automation in the wood processing industry - diva portal

104
Examining Levels of Automation in the Wood Processing Industry Christian Schneider & Oscar Andersson Jönköping 2016-11-18 A case study

Upload: others

Post on 13-May-2022

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Automation in the Wood Processing Industry - DiVA portal

Examining Levels of Automation in the Wood

Processing Industry

Christian Schneider & Oscar Andersson

Jönköping 2016-11-18

A case study

Page 2: Automation in the Wood Processing Industry - DiVA portal

Postadress: Besöksadress: Telefon: Box 1026 Gjuterigatan 5 036-10 10 00 (vx) 551 11 Jönköping

This thesis has been carried out within the Department of Industrial Organization and Production at the School of Engineering in Jönköping as part of the M.Sc. program Production Development and Management. The authors take full responsibility for the opinions, conclusions and findings presented. Examiner: Kristina Säfsten Supervisors: Anette Karltun & Roaa Salim Scope: 30 credits Date: 2016-11-18

Page 3: Automation in the Wood Processing Industry - DiVA portal

ii

To our inspiration from the gym, thank you for your continuous motivation!

Page 4: Automation in the Wood Processing Industry - DiVA portal

Keywords

iii

Abstract

Companies operating in the wood processing industry need to increase their productivity by implementing automation technologies in their production systems. An increasing global competition and rising raw material prizes challenge their competitiveness. Yet, too extensive automation brings risks such as a deterioration in situation awareness and operator deskilling. The concept of Levels of Automation is generally seen as means to achieve a balanced task allocation between the operators’ skills and competences and the need for automation technology relieving the humans from repetitive or hazardous work activities.

The aim of this thesis was to examine to what extent existing methods for assessing Levels of Automation in production processes are applicable in the wood processing industry when focusing on an improved competitiveness of production systems. This was done by answering the following research questions (RQ):

RQ1: What method is most appropriate to be applied with measuring Levels of Automation in the wood processing industry?

RQ2: How can the measurement of Levels of Automation contribute to an improved competitiveness of the wood processing industry’s production processes?

Literature reviews were used to identify the main characteristics of the wood processing industry affecting its automation potential and appropriate assessment methods for Levels of Automation in order to answer RQ1. When selecting the most suitable method, factors like the relevance to the target industry, application complexity or operational level the method is penetrating were important. The DYNAMO++ method, which covers both a rather quantitative technical-physical and a more qualitative social-cognitive dimension, was seen as most appropriate when taking into account these factors. To answer RQ 2, a case study was undertaken at a major Swedish manufacturer of interior wood products to point out paths how the measurement of Levels of Automation contributes to an improved competitiveness of the wood processing industry. The focus was on the task level on shop floor and concrete improvement suggestions were elaborated after applying the measurement method for Levels of Automation.

Main aspects considered for generalization were enhancements regarding ergonomics in process design and cognitive support tools for shop-floor personnel through task standardization. Furthermore, difficulties regarding the automation of grading and sorting processes due to the heterogeneous material properties of wood argue for a suitable arrangement of human intervention options in terms of work task allocation. The application of a modified version of DYNAMO++ reveals its pros and cons during a case study which covers a high operator involvement in the improvement process and the distinct predisposition of DYNAMO++ to be applied in an assembly system.

Keywords

Hierarchical task analysis, literature review, level of competence, level of information, value stream analysis

Page 5: Automation in the Wood Processing Industry - DiVA portal

Acknowledgement

iv

Acknowledgement

First and foremost, we would like to offer our most sincere gratitude to our supervisor Roaa Salim for her guidance, patience and valuable help during the course of this thesis. Additionally, special thanks are dedicated to our devoted professor and senior supervisor Anette Karltun for her knowledge and support during the writing process. Her commitment and helpful comments were essential for the completion of this project.

Not to forget, we would like to express our gratitude towards the case company and its staff, who made this project possible. Thank you for all your commitment and support!

Finally, we want to acknowledge professor Mats Winroth for his expertise and guidance during the second phase of the case study.

Christian Schneider

Oscar Andersson

Page 6: Automation in the Wood Processing Industry - DiVA portal

Contents

v

Contents

Abstract ......................................................................................... iii

Keywords ....................................................................................... iii

Acknowledgement ........................................................................ iv

Contents ......................................................................................... v

List of abbreviations ...................................................................... ix

1 Introduction .............................................................................. 1

1.1 BACKGROUND AND PROBLEM DESCRIPTION .............................................................................. 1

1.2 AIM AND RESEARCH QUESTIONS ................................................................................................ 2

1.3 DELIMITATIONS ......................................................................................................................... 3

1.4 THESIS OUTLINE ........................................................................................................................ 4

2 Theoretical background ........................................................... 6

2.1 CHARACTERISTICS OF THE WOOD PROCESSING INDUSTRY ......................................................... 6

2.1.1 Quality criteria ................................................................................................................. 6

2.1.2 Work place ........................................................................................................................ 7

2.1.3 Job profile ......................................................................................................................... 7

2.1.4 Production aspects............................................................................................................ 7

2.1.5 Product categories ............................................................................................................ 8

2.1.6 Supply chain characteristics ............................................................................................. 8

2.1.7 Factors of competitiveness in the wood processing industry ............................................ 8

2.2 LEVELS OF AUTOMATION ........................................................................................................ 11

2.2.1 Challenges with the automation of production processes............................................... 12

2.2.2 What is Levels of Automation? ....................................................................................... 13

2.2.3 Methods for measuring Levels of Automation ................................................................ 16

Cognitive Reliability and Error Analysis Method ............................................................. 17

The Delphi method ............................................................................................................. 17

DYNAMO++ ....................................................................................................................... 17

KOMPASS .......................................................................................................................... 20

Rapid Plant Assessment ..................................................................................................... 21

Lean Customization Rapid Assessment ............................................................................ 21

MABA-MABA task allocation ............................................................................................ 21

Productivity Potential Assessment .................................................................................... 22

Systematic Production Analysis ........................................................................................ 22

Task Evaluation and Analysis Methodology ..................................................................... 22

TUTKA ................................................................................................................................ 22

Unit cost related approaches ............................................................................................. 23

2.3 LEVELS OF COMPETENCE ........................................................................................................ 23

2.3.1 The SRK model ............................................................................................................... 23

2.3.2 Operator roles ................................................................................................................ 24

2.4 LEVELS OF INFORMATION ........................................................................................................ 25

2.5 HIERARCHICAL TASK ANALYSIS ............................................................................................. 26

2.6 VALUE STREAM ANALYSIS ..................................................................................................... 27

3 Research design....................................................................... 28

3.1 RESEARCH STRATEGIES ........................................................................................................... 28

3.1.1 Theoretical methods........................................................................................................ 29

Page 7: Automation in the Wood Processing Industry - DiVA portal

Contents

vi

Traditional literature review .............................................................................................. 29

Systematic literature review .............................................................................................. 29

3.1.2 Empirical methods .......................................................................................................... 29

Case study ........................................................................................................................... 29

Applied DYNAMO++ ........................................................................................................ 30

Observation ......................................................................................................................... 31

Document analysis ............................................................................................................. 31

Focus group ........................................................................................................................ 31

3.2 METHOD APPLICATION ............................................................................................................ 32

3.2.1 Application of theoretical methods ................................................................................. 32

Traditional literature review .............................................................................................. 32

Systematic literature review .............................................................................................. 34

3.2.2 Application of empirical methods in a case study company ........................................... 36

Applied DYNAMO++ ......................................................................................................... 37

4 Findings and analysis ............................................................... 40

4.1 ISSUES WITH AUTOMATION IN THE WOOD PROCESSING INDUSTRY BASED ON THE TRADITIONAL

LITERATURE REVIEW .......................................................................................................................... 40

4.2 SELECTION OF LOA ASSESSMENT METHOD BASED ON THE SYSTEMATIC LITERATURE REVIEW 43

4.3 CASE DESCRIPTION .................................................................................................................. 51

4.4 FINDINGS ‘APPLIED DYNAMO++’......................................................................................... 53

4.4.1 Phase I – pre-study ......................................................................................................... 53

4.4.2 Phase II – measurement .................................................................................................. 55

4.4.3 Phase III – Analysis ........................................................................................................ 58

Workshop ............................................................................................................................ 58

SoPI matrices ...................................................................................................................... 61

4.4.4 Phase IV – Implementation ............................................................................................. 62

5 Discussion and conclusion ...................................................... 69

5.1 METHOD DISCUSSION .............................................................................................................. 69

5.1.1 Discussion of traditional literature review ..................................................................... 69

5.1.2 Discussion of systematic literature review ..................................................................... 69

5.1.3 Discussion of ‘Applied DYNAMO++’ ............................................................................ 70

5.2 DISCUSSION OF FINDINGS ........................................................................................................ 72

5.2.1 Research question 1 ........................................................................................................ 72

5.2.2 Research question 2 ........................................................................................................ 74

5.3 VALIDITY AND RELIABILITY .................................................................................................... 75

5.4 TRIANGULATION ASSESSMENT ................................................................................................ 76

5.4.1 Data triangulation .......................................................................................................... 77

5.4.2 Investigator triangulation ............................................................................................... 77

5.4.3 Theory triangulation ....................................................................................................... 77

5.4.4 Methodology triangulation ............................................................................................. 77

5.5 SUGGESTED FUTURE RESEARCH............................................................................................... 78

5.6 CONCLUSION ........................................................................................................................... 78

References .................................................................................... 80

Appendices ................................................................................... 85

Page 8: Automation in the Wood Processing Industry - DiVA portal

Figures

vii

Figures

Figure 1-1: How the research questions correlate to each other ................................................. 3 Figure 1-2: Overview of the chapter structure with sections linked forming a logical entity ....... 5 Figure 2-1: The LoA taxonomy and relevant implications on production characteristics ......... 19 Figure 2-2: Task optimization (left) and possible operation optimization .................................. 20 Figure 2-3: The SRK model .......................................................................................................... 24 Figure 2-4: Example of how an HTA is conducted ...................................................................... 26 Figure 2-5: Steps in the value stream analysis procedure.......................................................... 27 Figure 2-6: Value stream analysis can be applied in various system levels ............................. 27 Figure 3-1: Overview of the Applied DYNAMO++ approach ..................................................... 30 Figure 4-1: Number of hits for the first section of keywords of the traditional literature review

................................................................................................................................................ 40 Figure 4-2: Number of hits for the second keyword section of the traditional literature review 41 Figure 4-3: Number of hits for the third keyword section of the traditional literature review .... 42 Figure 4-4: Distribution of reviewed papers regarding assessment method ............................. 44 Figure 4-5: Grouping of assessment methods according to dimensional viewpoint derived from

Fasth (2012) ........................................................................................................................... 47 Figure 4-6: Examples of mouldings with high geometric complexity (left) as in Line B and low

(right) as it is the case in Line A ............................................................................................ 51 Figure 4-7: SoPI for rework ........................................................................................................... 62 Figure 4-8: SoPI for planing/ feeding ............................................................................................ 62 Figure 4-9: SoPI for primer/ top coat ............................................................................................ 62 Figure 4-10: SoPI for stacking/ plastic wrapping ......................................................................... 62 Figure 4-11: Layout of a rework loop to improve LoA in material handling .............................. 64 Figure 4-12: Suggestion 4 - Parallel processing of rework activities with transport conveyors

and puttying machine ............................................................................................................. 65 Figure 4-13: Schematic layout for a moulding selection according to quality levels before the

first processing step ............................................................................................................... 66

Page 9: Automation in the Wood Processing Industry - DiVA portal

Tables and appendices

viii

Tables

Table 2-1: Typical causes of quality costs in the wood processing industry ............................... 6 Table 2-2: Overview of OEE structure modified according to a case study in wood processing

................................................................................................................................................. 10 Table 2-3: Sheridan's Levels of Automation offers an overview about common degrees of

automation .............................................................................................................................. 14 Table 2-4: The integrated concept of Levels of Mechanization .................................................. 15 Table 2-5: Levels of physical and cognitive automation according to ....................................... 16 Table 2-6: DYNAMO++ overview according to Fasth et al. (2008) ............................................ 18 Table 2-7: The MABA-MABA list ................................................................................................. 21 Table 2-8: Table 2 9: Evaluation matrix illustrating supervisory control roles and Rasmussen’s

human behavior levels .......................................................................................................... 23 Table 2-9: Model of the relation between conscious and automatic behavior, based on

Rasmussen and Vicente ....................................................................................................... 24 Table 2-10: Operator roles and tasks .......................................................................................... 25 Table 2-11: Abstraction hierarchy for information requirements ............................................... 26 Table 3-1: Methods used for answering each of the research questions ................................. 28 Table 3-2: Overview of the research process .............................................................................. 32 Table 3-3: First keywords used during the traditional literature review...................................... 33 Table 3-4: Keywords comparing the degree of exploration of automation in different

manufacturing industries ........................................................................................................ 33 Table 3-5: Keywords showing the degree of exploration of the concept Levels of Automation

................................................................................................................................................. 33 Table 3-6: Inclusion/ exclusion criteria for the systematic literature review ............................... 35 Table 3-7: Overview of author related research (keyword 2) in Google Scholar ...................... 36 Table 3-8: Overview over the data extraction categories of the reviewed articles .................... 36 Table 4-1: Overview of the evaluation criteria used for figuring out the most suitable LoA

assessment method ............................................................................................................... 45 Table 4-2: Operational and space levels of a factory by Fasth (2012) ...................................... 46 Table 4-3: Comparison of assessment models according to evaluation criteria in table 4-1 .. 48 Table 4-4: Comparing the analysis units line A and B ................................................................ 52 Table 4-5: Comparing Line A and B after the Value Stream Mapping ....................................... 54 Table 4-6: Overview how rework is performed depending on operator ..................................... 54 Table 4-7: Scrap data regarding the scanner/ operator interface .............................................. 55 Table 4-8: Comparison operator - machine work task division Line A ....................................... 56 Table 4-9: Comparison operator - machine work task division Line B ....................................... 56 Table 4-10: LoA taxonomy for line A ............................................................................................ 56 Table 4-11: LoA taxonomy for line B ............................................................................................ 56 Table 4-12: LoC matrix for line A .................................................................................................. 57 Table 4-13: LoC matrix for line B .................................................................................................. 57 Table 4-14: LoI matrix for line A .................................................................................................... 57 Table 4-15: LoI matrix for line B .................................................................................................... 57 Table 4-16: Overview of how the processes of line A affected the competitive factors ........... 58 Table 4-17: Results from the workshop and the subsequent analysis of future LoA ............... 60 Table 4-18: Graphic results from the workshop with relevant minimum and maximum LoA

values ...................................................................................................................................... 61 Table 4-19: Schematic Fault - Symptom matrix for the packaging processes ........................ 68

Appendices

Appendix 1: Characteristics of the wood processing industry .................................................... 85

Appendix 2: LoA-, LoC- and LoI taxonomy for line A ................................................................. 88

Appendix 3: LoA-, LoC- and LoI taxonomy for line B .................................................................. 91

Appendix 4: Gantt chart of the thesis work .................................................................................. 94

Page 10: Automation in the Wood Processing Industry - DiVA portal

List of abbreviations

ix

List of abbreviations

AMS Advanced Manufacturing System

CNC Computerized Numerical Control

CREAM Cognitive Reliability and Error Analysis Method

DYNAMO Dynamic automation in manufacturing

FSM Fault Symptom Matrix

HRA Human Reliability Assessment

HTA Hierarchical Task Analysis

KOMPASS Complementary Analysis and Design of Production in Socio-technical Systems

KPI Key Performance Indicator

LCRA Lean Customization Rapid Assessment

LoA Levels of Automation

LoC Levels of Competence

LoI Levels of Information

MABA Man/ Machines Are Better At

OEE Overall Equipment Effectiveness

PPA Productivity Potential Assessment

REBA Rapid Entire Body Assessment

RPA Rapid Plant Assessment

RULA Rapid Upper Limb Assessment

SoPI Square of Possible Improvements

SPA Systematic Production Analysis

SPC Statistical Process Control

SRK Skill-, Rule- and Knowledge based behavior

TEAM Task Evaluation and Analysis Methodology

TPS Toyota Production System

TUTKA Name of the production system assessment tool derived from the Finnish “tuotantojärjestelmän kehittäminen ja arviointi”, meaning production system improvement and assessment

VSA Value Stream Analysis

VSM Value Stream Mapping

Page 11: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

1

1 Introduction

This chapter explains the background and problem description of the study which is followed by the aim and research questions, the scope of the study and the outline of the report. The reader receives a reasonable overview about the targets in the course of the research process, which delimitations are given and how the thesis is structured.

1.1 Background and problem description

Automation is seen as an efficient way to achieve cost-effective production systems in various industries (Satchell, 1998). In a production context, it is generally extended to work activities and functions that workers cannot perform as accurately and reliable as automated machines (Parasuraman, 2000). Automation can help to achieve enhanced product quality, improved handling of broad product ranges, higher process safety and more efficient resource utilization (Jämsa-Jounela, 2007). With an increasing global competition and off-shoring to low-cost countries, the design of competitive production systems has a high priority for manufacturers, especially in highly industrialized countries like Sweden (Säfsten, Winroth & Stahre, 2007).

Groover (2001) names increased labor productivity and the reduction of labor cost next to the elimination of repetitive routine tasks as reasons for pursuing automation strategies. Yet, the decision for the implementation of automated process technology seems to be rather taken without any support system guiding the decisions (Lindström & Winroth, 2010). Säfsten, Winroth and Stahre (2007) describe cases where initiatives from top management to install automation technologies without linkage to the manufacturing capabilities, such as operators’ skills, have become massive failures.

Sheridan and Parasuraman (2000) present several perspectives on which the decisions for automation can be based. Economic (automation cheaper than human labor), technical (everything gets automated whenever it is technical possible) or humanist perspective (repetitive, risky and boring tasks get automated) are quite common applied criteria. An adaptive approach by Scallen, Hancock and Duley (1995) focuses on human operator workload and situation awareness.

Increasing product customization leads to more complex products and therefore to an increased extent of automation (Youtie, Shapira, Urmanbetova & Wang, 2004; Sheridan, 2002). Yet, as described by Parasuraman (2000), extensive Levels of Automation do not necessarily result in a high performance of production systems. Connors (1998) mentions a strong rise in mental workload due to a more complex situation awareness. In such systems, i.e. complex manufacturing systems, present and future situations have to be understood during the static monitoring of production activities, as a higher probability of catastrophic failure has to be accepted (Frohm, 2008).

The concept of Levels of Automation (LoA) has been discussed as means of achieving a sufficient operator involvement which leads to an improved situation awareness and better control of abnormal scenarios within a production system (Endsley & Kaber, 1999). This represents a more flexible and dynamic approach as the traditional binary decision making procedure of automation investments used

Page 12: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

2

to implement those technologies according to the motto “all-or-none” (Fasth, 2012). Frohm, Lindström and Bellgran (2005) argue for the LoA concept as tool for achieving the most optimal production system with regard to robustness and flexibility. Following this approach, Frohm, Granell, Winroth and Stahre (2006) name both skilled human workers and technical systems as evident components for highly competitive production processes. The major concern of the LoA approach is to examine the pros and cons of combined human – technique task allocation designs with regard to different automation solutions (Fasth, 2012).

A need for automation as means for increasing competitiveness through increased productivity and product quality is also present in the wood processing industry. A relatively low value of outcome products and value added during manufacturing compared to the metal or other similar industries as well as high labor costs are reasons for that (Sandberg, Vasiri, Trischler & Öhman, 2014). Outsourcing tendencies regarding production facilities in low-cost countries strengthen this view (Schuler & Buehlmann, 2003).

Increases in material cost and difficulties regarding the recruitment of competent personnel put additional pressure on firms within the wood processing industry (Sandberg et al., 2014). As a survey among Swedish wood processing blue-collar workers reveals, a majority names ‘underdeveloped production technology’ as main issue in terms of working conditions (Karltun, 2007). Indeed, a challenge regarding process automation in wood processing industry lies in the anisotropic and variant character of the material (Karltun, 2007). Teischinger (2010) argues for a roadmap of new technologies to be applied in wood processing industry in order to stay competitive.

These factors build up the circumstances of the wood processing industry in which investments in new automation technology have to be thoroughly elaborated based on the specific industry profile and the balancing of operators’ skills and technological capabilities. Measuring Levels of Automation as supportive element to achieve this strategic fit when designing competitive production systems is subject to review in this thesis.

1.2 Aim and research questions

The aim of this thesis was to examine to what extent existing methods for assessing Levels of Automation in production processes are applicable in the wood processing industry when focusing on an improved competitiveness of production systems. The following research questions (RQ) are answered:

RQ1: What method is most appropriate to be applied with measuring Levels of Automation in the wood processing industry?

It is aimed for finding the most suitable method for measuring LoA taking into account the prerequisites regarding the wood processing industry’s ability to automate production processes. Therefore, particularities must be identified which influence the implementation of the LoA concept in this context.

The authors assume that applying a method how to measure LoA in industrial practice represents a contribution for the whole industry, as it is formulated in RQ2.

Page 13: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

3

RQ2: How can the measurement of Levels of Automation contribute to an improved competitiveness of the wood processing industry’s production processes?

Answering the second research question is supposed to point out paths where the application of the LoA concept in the mentioned industry leads to tangible competitiveness enhancements. This is achieved by formulating concrete improvement suggestions after having tested the LoA measurement in the wood processing industry. From these, generalizing conclusions are drawn out based on the Key Performance Indicators (KPIs) which are analyzed and soft factors, which are not directly measurable.

Figure 1-1 below illustrates the relation between the research questions and the contributions which are presented in the research process. A thorough study of each part of the research strategy is therefore highly important as the results of the first research question build a foundation for tackling the second question.

Figure 1-1: How the research questions correlate to each other

1.3 Delimitations

The focus regarding the industrial sector is, as mentioned in the previous part, the wood processing industry. Both the traditional-, and scientific literature reviews focus on scientific publications in English language which are not older than 1990. Regarding the method selection, aspects dealing with automation software and relevant computer simulation are excluded. The method evaluation focuses on production processes in the manufacturing industry, which leaves considerations regarding end customers and their interfaces to automated self-service applications aside.

The testing of a method for measuring Levels of Automation in order to properly answer RQ2 is limited to two production lines in one plant. Furthermore, the relevant level of analysis is on a production level, which indicates that questions regarding management operations are not included in this study.

The study of work tasks and task allocation between operators and machines is seen as crucial for the project as the focus is lying on the assessment of human – technology interaction. This motivates also the fact why product design considerations are not dealt with in this thesis. Furthermore, ergonomic issues are recognized as such during the case study, but not addressed in-depth with a specific background theory or even an own assessment method.

RQ1

• Evaluation and selection

of a suitable assessment

method in the context of

the wood processing

industry

RQ2

• Examining the potential

of the LoA measurement

in the wood processing

industry with regard to

competitiveness

improvement in a

production system

Page 14: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

4

1.4 Thesis outline

Chapter 2 presents the theoretical foundation of the thesis. This chapter deals with an industry profile of the wood processing industry with focus on the aspect of competitiveness. Furthermore, the concept of Levels of Automation (LoA) is presented in detail together with the assessment methods reviewed for measuring LoA. Level of Competence (LoC), Level of Information (LoI), Hierarchical Task Analysis (HTA) and Value Stream Analysis (VSA) are additional concepts which become important when modifying a method for the case study which is described in subsection 3.1.2.

Chapter 3 deals with a detailed presentation of the research strategy as well as the research methods applied. A separation according to theoretical and empirical methods provides a distinctive classification regarding the means for studying each research question. Literature reviews and case study design are described in order to give an understanding how the overall reasoning is built up. The data collection and analysis described in the method application encompasses various methods which raises validity and reliability of the study.

In chapter 4, the findings from the method application are described. The distinct particularities of the wood processing industry and a well-thought motivation for a selected assessment method as finding from the systematic literature study lay ground for the presentation of the case study results. These are analyzed and presented according to the phase model of the chosen methodology. The chapter is completed with a description of the case study conducted in collaboration with a major Swedish wood processing company.

Chapter 5 deals with a discussion and conclusion of the methods applied within the research strategy and the research questions. Initially the application of the literature reviews is evaluated. Thereafter, the method chosen for the case study is discussed in detail. This is followed by a structured discussion of each research question in terms of the extent they have been answered. After this, the reliability and validity of the thesis work is discussed. The chapter ends with suggestions for further research and concluding remarks concerning the main generalization aspects.

Figure 1-2 illustrates the framework of the thesis with the respective sections which complement each other in order to finally discuss the fulfillment of the research questions in chapter 5.

Page 15: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

5

2.1 Characteristics of the wood processing industry

2.2 Levels of Automation

2.3 Levels of Competence

Chapter 2 Theoretical background

2.4 Levels of Information

2.5 Hierarchical Task Analysis

2.6 Value Stream Analysis

Section 3.1 Research strategies – What is the content of the study

Section 3.2 Method application – How is the study conducted

Step I Traditional literature review

Step II Systematic literature review

Step III ‘Applied DYNAMO++’

Chapter 4 Findings and analysis

4.1 Issues with automation in the wood processing industry

4.2 Selection of LoA assessment method

4.4 Findings ‘Applied DYNAMO++’

Chapter 5 Discussion and

conclusion

5.2.1 Research question 1

5.2.2 Research question 2

5.1 Method discussion

5.3 Validity and reliability

5.4 Triangulation assessment

5.5 Suggested future research

5.6 Conclusion

Figure 1-2: Overview of the chapter structure with sections linked forming a logical entity

Page 16: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

6

2 Theoretical background

This chapter aims to familiarize the reader with some critical concepts and terms for this thesis. Characteristics of the wood processing industry are presented as well as important factors of competitiveness within this industry. Furthermore, automation challenges and various methods for the measurement of LoA are discussed. Also the concepts of Levels of Competence (LoC), Levels of Information (LoI), Hierarchical Task Analysis (HTA) and Value Stream Analysis (VSA) are presented which forms the theoretical base for the case study.

2.1 Characteristics of the wood processing industry

About 75% of the Swedish woodlands is used industrially to produce raw material for various wood refining industries (Sandberg et al., 2014). The industries processing these goods can be described as “production activities that transform primary wood products (i.e. lumber and panels) into other wood products” (Kozak & Maness, 2001, p. 47). As part of the traditional literature review conducted in the context of the wood processing industry and automation issues, the important characteristic areas are described in the following subsections. Appendix 1 provides the complete overview of the wood processing industry’s profile.

2.1.1 Quality criteria

A big issue in wood processing which also affects the automation potential is the heterogeneous character of the raw material (Kozak & Maness, 2001). Depending on origin, which type of tree it is and which season, there is an impact on the overall raw material quality. Kozak and Maness (2001) define four different causes of quality costs in the wood processing industry, namely quality related to raw material, people, processes and products. Table 2-1 provides on overview about four common deficiency categories.

Table 2-1: Typical causes of quality costs in the wood processing industry (Kozak & Maness, 2003)

Raw material Processes

Knots and other natural defects Improper drying (moisture content)

Splits and cracks Poor sizing and machining marks

Variable moisture content Incorrect drilling

Streaks and discoloration Purchasing defective products

People Products

Hiring the wrong people Poor finishing quality

Lack of training Inconsistent color

Poor morale Performance failures

Overbearing supervisors Improper home assembly

Page 17: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

7

In addition, Eliasson (2014) mentions a varying moisture content, natural dimension and shape as well as biological effects such as fungi-, bacterial- and insect infestations as crucial factors when it comes to the raw material quality. Eliasson (2014) also emphasizes the importance of mechanical factors, such as tensile strength, stiffness and durability.

2.1.2 Work place

The factor work place refers to the physical setting in which the workers operate. Physical work conditions are described as hard which is a result of lacking implementation of automation (Karltun, 2007). The same author names also the widely presence of handicraft work tasks which refers to a rather low automation level and noise, dust, heavy lifting, repetitive motions and solvents as discomfort factors experienced by blue collar workers in the Swedish secondary woodworking industry. Tuntiseranee and Chongsuvivatwong (1998) mention exposure to chemicals as significant aspect affecting the operators.

2.1.3 Job profile

The staff often has a lower education level compared to operators within other comparable industries and are commonly missing certificates and operation licenses (Tuntiseranee & Chongsuvivatwong, 1998; Karltun, 2007; Sowlati & Vahid, 2007). Karltun (2007) also describes a rapid turnover of labor force. Leschinsky and Michael (2004 ) conducted a survey among employees of the wood product industry and concluded that steady employment, good pay and security benefits such as pensions are ranked highest among motivators. As highest ranked company values, fairness, respect for the individual’s right and carefulness regarding work place conditions are named (Leschinsky & Michael, 2004 ).

Pirraglia, Saloni and van Dyk (2009) mention one way to gain competitive advantage, explicitly to implement training and education about lean implementation. Not only training about lean, but regarding new technologies applied in production is beneficial for the operator (Wiedenbeck & Parsons, 2010). This has a positive effect on both manufacturing productivity and the individual worker, as it is a beneficial factor in terms of ergonomics to educate workers and to give them more individual responsibility (Bohgard, 2009).

2.1.4 Production aspects

Hoff, Fisher, Miller and Webb (1997) indicate that there has been an increase in offshore production in the wood processing industry. They mention manufacturers in the U.S.A. who export logs (i.e. oak, ash and various tree varieties.) to production facilities in Taiwan, which in turn send the processed wood products back stateside. Immense savings in labor and production costs result from this procedure (Hoff et al., 1997). Eliasson (2014) writes about automation in relation to the wood processing industry. More stringent control of incoming raw material is proposed, i.e. introducing X-ray and ultrasound scanners as a means of measuring moisture content in a more accurate and effective way.

Page 18: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

8

Karltun (2007) compares the wood processing industry to the metal industry and states varying material characteristics as a problematic factor regarding the implementation of automation. Furthermore, cutting forces and processing speed are significantly lower than within the metal industry (Karltun, 2007). Nevertheless, industrial robots as ‘third arm’ to support handling of the bulky material is seen as a way to introduce more efficient production techniques (Eliasson, 2014). Yet, sorting and grading processes are more difficult to automate (Karltun, 2007). As a consequence, when pursing the implementation of automation technologies, tighter acceptance tolerances regarding the specifications of incoming raw wooden material have to be understood, which is associated with higher rejection rates (Eliasson, 2014).

2.1.5 Product categories

The articles reviewed focuses on various product categories, which are part of the wood processing industry. These are listed below.

Furniture (i.e. office furniture)

Fences

Palettes

Kitchen and bath cabinetry

Door and window frames

Shakes and shingles

Flooring

Mouldings and fittings

Timber houses

Miscellaneous categories such

as toys or ladders.

2.1.6 Supply chain characteristics

Schuler and Buehlmann (2003) identify that successful wood furniture manufacturers are often clustered within the same region. They mention Denmark and northern Italy as successful examples. However, there are also companies that pursue an opposing strategy and choose to outsource production to countries with lower production costs, as it is mentioned with Schuler and Buehlmann (2003). Karltun (2007) characterizes the Swedish secondary woodworking industry, which consists to a large extent of small firms with less than 50 employees.

Strategic alliances with suppliers of adhesives, packaging, steel, plastics, fabrics, lumbers or chips are favored by these firms (Schuler & Buehlmann, 2003). A demand from customers to reduce product lead time while simultaneously adding options to customize the products can be noticed (Teischinger, 2010).

2.1.7 Factors of competitiveness in the wood processing industry

The following subsection deals with the question how competitiveness is defined in general and which factors determine a company’s ability to act successfully on the market in the wood processing industry.

In general, the term competitiveness can be applied to companies, industries and nations (Hoff et al., 1997). Hopper, Jazayeri and Westrup (2008) mention sectorial competitiveness, relative cost competitiveness dealing with the real national exchange rate and productivity as three competitiveness measures on various aggregation levels.

Page 19: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

9

Dealing with the firm’s level, external as well as internal components of competitiveness can be found (Hoff et al., 1997). Whereas the first includes aspects like comparative advantage, price and market distortions and the latter encompasses considerations associated with industrial organization, like efficiency and quality (Hoff et al., 1997).

Especially diverse quality management concepts such as Statistical Process Control (SPC), the Toyota Production System (TPS) or Lean Management with its focus to reduce waste for an optimized production have been extended to the entire organization (Schmitt, Stiller & Falk, 2013). Following this approach, there is the view that an evaluation of competitiveness is build up on both product and process features (Hoff et al., 1997).

An improved quality typically leads to higher production cost, but concurrently with a more stable delivery quality, the costs due to deficiencies will be reduced enormously and higher prices or an increased sales volume can be achieved (Bisgaard, 2008). This emphasizes the connection between internal and external aspects, as price competition in a long term view must also include quality considerations (Deming, 1982).

However, Porter reduces the term competitiveness to national productivity which seems to be “the only meaningful definition of competitiveness at the national level” and names relevant factors such as product quality, design, technology and the efficiency of manufacturing (Ellis & Davies, 2000, p. 1190). Taking over a firm perspective, if a company wants to achieve a more efficient manufacturing system, Fasth, Stahre and Frohm (2007) present various parameters to consider. Efficiency, flexibility, complexity, robustness and proactivity are seen as indicators shaping a production system here (Fasth et al., 2007).

Bellgran and Säfsten (2009) link the common competitive factors cost, quality, speed, dependability and flexibility to the performance of production systems and use a polar to illustrate the extent of fulfillment of market requirements. According to Winroth, Säfsten, Stahre, Granell and Frohm (2007) automated manufacturing systems are seen as highly productive, associating it directly with an improved competitiveness.

In this context, the Overall Equipment Effectiveness (OEE) is regarded as common measure variable for determining the performance of semiautomatic and automatic production systems (Bellgran & Säfsten, 2009). It is calculated by using various generic variables which determine material utilization, availability and machine utilization (see table 2-2). The yellow frames in the table represents the three constituents of OEE; whereas the red squares illustrate losses. Kent, Bakker, Hoyles and Noss (2011) describe the purpose of this measure as balance between performance of a process and the quality whilst keeping it in a sound and sustainable condition with a proper availability.

Page 20: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

10

Table 2-2: Overview of OEE structure modified according to a case study in wood processing

Overall Equipment Effectiveness

Availability Set-up

Machine stop

No planned

production

Utilization Short stops

Reduced pace

Set-up

Machine stop

No planned

production

Quality/ Material

utilization Waste

Short stops

Reduced pace

Set-up

Machine stop

No planned

production

Approved running meters

Lead-times and on-time deliveries are effects of a better flow, whereas the reduction of defects increases the reliability of the production processes and reduces variability (Lander & Liker, 2007). Slack (2005) reviews different kinds of manufacturing flexibility as contributors to overall company performance and names product flexibility as ability to introduce or modify products, mix flexibility as measure of the range of products made within a specific time period and volume flexibility as the ability to change the total output level.

Focusing on the wood processing industry, factors which are examined in the reviewed articles include the total factor productivity, labor cost on average and production value per firm (Koebel, Levet, Nguyen-Van, Purohoo & Guinard, 2016). This leads to the assumption that economies of scale and new technology or innovative production processes have a special importance. However, Hoff et al. (1997) argue that small firms are able to capture economies of scale in collaboration with suppliers and distributors and at the same time retain a critical flexibility, so that small firm size is not a disadvantage in the wood processing industry. As Sandberg et al. (2014) describe, the wood processing industry is seen as ‘small-scale industry’ with strong differentiated activities and products, which on the other hand complicates the adaption to new conditions which occur on a global market.

Välimäki, Niskanen, Tervonen and Laurila (2011) study the relationship of innovativeness on the competitiveness of wood product companies and conclude that indicators such as patent applications, new products and processes and spending for R&D are more important for the birth of innovations compared to i.e. the level of education of personnel. Also Diaz-Balteiro, Heruzo, Martinez and Gonzalez-Pachon (2006) mention the reduction of production costs by introducing new and more efficient production processes based on technological innovation in Spain’s wood-based industry. An aspect which fastens this development are rising raw material prizes. Investments in new technologies enable firms in the wood processing industry to gain a faster customer response, achieve a quicker production and more customization as well as a greater product variety (Hoff et al., 1997).

Page 21: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

11

The implementation of management systems which provide production data is crucial for the quantification of productivity and quality figures. Nowadays, the development of improved integrated sensors in production processes allows the implementation of so-called “cyber physical systems” which enable the application of i.e. accurate efficiency and quality deficiency measures within production processes (Schmitt, Stiller & Falk, 2013, p. 309). This concept is grounded on the perception that ‘Industry 4.0’, the term by which the future step of industrialization is called in Germany, leads to an improved quantification of relevant measurable variables associated with productivity as it is mentioned by Schuh, Potente, Varandani, Hausberg and Fränken (2014) and Wesch-Potente, Weber, Prote, Schuh and Potente (2014).

Especially when regarding the anisotropic structure of wood, the control over quality deficiencies which are determined according to common property parameters such as moisture content, shape stability, density and knots is important to retain profit margin (Gustafsson & Eliasson, 2014). Fasth, Stahre and Frohm (2007) name in this context material flexibility as “the ability to handle unexpected variations in dimensions or quality in part material”. Eliasson (2014) mentions the example of timber houses manufacturers in Sweden who have focused on process development or introduced Lean principles in order to attain efficient production processes.

Summarizing the factors named by various authors it becomes clear that a specific definition of competitiveness for the wood processing industry is not given. In fact, a focus on internal indicators such as efficiency and quality, also named as process factors, is considered as a suitable approach to achieve long-term competitiveness. This is combined with the implementation of new technologies concerning products and production processes, which increases material flexibility and gains quality as well as cost structure improvements.

When noticing that only one third of the raw material which is processed in the wood processing industry becomes a product with essential higher value than the source material, a higher need for process automation as means for an increased competitiveness in the wood processing industry can be noticed (Sandberg et al., 2014). Based upon this, Frohm (2008) draws a direct line from the implementation of automation technologies to cost reductions and increased efficiency which leads to an improved competitiveness.

2.2 Levels of Automation

Relevant information about the LoA concept is necessary to understand the background, the purpose and the methods which exist within the umbrella term “Levels of Automation”. Introductory, common challenges within the domain of automation implementation in production systems are described. This clarifies the motives why LoA is used for achieving an optimal task allocation between humans and machines. Furthermore, various definitions of LoA are discussed and relevant means for measuring LoA are presented which form the theoretical base for the systematic literature review.

Page 22: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

12

2.2.1 Challenges with the automation of production processes

This subsection gives the reader hints what companies’ problems associated with automation challenges are about.

When using the term automation, it is often used to describe mechanical, electronic and computer-based systems that are used for carrying out inspection tasks and controlling operations in production (Groover, 2001). Discussing process automation and the inherent consequences for the production system, there is usually a balance to determine between people and technology. On the one side there are the power, speed and physical abilities of automation systems, which stand against the flexibility and intuitiveness as well as analytical ability of humans (Slack, 2005). There is also a distinction about the degree of human involvement, which separates manual, semi-automatic and automatic tasks (Bellgran & Säfsten, 2009).

Frohm (2008) mentions producing with a minimum of employees, a better working environment and improved quality as benefits of automation, whereas investment costs, an adaption of product design to automatic manufacturing and a too broad product portfolio are named as disadvantages. In general, automation aims to extend the technical feasibility and human capability and to overtake impossible or hazardous work tasks for humans (Lindström & Winroth, 2010). Also an increased flexibility is a reason to automate according to Osvalder and Ulfvengren (2009), even if it is emphasized that human contribution is generally still required in a technical system.

This aspect includes the human intervention in so-called advanced manufacturing systems (AMS), which – according to (Frohm, 2008) – cannot be reliable to 100 % throughout their operation time. Harlin, Frohm, Berglund and Stahre (2006) mention the operators’ education levels, knowledge about support tools and system interfaces as well as about work procedures as relevant capabilities to consider in the design of task allocations between humans and machines. Not only the best technical system is decisive for an improved performance, but also the division of work tasks between technology and its operators (Frohm, 2008).

A fractional loss of specific working skills and a degradation of the cognitive awareness are common consequences of too excessive Levels of Automation (Parasuraman, 2000). This can in turn lead to decreases in the Overall Equipment Efficiency (OEE) (Ylipää, 2000). Factors which influence the relative situation awareness in a human operated AMS are the perception of critical factors, the relation of these to the process goals and the probable system shift in near future (Endsley, 1996). As Osvalder and Ulfvengren (2009) emphasize in their work, dealing with disturbances and disruptions is critical within automated technical systems.

Following this idea, Ohno (1988) defines ‘autonomation’ as automation concepts with a human touch. The integration of devices which can distinguish normal and abnormal conditions prevents the production of defective parts, which is interpreted as human intelligence given to automated systems (Ohno, 1988).

Page 23: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

13

With a highly automated production system, which cannot perform these decisions choices on its own, the operator could be at risk to fail the detection of the few occasional times when the automation fails (Parasuraman, 2000). Apart from this, trust in automatic systems relies heavily on the overall task complexity as Endsley and Kaber (1999) report: different users adopt diverging task execution strategies which are influenced by individual perceptions and understandings. These form an individual mental workload.

Connected to the issue of situation awareness and mental workload is the risk of physical or cognitive skill degradation which consequently turns the operators into static or passive monitoring roles and further deteriorates the ability to maintain direct control (Frohm, 2008).

The concept of human - centered automation is relevant in this context, which focuses on human capabilities to be complemented with automation and not vice versa (Satchell, 1998). This approach sees operators in a production system actively involved in control or decision making activities, allocation of resources and the evaluation of alternatives (Satchell, 1998). Otherwise, with the operators feeling ‘deskilled’ as professionals due to the loss of process knowledge, the systems become vulnerable if the technology fails.

2.2.2 What is Levels of Automation?

The following subsection presents the concept of LoA and contributes to establish an understanding for the background and the purpose of this approach. Certain diverging definitions are discussed and compared based on their perspective on automation.

Several authors have written about the concept of LoA in the past. Yet, they do not combine the same understanding of this term and apply different perspectives. Groover (2001) uses the expression LoA in a pure physical sense to describe five different levels where automated systems are installed within factory operations; enterprise level, plant level, cell or system level, machine level and device level. These levels build up an automation hierarchy which categorizes a number of components for automation and technologies applicable for different automated control, processing and handling purposes within a firm.

Osvalder and Ulfvengren (2009) raise up the question to which degree automation is suitable when taking into consideration the competitive profile of a company, i.e. high productivity or high responsiveness. A task allocation between humans and machine which is based on the relative benefits of automation was developed in the 1950s and has still a big relevance when designing today’s production systems. The so-called MABA–MABA list is dealt with in subsection 2.2.3.

Looking back in history, LoA was mentioned by Sheridan and Verplank in 1978, who described it as a continuum involving decision making and action processes. The range of options encompasses complete manual control to complete automatic control (Osvalder & Ulfvengren, 2009). Table 2-3 provides an overview about the classifications in Sheridan’s scheme.

Page 24: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

14

Table 2-3: Sheridan's Levels of Automation offers an overview about common degrees of automation (Osvalder & Ulfvengren, 2009)

Sheridan’s Levels of Automation

1. Automation offers no assistance. The human must do it all.

2. Automation suggests multiple alternatives, filters and highlights those which are considered to be the best.

3. Automation selects an alternative, one way to do the task and suggests this to the human.

4. Automation carries out the action if the person approves.

5. Automation provides the human with limited time to veto the action before automatic execution.

6. Automation carries out an action, and then informs the human.

7. Automation carries out an action and informs the human only if asked.

8. Automation selects method, executes task, and ignores the human.

Satchell (1998) describes each of his seven automation levels as a way of humans and machines sharing task to achieve outcomes. The degree of human involvement reaches from assisted manual control, shared control to management by exception and autonomous operation. Consequently, both Satchell’s and Sheridan’s concepts focus on the cognitive workload which is allocated among humans and technology.

‘Level of Mechanization’ as a variation to the Level of Automation is considered as the technical level of a manufacturing system. It consists of three levels and nine steps with a continuum from manual to automated manufacturing (Frohm, Lindström, Stahre & Winroth, 2009). Table 2-4 illustrates the concept, which is built upon the approach that the automation of physical tasks is categorized according to subsystems like a Computerized Numerical Control (CNC) machine. These can be further integrated with e. g. automated transportation systems to form a manufacturing system (Frohm et al., 2009).

Page 25: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

15

Table 2-4: The integrated concept of Levels of Mechanization (Frohm et al., 2009)

Levels of Mechanization according to Kern and Schumann

Pre-mechanization Manual

Line flow

Mechanization

Single units with manual work

Single units with mechanical control

Multi-functional units without manual control

Systems of units

Automation

Partly automated singe units

Partly automated systems

Automated manufacturing

Considering the dimensional perspectives which have been referred so far regarding Levels of Automation, it becomes clear that a main distinction concerning research focus can be set between the physical and the cognitive aspect of automation. (Frohm, 2008) provides a definition of LoA which combines both approaches:

“The relation between human and technology in terms of tasks and function allocation, which can be expressed as an index between 1 (total manual work) and 7 (total automation) of physical and cognitive support.” (Fasth et al., 2007)

Their work results in a scheme, which classifies LoA in terms of a ‘Mechanical’ and ‘Cognitive’ perspective, as it can be drawn out from table 2-5.

Naming a definition of automation from (Lee, 2008), which deals with the credo that automation should aim for extending the physical and cognitive ability of people to achieve what is not possible otherwise, this illustrates again the complexity of automating. However, increased information flows in modern mass customization surroundings justify that not only technological feasibility and cost need to be considered when applying automation (Fasth, 2012).

Page 26: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

16

Table 2-5: Levels of physical and cognitive automation according to (Frohm, 2008)

Levels Mechanical Cognitive

1

Totally manual

Totally manual work, no tools are used, only the users own muscle power. I.e. The user’s

own muscle power

Totally manual

The user creates his/ her own understanding of the situation and develops his/ her course of action based on his/ her earlier experience

and knowledge. I.e. The user’s earlier experience and knowledge

2

Static hand tool

Manual work with support of a static tool. I.e. Screwdriver

Decision giving

The user gets information about what to do or a proposal for how the task can be achieved.

I.e. Work order

3

Flexible hand tool

Manual work with the support of a flexible tool. I.e. Adjustable spanner

Teaching

The user gets instruction about how the task can be achieved. I.e. Checklists, manuals

4

Automated hand tool

Manual work with the support of an automated tool. I.e. Hydraulic bolt driver

Questioning

The technology questions the execution, if the execution deviates from what the technology

considers suitable. I.e. Verification before action

5

Static machine/ workstation

Automatic work by a machine that is designed for a specific task. I.e. Lathe

Supervision

The technology calls for the users’ attention, and directs it to the present task. I.e. Alarms

6

Flexible machine/ workstation

Automatic work by a machine that can be reconfigured for different tasks. I.e. CNC

machine

Intervene

The technology takes over and corrects the action, if the executions deviate from what the technology considers suitable. I.e. Thermostat

7

Totally automatic

Totally automatic work. The machine solves all deviations or problems that occur by

itself. I.e. Autonomous systems

Totally automatic

All information and control are handled by the technology. The user is never involved. I.e.

Autonomous systems

The cognitive perspective on automation research aims to speed up information flow and to provide a sufficient decision support, which enables the human to monitor the situation (Frohm, 2008). Together with the mechanical/physical perspective of automation which covers the replacement or support of operator’s muscle power for a faster as well as enhanced performance of repetitive tasks, a more holistic view on LoA is achieved, compared to regarding automation as merely physical (Fasth, 2012).

2.2.3 Methods for measuring Levels of Automation

In order to perform an assessment of LoA in a production system, different methods have been developed in a number of publications. The further parts deal with a number of those, which are presented with their aim, their scope within an organization and relation to automation as means for production performance

Page 27: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

17

optimization. Within the course of the systematic literature review which is presented in chapter 3 of this thesis, the methods for LoA measurement are compared by selecting and applying relevant evaluation criteria.

Cognitive Reliability and Error Analysis Method

Cognitive Reliability and Error Analysis Method (CREAM) is a method for risk assessment belonging to Human Reliability Analysis (HRA), with the aspects of cognitive errors and error mechanisms (Fasth, 2012). It is mentioned by Hollnagel (1998) and aims to model control factors in their context, i.e. categorize human cognition in terms of what competence is required in order to perform certain tasks or operations, and how the actions are controlled. CREAM is mainly used to identify the cause of an observed happening. The cause is categorized as either an accident or as an erroneous action. The need and purpose of this qualitative method has its origins in the air traffic control (Hollnagel, 1998).

The Delphi method

The Delphi method was developed in the 1950s by a private research institute in U.S.A. and is named after the oracle in Delphi. The original intention for the U.S. department of defense was to work out prognoses for the intention of planning the nation’s defensive strategies (Linstone & Turoff, 2002). The methodology is about involving a panel of experts and asking each of them a series of questions. The answers are collocated and are given back to the panel experts, who now have the ability to change their answers depending on the other experts’ opinions. This is an anonymous form of an interview, with the intention of a group reaching common ground regarding an issue. As mentioned previously this method is used when it comes to predictive actions (Linstone & Turoff, 2002).

DYNAMO++

The original DYNAMO method is the result of a project (called the DYNAMO project) conducted between 2004 - 2007 and consisted of eight steps, which were verified and validated through the conduction of six case studies (Granell, Frohm, Bruch & Dencker, 2007). The further adaption of the methodology, given the name DYNAMO++, was developed between 2007 - 2009 and was validated using information gathered from additional five case studies within the Swedish manufacturing industry (Fasth, Stahre & Dencker, 2008; Granell et al., 2007).

The modifications encompassed conducting a Value Stream Analysis (VSA) in order to gather information related to time parameters and material as well as information flow. A further part is video documentation to facilitate analysis of the assembly system (Fasth et al., 2008). DYNAMO++ contains four phases with twelve steps presented in table 2-6 below. This is followed by a step-by step walkthrough of the methodology.

Page 28: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

18

Table 2-6: DYNAMO++ overview according to Fasth et al. (2008)

Overview of DYNAMO++

Phase I - Pre-Study

1 Choose a system

2 Walk the process

3 Identify flow and time parameters (VSA)

Phase II - Measurement

4 Identify main operations (HTA)

5 Measure LoA (mechanical and cognitive)

6 Document the results

Phase III - Analysis

7 Decide the minimum and maximum levels for the

identified operations (Workshop)

8 Design the Square of Possible Improvements

(SoPI) based on results from the workshop

9 Analysis of the SoPI

Phase IV - Implementation

10 Visualize suggestions of improvements

11 Implementation of the suggested improvements

12 Follow-up on the implementation

Initially the system and its boundaries have to be selected. This can be done off-site and includes discussing goals and purposes of the measurement. Relevant delimitations deal with production flow considerations and aim to identify a suitable system (Frohm, 2008). The scope of the system also defines the range of possible improvements, as a too narrow scope allows only limited changes regarding size, number and variety of task arrangements (Edwards & Jensen, 2014). Based upon this determination, the process is walked through, in order to visualize the production flow and define the relevant work stations where production activities are performed (Frohm, 2008).

The Value Stream Analysis (VSA) identifies flow and time parameters which are relevant for performing and creating a Value Stream Map (VSM) (Fasth et al., 2008). As the LoA is measured on task level with the DYNAMO++ method, each individual work task performed needs to be identified. This is done with Hierarchical Task Analysis (HTA). It simplifies the breakdown from main tasks into sub-tasks (Frohm, 2008).

The matrix which constitutes figure 2-1 illustrates two reference scales, encompassing 49 types of possible solutions for task allocation (Fasth & Stahre, 2011). The mechanical LoA on the x-axis describes ‘with what’ to assemble, whereas the cognitive LoA on the y-axis deals with ‘how’ to assemble on lower levels (1-3) and ‘situation control’ on the higher level (4-7) (Fasth-Berglund & Stahre, 2013).

The taxonomy matrix quantifies the measuring of LoA, allows a comparison of various work tasks on a ranking scale and acts therefore as reference point when discussing possible improvements (Fasth & Stahre, 2011).

Page 29: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

19

Figure 2-1: The LoA taxonomy and relevant implications on production characteristics (Fasth-Berglund & Stahre, 2013)

The results are then documented and discussed in a focus group. Fasth and Stahre (2011) name production operators, logistics, engineers and production managers as suitable target group. As a result, the relevant minimum and maximum levels of critical tasks are mapped in the reference scale in order to illustrate the possible improvements (Fasth et al., 2007).

Thus, the Square of Possible Improvements (SoPI) sets the boundaries for the company’s system to develop in the future (Fasth et al., 2008). An example of this matrix is presented in figure 2-2. The left matrix in the figure represents the SoPI for one task in a process. The dark green box shows the current LoA for that task while the area marked with a lighter green shade shows the outcome of possible improvements regarding alterations in the LoA. The matrix to the right in the figure displays the operation optimization, which means the common base of improvement for one process and its included tasks.

Page 30: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

20

Figure 2-2: Task optimization (left) and possible operation optimization (right) (Fasth & Stahre, 2008)

The following analysis of SoPI deals with the issue which measures are most feasible for a task optimization (Fasth & Stahre, 2008). In order to achieve this, companies need to consider an amount of parameters and performers which are relevant for their system (Fasth et al., 2007). After having done an analysis concerning the improvement options, an action plan is worked out which includes detailed measures to consider in the implementation phase and dates for activation (Frohm, 2008). As a last step after implementation of the chosen improvements, a follow-up reveals which effect the suggestions have had on time and relevant flow parameters (Fasth et al., 2008).

Summarizing DYNAMO++, it is about identifying a production system’s current state, visualizing how it can be improved with regard to automation measures, and finding a way to successfully implement the improvements in order to achieve -based on the selection of relevant performance parameters - an increased efficiency or flexibility level or reduce production related disturbances (Fasth & Stahre, 2008).

KOMPASS

The Complementary Analysis and Design of Production Tasks in Socio-technical Systems (shortened KOMPASS) method aims to support design teams in determining the task allocation in automated systems by guiding them with the KOMPASS criteria, whose main attributes are work system issues like task completeness or independence of work stations, human work tasks (communication requirements, time pressure) and human – machine interaction (i.e. decision authority) (Grote, Ryser, Wäfler, Windischer & Weik, 2000).

With its consideration of people-related, technological and organizational factors, it encompasses a holistic overview what to consider when designing work stations and can be validated statistically via a questionnaire and correlation analysis (Grote et al., 2000).

Page 31: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

21

Rapid Plant Assessment

Rapid Plant Assessment (RPA) analyzes a plant according to eleven categories, which encompass amongst others equipment condition or safety and environment. A questionnaire is filled in by relevant experts and as a result, a rating sheet illustrates the Leanness of the diverse sectors, revealing improvement areas (Comstock & Bröte, 2005). The method is proven for its fastness to assess a plants Lean Competence, but also requires expertise knowledge from inside the firm or industry sector to be applied in a suitable manner (Goodson, 2002).

Lean Customization Rapid Assessment

This method is derived from the already mentioned RPA method. The aim of Lean Customization Rapid Assessment (LCRA) is to provide support in the analyzing process and the design of a production system for the implementation of mass customization (Comstock, 2004). This is undertaken with the usage of three sheets; customer elicitation, engineering and manufacturing.

MABA-MABA task allocation

(Frohm, 2008) also provides an overview about the allocation of tasks between humans and machines. This highlights automation levels as production stages where both factors complement each other. Table 2-7 shows where the performance of the one actor exceeds that of the counterpart, also named as MABA-MABA (‘Man/ Machine Are Better at…’) criteria.

Table 2-7: The MABA-MABA list (Fasth, 2012; Frohm, 2008)

Humans surpass machines in the Machines surpass humans in the

Ability to detect small amounts of visual or acoustic energy

Ability to respond quickly to control signals and to apply great force smoothly and

precisely

Ability to perceive patterns of light or sound Ability to perform repetitive, routine tasks

Ability to improvise and use flexible procedures

Ability to store information briefly and then to erase it completely

Ability to store very large amounts of information for long periods and to recall

relevant facts at the appropriate time

Ability to reason deductively, including computational ability

Ability to reason inductively Ability to handle highly complex operations,

i.e. to do many different things at once

Ability to exercise judgment

The table above reveals that human nature often leaves the operator of a machine in case of an increased automation with the tasks to detect process deviations or the diagnosis of failures and similar abnormalities (Harlin et al., 2006).

Page 32: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

22

Productivity Potential Assessment

Almström’s and Kinnander’s Productivity Potential Assessment Method (PPA) provides detailed instructions how to analyze a companies’ manufacturing performance. It is based on a 4-level view, which sees OEE on the narrowest level, followed by some broader parameters like turnover rate, scrap rate and delivery accuracy on level 2. As level 3, a list of 40 questions about 11 topics (from strategy–goals to quality) assess a companies’ ability to develop and maintain a production. Level 4 encompasses the productivity increase through method improvement, therefore it reflects only an estimation of present potential (Almström & Kinnander, 2007).

Systematic Production Analysis

Systematic Production Analysis (SPA) is a tool which was developed between 2007 and 2008. The goal with SPA is to reduce costs by measuring production condition and simulating various scenarios where three parameters, i.e. quality, production tact time and down-time parameters, vary (Jönsson, Andersson & Ståhl, 2008). This method uses two automation levels (automatic and manual) to categorize operations and assembly or production stations.

Task Evaluation and Analysis Methodology

The aim of Task Evaluation and Analysis Methodology (TEAM) is to assess current advanced manufacturing systems from the users’ perspective. This main purpose is to gain perspective and to pinpoint areas which experience efficiency related problems. As a result, an improved interaction between human operators and advanced technology is achieved (Wäfler, Johansson, Grote & Stahre, 1997). The evaluation is done with help of a matrix, developed by Stahre (1995), which is based on Rasmussen’s behavior levels and Sheridan’s operator roles (see table 2-8). In this model, four factors are examined: work environment, work tasks, information flow and system performance (Johansson, 1994).

TUTKA

TUTKA was developed in the late 2000s by Koho (2010) with the aim of assessing the current state of a production system and identifying potential improvement means. It compares the current state with the desired one (i.e. a well-functioning production system) by using a matrix consisting of 33 selected key characteristics divided over six different decisions areas. Each key characteristic is weighed against six production objectives, which are defined as quality, lead time, product flexibility, volume flexibility, delivery reliability and cost (Koho, 2010). Examples of key characteristics include a reliable production equipment, positioning of customer differentiation points and cross-trained operators. The qualitative judgment assesses the degree of according to four levels – encompassing no correspondence, partial correspondence, full correspondence and adaptability. As a results, a comparison of current and desired values in specific key characteristic reveal action potential (Koho, 2010).

Page 33: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

23

Unit cost related approaches

Various authors like Gorlach and Wessel (2008) or Windmark, Gabrielson, Andersson and Støhl (2012) focus on the aspect of unit cost when trying to assess LoA. This is done by describing the LoA as relationship between equipment cost and workers’ salary per hour and expressing the ‘right’ level of automation with relevant function models. These methods use comprehensive input data to compute a model and focus on an economic perspective. Gorlach and Wessel (2008) compare productivity indices, like the amount of products per employee, with the equipment utilization. Yet, they also confirm that flexibility of production equipment is complex to quantify.

2.3 Levels of Competence

Fasth et al. (2009) present LoC, which is closely linked to the operators’ role in the assembly system. This concept takes the cognitive support of work tasks into consideration and combines it with Sheridan’s five operator roles. The evaluation matrix (table 2-8), developed by Stahre (1995) aims to isolate task which are too complex for human nature and specify the requirements for cognitive support.

Table 2-8: Table 2 9: Evaluation matrix illustrating supervisory control roles and Rasmussen’s human behavior levels (Stahre, 1995)

Behavior

Roles Skill Rule Knowledge

Plan Task 2

Teach Task 3

Monitor Task 1

Intervene Task n

Learn

By using the matrix for task analysis, a reference for empirical data gathering is achieved, mapping the operator’s action space (Fasth et al., 2009).

2.3.1 The SRK model

The SRK model was developed by the Danish professor Jens Rasmussen in the early 1980s and aims to describe three levels of behavior, which in turn describes three different types of decision-making. The three levels of behavior are called Skill-, Rule- and Knowledge-based behavior. The levels and how they interact respectively are presented in table 2-9 below.

This describes to what extent the human mind is involved in a decision-making process. Knowledge-based decision-making requires a high level of conscious awareness, i.e. acting in an unfamiliar environment, while skill-based decision-making is performed in a more autonomous way, i.e. performing frequently occurring and accustomed tasks. Rule- based decision-making relies on logic, as the user, although conscious aware, still has certain guidelines to follow.

Page 34: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

24

Table 2-9: Model of the relation between conscious and automatic behavior, based on Rasmussen and Vicente (1989)

Figure 2-3 presents a schematic sketch of the SRK model. More specifically, it illustrates how the decision-making is performed with regard to certain stimuli and carried out through physical actions. All actions carried out by people in their daily life can be broken down into decisions and responses at various levels of consciousness (Bohgard, 2009). To execute actions based on rule-based decisions often includes skill-based actions. When performing knowledge-based decisions, the user often includes both skill- and rule-based actions, as shown in figure 2-3.

Figure 2-3: The SRK model (Rasmussen, 1983)

2.3.2 Operator roles

(Sheridan, 1992) defined five types of operator roles, which are further improved and adapted for the manufacturing industry by (Stahre, 1995). These are presented in table 2-10 below. The model consists of five levels of operator supervisory control; plan, teach, set up, intervene and learn.

Plan means that the operator is in control and decides what to do in the system. Teach assures that the right tools and material are available in order to perform a certain operation, as well as how to relocate a plan into the technical system, i.e. through programming. Monitor means that the operator starts and performs the process, and performs deviation handling. When disturbances occur, the operator

Page 35: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

25

has to intervene in the process, by stopping the flow and rectifying the issue at hand. The operator learns from every new event and can use this for the planning of the next batch. Thus two different operating loops are created: the small loop for products which are already known (i.e. familiar products) and the big loop for products of more unfamiliar kind.

Table 2-10: Operator roles and tasks (Stahre, 1995)

2.4 Levels of Information

Levels of Information (LoI) is described as a dynamic flow of predictable information between assembly and production management or other relevant actors (Dencker et al., 2009). Several qualitative criteria are given for an efficient information flow in assembly lines. These are relevance, timeliness, accuracy, format, comprehensiveness and accessibility (Fasth et al., 2010). In order to perform well at a certain task, the operator needs information on a specific level, which is illustrated in table 2-11.

Page 36: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

26

Table 2-11: Abstraction hierarchy for information requirements (Sheridan, 1995)

After mapping the processes according to these criteria, four levels of information content are concluded: Insufficient, sufficient for expert operator, sufficient for novice operators and too much (Fasth et al., 2007).

2.5 Hierarchical Task Analysis

Hierarchical Task Analysis (HTA) is a methodology used for identifying tasks and subtasks of a greater process by analyzing how the main activities are performed (Frohm, 2008). The same author also mentions HTA as a structured and objective approach of describing tasks performed by the user. Figure 2-4 exemplifies what an HTA looks like. Each process has a goal (often expressed as a verb-noun, i.e. apply putty to the moulding) and in order to reach this, the relevant sub-goals which consist of tasks and subtasks need to be identified (Frohm, 2008). The process goal is therefore divided into relevant sub-goal with underlying tasks.

Figure 2-4: Example of how an HTA is conducted (Frohm, 2008)

Page 37: Automation in the Wood Processing Industry - DiVA portal

Theoretical background

27

Once created, the model serves as an effective form of system documentation and provides a thorough overview of a user’s interaction with a system.

2.6 Value Stream Analysis

Value stream analysis (VSA) is an elementary tool in lean production and aims for improving value streams (Bellgran & Säfsten, 2009). VSA is a crucial part of the Toyota Production System where it is used to picture the current- and future state of a production system. Material and information flow is mapped according their movement profile in the system. Figure 2-5 summarizes the procedure of undertaking a VSA.

Initially a certain product which is subject to the VSA has to be identified. Secondly, a current state map of the relevant production system is drawn. The flow of material and information is analyzed which results in a map where the improvement areas of the system can be identified, i.e. bottlenecks of the production system. As a next step, the future state map is drawn considering these areas of improvement.

Figure 2-5: Steps in the value stream analysis procedure (Bellgran & Säfsten, 2009)

A VSA can be conducted in various levels (labeled 1-4) of a production system, as seen in figure 2-6 below. Measuring a sub-process corresponds to the measurement of material and information flows in a workstation, i.e. a planing machine. Parameters are only measured for this specific machine. Creating a value stream map (VSM) at a factory level is a comprehensive task to improving the factory’s performance (Bellgran & Säfsten, 2009). In addition, VSM can also be done on several factories and even entire companies, as seen in level three and four of the figure below.

Figure 2-6: Value stream analysis can be applied in various system levels (Bellgran & Säfsten, 2009)

Product family

Current state map

Future state map

Plan of action

1. Sub processes

2. Single factory (door to door)

3. Several factories

4. Between and through several companies

Page 38: Automation in the Wood Processing Industry - DiVA portal

Research design

28

3 Research design

This chapter aims to introduce the reader to the research strategies and methods chosen. Initially the strategies for each research question are presented and motivated, which is followed by a walkthrough of the research methods. Both the reasoning for the choice of each method in the first part and the description of the method application in the second, are divided according to theoretical and empirical viewpoints.

3.1 Research strategies

Research strategies were set up as formulated plans with the purpose of answering the posed research questions (Saunders, Lewis & Thornhill, 2012). For this thesis, two research strategies were formulated in order to meet the requirements. More accurate date can be obtained by implementing more than one research strategy (Bryman & Bell, 2011). Literature review and case study research were the chosen research strategies for this thesis to achieve an in-depth penetration of the research questions:

RQ1: What method is most appropriate to be applied with measuring Levels of Automation in the wood processing industry?

RQ2: How can the measurement of Levels of Automation contribute to an improved competitiveness of the wood processing industry’s production processes?

The reasoning for the method selection and the method application are further explained in the two following sections. Table 3-1 provides an overview of the relation between the single methods and each research question.

Table 3-1: Methods used for answering each of the research questions

Research questions

Methods RQ1 RQ2 Comments

Traditional literature review

The traditional literature review aimed to reveal the characteristic properties of the wood processing industry regarding automation of production processes.

Systematic literature review

With the systematic literature review the objective was to gather and structure appropriate methods for empirical data collection to find the most appropriate method to measure LoA.

Applied DYNAMO++

Applied in a case study context, the use of the ‘Applied DYNAMO++’ method made it possible to answer how the LoA measurement is related to competitiveness of the wood processing industry’s production processes.

Whereas the traditional literature review was applied in order to gain a broad understanding of the wood processing industry’s characteristics regarding process automation, the systematic literature review was intended to reveal specific insights about methods for the assessment of Levels of Automation. The ‘Applied DYNAMO++’ was the selected approach used in a case study to tackle research question 2 in an industrial real-life context. The reasoning for selecting this method is discussed in section 4.2.

Page 39: Automation in the Wood Processing Industry - DiVA portal

Research design

29

3.1.1 Theoretical methods

Theoretical methods are essential to fulfill the requirements which arise from formulation of the research questions. The following subsection deals with approaches for literature analysis to gain a broad knowledge and achieve a degree of theory triangulation.

Traditional literature review

The traditional literature review is regarded as suitable method for creating a broad understanding of the wood processing industry (Armitage & Keeble-Ramsay, 2009). It was conducted in order to investigate various characteristics in the wood processing industry with regard to process automation. This built the base for screening the literature to find a suitable method for LoA measurement. Jesson et al. (2011) argue that the traditional literature review allows for a high degree of freedom and flexibility, which is suitable for the initial situation. The explorative nature of the approach enabled to present a current state map of the characteristics of the wood processing industry.

Systematic literature review

The systematic literature review is generally seen as a more objective method for compressively investigating the available literature (Jesson et al., 2011). A systematic approach for reviewing the literature was suitable to the need for synthesizing and appraising the findings of relevant studies dealing with assessment methods related to the automation of production processes (Jesson et al., 2011). It also helped to identify the absence of coverage in the literature and gave us an indication about possible research gaps (Williamson, 2002). Specified keywords, inclusion and exclusion criteria as well as headings for data classification into an extraction protocol act as quality criteria and support the rigorous way of working (Jesson et al., 2011). Without this research focus, the sheer volume of data can be perceived as overwhelming (Eisenhardt, 1989). Therefore, it enables a transparent and replicable investigation of the methods which exist for assessing LoA.

An in-depth quality assessment, connected with extraction of data into a specific extraction protocol builds a solid theoretical base for decision making for the final stage of the study, which is the second research question.

3.1.2 Empirical methods

The empirical methods were used for examining the application of a modified LoA assessment method in an industrial context. Several methods for analyzing process automation are included in the chosen approach ‘Applied DYNAMO++’, which aims to strengthen methodology triangulation.

Case study

According to (Dubois & Gadde, 2002), case study strategy is undoubtedly the best way to obtain accurate data from a study of real-life events and happenings as well as to gain an understanding of how the event interacts with the context. As research strategy, case studies are suitable in more complex situations where the researcher has limited or no control (Yin, 2003). These conditions apply within the

Page 40: Automation in the Wood Processing Industry - DiVA portal

Research design

30

sourrounding of a production line in the wood processing industry, where numerous factors influence the outcome of the process and a high degree of uncertainty has to be assumed.

The dynamics of a production line can be made more understandable with this approach (Eisenhardt, 1989). It is also an appropriate choice of method when there is little knowledge to be gathered in forehand regarding the phenomenon (Williamson, 2002). This applies to the wood process industry, as the concept of LoA represents a research gap in relation to this industry (see also subsection 3.2.1).

Applied DYNAMO++

The ‘Applied DYNAMO++’ approach was considered to study the relationship between production processes’ competitiveness in the wood processing industry and LoA assessment. By extending the original version with the concept of LoC and LoI, the method provides a step by step procedure to analyse the task allocation between machine and human (Fasth, 2012). The reasons for the choice of this modified version of the DYNAMO++ method are presented as findings of the systematic literature review in section 4.2.

Figure 3-1 provides an overview about the single steps of the method, which are explained in more detail in subsection 2.2.3 (see DYNAMO++). LoC and LoI are described in sections 2.4 and 2.5 respectively. The various data collection and analysis methods applied during the four steps are motivated in the following subsections.

Figure 3-1: Overview of the Applied DYNAMO++ approach

Page 41: Automation in the Wood Processing Industry - DiVA portal

Research design

31

Observation

Within the course of this study, several observation techniques are applied. Hereby, the continuum reaches from complete observer to a participant as observer (Williamson, 2002).

Participant observation was an appropriate choice for gathering the necessary data in this study as the degree of uncertainty within a production process demands high flexibility to obtain data (Williamson, 2002). The researcher is hereby not limited to constraints regarding funding, time and resources (Williamson, 2002). The relevance of this data collection method is given, since the informality of production processes – i.e. procedures and happenings between humans and machines which are not spoken - can be encountered with this approach.

During the Value Stream Mapping (VSM) non – participant observation is used to map the product and material flow within the case company (Fasth, 2012).

In order to broaden the amount of data, objective recordings of subject behavior can be collected with video material, which is referred as direct observation (Frohm, 2008). HTA, a task description method which aims to identify sub-tasks, was used amongst others in this observation method (Frohm, 2008).

Document analysis

To gain an understanding of the policies and guidelines which are applied in an organization (Williamson, 2002), document analysis is a commonly adopted method. Quantitative and qualitative evidence can be drawn out from instructions for work stations, safety instructions, maintenance check lists or technical drawings (Eisenhardt, 1989). As these types of documents reveal valuable insights for the researchers, they were considered for the case study.

Data gained from internal documents such as project reports or production data sheets can be analyzed with statistical means. Commonly applied data collection approaches like Value Stream Mapping (VSM) uses data from various document sources, which reveals information about process parameters or productivity figures. The need for a broad empirical data base justifies thereby the inclusion of this data collection method.

Focus group

Since the participants of a focus group are only to some extent homogenous, a group interaction can reveal insghts about concrete opportunities or issues (Williamson, 2002). In a case study context, this method is suitable due to the concentrated amounts of data produced and the human interaction aspect, which stimulates attitudes and perceptions (Williamson, 2002). When doing this, the involvement of experienced professionals is important and only an intensive group evaluation about a pedetermined topic triggers optimal suggestions.

When evaluating the square of possible improvements in phase 3 during the application of the ‘Applied DYNAMO++’ method, a focus group was choosen in order to concentrate the know-how of the professionals involved in the production line of the case study on the issue of a suitable Level of Automation.

Page 42: Automation in the Wood Processing Industry - DiVA portal

Research design

32

3.2 Method application

This section describes the methods which were selected for the research process. The research methodology process was carried out in three research steps, as displayed in table 3-2.

Table 3-2: Overview of the research process

Research steps

Step I Step II Step III

Method Traditional literature

review Systematic literature

review Applied

DYNAMO++

Procedure a. Keywords b. Database

a. Keywords b. Database c. Inclusion-/ exclusion criteria d. Sample

a. Case study b. No. observations c. No. document analyses d. Focus group

Scope

a. See tables 3-3, 3-4 and 3-5

b. Primo

a. See page 35 b. Primo/ Google

Scholar c. See table 3-6 d. 53 articles

considered for further review

a. Two processing lines of interior wood products b. approx.90 hours, 34 different shifts c. 6 work instructions/ checklists, 11 reports of internal IT system d. 7 participants, 2 hours duration

The first step aimed to give a theoretical foundation regarding the industry profile and which specific particularities existing within the wood processing industry. For that purpose, a traditional literature review was selected. In a next step, a systematic literature review was performed which dealt with relevant LoA assessment methods. Both literature reviews together aimed to answer RQ1, which is to figure out the most appropriate LoA measurement method to be applied in the wood processing industry. Lastly, the third step was designed for answering RQ2 by the use of an applied version of the DYNAMO++ method in form of a case study. The relevant units of analysis were two processing lines of interior wood products.

3.2.1 Application of theoretical methods

Theoretical methods were essential in order to respond to the first research question. In this thesis two literary reviews are conducted, these are described in the two following subsections.

Traditional literature review

In order to lay a theoretical base for answering RQ1, relevant characteristics of the wood processing industry with regard to automation had to be identified. Revealing information regarding the wood processing industry was a first step in the process of mapping the characteristics, utilizing the database ‘Primo’ at Jönköping University library.

Page 43: Automation in the Wood Processing Industry - DiVA portal

Research design

33

Due to the exploring nature of the traditional literature review approach, keywords and search terms could be configured fairly independently. Table 3-3 presents the keywords which were used for creating an initial understanding of the industry’s scope and variety.

Table 3-3: First keywords used during the traditional literature review

Initial keywords: ‘Characteristics’ AND ‘X industry’

Wood working manufacturing Wood product industry

Wood production industry Secondary wood working industry

Wood processing industry Primary wood working industry

Several synonyms for ‘wood processing industry’ were used in combination with the key search word ‘characteristics’. The keywords represent the same industry; the secondary wood processing industry which is transforming raw wood material into refined products, i.e. furniture, cabinetry, flooring and mouldings.

The following criteria were applied when conducting the search for literature in Primo during January and February 2016:

Language: Only English

Type: Only scientific publications

Year: After 1990

The keywords automation and Levels of Automation in combination with common manufacturing industries were used in a second section. This was likely to reveal new insights about the degree of exploration within these industries and allows cross-sectoral comparisons in what sense automation and the concept of Levels of Automation have been implemented in different major manufacturing industries. The tables 3-4 and 3-5 show the keywords used in the database ‘Primo’.

Table 3-4: Keywords comparing the degree of exploration of automation in different manufacturing industries

Second section of keywords: ‘Automation’ AND ‘X industry’

Automotive Industry Plastics Industry

Electronics Industry Chemical Industry

Wood Product Industry Metal Industry

Wood Working Industry Wood Processing Industry

Table 3-5: Keywords showing the degree of exploration of the concept Levels of Automation

Third section of keywords: ‘Levels of Automation’ AND ‘X industry’

Automotive Industry Plastics Industry

Electronics Industry Chemical Industry

Wood Product Industry Metal Industry

Wood Working Industry Wood Processing Industry

Page 44: Automation in the Wood Processing Industry - DiVA portal

Research design

34

The data analysis part was done by reading the articles independently of each other and discussing the relevance of the articles. The occurrence of recurring patterns such as the description of similar circumstances with several authors and the number of citations of a certain article were crucial in this phase.

The widespread recognition of a specific characteristic within greater parts of the wood processing industry and a relation or influence onto the topic of automation were the criteria applied. Due to the personal background of the researchers, who have no professional experience within the wood processing industry, subjective evaluation played a role when selecting relevant articles for the further review.

In order to achieve an overview about relevant characteristics, loose categories were defined where relevant statements of the authors were grouped. These included aspects like supply chain considerations, quality issues, job profiles, challenges in production and factors for achieving competitiveness (see also Appendix 1). Section 4.1 presents the results in form of the hits of the respective keywords and the main factors which were analyzed from the reviewed articles.

Systematic literature review

The choice of a relevant method to be applied in the LoA assessment was based on the systematic literature review. Therefore, a certain degree of rigor had to be applied. This subsection describes the data collection and analysis followed within the systematic literature review. The first research question, which deals with the sourcing and selection of an appropriate method for assessing LoA in the wood processing industry, acted as pathway.

As a first step, available knowledge concerning Levels of Automation and existing assessment methods was identified. For this purpose, several keywords were used for mapping the field in the database server ‘Primo’ at Jönköping University library.

The following criteria were applied when doing the search for relevant data during February and March 2016:

Only articles in English

Only scientific articles

Starting year: 1990

Due to the rapid changing technological progress which also affects production facilities and human-related work tasks in a shop floor context (see subsection 2.2.1.), certain assessment methods can become inappropriate by time. New technological concepts and new operator roles as well as required skills are aspects which endorse the limited search of publications after year 1990. Following the rigor standards of a systematic literature review, the restriction on scientific articles is a consequence in order to comply with a thorough quality assessment of the articles.

Page 45: Automation in the Wood Processing Industry - DiVA portal

Research design

35

Due to the research gap regarding the concept Levels of Automation and the wood processing industry, which was concluded in the preceding traditional literature review, the keywords for the comprehensive research were not related to the wood processing industry. Some other selection criteria for the choice of keywords could be deducted as results from the traditional literature review. Internal competitiveness factors like productivity were seen as decisive variables which shape a company’s ability to compete on a market with a long-term perspective. This forms the reasoning why the terms ‘production performance’ and ‘manufacturing performance’ are applied as further keywords.

The keywords complied for the sourcing of LoA assessment methods encompassed:

Levels of Automation AND assessment method

Levels of Automation AND production performance

Levels of Automation AND manufacturing performance

Levels of Automation AND assessment method AND Production

As a next step, relevant inclusion and exclusion criteria were applied as shown in table 3-6 in order to identify a body of relevant articles and achieve a suitable quality assessment of the number of hits.

Table 3-6: Inclusion/ exclusion criteria for the systematic literature review

Inclusion criteria Exclusion criteria

Production system Ergonomics

Scientific articles Software

English Computer programs

Simulation

Management

Before 1990

Self-service applications

When applying the above factors and judging the articles against the pre-defined criteria quite a small number of articles remained. In this stage, reference tracking was used to achieve a higher number and especially more detailed publications dealing with the identified methods. Hereby, the references used in these articles are checked for frequent occurrence. After going through the citations in the remaining articles, several key authors could be identified which have regularly published articles dealing with LoA assessment methods and their results were regarded likely to be suitable.

‘Google Scholar’ was used in this further screening step as instrument to gain access to databases which include scientific articles about the LoA in combination with the key contributors. This allowed to continue the review with a narrow focus on a few authors in a broader database.

Page 46: Automation in the Wood Processing Industry - DiVA portal

Research design

36

Table 3-7: Overview of author related research (keyword 2) in Google Scholar

Keyword 1 Keyword 2

Levels of Automation Frohm

Levels of Automation Fasth

Levels of Automation Sheridan

Levels of Automation Almström

Levels of Automation Endsley

The inclusion and exclusion criteria were also applied when performing a quality assessment of the papers found in this step. An obstacle regarding this was the non-accessibility of certain sources.

As a next step, the articles considered as useful according to the criteria were grouped according to categories which were seen as helpful for the later method choice. In this case, the extraction step allowed a profound comparison of the methods fund. Table 3-8 shows the categories which were applied for the extraction of data from the articles.

Table 3-8: Overview over the data extraction categories of the reviewed articles

Data Extraction Categories

Author Industry

Title of publication Country in focus

Type of publication Estimation of the complexity of

method application

Year Method strengths/ weaknesses

Method described Relevance for the case study

The extraction of relevant data from the research papers enabled a thorough analysis regarding a rating of the assessment tools according to defined evaluation criteria which is undertaken in the analysis section 4.2.

3.2.2 Application of empirical methods in a case study company

This thesis applies empirical methods for gaining a profound understanding of the real life context, in which the Levels of Automation concept is utilized. The application of data collection and analysis methods is described in the following subsections, which comprise the extent of observation activities, document analysis

Page 47: Automation in the Wood Processing Industry - DiVA portal

Research design

37

and focus group and how these are embedded in the ‘Applied DYNAMO++’ approach.

Applied DYNAMO++

The four steps of the DYNAMO++ approach chosen for the case study are described in detail in subsection 2.2.2 (see DYNAMO++). In the following, the method application of the slightly modified version is demonstrated which includes several data collection and analysis techniques named in subsection 3.1.2.

Phase I – Pre-Study

The initial stage of the ‘Applied DYNAMO++’ approach includes the selection of a suitable system. In the case study, this refers to a processing line (line A) of interior wood products and an alternative line (line B) which consisted of several machining lines connected via forklift transports. During a walkthrough of the processes the system borders were set. The physical and spatial separation of the processing line A enabled a clear determination of the production activities to include, whereas the alternative line B requires observation activities in different production halls.

The following flow and time parameters were collected during the Value Stream Mapping at line A and line B with relevant units in brackets.

No. of operators

Cycle time (s)

Lead time (s)

Set-up time (time unit/ shift)

Up time (time unit/ day)

Downtimes (min)

Batch size (x)

Scrap rate (%)

Process time (time unit/ day)

No. of shifts

Approximately 18 hours of observation during the early and late shifts were spent at each line. This included participant observation in form of conversations with operators, participation in shift handover meetings or as complete observer on the shop floor. Furthermore, the document study regarding the printed shift reports of the IT–system revealed valuable data. These were compared with the manually measured values in order to achieve source triangulation and check the accuracy of measurements. Work instructions and checklists concerning the set-up of machine parameters are studied in order to gain an understanding of the processes.

Additionally, downtime analyses which were retrievable from the IT-system installed at the production lines supported an estimation about machine availabilities. With the Value Stream Mapping the buffers between operations and the determination of (non-)value adding activities were focused on in order to figure out the influence of the bottleneck onto the system’s performance. A deeper investigation was performed regarding the bottleneck station. Detailed records about the number of products arriving and leaving this work station gave indications regarding underlying causes for problems and the occurrence of intermediate buffers.

Page 48: Automation in the Wood Processing Industry - DiVA portal

Research design

38

Phase II – Measurement

The identification of main operations regarding specific sub-tasks is a necessity to be able to determine Levels of Automation with the LoA taxonomy and was performed with observation methods. Comprehensive video material comprising the operators’ activities were recorded which facilitated the breakdown into single tasks. The observation duration for the Hierarchical Task Analysis (HTA) performed at both line A and B was scheduled for 27 hours each.

The intention was to draw conclusions regarding the LoA taxonomy as well as the LoI and LoC matrices based on the impressions observed on the shop – floor, recorded with the video material and gained from relevant documents. These were:

Guidelines for the set-up of process parameters with a specific machine (i.e. machine speed)

Operator instructions dealing with safety regulations at a machine or behavioral advices in emergency situations

Handling instructions guiding the operators (i.e. mixing ratio of the components for painting processes)

The identified sub-tasks were considered to be measured on the three assessment scales of Levels of Automation, Information and Competence. Additional factors shaping the results of phase II were conversations with operators. These deal with the operators’ level of work experience, duration of employment at the case company and to which extent they approach their work tasks based on instructions, oral information passing by or trained skills.

Phase III – Analysis

A focus group in form of a workshop was performed as a suitable approach to determine relevant minimum and maximum levels for the work tasks involved in the case study. As line B acted as reference concerning the assessment of the LoA, LoI and LoC scale, line A was focused on when planning phase III and phase IV of the ‘Applied DYNAMO++’ method.

The implementation of improvements in phase III of the method aimed to find suitable LoA for specific work tasks. Therefore, a reference to specific work task had to be established when discussing relevant approaches for enhancement. Persons involved in the workshop were three machine operators of line A, the responsible production supervisor, a production technician, the plant manager as well as the responsible technique and product development manager.

In order to deal with the large number of sub-tasks analyzed within the HTA, a pre-selection of critical processes had to be performed. Within a workshop duration of two hours, the most important tasks were selected for further consideration and discussion about improvement concepts. Points, which shape the choice of critical work activities within the processes were the effects on competitiveness factors like the OEE and flexibility considerations.

Page 49: Automation in the Wood Processing Industry - DiVA portal

Research design

39

The square of possible improvements (SoPI) was elaborated for a number of tasks of five different processes. In the following, the ‘Applied DYNAMO++’ approach was used to perform an analysis of the SoPI regarding the effects on line A’s production performance, which prepared for the improvement suggestions developed in phase IV.

Phase IV – Implementation

The visualization of improvement suggestions regarding line A is performed in section 4.3. A focus is on a reference between the increase of LoA and effect on the OEE as well as material flexibility. Furthermore, a report of findings related to changes in LoA which affected the competitiveness of line A in a positive way was written for the case company.

Due to time reasons and lack of adequate resources, the case study as it is described in this thesis does not aim to pursue further activities regarding an implementation of the improvement suggestions.

Page 50: Automation in the Wood Processing Industry - DiVA portal

Research design

40

4 Findings and analysis

In this chapter the findings from each step of the research process are presented and analyzed. Relevant characteristics from the traditional literature review represent a foundation to build a reasoning for a LoA assessment method upon. The choice of the approach to be applied in the case study is motivated as analysis drawn from the systematic literature review. After a detailed case description, the results from the proceedings of the ‘Applied DYNAMO++’ methodology are presented phase by phase, which gives a detailed introduction to the improvement measures resulting from that.

4.1 Issues with automation in the wood processing industry based on the traditional literature review

In the following section, the particular characteristics of the wood processing industry with regard to the automation of production processes are presented. The findings from the conducted traditional literature review are developed further to conclusions regarding the automation of production processes in the wood processing industry. Following the exploratory nature of the traditional literature review, this is thought to create an understanding of the circumstances which affect process automation issues in the wood processing industry.

The industry is fairly hard to grasp as a number of different product segments and local peculiarities depending on national factors create a manifold picture. Figure 4-1 illustrates the difficulty to find even a suitable search term for the industry.

Figure 4-1: Number of hits for the first section of keywords of the traditional literature review

273

35

287

186

568 562

80 4 10

Number of hits for keywords:'Characteristics' AND 'X industry'

Page 51: Automation in the Wood Processing Industry - DiVA portal

Research design

41

As seen in the figure above, the number of resulting articles and reports vary depending on which word substitute was used. Using the search terms ‘wood product industry’ and ‘wood manufacturing industry’ together with ‘Characteristics’ generated most of the results. There was a distinct difference in the occurrence frequency amongst the presented keywords, as results pended from 4 to 568 hits.

The results for the next keyword sections illustrated in figure 4-2 and 4-3 indicate that a literature research based on the search term ‘wood processing industry’ or familiar meanings (‘wood product industry’, ‘wood working industry’) together with ‘automation’ did not reveal many hits.

Figure 4-2: Number of hits for the second keyword section of the traditional literature review

The respective numbers of hits for comparable industries were significantly higher than those of the wood processing industry, so that a research gap could be assumed. This covers knowledge about automation in the wood processing industry in general as well as the application of the LoA concept in the wood processing industry which is concluded from figure 4-3.

848

1410

62 2 20

498426

853

Automotiveindustry

ElectronicsIndustry

Woodproductindustry

Woodworkingindustry

Woodprocessing

industry

Metalindustry

Plasticsindustry

Chemicalindustry

Number of hits for keywords:'Automation'

AND 'X industry'

Page 52: Automation in the Wood Processing Industry - DiVA portal

Research design

42

Figure 4-3: Number of hits for the third keyword section of the traditional literature review

Again here in the figure above, the number of hits regarding the wood processing industry or familiar acronyms was significantly lower than those of comparable industries. Automation in combination with wood processing has thus not been studied to a greater extent in recent research.

Section 2.1 deals with the characteristics of the wood processing industry and is sub-divided into six areas. The choice for this clustering was done after screening the database ‘Primo’ with the three sections of keywords and identifying 21 articles which cover general characteristics of the wood processing industry and have a relevance to the topic of process automation. Appendix 1 provides the complete overview of articles reviewed in this context.

Quality criteria are likely to affect automated processes much more than manual production processes, especially with a heterogeneous material character as it is the case with wood (Eliasson, 2014). High variances of incoming raw material are more likely to cause problems with an automated process because it cannot deal with a wide range of sizing tolerances. Also knots and other defects such as splits and cracks are unfavorable properties for automated processing. Both factors named above require improvisation and flexible procedures, which is according to the MABA-MABA task allocation (table 2-4) a domain where humans surpass machines.

A large proportion of firms within the wood processing industry in Sweden can be regarded as small-sized and there is a trend for these firms towards penetrating niche markets (Karltun, 2007; Bumgardner, Buehlmann, Schuler & Crissey, 2011). These factors do not favor process automation, as economics of scale might not have the same importance as in mass production. Work place descriptions encompass the presence of handicraft work tasks and exposure to solvents, dust and noise as well as repetitive motions, which raises the conclusion that automation is lacking the industry (Karltun, 2007).

56

41

2 1215

41

Automotiveindustry

Electronicsindustry

Woodproductindustry

Woodworkingindustry

Woodprocessing

industry

Metalindustry

Plasticsindustry

Chemicalindustry

Number of hits for keywords:'Levels of Automation'

AND 'X industry'

Page 53: Automation in the Wood Processing Industry - DiVA portal

Research design

43

Lacking research about automation in combination with the wood processing industry is an evident conclusion drawn out from the literature review (see also figure 3-3). It can be assumed that the relation between both has rather rarely been penetrated be relevant scientific studies and articles.

The competitiveness factors within the wood processing industry mentioned in subsection 2.1.7 reveal that a general strive for process automation to achieve an improved competitiveness exists. This is due to a relatively low added value to the outcome products in the wood processing industry if a comparison is drawn to other industries (Sandberg et al., 2014). The work of Eliasson (2014) who describes a higher productivity and the aspect of flexibility in terms of material deficiencies as biggest influencing factors for process automation summarizes the common viewpoints within the wood processing industry.

In total, the various characteristics presented above illustrate both the technical and cost-related challenges with the automation of production processes in the wood processing industry. The results of this analysis form on own assessment criterion when choosing a LoA assessment method and possess therefore further importance for answering research question 1.

4.2 Selection of LoA assessment method based on the systematic literature review

In order to find the most suitable method for measuring LoA in the wood processing industry, the systematic literature review was conducted which resulted in the following number of hits for the keywords below:

Levels of Automation (exact) AND assessment method: 56 (1) hits

Levels of Automation (exact) AND production performance: 81 (12) hits

Levels of Automation (exact) AND manufacturing performance: 94 (10) hits

Levels of Automation (exact) AND assessment method OR Production: 661 (7) hits

A profound refinement was reached when using the term Levels of Automation in an exact correspondence, as the figures in brackets indicate.

After having screened the ‘key authors’ contributions and the frequency of citations to specific methods, there were 53 articles, which were considered for the data extraction into the categories described in section 3.2. The following figure 4-4 illustrates the distribution of papers dealing with specific methods.

Page 54: Automation in the Wood Processing Industry - DiVA portal

Research design

44

Figure 4-4: Distribution of reviewed papers regarding assessment method

The DYNAMO method seems to be fairly known when screening the databases for LoA assessment methods, which reflects the high number of articles found for this method. Yet, these include also the articles which are summarized in the doctoral theses of the Swedish founders of the method so that there are certain overlaps. Also the high availability of these articles is a factor influencing the distribution of papers in total. Other methods like MABA-MABA are regularly cited and act as foundation for others, like TEAM. The methods are explained in detail in subsection 2.2.1.

In order to perform an analysis of the articles reviewed during the systematic literature review, defined evaluation criteria had to be applied. These helped to determine the most suitable method for assessing LoA in the wood processing industry. The criteria were chosen as a conclusion of the data extraction phase which was done with a number of 53 articles dealing with LoA assessment methods. The table below illustrates each criterion and why it is considered for the evaulation of all methods.

4 3

16

5

2

6

3 42

42 2

02468

1012141618

Number of reviewed articles dealing with LoA assessment

methods

Page 55: Automation in the Wood Processing Industry - DiVA portal

Research design

45

Table 4-1: Overview of the evaluation criteria used for figuring out the most suitable LoA assessment method

Evaluation criteria Why is this criterion applied?

Assessment objective

The meaning with this criterion is it to reconcile the various objectives of the LoA measurement which exist in the reviewed articles with the most prominent factors of competitiveness of the wood processing industry.

Operational and space level

Inspired by Fasth (2012), who conducted a review of known assessment methods for LoA earlier, this criterion shows the deepness of the method and therefore helps to maintain a focus on production processes and task level.

Application complexity

The complexity to apply a certain method is a pragmatic choice for a factor which encompasses time-, knowledge- and experienced-related issues of the researchers when conducting a case study. It ensures that the chosen method can be applied with the available resources in a satisfying way.

Relevance regarding the wood processing industry

In order to take into account the conclusions of the traditional literature review, the specific characteristics of the industry are considered.

Dimension view

The dimension view used by Fasth (2012) categorizes the methods according to a technical-physical or social-cognitive focus and achieves a classification regarding the engineering approach. This ensures an adequate grouping of the methods according to a scale which comprises quantitative up to qualitative viewpoints.

Assessment objectives as first evaluation criterion deal with the strategic goal which the methods tries to formulate and to pursue. To a great extent this factor is related to the factors for competitiveness which is dealt with in section 2.1.7. Different parameters form the decision why to change a production system and to what extent improvements are achieved. The formulation of RQ2 required to choose a method which can improve the competitiveness of production processes. Therefore, it is aimed for to assess the assessment objectives of a specific method against the most prominent factors which determine competitiveness in the wood processing industry.

The operational and space level by Fasth (2012) refer to penetration of the method in an organisation. Table 4-2 illustrates the different categories in which assessment methods can be located on an hierarchical stage from site to work station or from building to the single working place. Often methods process data from various levels and therefore are not located to a certain field. The levels also represent an aggregation of human, infrastructural and organisational elements within a company. In this thesis, the level factor assess on which level the method focuses on when aiming to improve production systems.

Page 56: Automation in the Wood Processing Industry - DiVA portal

Research design

46

Table 4-2: Operational and space levels of a factory by Fasth (2012)

Operational level Space level

Network Network

Site Location

Master plan

Segment Building

System Working area

Cell

Station Working place

Task

As an additonal factor, complexity to apply a method deals with the extent to which expert know-how, specific skills and methodical prior knowledge or access to company representatives and internal IT-systems are required when applying a certain method. Therefore, this aspect encompasses time, resource and skills related issues and is interlinked with the hierarchical aspect of operational and space levels. The criterion was taken as the majority of reviewed articles deal with case studies in industrial real life context. The category ‘Estimation of the complexity of method application’ in the data extraction step of the systematic literature review reflects the importance of an estimation to which extent the methods can be applied with the availible resources. A simple self-estimation was performed on three levels in the data extraction protcol to take into account for this factor.

In order to consider the conclusions from the traditional literature review as well, which reveal valuable insights regarding the characteristics of the wood processing industry, the factor relevance is included in the evaluation. It concerns to which extent aspects like small company sizes, ‘under developed’ production technologies and rather low value adding are suitable prerequsities for conducting a case study with a certain method. Similar to the application complexity factor, a three level scale rating was performed during the data extraction of the articles when assessing the relevance according to the categories low – medium – high. This ensured that the method chosen fits well with the industry’s characteristics during the case study.

The question in which dimension a method can be grouped refers to the socio-technical viewpoint regarding production systems. The combination of human, organizational and technical asepects in order to improve the competitiveness of a production system is divided into four areas (Fasth, 2012). Whereas the socio-cognitive engineering tries to investigate the interaction of people and computer-based technology and therefore focuses on human based systems with mostly qualitative methods, the technical physical is seen as more rational, (Fasth, 2012). Quantitative data collection tools are used here and quantifiable parameters related to time and cost act as assessment criteria within this engineering approach.

Page 57: Automation in the Wood Processing Industry - DiVA portal

Research design

47

In order to achieve a classification of the assessment methods, the dimensions of socio-cognitive and technical-physical were applied. When doing the grouping into the respective fields, the method description availible in the data extraction protocol provided assistance. Crucial for a classification into the socio-cognitive category is the presence of a human-centred approach which takes into account cognitive burdens. Addressing the topics of human-machine interaction or individual and group specific demands also makes the method to be grouped into this category. On the other side, a LoA assessment according to pre-defined criteria and checklists with measurable parameters like OEE or value flow figures supports the grouping into the technical-physical field.

Figure 4-5 shows the categorisation of each method into the respective field, except for DYNAMO++. This method provides elements of both dimensions, as it deals with measurable parameters like the VSM, but also comprises ways to assess the cognitive load. A two-sided scale measuring both mechanical and cognitive automation level shows the double-sided character of this method.

Figure 4-5: Grouping of assessment methods according to dimensional viewpoint derived from Fasth (2012)

Table 4-3 below illustrates the evaluation criteria in a matrix with relevant assessment methods from the systematic literature review. This overview reveals the suitability of the methods in accordance with the first research question. The method highligths include also the assessment objective of the specific method.

Socio-Cognitive

•CREAM

•Delphi

•KOMPASS

•MABA-MABA

•TEAM

Technical-Physical

•LCRA

•PPA

•RPA

•SPA

•TUTKA

•Unit cost model

DYNAMO++

Page 58: Automation in the Wood Processing Industry - DiVA portal

Research design

48

Table 4-3: Comparison of assessment models according to evaluation criteria in table 4-1

Assessment Method

Relevance regarding the wood

processing industry

Application complexity

Dimension view

Operational and space

level (starting

from)

Method highlights including assessment

objectives

CREAM Low Hard Socio-

Cognitive Task

Risk assessment

Contextual control models

Error avoidance

Identifies likely cause of an event

Delphi Medium Easy Socio-

Cognitive Network

Event prognosis

Survey

Panel of experts

Goal is to reach common ground

DYNAMO++ High Medium

Socio-Cognitive / Technical-Physical

Task

Assess automation levels with taxonomy

Four phase methodology

Mech./ cog. Automation levels

Task improvements

KOMPASS Medium Hard Socio-

Cognitive Cell

Task allocation

Holistic view (HTO)

Statistic validation via questionnaires or correlation analysis

LCRA Low Easy Technical-Physical

Segment

Provide support in analysis processes

Mass customization

Customer, Engineering, Manufacturing

MABA-MABA

Medium Easy Socio-

Cognitive Task

Task allocation

Man/Machine complement each other

Highlights automation levels

PPA Low Easy Technical-Physical

Segment

Analyzing manufacturing performance

Focus on several company areas

Assessed on four levels

Uses direct and indirect parameters

Page 59: Automation in the Wood Processing Industry - DiVA portal

Research design

49

Assessment Method

Relevance regarding the wood

processing industry

Application complexity

Dimension view

Operational level (starting

from) Method highlights

RPA Medium Easy Technical-Physical

Site

Plant analysis according to 11 factors

Questionnaire

Fast lean assessment

SPA Medium Medium Technical-Physical

Cell

Goal is to reduce costs by measuring three factors; i.e. quality, down time and tact time

TEAM Medium Easy Socio-

Cognitive Segment

Assess manufacturing systems from users’ perspective

Purpose to gain perspective and find problematic areas

TUTKA Low Hard Technical-Physical

System

Assessing current state of production systems

Identifying potential improvement means

Unit cost model

Low Hard Technical-Physical

System

LoA described as relation between equipment cost and worker salary

Improve cost structure

By grouping the methods reviewed during the systematic literature according to the evaluation criteria, the motivation for the choice of the ‘DYNAMO++’ approach becomes evident. The relevance to the wood processing industry can be regarded as high, since the LoA taxonomy provides a continuum how to assess the human-machine interaction which is suitable when considering the necessity for disturbance handling in wood working processes due to the heterogenous material.

Together with other methods such as MABA-MABA or CREAM, DYNAMO++ has a focus on the task level, which is suitable in terms of the formulation of RQ2, being explicitly about improving production processes. But compared to CREAM which focuses primarily on error avoidance, DYNAMO++ shows itself superior in terms of application complexity (medium) and assessment objective which is task improvement.

Page 60: Automation in the Wood Processing Industry - DiVA portal

Research design

50

Other concepts such as PPA and TUTKA are considered as to comprehensive and not solely focused on production processes, when looking at the dimensional level, or they exceed the resources for this thesis, such as KOMPASS. The disadvantage of quantitative methods such as RPA and LCRA against DYNAMO++ is that they consist out of a moreless static checklist for the analysis process, whereas DYNAMO++ includes also tools to elaborate solutions for detected failings.

As the only method, DYNAMO++ combines both the socio-cognitive and technical-phyiscal viewpoints when aiming to improve task allocation between operators and automation technology. The qualitative evidence gained from observations is complemented by the quantitative data of the Value Stream Mapping as well as by the LoA taxonomy. Eisenhardt (1989) describes this constellation of supplementing data collection techniques as favourable within a case study research. In general, this leads to more meaningful and more reliable results. The two-sided assessment scale of DYNAMO++ allows a holistic understanding of the work activities from a mechanical and cognitive viewpoint. Furthermore, by the use of a workshop during phase III, the elaboration of improvement suggestions is based on the widely acceptance of the involved personnel.

The reasoning for the choice of the ‘Applied DYNAMO++’ approach steems from the traditional literature review and the comprehension of LoA summarized in subsection 2.2.2, which both motivate a slight modification of the DYNAMO++ method in particular.

The implementation of process automation in the wood processing industry is described as problematic due to the heterogeneous character of the raw material (Eliasson, 2014; Karltun, 2007). Therefore, it can be concluded that automation concepts for this industry have to make human intervention in case of disturbance handling possible or include devices which master normal and abnormal conditions (Ohno, 1988). The LoA, LoC and LoI taxonomies included in the ‘Applied DYNAMO++’ concept provide a detailed allocation of tasks and control between operator and machine and is therefore favored against the other concepts.

Karltun (2007) mentions the rather low educational status of employees, a large proportion of handicraft tasks and repetitive motions. Automation in the wood processing industry should aim to reduce these kind of tasks but also needs an understanding of the cognitive aspects regarding the technical interfaces for the operators. Guidelines, instructions and trainings are needed with regard to the implementation of automation technologies in production (Wiedenbeck & Parsons, 2010). The ‘Applied DYNAMO++’ approach combines the higher weight of information and competence aspects with an increased mental workload and higher cognitive skills needed. The inclusion of the LoI and LoC taxonomies during phase II of the method supplements the LoA perspective and supports the development of appropriate improvement suggestions.

Page 61: Automation in the Wood Processing Industry - DiVA portal

Research design

51

The modification of the DYNAMO++ encompasses the inclusion of the LoC and LoI concept in phase II of the DYNAMO++ method. Next to the LoA taxonomy, also these methodological tools support the measurement phase and enable to make analyses based on a broad data collection. Figure 3-1 provides an overview of the modified DYNAMO++ concept, which is called ‘Applied DYNAMO++ approach’.

4.3 Case description

This subsection gives an understanding about the framework conditions of the case study and presents the units of analysis in more detail. It is conducted in collaboration with a major Swedish wood processing firm which manufactures interior wood products. The company runs manufacturing facilities in Sweden and in several northern European countries. Main customer groups are wholesalers such as do-it-yourself chain stores.

At the facility studied in the case study, the company manufactures wooden interior details such as panels and mouldings in various dimensions and sizes. The product range consists of mouldings out of finger-jointed and solid wood material which are used for both inner wall lining and for aesthetic decoration.

Figure 4-6: Examples of mouldings with high geometric complexity (left) as in Line B and low (right) as it is the case in Line A

The production plant runs several production lines which are specialized on certain product ranges. This case study studies two production lines of varying composition and structure.

Line A processes large production volumes for products with low diversity and relatively simple dimensions (figure 4-6, right). Line B consists of five separate lines which work jointly in order to obtaining a higher degree of routing flexibility and be able to produce mouldings with complex shapes (figure 4-6, left). During the course of this case study, one article in each production line was observed and the resulting characteristics are illustrated in table 4-4 below.

As the company’s focus was on improving line A, improvement suggestions were developed solely for this line in the course of the LoA assessment. Line B acted as reference in order to obtain comparable data. Table 4-4 describes which characterizing factors determine the differences between the two production lines.

Page 62: Automation in the Wood Processing Industry - DiVA portal

Research design

52

Table 4-4: Comparing the analysis units line A and B

Category Line A Line B

Processing principle Continuous flow Multiple lines

Variant diversity Low High

Product complexity Low Medium - High

Batch size High Low

Number of operators 3 9

Line A consists of 13 connected workstations and machines, such as a manual feeding station, planing, puttying, sanding, painting, scanning quality control, rework, labeling and plastic wrapping. As these are connected with conveyor belts, a continuous flow is given for the input material. Line A is manned by three operators working on two shifts respectively.

When describing the material flow within line A, initially the raw material is fed manually to the conveyor. The planing machine as first processing step give the dimensional shape to the mouldings. These move along the conveyor to the putty application machine and through an infrared oven. After an intermediate sanding process, the prime color is applied. In the following a second heat treatment takes place in an air drying station for a longer time period.

The mouldings run through a laser scanner which is programmed to detect surface deficiencies. A polishing station removes dust and minor irregularities after the scanning. The rework station is operated manually with moulding racks for defective parts sorted out by the system and for the reworked mouldings which are put back on the conveyor. A top color layer is applied as well as a second air drying process. After this, an individual moulding labelling takes place and the parts are automatically stapled to 10 pieces each. During the plastic wrapping, the bundle gets packaged and a label is applied in the end of the process line.

This layout proves appropriate for producing larger batches (n ≈ 100 000 running meters) of a narrow product range with similar profile shapes and slightly diverging dimensions. Within the final bundle package the categories fixed and declining lengths are applied. This means that either the pre-set length is bundled or end-cut mouldings which come from rework activities are accepted for further processing.

Line B contrasts from line A as it consists of five separate machining lines which are linked via forklift transports. This results in a processing line with a higher ability to handle product diversity. Different products can be routed in a flexible way throughout the system, depending on batch-size, delivery date or required operations. Line B consists of the following machining lines:

3x planning lines (two operators per line)

2x painting lines (two operators per line)

3x reworking stations (one operator per station)

Page 63: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

53

2x sorting lines (two operators per line)

As in line A, the raw material is inspected visually regarding quality deficiencies before it is fed to the planing machine. Regarding the next process step there is to mention that both of the painting lines can be used for primer and top color application which increases the routing flexibility of line B. The mouldings receive their color application on a conveyor belt in similar process steps as in line A. Manual feeding and quality inspection tasks are performed at each machining line as no surface scanner is installed in line B.

A rework process is usually scheduled for the complete batch in line B. Defects are treated with putty or the moulding is cut if the deficiencies are deemed as too severe to fill. In the sorting station, the mouldings are visually inspected by the operators and sorted according to their length. Similar packaging and labelling steps as in line A are applied by an operator being supported by a flexible wrapping station.

4.4 Findings ‘Applied DYNAMO++’

The following subsections deal with the findings and relevant data analysis gained from the conduction of the ‘Applied DYNAMO++’ method in the case study. The subsections are built up according to the phase model of DYNAMO++. Phase I presents the findings from the initial walkthrough of the production lines and the Value Stream Mapping. Phase II deals with the HTA and the assessment of the individual work activities according to the taxonomies and matrices of LoA, LoC and LoI. SoPI for chosen critical processes in phase III point out suitable improvement suggestions for enhanced automation levels which are described in detail in phase IV.

4.4.1 Phase I – pre-study

By gathering quantitative data about the relevant production line’s material and information flows, the respective bottlenecks of both production lines became evident and key figures were collected to visualize the current state of the systems.

A Value Stream Analysis was conducted at line A with its 13 connected stations and a separated rework station. After having followed a certain batch and measuring parameters like cycle times, lead times and scrap rates directly on the shop-floor, a software for measuring the OEE was used. This tool provided data regarding set-up times and downtimes.

The rework station which represented a manual workbench was identified as bottleneck of line A. In this station, the mouldings considered as faulty by the scanner were bumped out of the conveyor system with a slide and had to be put back on the conveyor of the manual rework. This practice delayed the completion of the batch in its full size considerably and led to a storage area for trolleys which represented a significant buffer. At its maximum size, this rework buffer encompassed 15% of the total batch size. Apart from this, there were no evident divergences concerning the cycle times, which means that Line A could be seen as fairly balanced.

As already mentioned in subsection 3.2.2, line B encompassed a more complex and in general broader product portfolio. As the stations were located in three different

Page 64: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

54

buildings, the information flow regarding order sequences was displayed for each workstation individually. Compared to line A, similar processes were applied with the mouldings at line B, whereas the number of operators were two per machine. A smaller batch, which was only around 13 % of the one of line A was followed and data gained in a similar way through shop-floor measurements and the OEE – tool.

Again the rework station was identified as bottleneck with a considerably higher cycle time than the remaining process steps. As line B operated according to the workshop principle, the time span in which the batch size remained in an idle state was several days. Table 4-5 summarizes some characteristics of both lines in comparison.

Table 4-5: Comparing Line A and B after the Value Stream Mapping * Line B is taken as reference to Line A in this case

Category Line A Line B*

Processing principle Continuous flow Multiple lines

Variant diversity Low High

Product complexity Low Medium - High

Batch size High Low

Number of operators 3 9

Ratio of value adding

time Line A to B 0,59 1*

Ratio of scrap rate 0,77 1*

Noteworthy data regarding the bottleneck station rework was collected in line A. Table 4-6 shows scrap data acquired from observing four operators at the rework station. It displays what kind of rework activity was carried out with the mouldings arriving at the rack. The last column is of interest here, and verifies the need for increased cognitive automation levels. The proportion of mouldings which do not require rework after a visual inspection vary from 9% to 33% depending on the operator and his view on the approved condition of a moulding.

Table 4-6: Overview how rework is performed depending on operator

Operator Rework

by cutting Rework

by puttying No performed

rework

1 67% 0% 33%

2 90% 0% 10%

3 72% 9% 19%

4 61% 30% 9%

It is also of interest to analyse the variations between rework with puttying and cutting. Cutting is the most commonly used method, even if some moulding can be

Page 65: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

55

treated by applying putty. Two workers performed no putty application at all due to time constraints. Cutting proves to be a faster method, although it induces a higher degree of scrap. The decision is made by the individual operator and his personal perception since there are no work instruction covering these tasks.

Table 4-7 shows the aggregated data regarding the operating of the rework station. The scanner regarded 17% of the mouldings to be defect and sent them to the rework rack. Out of these 17%, 23% were sent back without any performed rework actions due to the operator being unable to find any surface deficiencies. This means a total of 13% of the mouldings were treated with rework activities.

Table 4-7: Scrap data regarding the scanner/ operator interface

Category Proportion in percentage

Identified as scrap by scanner 17%

From this, put back by operators 23%

Proportion of mouldings reworked after operators' inspection

13%

After completion of the pre-study phase, the following conclusion can be drawn:

The system borders for both lines started with the input buffer at the first conveyor, which fed the planing machines

The system borders ended with a packaged bundle of mouldings at the packaging station which was ready for the transport to the final stock

The work tasks to be observed in phase II thereby dealt with the processes within these borders as well as the transportation between the stations

Scrap rate varied among rework operators, given the fact that the operators rework according to their own experience and perception

4.4.2 Phase II – measurement

Using the observation techniques described in subsection 3.1.2, the relevant work tasks can be listed and assessed according to the LoA scheme presented in subsection 2.2.2. An HTA analysis supports the process. In total there were 93 work tasks identified for line A and 99 work tasks for line B. Appendices 2 and 3 show an overview about the identified tasks and their LoA, LoC and LoI assessments in line A and B respectively.

Page 66: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

56

Table 4-8: Comparison operator - machine work task division Line A

Who operates task?

Number of tasks

Total Proportion

Operator 56 60%

Operator/ Machine

3 5%

Machine 33 35%

Table 4-9: Comparison operator - machine work task division Line B

Who operates task?

Number of tasks

Total Proportion

Operator 62 63%

Operator/ Machine

13 13%

Machine 24 24%

An initial rough categorization of the work tasks according to the automated, semi-automated and manual tasks was performed which supported the assignment of automation levels on the 7x7 scale. These include the seven mechanical and seven cognitive Levels of Automation. The tables 4-8 and 4-9 showed a higher proportion of machine operated tasks at line A, which was evident when considering the continuous processing working principle applied. This means that i.e. transportation tasks are machine operated over here, which is not the case with line B.

When dealing with the Levels of Automation taxonomy it is not possible to decide an average LoA, as the concept deals with ratio scales. The LoA taxonomies for both lines reveal that a high proportion in both systems are done purely manual on level (1;1). Also the LoAmech. is generally higher than LoAcog. which is stable for both lines. The following tables 4-10 and 4-11 show the LoA taxonomies of line A and line B.

Table 4-10: LoA taxonomy for line A

LoAmech

LoAcog 1 2 3 4 5 6 7

7 6 1 5 1 26 5 4 4 3 3 5 2 7 5 1 33 3

Table 4-11: LoA taxonomy for line B

LoAmech

LoAcog 1 2 3 4 5 6 7

7 6 1 5 1 4 17 3 4 15 1 3 6 1 2 3 5 1 41 1

Both the LoC and LoI matrices (Tables 4-12 and 4-13) are built up solely out of the semi-automated and manual work tasks. When interpreting the LoC matrices, it can be said that knowledge related tasks were mainly related to disturbance handling

Page 67: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

57

tasks, whereas planning and learning tasks were not performed within the production lines. This is basically due to the fact that the objects of analysis are executing systems and do not plan production on its own.

Table 4-12: LoC matrix for line A

Behavior

Role Skill Rule Knowledge

Plan

Teach 4

Monitor 22 16

Intervene 6 11

Learn

Table 4-13: LoC matrix for line B

Behavior

Role Skill Rule Knowledge

Plan

Teach 3

Monitor 23 29

Intervene 6 14

Learn

Analyzing the LoI gathered for both lines, it becomes clear that expert knowledge is necessary for a majority of work tasks. A few existing working instructions and checklists give information how to proceed regarding specified work tasks or in emergency situations which is likely to ensure a categorization into the ‘sufficient for novice’ group.

As line B has a higher variant diversity and product complexity compared to line A, which affects for example machine set-up or quality inspection tasks, a higher proportion of tasks which require an expert level could be assumed. Yet, this evidence cannot be drawn from table 4-12 and 4-13. A higher proportion of tasks with a ‘novice’ level for line B with and higher total number of tasks assessed in the LoI matrix (60 tasks for line A to 75 tasks in line B) indicate that the corresponding activities for line A have been automated.

Table 4-14: LoI matrix for line A

LoI Level Number of

tasks

Proportion of work tasks

Insufficient 4 6%

Sufficient for Expert

40 67%

Sufficient for Novice

16 27%

Too much 0 0%

Table 4-15: LoI matrix for line B

LoI Level Number of

tasks

Proportion of work tasks

Insufficient 2 3%

Sufficient for Expert

41 55%

Sufficient for Novice

32 42%

Too much 0 0%

The design of the case study with line B as a reference line makes an impact when comparing the conduction of work tasks at both lines. The frequency of occurrence of similar work activities allows a more accurate classification according to the LoA, LoI and LoC levels. This is especially important, as both case study investigators and the case company possess no prior experience with the DYNAMO++ methodology.

Page 68: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

58

4.4.3 Phase III – Analysis

In this subsection the results from the third phase of the ‘Applied DYNAMO++’ method are presented. A workshop which consisted of employees working at varying hierarchical levels at the case company was performed to identify relevant minimum- and maximum values regarding LoA of selected processes and tasks. The results from the workshop were used for developing SoPI matrices for the most critical processes.

Workshop

The workshop preparations consisted of identifying the processes with the highest impact on selected competitive factors within the wood processing industry (see also subsection 2.1.7), as seen in table 4-16 below. Material utilization, productivity and availability make up OEE. Material flexibility refers to the line’s flexibility to handle material variations (i.e. tolerances).

Four processes were discussed during the workshop which seemed to have the highest impact on the four competitive factors. Table 4-16 marks these processes with bold and underlined text. A process given the value ‘3’ in the table has a high impact on that specific competitive factor, while ‘0’ indicates no impact. This subjective scale is chosen to provide a clearly arranged ranking for the workshop participants, taking into account the resources available for the case study

Table 4-16: Overview of how the processes of line A affected the competitive factors

Competitive factors

OEE

Process Material

utilization Productivity Availability

Material flexibility

Feeding/ Planing 2 3 1 3

Flipping 0 0 0 0

Puttying 2 1 2 1

Heat treating 0 1 1 0

Sanding 0 0 0 0

Priming 2 2 3 2

Drying 1 1 1 1 1

Scanning 3 1 1 2

Polishing 0 0 0 0

Reworking 3 3 0 0

Top coating 2 2 3 2

Drying 2 1 1 1 1

Individual labeling 0 0 0 1

Plastic wrapping/ Stacking

1 1 3 3

Page 69: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

59

The selected processes were planing, priming, rework, top coating and plastic wrapping. The planing process had a direct impact on the line’s OEE, as it set the process speed parameter. Priming and top coat application were selected because they both impacted all competitive factors at a high degree. A number of cleaning and maintenance tasks of the color application processes had an effect on the availability of process line A, as downtimes had to be scheduled for this.

Referring to the results from the VSM, the choice of the rework station as bottleneck station was evident. Mouldings which were indicated as defective by a scanner were put out of the system to a rework rack. As shown in subsection 4.3.1, the visual quality inspection of the mouldings followed different standards, depending on the operator who performed the task.

The plastic wrapping process was chosen as a critical process due to the high number of unplanned stops which occur. The operators need to handle and correct occurring disturbances based on their experiences, as there is only an alarm indicating an abnormal situation. Due to the continuous processing principle of line A, a disturbance or short stop at one work station as the plastic wrapping process affects the whole line’s OEE.

The workshop aimed to locate key tasks within the critical processes regarding the effect on the competitive factors. The seven attendees (excluding the researchers) consisted of three machine operators from line A, the responsible production supervisor, a production technician, the plant manager as well as the responsible technique and product development manager. The participants were introduced to the concept of LoA and familiarized with terms necessary to comprehend the context properly.

Each of the four processes determined as critical was discussed and improvement approaches were generated during the workshop considering the alteration of LoA. As tables 4-17 and 4-18 show, ten improvement approaches were brought forth and their effect on the respective LoA scales was discussed. Each suggestion had an effect on one specific task within the related process.

Page 70: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

60

Table 4-17: Results from the workshop and the subsequent analysis of future LoA

Approach #

Suggestion Process Task LoAcurrent LoAfuture

1 Scanner marks

defective parts with ink jet technology

Rework Quality

inspection 1;1 6;6

2 Improved material handling in rework

Rework Move from

rework Move to rework

1;1 5;4

3 Error/ boundary images of parts

Rework Quality

inspection 1;1 1;3

4 Parallel rework processing, 2nd

loop Rework

Move from rework

Move to rework 1;1 5;4

Putty application 2;1 5;4

5 Automatic feeding Feeding/ Planing

Feed material 1;1 4;1

6 Scanner checks

input material Feeding/ Planing

Quality inspection

1;1 6;6

7 Scanner after

planing process Feeding/ Planing

Quality inspection

1;1 6;6

8 Adjustment device

for spraying pressure and angle

Primer and top coating

Priming top coating

5;4 5;6

9 Surveillance camera Primer Monitor process 1;1 2;1

10 FSM Stacking/

Plastic wrapping

Disturbance handling

1;1 1;3

Table 4-18 differentiates the LoA alterations with regard to mechanical and cognitive automation. It provides a clear overview of how each approach (A1-10) is likely to affect the mechanical and cognitive automation level. The respective LoAcurrent value is framed. It represents the minimum value of the LoA scale whereas the intended change, portrayed in the improvement approach, symbolizes the maximum automation level for the specific task.

A big proportion of time during the workshop was spent on discussing the rework process, as it had a high impact as the bottleneck of line A. Approaches for distinct increases in both sections of the LoA scale were proposed. Furthermore, concepts for improving the LoA with the planning, primer/ top coat and the stacking/ plastic wrapping process were generated.

Page 71: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

61

Table 4-18: Graphic results from the workshop with relevant minimum and maximum LoA values

LoAmech

Approach #

1 2 3 4 5 6 7

1 min max

2 min max

3 min/ max

4 min max

5 min max

6 min max

7 min max

8 min/ max

9 min max

10 min/ max

LoAcog

Approach #

1 2 3 4 5 6 7

1 min max

2 min max

3 min max

4 min max

5 min/ max

6 min max

7 min max

8 min max

9 min/ max

10 min max

SoPI matrices

The improvement approaches were translated into task optimization matrices (figures 4-7 through 4-10). The original LoA was marked in a lighter color while the resulting LoA after the intended change was marked in a darker color. The suggestions of improvement are described in detail in subsection 4.4.4, covering phase IV of the ‘Applied DYNAMO++’ methodology.

Approaches 1-4 (A1-4), displayed in figure 4-7, deal with changing the rework process. The concepts regarding improvements of the rework process are affecting the LoA on both axes, i.e. both mechanically and cognitively. Figure 4-8 displays suggestions of how to improve the planing process.

Page 72: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

62

A5-7 varies in their effect on the LoA scope. A5 solely provides an increase in mechanical automation, while A6 and A7 result in a change of both the mechanical and cognitive aspect. A8-10 presented in figures 4-9 and 4-10 only result in increased cognitive automation for both color application processes and stacking/plastic wrapping respectively

Figure 4-7: SoPI for rework

Figure 4-8: SoPI for planing/ feeding

Figure 4-9: SoPI for primer/ top coat

Figure 4-10: SoPI for stacking/ plastic wrapping

4.4.4 Phase IV – Implementation

The last phase of the Applied DYNAMO++ approach deals with the development of improvement suggestions based on the analysis of the SoPI matrices from the previous subsection. In the following, ten proposals are presented which aim to increase the Levels of Automation within the critical processes in line A in accordance with the competitiveness factors OEE and material flexibility. Improvement suggestions beyond the task level which do not improve the execution of a work activity itself are not considered in this section. The suggestions 1-10 correspond to the minimum and maximum values shown in table 4-17 and 4-18 as well as in figures 4-7 through 4-10. The LoA changes described in the running text can be made understandable when studying the SoPI matrices.

Page 73: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

63

Suggestion 1: Scanner marks defective parts

As the analysis of the Value Stream Analysis in subsection 4.4.1 reveals, the 2D-scanner installed for in-line quality assurance in the processing line seemed to have a discrepancy regarding the rate of mouldings detected as defective. As a common practice within the rework station the operators checked each product for defects such as cracks, raw surface or edge damage. As the data collection reveals further, an average of 23% of the mouldings were not assessed as defective by the operators and put back into the conveyor system. It has to be concluded that the verification after defective points on the mouldings was not successful in these cases.

Both this discrepancy concerning diverging degrees of quality assurance of the scanner and the operator as well as the second quality inspection performed by the operator influenced the OEE negatively. It is therefore recommended to extend the scanners function with a marking function. Possible deficiencies with the mouldings passing through the scanner can be indicated with a direct part marking method such ink jet. This method offers low initial investment compared to other non -contact marking methods available on the market and the dots can be removed easily when applying putty on the wooden surface.

The quality inspection performed by the operator which is ranked on level (1; 1) in LoA scale becomes (6; 6) with the installment of an additional scanner function. A reduction of the cycle time within the rework process reduces the intermediate buffers occurring due to the comprehensiveness of the reworking activities.

Suggestion 2: Improved material handling in rework

In order to achieve considerable improvements concerning the material handling within the reworking process, a conveyor system is recommended as a second loop encompassing the reworking activities quality inspection and a respective puttying or cutting process. An additional saw for a quicker handling of the cutting process and conveyors for back and forth transport of the mouldings to rework improve the LoA of the handling process from (1; 1) to (5; 4). Figure 4-11 illustrates the suggestions implemented in the current layout.

Due to a quicker processing time within rework, savings with regard to the cycle time are achieved which increases the productivity ratio is a countermeasure against the emerge of intermediate buffers. Also the ergonomic aspects play an important role here, as harmful regular lifting of material above shoulder height is avoided which effects employees strain and concentration in a positive way.

Page 74: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

64

Figure 4-11: Layout of a rework loop to improve LoA in material handling

Suggestion 3: Error/ Boundary images of parts

This proposal includes the establishment of A - and B - quality mouldings. Material arriving at the in-line moulding rack in front of the rework station can be sorted by the operator according to pre-defined criteria which are fixed in an operator’s guideline. Error patterns and boundary parts are illustrated on boards which make the grading activity clear. Mouldings with minor deficiencies like small edge damage or surface unevenness can be collected without rework and hold back for further processing as B-quality material. This can be sold for a lower price to the customers.

The data collection presented in subsection 4.4.1 reveals that the quality assessment of different operators regarding which mouldings to put back into the system differs to a high extent. Therefore, common standards which deficiencies are considered necessary for reworking and which parts can be regarded as B-quality material. The suggestion improves the cognitive automation level of the quality inspection task (from 1 to 3). A higher material utilization results from this proposal, as the number of mouldings to be cut during rework decreases.

Suggestion 4: Parallel rework process loop

In a similar meaning as suggestion 2, the installment of a second puttying machine for reworking purposes enables a parallel processing of mouldings in the reworking loop. Figure 4-12 shows a parallel loop of puttying and cutting equipment next to the processing line A. The puttying station has to be equipped with a selective puttying application device which ensures that eventual cracks are covered with a layer. The following changes regarding tasks in the reworking process are achieved (LoA levels in brackets):

Transport from/to the in-line moulding rack (1; 1) (5; 4)

Puttying application in rework (2; 1) (5; 4)

An improved productivity is reached with this approach, as the cycle time within the rework loop decreases and relevant buffers can be processed with less manual tasks.

Page 75: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

65

Figure 4-12: Suggestion 4 - Parallel processing of rework activities with transport conveyors and puttying machine

As already indicated in suggestion 4, the combination of effective proposals for reworking reveals improvements regarding the LoA scale in several tasks of the bottleneck station. With the reworking process in line A, an additional scanner feature with a direct part marking function next to boards with error patterns and the possibility to sort out B - quality material in combination with a parallel rework loop shows the LoA changes illustrated in table 4-17.

Suggestion 5: Automatic feeding of material

The material feed process on the conveyor is done manually with an operator turning each moulding before placing it on the conveyor. An elevator system which is steered with a control panel makes it possible to free the operator from this monotonous and repetitive activity. The corresponding Level of Automation within the feeding process is raised from (1; 1) to (4; 1), as the elevator system represents an automated tool with manual steering. Also from an ergonomic point of view this measure represents a facilitation of the operator’s work. It allows the operator to focus more on quality inspection of the incoming raw material.

Suggestion 6: Scanner inspects input material

The installment of a second scanner for quality inspection purpose of the incoming material is likely to have a positive effect on the material flexibility. The grading of mouldings before the planing process into A-, B- and C categories enables to extend the material flexibility and achieve an improved material utilization, as less material is cut away in this case. Figure 4-13 illustrates the recommendation in a schematic way with A quality products being further processed.

Page 76: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

66

Figure 4-13: Schematic layout for a moulding selection according to quality levels before the first processing step

The anisotropic nature of the wooden material leads to varying quality of the lots, which is observed during the Values Stream Analysis. A visual control of the operator who conducts the feeding task currently aims to filter out the worst deficiencies. A more thorough control with a non-contact scanner leads to a considerable increase in the LoA scale from (1; 1) to (6; 6).

Suggestion 7: Quality inspection scanner placed after planing

The planing process shapes the mouldings’ profile and therefore needs a consistent quality assurance in order to ensure a smooth processing in the remaining process steps. As it is situated in the beginning of the processing line A, the detection of defective parts only at the scanner would represent a waste.

Therefore, it is recommended to install a second scanning system after the planing process which ensures the conformance of the dimensional and geometrical tolerances with the drawing specifications. Non-contact measuring technique such as laser focusing, interferometry or optical triangulation can perform fast inspections without interfering the current manufacturing process. Within the present state of line A, the operators perform quality checks with random samples with the mouldings after the planing process. During the observation phase the researchers experienced the necessity of this measure when a long series of defective parts due to a faulty planing tool where identified only at the scanning process, which does not enable fast countermeasures.

This improvement measure increases the LoA of the quality assurance performed at the flipping station directly after the planing process in a considerable way, from (1; 1) to (6; 6). An improved material utilization results from this suggestion, as corrective measures can be implemented at once as soon as the scanning system displays a faulty part.

Page 77: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

67

Suggestion 8: Adjustment device for spray pressure and angle

The priming and top coating process are triggered with sensors which give a signal with a passing moulding. The spraying jet is coupled to this signal and apply paint on the parts with the pressure and the angle of the pistols being adjusted manually.

With a coupled device controlling and adjusting those parameters according to the product being processed, an improved material utilization is reached. The consumption of paint in line A is a factor which influences the overall productivity and the quality level. With an adjustment device which sets up spraying pressure and angle according to pre-defined data about product dimensions, the issue of using too much or too less paint is solved by applying higher Levels of Automation.

The respective priming and painting processes in both spraying chambers raise their Level of Automation from (5; 4) to (5; 6). The technology is able to supervise the coating of paint and takes corrective action meaning parameter adjustment in case of a deviation. As this automation level has an intervening character, the assessment according to LoAcog 6 is justified.

Suggestion 9: Surveillance camera in primer station

In order to monitor the painting process, a camera is installed near the chamber which is coupled to a monitor at the individual labelling station. A quality assurance task with an operator takes place here. This monitor represents a static tool, which supports the monitoring of the top coat application, as a disturbance can be identified by the operator without having to move to the spraying chamber. This ensures a faster problem identification in case of line stoppages and has an influence on the productivity level. Short stops occurring in both spraying chambers are ranked among the top five taking into account an analysis of the aggregated time loss due to disturbances at various station in line A.

It is therefore recommended to install a second camera and monitor which facilitate the surveillance of the priming process, especially as this spraying chamber is located more apart from the work stations where the operations are continuously involved in work activities. The task ‘monitor the process’ of the primer can receive a higher Level of Automation with this simple measurement, which increases the task from level (1; 1) to (2; 1).

Suggestion 10: Fault-symptom matrix

Occurring disturbances within the plastic wrapping and stacking of the mouldings can have severe consequences for the availability of the machine. The high temperature in the oven damages the material in the case of a stop and the accumulated line stoppage time due to the stacking machine is amongst the highest of all work stations in line A. An alarm at the wrapping station calls for attention if the process is interrupted. The operators have to react quickly if disturbances occur and are dependent on their manual abilities as well as no cognitive support how to solve the problem.

It is recommended to develop a so-called fault-symptom matrix for these critical processes. This cognitive support manual can also be extended other relevant processes where disturbances and line stoppages have a major impact on the

Page 78: Automation in the Wood Processing Industry - DiVA portal

Findings and analysis

68

OEE–figure. A fault-symptom matrix (FSM) is described by Embrey (2009) is a strategy to tightly focus on knowledge which is readily applicable to support scenarios of malfunctions. In an FSM, each row refers to a failure category and the column stand for various symptoms which are accessible for the operator. Table 4-19 illustrates the example of such a matrix as document of failure diagnostics which has a use for training purposes.

Table 4-19: Schematic Fault - Symptom matrix for the packaging processes 1 – Relation, 0 – No relation

Symptoms

No. Scenario

Plastic wrapping machine alarms

Stacking machine blocked

Plastic wrapping - edge offset

blocked

Stacking machine -

edge offset blocked

1 Remaining plastic

residue 1 0 0 0

2 Uneven edge offset 0 0 1 1

3 Short moulding in the bottom of the stack

0 1 0 1

4 Moulding is bent 1 0 1 0

5 Insufficient grip of the

mouldings 1 0 1 0

6 Mouldings are fed

askew 1 1 0 1

The development of such a matrix should use the inputs of the operators from the successful management of previous instances of a specific scenario (Embrey, 2009). The Level of Automation of the work task ‘disturbance handling’ is therefore raised from (1; 1) to (1; 3).

Page 79: Automation in the Wood Processing Industry - DiVA portal

References

69

5 Discussion and conclusion

The final chapter in this thesis aims to discuss the work conducted in a structured way with a detailed answer of each research question. The choice of the applied methods during this study is discussed in section 5.1 and the research questions are discussed in the following section 5.2. Validity and reliability as well as triangulation aspects within the frame of the research strategy are also debated. The next section discusses how the work can be further improved. A short conclusion constitutes the closing end of the thesis.

5.1 Method discussion

The following section deals with a discussion about the methods which are applied in the course of the research process. The traditional and systematic literature reviews are evaluated based on the quality of the findings and the problems which occurred during the method application. A discussion of the ‘Applied DYNAMO++’ method concludes this section.

5.1.1 Discussion of traditional literature review

In retrospect, the choice of the traditional literature review is an appropriate means to give an initial overview about the target industry and its characteristics affecting automation. Regardless of the multitude of product segments which made it difficult to find a suitable term for the industry, the high degree of freedom and flexibility of the traditional literature review supported in finding relevant articles. Furthermore, the fact that both researchers had little experience dealing with the wood processing industry, the exploring nature of the traditional literature review was appropriate when a clear picture of the subject of study – in this case the industry itself – was not very detailed.

As a research gap was concluded concerning the wood processing industry and automation, the results of this review justify why alternative ways in finding suitable articles later during the systematic literature review had to be pursued. A subjective assessment according the article relevance corresponds to the nature of the traditional literature review. Also the selection of a number of keywords framing the target industry simplified the search.

Regarding the quality of the findings which were drawn out of the traditional literature review, it can be said that a several authors stating similar views and describing the same developments within the industry is advantageous in terms of triangulation aspects. The recurring description of characteristics affecting automation issues also prepared for the case study, as an extensive preparation stage was included here prior to the LoA assessment in phase II of the ‘Applied DYNAMO++’.

5.1.2 Discussion of systematic literature review

The systematic literature review undertaken to figure out the most appropriate method also includes the findings of the traditional literature review. Consequently, the method evaluation and the reasoning for choosing the ‘Applied DYNAMO++’ method was partly based on the relevance regarding the target industry. As an initial keywords search (figure 3-4) reveals, a research gap could be concluded concerning the concept Levels of Automation and the application in the wood processing

Page 80: Automation in the Wood Processing Industry - DiVA portal

References

70

industry. This indicated that the most appropriate method has to be found on a more general level. A broadening of the review frame also yielded difficulties regarding the rigorous manner in which the literature review was supposed to be performed.

However, a quite narrow focus was established after applying relevant inclusion/ exclusion criteria and screening the database server ‘Primo’ for a number of keywords. Author related research with screening of references used in the respective articles revealed a pool of suitable papers for the review. In a retrospective, the focusing on a few key authors who regularly were cited by other authors working with similar methods created a profound understanding of how the methods relate to each other, which the underlying concepts are and which dimensional perspective they represent.

The quality of the data drawn out from the articles into an extraction protocol was strengthen by the key authors contributions being present in several other articles. This created a quite deep understanding of the respective methods’ purpose, procedure and the underlying background. A high proportion of reviewed articles dealing with the DYNAMO++ concept is justified by the availability of the doctoral theses of the Swedish researchers involved in the DYNAMO project. Another reason is the need for getting acquainted with the method by studying articles dealing with case examples. In general, the systematic procedure, especially the data extraction step supported in finding and grouping the various assessment methods, for example through the application of inclusion and exclusion criteria.

5.1.3 Discussion of ‘Applied DYNAMO++’

A review on the method used for assessing Levels of Automation in a production system revealed key points for the applicability in the context of a processing line. Furthermore, the conduction of the four-phase model of DYNAMO++, which included a slight modification in phase II with the additional assessment of Levels of Competence and Levels of Information, allowed to point out its strengths and weaknesses.

Performing a Value Stream Analysis in phase I is considered an ideal step for gaining a profound understanding of the material and information flow as well as to collect data about relevant time and machine parameters such as performance figures. Yet, the inclusion of the reference system line B in order to be able to compare the actual findings with an alternative processing principle made phase I time consuming. Especially as data collected at line B was not intended to be used for the development of Square of Possible Improvement (SoPI) matrices or improvement suggestions, the data collection was not necessary to such a comprehensive extent. The thorough data collection at line A enabled the identification of critical processes in phase III and established process and task knowing for the LoA assessment.

During the second phase of the ‘Applied DYNAMO++’ method, which encompassed a Hierarchical Task Analysis (HTA) as well as the measurement of LoA, LoC and LoI, the acceptance and approval concerning the recording of observed tasks was given. This facilitated the later analysis of the LoA measurement if something remained unclear.

Page 81: Automation in the Wood Processing Industry - DiVA portal

References

71

A discrepancy between the organizational practice on the shop floor and the taxonomy concept illustrated in table 2-2 emerged during the measurement phase. This was mainly related to the cognitive automation level, i.e. how to assess the current task with the operator role model in the LoC concept (table 2-8).

The distinction between skill -, rule – and knowledge based activities as well as the assessment regarding the information content in the LoI concept was complicated due to missing work instructions and checklists. It can be concluded that, once the system borders were fixed, a determined scheme needed to be followed in phase II which funneled the specific work task allocation into the continuum schemes of the LoA taxonomy.

However, there was a big acceptance by the operators to tell about own skills and experience in the job which supported the assessment in the measurement scales considerably. Regarding the refinement of the LoA measurement, it is recommended to get advice from an expert who can provide knowledge, as it was the case in this study. The results from the measurement phase indicate that a one-sided focus was lying on the mechanical automation level so far, as the LoAmech was considerably higher than the LAcog for both line A and B (see tables 4-8 and 4-9).

Regarding phase III, it has to be mentioned that the DYNAMO++ method provides no guidance in determining the critical processes for the workshop. The matrix developed in table 4-14 combines the processes with their underlying tasks and the assumed competitiveness factors of the wood processing industry, but DYNAMO++ leaves space for self-determination. Furthermore, there is no guidance provided in which direction to improve the tasks. The underlying performance criteria which make up the decision for minimum and maximum levels can be arranged according to own evaluations which is also discussed by Fasth et al. (2007). This fact allows an adaption to an industry or company specific focus in terms of improvement. However, the participation of different employee groups regarding experience, field of expertise and job role ensure a considerably higher acceptance of the measures discussed during the workshop.

Discussing phase IV of the ‘Applied DYNAMO++’ method, the application complexity assessment ‘medium’ is realized when being confronted with the implementation of suggested improvements. The resources for taking further measures to execute some proposals were not given in the case study from the beginning. This leaves the researchers with a presentation of the suggested improvements at the case company.

Previous descriptions about DYNAMO++ deal with its application in assembly lines, i.e. (Fasth et al., 2008; Fasth & Stahre, 2011). Frohm et al. (2009) describe an example of a LoA assessment in an assembly system of the Swedish furniture industry. The supervision of automated machines and robots as well as quality inspections are the main activities in that study which results in similar LoA compared to the situation in the case study of this thesis. A majority of tasks performed by the operators in line A had to be classified in the ‘monitor’ range of the LoC scheme which is shown in table 4.10. Yet, these results do not refer to a system where several part components are used for assembly.

Page 82: Automation in the Wood Processing Industry - DiVA portal

References

72

The processing line of the case study included a different material flow as well as other product complexity and the routine of the operators was fixed to a lower degree. As it was perceived during the case study, occurring disturbances in line A made flexible corrective actions of one or several operators necessary. Furthermore, frequent relocation between different workstations were common practice for the operators, which also distinguished line A to the assembly systems studied in prior cases.

The application of the LoC concept in phase II reveals a limited scope of operators’ roles considering the different work tasks. This limitation mostly refers to ‘monitor’- and ‘intervene’ tasks and allows only minor conclusions regarding the contribution to improvement suggestions. Probably due to the wide spread one-sided perspective on the implementation of automation technologies, which focuses solely on the mechanical/ physical view, a more passive role of operators was accepted with prior installment of automation technology. The inclusion of the LoI scheme revealed that a majority of the work tasks within line A require expert knowledge (table 4-12). A high system complexity due to a number of coupled work stations in the processing line and an increased cognitive workload due to the probability of catastrophic failures can be seen from this. The incorporation of this concept in the method points out in general the need for more cognitive support of work activities, as it is proposed in the findings subsection 4.3.4.

The ‘Applied DYNAMO++’ method as it was used in the case study combines quantitative and qualitative data collection techniques and has a two-sided view on mechanical and cognitive aspects which none of the other methods provided and which made it the most favorable method to choose. This allows a holistic understanding of the work activities and elaboration of improvement suggestions based on the widely acceptance of the involved personnel. A degree of adaptability towards various improvement targets based on specific performance criteria enables the application across industry borders and firms’ sizes. Yet, the comprehensiveness from planning stage to the implementation of solutions requires high effort and sufficient resources. Furthermore, the method was mostly verified in production system with a more active operator role, i.e. in assembly systems. The limited action space regarding operators’ choice of actions and material flow have to be considered when using the method in an industry case.

5.2 Discussion of findings

The discussion of this section regards the answering of the research questions and to what extent the research questions have been answered.

5.2.1 Research question 1

When penetrating the first research question, the term methods for measuring LoA and the reference to the wood processing industry set a tightly specified frame:

RQ1: What method is most appropriate to be applied with measuring Levels of Automation in the wood processing industry?

Section 2.1 contains the theoretical background for getting acquainted with the characteristics of the wood processing industry which affect the ability to automate

Page 83: Automation in the Wood Processing Industry - DiVA portal

References

73

production processes. This information can be seen as reproduction of the content of the traditional literature review which reveals several aspects. The structure of the section with seven subdivisions shows the broad perspective which is taken in order to gain a holistic picture of the characteristics. Section 4.1 presents the most prominent factors dealing with the wood processing industry and automation which were figured out during the review. It can be said that some of these factors have a function as enabler of automation investments such as the strive for productivity improvement to reach a more efficient production. Others restrain the implementation of further automation technologies for material handling or quality inspection purposes such as the anisotropic material character and a rather small firm size in average.

Based on this knowledge, a method which focuses on single task operations in production, which takes into account the operator’s skills regarding the assessment of material state and process conditions has a higher relevance during the method evaluation part. The adaptability of the method regarding how to improve production processes was a further factor which played a role. The extraction protocol of the systematic literature review acted as data pool where necessary information regarding the industry in focus, eventual case examples and the strengths and weaknesses were available. Therefore, the required breadth and profundity to be able to find the most suitable method was established. As relevant case studies dealing with measuring Levels of Automation in the wood processing industry are not present in the reviewed literature, a method selection focusing exclusively on this industry breaks fresh ground and represents a contribution for future case studies.

The choice of relevant assessment criteria for a holistic quality verification was based on different reasons. The first criterion, as already mentioned, is the relevance to the wood processing industry with its particular characteristics which is motivated in section 4.1. The choice of the second criterion, the application complexity referred mainly to the investigators’ resources and skills within the case study designed for the method implementation. To consider dimension view as well as operational and space level as third and fourth criterion was a decision inspired by Fasth (2012), who performed a review about assessment methods for task allocation in a production system earlier. Assessment objectives as last criterion illustrates the impact direction of the method, which correlates with the factors for competitiveness in the wood processing industry given in subsection 2.1.7.

Productivity figures such as OEE and material deficiencies are most prominent in this industry, which argues for the choice of a quantitative LoA assessment method. Yet, the already mentioned importance of human-machine interaction in the context of the industry profile made the inclusion of qualitative-cognitive studies necessary. Having in mind these two aspects, the reasoning for DYNAMO++ as only method combining the engineering approaches of a technical-physical and socio-cognitive focus becomes clear.

Section 4.2 presents the motivation for the selection of a modified version of DYNAMO++ in detail, which is discussed in the method discussion subsection 5.1.3. Yet, the decision for a modification of the method was based on the

Page 84: Automation in the Wood Processing Industry - DiVA portal

References

74

researcher perception during the literature study that competence and information aspects are becoming increasingly important when automating processes (see also subsection 2.2.1).

RQ1 has been answered by an in-depth evaluation of pre-defined assessment criteria regarding reviewed methods from a systematic literature study, which is summarized in table 4-3. The consideration of relevant triangulation methods such as the involvement of two researchers and different underlying theoretical concepts like the two diverging literature review strengthen the reliability of the results. Furthermore, the collection of relevant statements from the articles regarding the industry profile and a comprehensive data extraction protocol verified a transparent procedure within both literature reviews.

5.2.2 Research question 2

The second research question aimed to establish a connection in industrial practice between the measurement method for LoA, figured out in the systematic literature review, and the wood processing industries necessity for more automated production processes, as it is concluded from the traditional literature review. In order to achieve meaningful results, a case study was designed involving a major Swedish wood processing company, which collaborated in the following question:

RQ2: How can the measurement of Levels of Automation contribute to an improved competitiveness of the wood processing industry’s production processes?

Examining LoA at a firm within the wood processing industry will hardly solve general issues like a low value adding of wooden products. Yet, the concept offers options to rethink work station’s design to achieve more attractive work environments. Implications regarding the competitiveness of the production processes can be broken down to improved efficiency and quality figures.

Several facts gained during the case study indicate that the focus of the wood processing industry regarding the implementation of automation technology is on the mechanical level (LoAmech). During the workshop in phase III of the Applied DYNAMO++ methodology, it was perceived as difficult to relate the need for a higher automated processes to both the mechanical and cognitive level. It became clear that many practitioners mostly regard the technical-physical dimension as important when discussing solutions.

The aspect that operators in a higher automated production system are typically left with passive roles including observing processes and managing of disruptions is described by (Osvalder & Ulfvengren, 2009). Yet, the accumulation of work tasks in the studied production line regarding ‘monitor process’ and ‘disturbance handling’ with LoA (1;1) indicates that the aspect of the human integration into automation concepts is neglected. Considering this, the measurement of LoA can contribute to an increased awareness of the operator’s role within the production system to strive for a higher OEE. The improvement suggestion 3 (error pattern boards and boundary parts) and 10 (Fault – Symptom Matrix) presented in subsection 4.4.4 point out possibilities to raise the LoAcog of operator driven work tasks.

Page 85: Automation in the Wood Processing Industry - DiVA portal

References

75

The measurement of LoA can also achieve the standardizing of working procedures which contribute to a stable quality level in production. During the case study data collection at the bottleneck station has revealed that, due to the lack of binding instructions, operator’s experience plays a major role in quality inspection. As it can be concluded from conversations with the operators on the shop floor, work within the wood processing industry might be seen more as craft than automated machining.

Examining LoA in this industry points out where the user’s experience and knowledge (LoAcog 1) of a process can be extended with work orders (LoAcog 2) or manuals (LoAcog 3). Following this approach, the effects of rapid turnovers of workforce as described by Karltun (2007) can be mitigated.

Operator involvement is a key aspect when designing production systems and determining task allocation with a high Level of Automation. By applying a more holistic perspective which counteracts an impression of ‘skill degradation’, an improved acceptance among the shop floor personnel is achieved. The possibility to create solutions on a two-scaled continuum from manual to automatic gives the operators the certainty that new automated processes are not designed without the consideration of their skills.

The anisotropic structure of wooden raw material requires necessarily human involvement, i.e. during rework or quality inspection processes. The configuration of these work tasks with appropriate ergonomics and cognitive support tools such as clear guidelines and an easier monitoring of automated running operations ensures a higher attractiveness of the wood processing industries’ workplaces.

To sum up, the case study has revealed paths how the measurement of Levels of Automation can lead to an improved competitiveness within a representative firm of the wood processing industry. Issues discussed such as workplace attractiveness and cognitive workload as well as the importance of human – machine task allocation in terms of the heterogeneous material properties point out that aspects for generalization have been identified. These conclusions allow the judgment of research question 2 having been answered as far as the limited scope of this thesis work allows for that.

5.3 Validity and reliability

This section discusses to what extent validity and reliability have been achieved in this study. The results obtained from the literature reviews and the testing of the ‘Applied DYNAMO++’ method can be assessed from a quality perspective by analyzing their validity and reliability respectively.

Internal validity was achieved by using triangulation assessment (described in section 5.4) and by assuring that the required factors were in fact those being observed and analyzed. Authors describing similar circumstances regarding the wood processing industry during the traditional literature review point out that internal validity is achieved. Appropriate keywords which were detached from the industry profile during the systematic literature review led to a general method collection with a later funneling regarding the relevance towards the industry. This prevented the overestimation of single case results.

Page 86: Automation in the Wood Processing Industry - DiVA portal

References

76

Document analyses, observations of crucial tasks and events in the case study contributed to that. Furthermore, by testing the modified methodology at two production lines with varying processing principles and products, internal validity was increased.

External validity was achieved as well, given that the LoA assessment method DYNAMO++ is verified in many various industries with similar results. The LoA assessment method has been used by several scholars previously, thus it can be seen as appropriate for measuring Levels of Automation. Generalization aspects are mostly limited to the scope of the case study, but by including particular characteristics of the wood processing industry, conclusions can be drawn which affect the whole industry’s potential to improve production processes with the LoA concept. Furthermore, the statement protocol and the data extraction protocol verify a transparent procedure during both literature reviews.

Construct validity is not achieved to the same extent due to DYNAMO++ not being designed especially for the wood processing industry. Furthermore, a number of previous validations deal with assembly lines (see also subsection 5.1.3). However, judging the method implementation on task level it performs as expected. The ‘Applied DYNAMO++’ methodology designed for the purpose of this thesis works well when answering the posed RQ2.

The traditional literature was a suitable choice when exploring the industry’s characteristics, as it is described in subsection 5.1.1. A systematic procedure when reviewing LoA assessment methods ensured a transparent decision making which includes well-defined evaluation criteria and is easy to track. Therefore, subsection 5.1.2 motivates why this data collection method was the most suitable in the context of this thesis.

Reliability is achieved due to the various data collection techniques utilized in this research. Research methods consisting of both theoretical and empirical approaches such as literature reviews and a number of common data collection and analysis techniques such as VSM and HTA ensure a trustworthy procedure. Detailed transcriptions of the results from the traditional literature review as well as the data extraction protocol for the systematic literature review establish reliability in the context of the research strategy presented in section 3.1.

5.4 Triangulation assessment

Four triangulation categories are discussed in the subsections below as to what extent they aid in increasing the validity and reliability of the thesis.

Page 87: Automation in the Wood Processing Industry - DiVA portal

References

77

5.4.1 Data triangulation

By exploiting various data sources, i.e. collecting both qualitative and quantitative data, data triangulation is achieved. Regarding the traditional literature review, the provision of different data sources is not a primary focus. But due to the fact that a research gap exists concerning automation and the wood processing industry, a review consisting of various authors describing research results from different countries is conducted. Difficulties regarding the extent of generalization of the described characteristic for the whole wood processing industry are remedied by reviewing authors stating similar conditions and perceptions. The systematic literature review uses the servers of both ‘Primo’ and ‘Google Scholar’ as well as pre-defined inclusion/exclusion criteria. This ensures a variety of authors and methods being reviewed.

Document review and observation techniques are used extensively during the case study. Next to the collection of statistical data, i.e. about the rework process, a focus is on informal conversations with operators who share their experiences and opinions. Together these make out the data base during the conduction of the ‘Applied DYNAMO++’ method. The workshop conducted during phase III utilizes the experience of individual workers at various ranks within the hierarchical system in order to generate data of qualitative nature, i.e. user perception regarding the critical processes in production line A and how this can be improved.

5.4.2 Investigator triangulation

The involvement of researchers with varying knowledge and experiences creates researcher triangulation and assists to increase the internal validity and reliability of the report. For this reason, the research is conducted in collaboration with a Ph.D. student from the department Industrial Engineering and Organization at Jönköping University. The assessment of work tasks in the LoA scale provides difficulties to unexperienced researchers, especially since there is a lack of LoA implementation in the wood processing industry. To mitigate this effect, an experienced researcher gives advices in phase II of the ‘Applied DYNAMO++’ procedure.

5.4.3 Theory triangulation

Theories from various authors working within different scientific areas have been used in order to attain theory triangulation. Exemplary, the dimension schools of socio-cognitive and technical-physical engineering can be referred. However, the lack of information regarding the implementation of LoA in the wood processing industry has affected the research.

5.4.4 Methodology triangulation

Table 3-1 shows that the research questions are targeted with different methods to ensure a thorough study of the phenomenon ‘LoA in the wood processing industry’. The use of both quantitative and qualitative methods strengthen the methodology triangulation within the ‘Applied DYNAMO++’. Furthermore, the systematic literature review provides overview and generates a comparison between several methods according to their applicability within the wood processing industry. The method which was deemed most suitable was modified and used within a case study.

Page 88: Automation in the Wood Processing Industry - DiVA portal

References

78

5.5 Suggested future research

Results from the traditional literature review point out the aspect of quality standards within the processing of wood in secondary industries such as the case company as highly important for improved process automation and optimization. Therefore, the relations between wood specific quality criteria and automation technologies need to be studied. Especially as several in-line scanning options are proposed to the case company, the future applicability of these technologies to wooden surfaces is of interest.

This thesis shows also that there is a lack of study regarding the implementation of the LoA concept in the wood processing industry. The research gap identified in section 4.1 encompasses the wood processing industry and its ability to design levelled automated processes beyond the plain separation according to automated, semi-automated and manual. A contribution of the study conducted at the case company is that improvement suggestions are shown which reveal paths leading to an improved competitiveness of the wood processing industry’s production processes. Further research can supplement these results with studies conducted in other product sectors or production system (i.e. assembly system or in a functional layout) of the wood processing industry.

Taking into account the DYNAMO++ method, it can be concluded that an implementation of the method in a production system where the operator has a more active role than in the processing line studied during this project limits the space of possible improvements. Therefore, method modifications regarding the ability to implement DYNAMO++ on a system level rather than task level are an alternative consideration when aiming for competitiveness enhancement. This provides also a better holistic overview of a production system’s LoA.

In order to cover ergonomic issues related to the working conditions in the wood processing industry, the DYNAMO++ method could be extended by assessing the physical work load. Exemplary, the improvement suggestions 2 and 5 presented in subsection 4.4.4 focus on the elimination of hazardous work task. No job rotation during shifts lead to a situation where an operator performs an activity which can lead to tensions in arms, neck and back. Numerous methods for assessing physical workload exist, i.e. ‘Rapid Upper Limb Assessment’ (RULA) and ‘Rapid Entire Body Assessment’ (REBA) which are easy to implement and which provide a solid situation analysis. A modification of the DYNAMO++ method in phase II with the inclusion of a physical workload assessment method mentioned from above reveals improved results regarding how to argue for improvement measures. Further studies concerning the combination of LoA and physical ergonomics within the wood processing industry are therefore recommended.

5.6 Conclusions

The aim of this thesis was to examine to what extent existing methods for assessing Levels of Automation in production processes are applicable in the wood processing industry. Literature reviews give a description of the specific industry profile regarding automation of production processes and suitable assessment methods for LoA. The concept of Levels of Automation was then implemented in

Page 89: Automation in the Wood Processing Industry - DiVA portal

References

79

a case study to find out how it can contribute to an improved competitiveness. By using a modified version of the DYNAMO++ approach, several contributions can be made regarding the applicability of the method DYNAMO++ and the LoA concept in general:

The main competitiveness factors for the wood processing industry in terms of production processes are productivity and a stable quality level.

The industry’s workplaces in production are considered as under automated with a number of handicraft activities.

The ‘Applied DYNAMO++’ method for measuring LoA on a task level context involves the operators in finding appropriate improvements and achieves a considerable acceptance and understanding.

The DYNAMO++ method is thought rather to be applied in an assembly system, which provides more active operator roles and complexity regarding converging product flows and therefore a lower proportion of monitoring activities.

The results from the measurement of LoA on a two-sided scale reveal that cognitive aspects regarding disturbance handling, process monitoring or quality inspection are neglected towards technology-focused solutions, which consider mostly the technical feasibility of fully automated processing.

Improvement potential is assumed regarding the physical ergonomic conditions which reveal a lower sickness absence rate and thereby improved productivity.

Due to the anisotropic nature of wood, high LoA for crucial quality inspection tasks (i.e. with scanning systems) can be only considered in context of a sufficient reliability and process capabilities, which points out the need for a well-balanced task allocation including operators’ reasoning skills in a coupled human – machine process solution.

The pathways for an improved productivity and quality level with the LoA concept are given by the standardization of working procedures and an improved ergonomic workplace design described in the improvement suggestions in subsection 4.3.4. The LoA concept is therefore an appropriate tool for directing the investment considerations in new process technology within the wood processing industry towards a balanced task allocation.

This includes a decent cognitive workload with the ability to have full awareness of the situation without perceiving too high stress levels and the incorporation of automated processing and quality checking technologies brought in-line with the unique heterogeneous material properties of wooden raw material.

Page 90: Automation in the Wood Processing Industry - DiVA portal

References

80

References

Almström, P., & Kinnander, A. (2007). Productivtiy Potential Assessment of the Swedish Manufacturing Industry. Chalmers University of Technology, Department of Materials and Manufacturing Technology. Göteborg: Chalmers University of Technology.

Armitage, A., & Keeble-Ramsay, D. (2009). The Rapid Structured Literature Review as a Research Strategy. Education Review, 6(4), 27-37.

Bellgran, M., & Säfsten, K. (2009). Production Development: Design and Operation of Production Systems. London: Springer London Ltd.

Bisgaard, S. (2008). Quality Management and Juran's Legacy. Quality Engineering, 390-402.

Bohgard, M. (2009). Work and technology on human terms (1st uppl.). Stockholm: Prevent.

Bryman, A., & Bell, E. (2011). Business Research Methods. New York: Oxford University Press.

Buehlmann, U., Bumgardner, M. S., & Sperber, M. (2013). How Small Firms Contrast with Large Firms Regarding Perceptions, Practices and Needs in the U.S. Secondary Woodworking Industry. BioResources, 8(2), 2669-2680.

Bumgardner, M. S., Buehlmann, U., Schuler, A., & Crissey, J. (2011). Competitive actions of small firms in a declining market. Journal of Small Business Management 2011, 49(4), 578-598.

Comstock, M. (2004). Production systems for mass customization - bridging theory and practice. Linköping University, Department of Mechanical Engineering. Linköping: Linköping University.

Comstock, M., & Bröte, S. (2005). Beyond 'Read a plant - fast' (for lean): read an enterprise for mass customization? Linköping University, Institute of Technology, Produktionssystem, IKP. Linköping: Linköping University.

Connors, M. M. (1998). Teaming Humans ans Automated Systems in Safely Engineered Environments. Life Support & Biosphere Science Vol. 5, 453-460.

Deming, W. E. (1982). Out of the crisis. Cambirdge: Massachusetts Institute of Technology , Center for Advanced Engineering Study.

Dencker, K., Fasth, Å., Stahre, J., Martensson, L., Lundholm, T., & Akillioglu, H. (2009). Proactive assembly systems-realising the potential of human collaboration with automation. Annual reviews in control, 230-237.

Diaz-Balteiro, Heruzo, A. C., Martinez, M., & Gonzalez-Pachon, J. (2006). An analysis of productive efficiency and innovation activity using DEA: An appliation to Spain's wood-based industry. Forest Policy and Economics, 762-773.

Dubois, A., & Gadde, L.-E. (2002). Systematic Combining: An Abductive Approach to Case Research. Journal of Business Research, 55(7), 553-560.

Edwards, K., & Jensen, P. L. (2014). Design of systems for productivity and well being. Applied Ergonomics, 26-32.

Eisenhardt, K. M. (1989). Building Theories from Case Study Research. The Academy of Management Review, 532-550.

Eliasson, L. (2014). Some aspects on quality requirements of wood for use in the industry manufacture of single-family timber house. Växjö: Linnaeus University Dissertations.

Ellis, P., & Davies, H. (2000). Porter's Competitive Advantage of Nations: Time for a final judgement? Journal of Management Studies, 1189 - 1212.

Embrey, D. (2009). A human factors approach to managing competency in handling process control disturbances. Institution of Chemical Engineers, 440-446.

Endsley, M. R. (1996). Automation and situation awareness. i R. Parasuraman, & M. Mouloua, Automation and human performance: Theory and applications. Mahawah, USA.

Endsley, M. R., & Kaber, D. B. (1999). Level of automation effects on performance and workload in a dynamic control task. Ergonomics, 462 - 492.

Page 91: Automation in the Wood Processing Industry - DiVA portal

References

81

Fässberg, T. (2012). Cognitive Automation in Mixed-Model Assembly Systems: Current and Future Use in Automation Industry. Chalmers University of Technology, Department of Product and Production Development. Göteborg: Chalmers University of Technology.

Fasth, Å. (2012). Quanitfying Levels of Automation - to enable competitive assembly systems. Chalmers University of Technology, Department of Product and Production Development. Gothenburg: Chalmers University of Technology.

Fasth, Å. (2012). Reviewing Methods for Analysing Task Allocation in a Production system. Journal of Logistics Management, 1-8.

Fasth, Å., & Stahre, J. (2008). Does Levels of Automation need to be changed in an assembly system? - A case study. Göteborg: Department of Product and Production Development, Chalmers University of Technology.

Fasth, Å., & Stahre, J. (2011). Task allocation in assembly systems - Measuring and analysing Levels of Automation. Jounal of Theoretical Issues in Ergonomics Science.

Fasth, Å., Stahre, J., & Dencker, K. (2008). Measuring and analyzing Levels of Automation in an assembly system. Proceedings of the 41st CIRP International Conference on Manufacturing Systems (ICMS). Tokyo, Japan.

Fasth, Å., Stahre, J., & Frohm, J. (2007). Relations between parameters/ performers and levels of automation. IFAC.

Fasth, Å., Stahre, J., & Frohm, J. (November 2007). Relations between parameters/performers and levels of automation. IFAC workshop on manufacturing modelling, management and control.

Fasth, Å., Stahre, J., Bruch, J., Dencker, K., Lundholm, T., & Mårtensson, L. (2010). Designing proactive assembly systems (ProAct) - Criteria and interaction between automation, information, and competence. Asian International Journal of Science and Technology in Production and Manufacturing Engineering (AIJSTPME), 2(4), 1-13.

Fasth-Berglund, Å., & Stahre, J. (2013). Cognitive automation strategy for reconfigurable and sustainable assembly systems. Assembly Automation, 33(3), 294-303.

Frohm, J. (2008). Levels of Automation in Production Systems. Chalmers University of Technology, Department of Product and Production Development. Göteborg: Chalmers University of Technology.

Frohm, J., Granell, V., Winroth, M., & Stahre, J. (2006). The industry's view on automation in manufacturing. Proceedings of the 9th symposium IFAC on Automated Systems based on Human Skills and Knowledge. Nancy, France.

Frohm, J., Lindström, V., & Bellgran, M. (2005). A Model for parallel Levels of Automation within Manufactuing. Proceedings of the 18th International Conference on Production Research. Salerno, Italy.

Frohm, J., Lindström, V., Stahre, J., & Winroth, M. (2009). Levels of Automation in Manufacturing. Ergonomia.

Goodreau Sawmill & Woodworking. (2016). Crown Mouldings. Hämtat från Goodreau Sawmill & Woodworking: http://www.goodreausawmill.com/h/crown_mouldings

Goodson, E. (2002). Read a Plant - Fast. Harvard Business Review.

Gorlach, I., & Wessel, O. (2008). Automation or De-automation. AIP Conference Proceedings.

Granell, V., Frohm, J., Bruch, J., & Dencker, K. (2007). Validation of the DYNAMO methodology for measuring and assessing Levels of Automation. Proceedings of the 1st Swedish Production Symposium (SPS). Gothenburg.

Groover, M. P. (2001). Automation, Production Systems and Computer-Integrated Manufacturing. New Jersey: Prentice Hall.

Grote, G., Ryser, C., Wäfler, T., Windischer, A., & Weik, S. (2000). KOMPASS: a method for complementary function allocation in automated work systems. International Journal Human-Computer Studies, 267-287.

Page 92: Automation in the Wood Processing Industry - DiVA portal

References

82

Gustafsson, Å., & Eliasson, L. (den 11 September 2014). Product quality deficiencies in the prefabrication industry for single-family houses. Wood Material Science and Engineering.

Harlin, U., Frohm, J., Berglund, M., & Stahre, J. (2006). Towards efficient automation implementation. Proceedings of the 9th symposium IFAC on "Automated Systems Based on Human Skills and Knowledge". Nancy, France.

Hoff, K., Fisher, N., Miller, S., & Webb, A. (1997). Sources of competitiveness for secondary wood products firms: A review of literature and research issues. Forest Products Journals, 47(2), 31-37.

Hollnagel, E. (1998). Cognitive Reliability and Error Analysis Method. Oxford: Elsevier Science Ltd.

Hopper, T., Jazayeri, M., & Westrup, C. (2008). World class manufacturing and accountability - How companies and the state aspire to competitiveness. Journal of Accounting & Organizational Change, ss. 97 - 135.

Jämsa-Jounela, S.-L. (2007). Future trends in process automation. Annual Reviews in Control, 211-220.

Jesson, J. K., Matheson, L., & Lacey, F. M. (2011). Doing Your Literature Review - Traditional and Systematic Techniques. London: SAGE Publications Ltd.

Johansson, A. (1994). Decision support tools in manufacturing systems: a task evaluation method. European conference on integration in manufacturing. Amsterdam.

Jönsson, M., Andersson, C., & Ståhl, J.-E. (2008). Implementation of en Economic Model to Stimulate Manufacturing Costs. The 41st CIRP conference on manufacturing. Tokyo, Japan.

Karltun, J. (2007). On stage: Acting for development of businesses and ergonomics in woodworking SMEs.

Kent, P., Bakker, A., Hoyles, C., & Noss, R. (2011). Measurement in the workplace: the case of process improvement in manufacturing industry. ZDM Mathematics Education, 747-758.

Koebel, B. M., Levet, A.-L., Nguyen-Van, P., Purohoo, I., & Guinard, L. (2016). Productivity, resource endowment and trade performance of the wood sector. Journal of Forest Economics , ss. 24-35.

Koho, M. (2010). Production System Assessment and Improvement: A Tool for Make-to-Order and Assemble-to-Order Companies. Tampere University of Technology. Tampere: Tampere University of Technology.

Kozak, R. A., & Maness, T. C. (2001). Quality assurance for value-added wood producers in British Columbia.

Kozak, R. A., & Maness, T. C. (2003). A system for continous process improvement in wood products manufacturing. Holz als Roh- und Werkstoff, 95-102.

Lander, E., & Liker, J. (2007). The Toyota Production System and art: making highly customized and creative products the Toyota way. International Journal of Production Research, 3681-3698.

Lee, J. D. (2008). Review of a Pivotal Human Factors Article: "Humans and Automation: Use, Misuse, Disuse, Abuse". Golden Anniversary Special Issue, 404-410.

Leschinsky, R. M., & Michael, J. H. (2004 ). Motivators and desired company values of wood products industry employees: Investigating generational difference. Forest Products Journal, 34-54.

Lindström, V., & Winroth, M. (2010). Aligning manufacturing strategy and levels of automation: A case study. Journal of Engineering and Technology Management, 148-159.

Linstone, H. A., & Turoff, M. (2002). The Delphi Method: Techniques and Applications. Newark, NJ: New Jersey Institute of Technology.

Page 93: Automation in the Wood Processing Industry - DiVA portal

References

83

Ohno, T. (1988). Toyota Production System. New York: Productivity Press.

Olsen, W. (2004). Developments in Sociology. Ormskirk: Causeway Press.

Osvalder, A.-L., & Ulfvengren, P. (2009). Human-technology systems. i M. Bohgard, S. Karlsson, E. Loven, L.-Å. Mikaelsson, L. Mårtensson, A.-L. Osvalder, . . . P. Ulfvengren, Work and technology on human terms (ss. 339-461). Stockholm: Prevent.

Parasuraman, R. (2000). Designing automation for human use: empirical studies and quantitative models. Ergonomics, 931-951.

Pirraglia, A., Saloni, D., & van Dyk, H. (2009). Status of lean manufacturing implementation on secondary wood industries including residential, millwork and panel markets. BioResources, 4(4), 1341-1358.

Rasmussen, J. (1983). Skills, rules and knowledge: Signals, signs and symbols, and other distinctions in human performance models. IEEE Transactions on Systems, Man and Cybernetics (SMC), 13(3), 257-266.

Rasmussen, J., & Vicente, K. J. (1989). Coping with human errors through system design: implications for ecological interface design. International Journal of Man-Machine Studies, 31, 517-534.

Säfsten, K., Winroth, M., & Stahre, J. (2007). The content and process of automation strategies. International Journal of Production Economics, 25-38.

Sandberg, D., Vasiri, M., Trischler, J., & Öhman, M. (2014). The role of the wood mechanical industry in the Swedish forest industry cluster. Scandinavian Journal of Forest Research, 29(4), 352-259.

Satchell, P. (1998). Innovation and Automation . Vermont: Ashgate Publising Limited.

Saunders, M., Lewis, P., & Thornhill, A. (2012). Research Methods for Business Students (6th uppl.). Harlow: Pearson.

Scallen, S. F., Hancock, P. A., & Duley, J. A. (1995). Pilot performance and preference for short cycles of automation in adaptive function allocation. Applied Ergonomics, 397 - 403.

Schmitt, R., Stiller, S., & Falk, B. (2013). Introduction of a Quality Oriented Production Theory for Product Realization Processes. 5th International Conference on Changeable, Agile, Reconfigurable and Virtual Production (CARV 2013) (ss. 309 - 314). Munich, Germany: Springer International Publishing Switzerland 2014.

Schuh, G., Potente, T., Varandani, R., Hausberg, C., & Fränken, B. (2014). Collaboration moves Productvity to the next level. Variety Management in Manufacturing Engineering, 3-8.

Schuler, A., & Buehlmann, U. (2003). Identifying Future Competitive Business Strategies for the U.S. Residential Wood Furniture Industry: Benchmarking and Paradigm Shifts. General Technical Report, United States Department of Agriculture, Forest Service, Newton Square, PA.

Sheridan, T. B. (1992). Telerobotics, automation, and human supervisory control. Cambridge, MA: The MIT Press.

Sheridan, T. B. (1995). Human centered automation: oxymoron or common sense? IEEE International Conference on Systems, Man & Cybernetics (SMC): Intelligent Systems for the 21st Century. 1, ss. 823-828. Vancouver, BC: IEEE.

Sheridan, T. B. (2002). Humans and Automation: System Design and Research Issues. Santa Monica, USA: John Wiley & Sons, Inc.

Sheridan, T. B., & Parasuraman, R. (2000). Human versus Automation in Responding to Failures: An expected-value analysis. Human Factors, 403 - 407.

Slack, N. (2005). The flexibility of manufacturing systems. International Journal of Operations & Production Management, 1190 - 1200.

Page 94: Automation in the Wood Processing Industry - DiVA portal

References

84

Sowlati, T., & Vahid, S. (2007). Malmquist productivity index of the manufacturing sector in Canada from 1994 to 2002, with a focus on the wood manufacturing sector. Scandinavian Journal of Forest Research, 21(5), 424-433.

Stahre, J. (1995). Evaluating human/machine interaction problems in advanced manufacturing. Computer Integrated Manufacturing System, 143-150.

Teischinger, A. (2010). The development of wood technology and technolofy developments on the wood industries - from history to future. European Journal of Wood Processing, 281-287.

Tuntiseranee, P., & Chongsuvivatwong, V. (1998). A survey into process and worker's characteristics in the wood furniture industry in Songkhla province, southern region of Thailand.

Välimäki, H., Niskanen, A., Tervonen, K., & Laurila, I. (2011). Indicators of innovativeness and enterprise competitiveness in the wood products industry in Finland. Scandinavian Journal of Forest Research, 19(5), 90-96.

Wäfler, T., Johansson, A., Grote, G., & Stahre, J. (1997). Complementary interaction of humans and machines is highly automated production systems - KOMPASS and TEAM. 13th Triennial Congress of the international Ergonomic Association (ss. 227-279). Finnish institute of occupational health.

Wesch-Potente, C., Weber, A. R., Prote, J.-P., Schuh, G., & Potente, T. (2014). Collaboration Mechanisms to increase Productivty in the context of Industrie 4.0. Robust Manufacturing Conference, 51-56.

Wiedenbeck, J., & Parsons, J. (2010). Digital Technology Use by Companiesin the Furniture, Cabinet, Architectural Millwork, and Related Industries. Forest Products Journal, 60(1), 78-85.

Williamson, K. (2002). Research methods for students, academics and professionals - Information management and systems (2nd uppl.). Wagga Wagga: Centre for Information Studies.

Windmark, C., Gabrielson, P., Andersson, C., & Støhl, J. E. (2012). A Cost Model for Determining an Optimal Automation Level in Discrete Batch Manufacturing. 45th CIRP Conference on Manufacturing Systems 2012. 3, ss. 73-78. Procedia CIRP.

Winroth, M., Säfsten, K., Stahre, J., Granell, V., & Frohm, J. (2007). Strategic automation - Refinement of classical manufacturing strategy. Jönköping: 1st Swedish Production Symposium.

Yin, R. K. (2003). Case Study Research: Design and Methods (3rd uppl.). Thousand Oaks: Sage Publications.

Ylipää, T. (2000). High-Reliability Manufacturing Systems. Göteborg, Sweden: Department of Human Factors Engineering, Chalmers University of Technology.

Youtie, J., Shapira, P., Urmanbetova, A., & Wang, J. (2004). Advanced Technology and the Future of U.S. Manufacturing. Proceedings of the Georgia Tech research and policy workshop. Georgia: School of Public Policy and the Georgia Tech Economic Development Institute.

Page 95: Automation in the Wood Processing Industry - DiVA portal

Appendices

85

Appendices

Appendix 1: Characteristics of the wood processing industry affecting the automation potential

What are the distinct characteristics of the Wood Processing Industry?

Quality Criteria Workplace Description

Job profile Production Product

categories Supply Chain

Characteristics Processes

Competitive factors

Rela

ted

to ra

w m

ate

rial (K

oza

k, R

.A. &

Ma

ne

ss

, T. C

. 20

03)

Kn

ots

an

d n

atu

ral d

efe

cts

D

ime

nsio

ns a

nd s

ha

pe

erro

r (Elis

as

so

n, 2

014

)

Sp

lits a

nd

cra

cks

mo

istu

re c

on

ten

t (8 - 2

2%

) (Elis

as

so

n, 2

01

4)

Stre

aks a

nd

dis

co

lora

tion

Bio

log

ical a

ttacks (fu

ng

al, b

acte

rial, in

se

ct d

am

ag

e) (E

lisa

sso

n, 2

01

4)

Exp

osu

re to

che

mic

als

(Tu

ntis

era

ne

e, P

. & C

ho

ng

su

viv

atw

on

g, V

.

19

98

)

Lo

w e

du

ca

ted

sta

ff, mis

sin

g re

leva

nt c

ertific

ate

s (T

un

tise

ran

ee, P

. &

Ch

on

gs

uv

iva

two

ng

, V. 1

99

8), (S

ow

lati, T

. & V

ah

id, S

. 200

6) &

(Karltu

n, 2

00

7)

Ve

ry lo

w in

cre

ae

in e

fficie

ncy a

nd

me

diu

m in

cre

ase

ba

se

d o

n fo

ntie

r

sh

ift (tech

nic

al c

ha

ng

e) in

com

pa

rison

with

oth

er in

du

strie

s (S

ow

lati, T

.

& V

ah

id, S

. 200

6)

Jo

ine

ry a

nd F

urn

iture

(i.e. o

ffice)

Stra

teg

ic a

llican

ce

s w

ith s

up

plie

rs o

f ad

he

siv

es, p

acka

gin

g, s

tee

l,

pla

stic

s, fa

bric

s, lu

mb

er, c

hip

s, h

ard

ware

… (S

ch

ule

r, A. &

Bu

eh

lman

n,

U., 2

00

2)

Sa

wm

ill

larg

er c

om

pa

nie

s: s

ho

rt lea

d tim

e, a

imin

g to

de

cre

ase

co

st b

y p

urs

uin

g

offs

ho

ring s

trate

gie

s (B

um

ga

rdn

er, M

. S. e

t .al., 2

00

4) &

(Bu

mg

ard

ne

r,

M. S

. et .a

l., 20

11

)

Rela

ted

to p

eo

ple

(Ko

za

k, R

.A. &

Man

es

s, T

. C.

20

03

)

Hirin

g u

na

ppro

pria

ted

em

plo

ye

es

Po

or m

ora

l

La

ck o

f train

ing

Ove

rbe

arin

g s

upe

rvis

ors

Hard

ph

ysic

al w

ork

ing

co

nd

ition

s, w

ide

ly p

esen

ce

of

ha

nd

icra

ft wo

rk ta

sks (K

arltu

n, 2

00

7)

Rap

id tu

rno

ve

r of la

bo

ur fo

rce (T

un

tise

ran

ee

, P. &

Ch

on

gs

uv

iva

two

ng

, V. 1

99

8) &

(Ka

rltun

, 200

7)

de

clin

ing

inve

stm

en

ts in

to m

ach

ine

ry a

nd

eq

uip

me

nt in

Can

ad

ian

wo

od m

anu

factu

ring

ind

ustry

(So

wla

ti, T. &

Va

hid

, S. 2

00

6)

Fe

nce

s

Inte

rna

tion

al - n

on

-glo

ba

l - so

urc

ing

(Sc

hu

ler, A

. &

Bu

eh

lma

nn

, U., 2

00

2)

Cuttin

g

sm

alle

r co

mp

an

ies: g

rea

ter u

se o

f ou

tso

urc

ed la

bo

ur/

se

rvic

es; p

ene

tratin

g n

iche

ma

rke

ts, a

min

g to

ge

ne

rate

ad

ditio

na

l reve

nu

es b

y o

fferin

g fle

xib

le o

rder q

ua

ntitie

s

(Bu

mg

ard

ne

r, M. S

. et .a

l., 200

4) &

(Bu

mg

ard

ne

r, M.

S. e

t .al., 2

01

1)

Rela

ted

to p

roc

ess

es

(Ko

za

k, R

.A. &

Ma

nes

s, T

. C. 2

00

3)

Imp

rope

r dry

ing

(mo

istu

re c

on

ten

t)

Po

or s

izin

g a

nd

ma

ch

inin

g m

ark

s

Inco

rrect d

rilling

Fa

cto

res

ran

ked

hig

hes

t am

on

g re

lev

an

t mo

tivato

rs:

Havin

g s

tea

dy e

mp

loym

en

t

Go

od

pa

y

Pe

nsio

n a

nd

oth

er s

ecu

rity b

en

efits

(Le

sch

ins

ky

, R. M

. & M

ich

ae

l, J. H

., 20

04

)

Wo

od

pro

du

ct m

an

ufa

ctu

rers

in T

aiw

an

imp

ort lo

gs (o

ak e

tc.) fro

m th

e

U.S

., pe

rform

va

lue-a

dd

ing a

ctiv

ities b

efo

re s

hip

pin

g th

e p

rodu

cts

ba

ck to

the

U.S

. du

e to

lab

or c

ost re

aso

ns (H

off, K

., Fis

he

r, N.,

Mille

r, S. &

We

bb

, A., 1

997

)

Pa

letts

Vo

lum

ne

leve

rag

ing

(Sc

hu

ler, A

. & B

ue

hlm

an

n, U

., 20

02

)

Imp

regn

atio

n (to

ma

ke

ma

teria

l resis

ten

t ag

ain

st d

estru

ctiv

e

org

an

ism

s) (E

lisas

so

n, 2

01

4)

late

pro

du

ct d

iffere

ntia

tion

, i.e. R

ead

y-to

-Asse

mb

ly fu

rnitu

re (S

ch

ule

r,

A. &

Bu

eh

lma

nn

, U., 2

00

2)

Page 96: Automation in the Wood Processing Industry - DiVA portal

Appendices

86

Quality Criteria Workplace Description

Job profile Production Product

categories Supply Chain

Characteristics Processes

Competitive factors

Rela

ted

to p

urc

ha

sin

g (K

oza

k, R

.A. &

Ma

ne

ss

, T. C

. 20

03)

To

o h

igh

va

rian

ce

s (b

ad

ch

arg

e) o

f p

urc

ha

se

d ra

w m

ate

rial

Fa

cto

res

ran

ked

hig

hes

t am

on

g re

lev

an

t

co

mp

an

y v

alu

es

:

Se

cu

re e

mp

loym

en

t B

ein

g fa

ir

Be

ing

ca

refu

l

Resp

ect fo

r the

ind

ivid

ua

l's rig

hts

(Le

sch

ins

ky

, R. M

. & M

ich

ae

l, J. H

., 20

04

)

Req

uire

men

t sp

ecific

atio

ns fo

r inco

min

g w

oo

d

ne

ed

ed

for fu

rthe

r pro

ce

ss a

uto

ma

tion

, i.e.

with

reg

ard

to d

irt or le

ng

th d

istrib

utio

ns

(Elis

as

so

n, 2

01

4)

Kitc

he

n a

nd

Ba

th C

ab

ine

try

Fe

w, s

ucce

sfu

ll clu

ste

rs in

woo

de

n fu

rnitu

re

ind

ustry

(i.e. N

orth

Italy

, De

nm

ark

) (Sc

hu

ler,

A. &

Bu

eh

lma

nn

, U., 2

00

2)

Pla

nin

g

ma

ss c

usto

miz

atio

n (B

um

ga

rdn

er, M

. S. e

t

.al., 2

00

4) &

(Sch

ule

r, A. &

Bu

eh

lma

nn

, U.,

20

02

)

Rela

ted

to p

rod

uc

ts (K

ozak

, R.A

. & M

an

ess,

T. C

. 20

03

)

Imp

rope

r hom

e a

ssem

bly

Tra

inin

g a

nd

ed

uca

tion

ab

ou

t lea

n p

rincip

les

an

d L

ea

n m

an

ufa

ctu

ring

to re

ga

in

co

mp

etitiv

en

ess a

nd

ga

in a

dvan

tag

e

(Pirra

glia

, A., S

alo

ni, D

. & v

an

Dy

k, H

., 20

09

)

Au

tom

atio

n o

f pro

ce

sse

s re

qu

ires m

ore

strin

ge

nt g

rad

ing

of in

com

ing

raw

ma

teria

l

(i.e.s

aw

n tim

be

r reg

ard

ing

timb

er h

ou

se

s) ->

tigh

ter m

ea

su

rme

nt to

lera

nce

s (E

lisas

so

n,

20

14

)

Doo

rs a

nd

Win

do

ws

Cra

dle

-to-C

rad

le c

ha

in d

eve

lop

men

t

(Sa

nd

be

rg e

t.al., 2

01

4)

Sp

rayin

g

incre

ase

d p

rod

uctiv

ity th

rou

gh

auto

ma

tion

,

lea

din

g to

de

cre

asin

g c

osts

(AW

ISA

, 201

4) &

(Bu

mg

ard

ne

r, M. S

. et .a

l., 200

4)

asse

ssm

en

t of tim

ber w

oo

d q

ua

lity

reg

ard

ing

ae

ste

thic

s, m

ech

an

ical

qu

alitie

s (s

tren

gh

ts, s

tiffne

ss, s

ha

pe

sta

bility

) or d

ura

bility

(Elis

as

so

n, 2

01

4)

Incre

ase

d tra

inin

g re

ga

rdin

g n

ew

tech

no

log

ies a

nd

me

an

s o

f au

tom

atio

n

(Wie

den

be

ck

, J. &

Pa

rso

ns

, J., 2

01

0)

Cuttin

g fo

rce

s a

re lo

w, c

uttin

g p

roce

sse

s

ve

locitie

s a

re h

igh

, co

mp

are

d to

me

tal

ind

ustry

(Karltu

n, 2

00

7)

Sh

ake

s a

nd

Sh

ing

les

Ma

jority

of s

ma

ll -siz

ed

(> 5

0 e

mp

loyee

s)

of s

eco

nd

ary

wo

od

pro

ce

ssin

g

co

mp

an

ies (K

arltu

n, 2

007

)

Asse

mb

ly w

ith a

mo

ng

st o

the

rs p

lastic

, m

eta

l co

mp

on

en

ts

Exte

nsiv

e tra

inin

g to

incre

ase

pro

du

ct

co

mp

ete

nce

(Bu

mg

ard

ne

r, M. S

. et .a

l.,

20

04

)

Esta

blis

hm

en

t of W

QS

(Woo

d Q

ua

lity S

yste

m) to

ha

nd

le q

ua

lity is

sue

s th

rou

gh

co

ntin

ou

s

imp

rovem

en

t (Ko

za

k, R

.A. &

Ma

ne

ss

, T. C

. 20

03

)

Lo

wer d

egre

e o

f au

tom

atio

n th

an

com

pa

rab

le

ind

ustrie

s d

ue

to a

nis

otro

pic

an

d v

aria

nt c

ha

racte

r

of ra

w m

ate

rial w

hic

h m

ake

s s

ortin

g a

nd

gra

din

g

pro

ce

ss h

ard

to a

uto

ma

te (K

arltu

n, 2

00

7)

(Elis

as

so

n, 2

01

4)

Flo

orin

g (i.e

. pa

rque

t)

mo

du

laris

atio

n o

f pro

du

ct a

rch

itectu

re (S

ch

ule

r, A.

& B

ue

hlm

an

n, U

., 20

02

)

Big

va

riety

of re

leva

nt m

ea

sure

s in

cu

sto

me

r-

to-o

rde

r pro

du

ctio

n d

ete

riora

tes a

uto

ma

tion

po

ten

tial (K

arltu

n, 2

00

7)

Mo

uld

ing

s, F

itting

s

Th

ird p

arty

qu

ality

assu

ran

ce

(Ko

za

k, R

. A. &

Ma

ne

ss

, T. C

., 20

01)

Page 97: Automation in the Wood Processing Industry - DiVA portal

Appendices

87

Quality Criteria Workplace Description

Job profile Production Product

categories Supply Chain

Characteristics Processes

Competitive factors

A m

ajo

rity n

am

es 'u

nd

erd

eve

lop

ed

pro

du

ctio

n

tech

no

log

y' a

s a

resu

lt of a

su

rve

y a

mon

g

Sw

ed

ish w

oo

d p

roce

ssin

g b

lue-c

olla

r work

ers

(Karltu

n, 2

00

7)

Mis

cella

ne

ou

s (to

ys, la

dd

ers

)

Le

an

man

ufa

ctu

ring

he

lps in

du

strie

s a

ch

ieve

op

era

tiona

l an

d m

an

ufa

ctu

ring e

xce

llen

ce

by

incre

asin

g p

rod

uctiv

ity a

nd

qua

lity fa

cto

rs, w

hile

red

ucin

g w

aste

an

d c

ost. (P

irrag

lia, A

.,

Sa

lon

i, D. &

va

n D

yk, H

., 20

09

)

Big

diffe

ren

ce

s c

on

ce

rnin

g v

alu

e a

dd

ed

on

resp

ectiv

e p

rod

uct c

ate

gorie

s: fro

m 1

.5 (tim

ber

ho

use

s) to

20

-30

(join

ery

/ furn

iture

) (San

db

erg

et. a

l., 20

14

)

Tim

be

r Ho

use

s

Incre

asin

g a

ttractiv

en

ess a

nd

co

mpe

titive

facto

rs b

y in

trod

ucin

g m

ore

recyclin

g a

nd

reu

sin

g o

f ma

teria

l. Th

is to

pre

se

rve

the

forre

sts

. (Sa

nd

be

rg e

t. al., 2

014

)

Lo

w v

alu

e a

dd

ed

to p

rodu

cts

in c

on

sid

era

tion

of h

igh

wa

ge le

ve

ls in

Sw

ed

en

-> fu

rthe

r

pro

du

ctiv

tiy in

cre

ase

ne

ed

ed to

sta

y p

rofita

ble

(Dic

k S

an

db

erg

et. a

l., 20

14)

Offs

ho

ring

pro

du

ctio

n in

ord

er to

com

pe

titive

facto

rs, s

uch

as d

ecre

asin

g c

ost w

hile

focu

sin

g

mo

re o

n c

usto

me

r sa

tisfa

ctio

n a

nd

be

ing a

ble

to p

rovid

e w

ha

t is d

em

an

ded

, in th

e s

en

se

of

cu

sto

miz

atio

n (B

ueh

lma

nn

, U. e

t al, 2

01

3)

Ind

ustria

l rob

ots

as 'th

ird a

rm' a

wa

y to

intro

du

ce

mo

re e

fficie

nt p

rod

uctio

n te

chn

iqu

es,

ne

xt to

sca

nn

ing

techn

olo

gie

s lik

e x

-ray a

nd

ultra

so

un

d fo

r imp

ovin

g th

e q

ua

lity le

ve

l (i.e.

reg

ard

ing

mo

istu

re c

on

ten

t) (Elis

ass

on

, 20

14)

Usin

g fu

ture

tech

no

log

y to

imp

rove

pro

du

ct

qu

ality

, i.e. M

ate

rial d

ua

rbility

, ae

sth

etic

s a

nd

en

gin

ee

red

wo

od

co

mp

osite

ma

teria

ls

(Te

isch

ing

er, A

., 201

0)

Com

pe

titive

ne

ss in

flue

nced

by fa

sh

ion

an

d

cu

lture

tren

ds a

s w

ell a

s p

olitic

al c

ircum

sta

ncs

(i.e. lo

an

s fo

r hou

se

s a

nd

co

mpo

ne

nts

)

(Karltu

n, 2

00

7)

Page 98: Automation in the Wood Processing Industry - DiVA portal

Appendices

88

Appendix 2: LoA-, LoC- and LoI taxonomy for line A

Process Task

# Task

LoA Mech.

LoA Cog.

Operator / Machine

performed Behavior

Control roles

LoI (Information

sufficient for…)

Feeding

1 Feed material 1 1 O Rule Monitor Novice

2 Quality assurance 2 1 O Skill Intervene Expert

3 Cutting 4 1 O/M Skill Monitor Expert

4 Manual adjustment 1 1 O Rule Monitor Novice

5 Vertical pallet movement 4 1 O/M Rule Monitor Novice

6 Horizontal pallet movement 4 1 O/M Rule Monitor Novice

7 Liner de-application 2 1 O Rule Monitor Novice

8 Replace scrap container 3 1 O Rule Monitor Novice

Transport 9 Conveyor 1 5 4 M

Planing

10 Cleaning/ Maintenance 3 3 O Rule Monitor Expert

11 Planing 5 5 M

12 Machine set-up 3 3 O Skill Teach Expert

13 Pace set-up 1 3 O Skill Teach Expert

14 Monitor process 1 1 O Skill Monitor Expert

15 Disturbance handling 3 1 O Knowledg

e Intervene Expert

Transport 16 Conveyor 2 5 4 M

Flipping

17 Monitor process 1 1 O Skill Monitor Novice

18 Flipping 5 4 M

19 Quality assurance 1 1 O Skill Intervene Expert

Transport 20 Conveyor 3 5 4 M

Puttying

21 Monitor process 1 1 O Skill Monitor Expert

22 Puttying 5 4 M

23 Machine set-up 2 1 O Skill Teach Expert

24 Pace set-up 1 3 O Skill Teach Expert

25 Cleaning/ Maintenance 2 3 O Skill Monitor Expert

26 Refill/Mixing 3 3 O Rule Monitor Novice

27 Disturbance handling 2 1 O Knowledg

e Intervene Expert

Heat treatment

28 Heat treatment 5 5 M

29 Monitor process 1 1 O Skill Monitor Expert

30 Pace set-up 1 3 O Skill Monitor Expert

31 Disturbance handling 1 1 O Knowledg

e Intervene Insufficient

Transport 32 Conveyor 4 5 1 M

Sanding

33 Monitor process 1 1 O Skill Monitor Expert

34 Maintenance 1 1 O Skill Intervene Expert

35 Disturbance handling 1 1 O Knowledg

e Monitor Expert

36 Sanding 5 4 M

Transport 37 Conveyor 5 5 4 M

Page 99: Automation in the Wood Processing Industry - DiVA portal

Appendices

89

Primer

38 Cleaning/ Maintenance 2 3 O Skill Intervene Expert

39 Refill/Mixing 3 3 O Rule Monitor Novice

40 Priming 5 4 M

41 Weighting additives 2 3 O Rule Monitor Novice

42 Disturbance handling 2 1 O Knowledg

e Intervene Expert

43 Monitor process 1 1 O Skill Monitor Expert

Dryer 1

44 Drying 5 4 M

45 Disturbance handling 1 1 O Knowledg

e Intervene Insufficient

46 Monitor process 1 1 O Skill Monitor Expert

Transport 47 Conveyor 6 5 4 M

Scanner

48 Error detection 6 6 M

49 Monitor process 1 1 O Skill Monitor Expert

Polishing

50 Polishing 5 4 M

51 Monitor process 1 1 O Skill Monitor Expert

Transport

52 Conveyor 7 to rework 5 4 M

53 Conveyor 8 from rework 5 4 M

Rework

54 Quality assurance 1 1 O Skill Intervene Expert

55 Cutting 4 1 O/M Skill Monitor Expert

56 Puttying 2 1 O Skill Monitor Expert

57 Move from rework 1 1 O Rule Monitor Novice

58 Move to rework 1 1 O Rule Monitor Novice

59 Replace scrap container 3 1 O Rule Monitor Novice

Painting

60 Cleaning/ Maintenance 2 3 O Skill Monitor Expert

61 Refill/Mixing 3 3 O Rule Monitor Novice

62 Disturbance handling 2 1 O Knowledg

e Intervene Expert

63 Weighting additives 2 3 O Rule Monitor Novice

64 Monitor process 2 1 O Skill Monitor Expert

65 Painting 5 4 M

Dryer 2

66 Disturbance handling 1 1 O Knowledg

e Intervene Insufficient

67 Monitor process 1 1 O Skill Monitor Expert

68 Drying 5 4 M

Transport 69 Conveyor 9 5 4 M

Individual labeling

70 Manual adjustment 1 1 O Skill Monitor Expert

71 Quality assurance 1 1 O Skill Intervene Expert

72 Length measuring 5 4 M

73 Disturbance handling 1 1 O Knowledg

e Intervene Expert

74 Monitor process 1 1 O Skill Monitor Expert

75 Moulding Labeling 5 5 M

Transport 76 Conveyor 10 5 4 M

Stacking

77 Stacking 5 4 M

78 Flipping the 10th moulding 5 4 M

Page 100: Automation in the Wood Processing Industry - DiVA portal

Appendices

90

79 Monitor process 1 1 O Skill Monitor Expert

80 Edge offset 5 4 M

81 Disturbance handling 1 1 O Knowledg

e Intervene Expert

Transport 82 Conveyor 11 5 4 M

Plastic wrapping

83 Wrapping 5 5 M

84 Monitor process 1 1 O Skill Monitor Expert

85 Disturbance handling 1 1 O Knowledg

e Intervene Insufficient

86 Edge offset 5 4 M

Transport 87 Conveyor 12 5 4 M

Stack labeling

88 Push stack 5 4 M

89 Stack Labeling 5 5 M

90 Move 17 stacks to pallet 5 4 M

91 Monitor process 1 1 O Skill Monitor Expert

92 Liner application 1 1 O Rule Monitor Novice

93 Disturbance handling 1 1 O Knowledg

e Intervene Expert

Page 101: Automation in the Wood Processing Industry - DiVA portal

Appendices

91

Appendix 3: LoA-, LoC- and LoI taxonomy for line B

Process Task

# Task

LoA Mech.

LoA Cog.

Operator / Machine

performed Behavior

Control roles

LoI (Information

sufficient for…)

Planing

Feeding

1 Feed material 4 1 O Rule Monitor Novice

2 Quality Assurance 1 1 O Skill Monitor Expert

3 End-cutting 4 1 O/M Skill Monitor Expert

4 Monitoring process 1 1 O Skill Monitor Expert

5 Disturbance handling 3 1 O Knowledge Intervene Expert

6 Manual adjustment 1 1 O Rule Monitor Novice

7 Vertical pallet movement 4 1 O/M Rule Monitor Novice

8 Horizontal pallet movement 4 1 O/M Rule Monitor Novice

9 Liner de-application 4 4 M

10 Replace scrap container 3 1 O Rule Monitor Novice

Transport 11 Conveyor 1 5 4 M

Planing

12 Cleaning 3 1 O Rule Monitor Novice

13 Machine set-up 3 3 O Skill Teach Expert

14 Disturbance handling 3 1 O Knowledge Intervene Insufficient

15 Monitoring process 1 1 O Skill Monitor Expert

16 Quality Assurance 2 1 O Skill Intervene Expert

17 Pace set-up 1 1 O Skill Teach Expert

18 Planing 5 5 M

Transport 19 Conveyor 2 5 4 M

Stacking

20 Fall angle adjustment 4 1 O Rule Monitor Novice

21 Monitor process 1 1 O Skill Monitor Novice

22 Position mouldings 1 1 O Skill Monitor Novice

23 Disturbance handling 1 1 O Knowledge Intervene Expert

Transport 24 Truck 1 5 2 O/M Skill Monitor Expert

Surface treatment 1

Feeding

25 Height positioning 4 1 O Rule Monitor Novice

26 Manual feed 1 1 O Rule Monitor Novice

27 Monitor process 1 1 O Skill Monitor Expert

28 Disturbance handling 1 1 O Knowledge Intervene Expert

Transport 29 Conveyor 3 5 4 M

Painting

30 Painting 5 5 M

31 Monitor process 1 1 O Skill Monitor Expert

32 Disturbance handling 1 1 O Knowledge Intervene Expert

33 Cleaning 2 3 O Skill Monitor Expert

34 Refill / Mixing 2 3 O Rule Monitor Novice

Page 102: Automation in the Wood Processing Industry - DiVA portal

Appendices

92

35 Weighting additives 2 3 O Rule Monitor Novice

36 Quality Assurance 1 1 O Skill Intervene Expert

37 Disturbance handling 1 1 O Knowledge Intervene Expert

Transport 38 Conveyor 4 5 4 M

Drying

39 Drying 5 4 M

40 Monitor process 1 1 O Skill Monitor Expert

41 Disturbance handling 1 1 O Knowledge Intervene Expert

Transport 42 Conveyor 5 5 4 M

De-applicatio

n

43 Monitor process 1 1 O Skill Monitor Expert

44 Disturbance handling 1 1 O Knowledge Intervene Expert

45 Move objects to pallet 4 1 O Rule Monitor Novice

46 Quality Assurance 1 1 O Skill Intervene Expert

Transport 47 Truck 2 5 2 O/M Skill Monitor Expert

Surface treatment 2

Feeding

48 Manual feed 1 1 O Rule Monitor Novice

49 End-cutting 4 1 O/M Rule Monitor Novice

50 Vertical pallet movement 4 1 O/M Rule Monitor Novice

51 Horizontal pallet movement 4 1 O/M Rule Monitor Novice

52 Monitor process 1 1 O Skill Monitor Expert

53 Disturbance handling 1 1 O Knowledge Intervene Expert

54 Replace scrap container 3 1 O Rule Monitor Novice

Transport 55 Conveyor 6 5 4 M

Coloring station

56 Painting 5 4 M

57 Monitor process 1 1 O Skill Monitor Expert

58 Disturbance handling 1 1 O Knowledge Intervene Expert

59 Cleaning 2 1 O Skill Monitor Expert

60 Quality Assurance 1 1 O Skill Intervene Expert

61 Refill / Mixing 2 3 O Rule Monitor Novice

62 Weighting additives 2 3 O Rule Monitor Novice

Transport 63 Conveyor 7 5 4 M

Drying

64 Drying 5 4 M

65 Monitor process 1 1 O Skill Monitor Expert

66 Disturbance handling 1 1 O Knowledge Intervene Expert

Transport 67 Conveyor 8 5 4 M

Stacking

68 Fall angle adjustment 4 1 O Rule Monitor Novice

69 Monitor process 1 1 O Skill Monitor Novice

70 Disturbance handling 1 1 O Knowledge Intervene Expert

71 Adjust mouldings on pallet 1 1 O Rule Monitor Novice

Transport 72 Truck 3 5 2 O/M Skill Monitor Expert

Page 103: Automation in the Wood Processing Industry - DiVA portal

Appendices

93

Rework

Rework

73 Add moulding to workstation 1 1 O Rule Monitor Novice

74 Quality Assurance 1 1 O Skill Intervene Expert

75 Apply putty 2 1 O Skill Monitor Expert

76 Cut ends 4 1 O/M Rule Monitor Novice

77 Remove moulding from

workstation 1 1 O Rule Monitor Novice

78 Empty waste container 3 1 O Rule Monitor Novice

Transport 79 Truck 4 5 2 O/M Skill Monitor Expert

Sorting

Feeding

80 Vertical pallet movement 4 1 O/M Rule Monitor Novice

81 Horizontal pallet movement 4 1 O/M Rule Monitor Novice

82 Monitor process 1 1 O Skill Monitor Expert

83 Disturbance handling 1 1 O Knowledge Intervene Expert

84 Quality Assurance 1 1 O Skill Intervene Expert

85 Feed material 1 1 O Rule Monitor Novice

86 Cut ends 4 1 M

87 Remove waste 5 4 M

Transport 88 Conveyor 9 5 4 M

Labeling

89 Length measuring 5 4 M

90 Machine set-up 1 3 O Skill Teach Expert

91 Individual label application 5 5 M

92 Monitor process 1 1 O Skill Monitor Expert

93 Disturbance handling 1 1 O Knowledge Intervene Expert

94 Automatic transport to pocket 6 6 M

95 Manual Adjustment 1 1 O Rule Monitor Novice

96 Cable tie application 5 4 M

97 Edge alignment 5 1 M

98 Stack label application 5 4 M

99 Move full row of stacks to

pallet 5 4 M

Page 104: Automation in the Wood Processing Industry - DiVA portal

Appendices

94

Appendix 4: Gantt chart of the thesis work