Exploring the Potential of Industry 4.0 for Production Planning and Control of Flexible Manufacturing Systems

(1)

Exploring the Potential of Industry 4.0

for Production Planning and Control of Flexible Manufacturing Systems

A Case Study

MSc. Technology and Operations Management University of Groningen,

Faculty of Economics and Business, Academic Year 2017 - 2018

CLAUDIA STEINHART S3290395

Supervisor / University of Groningen (RUG) Dr. J.A.C Bokhorst

Co-assessor / University of Groningen (RUG) Dr. ir. D.J. van der Zee

29-01-2018

(2)

(3)

Abstract

The ever fast changes of market demands intensify the need for manufacturing firms to run an efficient production with a high degree of flexibility. While flexible manufacturing systems (FMSs) have been praised as the ultimate solution to this, FMS constraints challenge production planning and control with adverse effects on production performance. New opportunities evolve with the rise of the fourth industrial revolution, known as Industry 4.0. This paper explores the potential of Industry 4.0 technologies to facilitate production planning and control activities in line with FMS constraints. As part of it, it specifies FMS constraints for production planning and control (PPC) in a high-variety-medium/low-volume production environment. A single case study was conducted in a mid-sized manufacturing firm in the northern Netherlands. It was found that a lack of shared capacity, interference, resource availability and routing leads to lower production efficiency, longer cycle times and a low due date performance. Reasons for a limited addressing of these constraints in PPC are a lack of data, transparency and decision support. The application of Industry 4.0 technologies provides potential to bridge this gap. While to date, Industry 4.0 is still at a concept level, this study illustrated that already existing technologies in the context of Industry 4.0, such as real-time monitoring and visualisation can largely facilitate an addressing of these constraints in PPC. Further research is encouraged to substantiate the potential impact with increasing maturity of Industry 4.0.

(4)

II

List of Figures

Figure 1 Conceptual Model ... 2

Figure 2 Industry 4.0 Maturity Index ... 12

Figure 3 Research Framework ... 13

Figure 4 Research Activities ... 19

Figure 5 Overview Process and Routing in the Gear Department ... 21

Figure 6 Overview FMS Constraints ... 26

Figure 7 Specification of FMS Constraints in the Research Framework ... 27

Figure 8 Relation between FMS Constraints and Production Performance ... 31

Figure 9 Resource Constraints ... 35

Figure 10 Overview Process Steps – Machine level ... 48

Figure 11 Excerpt Production Records GP_FREES ... 53

Figure 12 Excerpt Production Records RZ400 ... 53

Figure 13 Analysis of Cycle Time in the Gear Department ... 54

Figure 14 Analysis of Improvement Potential Shop-Floor-Control System ... 55

List of Tables

Table 1 Constraining Factors per Implementation Issue ... 5

Table 2 Overview of Data Sources ... 17

Table 3 Overview Characteristics and Constraints Toothing ... 37

Table 4 Overview Machine Characteristics and Constraints in the Gear Department ... 51

Table 5 Overview FMS Constraints ... 54

(7)

Abbreviations

AGV. Automated Guided Vehicle CNC. Computer Numerical Control CPS. Cyber-Physical System

ERP. Enterprise Resource Planning FMC. Flexible Manufacturing Cell FMS. Flexible Manufacturing System I4.0. Industry 4.0

MRP. Material Requirements Planning MTO. Make-to-Order

MTS. Make-to-Stock

(8)

VI

Preface

I would like to thank my supervisor, Dr. J.A.C. Bokhorst, for his significant support during the completion of this research. I am very grateful for his valuable time spent on providing me with constructive feedback and guidance. I would like to thank also my co-assessor, Dr. ir. D.J. van der Zee for his input during the research proposal.

(9)

1 Introduction

In the 1970s, technological advancements with the rise of numerically controlled machines, robotsand computers, triggered the trend for automated, unmanned production systems (Adlemo et al. 1994). This trend resulted in a transformation of the manufacturing landscape and eventually gave rise to a new manufacturing concept, namely Flexible Manufacturing Systems (FMSs). The characteristics of FMSs enable the production of multiple part types at the same time which allows a firm to react more flexible to market changes and further improve production efficiency (Slomp et al. 1993). Hence, the concept of FMSs is considered to be amongst the most important industrial applications of information technology and has been hailed as the ultimate solution for increased productivity and competitiveness (Raj et al. 2007).

In practice, implementation issues have often outperformed the numerous advantages of FMSs over a classical machine job shop production (Raj et al. 2007). Part and machine family selection, resource allocation, tool loading and machine allocation are examples of the difficulties manufacturing firms are confronted with (Balogun et al. 1999).

Evidently, these problems are first, highly related to production planning and control, and secondly, caused by several constraining factors of the systems. Physical constraints, causal constraints and availability restrictions (Fox et al. 1983) negatively impact production performance unless properly addressed.

Since traditional production planning and control approaches allow little flexibility to address these constraints, researchers have developed frameworks which are better suited for PPC in an FMS environment (e.g. Stecke 1985). Concurrently, a myriad of papers have been published, proposing solution approaches to these issues, ranging from the use of algorithms, simulations, mathematical programming to artificial intelligence. Raj et al. (2007) found that these techniques are difficult to apply in practice.

(10)

2 | P a g e The presumption is made that an increasing application of I4.0 enables better addressing of FMS constraints in PPC which in turn increases production performance. With this research, we aim to reveal how the application of I4.0 technologies enhances production planning and control activities in line with FMS constraints. As part of this, we intend to specify FMS constraints in PPC to investigate the relation between these constraints and production performance.

Figure 1 Conceptual Model

(11)

2 Theoretical Background

This section means to establish a common understanding of the key concepts of this research study which are first, flexible manufacturing systems, second, production planning and control and third, the concept of Industry 4.0. Lastly, it serves to outline the research framework.

2.1 Flexible Manufacturing Systems

For the introduction of the concept of flexible manufacturing systems the funnel approach is used, starting generic with a definition and concept description, and subsequently narrowing down to the problem situation.

2.1.1 Definition and Concept

Several definitions of FMSs are proposed in the literature (e.g. Stecke 1985, Brown et al. 1984). MacCarthy and Liu (1993) define an FMS as a computer-controlled system of automated workstations and material handling devices which enable to simultaneously manufacture different jobs. According to MacCarty and Liu (1993) an FMS consists of three elements:

1. A processing system with a set of CNC machines capable of changing tooling automatically,

2. a material handling and storage system,

3. a computer control system automatically operating the system.

Browne et al (1984) provide a classification of FMSs into four types which differ in their design. These are:

1. The flexible machining cell (FMC), consisting of a general-purpose CNC machine tool with automated material handling,

2. the flexible machining system which consists of FMCs of different types and additionally enables different routing,

3. the flexible transfer line that can be operated akin to a dedicated transfer line with fixed routing,

(12)

4 | P a g e 2.1.2 FMS Advantages

A flexible manufacturing system is supposed to allow a firm to achieve both the efficiency of automated high-volume mass production and the flexibility to machine multiple part types at the same time, similar to low-volume production (Browne et al. 1984).

The term “flexibility” indicates the ability of the systems to adjust to changes (Das 1996). Depending on the design of the FMS, an FMS provides machine flexibility (types 1-4) and routing flexibility (types 1-2). Machine flexibility refers to “the versatility of the machining centres as well as the capability of performing many different operations.” (Hansmann et al. 1997 p.2). Routing flexibility describes the “ability of a production system to manufacture a product using several alternative routes in the system” (Guo et al. 2009 p. 4269).

Aside from increased flexibility, some of the most important advantages of FMSs for a firm are a reduction of the total lead time, a reduction of the work in process, an increased ability to incorporate specific customer wishes and a more efficient production (Slomp et al. 1993). High production efficiency is derived from increased throughput and the low number of operators required to run an FMS. Frequently, FMSs can be operated unmanned for a definite period.

Yet, the realisation of these mentioned advantages has often been impeded by numerous implementation issues. For this reason, we interrogate the main issues companies are facing with the implementation of FMSs in section 2.1.3.

2.1.3 General Implementation Issues of FMSs

In the last decades, much attention has been paid to the concept and difficulties related to FMSs. Raj et al. (2007) conducted a literature review concerning planning, design and implementation of FMSs. They found that the major implementation problems of FMSs concern (1) the loading of parts, (2) the scheduling, (3) the material handling, (4) flexibility and its measurement, (5) machine tools, (6) operation, control and maintenance techniques, (7) the human element and culture. We briefly explain each of these seven classifications and refer to their article for a more in-depth description.

(13)

(2) The scheduling of production orders is challenged by the variety of products that can be produced on an FMS. Hence, production orders need to be sequenced to achieve production objectives, considering production constraints.

(3) The material handling issues are related to the transportation of parts. In FMSs typically automated-guided-vehicles (AGV) are used. The main challenge is to find an optimal schedule and route of AGVs.

(4) In literature, several views of flexibility are presented which it makes a challenge to measure the degree of flexibility of an FMS.

(5) Machine tools in FMSs are typically the CNC-machines, special purpose machines and cutting tools. One issue is that it may happen that a CNC-programme or cutting tool is not yet ready. (6) The increased complexity of the system challenges its operation and control.

(7) Socio-organisational aspects are related to the man-machine interaction, the timing of activities, such as the loading and unloading of parts or the supervision and the division of labour.

Throughout the years, a vast amount of researchers have published similar reviews, e.g. Stecke (1985) scrutinised and summarised design, planning, scheduling and control problems of flexible manufacturing systems, or Kouvelis (1992) who reviewed design and planning problems in flexible manufacturing systems.

Evidently, these implementation issues are highly related to production planning and control constraints, summarised in table 1.

Table 1 Constraining Factors per Implementation Issue

Implementation issue Constraining Factors PPC

Loading problem Availability of machine, capacity of machine tools

Scheduling of orders Capacity of machine, alternate routings, pallet and fixture limitations, finite-in-process buffers, shared tools

Material handling

(if automated with AGVs)

Number of vehicles, battery of AGV, carrying capacity, queue capacity

Machine tools Limitations of CNCs, tool lives

(14)

6 | P a g e For this reason, we devote particular attention to PPC constraints in this thesis. Fox et al. (1983) categorise five types of production planning and control constraints in an FMS environment. These are as follows:

 Organisational goals which represent a firms objectives concerning the FMS  Physical constraints encompassing machine capacity, size, and utensil constraints  Causal constraints based on, inter alia, tool and material exigencies. Causal constraints

represent the prerequisites to be fulfilled that production can start

 Availability restrictions, i.e. the availability of operators and shared utensils  Preferences, e.g. sequencing or machine preference of the operators

FMS planning and control is challenged to take these constraints into account. Failure to address causal constraints can, for instance, impede the processing of an order. As a consequence, machine downtime leads to lower throughput, longer waiting times and perhaps even to the exceeding of the respective due date. Likewise, an insufficient consideration of availability constraints result in a performance loss. For instance, in case of operator activities, e.g. setups, are required by multiple machines at the same time, one machine is inevitably down for a certain period.

(15)

2.2 Production Planning and Control

In this thesis, we regard production planning and control as a system responsible for the planning, scheduling and control of activities within a manufacturing unit. In section 2.2.1 the objectives and elements of traditional production planning and control approaches are presented. In section 2.2.2, we narrow down to production planning and control in an FMS environment.

2.2.1 Traditional Approach

Objectives of a production planning and control system as summarised by Speith (1982) are:

 a high due date performance,

 a high and balanced capacity utilisation,  short throughput times,

 low inventory costs,

 a low work-in-progress level,  a high delivery performance,

 high accuracy of information provision,

 high production flexibility to rapidly react to customer demands and disturbances,  low costs of ordering,

 high material availability and

 reliable planning and scheduling results.

For definitions of these objectives, we refer to Speith (1982 p.32-34). Depending on dynamic environmental situations, one or some production objectives might obtain higher relevance than others. Sometimes this means to pursue one objective at the expense of another one (Slomp 1993).

(16)

8 | P a g e maximum workload which represents the maximum capacity utilisation without a violation of norms (Bertrand et al. 1984).

2.2.2 Production Planning and Control of FMSs

While the traditional approach is appropriate for production planning and control activities of conventional machines, it allows little flexibility to deal with the numerous constraints that come along with flexible manufacturing systems. For instance, in FMS planning and control, a trade-off has to be made between a high and balanced capacity utilisation with lowest possible machine setups and the realisation of due dates. This is because of order postponements to reduce machine downtime caused by a high number of setups. Several researchers have suggested production control frameworks with multiple decision levels to facilitate addressing FMS constraints. Stecke (1985) proposed the establishment of three decision levels, namely the assignment level, the offline level and the online level. At the assignment level, the workload of FMSs is determined. At this level, orders are distributed, most often with the use of an MRP system. Since MRP systems are not designed to deal with FMS constraints, a matching of the orders with the FMS constraints is done at the offline level. The offline level builds the basis for the online level at which decisions are made based on real-time information (Slomp 1993, Stecke 1985). From an organisational perspective, the three levels correspond to the planning department, the supervisor of FMSs and the operators of FMSs, respectively (Slomp 1993).

(17)

2.3 Industry 4.0

The third major concept in this thesis is the concept of Industry 4.0. In section 2.3.1 we firstly provide an overview of the concept, followed by an elaboration on production planning and control within the context of I4.0 in section 2.3.2. Lastly, we present the maturity stages of I4.0.

2.3.1 Definition and Concept

With the rise of new digital industrial technology, industry is now on the threshold of the next big shift, known as Industry 4.0, triggered by the demand for short development periods, individualisation of products, flexibility and resource efficiency on the one hand, and technological advancements on the other hand (Ruessmann et al. 2015, Brettel et al. 2014). The “4.0” insinuates the potential impact of this trend to tie up to the three previous industrial revolutions (Schuh et al. 2017). Similar to the goals of flexible automation, Industry 4.0 with its technological advancements strives to enable a flexible, efficient, and sustainable production with highest quality at low costs (Wang et al. 2016).

(18)

10 | P a g e 2.3.2 Production Planning and Control in the Context of Industry 4.0

According to Almada-Lobo (2016), the application of Industry 4.0 technologies is likely to trigger major changes in production planning and control towards a fully decentralised system with agents/objects. He describes the shop-floor to most likely become a marketplace for capacity (supply) while CPSs represent the production demand. This multi-agent-like system enables self-organisation in which “competing targets and contradicting constraints will generate a holistically optimised system, ensuring only efficient operations will be conducted.” (Almada-Lobo 2016 p.17). Similarly, Hermann et al. (2016) envision that with the introduction of CPSs, every part of a factory will be represented as an object able to communicate with each other in the network. Through virtualisation a virtual model is created involving the condition of all CPSs, which allows the immediate detection of failures and provides support for decision-making processes. Decentralised control systems use RFID tags to determine the subsequent working step and lastly, real-time capability allows the immediate detection and adaption to disruptions. Thus, central planning and controlling is redundant (Hermann et al. 2016).

(19)

2.3.3 Maturity Index of Industry 4.0

Envisioning an incremental development of Industry 4.0, several researchers have proposed I4.0 maturity models (Porter et al. 2015, Lee et al., 2015). According to Schuh et al. (2017), an I4.0 production can be build up in four major stages. These are visibility, transparency, predictive capacity and adaptability, noting that as stated by Schuh et al. (2017) digitalisation of manufacturing is the prerequisite for Industry 4.0. In this model, digitalisation is achieved in two stages which are first computerisation and secondly connectivity. Connectivity encompasses the full integration of business processes, hence the integration of IT with operational technology. Some authors categorise this digitalisation process as part of Industry 4.0 (e.g. Lee et al. 2015, who present an analogous model).

In the following, we briefly elaborate on the four maturity stages of Industry 4.0 according to Schuh et al. (2017).

1. The first stage in the model is the establishment of visibility. This stage enables “seeing”, i.e. monitoring what is happening on the shop-floor. In this first stage, visibility is created through sensor technology which enables the capturing of processes. This allows real-time monitoring throughout the entire company and beyond. Thereby a “digital shadow” is created which equals “an up-to-date model of the entire company at all times that is not tied to individual data analyses” (Schuh et al. 2017 p.17).

2. The second stage is transparency which supports the “understanding” of happenings to identify their cause. Thus, process knowledge is gained which can be used to provide rapid decision-making support. Evolving technologies facilitate the analysis of large volumes of data which is executed in parallel to existing systems such as ERP systems.

3. Stage three is referred to as the predictive capacity stage and intends to predict what is happening next to “being prepared”. In this stage, simulation techniques enable the depiction of multiple future scenarios and their likeliness to materialise. Predictive capacity endeavours the anticipation of future developments to make appropriate decisions and take actions.

(20)

12 | P a g e

Figure 2 Industry 4.0 Maturity Index adapted from Schuh et al. (2017)

2.4 Summary and Research Framework

FMS constraints represent a challenge for production planning and control of manufacturing firms. Failure to address these FMS constraints in production planning and control can impede the materialisation of FMS advantages of a more efficient and flexible production, eventually affecting production performance negatively. Fox et al. (1983) classified these constraints in, inter alia, physical, causal and availability constraints1.

With Industry 4.0, new opportunities evolve, bringing incrementally visibility, transparency, predictive capacity and lastly adaptability to the shop-floor. While in the last stage of Industry 4.0 constraining factors become subject to negotiation processes between agents, there is no information in the literature of the potential to better addressing FMS constraints in lower maturity stages.

Based on these findings, the research framework was developed, presented in figure 3.

(21)

Figure 3 Research Framework

(22)

14 | P a g e

3 Methodology

The question of interest in this research is whether the application of Industry 4.0 technologies enhances production planning and control activities in line with FMS constraints. As a natural consequence, the question of “how” arises. According to Yin (1994), the qualitative approach of a case study is particularly suited to research the “how” aspect of phenomena. Meredith (1998) summarises two additional strengths of case studies: First, the phenomenon of interest can be studied in its original setting, which means that insights can be gained through observing the actual practice. Secondly, a case study allows exploratory investigations. A case study therefore, seems the most suitable method for this research topic.

The research is divided into three parts, discussed below. PART I:

Based on the explorative character of this case study, the first step in this research is to study FMS characteristics and their surrounding production environment in practice. Thereby we aim to specify the constraints to be addressed in PPC. Part I further aims to reveal the impediments to properly address these constraints in PPC.

PART II:

Part II means to reveal the potential of industry 4.0 technologies to enhance production planning and control in line with FMS constraints. Part II is grounded on the results of part I and is executed in alignment with the four stages of the maturity index presented by Schuh et al. (2017) (see section 2.3.3).

PART III:

(23)

3.1 Case Selection and Sampling

PART I:

For the first part of this research, a single case was used. The case selection was executed in cooperation with the Lectoraat Lean/World Class Performance of the University of applied sciences Arnhem and Nijmegen. The selected company participates in the project RAAK KIEM Smart Industry and is for this reason particularly suited for this research. The case company is a mid-size manufacturing company with its engineering and production of all products and parts located at a single production site in the northern Netherlands. The focus in this research is set on the gear department of this company as this department represents all relevant aspects of this study. The company has rapidly grown in the past years while constantly increasing their production portfolio as a reaction to changing market demands. As a consequence, series size decreased and product variety increased. Hence, there has been a shift from large series production with a low number of variants towards a high-variety production. From initially 50 different gear wheels, the number rose to 550 by today. Hence, there is a high variety of product types with medium to low volumes. In order to cope with the increased capacity and flexibility demand, the company invested in flexible automation. While the technology to meet the changing market demands has been established, the PPC approach was not adjusted accordingly.

PART II:

The second part of this research builds on the results derived from the first part. While this part is linked to the results of part one, it is conceptual in nature. To gather practical input of the state of the art of industry 4.0 related technologies used in production planning and control, we selected a representative shop-floor control system supplier whose products are developed for an Industry 4.0 production.

PART III:

(24)

16 | P a g e recommended in case research (Yin 2014, Eisenhardt 1989). The selection of these cases enabled the application of replication techniques in the data analysis phase (see chapter 3.3).

3.2 Data Collection

In this case study, the concept of data triangulation was employed, which is defined as “the use and combination of different methods to study the same phenomenon” (Voss et al. 2002 p.206). Using different sources of evidence allows developing “converging lines of inquiry” (Yin 2014 p.140). Three of the six main data sources were used in this case study, namely documentations, interviews and direct observations. Archival records, participant observations and physical artefacts were not utilised due to a lack of availability or as considered as not applicable to derive meaningful insights for this case study.

PART I:

In part I, the main data source was represented by interviews. Additional methods for the data gathering in this part were observations and the review of internal documentations. Documentations comprised machine capacities, bill of materials, excerpts of production plans, actual machine hours and scan data. Besides of that, we leveraged performance analyses presented in a former research study in the case company (Bizimana 2016).

To gather the data, the researcher visited the case company twice with a time frame of two weeks between both visits. The first site visit was divided up into interviews with the production planner and the production manager and a walk-through the shop-floor with the production manager. For the first visit, the approach of largely unstructured interviews with a framework of a priori prepared questions led to an unbiased understanding of products, process and the PPC approach. During the walk-through, the researcher was equipped with paper and pen to record all observations manually. For the first visit one full working day was used.

(25)

PART II:

The data collection in the second part involved an expert interview and a review of literature. The Interview was conducted with the managing director of a shop-floor control system supplier. This interview served beyond that as initial validation of the findings of research part I. Analogous to the interviews in part I, the interview was audio-recorded and transcribed. Again, an interview protocol with semi-structured questions was developed in advance, aligned with the outcomes of part I. The questions were formulated as open-questions which allowed to gain a deeper knowledge of the phenomenon. The interview was conducted via phone and had a duration of one hour.

The literature review was conducted to investigate the potential of future technologies of Industry 4.0. Google Scholar was utilised for this search. Following key terms were used: Industry 4.0 technologies, Industry 4.0 advanced data analytics and Industry 4.0 smart shop-floor.

PART III:

The data gathering of the third part was grounded on the findings of parts one and two. The approach in this part was again semi-structured interviews with two company representatives per case. For reasons of availability, the questions were answered by two students that have been investigated production planning and control in the companies for a duration of six months. Therefore, we judged their knowledge to precisely answer the questions as sufficient. The duration of each interview amounted to approximately 30 minutes and was conducted at the University of Applied Sciences in Arnhem.

An overview of the sources per research part is presented in table 2.

Table 2 Overview of Data Sources

Part Interviews Observations Internal Documentations Literature

1 Largely

unstructured and semi-structured

Notes of field observations

Machine Capacities, Bill of Materials, Excerpts of Production Plans, Actual machine hours, Scan data

-

2 Semi-structured - - Industry 4.0

technologies

(26)

18 | P a g e documents, interview transcripts and notes as collected in the field. The second collection included the researchers’ analyses and reports. To ensure maintaining a clear chain of evidence a case study protocol was utilised (Karlsson 2010).

The collection of data took place in accordance with the division into the three research parts as described above, hence was executed on a staggered basis. The respective interview protocols can be found in the Appendix.

3.3 Data Analysis

The data collected in the three parts were analysed separately. According to Yin (2014), there are five specific techniques for the case analysis. These are pattern matching, explanation building, time-series analysis, logic models, and cross-case synthesis.

PART I:

The data analysis of part one encompassed the aggregation of the data to identify major issues and difficulties in the production planning and control framework of the companies. The observations made during the walkthrough were directly processed into a drawing of production sequence and possible routings within the department. The aggregated data served as data set to specify the FMS constraints in the case company. A combination of an inductive and deductive approach was used.

PART II:

For the second part of this research, we chose an application-oriented approach grounded on the FMS constraints as identified in the case company. We analysed the data from literature and the answers from the interview for similarities and merged them with the problems discovered in part I. For the analysis of the distant future, we firstly investigated literature for the main technologies per maturity stage and subsequently investigated the potential for one constraining factor to give an idea of the potential of these technologies.

PART III:

(27)

(28)

20 | P a g e

4 Case Description

The case description functions as an introduction into the companies as-is situation. Hence, the nature of this section is descriptive. The focus is set on the key elements of this research. We first describe the production process and routing, followed by the organisation of production in section 4.2, and the production planning and control approach applied in the case company in section 4.3.

4.1 Process and Routing

The production in the gear department is divided into several sequential steps. The process steps are illustrated in figure 5. The routing of the products through the steps depends on the shape, i.e. form and diameter of the product. For most products, the first production step is the sawing of raw material. For products with high yearly volumes, forging material is used. Forging parts are obtained by the supplier on a call-off basis. The sawed parts and the forging material are processed at turning machines.

After the turning, the parts are processed into gears through toothing. This includes the milling of external toothing in a hobbing process and the punching of the internal toothing. Both processes can take place in an interchangeable sequence. In some exceptional cases, the external toothing is also done in a punching process. Subsequently, the majority of semi-finished goods are hardened externally. For some parts hardening is redundant, hence the parts are directly assembled.

The delivery to and from the heat treatment shop is organised with a milk run system. The parts are sent twice a week, Wednesdays and Fridays and returned latest on Tuesday morning and Thursday morning, respectively. Once returned, parts receive a finishing operation to achieve the required precision. Therefore, the parts are either hard-turned or cylindrical grinded. Solely one product type requires both treatments in an interchangeable sequence.

(29)

Figure 5 Overview Process and Routing in the Gear Department2 Sawing Hobbing Punching Hardening (external) Raw Material (Forging) Turning Raw Material

Hardturning Cylindrical grinding

Tooth grinding

Assembly

Interchangable sequence

Exceptional cases

(30)

22 | P a g e

4.2 Organisation of Production

Manned production takes place five days per week from Monday until Friday from seven in the morning to five in the evening with a one hour break per day. In case of a high production backlog, additional manned night or weekend shifts are arranged. While production takes place five days a week, employees are present four days a week with one day off. Operators are in charge of keeping the machines running. Their range of activities encompasses the setup of machines, solving of machine breakdowns, loading and unloading of parts for machines with manual material handling, the deployment of production utensils and tool preparation. Internal transportation of parts is supported by a logistician.

4.3 Planning and Control Approach

At the case company, production planning is centralised with a hierarchical structure. From an organisational view, three major parties are involved in the planning and control process. These are the planning department, the supervisor and the operators on the shop-floor. Once a week, production planner, supervisor and operators discuss the production plan for the respective week.

Planning Department

The production in the gear department is partly forecasted and partly order based, i.e. it involves make-to-order and make-to-stock products. The company works with an ERP software package called “Exact Globe Next”. Sales orders are planned with the use of an MRP programme which is part of the ERP software package. Production forecasting is done with a stock module called "Slim4", or "Slimstock" which helps to level out seasonal variations in demand. Production orders for stock parts are created directly from Slim4. An estimate of make-to-order parts is included in the stock module to facilitate planning. Material and component requirements are derived from the expected demand for finished products. Most of the components used for the gear production are internally produced.

(31)

respective machine, an additional day is granted for the corresponding process step. Orders are allocated to the machines accordingly.

For the detailed planning, the data is loaded manually into a company internal Excel-tool. The tool provides a production plan for each machine and can be manually adjusted by the planning department. For this allocation, a theoretical value of machine capacity is used that equates 90% of the machines’ design capacity.

Supervisor

According to the production plan, as determined in the planning department, orders are provided to the supervisor in printed format. The supervisor distributes production orders based on the production steps to the operators on the shop-floor. Each production order is printed on A4 and attached to the physical product. It provides order relevant data such as production steps, planned production dates, batch size and the end location. The supervisor is further responsible for planning and allocating operator requirement.

Operators

Operators are subsequently responsible for the execution of production. Displays at each machine show the respective production plan for the machine. As soon as the operator starts with a new order, he scans the barcode on the A4-paper. In addition to that, the progress of order within the production process can be traced.

(32)

24 | P a g e

5 Results

5.1 Part 1: Specification of FMS Constraints

To specify the major constraints of the automated machines in the gear department we used a combination of an inductive and deductive approach based on PPC constraints as summarised by Fox et al. (1983). We recall that the focus in this research is set on physical, causal and availability constraints.

5.1.1 Analysis of FMS Characteristics and FMS Environment

The automated machines in the gear department are of FMS type 1, encompassing the following elements:

 general-purpose CNC-machines,  automated material handling,

 a computer control system that is automatically operating the system.

AGVs are not used at the case company, hence, are not further considered in this research. Physical Constraints

It was found that the main physical constraint at all machines is machine capacity. Machine capacities are determined by the machine design and therefore differ per machine and function. Since machine capacity is shared among multiple, different products, setups are required. Setups include the loading of the CNC-programme, adjustment of parameters, tool changes and changes of chucking devices, and the assurance of material supply. Setups also include a machine warm-up period and a setwarm-up check. During a setwarm-up order processing is not possible. The duration of setwarm-ups gets higher with increasing production progress. While setup times average to around 10 minutes at the first production steps, they can amount to up to six hours at the last one, right before the assembly.

(33)

Causal Constraints

Causal constraints of the automated machines at the case company are represented by the production utensils needed. Most of the automated machines require tools, chucking devices, (raw) material, the required CNC-programme and fixtures/boxes. Presence of production utensils is crucial for machine operation. As different products require different utensils, the number of different utensils - particular tools, is respectively high. For each function utensil requirements differ, so do utensil constraints. While chucking devices are most crucial for the toothing, tools are the main constraint at the turning machines. Tools and chucking devices are changed by the operators. Therefore, at the case company, problems concerning the capacity of tool magazines are not present. In the gear department, tools and chucking devices are prepared beforehand to keep machine downtime during setup at a low level. This preparation is particularly relevant for new product types due to the increased amount of time needed. Apart from the availability of the required resources, we note that the number of fixtures and boxes is limited, restricting unmanned operation capability of the machines.

Availability Constraints

Availability constraints in the gear department are represented by first, availability of skilled operators and secondly, availability of shared production utensils: While sharing of production utensils is rare in the gear department with the exception of some tools and chucking devices, availability of operators is highly relevant.

(34)

26 | P a g e Based on this analysis, we specified the following constraints for PPC of FMS type 1. Physical Constraints  Shared machine capacity

Causal Constraints  Interference Availability Constraints  Resources

Zooming out from the perspective of single machines, we found that flexibility in routing as encouraged by the flexible machines, is an additional constraint to be considered in the planning and control in an FMS environment. Based on the product type, shape and size, a product has its routing from raw material to the final part. Hence, pre- and post-processes differ. Therefore, schedule changes at one machine inevitably affect the schedules of multiple subsequent machines. Thus, we added routing as fourth FMS constraint. We note that the importance of each constraint differs among functions (see Appendix C for an elaboration). A detailed overview of the machine constraints is illustrated in figure 6.

(35)

The refined research framework is shown in figure 7.

Figure 7 Specification of FMS Constraints in the Research Framework

5.1.2 FMS Constraints in PPC

(36)

28 | P a g e Shared Capacity

Planning department:

Capacity planning is based on a theoretical value of machine capacity. This value is adjusted in case of weeks with low batch sizes. Further deviations from this value are neglected in production planning. Even though there is no infinite machine loading, often machines are planned over the 90%-capacity value.

Similarly, the planning calculates fixed time slots for processing and setup times. Processing and setup times are not deterministic and therefore can vary largely in both directions. Difficult parts may require a longer cycle time as new fixtures and programmes have to be made. Also, an operator might need extra time to study the product when it is new. For simple parts, the calculated time frame might be too high.

Hence, appropriate matching of orders with machine capacity is impeded.

Furthermore, a pooling of orders is barely done in the production plan (see Appendix E). Supervisor /Operators:

Based on experience, supervisor and operators can suggest capacity adaptations and order pooling in a weekly meeting with the production planning and operators. Nonetheless, an inappropriate capacity planning on the shop-floor is experienced.

Consequence: A previous study in the company (Bizimana 2016) found a high utilisation

of the machines in the gear department that results in long waiting times for production orders. Orders require a longer processing time than estimated which generates waiting time for other production orders. This means production orders are blocked by other production orders. Ultimately, this results in a chain reaction in which several production orders have to wait for the machine. Internal performance records of 2016 show that waiting time is the main component of total order cycle time, amounting to 90.13% while actual product touch time took 9.87% of the time (see Appendix F).

(37)

Interference Planning department:

Currently, operators and their skills are not considered in production planning. As a consequence, task interference is possible. Neither is interference of production utensils considered in the current production planning.

Supervisor:

The supervisor is in charge to match the received plan/schedule with operator constraints. Operators:

Operators have limited influence to change production sequences beforehand. In case of interference, the respective order or task is antedated.

Consequence: According to the production manager, both interferences of operators and

utensils occur, leading to waiting times that inflates order cycle time. Resources

Planning department:

As mentioned above, orders are released to the shop-floor without consideration of operator and utensil availability.

Production planners do not account unavailability of certain utensils except unavailability of raw material. Still, also in case of raw material unavailability, schedules remain mostly unchanged.

Supervisor:

The supervisor is in charge to match the received plan/schedule with operator constraints. Operators:

Availability of utensils is ensured by the operators on the shop-floor. In case of unavailability of orders, operators again antedate the respective order and continue with the subsequent one. Further, as is the case with the production supervisor, operators can suggest schedule changes at the weekly meeting.

Consequence: Internal recordings show that the machines are scheduled over the availability of operators at the respective time, which results in overtime. The average overtime hours within the department amounted to 1450 hours in 2015 (Bizimana 2016). Thus, with the current approach, we observe a negative impact on production efficiency.

(38)

30 | P a g e synchronisation. As a result, cycle times of orders become uncertain. Secondly, this results in an impediment of proper preparation of production tools as the schedule provided by the system differs from the actual one. This is particularly problematic in the case of new product types which call for a high preparation effort. Particularly in the later process steps, setups require a high amount of time.

Thus, we observe that limited consideration of the availability of production utensils and order pooling in the planning has a negative effect on production efficiency.

Routing Planning department:

The routing of an order from raw material until final assembly is determined in the bill of materials of a product. The allocation of orders to machines is automatically done by the system based on the product characteristics of the order, hence is considered in the current production planning approach.

Schedule changes done in the system are automatically incorporated for subsequent machines. Supervisor/Operators:

In case of unplanned disruptions, the supervisor and operator are asked to react to these changes. Usually, this means that another order is chosen. However, supervisor and operators have limited insights into the planning. Hence, it is hard to judge the consequences of sequence changes on production performance.

Consequence: If changes are made on the shop-floor without system synchronisation, the effect of the respective change is not visible, and consequences of this change remain uncertain. Therefore, the cycle time of orders remains uncertain.

The cycle time of production orders in the gear department is planned with around 21 days. Bizimana (2016) found that actual product cycle times in the gear department vary between 25 and 50 days, illustrating the large variability of cycle times.

(39)

5.1.3 Problem Summary and Analysis

The limited consideration of FMS constraints in the production planning and control approach impacts production performance negatively. In figure 8, the relation between addressing FMS constraints in PPC and production performance is summarised:

Figure 8 Relation between FMS Constraints and Production Performance

As stated by the planners and production manager, reasons for not addressing FMS constraints in the current production planning are as follows:

 Time capacity to manually incorporate constraints

 Insights and data available about the happenings on the shop-floor, such as actual capacity, processing times, setup times, operator availability and skills, production utensils

 The lack of decision support to determine a feasible sequence in line with FMS constraints due to its high complexity

Supervisor and operators subsequently have to cope with a relatively fixed, but often infeasible production plan on the shop-floor. Due to the lack of addressing FMS constraints in the production plan, the supervisor and operators are in charge to achieve a matching of the received plan with production constraints. However, supervisor and operators on the shop-floor struggle with the following issues:

 Lack of flexibility in the current approach to make major changes  Lack of transparency to prevent interference problems

(40)

32 | P a g e

5.2 Part 2: Potential of Industry 4.0

With reference to the maturity index of Schuh et al. (2017), we allocate the gear department of the case company to the digitalisation stage. While connectivity has been partly achieved by connecting production planning with machines, the production plan is – to mention one example, still distributed in paper format to the shop-floor. Hence, once the digitalisation stage has been completed, we declare visibility (“seeing”) as first and next stage of Industry 4.0 (near future). We subsequently regard the three remaining stages (transparency, predictive capacity and adaptability) as distant future. Near and distant future were analysed separately. The results are presented in sections 5.2.1 and 5.2.2, respectively.

5.2.1 Visibility (Near Future)

Recall that in the visibility stage an increased amount of data, captured by sensors, RFID tags, automatic localisation and identification tools, and others, enables real-time monitoring of all activities on the manufacturing shop-floor. The potential of I4.0 technologies in the visibility stage was determined based on the findings of part I, complemented with literature and the input of a shop-floor control system supplier. We specified the impact of the application of above-mentioned technologies for production planning and supervisor/operators separately for each of the prior identified FMS constraints (to-be). We included the column “as-is” to revisit the findings presented in section 5.1.1. Additionally, we marked the additional potential through the implementation of shop-floor control systems in blue colour.

Shared Capacity: As-is To-be Planning Department:  Estimated values for capacity, process and setup times  Order pooling barley done

 The availability of precise data allows a more accurate matching of orders with actual capacity.  Deviations of production planning become visible

which allows to reduce capacity over-planning.  Order pooling could be stimulated.

Supervisor/ Operators:

 Suggestion of capacity adaptations

 Operators can react faster to capacity bottlenecks with sequence changes.

(41)

 Decisions based on feeling and experience

 More precise allocation of orders with actual

capacity.

 Visualisation of deviations in real-time allows fast

reaction to capacity bottlenecks and unplanned events.

Operator and Utensil Interference: Planning Department:  No consideration of operator and utensil interference

 No considerable improvement due to complexity of data analysis (in real-time) and time-intensity to manually incorporate constraints.

Supervisor/ Operators:  Attempt to avoid operator and utensil interference as good as possible

 Real-time data enables better matching of

production plan with operators and with utensils.  Faster reaction to interference problems.

 Allocation of resources to orders in the shop-floor

control system can help to reduce interference problems. Resources: Planning Department:  No consideration of operator and utensil availability

 Status and location of resources allows the release of orders only provided that production utensils are available (and in good state).Yet, manual adaptation of production plan remain time-intensive (analogous to interference and capacity)

Supervisor/ Operators:  Supervisor matches operators with production plan as good as possible  Assurance of utensil availability as good as possible  Reaction to unavailability with order antedating

 Similar to interference, real-time data enables better matching of production plan with operators and utensils on the shop-floor.

 Localisation and identification of utensils allow a fast location of production utensils.

 Allocation of resources to orders facilitates matching

of orders with operator availability (incl. skills).

 Allocation of resources to orders can support

(42)

34 | P a g e Routing: Planning Department:  Considered in initial planning

 For the planning department, this means that deviations become real-time visible (yet, similar to above, re-scheduling remains time intensive) Supervisor/ Operators:  Sequence changes without knowledge of impact on subsequent machines (based on feeling and experience)

 Real-time data facilitates decision-making as the effect of changes on further machines is illustrated.

 Visualisation of production plan allows to better

judge the impact on subsequent machines.

Conclusion Planning Perspective (Planning Department):

Contrasting the potential of visibility with the problem analysis in section 5.1.3, we note that even though production planning is provided with real-time data, a consideration of all FMS constraints by the planning department would still be highly complex and time-consuming without respective data analysis tools. Adaptations would have to be made manually and in real-time. Production planner and manager at the case company substantiated that the availability of all data in real-time would not necessarily facilitate production planning activities due to the long time horizon of production plans. Nonetheless, we argue that production planning can approximate their planning activities with actual data.

Conclusion Shop-Floor Perspective (Supervisor/Operators):

The application of Industry 4.0 technologies in the visibility stage can particularly facilitate an addressing of FMS constraints on the shop-floor. More precisely, supervisor and operators can rapidly react to capacity bottlenecks, interference problems and resource constraints, eventually lowering machine downtime. Yet, with the current planning and control approach, there is limited freedom of action on the shop-floor. As a consequence, operators can only improve their reaction capabilities which impedes an exploitation of the full potential of these technologies.

(43)

system in the gear department resulted in an estimated cycle time reduction of 30% or more (Bizimana 2016). See Appendix H for an illustration.

5.2.2 Transparency to Adaptability (Distant Future)

To investigate the potential for FMS planning and control with growing maturity of Industry 4.0, we first derived some major technologies3 per maturity stage from literature. Secondly, we explored the potential of these technologies exemplary for resource constraints (see figure 9, operators excluded).

Figure 9 Resource Constraints

Major Technologies Transparency – Adaptability Stages

(44)

36 | P a g e According to Almado-Lobo (2016) two types of analyses can be differentiated. These are first offline analyses and second real-time analyses. Offline analyses are based on sophisticated statistical process models. While offline analyses usually have a relational database or warehouse cubes, real-time analyses use techniques such as “inmemory” and complex event processing. This enables to trigger actions even before the data is stored (Almada-Lobo 2016).

Predictive Capacity: As prior mentioned, the predictive capacity stage involves simulation techniques that enable an anticipation of future developments. Thereby multiple scenarios can be tested with respect to their likeliness to materialise. This allows reacting to future events promptly. Thus, the number of unexpected disruptions will be reduced, increasing robustness in production (Schuh et al. 2017).

Adaptability: The adaptability stage represents the last stage in the maturity model and depends on a successful establishment of the previous stages. We refer to section 2.2.2 in which we described this stage in greater detail.

Resource Constraints with growing Industry 4.0 Maturity

In the following, we illustrate the potential of evolving technologies to handle resource constraints in PPC in the stages transparency, predictive capacity and adaptability. To approach this stage, we took a closer look on the toothing function in the gear department. The characteristics of this function are briefly summarised in table 3. Further, we presume for a short moment that all processes are digitalised and visibility has been established.

For the toothing function, the following production utensils are needed: the respective CNC-programme, chucking devices, tools and blue boxes. Resource constraints trigger the difficulty to assure that the right utensils are at the right area at the right moment in the right quantity.

(45)

required decision-support to the operators on the shop-floor. Thereby waiting times can be reduced to a minimum.

Predictive Capacity: The application of simulation techniques in the predictive capacity stage allows to determine a near to optimal production plan which includes smart pooling of orders while maintaining a high due date performance. Thereby, the effort of utensil changes can be reduced. Considering the high setup times at toothing machines and the fact that order pooling is barely done in the current planning (we refer here again to Appendix E), this could lead to a significant performance improvement. Secondly, simulation techniques help to assure that the respective tools are prepared in time. This is particularly relevant for new parts as they require a long preparation time.

Adaptability: In the adaptability stage, production utensils are represented by smart objects which communicate with the automated machines. Central planning is no longer required, neither is a manual preparation of utensils. In case of disruption, the system itself determines a new production strategy to ensure highest possible production performance.

Table 3 Overview Characteristics and Constraints Toothing

(46)

38 | P a g e

5.3 Part 3: Validation of Results

5.3.1 Validation Part 1

Besides of serving as input for the analysis of part two, the interview with the managing director of the shop-floor-control system supplier was leveraged to match the findings from part one with the practical experience of the company. We found a large overlap of the constraints experienced by the shop-floor control system supplier and the ones we found in the case company (see Appendix G). Therefore, we argue that a high number of companies experience problems with their FMSs similar to the ones identified in the case company.

5.3.2 Validation Parts 1 and 2

The interviews with the validation companies delivered the following outcomes:

In company 1 “lots of elements are based on feelings.” The interviewees report difficulties to “keep turnaround time short and get a good loading of the resources”. Machines are planned with a capacity of 80%. Still, machines are often over-planned. Setup times amount to around 40% of the production time. Hence, order pooling would be highly beneficial to the case company. However, it is not stimulated. Operators and production utensils alike are barely considered in the companies’ planning and control approach.The interviewees’ state that is hard “to find a good mix of operators and machines”. The order sequence is determined by the operator on the shop-floor, yet the “operator only thinks for his machine and is not supported to make a choice […].” This implies that production sequence is determined regardless of due date realisation. Interference and unavailability of resources do not represent an issue in this company.

In company 2, the nester can be regarded as “planner”. Based on his experience and feeling

he picks an order. Yet, machines are mostly planned over capacity. His main objective is waste reduction and secondly the due date, hence order pooling is largely done. This also implies that routing is not taken into account. Resource and interference issues are not of an issue in this company.

(47)

(48)

40 | P a g e

6 Discussion

The objective of this study was to explore the potential of industry 4.0 technologies for production planning and control of FMSs. In section 6.1 the key findings are compared with prior research activities. Section 6.2 discusses the practical implications of the findings of this study.

6.1 Theoretical Implications

The conceptual model in this study is built upon two predictions. Prediction one is that the lack of addressing physical, causal and availability constraints (see Fox et al. 1983) in production planning and control impedes the exploitation of FMS advantages, hence affects production performance negatively. With the use of a single case study, we were able to substantiate that a lack of addressing these constraints in production planning and control leads to an inflation of cycle times, lower throughput and eventually a poor due date performance. It further allowed to specify these constraints for type 1 FMSs in a medium/low-volume-high-variety production environment and revealed that routing represents an additional constraining factor to be addressed in FMS planning and control.

Furthermore, this research endorsed that traditional production planning and control approaches are too static and impede proper addressing of FMS constraints in PPC. While decision hierarchies as proposed by Stecke (1985) allow an increased amount of flexibility to address FMS constraints, further adaptations in a production planning and control framework will most likely be needed to handle the shift from a centralised towards a completely decentralised production as triggered by the application of Industry 4.0 technologies.

(49)

implementation issues manufacturing firms face with FMSs. Neither does this research aim to replace current solution approaches as presented in the literature. We note, that for highly complex flexible manufacturing systems, these solution approaches do have their legitimacy. Further, this research approached planning and control problems from a more general perspective, while solution approaches in literature are highly focussed on one specific problem. Nonetheless, this research showed that existing solution approaches to production planning and control issues as proposed in the literature, such as mathematical programming or artificial intelligence yet found little propagation in small/medium-sized manufacturing firms, which resonates with the conclusion drawn by Raj et al. (2007) that these approaches lack practicability.

6.2 Practical Implications

With the adoption of new, advanced technologies, manufacturing companies anticipate an immediate improvement of production performance. The example of FMSs has demonstrated that technological advancement alone is not necessarily effective. Without proper consideration of the characteristics and respective constraints that come along with new technologies in production planning and control, an exploitation of the desired advantages is often impeded, once more demonstrated in this case study. The practical example further revealed that the main difficulties to address these constraints in PPC are engendered by a lack of data, transparency and decision support. Existing and new evolving technologies in the context of Industry 4.0 provide the opportunity to bridge this gap.

The case study furthermore demonstrated the importance of synchronised data throughout the departments to obtain a common data basis for planning and control activities. Therefore, we encourage manufacturing firms that have not yet started with the digitalisation of manufacturing processes, to successively drive this process forward. The same applies to companies that have not yet started implementing or discovering Industry 4.0 related technologies. As illustrated in this case study, already minor changes, such as the creation of more visibility, can largely facilitate an addressing of FMS constraints in production planning and control.

(50)

(51)

7 Conclusion

7.1 Answering the Research Question

The question of interest in this research was whether the application of Industry 4.0 technologies enhances production planning and control in line with FMS constraints. The short answer to this question is yes. To provide an answer to the how, we first draw the attention to the constraining factors that impact production performance. We found that shared capacity, interference, availability of respective resources and additionally, routing are the major constraints to be addressed in PPC to ensure a high production performance in an FMS (Type 1) environment. The how question can subsequently be answered as follows: This research illustrated that already early stages of Industry 4.0 enable a better addressing of FMS constraints in PPC: While the interpretation of data at the offline level in the visibility stage remains highly complex, there is significant potential to address FMS constraints on the shop-floor: On the shop-floor, real-time data for instance enables a fast reaction to capacity bottlenecks; localisation of utensils allows an immediate reaction to interference problems and unavailability of utensils; data synchronisation ensures that the right tools are prepared at the right time. These benefits can be augmented with the use of visualisation technologies. Based on this research, we expect that with growing maturity of Industry 4.0, advanced data analytics support planning and control activities in line with FMS constraints by providing decision-support. The reduction of human misjudgments and wrong decisions is reduced to a minimum which evidently affects production performance positively. Concurrently, production planning becomes increasingly aligned to the actual production and process steps, reducing the complexity to match production planning with FMS constraints. Eventually, in the adaptability stage, FMS characteristics (note: we do not longer speak of constraints) become part of an inter-connected system.

7.2 Limitations and Further Research

Exploring the Potential of Industry 4.0 for Production Planning and Control of Flexible Manufacturing Systems