Management summary
This research is the result of a graduation assignment for the master technology management at the Rijksuniversiteit Groningen. This research was conducted within VSH Fittings B.V. in Hilversum (VSH). VSH is part of the Aalberts group and produces a wide range of fittings for the distribution and regulation of liquids and gases.
(confidential) The output of the current production resources has to increase.
Analyses of the current OEE of VSH shows that most losses are located in the availability of machines and the bochtenlijnen (BLs) in specific. The BLs are accountable for 33,2% of all the downtimes (availability loss) of the two departments under consideration (Xpress). Therefore the goal of this research is formulated as follows.
How can VSH Fittings B.V increase the availability of the bochtenlijnen in order to improve the Xpress productivity?
First, the availability of the BLs was analyzed. The highest loss in availability is waiting for personnel and Technical Department (TD), namely 9,52% of the available production time. Waiting times stand for the time a BL is not producing and is waiting for a worker to start an activity. These waiting times are caused by a work organization which is not able to cope with fluctuations in demand on human resources.
In the current work organization BLs are controlled by a dedicated team and roaming TD engineers. Each worker has specific qualifications and it can occur that there are more tasks pending than idling qualified workers available, which leads to waiting times. The randomness of downtimes at BLs is the main cause for this fluctuation in demand. VSH currently has seven BLs divided over two production halls, controlled by two BL teams. Furthermore, VSH planned to introduce an additional BL in each production hall in the coming months. The additional BLs increase the randomness of downtimes which affects the work organization.
FIGURE 1: RESEARCH PHASES
As expected, simulations revealed that VSH loses 406,4 BL production days due to waiting times in the future state with the current work organization. This is an increase in waiting times of 158% and comparable to 17,4% of lost production time in the future state (Each BL has 260 production days a year). The availability of the system will drop due to the additional BLs and the work organization needs to be improved.
Three concepts to improve the work organization were selected. (1) Adding capacity, hire more personnel to increase the capacity of the work force, (2) workload sharing, allow workers to perform tasks of other workers, and (3) workload balancing, changing the responsibilities of workers in order to create a more balanced workload for all. These concepts individually and combinations of the three form seven solution alternatives. These alternatives are translated into experiments and tested in the simulation model.
Simulations reveal that the most significant improvement is achieved by the concept which combines workload sharing and balancing. This concept indicates a reduction of waiting times of 290,1 production days a year, which is comparable with 12,4% of extra production time. The gain of this increase in availability is estimated at €190.734,95 additional profit a year, excluding investments.
Preface
In the light of my graduation for the MSc. Technology Management I have conducted this research at VSH fittings BV in Hilversum. This thesis is the result of a research project which started in August 2011 and was finalized in May 2012. During these months I have been investigating the productivity of the organization.
Being a part of the VSH organization during this research has been a great experience for me. Other than the previous projects which were part of my education I was highly involved in the daily activities of the organizations and this confronted me with other aspects than I had experienced before.
I M P R O V I N G T H E P R O D U C T I V I T Y
O F V S H F I T T I N G S B . V .
A N T I C I P A T I N G G R O W T H
Author : Karel Reinders
Study : MSc Technology Management
Student number : 1516361
Email : k.a.reinders@student.rug.nl
Faculty : Faculty of Economics and Business
University Supervisor : dr. W.M.C. van Wezel University Co-assessor : dr. ir. D.J. van der Zee
Company : VSH Fittings b.v. Hilversum
Company Supervisor : ir. J. Ruissen, Technical Director
Date : June 2012
Table of contents
1. INTRODUCTION ... 5
2. COMPANY INTRODUCTION ... 6
2.1HISTORY ... 6
2.2VSH FITTINGS STRATEGY ... 6
2.3VSH FITTINGS PRODUCTION SYSTEM ... 7
3. RESEARCH FRAMEWORK ... 9 3.1CONDITIONS ... 9 3.2PROBLEM CONTEXT ... 9 3.3PROBLEM STATEMENT ... 11 3.4FOCUS AREA ... 12 3.5SUB QUESTIONS ... 14 3.6RESEARCH METHODOLOGY ... 15 3.7DATA GATHERING ... 16 DIAGNOSIS PHASE 4. SYSTEM DESCRIPTION ... 19 4.1BOCHTENLIJNEN ... 19 4.2WORKERS ... 19 4.3WORK ACTIVITIES ... 20 4.3.1 Downtime activities ... 20 4.3.2 Uptime activities ... 22 4.4PROCESSING ORDER ... 23 5. CONCEPTUAL MODEL ... 24 5.1PROBLEM SITUATION ... 24 5.2MODELLING OBJECTIVES ... 24
5.3INPUTS AND OUTPUTS ... 24
5.4MODEL CONTENT ... 25 5.4.1 Scope ... 26 5.4.2 Level of detail ... 27 5.5ASSUMPTIONS ... 27 5.6SIMPLIFICATIONS ... 27 6. DATA COLLECTION ... 28 6.1WORKERS ... 28
6.2SETUP AND ADJUSTMENT ... 29
6.3REACTIVE MAINTENANCE ... 30
6.4PREVENTIVE MAINTENANCE ... 31
6.5MAJOR REPAIRS (NOT BL) ... 31
6.6TASK PRIORITIES ... 32
7. SIMULATION MODELING ... 33
7.1SOFTWARE ... 33
7.2MODEL DESIGN ... 33
7.3VERIFICATION AND VALIDATION ... 35
8. SIMULATION ANALYSIS ... 37
8.1FUTURE STATE SCENARIO ... 37
8.1.1 Goal ... 37
8.1.2 Settings ... 37
8.1.3 Observations and analysis ... 37
8.1.4 Conclusion ... 38
8.2PROBLEM ANALYSIS ... 39
8.2.1 Observation and analysis ... 39
8.2.2 Conclusion ... 41
DESIGN PHASE 9. EXPERIMENTS ... 44
9.1THEORETICAL CONCEPTS ... 44
9.1.1 Increase workforce capacity ... 44
9.1.2 Workload sharing ... 44 9.1.3 Workload balancing ... 45 9.1.4 Summary ... 45 9.2SCENARIOS ... 46 9.2.1 Outcomes ... 46 9.2.2 Decision criteria ... 48 9.2.3 Evaluation ... 49 9.2.4 Conclusion ... 51 10. DESIGN... 53 10.1WORKLOAD SHARING ... 53
10.1.1 Centralized labour control ... 53
10.1.2 Physical separation ... 54 10.2WORKLOAD BALANCING ... 54 10.2.1 Team composition ... 54 10.2.2 Cross-training ... 55 10.2.3 Task transfer ... 55 10.3IMPLEMENTATION ... 56 11. CONCLUSION ... 57
LIMITATIONS AND FURTHER RESEARCH ... 57
12. REFERENCES ... 59 13. APPENDICES ... A
List of abbreviations
AMT - Advanced Manufacturing TechnologyBD - Buigdoorn
BL - Bochtenlijn
BM - Benchmark
BPS - Business Problem Solving
BU - Business Unit
CAPA - Corrective and preventive actions C-Press - Carbon steel press fitting
DOV - Diagnose Ontwerp Verandering DRBL - Dagrapporten Bochtenlijnen FIFO - First in First out
FTE - Full Time Employee
OEE - Overall Equipment Effectiveness
OTK - OEE Toolkit
PM - Preventive Maintenance
SIS - Schoonmaken Inspectie Smeren SMED - Single Minute Exchange of Dies S-Press - Stainless steel press fitting TD - Technical service Department TPM - Total Productive Maintenance VSH - VSH Fittings B.V.
V&V - Verification and Validation
WP - Wisselplaat
1. Introduction
VSH Fittings B.V. (VSH) is a subsidiary of the International operation, Aalberts Industries N.V., a group of companies listed on the Amsterdam AEX stock exchange. VSH Fittings is a company within the Flow Control activities of Aalberts Industries.2 The activities of Flow Control cover the
development, the production and the assembly of products and systems for the distribution and regulation of liquids and gases. Flow Control activities is focused on for instance the housing market, commercial buildings, private and public new-builds and renovation, utility networks, district heating and cooling, fire protection, irrigation systems, the beer and soft drinks industry, laboratory systems and industrial markets.1 Aalberts Industries is the largest European
manufacturer of fittings with sales offices and production facilities throughout Europe. VSH offers a complete range of brass (Brass) compression fittings, stainless steel (S-Press) and carbon steel (C-Press) press fittings. The complete range of products is manufactured in Hilversum.2
Due to the financial crisis of 2008 orders dropped for Aalberts Industries. Instead of cut backs the focus of Aalberts Industry turned to the acquisition of new customers in order to maintain the production level prior to the crisis. Currently world economies are recovering and probably will continue to do so in the coming years. Consequently the demand will rise accordingly and with the additional acquisitions in mind a reasonable increase in demand can be expected. VSH is anticipating this prediction and is planning to double their output in four years from now to fulfill this demand.
With a long and rich history the organisation has a lot of experience and knowledge. Machines are designed in cooperation with professional machine-builders but the tools for the machines are developed by the organisation itself. The great amount of shared experience and knowledge that is integrated in these machines and tools make them sophisticated, hard to copy or even unique. Although possessing the technical characteristics for cost leadership and operational excellence, productivity is still disappointing due to possible inefficiencies in production and needs to be reviewed. Improving the productivity for VSH should be the first step to doubled output and is therefore reviewed in this research project.
2. Company introduction
2.1 History
VSH originated in the beginning of the 1930´s from a merger between a couple of metalworking companies. In the beginning the company acted as a subcontractor making turned parts for OEMs. In 1957 VSH started producing plumbing components. Brass end-feed fittings were one of the first in a rapidly expanding range of products. In 1975 VSH marketed the famous VSH Super Compression Fitting. This was an instant success. The combination of a conical bored compression nut, a compression ring and a housing with a conical bore guarantee ease of installation. After all these years the well-known VSH Super Compression product range is still the benchmark in quality and performance.
Over time the VSH product range expanded as more and more fittings and accessories such as stop valves, gas valves, and check valves were added to the range. Each time the deciding factor behind the product development of VSH were quality and ease of use for the installer. In 1991 VSH was taken over by Aalberts Industries, and is part of the Flow Control activities within this organisation. Within Flow Control there is cooperation between companies which are producing products for similar market segments. The main components are fittings and accessories for water, gas, central heating and cooling installations. 2
In the last decade VSH has been developed into a market focused company. Customer’s wishes and needs are the starting point for the development and marketing of new products. The products that are produced by VSH are sold outside the Benelux region by an extensive network of sister companies and agencies. The focus in these international markets is on the brass VSH Super Compression Fittings and the C-Press and S-Press Fittings.
2.2 VSH fittings strategy
Understanding the strategy of VSH explains and validates decisions and activities performed within VSH.
2.3 VSH fittings production system
Organisation structure: VSH Fittings b.v. Hilversum is the only company of its kind in the Netherlands and is responsible for the production of 2.000 different products varying in type, material and dimensions. VSH produced a total amount of approximately 40 million products last year. These products are divided over the C-Press, S-Press and Brass fittings and are produced in three loosely interconnected Business Units (BU) each in more or less their own production hall. Generally the BUs work independently from each other and have their own team, although when needed single employees can be transferred to other BU´s, however this is avoided when possible. Machines are not shared to prevent cross-contamination, carbon steel and stainless steel affect each other and because of diverging production processes. The production organogram in Figure 2 shows the structure of the production at VSH.
FIGURE 2: ORGANOGRAM
Each BU is controlled by a production leader which is in charge of the management of the production and the personnel. Daily activities are performed by the team which is dedicated to the BU. Exceptional activities such as heavy machine breakdowns are executed by the Technical Service Department (TD) which covers all three BU´s. The production manager is responsible for the management of the production, adjustment and cooperation between the different business units.
covering 85, 120 or 168 hours a week, respectively. The amount of shifts depends on the demand and capacity of VSH. Products are produced in batches based on forecasts.
The brass department is not included in this research and will therefore not be elaborated any further. From now on can be assumed that given information is only concerning C-Press and S-Press (called Xpress) unless indicated otherwise.
Product: The VSH XPress system is a complete piping system with press fittings and tubes. The VSH XPress fitting system is applicable for different applications, for example potable water installations, heating and cooling installations and gas
installations. Press systems offer a lot of advantages compared to other connection techniques. Installation times can be significantly reduced when you use the smart connection technique of the XPress system (Figure 3). The
fitting is placed over a tube and is installed by tightening the fitting using a automated installation claw. The rubber ring and deformed material form a perfect seal. The fittings are produced in different shapes, straight, T-shape and elbow fittings. These types can vary in diameter from 12mm up till 54mm. The elbow Fittings can be produced up till 108mm.
3. Research framework
A structured research requires a suitable methodology, which supports the design of an effective, professional and efficient way of research. This will increase the quality of the outcomes (De Leeuw, 2000). In this chapter the research framework is elaborated starting with the complaints, followed by the problem statement, the conditions and the research approach.
3.1 Conditions
To assure a high quality research, boundaries have to be set within which this research is executed. The conditions linked to this research project comprise these limitations. These limitations influence the results of this research and the process of the research itself. VSH management has formulated limitations and this research being a master thesis leads to additional limitations as well.
Investments in new machines, space and radical modifications to existing machines is no focus area for this research project. Instead the design should fit the current production resources.
The design should focus on the departments C-Press and S-Press (called Xpress). The Brass department is not reviewed during this research.
Considering the time limitations the implementation of the proposed design is not part of this research project.
3.2 Problem context
OEE = Availability (A) x Performance (P) x Quality (Q) [1] TABLE 1: LOSS CATEGORIES
Six Big Loss Category OEE Loss Category OEE Factor
Equipment Failure Setup and Adjustment Idling and Minor Stoppages Reduced Speed
Reduced Yield Quality Defects
Downtime Losses Availability (A) Speed Losses Performance (P) Defect Losses Quality (Q)
An improvement of one or more of the OEE Factors should consequently lead to an improvement of the machine efficiency which is beneficial for the productivity of VSH. The OEE data covers 101 weeks of production and ends in week 17 of 2011. It has to be noted that the data is probably inaccurate. The inaccuracies are highlighted by the quality manager and acknowledged by operators, production leaders and production manager. Data inputs are often a very rough estimation of what actually happened. However, in this research it is assumed that these inaccuracies concern all machines and are evenly distributed. Therefore the data is used to make comparisons.
FIGURE 4: OEE VALUES XPRESS
Nakajima (1989) mentions ideal values for availability (90%), performance (95%) and quality (99%). He states that achieving these values yields a world-class performance. Figure 4 shows the OEE values of Xpress and it is clear that the factor availability is the lowest and deviating the most of the ideal values of Nakajima and will therefore be selected for improvement.
The Xpress OEE of VSH over the 101 weeks was 59,1%. Appendix A shows the separated OEE timeline of C-Press and S-Press and the individual OEE values which are 65,7% for C-Press and 52,6% for S-Press respectively.
machines with the highest downtime losses (machine 2031, 2037, 2064, 3010, 2070, 2063 and 2067) and responsible for a total of 33,2% of all downtime losses of Xpress.
The focus area of this research will be at the BLs for several reasons. First, the seven BLs are accountable for 33,2% of the downtime losses of Xpress. Secondly, the BLs are responsible for approximately half of the production output of Xpress and therefore a big element of production. Finally, the BLs are similar machines and therefore in the perspective of Dal et al. (2000) comparable using OEE.
To quantify downtime at a BL a simple sensitivity analysis is executed. The sensitivity analysis (appendix D) shows that, based on a single week, one lost production hour at one BL leads to an average output loss of 221 products which is equal to estimated opportunity costs of €179,55. The estimated profit per production hour can be derived from the sensitivity analysis and is determined at €37,57. There has to be noted that this sensitivity analysis is based upon assumptions and estimations which affects its reliability.
3.3 Problem statement
According to De Leeuw (2000) a problem statement consists of a research objective and a research question. The research objective represents the objective of the investigation, whereas the research question describes the knowledge needed to achieve the objective. The research objective can be formed as following:
Improving the Xpress productivity of VSH Fittings. B.V.
In this research project productivity will be defined as the ratio of realized output / theoretical output. The improvements will be focused at improving the availability of the BLs. The research question is formulated as following:
How can VSH Fittings B.V increase the availability of the bochtenlijnen in order to improve the Xpress productivity?
As with many other shop-floor indicators, data availability and integrity, along with the question of how to leverage data for corrective and preventive actions (CAPA), are the primary challenges of OEE-based performance management (London & Segev, 2003). Ljungberg (1998) states that OEE does not take into account all factors that reduce the capacity utilization e.g. lack of material input, lack of labour or personnel. Reasons for losses should be known, or else the activities will not be allocated towards solving the major losses in an optimal way.
3.4 Focus area
An analysis of the losses of BLs should highlight the most significant problem areas. According to a monitoring tool of the BLs (OEE Toolkit) the two highest loss factors are waiting for TD and waiting for personnel and represent 9,52% of lost production time (appendix E). Nakajima (1989) states that idling (waiting for) is considered as a loss and should be eliminated.
A BL is a highly autonomous machine cell which is a combination of multiple machines connected to each other by robots and conveyer belts. For organizations using advanced manufacturing technology (AMT), such as a BL, it has been suggested that suboptimal performance stems not so much from the in-adequacies of the technology itself but more from deficiencies in the associated work organization (Wall et al., 1992). The term work organization has no strict definition and is a broad concept which deals with how work is organized and managed. This concerns all possible elements which influence work, such as scheduling, personnel, job design, production floor lay-out etc. One definition which is adopted in this research, is that of Goubergen & van Landeghem (2002) who state that the organization of work can be described as ‘who does what when’.
During 57 random observations of all the BLs throughout the research period it became clear that during BL downtime often there was no human activity at all. This can be related with the 9,52% of waiting time at the BLs. According to work floor personnel these waiting times are caused by fluctuations in demand on human resources, due to simultaneous breakdowns of machines, or the lack of qualified personnel to cope with the demand, due to absenteeism. Consequently it occurs that the personnel experiences difficulties in handling the tasks efficiently and in time.
downtimes (appendix F) show that the demand is fluctuating. Figure 5 is a depiction of this fluctuation.
FIGURE 5: DEMAND FLUCTUATION OF THE PRODUCTION HALLS
Whenever the demand is high it will be more likely that tasks are waiting for human resources to become available. The high peaks indicate that there was a lot of downtime during that shift and logically can be stated that personnel experienced a high workload as their main responsibility is to keep the BLs operational. It is assumed that the demand peaks in Figure 5 have a significant influence on the work organization and are one of the main causes for the waiting times.
VSH allows personnel of different departments to assist each other when possible to cover for fluctuations in demand on human resources. However, during several observations personnel is idling in one hall, while the other hall is experiencing a too high demand on human resources. Due to the physical separation and the limited communication between the production halls personnel have almost no indication of when exchange of personnel is possible and beneficial. Work that needs to be done at the BLs consist of planned and unplanned activities. Planned activities are frequent and not labour intensive due to the short cycle times. Personnel has standard routines to manage these planned activities throughout the day. Without disturbances these tasks are easy to control.
In the coming months VSH will introduce two additional BLs identical to the current ones, which will increase the demand of the system and its randomness. The effect of this change is unknown and there is no indication that the current work organization will be able to cope with this changing demand. In order to give a quality advise this change should be included in the research.
This section showed that the focus area of the problem should be on reducing the waiting times. These waiting times are causes by the work organization which is having difficulties in coping with the fluctuating demand. The randomness of breakdowns seems to be the cause for this fluctuation. Moreover, VSH will introduce two additional BLs which will probably increase the fluctuation. In this particular situation a quality advise can be defined as a new work organization which handles this fluctuation of demand on human resources best. Logically the costs to establish this new work organization should not exceed the benefits.
3.5 Sub questions
The previous section highlighted the focus area for this research. In order to solve the stated problem in a structured way sub questions have been formed. These sub questions are as follows.
1. What is the current state of the work organization of the BLs?
This sub question will identify the current work organization. This work organization leads to waiting times which should be lowered.
2. How will the future state with the additional BLs affect the waiting times and what causes these waiting times?
The second sub question can be split in two sections. First, The sub question should analyze the effects of the additional BLs on waiting times. Because this change to the system is certain the advise should identify the work organization which suits this future state best. It is chosen to use a simulation model for this problem because the future state does not exist and therefore it is impossible to perform analyses in the real world. According to Robinson (2004) simulation models are able to explicitly represent the variability, interconnectedness and complexity of a system. The simulation model will be able to predict what the effects will be. Secondly, Once the effects are identified of the additional BLs on waiting times it has to be analyzed what causes the waiting times in the future state. Guided by theoretical concepts factors will be identified which cause the waiting times and the effects of changing these factors will be tested.
Finally, a redesign of the is given which improves the situation at VSH. In addition an implementation plan will be given for this redesign.
3.6 Research methodology
The design of this research project is based on the methodology of design-focused research of van Aken (2007). Design-focused business problem-solving (BPS) deals with improvement problems, not with pure knowledge problems. The methodology aims at the design of specific solutions for specific business problems, the design of the change processes needed to realize the solutions, and the development of support within the client organisations for these designs (van Aken, 2007).
Design-focused research is divided into three phases following the DOV model of de Leeuw (2000). DOV stands for Diagnose (Diagnosis), Ontwerp (Design), Verandering (Change). According to de Leeuw (2000) the diagnosis phase aims to transform signals of the organisation under consideration into a management problem, the design phase transforms the management problem into a tangible solution and subsequently the change phase transforms the solution into a tangible system. The focus of this research project is on the first two phases. The last phase is not included in this report. Figure 6 shows the research phases of this research report.
FIGURE 6: RESEARCH PHASES
First, in chapter four a general description of the system will be given based on observations, interviews and available data. This section will be the input of the simulation and will answer sub question one.
Once the simulation model is constructed sub question two will be answered using this model in chapter eight. Based on simulation results the factors that cause the waiting times for personnel and TD will be identified. This completes the diagnosis phase.
In the design phase possible solutions are identified and a solution for the problem will be given. In chapter nine simulation experiments will be created based on theoretical concepts. The suitability of these concepts will be evaluated with the simulation model. Afterwards all solution alternatives are compared and the best design for the stated problem will be chosen.
Chapter ten will answer sub question three. The chosen solution will be elaborated and translated into a redesign which fits VSH. In addition an implementation plan for the design is given, which completes the design phase.
Chapter eleven will be a conclusion for this research project with the most significant findings. The research project will be evaluated and limitations and directions for further analysis are highlighted.
3.7 Data gathering
During this research several sources have been used to collect data. This section briefly elaborates on the sources used.
The first data source is the OEE data of VSH. The OEE data is from a long period in which a lot of improvements have been implemented, such as Single Minute Exchange of Die (SMED) of Shingo (1985) and Total Productive Maintenance (TPM) of Nakajima (1989). Therefore it is plausible that the data does not represent the current state of the machine availability. In the remainder of this research the OEE data is not used for analyses due to this inaccuracy.
The second source is a more reliable data set of VSH. Since December 2010 VSH uses a new and more reliable data collecting system called “OEE toolkit” (OTK). Due to license restrictions this toolkit monitored only several critical machines within VSH and was idling several weeks because of expired licenses. In 2011 OTK monitored BL C1 (the oldest and slowest BL) and BL C5 (the most stable BL) for 39 weeks. OTK is more reliable as it is an automatic system which checks output a machine, and once a machine stops producing output a timer starts and has to be labeled by personnel in the correct category. At the BLs this lead to 39 different loss categories with reliable numbers, assuming that the labeling is done properly.
report the setups, major malfunctions and major downtimes during that shift. At last the personnel working during the shifts are tracked.
The fourth source is TD data. All repairs by TD are documented in a system called AS400. In May 2011 this system was replaced by another system called M3. A previous analysis of the malfunction reports of the AS400 system of Januari 2007 till May 2011 brings up 3583 reports distributed over 489 unique malfunctions for the BL´s.
4. System description
Before going into detail on specific areas mentioned in the problem statement, a general description of the system under investigation is given in the next section. In the subsequent sections more details are introduced to illustrate the used interpretation and to derive a conceptual model which sets the boundaries and functionality specifications of the simulation model.
4.1 Bochtenlijnen
A BL is a highly automated machine cell which produces elbow fittings out of pipes. This research is only interested in the activities performed at a BL when it is not functioning due to all possible factors. During production a BL does not require any external inputs, it is only when a BL stops functioning that activity is required from workers. For the sake of this research is not necessary to go into detail about the functioning and elements of a BL itself and it is chosen to use a high-level black-box approach for a BL from now on. This also means that minor differences between the BLs are ignored from now on.
VSH currently owns seven BLs which are located in two different production halls. Production hall 3 (C press) contains four BLs and production hall 1 (S Press) contains three BLs. As stated before each production hall will get an additional BL in the near future. The production halls are located within a two minute walk from each other. In both halls the BLs are located next to each other. A schematic layout of the production halls with additional BLs can be found in appendix I. Production is done in batches compiled by the planning department. Planning of orders is based on demand forecasts and each BL has always several orders queuing. Planning takes setup time reduction into account by assigning orders with similarities to the same BL. Normally BLs run on a 17,5 hours two shift schedule. If demands are exceptionally high for a specific product and deadlines are putting pressure on the production capacity it can be decided to switch that particular BL to a 24 hours three shift schedule to increase its capacity.
4.2 Workers
Each type of worker has its qualifications. Training programs are part of the daily routine, for instance, to turn operators into setters. This training can be seen as ‘watch and learn’. Operators join a setter on his tasks and that way they should get familiar with the setter’s tasks. This means that operators get more capabilities over time, however, they are still not qualified to perform setter tasks on their own unless there is a setter who supervises the activities. Consequently the multifunctionality of workers is not affected by the internal training at VSH until a training program is completed.
BLs have a dedicated team which is responsible for maintaining operationality of only the BLs. Due to physical separation of the BLs over the two production halls there are two BL teams in each shift, each responsible for the BLs in a production hall. This separation of workers is preferably maintained under normal conditions, but it is allowed to assist the other team when labour capacity allows it.
Teamsize is three or four (depending on the production hall) and are preferably composed out of a teamleader, setter and one or two operators to facilitate flexibility. The actual amount of each type of worker can vary due to absenteeism and strategic scheduling decisions. Technical engineers are not part of the team and are only called in when a BL is suffering a major breakdown. Each department has its own technical engineer per shift.
4.3 Work activities
A BL team can be seen as a workpool which moves around the BLs in the production hall. During the day there are several tasks to be performed. Whenever tasks needs to be performed a qualified worker is taken from the pool to perform the task. These tasks have different natures and the most significant activities are elaborated below.
4.3.1 Downtime activities
Even though a BL is a highly automated machine cell and capable of performing all the production steps by itself, it requires human intervention for supporting activities. Not all of these activities require a BL to be stopped. Due to safety regulations a BL has to be stopped whenever a worker is working inside the cell. In the light of this research project only activities that require downtime of a BL are significant and therefore all other activities will not be elaborated unless they have an presumed influence on the problem. The activities that do require downtime are the following:
When a setup is completed products often do not meet quality standards yet. The time from the first product until the first product which meets the quality standards is called adjustment and is inextricably connected with setup.
(2) Reactive maintenance; Repairs are needed because equipment fails. According to Hopp and Spearman (2008, p272), equipment failure refers to preemptive outages because they occur whether we want them to or not. Preemptive outages can greatly influence variability of a process and occur completely random. Consequently, in the perspective of VSH this can lead to high fluctuations in demand on human resources.
During production a BL can stop producing for several reasons. Once this happens the green light on a BL automatically switches to yellow or red based on the classification of error. A worker responds to this light and identifies the details of the malfunction. The problem should be fixed by the worker when possible or reported to the TD if it exceeds the capabilities of the present personnel. A distinction is made between minor maintenance, executed by regular personnel, and major maintenance, executed by TD.
(3) Preventive maintenance; Preventive maintenance is executed whenever there is enough labour capacity. Hopp et al. (2008, p275) categorize preventive maintenance as nonpreemptive outages because they will inevitably occur but for which we have some control as to exactly when. In contrast, a preemptive outage, which might be caused by catastrophic failure of a machine or when a the machine becomes radically out of adjustment, forces a stoppage whether or not the current job is completed.
In the light of TPM several preventive maintenance activities have been introduced in order to prevent machines from breaking down and to guarantee quality. Elements that cause a lot of downtime when malfunctioning are replaced before the breakdown occurs. Three different preventive maintenance activities can be identified.
A. Schoonmaken, Inspectie, Smeren (SIS); These activities contain cleaning, oil level monitoring/refill and lubricating. SIS is divided in daily and weekly schedules. In fact daily SIS is performed at the end of every shift because the BL has to be cleaned at the end of each shift. However checking oil levels and lubricating only happens once a day. B. Buigdoorn replacement; A buigdoorn (BD) is a crucial element in a BL which endures
C. Wisselplaat replacement; A wisselplaat (WP) is also a crucial element in a BL which endures a lot of wear. To prevent product quality to fall these elements are replaced on a periodic basis.
In the perspective of VSH this implies that a worker should not perform preventive maintenance when there are higher priority activities waiting which are not controllable. However, the guidelines force workers to execute preventive maintenance even if labour capacity does not allow it. This makes the preventive maintenance preemptive as reactive maintenance, which in turn can lead to wait times at other tasks.
4.3.2 Uptime activities
Besides the identified tasks that require downtime of the BL, several tasks which do not require BL downtime can be identified which affect the availability of workers. According to Molleman & Slomp (1999) the bottleneck worker determines the makespan, there will be more waiting time for the other team members if the workload of the bottleneck worker is too high. Therefore the other activities of workers have to be identified. The most frequent uptime tasks are as follows.
(1) Prepare setups; In the light of SMED VSH uses a setup cart which minimizes setup times. Before a setup is performed this cart has to be prepared by loading it with the appropriate tools and BL parts.
(2) Store material; After the setup is performed the setup cart is loaded with parts of the previous batch. These need to be stored at the correct spot in the storage area. The moment that the parts are needed for a new batch is usually far away due to the setup scheduling. Therefore there is no time pressure for this task at all and it can be postponed when needed.
(3) Loading and unloading; The input and output buffers of the BL need to be loaded and unloaded in time. These buffers have a high capacity and do not hamper production under normal conditions.
(4) Monitoring and measurement; Workers should monitor the BLs and identify disturbances and quality drops. Samples are taken from finished products and several tolerances are measured to ensure quality products.
coincidentally this particular activity can also occur during downtime of a BL. However in this research this will not influence the uptime classification for this activity due to the identical priority.
(6) Remaining; This rest category comprises all other irregular activities, for instance mandatory meetings, cleaning activities outside a BL, improvement projects etc.
4.4 Processing order
VSH employs several informal operating guidelines to control the processing order of tasks. A general description is as follows.
When a new task emerges a random idling worker which is qualified for the job is assigned to that job. It can occur that all qualified workers are already working on another tasks. If the new task has a higher priority than the current task the worker is transferred to the new task if possible.
Tasks can also be transferred over workers when needed, provided that multifunctionality allows it. For instance a setter is doing operator tasks and a new setter task emerges. The idling operator is assigned to the current tasks of the setter and the setter takes the new task. This way the workload can be evenly distributed over available personnel.
If two tasks with the same priority emerge it is up to the workers to decide which task is performed first. Which task is picked depends on the moment. For instance, a worker will prefer a task which will be completed within his own shift above a task that spreads over multiple shifts.
Tasks with the highest priority cannot be influenced by workers. These tasks have to be performed as soon as possible. All other tasks can be influenced by the workers themselves. It is up the them to decide which order is believed to be most effective and can be called as ‘the inspiration of the moment’.
5. Conceptual model
The conceptual model is a non-software specific description of the simulation model that is to be developed, describing the objectives, input, outputs, content, assumptions and simplifications of the model (Robinson, 2004). The purpose of the conceptual model is to set out the basis on which the computer based simulation is to be developed. It is in effect a functional specification of the computer software. The coming sections will form the conceptual model.
5.1 Problem situation
To create a model which adequately describes the real world it is necessary to have a good understanding of the problem situation (Robinson, 2004). Section 3.4 highlighted a problem situation and changes which most likely affect the availability of the BLs. A recapitulation will be given here.
The waiting times for personnel and TD are too high. These waiting times are causes by the work organization which is having difficulties in coping with the fluctuating demand. The randomness of breakdowns seems to be the cause for this fluctuation. Moreover, VSH will introduce two additional BLs which will probably increase the fluctuation. In this particular situation a suitable advise can be defined as a new work organization which handles this fluctuation of demand on human resources best. Logically the costs to establish this new work organization should not exceed the benefits.
5.2 Modelling objectives
According to Robinson (2004) the objectives are the means by which the nature of the model is determined, the reference point for model validation, the guide for experimentation, and a metrics by which the success of the study is judged. The objectives should provide three key features; (1) what has to be achieved, (2) what level of performance is required, and (3) what are the constraints? These specific features have been formulated as follows.
Identify the work organization which improves the utilization of the nine BLs best by reducing waiting times for personnel and TD.
5.3 Inputs and outputs
According to Robinson (2004) the inputs and outputs are the experimental factors and responses respectively. In effect these are the means by which it is proposed that the objectives are to be achieved.
factors for this model. The work organization was defined as ‘who does what when’. Experimenting with different settings of who does what when should compose the different scenarios which are tested under comparable conditions to identify performance levels of each work organization scenario.
According to Robinson (2004) the responses (output) have two purposes. The first is to identify whether the objectives have been achieved. The second purpose is to point to the reasons why the objectives are not being achieved. Like the inputs these are derived from the modeling objectives. The responses should give an insight in waiting times which can be seen as lost production time. The responses should also highlight bottleneck areas to identify why the objectives are not achieved. Table 2 lists the experimental factors and responses for this research.
TABLE 2: EXPERIMENTAL FACTORS AND RESPONSES
Experimental Factors
• Work organization scenarios
Responses (to determine achievement of objectives) • Waiting time for personnel / TD (lost production time) Responses (To identify reasons for failure to meet objectives)
• Total waiting times for each activity (bottleneck task identification) • Worker utilization (bottleneck worker identification)
• Task occurrence (Validation) • Average task duration (Validation)
The experimental factors and first responses are directly derived from the modelling objectives. The second responses are needed to identify why the desired results are not achieved. For this research four secondary responses have been formed to highlight the specific problem areas. The first two concern bottlenecks to identify which aspect creates the most waiting times, the last two are needed to validate the simulation model in chaper 7.3.
5.4 Model content
5.4.1 Scope
The scope determines which elements of the real world system have a significant impact on the link between the inputs and outputs. Chapter four illustrated a general description of the system under consideration. Table 3 lists the elements of this system and the decision to include or exclude it.
TABLE 3: SYSTEM ELEMENTS
Component Include/exclude Justification
Bochtenlijnen Include Are subjectable to wait times
Workers Include Experimental factor
Downtime activities Include Determines demand on labour capacity
Uptime activities Include Can influence worker utilization
Processing order Include Create waiting times
TABLE 4: LEVEL OF DETAIL
Component Detail Include/exclude Comment
Bochtenlijnen
Amount of BLs Include Experimental factor Location Include Needed for team allocation
Changing demand Exclude Continuous demand is used to identify utilization
Dynamic production
schedule Exclude
Fast majority of 2011 was on a two shift schedule and is taken as standard Workers
Type Include Experimental factor
Qualifications Include Experimental factor Team size Include Experimental factor Team composition Include Experimental factor Staff roster Include Experimental factor
Absenteeism Exclude
Not explicitly modelled, but could be represented by pertubations to the staff roster
Downtime activities
Setup and adjustment Include Preemptive activities causing waiting times Reactive maintenance Include Preemptive activities causing waiting times Preventive maintenance Include Preemptive activities causing waiting times Uptime activities
Prepare setup Include Crucial part of setup Store material Exclude No effect on waiting times Loading and unloading Exclude Buffers are sufficient Monitoring and measurement Exclude No effect on waiting times
Major repairs Include Preemptive activities causing waiting times for TD
Remaining Exclude Sporadic events and therefore irrelevant for modelling
Processing order
Task priorities Include Influences task processing order
5.4.2 Level of detail
The level of detail of the included components must be such that it represents the components defined within the scope and their interconnection with the other components of the model with sufficient accuracy (Robinson, 2004). Table 4 lists the details of the included components.
5.5 Assumptions
The conceptual model uses several assumptions which are necessary for the establishment of the simulation model. These assumptions are discussed with personnel and superiors of VSH and acknowledged for the purpose of this research. The assumptions are as follows.
• Even though data is presumed to be inaccurate it is assumed that the combination of different data sources compensate for inaccuracies and are a sufficient representation of reality
5.6 Simplifications
According to Robinson (2004) a good simplification is one that brings the benefits of faster model development and run-speed (utility), while maintaining a sufficient level of accuracy (validity). Simplifications may also be necessary if the original model design is deemed infeasible, for instance, because required data are not available. On the other hand simplifications can affect the credibility of the simulation results. These effects have to be kept in mind and are analyzed afterwards in chapter 11. The simplifications for this simulation model are as follows.
• Presumed efficiency differences between workers are excluded because there is no data available on this subject
• Strategic decisions which influence teams, work schedules or priorities are excluded • All BLs are always on the same production schedule of two shifts.
• Once a task emerges and a qualified worker is idling, the worker starts immediately with the task on hand. Thus neglectance or unawareness of personnel about queuing tasks is excluded.
• Adjustment activities are considered minor repairs because of the comparable preemptive nature.
•
Workers assisting each other when idling to speed up tasks is not modeled due to simulation software limitations.•
Differences between setters and teamleaders are ignored and both categories are seen as setters.6. Data collection
Now that the elements of the simulation model have been defined, specific data has to be gathered to specify model parameters and probability distributions. This chapter will list all the available facts and figures about the identified elements.
6.1 Workers
As stated before operators are qualified to perform preventive maintenance only. A setter/teamleader is allowed to perform preventive maintenance, minor maintenance and setups. Finally, TD is qualified to perform the major maintenance. Table 5 shows a depiction of the multifunctionality. Which worker performs the uptime activities depends on the nature of the task. TABLE 5: MULTIFUNCTIONALITY Operator Setter/teamleader TD Setup x Reactive maintenance Major x Minor x Preventive maintenance SIS x x WP x x BD x x Uptime activities - - -Task
The team sizes are four and three workers in production hall 3 and 1 respectively. These are the standard team sizes, but these can be different due to external influences as absenteeism. Even though there are two teams they can cooperate. Exchange and transfer of tasks within a single teams can also be applied between the two teams. However, this does not occur as often as it should. Due to the physical separation and the limited communication between the production halls personnel have almost no indication of when exchange of personnel is possible and beneficial.
VSH has set three different shifts. The start, end and break times of these shifts are listed in Table 6. This research is only interested in the two shift schedule which comprises the morning and afternoon shift. During the night shift there is no production.
TABLE 6: WORKING HOURS PER SHIFT
Shift morning afternoon night
Start 6:00 14:30 23:00
End 14:30 23:00 6:00
6.2 Setup and adjustment
Products are produced in large batches. Derived from the DRBL the average between setups was 20,7 shifts over all BLs. Assuming that production is on a two shift schedule only (520 shifts a year) this leads to 25,12 setups a year per BL. Every BL has multiple batches in a queue and should never be idle because there is no order waiting. There is a maximum of 1 setup at a time per production hall. Batches are scheduled to spread the setups over time. Whenever batches in a production hall end at the same time, one of the batches continues producing until the other setup is complete.
In the light of the SMED approach setups have to be prepared off-line by placing all needed tools and equipment on a setup-cart with fixed slots for every element to prevent mistakes. Activities that are performed during on-line setup depend on the difference between current and next product and comprise mechanical, calibration and programming activities. The bigger the difference between current and next product, the more elements that have to be changed over. The setup preparation is done by an operator or setter and the actual setup is performed by two workers, of which one is at least qualified as a setter.
The average duration for a setup is 228 minutes per setup. The collection of setup times is given in Figure 7 and shows a curve which shows similarities with an exponential distribution3.
According to Robinson (2004) the exponential distribution has only one parameter, the mean. It is derived from the characteristics of purely random arrival processes and suitable to represent the time to complete a task, for instance repairs.
FIGURE 7: SETUP TIMES
6.3 Reactive maintenance
As stated before there is a distinction to be made between major and minor reactive maintenance. Major maintenance comprise all repairs that are the responsibility for TD and minor maintenance are all repairs executed by the BL team. The AS400 analysis reported 489 unique malfunctions at the BLs which were repaired by TD. The repairs done by regular personnel should be added to this number. Consequently can be concluded that reactive maintenance has very diverse nature and no standard procedure because it is determined by the specific nature of the malfunction. For the sake of this research a black-box approach is used for the reactive maintenance.
As stated by Hopp et al. (2008) equipment fails on a completely random interval. In the perspective of VSH this is not different. Observations and interviews acknowledged that there is no pattern to be found in the way equipment fails.
Major:
After each repair done by TD a report is filed. In a previous research4 at VSH these
reports were analyzed. Based upon this research TD filed 1307 repaired malfunctions at seven BLs in 2010. This leads to 187 major repairs a year per BL. Derived from the OTK the average task duration of the major repairs is 176,7 minutes per repair.
The distribution of the cycle times are fluctuating severely. The 1307 malfunctions were spread over 489 unique malfunction classifications, which all require its unique repair time. This leads to huge outliers which are visible in the OTK reports. Consequently, the variation of repair times will be described as exponentially distributed.
Minor:
Aside from some informal notes the minor repairs are not documented at VSH. However, the OTK gives an approximation of the amount of repairs and a general description. Derived from the OTK the average amount of minor repairs is 1042 per BL. The average task duration is 32,9 minutes per repair.
As with the major repairs, the minor repairs are also a blackbox which covers many different classifications of malfunctions. Due to the lack of data about these malfunctions and the similar nature as major repairs it is chosen to use an exponential distribution for repair time variation as well.
4 Kuijl R. G. A., van der, 2010, De Onderhoudsfunctie: Een vernieuwde onderhoudsfilosofie en
6.4 Preventive maintenance
These tasks are scheduled and have clear guidelines about when and how to perform. The arrival rate of these tasks is determined by work schedules (SIS) and BL output rates (BD and WP).
SIS:
Every BL has to be cleaned at the end of each shift unless it was not operational during that shift. The shift determines if the cleaning, daily or weekly SIS has to be executed. For the sake of this research all three categories are seen as one category called SIS because the duration is comparable. Based on OTK the average duration is 17 minutes per SIS. BD:
Depending on the diameter of the produced batch the BD has to be replaced every X amount of processed pipes. The amounts are determined by VSH but on average a BD is replaced every 375 processed pipes. On average it takes 12 minutes to replace a BD based on OTK reports and multiple observations.
WP:
As stated before WP are replaced every 100 processed pipes. This is also determined by VSH. On average it takes 12 minutes to replace WP based on OTK reports and multiple observations.
The cycle times of preventive maintenance are normally distributed. The tasks are simple and due to its frequent arrival rate workers are highly familiar with the activities. Based on OTK reports the cycle times have a maximum deviation of the mean of 5 minutes for SIS and BD, and 4 minutes for WP.
6.5 Major repairs (not BL)
The major repairs performed on a machine other than a BL are important because these repairs are also performed by TD. It can therefore occur that a TD is busy repairing another machine while a BL fails. This may cause waiting times at the BL for the TD to become available.
The research of van der Kuijl4 also analyzed the malfunctions of all the other machines within
6.6 Task priorities
The task priorities determine which task is performed first if tasks are queuing. These guidelines are informal and can be altered by workers when this is believed to be beneficial. However this does not occur very often and a general task priority list can be given.
Reactive maintenance has the highest priority. Whenever a BL stops producing, restoring operationality becomes the highest priority of the BL team. Tasks that can be performed during BL uptime have to be postponed and resumed once the BL’s operationality is restored or is transferred to TD.
Setups and preventive maintenance require downtime of a BL but the moment of this downtime is controllable in most occasions. Except SIS, these tasks have the second highest priority. Shifts have to clean all used BLs before the shift ends. Consequently, SIS only gets a high priority when a shift is about to end. All lower priority tasks can be interrupted in favour of higher priority tasks.
In most cases uptime activities do not have a fixed moment for execution and are therefore highly controllable. Therefore uptime activities have the lowest priority. However, the uptime activities of TD (major repairs) have a different priority.
Priorities of TD are defined on the moment itself. There are many factors that can influence the priority of a task and in unusual situations the production leader determines which task gets the highest priority. This complicates the process, however, in most cases a first in first out (FIFO) principle is used. For the sake of this research the FIFO principle is adopted and strategic decisions will not be modeled in the simulation.
Table 7 summarizes the priorities per worker. A one stands for the highest priority and a four for the lowest.
TABLE 7: TASK PRIORITIES
7. Simulation modeling
This section will focus on underpinning modeling decisions and explaining the simulation model. Subsequently the model will be verified and validated to demonstrate that the model is sufficiently accurate for the problem at hand.
7.1 Software
The software used for the simulation model is ‘Corel iGrafx Process 2003’. The decision to use this simulation software is based on the familiarity and availability of this software to the author of this research report. The simulation model will only be used for this single application and does not require any other features than determined in the previous chapter. The chosen software has sufficient capabilities for modeling the stochastic problem at hand. Another favorable feature of the software is the very short simulation run-speed which enables more simulation possibilities within the limited time for this research and faster model development.
7.2 Model design
This section elaborates on the design of the simulation model. Figure 8 shows the visual depiction of the simulation model. The aspects that are significant for this research are given here.
The simulation model is divided in three different departments, Timer, Hall 3 (C), and Hall 1 (S). All activity generators are located in the timer department. This department has two functions. On the one hand it creates the demand for setups, major repairs, minor repairs, other TD (utilization TD) and production orders, and on the other hand it divides the demand over the two production halls.
Hall 3 (C) and hall 1 (S) have identical elements and simulate the processes which need to be executed for each activity. The only difference can be found in amount of BLs and workers. Hall 3 (C) has 4 BLs, 2 operators and 2 setters/teamleaders. Hall 1 (S) has 3 BLs, 1 operator and 2 setters/teamleaders.
The processes are modeled as parallel processes. Each process acquires a BL at the right moment which takes it out of production. The cumulative waiting times of the process steps after acquisition determine the wait times for personnel or TD.
elements in the model, the same result will be obtained. This is crucial for this research as the inputs are scenarios which need to be compared under identical conditions.
As stated before the inputs for this simulation model are scenario’s. Because the model in only used for this purpose the inputs are not separated from the data. The scenarios can be created by changing parameters of the model. iGrafx stores the results of each simulation run so that simulations can be compared.
The outputs give an insight in waiting times and utilization levels of the whole model. The outputs are reported in tables. These tables list more outputs than needed and for the sake of this research report only the significant numbers are extracted from the output tables and implemented in a custom made table.
7.3 Verification and validation
Verification is the process of ensuring that the model design (conceptual model) has been transformed into a computer model with sufficient accuracy; in other words, building the model right. Validation, on the other hand is the process of ensuring that the model is sufficiently accurate for the purpose at hand; in other words, building the right model (Robinson, 1997).
According to Robinson (1997) it is not possible to prove that a model is absolutely correct. Therefore, model verification and validation (V&V) is concerned with creating enough confidence in a model for the results to be accepted. This is done by trying to prove that the model is incorrect.
The advice of this research project will be based on the future state simulations. There is no data available of the future state which complicates the V&V. However, Robinson (2004) states that if there is no real world data then comparison can be made against the expectations and intuition of those who have a detailed knowledge of the real system, or against other models. Comparison against approximate real world data such as these may not give absolute confidence in the model, but it should help to increase confidence. It is chosen to V&V the current state model since there is enough data available of this system. Afterwards, expectations of changes to the results due to additional BLs will be tested to create confidence in the future state model.
A detailed description of the various V&V can be found in appendix G. Based on the results of the V&V is assumed that the model does represent the future state with enough accuracy to serve the goal of this research. The most important findings are selected and elaborated in this section.
7.4 Model application
This section elaborates on how the simulation model is used for the purpose of this research. This is briefly explained in order to create an understanding of how the scenarios are created. A detailed manual is not composed because the model is not designed for reuse or other problem statements different as the one considered in this research.
The model represents the current state of the BLs. The model can be used for identifying the effects of changes to the work organization by executing experiments. These experiments consist of a scenario which is tested under identical circumstances in order to identify differences in results. A scenario can be adjusted by changing several parameters which creates a different work organization. This is done before the simulation is executed and cannot be adjusted during the simulation.
8. Simulation analysis
This chapter will illustrate the effects of the future state on the current work organization. In order to identify these effects the outcomes of the current state model will be used as a benchmark. In order to generate the outcomes of the future state a scenario is simulated.
8.1 Future state scenario
This scenario should give an answer to the first element of sub question two. In order to create a structured scenario it is divided in four phases, the goal, the settings, observations and analysis, and the conclusion of the scenario.
8.1.1 Goal
The goal describes the desired achievement of the scenario. The goal of this scenario is to simulate the effects of adding two BLs to the demand on human resources. This in order to create understanding of how the current work organization will react on these changes.
8.1.2 Settings
This section will shortly explain the settings used for the scenario. The current state model serves as a foundation for the future state. To create this future state several elements have to be changed. This section will shortly explain how the current state model is translated into the future state model.
The difference between current and future state can be found in the number of BLs and its associated demand. To simulate these changes some parameters will be changed. First, each production hall gets an additional BL which can be acquired for the activities in the hall. Second, to simulate the extra demand each generator, except utilization TD, will generate 28,6% more activities. Finally, this demand has to be properly divided over the two production halls as the current ratio of 3:4 (hall 1 : hall 3) will be change to 4:5. All other parameters remain untouched. 8.1.3 Observations and analysis
A selection of the most important outcomes are given below. Table 8 shows the waiting times for the current state (Sim #1) and the future state (Sim #2) and the increase. These waiting times are the sum of all waiting times of all the activities.
TABLE 8
Sim #1 Sim #2 Increase
Wait for personnel 81,62 139,68 71,1%
Wait for TD 76,11 266,72 250,4%
Wait for TD (not BL) 12,23 13,82 13,0%
