partly due to the novelty of the concept, current literature is somewhat ambiguous in the use of the term Digital Twin. Some definitions limit the Digital Twin to being the virtual replica of a product (Grieves & Vickers, 2016), whilst others include systems (Söderberg, Wärmefjord, Carlson, & Lindkvist, 2017) or processes (Liu, Zhou, Tian, Liu, & Jing, 2019) as well. Moreover, other definitions describe Digital Twins as ‘the digital replica of a living or non-living physical entity’ (El Saddik, 2018), basically restraining any limitation to the definition of Digital Twin. Kritzinger et al. (2018) conducted a categorial literature review and classification on Digital Twins in manufacturing and concluded that there is no common definition of the Digital Twin and that the development of the Digital Twin is still at its infancy. Therefore, it is especially important to extensively describe how a research, article or thesis perceives Digital Twins and in which context it is placed.
Since this research focuses on examining the role of Digital Twins and their contribution to the maturity of new manufacturing technology, the concept of the Digital Twin should be placed in the context of manufacturing. Therefore, the following definition is proposed:
The Digital Twin is a virtual representation of a production system that runs on different simulation disciplines, characterized by the synchronization between the virtual and real system. With the help of mathematical models and real time data elaboration it becomes able to forecast and optimize the behaviour of the production system at each life cycle phase (Kritzinger, Karner, Traar,
Henjes, & Sihn, 2018, p. 1017).
This definition includes some important aspects relevant in this research. Firstly, the physical twin in this definition constitutes a production system, which is exactly what manufacturing technology is intended for; production. Secondly, the different life cycle phases are explicitly mentioned. As will be elaborated later in the chapter ‘Maturity of new manufacturing technology’, the scope of this research concerns the phase of new manufacturing technology from concept generation and formulation to industrialisation and operationalisation, which can be regarded as the very beginning of the lifecycle.
9
Nevertheless, due to the ambiguous use of the term in the current literature it may be beneficiary to go into more detail and elaborate further. When is a virtual model actually a Digital Twin and not just a virtual model? Kritzinger et al. (2018) propose to categorize Digital Twins into three subcategories, namely: ‘Digital Model’ (DM), ‘Digital Shadow’ (DS) and ‘Digital Twin’ (DT). Although these terms may be used synonymously in existing literature, a classification can be made based on the level of data integration.
Digital Model:
The Digital Model is the subcategory with the lowest level of data integration. It comprises the digital representation of an existing or planned physical object but does not use any form of automated data exchange between the physical and virtual objects. The digital representation might include a comprehensive description of the physical object and the models might include, but are not limited to, simulation or mathematical models. All data exchange is done manually, meaning that a change in state of the virtual model does not directly affect the physical object and a change in state of the physical object does not directly affect the virtual model.
Figure 1 Data Flow in a Digital Model
Digital Shadow:
A Digital Shadow is an extension of the definition of a Digital Model in the sense that there is a one-way automated data flow between the physical object and the virtual model. This would mean that a change of state of the physical object directly affects its virtual counterpart, but not vice versa.
10
Digital Twin:
The Digital Twin is the subcategory with the highest level of data integration. As one might expect, it is the extension of the definition of a Digital Shadow in the sense that there is a two-way automated data flow between both counterparts. The digital object might also induce changes of state of the physical object.
Figure 3 Data Flow in a Digital Twin
These subcategories make a distinction based on whether the data flow between the physical and digital object is automated or not. This should not, however, be confused with autonomy. Automated systems execute fixed engineered actions, whilst autonomous systems understand their tasks based on explicitly represented knowledge. Autonomous systems are able to execute high-level task specifications without explicitly being programmed for the specific task (Rosen, Von Wichert, Lo, & Bettenhausen, 2015).
11
Figure 4 Overview of a Digital Twin (Stark, Kind, & Neumeyer, 2017)
Comparing the above two definitions, we can already find a discrepancy. The first describes the Digital Shadow as a digital model of a physical asset with one-way automated flow of data and the latter uses the term Digital Shadow to describe the accumulated data of a physical asset over its lifetime. These are two entirely different definitions for the same term. Henceforth, the definition of Stark et al. (2017) is used to describe the ‘Digital Shadow’. For a Digital Twin to be able to analyse, simulate and optimize its own behaviour, there needs to be historic data of the physical asset. This lacks in the model provided by Kritzinger et al. (2018), which in comparison is more simplistic.
Additionally, as Madni et al. (2019) describe, there can be different levels of maturity or sophistication of the Digital Twin. This classification emphasises on the sophistication of the virtual model and includes a phase in which it may be possible that there is no Physical Twin yet, which may be the case in the maturing of new manufacturing technology. The following table shows these four levels – ‘Pre-Digital Twin’, ‘Digital Twin’, ‘Adaptive Digital Twin’ and ‘Intelligent Digital Twin’ – and their corresponding characteristics:
12
Level 1 Pre-Digital Twin
The first level of maturity of the Digital Twin is called the ‘Pre-Digital Twin’ and encompasses the traditional virtual prototype created during upfront engineering. It is a virtual prototype of the envisioned system that is created before the physical prototype is built, meaning that it does not refer to already existing physical models. The purpose of the Pre-Digital Twin is to identify and mitigate technical risks early on in the design process (Madni et al., 2019).
Level 2 Digital Twin
The level 2 Digital Twin describes a virtual system model which is already capable of incorporating performance, health and maintenance data from the physical counterpart. This data is sent to the virtual system model in batch updates. Thereafter, this data is used to support high-level decision making and feedback control actions can be induced. The purpose of the DT at this level is to explore the behaviour of the physical twin under different what-if scenarios (Madni et al., 2019).
Level 3 Adaptive Digital Twin
The third level of sophistication is called the ‘Adaptive Digital Twin’ and is an extension of the description of the level 2 DT in the sense that it adds the capability of learning the preferences and priorities of the user/operator in different contexts. The use of neural network-based supervised machine learning algorithms should facilitate this (Madni et al., 2019).
Level 4 Intelligent Digital Twin
13
Applications
Contrary to the exact definition of a Digital Twin, most literature does agree that there is a multitude of actual applications it may serve. Creating Digital Twins enables the efficient execution of product and process designing, manufacturing, servicing, and various other activities throughout the product life cycle (Schleich, Anwer, Mathieu, & Wartzack, 2017). Adding to that, a Digital Twin in a manufacturing context offers the opportunity to simulate and optimize the production system, including its logistics, whilst also enabling a detailed visualization of the manufacturing process from single component up to the whole assembly (Kritzinger et al., 2018).
In the context of production systems in manufacturing, Kritzinger et al. (2018) found that Digital Twins contribute in the following main disciplines:
• Production planning and control (Rosen et al., 2015) o Orders planning based on statistical assumptions
o Improved decision support by means of detailed diagnosis
o Automatic planning and execution from orders by the production units
• Maintenance (D’Addona, Ullah, & Matarazzo, 2017; Susto, Schirru, Pampuri, McLoone, & Beghi, 2015)
o Identify the impact of state changes on upstream and downstream processes of a production system
o Identification and evaluation of anticipatory maintenance measures
o Evaluation of machine conditions based on descriptive methods and machine learning algorithms
o Integrate, manage and analyse machinery or process data during different stages of machine life cycle to handle data/information more efficiently and further achieve better transparency of a machine’s health condition (Lee, Lapira, Bagheri, & Kao, 2013)
• Layout planning (Uhlemann, Lehmann, & Steinhilper, 2017)
o Continuous production system evaluation and planning
o Automatic and application independent data acquisition and variation
Apart from Production planning and control, Maintenance and Layout planning, other applications of the Digital Twin can be found as well. These include increasing the efficiency of product design and manufacturing implementation (Tao et al., 2018), enabling traceability and diagnostics (Cai, Starly, Cohen, & Lee, 2017), and technical risk mitigation (Madni et al., 2019).
14
Digital Twin model
Taking all of the above views on the definition of Digital Twins into consideration, the following model is proposed:
Figure 5 Digital Twin model
This model proposes that the Digital Twin consists of the Digital Shadow, which is fed with data from the Physical Twin, the Digital Master and intelligent linking between them. The intelligent linking consists of data analytics, simulation and high-level decision making (e.g. artificial intelligence). The feedback loop to the Physical Twin can either be direct, or with the interference of an operator. In this model, there is also a presence of a Physical Twin, which is not an integral component of the Digital Twin but must be present in order for the Digital Twin to exist. As the name ‘twin’ already suggests that there must be a unique, but inextricably counterpart.
Development of a Digital Twin
15
Figure 6 The “Digital Twin 8-dimension model” (Stark, Fresemann, & Lindow, 2019)
In this model a distinction is made between the Digital Twin context and environment (left side) and behaviour and capability (right side). Each dimension in this model provides a number of levels of realization. A higher level of realization does not necessarily depict a better DT, merely a different and/or unique realization space. Additionally, the eight dimensions in this model are not exclusive or exhaustive. They represent the most likely areas of importance to support the individual context situations of the specific DT in scope (Stark et al., 2019).
16
Figure 7 Digital Twin design elements (Stark et al., 2019)
Maturity of new manufacturing technology
Technology has been around longer than the Homo Sapiens. The oldest, deliberately made tools, date back two and a halve million years and were found in the Omo Valley of Ethiopia (Headrick, 2009). It is likely to assume that the sharpened cobbles, used for chopping wood and breaking bones, referred to in this example do not uphold to today’s standards of what we assume to be ‘technology’. The word itself is derived from two Greek words, being techne and logia. Which can be translated as the subject of study or interest in; means, art, skill craft or the way/manner. Today’s definition of technology has slightly shifted and refers to the application of scientific knowledge for practical (industrial) purposes (Oxford Lexico, 2019).
Organizations invest billions of dollars annually in the development of technologies and engineering products (W. Peters, Doskey, & Moreland, 2017). New technology is in many cases a key driver for product innovation, cost-reduction, or for fulfilment of customer demands concerning quality or sustainability (Schuh, Schubert, & Wellensiek, 2012). Nevertheless, the development of new technologies can be time-consuming and costly. Starting with an idea or the identification of an opportunity, it could take many years before the technology is developed, validated and implemented. There are many phases that can be distinguished in this process and in between these phases there are often decision points that determine whether to advance to the subsequent phase or not. See for example Figure 8, from the United States Government Accountability Office:
17
This example illustrates how technology development starts with formulation and after a series of different phases and key decision points reaches implementation. Although this an example from the United States Government Accountability Office, the basic outlines are presumably similar for different companies.
When zooming in, and specifically focusing on technology in the manufacturing context, we define manufacturing technology. That is; technology with the specific purpose of being used for production (S. Peters, 2015). In this context, technology development, too, passes different phases before it is ready to be implemented. Commonly, the term ‘maturity’ is used as a measure to describe how close new manufacturing technology is to implementation and operation. New manufacturing technology is assumed to be ‘mature’ when it is implemented and behaves as intended in practice2.
The criteria that need to be met in order for manufacturing technology to be mature depend on the intended purpose of said technology. A mature technology may sometimes also be referred to as ‘stable’ (S. Peters, 2015), meaning that the process, for which the manufacturing technology is intended, yields stable and predictable output. After implementation, the lifecycle of the technology (TLC) continues until finally it becomes obsolete and new technology replaces its necessity. Technology progresses slowly at first, then accelerates and then inevitably declines (Foster, 1986). A very generic graph of the business gain following the TLC is depicted in Figure 9.
Figure 9 Business Gain during the Technology Life Cycle (Park, Sung, & Kim, 2015)
In the early phases of the technology lifecycle, the stage in which the research and development (R&D) takes place, the business gain is negative. First, investments need to be made in order to develop and mature the new manufacturing technology. Only when the manufacturing technology is mature and ready to be commissioned in practice, may it generate revenue.
Tools to measure the maturity of new manufacturing technologies
Determining when manufacturing technology is actually stable or mature is often difficult due to the fact that the technology usually does not act individually but is part of a manufacturing process chain, thus affecting the whole chain (Schuh et al., 2012). Since assessing the maturity of new manufacturing technology is crucial for the organization’s ability to manage performance, cost and schedule (W. Peters et al., 2017), several tools have been developed over the past decades. Usually these tools consist of a scale-based measure. Most of these tools are derived from the Technology Readiness Level
18
(TRL) originally developed by the National Aeronautics and Space Administration in the 1970s (see Table 2). This scale consists of nine levels, with TRL 1 being the lowest and TRL 9 the highest. Each of these nine levels describes a level of maturity of a technology. By utilising maturity scales such as the TRL, the progression of technology development becomes quantifiable and thus the speed of maturing becomes measurable.
Table 2 NASA Technology Readiness Levels
19
Expected relation between Digital Twins and the maturing of new manufacturing
technology
The Digital Twin is currently a popular topic in industry. The expectations concerning the Digital Twin are high and many tech-companies and consultancy businesses advertise and parade with it. See for example: ROI-International3, Prespective4, Deloitte5 and PwC6. A simple web search provides
commercial examples of how Digital Twins positively contributed in the maturing of new technology. For example, Knorr-Bremse claim that with the help of Digital Twins there was a 30% reduction of resource-consuming hardware tests of railway braking systems (Bouskela, 2019). Another example is that of Mechanical Solution Inc. (MSI), which is specialized in analysing and testing rotating and reciprocating turbomachines. MSI claims that Digital Twins helped them predict performance earlier in the design cycle, analyse multiple designs, reduce reliance on prototypes and expensive testing and shorten design time and cost (Benett & Ivashchenko, 2017).
It seems straightforward that there is a positive relationship between the use of Digital Twins and the maturing of new manufacturing technology. However, as described above, there is no common definition of the Digital Twin, and different views on the subject exist. It can be questioned whether the examples indeed utilised Digital Twins.
As for previous technological advancements, the expectations are often high and far ahead of the actual technology in the beginning. Renowned global research, advisory and information technology firm Gartner even developed the technology hype cycle which shows how expectations usually grow extremely fast in the beginning of technology development, but will decrease as time passes, until people are through of disillusionment (see Figure 10).
20
Figure 10 The technology hype cycle by Gartner
Keeping in mind that Digital Twin technology, as it is still in its infancy, may currently find itself in the peak of inflated expectations, it is imperative to remain highly critical in assessing the expected relationship between Digital Twins and the maturing of new manufacturing technology, and not just simply follow the hype. What can Digital Twin technology offer which, for example, simulation tools cannot in the maturing of new manufacturing technology?
Combining the theory of Tao et al. (2018), which mention that the innovation and efficiency from product design, production planning and manufacturing implementation can be effectively enhanced by Digital Twins, and Madni et al. (2019), which describes the ‘Pre-Digital Twin’, a phase in which there is no existence of a Physical Twin, it is expected that there is a positive relationship between Digital Twins and the maturing of new manufacturing technology.
21
III
Method
Research design and case selection
To be able to investigate how Digital Twins contribute to the maturing of new manufacturing technology it is evident that two factors are of importance: the use of Digital Twins and a manufacturing setting. The combination of needing to answer a ‘how’ question, as well as the specific context in which this question can be researched led to the methodology choice of a case study. The case study embodies a strategy which focuses on understanding the dynamics within a single setting (Eisenhardt, 1989). In her article “Building Theories from Case Study Research”, Eisenhardt (1989) provides a comprehensive roadmap for executing case studies which can be found in Appendix B. This framework forms the guidance for the methodology adhered to in this study.
Operations Management is a very dynamic field in which new practices are continually emerging (Voss, Tsikriktsis, & Frohlich, 2002). The Digital Twin can be identified as one of such practices as the Digital Twin is still in its infancy and only just begins emerging in industry, with different definitions and applications. In the design of a case study a decision needs to be made on whether the purpose is either descriptive, exploratory, or explanatory (Salkind, 2010). Employing an exploratory conduct seems to fit best in this context. The goal of the exploratory case study is to develop working hypotheses and perhaps propose further research (Salkind, 2010).
Case Description
After the initial ‘getting started’ step of the case study framework, the next step embodies selecting the case. This study is conducted at a Dutch settled manufacturer of aerostructures. The company is part of a larger multinational concern, which operates globally in 15 different countries with 18.000 employees. For this research a plant in the Netherlands was selected at which over 1.150 employees are working.
At one of the facilities on this site, composite structures for the aerospace industry are built. Composites are materials made from two or more different materials. Together, these materials have different physical and chemical properties than individually (Elhajar, La Saponara, & Muliana, 2017). In the aerospace industry, composites are especially interesting as they can be used to produce strong but extremely lightweight structures. The aerospace industry is prepared to invest largely in composites as they lead to lightweight aircrafts that use less fuel, but can still withstand the forces inflicted on them (Marsh, 2012).
22
Data collection
Case study research is sometimes referred to as a versatile approach to research (Salkind, 2010) since there are multiple manners in which data can be collected. Common sources of data in case research are (semi-)structured interviews, interactions, observations and archival sources (Voss et al., 2002). To get a thorough understanding and stimulate objectivity, the use of triangulation can help get the most accurate picture in case research (McCutcheon & Meredith, 1993). Triangulation means using more than one method to collect data. Since the aerospace is a highly regulated industry and every process and protocol is documented, the use of archival records seemed to be an obvious and reliable source of data. Secondly, interviews were conducted to collect other (undocumented) pieces of information and validate the archival records.
For the archival records, the company’s digital database was used. For the aerospace industry it is common that all processes, procedures and actions are highly regulated. The company keeps an internal database which all employees can access to view these procedures. From this database, the procedure of technology development within the company was retrieved. This document describes the process of assessing and documenting a technology’s level of maturity and which parties are involved and responsible in this process.
After checking and analysing this internal document and speaking to project management, it was decided that there are two focus groups responsible within Fokker for the maturing of new manufacturing technology, which can be interviewed. The first of these is the Research and Development department and the second is the Industrialisation department. Therefore, interviews were conducted with these two groups to get a clear view on the current procedures in the maturing of new manufacturing technology and explore whether Digital Twin technology could be of aid. With each of these two groups, one interview was conducted.
The interview with the Research and Development department was conducted with two interviewees. One of the interviewees was an R&D engineer and the other interviewee was the head of the R&D department of the composite division. The questions were directed at both of them and the answers they gave were agreed upon by both interviewees. This gave a complete picture as persons from different echelons constituted one answer. The interview with the Industrialization department was conducted with one interviewee. The interviewee is an Industrialization Manager, who has worked for over 18 years at the company and for six years is the Industrialization Manager.
The interviews that were conducted were exploratory in nature since the interviewees had no prior knowledge of Digital Twins, and its potential applications. The predominant goal of these interviews was to discover what their functions entail, what role the interviewees have in the maturing of new manufacturing technology, what is currently already being done in the field of digitization and where they believe is room for improvement in the maturing process of new manufacturing technology.
23
Data analysis
For case research it is not uncommon that data collection and data analysis occur concurrently. This speeds up the analyses and reveal helpful adjustments to data collection (Eisenhardt, 1989). Thanks to interactions and conversations with employees at the company, the existence of the database with all internal protocols and procedures was revealed. From this internal database, the protocol of the technology progression through the TRL maturing process was retrieved. This internal document was used to develop a flowchart which illustrates the process. Developing the flowchart helps fathom the process and gain a clear understanding of how the company handles the maturing of new manufacturing technology.
Thereafter, the semi-structured interviews were conducted with the two focus groups. To analyse the interviews, they were audio-recorded, so during the interviews the interviewer could focus on the answers and if necessary, go more in-depth during the interview. From the audio-recordings the interviews were first transcribed completely. Since the interviews were conducted in Dutch, the transcriptions were translated to English and subsequently analysed according to the method described by Burnard (1991) with the help of a coding book (DiCicco-Bloom & Crabtree, 2006). This method describes how first the transcript is read through completely to become immersed in the data, then categories are created for all aspects of the content (open coding), then the categories are surveyed and grouped together under higher order headings (Burnard, 1991). This is an iterative process which is repeated until a final list of categories has emerged.
24
IV
Results
In this section the results of the research will be provided by means of a thorough and extensive description of the current situation of the maturing of new manufacturing technology and the outcome of interviews that were conducted. Firstly, the situation ‘as-is’ will be explained. The situation ‘as-is’ describes how the company currently handles the introduction and maturing of new manufacturing technology without the use of Digital Twins. Thereafter, the results of the interviews will be discussed, according to the themes that emerged from the analysis.
Situation as-is
New manufacturing technology is either developed at the company’s own Research and Development department, acquired from another third party, or developed in collaboration with the third party. The company has defined TRL 6 (see Table 2) as the maturity a technology needs to have in order to be used in production. To assess the maturity of the technology, the company has developed the Technology Readiness Assessment (TRA) procedure. This is a procedure specifically made for all companies under the larger concern to which the company belongs to. The TRA is used for different types of technology (i.e. product, manufacturing, material) and establishes the process for assessing and documenting a technology’s level of maturity relative to its readiness for transition into a product, development process, manufacturing process, repair process, or service. As mentioned above, TRL 6 is defined as the maturity level at which the technology is ready for this transition. Effectively this means that new manufacturing technology can be used for production as soon as it reaches TRL 6. Therefore, the TRA process stretches from TRL 1-6.
Figure 11 illustrates the framework of the TRA. It is an iterative process which begins with defining the scope of each cycle and ends with an action closure plan. Ideally, each cycle ends with the accreditation of a new TRL level. Thus, as technology matures, a series of TRAs are executed.
During this process there are a total of seven parties involved. These are: V.P. Technology, Designated Technologist, Project Lead, Review Chair, Reviewers, TRA Manager and the Chief
Engineers Office. Each of these parties has their own role and responsibility. For this thesis, the most interesting ones are the Designated Technologist and the Project Lead, as they are ultimately
responsible for the progression of the technology through the TRL Maturity Levels.
25
26
a) Pass – TRL Maturity Level has been successfully demonstrated.
b) Action identified – Further evidence is required to demonstrate the targeted TRL Maturity Level is achieved.
c) Hold – Technology has not achieved the target TRL Maturity Level, and there is no current identified route to meet the technological/financial requirements.
d) Stop – Technology has not achieved the targeted TRL Maturity Level, and it is considered that the technology will not be able to meet the technological/financial requirements.
The Project Lead and the Designated Technologist review any recommended actions and produce an action closure plan. Finally, once it is agreed by the Review Chair that all actions are closed, the TRL accreditation is assigned. This is an iterative process which continues until TRL 6 is achieved and all the risks of the technology have been reduced such that the technology can be used on a customer application.
Thus, to recapitulate; the Designated Technologist and the Project Lead need to provide evidence during the TRAM that the new technology has progressed to a new TRL Maturity Level. This evidence is predefined as a list of technological requirements and described in the TDP. This list of technological requirements is completely dependent on the specific technology but can be classified under the top-level definitions of the TRLs as illustrated in Table 2. If the evidence has been provided according to the review panel, a new TRL Maturity Level is assigned.
Next to the TDP, there is also a Technology Business Case (TBC) developed. The TBC contains information on why the technology is developed, and amongst other affairs, all the financial data, such as ‘Post / Pre Technology Cost’, forecast sales volume and non-recurring costs. This is an important document since there should be a clear business case for the company to invest in developing a technology.
Current working methods and procedures
The interviewees from the Research and Development department indicated that they are responsible for the development of new manufacturing technology. They develop technology that the company believes may be viable for future production. First the new technology is developed until the point the company believes it can be used for quotations. When a customer actually places an order, the technology is further developed so it can be used in production. However, there are differences between the requirements a technology needs to meet in the R&D environment and in the production environment. The requirements necessary for the production environment are provided by the industrialization manager. When a technology has proven to meet these requirements, which constitutes to a maturity level of TRL 6, the handover of the technology takes place. Thereafter, the technology is used in a production environment and further matures. For the development of new technology, the department has access to state-of-the-art modelling and simulation software. The software they use is advanced and of the highest quality.
27
using common sense to develop different scenarios and then calculating the costs for each of these scenarios. All information to calculate these different scenarios comes from multiple, different concepts. Sometimes a production line will not contain new manufacturing technology, but sometimes new manufacturing technology is required in the production line (e.g. new machine). When new manufacturing technology is to be included in a production line, it will need to be at least TRL 6. The Industrialization Manager is commonly part of the review panel and provides requirements to the R&D department about the requirements new manufacturing technology needs to meet before it can be implemented in the production environment. These requirements normally include performance measures such as capacity of the machine, deterioration rate, variation etc., as well as stress- and strength reports. When a production line successfully delivers a number of products and is proven to be stable, it is handed over to the production manager and used for (series) production.
Problems within the current working methods and procedures
The interviewees from the Research and Development department indicated that in the development of new technology the software they use is not always able to detect errors when the errors are hardware related. For example, if a drill chuck malfunctions, this will not be detected in the virtual testing of a process. Additionally, since the modelling, virtual testing and simulation of new technology is relatively time-consuming, physical tests are often preferred. Sometimes it takes more time to alter the models and run new simulations than to run physical tests.
The interviewee from the industrialisation department reported that he wishes to be able to make more accurate calculations of the non-recurring costs for the quotations. He mentioned that for the calculations he relies a great deal on the experience he has with former or similar projects. Currently he will devise multiple possible scenarios for a production line (number of machines, number of shifts, number of operators, throughput rate etc.) and the information for these scenarios come from different concepts. Whenever a minor change in one of these concepts occurs, he will need to repeat all the calculations for the scenarios. This is a highly iterative process and the outcome may not be very certain. The uncertainty increases when new technology is to be used.
Digital Twin
When the interviewees from the Research and Design department were asked about how digitization or Digital Twins may aid them in the process of maturing new technology the interviewees responded that they believe that there is no novelty for the design phase of development of new technology, since they already utilise state-of-art software packages. One of the interviewees thought that the Digital Twin is mainly a marketing term used for software applications which already exist. Nevertheless, they do believe that Digital Twins may aid in increasing the learning curve of technology when the technology is put to use (production). Lastly, it was mentioned that it would be interesting for them if Digital Twins could result in improved simulations.
28
Digital Twins and the Technology Readiness Levels
With the help of internal documents, the process of how new technology matures until it is employed for production was found. This process is described above and illustrated in Figure 11. Interviews with parties involved in this process were conducted to gain further knowledge on the current working methods and procedures. It can be concluded that the maturing of new technology, in the case of this research, is very structured and progresses in phases. Nevertheless, it is also very time-consuming: “This is a very time-consuming process and may take multiple years”7. The most time-consuming phase
is in between the TRA cycles. Therefore, to be able to move faster through this progress, potentially with the help of Digital Twins, one would like to meet the requirements in the TDP sooner.
The requirements defined in the TDP are specific to the specific technology, which is in development, however when looking at the top-level definitions of the TRL scale, there may be phases where the use of Digital Twins can speed up the process.
TRL 1 and 2
TRL 1 and 2 entail the observation of a basic principle and the formulation of an application respectively. During these two stages, the focus lies on the conceptualization of an idea and the formulation of what application(s) the concept could potentially be utilised for. During these stages, there seems no evident role for Digital Twins to play a role.
TRL 3 and 4
TRL 3 entails a proof of concept through analysis and experimentation. A Digital Twin of a new manufacturing technology can be used for virtual testing and simulation. According to Madnit et al. (2019) the ’Pre-Digital Twin’ (level 1) does not need to have a Physical Twin yet and is predominantly used for technical risk mitigation. It can be argued that the virtual models and simulation software used by the Research and Development department constitute such a ‘pre-Digital Twin’. Accordingly, the Pre-Digital Twin would not change anything for the company, as this is what is done already.
TRL 4 describes how a basic prototype is validated in a laboratory environment. Since there is a presence of a prototype in this phase, it means that there is a Physical Twin which the Digital Twin can communicate with. In a laboratory environment the prototype acts individually and self-contained. Employing a Digital Twin in this phase may aid in gaining a steeper learning curve of the new technology since it directly receives data from its physical counterpart.
TRL 5 and 6
TRL 5 and 6 describe how a basic prototype is validated and demonstrated in a relevant environment respectively. In the relevant environment the technology isn’t self-contained, but rather a part of a larger system. An especially interesting facet in a relevant environment would be if there already exists a Digital Twin of the manufacturing process, such that the Digital Twin of the new technology can be integrated within the Digital Twin of the larger system, allowing for easier and smoother transition. Accordingly, Digital Twin technology looks promising during these maturity phases.
TRL 7 – 9
During the last three phases of maturing, the new manufacturing technology is used in (series) production. Due to all the data that is generated in production and accumulated in the Digital Twin, the analytics become richer which allows for further optimization. Accordingly, Digital Twin technology looks very promising during these maturity phases.
29
V
Discussion
Answering the research question
This research began by aiming to discover the contribution of digital twins to the maturing of new manufacturing technology. When it became apparent that during the timespan of this research it would not be possible to observe the creation of a digital twin of a new manufacturing technology, the type of research changed from an explanatory to an exploratory one. To do so, it became imperative to first develop a comprehensive understanding of Digital Twins by means of a thorough literature review and second, find out how currently new technology is introduced and commissioned by the company of the case study. To discover this process, internal protocols were analysed which describe the process. Since the aerospace industry is highly regulated and extremely well-documented, it can be stated with certainty that the flow-diagram, resulting from this analysis, correctly displays the reality. Interviews were conducted with parties involved in the process to discover the underlying processes. Thereafter, it was explored how Digital Twin may provide opportunities to enhance the maturing process of new manufacturing technology.
Based on the findings of this case research it can be concluded that the potential contribution of Digital Twins to the maturing of new manufacturing technology, is that they promise opportunities to speed up the process. The findings show that the development of new technology is very time-consuming and follows a structured process. The procedure of this process is illustrated in Figure 11 and from the interviews it became clear that the most time consuming step of the procedure is in between each cycle when technology needs to be further developed so it meets the requirements formulated to reach a TRL maturity level. These requirements are technology-specific but can be classified under the top-level definition of the TRL illustrated in Table 2.
Examining the top-level definitions of the TRL, there are certain phases during which Digital Twins can play a role to speed up the maturing process. The potential contribution seems little during the early maturity levels but increases as the technology progresses to higher maturity levels.
In the very beginning of new technology maturing, meaning TRL 1 and 2, there seems no evident use for Digital Twins. These levels only entail the observation of the basic principle and the formulation of a technology concept. From TRL 3 onwards the first opportunities arise for Digital Twins to play a role. Nevertheless, the research and development department, which is responsible for the maturing of new technology during these stages already utilise already utilise state-of-the-art modelling and simulation software, which, according to Madni et al. (2019), can be regarded as the Pre-Digital Twin. Thus, there seems to be no novelty in this specific case in this phase, but generally speaking, utilising Pre-Digital Twins have a positive contribution in the early maturing phase of new technology development.
30
Limitations
This study was conducted according to validated research methodologies. The theory section was developed using a systematic literature review and for the case study triangulation, the use of multiple data sources, was employed to increase reliability and gain a more complete understanding of the case. Nevertheless, there are a number of limitations that need to be addressed concerning this research. First of all, due to the novelty and ambiguity of Digital Twins, there might be discrepancies with other literature. This paper extensively describes the concept and discusses different views on the topic to eventually come to a model. However, as the concept of Digital Twins still finds itself in its infancy, it may be possible that this model will not uphold indefinitely.
Secondly, the interviews that were conducted were exploratory in nature with the specific aim of discovering opportunities for Digital Twins to play a role. This type of interview can lead to subjective results. Contrary to the interview with the R&D department, which was conducted with two interviewees from different echelons, the interview with the industrialization department was conducted with just one interviewee, meaning that the answers provided in this research are possibly even more subjective. Perchance other employees from this department do not agree with the answers provided by the Industrialization Manager.
Thirdly, the scope of this research is relatively limited. The study was conducted at a single company who has only yet begun investigating the potential use of Digital Twins. More importantly, the company operates in the aerospace industry, which is highly regulated and the products they produce must be of a guaranteed quality. For other industries, the relevance of using Digital Twins might differ.
Fourth, in exploring the potential contribution of Digital Twins on the maturing of new manufacturing technology, the only focus was on the speed of maturing. There might be other contributions than just the speed of the process. For example, it might be interesting whether the use of Digital Twins could reduce the total expenses of the maturing process of new manufacturing technology.
Lastly, technology is very diverse and therefore the requirements for a technology to mature also vary enormously. This makes it difficult to pinpoint one or two single components that Digital Twins can enhance. The contribution of Digital Twins in the maturing of new manufacturing technology may therefore greatly vary for different cases.
Theoretical implications
31
Practical implications
Innovations lead to new technologies which may provide companies with a competitive advantage over their competitors. The practical implication of this research is that it may be beneficiary for companies to develop Digital Twins of their manufacturing technology whilst it is still in development. Nevertheless, criticality is paramount in deciding whether to utilise Digital Twins. There must be a clear objective for choosing to use Digital Twins, since they are versatile and can be used for different applications. If one wishes to develop a Digital Twin one should remain critical in deciding at which maturity phase a Digital Twin is developed. The potential contribution increases over the TRL levels, and there should always be a clear business case. One of the most promising contributions of Digital Twins is when new manufacturing technology is to be introduced in a relevant environment. If a Digital Twin of the relevant environment already exists, and a new manufacturing technology is developed which replaces one of the steps in this process, the integration of a Digital Twin of this technology in the Digital Twin of the relevant environment, could ensure a smooth and easier transition. Subsequently, any calculations concerning the non-recurring costs of designing a production line could become more accurate and the time-to-market is reduced. This ultimately results in more precise quotations the company can give to their customers, which may increase the company’s competitive advantage.
32
VI. Conclusion
This researched aimed at discovering the contribution of Digital Twins to the maturing of new manufacturing technology. Through a case study in the aerospace industry, the procedure of the maturing process of new manufacturing technology was analysed and described. Interviews with focus groups responsible in this process helped in exploring in which phases of technology development Digital Twins can play a role. Based on the findings of this research it can be concluded that there is a potential contribution of Digital Twin in the maturing of new manufacturing technology. The potential contribution is that Digital Twins speed up the maturing process and new technology can progress faster through the TRL Maturity levels. The greatest potential of Digital Twins lies in the last maturing phase when the manufacturing technology is to be used in production. Digital Twins can help to ensure a smooth transition of a technology from a research and development environment to a relevant production environment.
In early development phases, without the existence of a physical counterpart, the ‘Pre-Digital Twin’ can be used for technical risk mitigation. Nevertheless, there is little novelty in the Pre-Digital Twin, compared to already existing modelling and simulation software. If decided that there is going to be a Digital Twin of a manufacturing technology should be taken into account in the design process. For the Digital Twin to receive data from the Physical Twin it needs sensors, actuators and devices for network connecting (see figure 6). This should be taken into account in the design of new manufacturing technology. In the presence of a Physical Twin, the Digital Twin can help to quickly gain knowledge concerning the Physical Twin. Due to receiving feedback from its physical counterpart, the learning curve of the technology increases.
33
VII References
Benett, E., & Ivashchenko, A. (2017). Digital twin tackles design challenges.
https://doi.org/10.1016/S0262-1762(17)30139-6
Boote, D. N., & Beile, P. (2005). Scholars Before Researchers: On the Centrality of the Dissertation Literature Review in Research Preparation. Educational Researcher, 34(6), 3–15. https://doi.org/10.3102/0013189X034006003
Bouskela, D. (2019, November). ITEA Success story: MODRIO. ITEA Magazine, 18–21.
Burnard, P. (1991). A method of analysing interview transcripts in qualitative research. Nurse Education
Today, 11(6), 461–466. https://doi.org/10.1016/0260-6917(91)90009-Y
Cai, Y., Starly, B., Cohen, P., & Lee, Y. (2017). Sensor data and information fusion to construct digital-twins virtual machine tools for cyber-physical manufacturing. Procedia Manufacturing, 10, 1031– 1042. https://doi.org/10.1016/j.promfg.2017.07.094
D’Addona, D. M., Ullah, A. M. M. S., & Matarazzo, D. (2017). Tool-wear prediction and pattern-recognition using artificial neural network and DNA-based computing. Journal of Intelligent
Manufacturing, 28(6), 1285–1301. https://doi.org/10.1007/s10845-015-1155-0
DiCicco-Bloom, B., & Crabtree, B. F. (2006). The qualitative research interview. Medical Education,
40(4), 314–321. https://doi.org/10.1111/j.1365-2929.2006.02418.x
Drath, R., & Horch, A. (2014). Industrie 4.0: Hit or hype? [Industry Forum]. IEEE Industrial Electronics
Magazine, 8(2), 56–58. https://doi.org/10.1109/MIE.2014.2312079
Eisenhardt, K. M. (1989). Building Theories from Case Study. Academy of Management, 14(4), 532– 550. Retrieved from http://www.jstor.org/stable/258557
El Saddik, A. (2018). Digital Twins: The Convergence of Multimedia Technologies. IEEE Multimedia,
25(2), 87–92. https://doi.org/10.1109/MMUL.2018.023121167
Elhajar, R., La Saponara, V., & Muliana, A. (2017). Smart Composites: Mechanics and Design (1st ed.). CRC Press.
Flinn, P. (2019). Managing Technology and Product Development Programmes. In Managing
Technology and Product Development Programmes. https://doi.org/10.1002/9781119517283
Foster, R. N. (1986). Innovation: The Attacker’s Advantage. London: Macmillan.
Goossens, P. (2017). Industry 4.0 and the Power of the Digital Twin. 1–17. Retrieved from https://dupress.deloitte.com/dup-us-en/focus/industry-4-0/digital-twin-technology-smart-factory.html
Grieves, M., & Vickers, J. (2016). Digital Twin: Mitigating Unpredictable, Undesirable Emergent Behavior in Complex Systems. Transdisciplinary Perspectives on Complex Systems, 85–114. Guo, J., Zhao, N., Sun, L., & Zhang, S. (2019). Modular based flexible digital twin for factory design.
Journal of Ambient Intelligence and Humanized Computing, 10(3), 1189–1200.
https://doi.org/10.1007/s12652-018-0953-6
Headrick, D. R. (2009). Technology: A World History. Oxford University Press, Incorporated.
Kagermann, H., Wahlster, W., & Helbig, J. (2013). Recommendations for implementing the strategic
initiative INDUSTRIE 4.0.
34
categorical literature review and classification. IFAC-PapersOnLine, 51(11), 1016–1022. https://doi.org/10.1016/j.ifacol.2018.08.474
Lee, J., Lapira, E., Bagheri, B., & Kao, H. (2013). Recent advances and trends in predictive manufacturing systems in big data environment. Manufacturing Letters, 1(1), 38–41. https://doi.org/10.1016/j.mfglet.2013.09.005
Liu, J., Zhou, H., Tian, G., Liu, X., & Jing, X. (2019). Digital twin-based process reuse and evaluation approach for smart process planning. International Journal of Advanced Manufacturing
Technology, 100(5–8), 1619–1634. https://doi.org/10.1007/s00170-018-2748-5
Madni, A., Madni, C., & Lucero, S. (2019). Leveraging Digital Twin Technology in Model-Based Systems Engineering. Systems, 7(1), 7. https://doi.org/10.3390/systems7010007
Marsh, G. (2012). Aero engines lose weight thanks to composites. Reinforced Plastics, 56(6), 32–35. https://doi.org/10.1016/S0034-3617(12)70146-7
McCutcheon, D. M., & Meredith, J. R. (1993). Conducting case study research in operations management. Journal of Operations Management, 11(3), 239–256. https://doi.org/10.1016/0272-6963(93)90002-7
Negri, E., Fumagalli, L., & Macchi, M. (2017). A Review of the Roles of Digital Twin in CPS-based Production Systems. Procedia Manufacturing, 11(June), 939–948. https://doi.org/10.1016/j.promfg.2017.07.198
Oxford Lexico. (2019). Lexico.com. Retrieved November 12, 2019, from https://www.lexico.com/en/definition/technology
Park, H.-W., Sung, T.-E., & Kim, S.-G. (2015). Strategic Implication of Technology Life Cycle on Technology Commercialization. International Association for Management Technology, 2736– 2748.
Peters, S. (2015). A readiness level model for new manufacturing technologies. Production Engineering,
9(5–6), 647–654. https://doi.org/10.1007/s11740-015-0636-5
Peters, W., Doskey, S., & Moreland, J. (2017). Technology Maturity Assessments and Confidence Intervals. Systems Engineering, 20(2), 188–204. https://doi.org/10.1002/sys.21389
Rodič, B. (2017). Industry 4.0 and the New Simulation Modelling Paradigm. Organizacija, 50(3), 193– 207. https://doi.org/10.1515/orga-2017-0017
Rosen, R., Von Wichert, G., Lo, G., & Bettenhausen, K. D. (2015). About the importance of autonomy and digital twins for the future of manufacturing. IFAC-PapersOnLine, 28(3), 567–572. https://doi.org/10.1016/j.ifacol.2015.06.141
Salkind, N. J. (2010). Case Study. In Encyclopedia of Research Design (pp. 116–119). SAGE Publications, Inc.
Schleich, B., Anwer, N., Mathieu, L., & Wartzack, S. (2017). Shaping the digital twin for design and production engineering. CIRP Annals - Manufacturing Technology, 66(1), 141–144. https://doi.org/10.1016/j.cirp.2017.04.040
Schuh, G., Schubert, J., & Wellensiek, M. (2012). Model for the valuation of a technology established
in a manufacturing system. Procedia CIRP, 3(1), 602–607.
https://doi.org/10.1016/j.procir.2012.07.103
35 137–140. https://doi.org/10.1016/j.cirp.2017.04.038
Stark, R., Fresemann, C., & Lindow, K. (2019). Development and operation of Digital Twins for technical systems and services. CIRP Annals, 68(1), 129–132. https://doi.org/10.1016/j.cirp.2019.04.024 Stark, R., Kind, S., & Neumeyer, S. (2017). Innovations in digital modelling for next generation
manufacturing system design. CIRP Annals - Manufacturing Technology, 66(1), 169–172. https://doi.org/10.1016/j.cirp.2017.04.045
Susto, G. A., Schirru, A., Pampuri, S., McLoone, S., & Beghi, A. (2015). Machine learning for predictive maintenance: A multiple classifier approach. IEEE Transactions on Industrial Informatics, 11(3), 812–820. https://doi.org/10.1109/TII.2014.2349359
Tao, F., Cheng, J., Qi, Q., Zhang, M., Zhang, H., & Sui, F. (2018). Digital twin-driven product design, manufacturing and service with big data. International Journal of Advanced Manufacturing
Technology, 94(9–12), 3563–3576. https://doi.org/10.1007/s00170-017-0233-1
Thomé, A. M. T., Scavarda, L. F., & Scavarda, A. J. (2016). The Management of Operations Conducting systematic literature review in operations management. Production Planning & Control, 7287, 1– 13. https://doi.org/10.1080/09537287.2015.1129464
Tuegel, E. J., Ingraffea, A. R., Eason, T. G., & Spottswood, S. M. (2011). Reengineering aircraft structural life prediction using a digital twin. International Journal of Aerospace Engineering, 2011. https://doi.org/10.1155/2011/154798
Uhlemann, T. H. J., Lehmann, C., & Steinhilper, R. (2017). The Digital Twin: Realizing the Cyber-Physical Production System for Industry 4.0. Procedia CIRP, 61, 335–340. https://doi.org/10.1016/j.procir.2016.11.152
United States Government Accountability Office. (2009). NASA Assessments of Selected Large-Scale
Projects. Retrieved from http://purl.access.gpo.gov/GPO/LPS113423
Voss, C., Tsikriktsis, N., & Frohlich, M. (2002). Case research in operations management. 22(2), 195– 219. https://doi.org/10.1108/01443570210414329
36
VIII Appendices
Appendix 1 Conducting a systematic literature review of Digital Twins
The literature review is arguably the most important section of a thesis. An extensive and profound literature review is crucial for thoroughly understanding the topic and getting acquainted with the different terminology and aspects that play a role. Therefore, the literature review can be regarded as the foundation of the thesis. Similar to construction work, where a solid foundation ensures a solid building, the literature review should be solid to ensure a solid thesis. Moreover, knowing what has already been researched helps to prevent reinventing the wheel and may give insight of what other researchers argue is important to still be investigated.
To aid in conducting a sound literature review, a framework was used which embodies a detailed step-by-step approach. It is the Systematic Literature Review in Operations Management by Thomé et al. (2016) which describes a methodology that locates existing studies, selects and evaluates contributions, analyses and synthesizes data and finally reports the findings in a clear way. The framework consists of eight steps as shown in figure 12 and each step comprises a number of sub steps or tasks. Lastly, reliability checks among researchers should be performed between some steps.
37
