Knowledge Creation in the era of Cyber-Physical Systems: a Case Study

(1)

Knowledge Creation in the era of

Cyber-Physical Systems:

a Case Study

Author Björn Wokke Student number S2999005

Master thesis, MSc Technology and Operations Management University of Groningen, Faculty of Economics and Business

January 2018

Word count: 11.968

Supervisor – University of Groningen J.C. Wortmann

Co-assessor – University of Groningen S. Waschull

K.J. Roodbergen

(2)

(3)

Abstract

The fourth industrial revolution is at our doorstep, enabling a multitude of opportunities made available by advanced technologies that combined represent cyber-physical systems. However, we do not know yet what aspects of CPS technology affect the performance of manufacturing organizations and how. This paper adds to the body of literature by clearly describing how CPS technology affects organizational performance through its impact on knowledge creation. Additionally, findings are compared to the impact of the third industrial revolution. A single case study was conducted to identify the current state of knowledge creation in organizations and map expectations regarding the impact that CPS technologies would have on this situation. The resulting impact was reflected upon in literature to determine whether and how organizational performance was influenced. It was found that CPS technology enables the building of a digital explicit knowledge base, reinforcing the combination process implicating similar effects as the third revolution. However, influence on organizational performance can only be achieved by using human or software resources to structure the knowledge base and determining who benefits from having specific knowledge. Fortunately, contrarily to the third revolution, CPS technology also enhances externalization and internalization processes. Technologies enable better presentation of knowledge to employees allowing for proactive problem solving and new knowledge creation throughout the organization. Future research is encouraged to investigate whether these variables interact differently in smaller organizations.

(4)

Preface

Before you lie the research paper “Knowledge Creation in the era of Cyber-Physical Systems: A Case Study”, resulting from an empirical investigation in the aviation industry. It has been written to fulfill the graduation requirements of the Technology and Operations Management master program at the Rijksuniversiteit Groningen (RUG). I have been engaged in empirically investigating and writing the research paper from September 2017 to January 2018.

The research was conducted in cooperation with Fokker Aerostructures B.V., where I interned. Together we formulated a research plan that fulfilled both the practical needs for the internship company and the theoretical needs for the university. Continuously balancing practice and theory throughout the process has been challenging, but fortunately my supervisors S. Waschull (RUG) and S. Hengeveld (Fokker) were always available to answer questions and engage in discussion.

I would like to thank my supervisors from the university for their guidance and support during the researching and writing processes. Without their cooperation I would not have been able to present the research paper in its current state. To my colleagues at Fokker: I would like to thank you for your making me feel welcome. Your openness, cooperation and willingness to be involved has been wonderful. Our conversations and discussions have been very helpful in getting to know the organization and conducting my research. Debating issues with fellow students, friends and family further motivated me to stay focused. Particularly my parents, as always, have been a source of both wise and kind words which helped and encouraged me to succeed.

Björn Wokke

(6)

1. Introduction

The scientific and the industrial sector have identified the presence and importance of the fourth industrial revolution, referred to as industry 4.0, originally launched as a strategic initiative by the German government in 2011 (Bartodziej, 2017). Respectively the (1) mechanization based on water and steam, (2) mass production by using electric energy and (3) automation through electronics and information technology represent the first three industrial revolutions. Industry 4.0 focuses on integrating new digital technologies into industrial production systems, allowing for the interconnection of physical objects such as machines, products, people and computers through the internet (Kagermann, Wahlster, & Helbig, 2013). The collection of new digital technologies that organizations need to integrate to enable the Industry 4.0 initiative is typically referred to as a cyber-physical system (CPS) (Bartodziej, 2017). The CPS includes a network of embedded systems which are typically implanted in other computational systems (Monostori et al., 2016). These embedded systems have two functionalities: (1) interconnect multiple computational systems and (2) enable sensing, monitoring and actuating of physical elements. According to this conception, a CPS connects the physical and virtual worlds.

The CPS technology is claimed to form the basis for the transfer from the third to the fourth industrial revolution (Bartodziej, 2017). One important element that has been associated with the third ‘information technology’ revolution is the management of knowledge which is

(7)

For instances, literature highlights CPS opportunities such as a new digital source of knowledge through big data & data analytics (Kune, Konugurthi, Agarwal, Chillarige, & Buyya, 2016), real-time communication of knowledge through coordination between computational and physical processes (Monostori et al., 2016) and enhanced sharing or presenting of knowledge through human-machine interfaces (Stock & Seliger, 2016). However, challenges present themselves in which analytical tools to use, how to integrate data coming from multiple sources and necessity of data scientists (Sumbal, Tsui, & See-to, 2017). Furthermore, all of these possibilities indicate that the environment of the employee is going to change, causing employees to become decision makers instead of manual workers which indicates the need for a different skillset (Hirsch-Kreinsen, 2016) and optimal support while working on the job (Dworschak & Zaiser, 2014).

With all this in mind, there seems to be a limited amount of information available on how CPS technologies used by industrial organizations will impact knowledge creation, and in turn, how this will influence organizational performance. Therefore, this paper will commit to finding answers to the following question:

How does the transformation towards a CPS influence organizational performance through its impact on knowledge creation for manufacturing organizations?

More specifically, defined sub-research questions highlight the different dimensions:

1. How do manufacturing organizations currently apply knowledge creation processes? 2. How do manufacturing organizations expect that CPS will affect knowledge creation? 3. How can the affected knowledge creation enhance organizational performance?

Answering these questions results in an overview of how CPS changes knowledge creation and how this will result in advantages for the manufacturing organization. Additionally, it will display whether CPS has the same effects on knowledge creation compared to the third revolution.

The next chapter theoretically elaborates relevant variables and connections included in this study. Thereafter, the third chapter provides insight into the used methodology during the research process. Chapter 4 presents in detail what findings resulted from the investigation. Next, the fifth chapter discusses the equalities and differences between the practical findings and the theoretical background chapter, concluding into an answer to the main research question in chapter 6.

(8)

2. Theoretical Background

The acquisition and usage of knowledge is perceived as one of the most important factors in modern business life (Uden & He, 2017), making it difficult to imagine how organizations would compete without putting effort into knowledge creation. Researchers in the past and present typically use the model as made by Nonaka (1994) when referring to knowledge creation. As technologies continuously keep changing, the traditional way of looking at and applying knowledge creation needs to alter. One of the reasons for this is that from Nonaka’s viewpoint, knowledge creation only occurred through social interaction between individuals, whereas in the present new digital sources of knowledge become available through digital technologies. To illustrate the relationship between knowledge creation and transforming towards a CPS, section 2.1 will explain what knowledge is and what the traditional concept of knowledge creation looks like. Early section 2.2 defines the concept of CPS, eventually resulting in a detailed overview of changes and implications that an organization must cope with when transitioning towards a CPS in the context of knowledge creation. Finally, section 2.3 will highlight achievable organizational benefits through an enhanced continuous knowledge creation cycle resulting in an overview of all discussed variables in a conceptual model.

2.1 Knowledge and knowledge creation

Existing literature defines that individuals build knowledge from data, while first processing data into information (Uden & He, 2017). More specifically, Bellinger et al. (2004) write that information equals data given purpose, while knowledge represents the usage of information in the right context. According to Alavi & Leidner (2001), ‘knowledge is information possessed in the mind of individuals: it is personalized information (which may or may not be new, unique, useful or accurate) related to facts, procedures, concepts, interpretations, ideas, observations and judgement’.

(9)

aware of the fact that they have specific knowledge, therefore resulting in the fact that it is not expressible and making it tacit. In this case, conscious awareness is the element that separates tacit from explicit knowledge. Back to Nonaka, using four conversion processes between tacit and explicit knowledge which are known as socialization, externalization, combination and internalization (SECI) in succession, allows organizations to effectively transform existing knowledge into new knowledge (see figure 1). Table 1 represents key characteristics of each conversion process.

Figure 1. The four knowledge conversion processes (SECI)

Important to notice is that in this model, all conversion methods are social processes between individuals, meaning that people create knowledge through interaction. Nonaka expresses that this process amplifies the quality and quantity of knowledge organizationally. In more detail, knowledge created at the level of the individual can spiral into knowledge that is relevant for an entire section, department and organization through the SECI model.

2.2 From IT to Cyber-Physical Systems

(10)

According to Geisberger & Broy (2015), a CPS should be seen as a set of interconnected technological systems and devices, containing information and communication technologies. becomes more powerful due to continuous improvement of information and communication technology. The growing potential of modern computing, transmission and storage capacities is enabling the improvement of these technologies, enhancing their interconnective characteristics. Monostori et al. (2016) characterize a CPS as being a computing and communicating core built to monitor, control, coordinate and integrate operations performed by physical and engineered systems. Whereas, Kagermann, Wahlster, & Helbig (2013) argue that CPS are interconnected via digital networks and are equipped with human-machine interfaces. Based on the terminologies, the definition of CPS in this paper is “Interconnected technological systems and devices via digital networks, whose operations are monitored, controlled, coordinated and integrated by a computing and communicating core”.

Table 1 The SECI knowledge conversion processes and key characteristics (Nonaka, 1994)

Factor Description

Socialization - from tacit knowledge to tacit knowledge

Requirements Physical experiences, care and trust between members Goal Sharing experiences and creating common tacit knowledge Means Observation, imitation and meetings

Example Apprenticeship

Externalization - from tacit knowledge to explicit knowledge

Requirements Ability to see patterns, logical translation or conscious awareness Goal Crystallize tacit knowledge into explicit concepts

Means Conversion through sequential use of metaphors, analogies and models Example Concepts, manuals and models

Combination - from explicit knowledge to explicit knowledge

Requirements Collect and combine explicit knowledge and spread it in the organization Goal Process combined explicit knowledge to make it more useable

Means Document sharing, meetings, telephone conversations and network Example Break-down of concepts, operational business plans and instructions Internalization - from explicit knowledge to tacit knowledge

Requirements Gain know-how by embodying explicit knowledge as tacit knowledge Goal Broaden, extend and reframe members’ tacit knowledge

Means Learning by doing

(11)

Researchers now write a lot about what opportunities and challenges present themselves when an organization is transforming towards a CPS. Opportunities and challenges are highlighted in the work of Bartodziej (2017) and Kagermann et al. (2013) who both elaborately discuss CPS technology functionalities and provide related recommendations. From these papers, three major influences have been distinguished: (1) big data and data analytics, (2) real-time information and human-machine interfaces and (3) skillset and on the job support. Big data & data analytics

The acquisition, processing and sharing of data between diverse sources is enhanced by new digitalization technologies, allowing the possibility to arrive at new digital knowledge (Li, Xu, & Zhao, 2015). The development in data acquisition is typically identified as big data, explaining that technological advancements cause huge volumes of data to be generated from multiple sources (Waller & Fawcett, 2013). Logically, this causes business analytics to be seen as a major technology trend (Chen, Chiang, & Storey, 2012). Transforming these amounts of data into explicit knowledge through reasoning and pattern finding, could have strong implications for improved business performance. Based on this information and the findings of Sumbal, Tsui & See-to (2017) some implications on the linkage between big data and knowledge creation require attention. First, knowledge is extractable from many new (digital) sources. Second, data scientists become critical when properly analyzing data and knowledge. Third, it is unclear which tools are best suited to digitally integrate and process data, information and knowledge coming from multiple sources. Last, increasing advancement in algorithms might cut the need for tacit knowledge in the future.

Data as a potential source of explicit knowledge contradicts the statement that Nonaka (1994) made, claiming that knowledge creation processes only occur based on human interaction. Based on recent literature as mentioned above, gaining explicit knowledge from data could become of increasing importance for modern and future organizations, having implications for the combination process. Whether this is the case and how data should be translated to explicit knowledge will be reflected upon.

Real-time information and human-machine interaction

(12)

present information and explicit knowledge to the employee and translate the actions of employees to digital language, making them information presenters and collectors. Therefore, using HMIs directs expectations towards being able to provide employees with real-time information. This is also in line with the following statement: to support the employee in decision-making processes, some technologies function as personalized assistance systems that provide users with a comprehensible visualization of context-sensitive information (Monostori et al., 2016). Furthermore, by seeing HMIs as information collectors it becomes possible for employees to share their knowledge with other employees through the CPS. Both statements imply the enablement of sharing expressible and therefore explicit knowledge during the combination and internalization processes. However, literature also hints towards using these kinds of sharing mechanisms for the sharing of tacit knowledge, enhancing socialization and externalization efforts (Al-Qdah & Salim, 2013).

The ability of real-time information and HMIs to enhance explicit and tacit knowledge sharing between employees within an organization is therefore hypothesized, influencing all SECI knowledge creation processes.

Skillset & on the job support

(13)

2.3 Operational benefits of knowledge and knowledge creation

The outcomes of organizational knowledge creation were underdeveloped in Nonaka (1994). Based on the work of Nonaka & von Krogh (2009), two ways of organizational outcomes have been adopted, (1) knowledge outcomes and (2) social practice outcomes.

In terms of knowledge outcomes, enhanced understanding is useful for product and process innovation, while an enhanced capacity to act through growing levels of tacit and explicit knowledge better equips operators to look at problems and find solutions, which increases task performance. Social practice outcomes present themselves in the shape of intensified interaction between organizational members, facilitating the expansion of individual and group knowledge. Additionally, by going through the knowledge creation cycle, human activity during individual SECI processes standardizes when used in a specific setting. In more detail, organizational members will learn to identify and remove obstacles to knowledge creation. The enhancement of knowledge creation in general has its benefits. To clarify the impact of digitalization on organizational benefits through knowledge creation, the next section explores connections between mentioned changes in paragraph 2.2 to expected outcomes on the

knowledge and social practice perspective.

Expected benefits of upgraded knowledge creation cycle due to digitalization

Enhanced ability to gain knowledge from data coming from digital objects and processes including the potential usage of analytical tools is expected to increase an organizations capacity to act on preventively identifying problem situations and finding solutions based on growing levels of explicit knowledge in this case. To be able to act, organizations must use the growing explicit knowledge base for internalization purposes, causing the understanding of individuals to thrive. As mentioned by Nonaka & von Krogh (2009) these knowledge outcomes in the shape of acting capacity and understanding are essential for product and process innovation, and task performance due to amplified problem identification and solution finding. Additionally, increased sharing options and the ability to share real-time facilitates social practices which are interaction between individuals. Through this process the expansion of tacit knowledge of individuals and groups can be facilitated, providing an opportunity to further infuse acting capacity and understanding.

(14)

understanding to improve processes and products, identify problems and find solutions is compliant with the findings of North & Hornung (2003) who conclude that effectively managing knowledge results in process acceleration, knowledge transparency, error reduction and redundancy avoidance.

(15)

3. Methodology

According to Yin (2013), there are three characteristics that determine when a case study should be used: (1) the research question typically answers “how” or “why” research questions, (2) the investigator has little to no possibility to control behavioral events and (3) the general circumstances of a phenomenon that is to be studied are in a real-life context. Furthermore, knowledge creation in organizations is a well-known phenomenon. However, how the phenomenon reacts to an organizational change towards a CPS is relatively unexplored. It is unclear what opportunities the CPS technology will offer in terms of knowledge creation, implying the need for theory building. Therefore, this study investigates how technological opportunities affects organizational performance through the impact that they have on the individual knowledge creation processes of socialization, externalization, combination and internalization. Having all three characteristics mentioned by Yin (2013) and focusing on theory building means that this research suits best with an exploratory case study. Therefore, a single case study at Fokker Aerostructures in Hoogeveen is relevant.

3.1 Case description

Fokker Aerostructures, Fokker Elmo, Fokker Landing Gear and Fokker Services are the four components that together form the Fokker Technologies group. By being a market and technology leading organization, Fokker Aerostructures faces the need to accommodate a more flexible and efficient production location, producing higher volumes while maintaining a high-quality level. This is one of the first reasons why the organization is showing interest in the opportunities that come with new technologies. In more detail, the organization realizes that their method of sharing instructions with operators is outdated and impedes the entire production process. This realization motivated the organization to act in terms of using technologies to enhance the process of feeding instructions to the work floor. This section first elaborates how instructions are composed in the current situation and then goes into more detail as to what actions have been taken to improve.

The current process of instruction

(16)

for a specific production process. In turn, the TPD has its own set of files containing instructions on how to perform quality checks called ‘proces instructie’ or PI which are digitally obtainable by scanning the second barcode. When to perform what kind of quality checks is in then again presented in the BV. Completing a quality check results in a stamp on the BV while blanks on the BV indicate where operators should write down as-built data.

McKinsey consultancy

Realizing the impact that the upcoming industry 4.0 revolution might have on this process and the organization in general, Fokker hired McKinsey consultancy in 2016 to investigate what digital solutions have potential for their organization. McKinsey concluded their investigation with a one-page solutions document, recommending solutions such as using sensors, RFID and advanced analytics. Most importantly they advised integrating multiple IT systems through a central database because it has the potential to reduce the time needed for creating work orders and instructions, leading to 10% capacity freed in manufacturing engineering. The benefits of having this central database are: (1) standardization of product specifications by focusing on common processes, (2) design changes are traceable in a DPD (digital product dossier) as required by customers, (3) knowledge and best practices on processes are more easily captured and shared and (4) work instructions created based on specification allow for better visualization. To take advantage of these benefits, Fokker is planning a Fokker 4.0 pilot in the composite lay-up departments in Hoogeveen and Papendrecht in the first quarter of 2018. Fokker 4.0

(17)

situation. Starting on a small-scale, the pilot specifically focuses on the Apache lay-up department since it is a relatively small and well-developed process. Therefore, this study specifically focuses on the Apache lay-up process and its related employees.

Apache lay-up

The Apache lay-up team splits up in two. First, the performance team represents the group of operators that perform the manual lay-up work. Operators are required to have at least completed their preparatory middle-level applied education and need to finish predefined general and lay-up Fokker training within a year after starting. Second, the PITCH team represents the employees that are responsible for supervising the lay-up operators and process. The PITCH team includes four members with different functions: (1) team leader, (2) product engineer, (3) quality controller and (4) planner.

3.2 Data collection

Voss, Tsikriktsis, & Frohlich (2002) describe that triangulation is one of the most important aspects when collecting data. By using and combining multiple methods of data collection, the reliability and validity of eventual results and conclusions increases in strength. Distinguishable sources of evidence are documents, archival records, interviews, direct observations, participant observations and physical artefacts (Yin, 2013). To realize triangulation, this paper will collect evidence mainly through participant-observations by the researcher, documents made available by the company and interviews with employees. A full overview of the used data collection methodologies can be found in Appendix Awhich will be explained in more detail below. How tacit and explicit knowledge and interactions between these types of knowledge have been operationalized can be found in appendix B. Clear should be that expectations regarding the use of new digital technologies is based on the ‘expert’ opinions of employees who have had experience with the Fokker 4.0 pilot either in the shape of participating in the Proof of Concept (operators & team leader) or by being a member of the Fokker 4.0 project team.

Answering the first two sub-research questions

(18)

in daily meetings with the full PITCH team and having one-on-one conversations with individual members, the focus was on finding out how the organization (1) facilitates becoming conscious awareness, (2) organizes the collecting, analyzing and using of data and explicit knowledge and (3) facilitates explicit knowledge sharing. Participant observations and meetings with the individual Fokker 4.0 project team members further clarified the current process, while also validating previous findings. At the same time, it provided the opportunity to ask follow-up questions regarding changes they expect concerning the Fokker 4.0 pilot and future technological developments. All observations and resulting conversations or discussions were noted. Additionally, visualizing findings resulting from the notes in a flowchart allowed for feedback sessions with the same employees for validation purposes.

Combined with reviewing documents, the observations provided the necessary data to get an initial overview of the current performed activities by Fokker in terms of SECI processes. Furthermore, reviewing the available documentation was predominantly useful in better understanding how the Fokker 4.0 project team views the current flow of information, visualizes changes that occur in the pilot and projects future expectations regarding digitalization. Sets of work instructions were available to get to know what kind of explicit knowledge operators receive on a daily basis.

(19)

consultation, the number of interviews was set to three, provided that the last one would not give any new insights.

Answering the third sub-research question

Collecting data as mentioned above results in answers to the first two sub-research questions, regarding current knowledge creation activities and expected changes in a CPS situation. Constantly comparing these two situations results data in terms of changes that occur in knowledge creation due to CPS technology which provides the source for answering the third question. How these changes influence organizational performance is reflected upon in existing literature that explains the relationship between knowledge creation and organizational performance.

3.3 Data analysis

According to Eisenhardt (1989) and Yin (2013), general analysis of data is the heart of building theory when doing a case study, but also one of the most difficult parts. There always seems to be a creative leap when translating large volumes of data into results and conclusions. When starting an analysis it is important that the raw data is organized and prepared in the form of transcripts so that it can be labeled and coded (Miles, Huberman, & Saldaña, 1994).

(20)

4. Findings

The findings chapter splits up in three sections, respectively showcasing empirical evidence for the three sub-research questions. Each section ends with concluding remarks, giving answer to the related question.

4.1 SECI at Fokker now

To be able to say how organizations should apply SECI processes in a digitalized context it is important to investigate how they currently apply them. The current process provides a baseline with which the expectations can be compared. This baseline points out which SECI processes are currently struggling in a production process that does not use any CPS related technologies and why. The overview in table 2 illustrates the individual SECI processes and their main problems in the current situation at the end of this section. Combining this with the visualization of SECI knowledge creation activities in the current situation in Appendix E sketches a basic overview of the processes and problems that faced by the organization in a non-digitalized environment.

Socialization

Socialization turns out to be the most developed process. It becomes clear that the emphasis in the development of tacit knowledge is on socialization over internalization since operators do not need to follow specific lay-up training before working with the products: "I've been working here for 7 weeks now and I haven't followed any theoretical lay-up training yet”. To still be able to meet quality criteria, work of new employees is strictly overseen and checked by an appointed mentor: “In the first week I only got to watch. In the second week I did some simple work under constant supervision of the mentor who was also there to do quality checks”. This mentoring process is very effective because a single operator is selected to fulfill this position who knows exactly what he has to teach new operators and how he has to do it: “We only have one mentor in our team so that every new operator learns to lay-up the same way. He is just the best in passing on practical knowledge”.

(21)

open to discussion and feedback which enhances socialization and knowledge sharing efforts in general.

Externalization

In terms of externalization there are two ways by means of which individual operators gain conscious awareness, therefore making their knowledge of the situation explicit. The first method is by reading work instructions: “Sometimes I read things in the instructions that I would do very differently myself. If this is the case I discuss the situation with the product engineer and he will change the instructions if I am right.”. This example represents operators becoming consciously aware about performing tasks differently than described by the instructions, allowing them to be able to express themselves. However, three sources of evidence prove that experienced operators do not always read instructions. From observations it became clear that the most commonly used set of instructions contains three different documents. A lot of information spreads out over these three files, each referring to other documents that can be used as back up. This makes it easy to imagine that over time, if parts of the process become a habit for operators, it would be time consuming to read or find specific instructions and therefore appealing for the operator to skip this action. This statement is supported by the second source coming from an interview with one of the operators who claims that: “Officially you need to open the instructions, but I know them by heart now. I can imagine that not reading the instructions sometimes happens because of laziness”. Finally, multiple conversations with individual Apache PITCH team members confirm that sometimes operators do not read instructions at all or not accurately. Not reading instructions prevents an opportunity for the operator to become consciously aware therefore impeding the externalization process.

(22)

knew that the operators skipped a specific production step, probably indicating that it was not in the instruction. The fact that he knew while all operators did not, provides enough evidence that the knowledge was of a tacit nature. The only reason the employee became aware of the fact that he had this knowledge was by facing a problem in production, making it a reactive process. These kinds of problems are preventable by better support on the job or better training. Combination

The way the Apache lay-up team handles errors, non-conformances (NCs) and problems also presents the first bottleneck for the combination process. Operators triggered by differences in instructions usually follow this up by contacting the operator that has the most interaction with the production engineer or in some cases directly contacting the product engineer: “It is possible that my coworkers discuss issues with the product engineer when he walks by – or they might contact the engineer over the phone”. In this situation it depends on the product engineer whether he documents and shares the knowledge within the organization. If he chooses to do so it will be on the work instruction level keeping the knowledge within the Apache lay-up department. If the product engineer chooses not to document problems, knowledge is potentially leaking. Even though it is understandable that in some or even many cases it is the fastest way to solve issues, it is also important to realize that employees communicate valuable knowledge here.

(23)

Clearly, explicit knowledge is not structurally documented, therefore keeping the creation of knowledge very decentralized and limited to programs, departments, teams or even individuals. Nevertheless, newly created knowledge could be very useful for other parties within the organization, which is a very important part of the combination phase as modeled by Nonaka (1994). Experts at Fokker agree with this statement: "Knowledge that is being developed during the recurring [standard / well developed] phase could prove to be valuable for other recurring products but also for non-recurring [new / underdeveloped] ones". Looking further than the production level, other parts of the organization also seem to be interested in shared knowledge: “In the aviation sector it is so important to know what your capabilities are so that you can accurately respond to customer demands, requests and orders". Additionally, not documenting the contents of the mentioned interactions results in a situation where knowledge remains in the heads of employees and will not directly be useable for others that do not have this knowledge. Therefore, reflecting on historical solutions to specific problems when similar problems occur in the future becomes impossible. Equally important, not registering problems and found solutions prevents the possibility to accurately perform problem analyses in the long term. Moreover, individuals forgetting what happened, operators switching between departments or in the worst-case scenario employees leaving the company are examples of situations causing leaking of knowledge since it is difficult or even impossible to retrieve. Internalization

(24)

be different in syntax. Even though reading words and sentences is always dependent on semantics—or individuals giving their own meaning to a word or phrase—, controlling the internalization process becomes even more difficult if instructions that describe the same action have differences in syntax. Obviously, these issues do not matter when instructions are not read at all since explicit knowledge cannot be internalized if it is not read.

Internalization through training and education seems to be the process that suffers the most from the decentralized externalization and combination processes. In fact, this is not more than logical. While training and education focus on spreading knowledge over multiple departments, individual departments within production are busy containing the knowledge that they have. This is all in line with found evidence from conversations with any of the employees. Two general conclusions can be drawn from these conversations. The first is that there is barely any communication between the training and production departments and the second is that there is a clear difference between what is taught during educational sessions and the methods that are being applied on the work floor. These conclusions obviously have a cause and result relationship. Important to add here is that there is a difference in opinion on whether it is possible to teach practical techniques in an educational environment without being on the work floor, which might be part of the reason why these conclusions have been drawn in the first place. An incentive to at least try to spread knowledge throughout the organization over multiple programs through training and education is the fact that most of the times the work floor methods are faster and easier while still allowing the operator to meet customer specifications. Additionally, it is at the very least undesirable to let operators gain experience based on outdated training if better methods exist in practice. Not to mention that there have been earlier efforts to enhance the match between practice and theory: “A while ago there was even a time where people from the training department tagged along during production for a week or 2”. However, obvious and understandable counterarguments such as: “Some things you will only learn in practice” and “Different programs have to deal with different materials and customer requirements” present themselves. But one statement that draws attention is the following: “You basically don’t need any theoretical training to function”, supported by the fact that the latest hired operator has been working for 7 weeks now without any practical or theoretical background in the lay-up process.

(25)

Table 2 The SECI processes applied to Fokker and main issues per process

SECI Activities Main issues

Socialization Mentoring _- Trusting team

Externalization

Creating conscious awareness through reading instructions

Not reading instructions

No real-time instruction updating No defined way of alerting operators about digital changes

Creating conscious awareness through facing problems

Reactive – facing problems generates conscious awareness of tacit or ‘hidden’ knowledge

Combination _{Combining explicit knowledge}

from externalization process

Decentralized knowledge creation & sharing

No structural method to document knowledge.

Combining explicit knowledge from data collection

Internalization

Internalization through instructions

Not reading instructions

No real-time instruction updating No defined way of alerting operators about digital changes

Differences in syntax between instructions that describe the same activity

Internalization through training and education

No communication between

production and training departments Mismatch between what is taught and what is practiced

Conclusion current situation

(26)

departments in different programs or even from the entire organization. Therefore, an organizational knowledge base does not exist, and a lot of newly created knowledge cannot be shared with and internalized by the majority of the organization. This situation is demonstrated by the lack of knowledge sharing from operations to the training department. Not having new explicit knowledge available in training has the effect that it cannot be injected into the organization. This results in sharing ‘old’ knowledge through training which is less or even not useful in practice.

Furthermore, the favored internalization through instructions is impeded because sometimes instructions are not read. If operators do read instructions, it is not guaranteed that they are up to date. Next to obstructing the internalization process, these issues also apply to the externalization process hindering operators in becoming consciously aware through reading instructions. If anything, the general lack of internalization efforts and therefore limited support by explicit knowledge results in a higher possibility to make mistakes. Not documenting knowledge coming from problem solving reinforces this effect by preventing the possibility to make problem analyses and allowing employees to make mistakes that result in the same problems. Generally, all these issues cause reactive conscious awareness through problem solving instead of proactive conscious awareness through reflecting on instructions.

4.2 Impact of CPS technologies on SECI processes

This section will build further on the issues found in the current situation. Per CPS opportunity as mentioned in the theoretical background and illustrated in the conceptual model, this section describes what Fokker 4.0 pilot technology fits in. What changes these technologies have on knowledge creation at the case company are explained and argued in detail and represented by the coding tree in Appendix F. The technologies incorporated in the Fokker 4.0 pilot are the manufacturing process designer (MPD) tool, shop floor viewer (SFV), job manager (JM) and registration manager (RM).

Big data and data analytics

(27)

data will be collected. We do not yet have the resources to do something with registered data. Collected data will probably end up in reports, resulting in a relatively unchanged situation”.

Therefore, in terms of knowledge creation the fundamental structure of the combination process through documentation of data is being built. However, to directly affect the knowledge creation process by providing a new source of explicit knowledge, captured data must be processed and analyzed.

Real-time information and human-machine interfaces

The opportunities of real-time information and human-machine interfaces are made possible by the MPD-tool, JM and SFV. These technologies cooperate with each other to facilitate communication between employees. Multiple forms of communication can be distinguished. First, the instructions that are managed in the MPD-tool can be presented real-time on the SFV through the interconnection that is handled by the JM: “Changes in instructions can be implemented directly so that every next product that passes this stage will immediately be produced according to the newest instructions”. Since the JM can connect to multiple SFVs, explicit knowledge in the shape of instructions can be shared with multiple employees within the organization. Second, multiple human machine interfaces are connected through the JM allowing for direct contact and communication in the shape of chatting between employees that are scattered through the organization: “Direct communication will be available between the operators and the product engineer”.

Through enhanced digital communication efforts, another important opportunity is created in terms of registration of knowledge. As mentioned by one of the Fokker 4.0 project members:

(28)

to the expressed concern. Clearly the operators crave an introduction to the purpose and usage of the chat function. Even though initial response was hesitant, they had interesting input when discussing what such a chat function should look like. The main conclusion coming from this input is that it should be determined in what situations operators should use the chat function. Help should always be available in situations that require immediate attention, while less critical situations should leave the operator free to solve themselves. In the latter situation, operators understand that the chat function can serve informational purposes: “I understand that the chat gives us the opportunity to write down problems that we face”, which enables the documentation of knowledge.

In terms of sharing tacit knowledge through HMIs, mixed findings can be concluded. One of the Fokker 4.0 project members indicated that the SFV currently focuses on sharing instructions based on pictures and that in the future it might support sharing videos and 3D modeling, arguing for the ability to share tacit knowledge through this functionality. Even though the pilot does not directly include these functionalities, it does lay the groundwork for their possibility in the future. Contrarily, operators were reluctant towards this matter, expressing that tricks they have learnt on the work floor could only be shared by showing one another.

Reflecting on the problems that have been highlighted in the current situation, the Fokker 4.0 technologies directly structures explicit knowledge collection from operators. In turn, the increased level of documented explicit knowledge provides the opportunity for the organization to build a knowledge base. This opportunity enables two changes for the organization to be made. First, the explicit knowledge base can be used for new knowledge creation in the combination process. Newly created knowledge can be shared real-time in instructions, enhancing externalization and internalization efforts. Second, the knowledge base can be used for training and education purposes, enhancing internalization efforts throughout the organization.

Skillset and on the job support

(29)

manageable and read more often. This founds the expectation of an increase in becoming consciously aware proactively through the usage of instructions, while also providing better support on the job during the internalization process. Though, there is still no standardized method to alert operators about changes in instructions. Additionally, the organization should be mindful since providing better and more accurate instructions does not guarantee the fact that operators will read them: “Important to keep in mind is that work instructions in the new situation might still be skipped because operators already know what to do”. Currently, the organization focuses on visualizing instructions with the help of pictures. Multiple Fokker 4.0 members indicate that the technology can also support instructions by video and 3D modeling, implying the presence of a digital twin that exactly matches the real-life product.

Product engineers are also expected to benefit from support on the job: “Through the library functionality, instructions that are equal only have to be changed once at the right level”. All individual instructions become modules which process engineers assemble into one overview that clearly describes all tasks step by step for a specific product. This opportunity also removes the possibility to have differences in syntax between instructions that describe the same activity. Conclusion expected situation

The four technological changes provide the organization with the opportunity to tackle most of the problems that have been identified in the current situation. Table 3 summarizes what opportunities provide the organization with solutions through specific technologies.

(30)

Table 3 Opportunities and solutions provided per technology

Opportunity Technology Function(s) Solution

Big data and data analysis

SFV Require operators to type _{in as-built data} _{Structure data collection} method (= not explicit knowledge yet!) RM Collecting and reporting _{as-built data}

Real-time information and HMIs

MPD Real-time management _{and sharing of instructions} Real-time instruction _updating

JM Interconnect all _{technological devices}

Interdepartmental communication / knowledge sharing SFV Digital chat Structure explicit _{knowledge collection}

SFV

Accurate and concise presentation of instructions

Easier to read instructions (might still not be read though)

MPD Changing instructions _{based on library level}

1 – Only need to change instructions once

2 – No more differences in syntax between

instructions that describe the same activity

4.3 Organizational performance

The previous section highlighted the opportunities that become available due to CPS technology. However, it is still not clear how organizations should apply knowledge creation to enhance their performance by responding to CPS opportunities.

Combination and internalization through SFV knowledge documentation

The pilot chat function will directly result in increased knowledge documentation and therefore an explicit knowledge base. Having more explicit knowledge available could enhance collection and combination efforts if the process engineer has enough capacity available to execute these efforts. However, at this point the two people who have the knowledge are the operator who used the chat function and the process engineer who reads it. This situation requires follow up action to achieve sharing of knowledge through the organization. In terms of internalization, two options present themselves.

(31)

it. However, the current platform does not yet include multiple departments or programs, limiting knowledge creation to the Apache lay-up department. More departments and programs need to use the Fokker 4.0 technologies throughout the organization to be able to enhance organizational knowledge creation by taking advantage of CPS technology. When the organization does reach this stage, it needs to put a lot of effort into defining which instructions are generally applicable and which ones only in specific situations.

Second, the organization can use combined explicit knowledge in training and education. As mentioned in the current situation, no new knowledge is shared with training causing a mismatch between what is taught and what is practiced. This situation will not change with the pilot. However, the interconnectedness of technologies does allow for quick sharing, indicating that it is interesting use this opportunity to start changing. Through training, internalization can expand organizationally.

An alternative option is to open the chat function to different receivers. Instead of forcing the sharing of explicit knowledge with only the departments’ process engineer it might be interesting to share with a bigger public. Hence, barriers in sharing knowledge between departments are removed.

Proactive externalization

Externalization through reading instructions already thrives due to concise, accurate and manageable presentation of instructions in the SFV. Next to directly making less mistakes due to better support on the job, the documentation of explicit knowledge coming from problem solving allows for this knowledge to be shared. Once shared, knowledge reduces the chances of other employees making similar mistakes. Additionally, problem analyses can identify major issues that often result in NCs. Having this knowledge allows the organization to focus on tackling the most important NC causes. Overall, the technologies provide the opportunity to shift the focus from reactively becoming conscious aware through problem facing to proactively becoming conscious aware through reading explicit knowledge.

Transforming data into explicit knowledge

(32)

(33)

5. Discussion

The conceptual model in the theoretical background section concluded three potential influences that digitalization towards a CPS might have on the traditional SECI knowledge creation approach as modeled by Nonaka (1994). The three theoretical influences are (1) big data and data analytics, (2) skillset and on the job support and (3) real-time information and human-machine interfaces.

Big data and data analytics

Important to start with and as predicted, the traditional SECI processes should be expanded to accommodate a new source of knowledge. The original model, created by Nonaka in 1994, explicitly assumes that all knowledge creation processes are of a social nature and occur between human beings. Contrarily, more theories come up arguing for the presence of a new digital source of knowledge that does not require social interaction, but instead creates it from the acquisition and processing of data (Li, Xu & Zhao, 2015). These theories find confirmation in this study based on expectations made by experts. In the pilot phase, the company is building towards registering digital data, slowly creating a foundation on top of which they can build big data and advanced analytical applications on the long term. Even though the findings do not concretely conclude it, they do hint towards the need for data scientists and/or analytical software to be able to cope with increasing amounts of data. However, due the relatively underdeveloped stage in which the Fokker pilot finds itself in terms of big data no concluding remarks can be made considering what analytical tools should be used.

Real-time information and HMIs

Experts confirm the expectation based on Bartodziej (2017) that the human machine interface, or the SFV as called in the pilot, is going to play a critical role. Its functionalities of presenting digital instructions accurately and specifically per situation are in line with the findings of Monostori et al. (2016) and amplifies the ability of supporting employees on the job which will be described in more detail later.

(34)

tacit knowledge (Al-Qdah & Salim, 2013). In the pilot as it is, the SFV focuses on enhancing the sharing of explicit knowledge. However, some arguments hint that tacit knowledge can be shared through the SFV soon by means of videos and 3D models. Others strictly state that some tacit knowledge just cannot be shared without being face to face. Clearly, the SFV enables intensive interaction between organizational members. However due to the possibility of generating too much fuzz in a chat function, thought must be put in what communication lines should be created and which ones should be avoided.

Due to the relatively early stage of transforming towards a CPS, systems and technologies that operators will work with are not that complex yet at Fokker. The biggest part of the operator function will still be manual work. The most relevant change is the fact that the SFV will support them on the job which does require some training as predicted by Hirsch-Kreinsen (2016). The SFV enables operators to continuously work with and reflect on their own methods based on given instructions, enhancing internalization and externalization efforts. However, for this to be effective it is crucial that instructions are accurate and up to date. Additionally, through the SFV operators can share newly created knowledge in the chat function. This corresponds with the reasoning of Kagermann et al. (2016) who mention that employees become mediators of experience and continuous improvement experts requiring on the job support. The third characteristic of the future employee as mentioned by Kagermann is being a decision maker. This characteristic does not apply to the operators in this case situation at Fokker, who will in general remain manual workers. However, supporting staff and especially the product engineers will have to become decision makers. Due to the increasing levels of documented knowledge coming from the chat, amongst others more resources need to be available for collecting and combining explicit knowledge. Most importantly, the product engineer must decide who to share the new explicit knowledge with and through what channels for the cycle of knowledge creation to continue with the internalization process throughout the organization.

Third vs. fourth industrial revolution

(35)

The effect on externalization can be explained by the conscious awareness approach adopted from Chilton & Bloodgood (2007). If conscious awareness can be triggered by reading instructions, the presentation of explicit knowledge can be connected to the externalization process. In more detail, operators can reflect on what methods they use in their day to day activities while reading accurate and up-to-date instructions allowing them to express themselves.

The different effect on internalization is easier to explain. CPS technology clearly provides opportunities to present explicit knowledge to the operators, enabling them to gain new experiences upon. It even provides the luxury of being able to choose whether explicit knowledge should be internalized through training or instructions.

No direct empirical evidence has been found to support the statement that CPS technologies can enhance the socialization process. However, current technology has the potential to have digital twins available as video instructions or 3D models that accurately represent physical products, potentially enabling the replacement of sharing tacit knowledge through face-to-face communication.

Organizational performance

(36)

6. Conclusion

Huge amounts of available explicit knowledge will be generated when implementing CPS technologies. It will, however, take time and combined effort of individuals and technologies to achieve the potential of enhancing organizational performance through knowledge creation. Through data analysis and on the job feedback / documentation of knowledge, volumes of explicit knowledge will be generated. Resources in the shape of analysts or software are required to filter fuzz, redundancies and syntax differences structuring the explicit knowledge base. From here, it can be determined where in the organization specific explicit knowledge is applicable and through what channels it can be combined and internalized. If this has been realized, CPS technology enables enhanced internalization opportunities based on which problems can be solved proactively and new knowledge creation processes are triggered throughout the organization. This way, CPS technology will serve its purpose in terms of enhancing organizational performance.

Compared to the third revolution, CPS technology shows similar and different characteristics. Having the strongest influences on the generation of explicit knowledge is in line with the findings of Lee & Choi (2003). However, new technologies provide better opportunities in terms of using explicit knowledge for internalization and externalization purposes. Enhanced knowledge presentation and sharing technologies serve internalization through instructions and through training and education. The fact that externalization is affected, finds its roots in the conscious awareness approach.

Researching the effect of using technology to support conscious awareness and therefore externalization is new. Therefore, it could be interesting to further explore what employee characteristics drive the desire to become consciously aware in a digitalized environment. Additionally, to further explain the relationship occurring between industry 4.0 and knowledge creation, it is interesting to perform research in the future investigating other organizations in similar settings as well as organizations in entirely different industries. The need to centralize knowledge might be of less importance for companies that are smaller. Therefore, investigating the effect of CPS technology on knowledge creation in smaller companies is relevant in explaining whether findings can be generalized over companies in general or if different approaches should be used based on the size of the company.

(37)

the SECI knowledge creation processes model of Nonaka (1994) structures this study. Being a traditional point of view in terms of knowledge creation, theory and practice widely accept, approve and use this model. Additionally, employees in big organizations typically find themselves in similar environments as presented in the case study. Therefore, it is safe to generalize presented findings. A second factor that presents a limitation is the fact that the interviewed operators did not have full knowledge of changes that would come with the pilot. Therefore, they should not be seen as ‘experts’ in the comments they have made concerning the usage of a chat functionality. However, including the concerns and wishes they have spoken out during the implementation of the chat function is highly recommended.

Practical implications

(38)

References

Al-Qdah, M. S., & Salim, J. (2013). Managing tacit knowledge in MNCS and the role of ICT: Review paper. Research Journal of Applied Sciences, Engineering and Technology, 6(21), 4110– 4120.

Alavi, M., & Leidner, D. E. (2001). Review: Knowledge Management and Knowledge Management Systems: Conceptual Foundations and Research Issues. MIS Quarterly, 25(1), 107–136. Andreeva, T., & Kianto, A. (2012). Does knowledge management really matter ? Linking

knowledge management practices , competitiveness and economic performance. Journal of Knowledge Management, 16(4), 617–636.

Bartodziej, C. J. (2017). The Concept Industry 4.0 - An Empirical Analysis of Technologies and Applications in Production Logistics. Springer Fachmedien Wiesbaden GmbH.

Boon, A., Zailani, S., Iranmanesh, M., & Ramayah, T. (2015). Int . J . Production Economics Structural equation modelling on knowledge creation in Six Sigma DMAIC project and its impact on organizational performance. Intern. Journal of Production Economics, 168, 105– 117.

Chen, H., Chiang, R. H. L., & Storey, V. C. (2012). Business Intelligence and Analytics: From Big Data to Big Impact. MIS Quarterly, 36(4), 1165–1188.

Dworschak, B., & Zaiser, H. (2014). Competences for cyber-physical systems in manufacturing-First findings and scenarios. Procedia CIRP, 25(C), 345–350.

Eisenhardt, K. M. (1989). The Academy of Management Review Building Theories from Case Study Research. C Academy of Management Review, 14(4), 532–550. Retrieved from http://www.jstor.org/stable/258557

García-morales, V. J., & Martín-rojas, R. (2015). Knowledge Creation , Organizational Learning and Their Effects on Organizational Performance Performance, (July 2011).

Geisberger, E., & Broy, M. (2015). Living in a networked world - Integrated research agenda - Cyber-Physical Systems. München. Retrieved from https://books.google.nl/books?id=uaMZCAAAQBAJ&lpg=PA1&pg=PA2#v=onepage&q &f=false

Gioia, D. A., Corley, K. G., & Hamilton, A. L. (2012). Seeking Qualitative Rigor in Inductive Research. Organizational Research Methods, 16(1), 15–31.

(39)

Kagermann, H., Wahlster, W., & Helbig, J. (2013). Recommendations for implementing the strategic initiative INDUSTRIE 4.0. Acatech, (April), 4–7.

Kune, R., Konugurthi, P. K., Agarwal, A., Chillarige, R. R., & Buyya, R. (2016). The Anatomy of Big Data Computing. Software: Practice and Experience, 46(1), 79–105.

Lee, H., & Choi, B. (2003). Knowledge Management Enablers , Processes , and Organizational Performance : An Integrative View and Empirical Examination Knowledge Management Enablers , Processes , and Organizational Performance : An Integrative View and. Journal of Management Information Systems, 20(1), 179–228.

Li, S., Xu, L. Da, & Zhao, S. (2015). The internet of things: a survey. Information Systems Frontiers,

17(2), 243–259.

Miles, M. B., Huberman, A. M., & Saldaña, J. (1994). Drawing and verifying conclusions. In

Qualitative Data Analysis (pp. 275–322).

Monostori, L., Kádár, B., Bauernhansl, T., Kondoh, S., Kumara, S., Reinhart, G., … Ueda, K. (2016). Cyber-physical systems in manufacturing. CIRP Annals - Manufacturing Technology, 65(2), 621–641.

Nonaka, I. (1994). A Dynamic Theory of Organizational Knowledge Creation. Organization Science, 5(1), 14–37.

Nonaka, I., & von Krogh, G. (2009). Perspective—Tacit Knowledge and Knowledge Conversion: Controversy and Advancement in Organizational Knowledge Creation Theory. Organization Science, 20(3), 635–652.

Rabionet, S. E. (2011). How I Learned to Design and Conduct Semi-structured Interviews : An Ongoing and Continuous Journey. The Qualitative Report, 16(2), 563–566.

Stock, T., & Seliger, G. (2016). Opportunities of Sustainable Manufacturing in Industry 4 . 0. In

Procedia CIRP (Vol. 40, pp. 536–541). Elsevier B.V.

Sumbal, M. S., Tsui, E., & See-to, E. W. K. (2017). Interrelationship between big data and knowledge management : an exploratory study in the oil and gas sector. Journal of Knowledge Management, 21(1), 180–196.

Uden, L., & He, W. (2017). How the Internet of Things can help knowledge management: a case study from the automotive domain. Journal of Knowledge Management, 21(1), 57–70. Voss, C., Tsikriktsis, N., & Frohlich, M. (2002). Case research in operations management.

International Journal of Operations & Production Management, 22(9), 195–219.

(40)

Revolution That Will Transform Supply Chain Design and Management. Journal of Business Logistics, 34(2), 77–84.

Yin, R. K. (1994). Case Study Research. Design and Methods. Sage publications.

Zuboff, S. (1988). In the age of the smart machine : the future of work and power. Basic Books.

Retrieved from

(41)

(42)

Appendix B – Operationalization of tacit and explicit

knowledge

Conceptual definition tacit knowledge: (Takeuchi, 2006)

Not easily visible and expressible, highly personal, hard to formalize and difficult to share with others (such as subjective insights, intuitions and hunches). These are deeply rooted in an individual's actions, and experiences, as well as in the ideals, beliefs, values or emotions a person embraces.

Conceptual definition explicit knowledge: (Takeuchi, 2006)

Formal and systematic, expressed in words and numbers, easily communicated and shared. Anything digital, anything that easily can be processed by a computer or transmitted electronically.

Operationalization tacit & explicit knowledge: (Chilton & Bloodgood, 2007)

Tacit Knowledge

component

Explicit

• Not consciously aware

• Automatic use of knowledge base

• Unable to fully explain behavior

Conscious awareness • Consciously think through the steps required to apply knowledge

• Able to fully explain behavior

• Difficult to express

• Difficult to codify

• Difficult to directly communicate

Expressability (Written & oral)

• Easy to express

• Easy to codify

• Easy to directly communicate

• High ability to perform task based on seeing the activity or final outcome

Demonstrability • Requiring detailed information to perform a task

• Logic appears to be missing in an informal approach, because steps are already learned. (prior learning)

• Reliance on prior learning through experience, therefore committing to the

subconscious

• Low ability to recall steps used in performing specific tasks

Formal or informal • Logical step-wise formal approach, currently learning.

• Reliance on current learning, using specific, concrete steps.

• High ability to recall steps used in performing specific tasks

(43)

Based on Chilton & Bloodgood (2007) socialization is seen as the process of gaining tacit knowledge through experience.

Based on Chilton & Bloodgood (2007) externalization is seen as the process of transforming tacit knowledge into explicit knowledge by creating conscious awareness. Being consciously aware allows it to be able to express, codify and communicate the tacit knowledge, turning it into explicit knowledge.

Based on Chilton & Bloodgood (2007) combination is seen as the process of collecting and combining explicit knowledge resulting in new, logical, concrete and easy expressible explicit knowledge.

Knowledge Creation in the era of Cyber-Physical Systems: a Case Study