,
STOCKHOLM SWEDEN 2018
Industry 4.0 from a
technology adoption
perspective
A case study at Sandvik Coromant
EMIL WINBERG
JESPER AHRÉN
KTH ROYAL INSTITUTE OF TECHNOLOGY
T
HE
R
OYAL
I
NSTITUTE
OF
T
ECHNOLOGY
MASTER THESIS IN INDUSTRIAL PRODUCTION
Industry 4.0 from a technology adoption perspective
A case study at Sandvik Coromant
June 13, 2018
Master of Science Thesis
(IV) I
Sammanfattning
Produktionsindustrin står just nu inför den fjärde industrirevolutionen där en ökad grad av anslutning och datastyrd produktion har möjligheten att skapa självoptimerande fabriker där maskiner och system kommunicerar automatiskt i realtid. Införandet av ny teknik kan skapa stora förändringar men även konkurrensfördelar för organisationer, vilket även är fallet för Industri 4.0. Syftet med studien var att identifiera vilka faktorer som påverkar införandet av Industri 4.0 hos tillverkande företag och hur problem inom automatiserad cellproduktion kan reduceras genom att introducera Industri 4.0 koncept. Studien utfördes som en fallstudie hos Sandvik Coromant där interna observationer och intervjuer utfördes. Dessutom intervjuades fem externa organisationer verksamma inom industriell digitalisering.
Studien visade att det finns olika faktorer som påverkar införandet av Industri 4.0, kategoriserade i tekniska, organisatoriska och marknadsmässiga faktorer. För de tekniska faktorerna har småskaliga applikationer, ökad transparens genom anslutning av enheter och en ökad integration av olika informationssystem en positiv effekt på införandet av industri 4.0. För att möjliggöra införandet i organisatoriska sammanhang måste produktionsorganisationer skaffa digitala kompetenser, integrera sin IT-organisation i sin produktion samt förändra sin kultur och inställning till Industri 4.0. Dessutom är standardisering, skapande av digitala ekosystem och IT-säkerhet de viktigaste marknadsaspekterna som påverkar införandet av Industri 4.0.
Hos Sandvik Coromant har elva problem identifierats som kan reduceras med införandet av koncept från Industri 4.0. Studien föreslår att anslutning, visualisering och dataanalys används för att reducera dessa problem.
Nyckelord: Adoption av teknik, Industri 4.0, CPS, IoT, Automatiserade produktionsceller Examensarbete Industriell Produktion 2018
Industry 4.0 ur ett
teknikadoptionsperspektiv
En fallstudie på Sandvik Coromant
Jesper Ahrén Emil Winberg Godkänt 2018-06-13 Examinator Lihui Wang Handledare
Abdullah Alhusin Alkhdur
Uppdragsgivare
Sandvik Coromant
Kontaktperson
(IV) II
Master of Science Thesis Industrial Production 2018
Industry 4.0 from a technology adoption
perspective
A case study at Sandvik Coromant Jesper Ahrén Emil Winberg Approved 2018-06-13 Examiner Lihui Wang Supervisor
Abdullah Alhusin Alkhdur
Commissioner
Sandvik Coromant
Contact person
Michael A Sköld
Abstract
The fourth industrial revolution is emerging, where connection and data driven production has the potential to create self-optimizing factories, in which machines and systems can communicate in real-time. However, adopting new technologies can impose big changes but also create competitive advantages for organisations, which is certainly the case of Industry 4.0.
The purpose of the study was to identify what main factors that affects the adoption of Industry 4.0 for production organisations and how problems in automated production cells could be reduced by introducing Industry 4.0 concepts. The study was performed as a case study at Sandvik Coromant, where observations and interviews were conducted. In addition, five external organisations specialized in industrial digitalization were interviewed.
The study found that there are various factors affecting the adoption of Industry 4.0 categorized into technological, organisational and external/environmental factors. In terms of technology, small scale applications, increased transparency through connection and an increased integration of information systems have positive effect on the adoption of Industry 4.0. In organisational context, production organisations must acquire digital competence, integrate their IT organisation into their production and change the culture and attitude towards the adoption of Industry 4.0. Furthermore, standardization, creation of ecosystem and IT security are the main external/environmental aspects which affect Industry 4.0 adoption.
At Sandvik Coromant, eleven problems were identified which has the potential to be reduced by implementing concepts of Industry 4.0. The study proposes use of connectivity, visualization, data analysis to reduce these problems.
(IV) III
Foreword
This study was conducted as a master thesis at the Royal Institute of Technology in the spring of 2018 and comprises of 30 academic credits. The case study was performed at Sandvik Coromant in Gimo.
The report has been published in two identical copies, one published by the Department of Industrial Economics and Management, and one published by the Department of Production Engineering.
Acknowledgements
A special thanks and gratitude towards our commissioner Sandvik Coromant and in particular our supervisors, Michael A Sköld and Simon Åkerblad. Their experience and enthusiastic support has guided us in the right direction and truly improved the quality of the study.
We would also like to thank our academic supervisors at the Royal Institute of Technology, Anna Jerbrant, for her patient support and many encouraging discussions, as well as Abdullah Alhusin Alkhdur, for providing valuable input and criticism reviewing our work.
Furthermore, this thesis would not have been possible without the contribution of all interviewees. Therefore, we would like to thank the operators at Sandvik Coromant in Gimo, employees at Sandvik Coromant and the external interviewees who willingly contributed with their time and experience.
Lastly, we would like to thank our families and friends for their support and understanding throughout the research process.
The Royal Institute of Technology Stockholm, June 2018
(IV) IV
Abbreviation
AI Artificial Intelligence
BMBF Bundesministerium für Bildung und Forschung (German federal ministry of education)
CAD Computer-Aided Design
CAM Computer Aided Manufacturing
CODE Center of Digital Excellence
CPS Cyber-Physical System
CPPS Cyber-Physical Production System
CNC Computer Numerical Control
CMM Coordinate-Measuring Machine
ERP Enterprise Resource Planning
GH Gimo Hårda (The production unit for cutting inserts in Gimo)
GV Gimo Verktyg (The production unit for tools in Gimo)
GVP Gimo Verktyg Produktion (Gimo tool production)
H2H Human to Human communication
HMI Human Machine Interaction
IoT Internet of Things
KPIs Key Performance Indicators
M2H Machine to Human communication
M2M Machine to machine communication
MES Manufacturing Execution System
MRP Material Requirements Planning
PLC Programmable Logic Controller
SMS Sandvik machining solutions
TAM Technology-Acceptance Model
TOE Technology-Organisation-Environment framework
UX User Experience
VP Visual Planning
Table of Contents
1 INTRODUCTION ... 1 1.1 BACKGROUND ... 1 1.2 PROBLEM FORMULATION ... 2 1.3 PURPOSE ... 3 1.4 RESEARCH QUESTIONS ... 3 1.5 DELIMITATIONS ... 3 1.6 EXPECTED CONTRIBUTION ... 4 1.7 DISPOSITION ... 4 2 METHODOLOGY ... 5 2.1 RESEARCH APPROACH ... 5 2.2 RESEARCH DESIGN ... 62.3 DATA GATHERING METHODS ... 7
2.4 DATA ANALYSIS ... 11
2.5 RESEARCH QUALITY ... 12
3 LITERATURE STUDY ... 14
3.1 INTRODUCTION TO INDUSTRY 4.0 ... 14
3.2 CONCEPTS OF INDUSTRY 4.0 ... 18
3.3 RISKS AND REQUIREMENTS WITH INDUSTRY 4.0 ... 23
4 THEORETICAL FRAMING ... 26
4.1 TECHNOLOGY ADOPTION IN ORGANISATIONS ... 26
4.2 TECHNOLOGY ACCEPTANCE IN THE CONTEXT OF INDUSTRY 4.0 ... 30
5 SANDVIK COROMANT ... 31
5.1 COMPANY OVERVIEW ... 31
5.2 GIMO VERKTYG ... 32
5.3 AUTOMATED PRODUCTION CELLS ... 33
6 RESULTS AND ANALYSIS ... 44
6.1 PROBLEMS IN AUTOMATED PRODUCTION CELLS ... 44
6.2 TECHNOLOGICAL FACTORS AFFECTING INDUSTRY 4.0 ADOPTION ... 58
6.3 ORGANISATIONAL FACTORS AFFECTING INDUSTRY 4.0 ADOPTION ... 68
6.4 ENVIRONMENTAL FACTORS AFFECTING INDUSTRY 4.0 ADOPTION ... 77
6.5 ALIGNMENT BETWEEN PROBLEMS AND FACTORS AFFECTING ADOPTION ... 83
7 CONCLUSION ... 85
7.1 ANSWERING THE RESEARCH QUESTIONS ... 85
7.2 SUSTAINABILITY IMPLICATIONS ... 88
7.3 FUTURE WORK ... 89
LIST OF REFERENCES ... 91
Table of Figures
FIGURE 1.RESEARCH APPROACH. ... 5
FIGURE 2.PROJECT ACTIVITIES AND TIMELINE. ... 6
FIGURE 3.LITERATURE STUDY STRUCTURE. ... 14
FIGURE 4.THE FOURTH INDUSTRIAL REVOLUTIONS (ROSER,2015). ... 15
FIGURE 5.DEFINITION OF INDUSTRY 4.0(LEAPAUSTRALIA,2017). ... 15
FIGURE 6.CPSFRAMEWORK (LEE, ET AL.,2015). ... 17
FIGURE 7.MAIN INDUSTRY 4.0 CONCEPTS. ... 18
FIGURE 8.STAGES OF A BUSINESS ECOSYSTEM (CITED FROM JAMES F.MOORE (1993)). ... 29
FIGURE 9.SANDVIK GROUP:ORGANISATIONAL STRUCTURE. ... 32
FIGURE 10.MILLING, TURNING AND DRILLING TOOLS. ... 32
FIGURE 11.GV FACTORY LAYOUT. ... 33
FIGURE 12.CONCEPTUAL VIEW OF AN AUTOMATED PRODUCTION CELL. ... 34
FIGURE 13.WORK PROCEDURE STEPS. ... 34
FIGURE 14.OVERVIEW OF AN AUTOMATED PRODUCTION CELL. ... 39
FIGURE 15.AUTOMATED PRODUCTION CELL, INCLUDING:(1) WAGONS,(2) PALETTES,(3) ROBOT,(4)CNC AND (5) FIXTURE SHELF. ... 40
FIGURE 16.CELL MANAGER SOFTWARE HOME PAGE. ... 41
FIGURE 17.CELL MANAGER SOFTWARE VIEW FOR ORDER MANAGEMENT. ... 42
Table of Tables
TABLE 1.DISPOSITION OF THE REPORT WITH DESCRIPTIONS TO EACH CHAPTER. ... 4TABLE 2.KEYWORDS USED FOR OBTAINING LITERATURE. ... 7
TABLE 3.INTERVIEWS WITH PRODUCTION STAFF. ... 9
TABLE 4.INTERVIEWS WITH NON-OPERATIVE EMPLOYEES AT SANDVIK COROMANT. ... 10
TABLE 5.INTERVIEWS WITH EXTERNAL ORGANISATIONS. ... 10
(96) 1
1 Introduction
In this chapter, the background of the research is presented, followed by the problem formulation, purpose and research questions. Finally, limitations, expected contribution as well as report disposition is presented.
1.1 Background
Introduction of new technologies can enable competitive advantages for the organisations which manages to adopt these technologies, but it can also create big changes in the market and the way businesses is managed. This includes changes in both internal operations and customer offerings, consequently pressuring companies to adopt to these new technologies (Lanzolla & Suarez, 2012). Companies that fails to adopt new emerging technologies may be overtaken by more successful competitors. Tornatzky & Fleischer (1990) found that technological adoption is affected by more than just the technology itself, and that an organisation should consider other aspects as well (Baker, 2011). To describe technological adoption in organisations, Tornatzky & Fleischer (1990) created the Technology-Organisation-Environmental Framework which acknowledges that technology adoption needs to consider not only the technological context itself, but also the organisational and environmental context (Baker, 2011). In the case when a new technology is intended to be used by actors within an organisation, it is not sufficient for the organisation to only adopt the technology. Lanzolla & Suarez (2012) suggest that organisations also must make sure that the adopted technologies are accepted by the intended users. If not, organisations can risk making unnecessary efforts and investments in technologies that will eventually be perceived as redundant by its indented users (Lanzolla & Suarez, 2012). To enable technology acceptance, the Technology acceptance model (TAM model) can be used to explain how technologies are accepted by users. It states that acceptance is dependent on two main characteristics: perceived usefulness and perceived ease of use (Davis, 1989).
The production industry has historically faced many introductions of technology, the most important are categorized into the four industrial revolutions. These revolutions have enabled technological advancements and industrialization in many countries, leading up to today’s production systems (Liao, et al., 2017). The first industrial revolution started in the later part of the 18th century with the introduction of machines and mechanical production. In the beginning of the 20th century, the second industrial revolution took place and introduced the production line and electrical powered mass production. The third industrial revolution, which is still ongoing, started in the 1970s and includes the application of information technology and electronics in production system to increase the level of production automation. The next industrial revolution, Industry 4.0, is highly driven by internet, digitalisation and technology to enable a far smarter and connected industrial era (Kagermann, et al., 2013). While the fourth industrial have not yet been fully executed, it is still expected to create significant change in existing industries and organisations. Many countries have developed strategic initiatives and allocated billions of euros to support the digitalisation and transition towards Industry 4.0. This includes countries such as Germany, Japan, US, UK, Singapore and Sweden (Swedish Government Office, 2017) as well as the European Union (Liao, et al., 2017) to name a few.
(96) 2
has great potential in terms of reducing cost and improving efficiency. In a survey with 235 European manufacturers, 36 % believe that Industry 4.0 will increase efficiency between 11-20 % in 5 years while 37 % expect the efficiency to increase with over 20 %. In the same survey, 35 % of the production companies expect that Industry 4.0 will reduce cost with 11-20 % while 21 % expect costs to be reduced by over 20 %. Some of the most promising areas of Industry 4.0 are better control and planning of production processes, improved quality and increased flexibility by enabling data analysis, information exchange and use of real-time data (Geissbauer, et al., 2014). Clearly, adopting to Industry 4.0 have the possibility to enable huge advantages in terms of quality, cost benefits, flexibility and product customization if adopted properly (Baheti & Gill, 2011; MacDougall, 2014).
An enabler of Industry 4.0 is Internet of Things (IoT), which is a concept on the connection and information sharing of objects such as machines, equipment, sensors and other objects (Dorsemaine, et al., 2015). IoT in an industrial setting enables the connection of the physical and digital world, integrating them even further and creating a Cyber-Physical System (CPS). CPS offers the physical entities of a system to be represented in a virtual setting, where changes in the physical system affect the virtual equivalent and conversely (Baheti & Gill, 2011). For production processes, CPS enables the development of smart factory capabilities such as real-time & predictive control, learning ability, self-optimization and decentralized decision making (MacDougall, 2014; Tantik & Anderl, 2016; Pai, et al., 2018; Lu, 2017). The use of IoT and CPS in production creates vast amount of data and information, which has the potential to be processed and analysed to provide insight and decision support to humans in the organisation. The handling and analysis of big amount of data is referred to as Big Data. However, to be able to bring the full potential of Big Data, new capabilities may be required, and high quality of the data is essential (Yi, et al., 2014; Demchenko, et al., 2013)
Even though Industry 4.0 may increase the usage of machines that can process information and independently act upon it, thanks to smart communication and data analytics, humans is expected to be a central part of the future production systems (Lu, 2017; Chen, et al., 2017). By designing the production systems around individuals, strong human capabilities such as creativity and flexibility can be enhanced, while the amount of monotonous and redundant tasks can be reduced. Thus, data analysis and connectivity through IoT and CPS can be used to deliver information or decision support to operators in the right way, at the right time and to the right person (Gorecky, et al., 2014).
To summarize, in the context of Industry 4.0, IoT and CPS have the possibility to enable an increased connectivity and information sharing. At the same time, data analysis enables this information flow to be processed to provide insights and decision support. Finally, humans will need to act on many of these insights and interact with machines and systems, which requires human-machine interaction. These concepts, connection of the physical world, analysis of data and human-machine interaction are the focus areas of Industry 4.0 in this report.
1.2 Problem formulation
(96) 3
In the process of adoption new technologies, there are a lot of factors which will have an impact on the successfulness of such processes. Adoption of Industry 4.0 in production companies can be considered as the adoption of a bundle of technologies and concepts rather than just one technology. Furthermore, while the potential of Industry 4.0 is perceived as high, many organisations are not ready to implement Industry 4.0 and struggles to identify more specifically how Industry 4.0 can provide value to their production processes. Thus, there is a need to identify more in detail which factors production companies must consider to better adopt to Industry 4.0, but also to identify how production processes can be improved specifically by adopting Industry 4.0 concepts. Furthermore, to reduce investment costs and to make use of current production facilities, identification of current production processes may provide guidance regarding what concepts will be most suitable to introduce in the existing processes.
Sandvik Coromant’s plant in Gimo, is their biggest production plant for cutting tools (Sandvik, 2017). With a high degree of mechanical automation and use of IT systems, Sandvik Coromant has acquired many of the benefits of the third industrial revolution. The machine operators are however still responsible for process monitoring, order planning and quality assurance in which there are little guidance from the production systems. Thus, there is a need to investigate if and how Industry 4.0 has the potential to improve current production processes and support the machine operators. Furthermore, it is also important for Sandvik Coromant to understand which factors will affect and benefit the adoption of Industry 4.0.
1.3 Purpose
The purpose of this study is to identify areas in Sandvik Coromant’s automated production cells where Industry 4.0 can be adopted in order to increase the support and solve problems for the machine operators. Further, the purpose is also to identify what factors that will affect adoption of Industry 4.0 in production organizations and how they should be considered in order to enable adoption of Industry 4.0.
1.4 Research questions
To fulfil the purpose and structure the research, two research questions have been formulated:
RQ 1: What is the current state of Sandvik Coromant’s automated production cells
and how can Industry 4.0 concepts be used to reduce problems expressed by production staff?
RQ 2: What are the main factors affecting the adoption of Industry 4.0 in production
organisations?
1.5 Delimitations
In order to reach sufficient depth in the research within the time available, several delimitations were made, all of which are stated below.
(96) 4
The production unit, where the study is conducted and to which recommendations for future needs are provided, is delimited to on one specific type of automated production cell (milling cell). This is the most common one and has the most homogeneous layout.
1.6 Expected contribution
This research provides a comprehensive investigation of the Technology-Organisation-Environment framework (TOE framework) in the context of Industry 4.0, which contributes both to research and companies by identifying factors that are perceived as most important by various stakeholders in the production industry.
Furthermore, this research aims to provide an empirical contribution by concretizing Industry 4.0 applications in automated production cells. This may provide guidance to which ways Industry 4.0 can be introduced in factories, reducing the uncertainty around Industry 4.0 investments.
Lastly, the research is expected to contribute to the case company by identifying problems in the current automated production cells and present how Industry 4.0 can be introduced to eliminate these problems and increase operator support. Thus, this study highlights the important areas which future solutions should focus on solving.
1.7 Disposition
The report consists of six chapters which are presented in the following table (Table 1). Table 1. Disposition of the report with descriptions to each chapter.
Chapter
Description
Introduction
An introductory chapter presenting the background and the aim of the study. Aims to give an understanding of the topic and purpose of the study.
Methodology This chapter gives the reader the possibility to critically evaluate the chosen methods and assess the quality of the findings.
Literature study In this chapter, technology adoption in organisations is presented, followed by a presentation of Industry 4.0 and its applications and risks.
Sandvik Coromant Presentation of the case company, mapping of the current state. Lastly, the problems are presented and analysed using concepts from Industry 4.0.
Factors affecting adoption of Industry 4.0
(96) 5
2 Methodology
In this chapter, the research approach and design are presented, followed by data gathering methods and discussion regarding research quality and ethics.
2.1 Research approach
Blomkvist & Hallin (2015) recommends an exploratory research approach when the purpose is to find underlying relationships and characteristics of a newly researched phenomena. Especially when the research aims to answer the questions of “how” and “why”. The purpose of this study is to identify what factors affect the adoption of Industry 4.0 and how these should be considered, but also to identify if and how Industry 4.0 can be adopted into automated production cells at Sandvik Coromant. Due to these characteristics and uncertainty regarding potential findings, a study with an exploratory research approach was identified as most appropriate.
For the use of literature and theories, there are two strategies; either theory is used to formulate ideas which is tested during the study or it is used to understand the empirical findings (Blomkvist & Hallin, 2015). The first is referred to as deduction while the second strategy is referred to as induction. The later has the drawbacks that it can be time consuming, since theories might lose their relevance during the research progress. However, the benefits of induction are that it enables an exploratory approach and the ability to be flexible during the process. Due to the exploratory approach, the use of induction reduces the risk that priory selected theories and literature delimits the research as theories can be evaluated and replaced continuously (Blomkvist & Hallin, 2015). Therefore, the exploratory approach was combined with an inductive use of theory.
While using an inductive and exploratory approach, it is important that theory and literature is reviewed continuously to be relevant for the gathered data (Blomkvist & Hallin, 2015). Therefore, the literature study, the empirical data gathering, and analysis were conducted simultaneously throughout the project. This makes it possible to narrow the research focus and align theory with the empirical material (See Figure 1).
Figure 1. Research approach.
2.1.1 Case study
(96) 6
insight, since it allows many kinds of data gathering methods (Denscombe, 2009). Because there was a need to obtain understanding of Industry 4.0 adoption in the practical setting of production and automated production cells, it was necessary to have access to data from a variety of stakeholders. Firstly, to define the current cell production and its problems, access to the automated production cells and interviewing its stakeholders was needed. Secondly, to get a deeper understanding about Industry 4.0 and what factors that affects adoption, there was a need to access other stakeholders within the organisation. These stakeholders include managers, IT personnel, project leaders etc. Lastly, to be able to strengthen the findings within the company, and to be able to contribute to the case organisation, complementary data sources outside the organisations were identified. Not only will this approach strengthen the findings within the organisation by triangulation (Denscombe, 2009), it will also enable more perspectives to be considered.
The case study approach is suitable for two reasons. Firstly, much of the data required is easily obtained within the case company. Lastly, the accessibility to external organisation increased as the case company could help arranging these interviews.
2.2 Research design
The case study was divided into two parts: pre-study and main study. The pre-study was broader and more unstructured as it allowed for identifying the problem and finding the best areas of focus for the main study, which was more focused and structured. All activities and their duration are presented in Figure 2.
Figure 2. Project activities and timeline.
2.2.1 Pre-study
(96) 7
2.2.2 Main study
The main study was initiated with internal interviews allowing to further understand the processes and dynamics of the case company. These interviews were used to map the current state of the automated production cells and identify problems in the automated production cells. Once a more detailed picture of the case object was obtained, external interviews were conducted which made it possible to study the adoption of Industry 4.0 from various perspectives. This is especially useful in social science, where different perspectives can enable a better understanding of the processes and their relations (Blomkvist & Hallin, 2015; Denscombe, 2009). Another benefit of using different sources of data is that it allows us to triangulate the findings, which can strengthen the validity of the result (Blomkvist & Hallin, 2015; Denscombe, 2009). Since an exploratory approach is used when there was uncertainty about potential findings, the focus and problematising of the research can evolve and change throughout the research (Blomkvist & Hallin, 2015). To enable adoption to unexpected changes, the internal and external interviews where integrated and overlapped in time.
2.3 Data gathering methods
Many sources for gathering of data were used, including literature review, interviews and observations. These methods are presented in detail below.
2.3.1 Literature study
The literature study was conducted for three reasons. Firstly, to collect knowledge and know-how within the relevant areas to find answers to the research questions, this includes Industry 4.0 as well as technology adoption and acceptance literature. Secondly, to acquire knowledge on how these areas in theory should be implemented and adopted to take full advantage of the concepts and technologies. The aim was to compare the findings with the specific case in this study. The third and final purpose was to get a perception of the knowledge and research that already exist in this area and how this study could contribute to previously made research.
The search for literature have been performed mainly in KTH’s library database, Primo, that has access to all journals and publishers that KTH has agreement with. As a complement, searches in Scopus, Google scholar and Google have been conducted. All literature has been obtained using the keywords defined in Table 2 below.
Table 2. Keywords used for obtaining literature.
Area of study Used keywords
Technology adoption “Technology adoption”, “Technology acceptance”, “TOE”, “TAM”, “Business ecosystem”
(96) 8
2.3.1.1 Source criticism
The literature study was based on four types of literatures: academic articles, books, reports from consulting firms and governments as well as internal documentation (e.g. manuals and system specifications). Depending on the type of literature, the credibility and independence will differ, which is why sources has been evaluated to see if they have any other agendas to express their opinions. Primarily, only literature that has a high credibility was used. In case of biased sources, these were excluded or discussed in the report by the authors. The important aspect was that this always was considered when choosing and discussing sources.
The base of the literature study has been academic articles. They have been used to find theoretical findings related to the research questions and to define concepts relevant for the study. Articles from academic journals have the advantage of being subjected to peer review (Denscombe, 2009). This is however no quality guarantee and the study has therefore considered additional factors. Denscombe (2009) describes five factors for evaluating the credibility of the articles: journal age, status in research community, credibility of editors, if the journal has a specific focus when selecting its articles (e.g. a national focus), if the journal is publishing on the behalf of foundations and if the article using clear references to other quality articles. These criteria have been used continuously through the project to evaluate the used articles. Articles published in conference papers can be of lower quality, therefore peer-reviewed articles were used primarily to ensure sufficient quality. Books have been used to give a more holistic picture of some of the concepts described in the literature study. Since the Industry 4.0 area is under rapid development, only recently published books have been used. Books have many of the same advantages as academic articles if they have been reviewed before being published. To ensure reliability, only reviewed books were used. Reports from governments and consulting firms has been used for finding examples of areas that can be improved by Industry 4.0. Since the term Industry 4.0 was created by the German government to identify how their industry should be developed, their reports and publications has also been used. Reports from governments are generally high in credibility since they often are results from long projects with experts from industry and the academy involved (Denscombe, 2009). Reports from consulting firms and other companies may however have less credibility. The firms and companies that writes these reports are normally also the ones that offer solutions or expert advice in the subject to other companies. Thereby, there is a risk that they want to describe new concepts and their applications as more promising than they are. Therefore, triangulation between consultancy reports and academic articles has been used to ensure validity of the consultancy report. The reason for including consultancy reports is mainly that they are less theoretical, contributing regarding how Industry 4.0 can be realized more specifically.
Internal documentation has been used to evaluate the current state to answer the first research question together with the observations described in Section 2.3.3. By doing so, the result can be triangulated to strengthen the draw conclusions.
2.3.2 Qualitative interviews
(96) 9
due to its simple nature. Unstructured interviews are opposite in their nature as the interviewer interferes as little as possible by only introducing a subject and letting the respondent elaborate freely. This enables contextualization but enables little control of the respondent discussing relevant areas. Semi structured interviews use themes that aims to be covered during the interview and allows for flexibility and in-depth answers, but also greater possibility to compare and relate answers from different interviews (Denscombe, 2009). The interviews in this study was conducted as semi-structured interviews as it would allow both flexibility in the interview setting and enable use of common themes to relate different interviews to each other. This gives a good combination of structure by using themes, but also contextualization by letting the interviewee explain and focus on central aspects as they appear.
The interviews are divided into two categories: internal interviews and external interviews. The internal interviews were conducted with production staff, including both operators and production leaders, to get information about the current state, its problem and their attitude towards Industry 4.0. Furthermore, internal interviewees were also conducted with other types of stakeholders, such as change leaders, production engineers, IT staff, project leaders and mangers. This enabled us to receive information about the current state and its problem from an additional perspective, but also to investigate factors that may affect Industry 4.0 adoption. The external interviewees included actors from other organisations (including a Sandvik Group spin-off organisation) which are specialized in industrial digitalization. In these interviews, factors affecting adoption of Industry 4.0 were discussed. The categories are presented below.
2.3.2.1 Internal interviews
The internal interviewees were conducted with production staff such as production leaders and operators but also with other stakeholders who do not work on the shop floor. The two categories of internal interviewees are presented further below.
The first category of internal interviews consists of 22 interviews and was held with operators and production leaders. In these interviews, current state and problems of the automated production cells were discussed in combination with the perception and attitude towards adoption of new technologies. The interviews were conducted in pairs where one of the interviewees took notes while the other asked questions and discussed the central themes described earlier. For an interview overview, see Table 3. The result from these interviews was used to map the current state of automated production cells, identify its central problem and identify factors affecting attitude towards new technologies.
Table 3. Interviews with production staff.
Category No. of interviews Area discussed
Operators 18 Current state, problems and technology adoption Production leaders 4 Current state, problems and technology adoption
(96) 10
Table 4. Interviews with non-operative employees at Sandvik Coromant.
Name Role Location Date
Interviewee 1 Maintenance engineer Gimo GV 2017-02-16
Interviewee 2 Maintenance Engineer Gimo GH 2018-02-21
Interviewee 3 Change Leader Gimo GV 2018-02-27
Interviewee 4 Chief Architect Sandvik HQ 2018-03-05
Interviewee 5 Senior IT developer Gimo GT 2018-03-07
Interviewee 6 IT Developer Gimo IT 2018-03-07
Interviewee 7 Project Manager (Industry 4.0) Gimo GH 2018-03-13
Interviewee 8 Technical Manager Gimo GV 2018-03-20
2.3.2.2 External interviews
The external interviews were conducted with five external actors; Swerea IVF, Siemens, ÅF Digital Business, Roland Berger and Sandvik Center of Digital Excellence (CODE) (for a description of the companies see Appendix B). All of these work with strategy development, implementation and design of digitalisation concepts and services within the production industry. The interviewed people from these companies have vast experience in industrial change and digitalisation applied in production. CODE and Sandvik Coromant is part of the same division in the Sandvik Group, however, CODE is an independent organisation within the Sandvik Machining Solution (SMS) organisation, supporting all different brands. CODE also develops services and business towards SMS customer, compatible with both Sandvik Coromant and their competitor’s services. Thus, in this context, CODE is considered as external. For a detailed description about the external companies interviewed, see Table 5.
Table 5. Interviews with external organisations.
Name Company Department Role Date
Interviewee 9 Siemens Digital Factory Business Area Manager 2018-03-09 Interviewee 10 Siemens Plant Data Services Sales Manager 2018-03-12 Interviewee 11 Swerea IVF Digitalization Project Manager 2018-03-15 Interviewee 12 Sandvik Group CODE Innovation Manager 2018-03-19 Interviewee 13 Sandvik Group CODE President 2018-03-27 Interviewee 14 ÅF Digital Business Business Area Manager 2018-04-10 Interviewee 15 Sandvik Group CODE CTO 2018-04-12 Interviewee 16 Roland Berger - Consultant 2018-04-17
(96) 11
2.3.3 Observations
There was a need to fully understand the processes and identify the different parts of the cells in its natural setting, since the research aims was related to explore Industry 4.0 in relation to automated production cells. For this reason, interviews with operators and production leaders were complemented by observations. This choice is motivated by Denscombe (2009) who views observations as particularly useful when there is a need to maintain the natural dynamics of a situation. To ensure efficient documentation during the observations, they were always conducted in pairs, where one of the observers took notes. According to Denscombe (2009), taking notes can disrupt the natural setting, especially if the identify and objective of the observer is unknown. However, as the operators were informed for ethical reason, it was argued that the benefits of proper documentation were greater. Details about the conducted observations is presented in Appendix A table II.
Observations were conducted in two rounds (pre-study and main study), where in the first round, the objective was to get an overview of the current state. This approach is in line with Denscombe’s (2009) view, who suggest that an initial field study consisting of observations should have a broader, less focused character and be used as a foundation for more focused observations. To get a thorough overview and not risking missing any tasks or important events, the observations lasted for 8 hours each. The first round of observations indicated that a cell with a greater number of processes and interactions would make the analysis less dependent on the choice of cell type. Therefore, only automated production cells that performs milling operations were selected as focus for the study.
The second round of observations was conducted at automated production cells with a similar setup: calibration, milling and measuring. The observations were conducted continuously throughout the project, often in connection with interviews with the operators. Identification of the current state was primarily identified in the first round, while the second round was used to complement missing parts of the initial mapping. The observations also allowed to identify problems expressed by operators. During the second round of observations the aim was to direct focus on central issues and processes as they were identified and recommended by Denscombe (2009).
2.4 Data analysis
Gathered data cannot be presented in its original state when conducting qualitative research, it must be processed and analysed (Denscombe, 2009). The way that data is analysed can be conducted in different ways, however. Denscombe (2009) has identified main steps that is included in the process of analysing data. These are: preparation, familiarization, interpretation (coding, categorization etc.), verification and presentation. All these steps were used when analysing the data gathered in this study and is presented below.
(96) 12
categories each. In the verification phase, different types of sources were compared to identify common views on subjects, but also contradictive opinions. In the presentation phase, the data was presented in combination with the analysis, making it more integrated.
2.5 Research quality
One of the main drawbacks of using a qualitative case study as a method is that it is highly dependent on context and the specific situation (Saunders, et al., 2012). When conducting any kind of scientific research, it is important to make sure that the result is correct and trustworthy. Therefore, it is important that the result can be verified in some way to ensure high research qualities (Denscombe, 2009). The three main concepts of research quality are presented in the below section. These are; validity, generalizability and reliability (Denscombe, 2009). It is hard to control these aspects of research quality in qualitative studies, as it in many cases is impossible to repeat the research. This puts extra weight on the presentation regarding how these concepts have been handled in the report (Denscombe, 2009).
2.5.1 Validity
Validity is the parameter measuring whether the chosen theory and methods are suitable for what is being studied. Validity in scientific research means that all parts of the research is related to the phenomena that is being investigated. By assuring that the literature study, sources, data gathering methods and discussion are related and contextualized with the research purpose and research questions, validity can be increased (Blomkvist & Hallin, 2015). A source can have high reliability and trustworthiness, but if it is being used in the wrong context, the validity can still be low. This means that it is important to motivate why methods and data sources are suitable for the specific study (Denscombe, 2009).
To ensure that the empirical data are valid, data triangulation can be used. Data triangulation is no mean to ensure validity, it is rather a method to increase trustworthiness of one source by comparing it to others, both in time and in context (Denscombe, 2009). In this study, triangulation was done by complementing internal interviews with eight external interviews. This enabled confirmation of the empirical result, which consequently increases the validity (Denscombe, 2009). Secondly, data triangulation was used to gather data from sources with different perspectives and contexts. The risk that any result was highly correlated to another method or choice of sources was reduced by using different data gathering methods and study different perspectives (Denscombe, 2009). The main drawback with triangulation is for the case of contradictive results (Denscombe, 2009), this was not a problem in this research however. Furthermore, in this case, the results of the qualitative interviews could be strengthened by also conducting a quantitate questionnaire. However, a questionnaire would not have been realistic to conduct, given the time limit of this research project.
(96) 13
2.5.2 Generalizability
Lack of generalizability is one main drawback of qualitative research, especially in the setting of a single case study. The primary reason for this is that the sample size is too small to be able to draw any conclusions about a general population. It is not the purpose or strength of case study research, to provide statistical generalizability. Instead it aims to provide in-depth and comprehensive results (Denscombe, 2009). Despite this, for the research to be of importance or relevance to someone outside the case company, researchers should aim to increase the generalizability as much as possible. Unlike statistical generalizability, qualitative generalizability can be described as the possibility to transfer finding from one case to another and determining to what extent it would be reasonable to find similar results in another case (Denscombe, 2009).
To increase qualitative generalizability and the ability to transfer result from this study to other cases, presentation of the case study is important, and elaboration on how the case study is relevant in the context of the research (Denscombe, 2009). The purpose of this study is to explore Industry 4.0 adoption in production companies and automated production cells. Denscombe (2009) recommends using an independent case object with clear boundaries as it concretises the context further and therefore, one production unit with automated production cells was chosen. The need for data triangulation by studying different perspectives to receive a complete view of the case study and increase validity requires a case study object, which can provide access to different stakeholders than just operators of the automated production cells. Based on these characteristics, the case of Sandvik Coromant was selected. They have one production unit, where access to operating staff is enabled, but also business development and strategy departments, where other data sources can be collected. Further description of the case study and its appropriateness can be found in Section 1.3.
To increase generalizability eight different interviews with five different organisations were conducted. In some sense, this increases the generalizability since it includes more than one organisation. Even if actions have been taken to increase generalizability, the fact that case study research provides poor generalizability compared to other methods should not be underestimated (Denscombe, 2009). It can be strengthened to some extent by triangulation, by using data sources outside the organisation. This means that the result within the case company can be strengthened with external findings, but also that it can provide further perspectives to the research.
2.5.3 Reliability
(96) 14
3 Literature Study
In this chapter, literary background of the project is presented. First, Industry 4.0 is introduced followed by a more specific presentation of Industry 4.0 concepts as well as risks and requirements. Together, these sections provide the background for analysis and interpretation of the results. For a conceptual structure of the literature study, see Figure 3. Literature study structure.
.
Figure 3. Literature study structure.
3.1 Introduction to Industry 4.0
The production industry is currently facing the fourth industrial revolution, which was defined during the Hannover fair in Germany 2011 to boost the competitiveness of the German production industry through digitalization (Lee, 2013; Santos, et al., 2017; Santos, et al., 2017). Since then, the term has spread and has taken different notations, one of which are Industry 4.0 (Kagermann, et al., 2013; Schlaepfer, et al., 2015). In the United States and English-speaking countries, other terms are used for describing the future of production. For example, “Internet of Things”, “Internet of Everything” and “Industrial Internet” (Schlaepfer, et al., 2015). During the fourth industrial revolution, entire supply chains are expected to be further integrated using digital technologies, where all parts can communicate and make decentralised decisions independently (Kagermann, et al., 2013; Schlaepfer, et al., 2015). For an illustration of the four industrial revolutions see Figure 4.
(96) 15
Figure 4. The fourth industrial revolutions (Roser, 2015).
The general description of Industry 4.0 is a digital transformation of production industries with focus on automatic information collection, analysis and communication between factories, machines and humans (Schlaepfer, et al., 2015). Many different concepts are included under the notion of Industry 4.0, which can be found in Figure 5 (LEAP Australia, 2017). However, no exact definition of what concepts that are included in Industry 4.0 is established, instead it differs between researchers. One reason for this is that the notion Industry 4.0 is relatively new and thereby tend to get associated with many new production concepts (Qin, et al., 2016).
Figure 5. Definition of Industry 4.0 (LEAP Australia, 2017).
(96) 16
a more comprehensive approach on solutions and services, enable end-to-end integration of engineering processes across value chains (Kagermann, et al., 2013; Santos, et al., 2017). It is important to involve employees and have production staff, data scientists and software developer working together, especially since Industry 4.0 is bringing production closer to digital solutions (Gorecky, et al., 2014; Wittenberg, 2016). An example of the integration of the physical and digital world is CPS, which is further described below.
3.1.1 Cyber-Physical System
The main enabler of Industry 4.0 is the development and use of IoT and CPS:
“In essence, Industry 4.0 will involve the technical integration of CPS into production and logistics and the use of the Internet of Things and Services in industrial processes. This will have implications for value creation, business
models, downstream services and work organisation.” (Kagermann, et al., 2013).
IoT is a wide term used to describe the network of objects being connected, making it possible to share data and information from and to these objects (Xia, et al., 2012). Due to cheaper and better technology and its widespread possibilities, an increased interest in IoT have been developed, both from an academic and industrial perspective (Chen & Chen, 2016; Xia, et al., 2012). The use of IoT enables creation of a CPS. A CPS is the concept of creating a digital equivalent of the physical world, and consequently connecting these two, making it possible for the digital- and physical world to affect each other. Further, a CPS can be described as a system that involves the physical work that takes help from computational IT capabilities. Self-driving cars are one example of how the use of IT in a physical setting creates a CPS (Monostori, 2015). A CPS also enables the communication and information flow between physical objects, such as machines and humans. It is through CPS that industries will be able to enable horizontal-, vertical integration, which is described previously in this section (Kagermann, et al., 2013; Santos, et al., 2017).
The development of CPS enables massive amount of data being generated that can be analysed, visualized and consequently used to improve productivity, flexibility and customization. CPS can also enable processes across supply chains to be optimized based on the current situation, making continuously and self-optimizing decisions (Kagermann, et al., 2013; Lu, 2017; Pai, et al., 2018; MacDougall, 2014). One strategic area identified by the German Government is the use of existing production facilities in combination with new IT and ICT technologies, transforming production units into CPS. CPS can be realized in various ways, one of which is through the CPS platform which is a central platform that manages and connects all parts to each other. The platform further allows deployment of services and applications which enables collaboration in business network and ecosystems (Kagermann, et al., 2013).
(96) 17
Figure 6. CPS Framework (Lee, et al., 2015). ‘
The first level, Smart Connection Level, includes information transparency by allowing data to be gathered from different objects, collected through sensors or secondary system. It is important that the data can be collected seamlessly and transferred to a central system. Here, the first step towards combining the physical and digital world is initiated. In the next step,
Data-to-Information Conversion Level, the gathered data is processed in order make sense of it. In the
context of production, these steps involve machines and components that are aware of their current performance and can do simpler predictions of remaining lifetime. This results in self-aware objects that can inform the surroundings of their current and future performance, enabling the identifications of root cause and errors. In the Cyber Level, the data from all objects in the system are connected into a central platform. This enables overview and comparison between different machines to identify deviations, where the condition of machines can be related to a bigger fleet of machines. For the Cognition Level, the system can analyse the information to reach a higher level of knowledge of the system that is monitored, for example by suggesting how the system should perform to reduce downtime. This knowledge is presented to humans, who in a suitable way ultimately acts on these suggestions. Lastly, in the Configuration Level, the system itself will be able to provide feedback to the monitored system. This means that the system adjusts itself to make it as optimized as possible. In the last stage the system functions without human interference (Jiang, 2017; Lee, et al., 2015; Qin, et al., 2016; Vijayaraghavan, et al., 2008).
CPS can be utilized in three complexity levels, component-, machine- and production system-levels. On the component level, sensors can be attached to critical components to gather data and provide self-awareness and prediction to these components. In the next level, additional machine data, often more complex, is combined with information from these critical components to create machine CPS. Lastly, machine level CPS is combined to create CPS for processes, which can include a group of machines or an entire factory. These three steps provide a sequential methodology for increasing the intelligence level of factory processes (Lee, et al., 2015; Qin, et al., 2016).
(96) 18
different communication standards, which reduces the flexibility and scalability of the CPS (Pai, et al., 2018; Yao, et al., 2017). Lastly, current systems may not be adjusted to handle unknown knowledge, and do not have supporting interface to provide necessary insights (Yao, et al., 2017). However, one of the biggest obstacles is not with the CPS itself, but instead with the organisations that adopts it. Companies often have tight margins that do not allow for initiatives with great strategic uncertainties, which reduces the willingness to fully commit to CPS. To overcome this barrier, CPS can be introduced step by step on simpler applications that requires limited investments and first later can be gradually introduced to the factory floor (Wang, et al., 2015).
3.2 Concepts of Industry 4.0
CPS and Industry 4.0 are closely related and are in many ways the description of the same phenomena; connected and intelligent production systems. Further, the central aspects of Industry 4.0 in production can be described as digitalisation of all processes, which enables information to be exchanged and processed to make better decisions, often in form of a smart factory (Kagermann, et al., 2013). While the Industry 4.0 concept is wider and more conceptual, CPS concretizes the realization of Industry 4.0 in five sequential steps and in different levels with varying complexity by for example, applying it to single components, machines or entire processes. Further, both Industry 4.0 and CPS includes the concepts of connection of the physical world and how processing of data and information can provide decision support to humans and later enable self-optimizing abilities. For these reasons, focus is on the concepts of connectivity, data analysis and human-machine-interaction, which are presented in the following sub chapters, see
Figure 7.
Figure 7. Main Industry 4.0 concepts.
3.2.1 Connectivity
The concepts of Industry 4.0 are focused on how data can be leveraged to create factories and supply chain with an increased level of communication and self-optimizing characteristics. To enable many of these concepts, the physical parts within factories must be connected in order to generate data (Badarinath & Prabhu, 2017; Lee, 2013; Schlaepfer, et al., 2015). This is certainly the case of the CPS, whose deployment is dependent on the connection of the real world. To summarise, enabling communication in a factory is the foundation of the adoption of Industry 4.0 and CPS (Lee, 2013).
(96) 19
connectivity sources and enable safe and stable communication to prevent data losses and failures (Weyrich, et al., 2014).
One of the most common techniques for wireless connection is the radio frequency identification (RFID) technology, where a tag is placed on the object, making it possible to wirelessly communicate with surrounding receivers. This solution can be applied to various applications in production, such as inventory control, tracking of products in the supply chain where the collection of data is made autonomous (Chen & Chen, 2016). Further, connectivity has the possibility to enable communication between machines and data collection used for analysis, predictions and to enable increased decision support for humans in the factory (Lee, 2013).
The use of connection by sensors or any other techniques can enable different forms of information transmission from and between components, machines, systems and humans. Information will be transmitted in a factory and can generally be divided into three categories of communications: machine to machine (M2M), human to machine (H2M) and human to human (H2H) (Kalyani & Sharma, 2015). The definition of machines also includes applications, devices, data gathering systems and servers. Regardless of how machine information is communicated, connectivity can enable multiple use cases. For example, it can be used for monitoring of machine status and component health, real-time asset identification, remote control, predictive maintenance and digital simulation (Cochran, et al., 2016; Gorecky, et al., 2014; Deloitte, 2013). By connecting products or tools, more efficient changeover process when changing from one product or tool to another (Küpper, et al., 2017). Many of the driving forces of connectivity is to support humans in various ways, which enables them to focus on other, more value adding tasks (Coleman, et al., 2017).
While connectivity alone not necessarily brings any value, it has the potential to enable many of the Industry 4.0 concepts, which are dependent on an increased level of connection of the physical world. For example, to leverage the use of Big Data, connectivity enables gathering of necessary data (Demchenko, et al., 2014). In terms of supporting the operator in a future factory, connectivity can enable better human-machine interaction by increasing visualization and decision support (Lee, et al., 2015). In the next sections, the concepts of data analysis and HMI will be presented.
3.2.2 Data analysis
After the beginning of the third industrial revolution, the amount of data has steadily increased (Demchenko, et al., 2013; Yan, et al., 2017; Yi, et al., 2014). However, with the implementation of Industry 4.0 and Big Data, the traditional data analysis performed in industry´s has the possibility to change. Big Data will change the value creations in companies and their organisation as new insights can be created from the data. This will lead to the creation of new products and service strategies for both industrial and IT companies (to serve e.g. industry companies with data management). Big Data is also a main part of IoT. With all devices connected, Big Data volumes are created: leading to the need for management and analyses of large amount of data (Demchenko, et al., 2013; Santos, et al., 2017).
(96) 20
(Babiceanu & Seker, 2016; Demchenko, et al., 2013; Demchenko, et al., 2014; Lau, et al., 2016; Yan, et al., 2017).
The most distinguished feature of Big Data is the volume, which also is one of the challenges with implementing data analytics. The definition of volume includes the size, scale, amount and dimension of data. This is created mainly from that the number of things that generates data increases as a by-product of IoT and each connected process is becoming increasingly data rich. Specifications of the computer system are therefore required, compared to traditionally sized data volumes, to allow for all generated data to be stored locally or in databases that are manageable and accessible for the users and searchable between them (Demchenko, et al., 2013; Demchenko, et al., 2014). They also need to prioritise information to ensure that the most critical information can be handled first and use scheduling of data transfer to ensure the capacity of the network. In the Big Data environment, data is generated in high rate due to e.g. data created in real-time by sensors in machines. This creates another requirement for implementation. How data is stored, analysed and compressed has to be considered, since the data that are generated needs to be processed directly in real -time or close to real-time. This also adds specific requirements for the transfer speeds and processing power needed for the data systems (Demchenko, et al., 2013; Demchenko, et al., 2014).
The increasing variety of data, created by the increasing number of connected devices, results in an increasing complexity when storing data in the same system (Demchenko, et al., 2014). Generally, data can be divided into three categories: structured (information from e.g. sensors and measuring equipment), semi-structured (e.g. customers feedback in XML format information) and unstructured data (media data e.g. audio and videos). These different types of data are creating new requirements for databases (Demchenko, et al., 2013; Yan, et al., 2017).
The value of the collected data is an important factor to identify. The data value will depend on the processes they represent and which type of data (stochastic, probabilistic, regular, random, etc.). The value of the data will determine how the information is stored. Valuable data will be stored in its entirety for long periods and less valuable information will be stored for a shorter time or be compressed to only store the general trends of the data (Demchenko, et al., 2013; Demchenko, et al., 2014).
Computer software and service companies have already started to take advantage of Big Data. Netflix for instance already used Big Data analytics before creating the TV series House of Cards. With all their user information, they knew that the British series were popular and which actors and directors that were most appreciated of their users and could combine this to create a series that they knew would be popular before they even started to produce (Rondeau, 2001; Yi, et al., 2014).
(96) 21
and has the potential to inform the operator on when action is needed before failure occurs (Küpper, et al., 2017).
Quality inspection is commonly conducted post-processing, often in an external CMM or manually by operators. 3D scanning methods that are automated provide accurate measurements, are fast and easy and eliminates the need for post-processing, either manually or in CMM. The technology uses scanners to create a 3D picture of the machined part that then are compared to the parts CAD model by data algorithms. However, these technologies have high costs and needs high capacity hardware (Pai, et al., 2018). The benefits are that higher quality can be assured more effectively, and time spent on inspection is reduced (Deloitte, 2017).
Data analysis has the capability to convert sensor information into information that can be used either by automated systems or operators. However, if operators should use the created information, it needs to be assessable in a way that enables them to absorb the information. As a consequence, human-machine interaction has to be considered, which is explained further in the following section.
3.2.3 Human-machine interaction
Even though the visionary future of CPS and Industry 4.0 includes self-learning and optimizing factories, there is still development needed to reach this state. During this process, the human is expected to play an important role, acting as a decision maker (Lee, et al., 2015). This is especially evident in the first four of the five intelligence levels of a CPS, where the system is used to provide better support to humans (Lee, et al., 2015; Qin, et al., 2016). The first four steps include collecting, sorting, analysing and finally visualising data to provide smart decision support to the humans within smart production (Lee, et al., 2015). Research also shows that there will be a decrease in basic and repetitive tasks, but that humans will be extremely important in the future production industry, especially for tasks that cannot be automated or where the flexibility and creativity of humans is valuable (Becker & Stern, 2016; Gorecky, et al., 2014). This could for example be tasks with a more managerial role (Chen, et al., 2017), for example by monitoring multiple machines and mainly solve problems as they appear (Cochran, et al., 2016; Gorecky, et al., 2014; Wittenberg, 2016).
Due to the increase use of data and IT in Industry 4.0, the production system may be even harder for humans to control and operate. It is therefore important to develop Industry 4.0 around human capabilities, to strengthen them as much as possible (Pellicciari & Peruzzini, 2017). This puts pressure around the design of HMI to support the production staff and provide them with information in the most appropriate way (Gorecky, et al., 2014). Production staff can be supported physically, for example by robots, but also in tasks that involve decision making. The later requires consideration of how visualization is constructed to be as supportive as possible (Gorecky, et al., 2014; Wittenberg, 2016). The main enabler of such supportive visualization is factors as connectivity and data analysis, which can provide operators with real-time data and decision support that enables operators to be supported in unusual or new situations (Chen, et al., 2017; Gorecky, et al., 2014).
(96) 22
require supporting infrastructure that provides the ability to process and analyse data as well as communicate information (Gorecky, et al., 2014; Wittenberg, 2016).
Through visualization, the increased information in Industry 4.0 have the opportunity to improve operator support. However, it is important that visualization is efficient so that the operator do not get overwhelmed by information which reduces the effectiveness of the visualization. Therefore, interfaces have to be intuitive and easy to use. Traditionally, the design of industrial visualization interfaces has not been prioritized. However, with the growth of smartphones and tablets, simple interfaces have become a norm, which is diffusing into industrial applications. It is important to use standard interfaces when developing different applications to improve usability and shorten learning curves for new users or deployment of new systems. This includes how information is placed on the interface as well as how standard icons and colours are used, which should be as standardised as possible. Another way to improve visualization is to consider when, where and how information is displayed. In total, there are three ways to enable this. First, notification based information, which is when information is presented only when an action is triggered, for example in the case of a machine break. Second is location based information, which is when information is distributed only to the right location, reducing the amount of unnecessary information at places where it is not needed. The last type is user based information which is when different types of users receive different types of information. Example of different users are operators, managers and maintenance staff (Gorecky, et al., 2014).
One type of application where HMI and visualization can be used is in real-time monitoring of equipment and processing parameters, to indicate how well something is performing (Küpper, et al., 2017). This could for example be enabled by placing sensors on parts within a machine that measures force, pressure and temperature. Such parts could be spindles, bearing, cutting tools etc. Not only can it indicate to the operator when something deviates from its normal performance (Pai, et al., 2018), the data can also be collected and analysed, allowing for predictions, trend identification or automatically triggering of service activities such as remote diagnostic (Herterich, et al., 2015).
Another area where new visualisation can improve the operator’s workflow and flexibility is with remote control and monitoring of the machines (Gorecky, et al., 2014). One solution for implementing remote controls is to use a remote programmable logic controller (PLC) system (Chen, et al., 2017). With remote PLC system, it is possible to access automated machine and its controllers from a remote computer or tablet. This makes it possible for the operator to control and monitor the machine from a distance, enabling the operator to overlook additional automated machines (cells) at the same time and still have up to date information on all systems simultaneously. Remote PLC have the benefit of relatively low investment cost. Simplified, only hardware upgrade is needed; a new PLC unit with possibility to network connection and an application for the mobile device to mirror the interface of the control computer (Chen, et al., 2017).
