at
Meijer Plaatbewerking
(Make-to-order Industry)
Author: Remon Peters
Sint Jacobiparochie, May 2009
ii
Restructuring Planning and Production Control
at
Meijer Plaatbewerking
Author: Remon Peters
Student number: s1463489
Study: Technology Management
Specialisation: Process Technology
University: University of Groningen
Faculty: Economics and Business studies
Primary supervisor: J. Riezebos
Secondary supervisor: I. ten Have
Company: Meijer Plaatbewerking
Location: Sint Jacobiparochie, Netherlands
Primary supervisor: J. Oost
Secondary supervisor: H. Spoelstra
iii
MANAGEMENT SAMENVATTING
De motivatie van dit onderzoek is om een manier te vinden om Meijer Plaatbewerking minder afhankelijk te maken van de planner dan in de huidige situatie. Om dit te onderzoeken is eerst gekeken naar waar deze afhankelijkheid uit bestaat. De mogelijke aspecten waarin deze afhankelijkheid zit, zijn het planning proces, de planner zelf, het productie proces en de onderlinge relaties daartussen.
Voordat er aan het daadwerkelijke onderzoek kon worden gestart, heeft er eerst een vooronderzoek plaatsgevonden. Hierin is kennis gemaakt met het productieproces, het planningproces en de relaties daartussen. Dit vooronderzoek heeft een beeld gevormd over hoe de organisatie van deze processen in elkaar steekt. Deze fase was de verkennende fase en hieruit volgde het ‘management probleem’. Hierin komt naar voren dat in het uit te voeren onderzoek de prestaties van het huidige werkvloer beheersingssysteem, met de focus op de beheersing van de convergente productie stromen moeten worden onderzocht. Verder moet de methode voor het opstellen van het productie schema en het bijhouden van de beschikbare capaciteit onder de loep worden genomen.
Om de prestaties van deze systemen te onderzoeken, moet het vergeleken worden met andere, voor Meijer relevante, systemen. Deze vergelijking is gemaakt op basis van eerder wetenschappelijk onderzoek. Hieruit bleek dat Meijer Plaatbewerking is in te delen als een ‘Job shop’ organisatie dat opereert in de ‘Make-to-order’ omgeving. De producten ondergaan verschillende bewerkingen die sterk in productietijd kunnen variëren. Een opvallende eigenschap bij dit soort organisaties is dat er lange wachtrijen van producten en half fabricaten aanwezig zijn in het productieproces. Door deze lang wachttijden en de hoge variëteit in productietijd is het lastig om van te voren in te schatten hoeveel capaciteit er beschikbaar is. Verder blijkt vanuit de eerdere wetenschappelijke onderzoeken dat in ‘Make-to-order’ omgevingen de doorlooptijd en de hoogte van onderhanden werk niveaus lastig te beheersen zijn.
iv punten van de planning en productie systemen tegen elkaar uitgezet en vergeleken met de (mogelijke) benodigdheden voor Meijer Plaatbewerking.
Figuur 1 Theoretisch raamwerk (zie hoofdstuk 3.3 voor verdere uitleg en referenties)
De aspecten in de grijze vlakken moesten verder worden onderzocht bij Meijer in een verdiepend onderzoek om te kunnen concluderen of het huidige planning en productie beheersing systeem hier mee om kan gaan, en zo niet, hoe dit in te passen in een nieuw planning en productie beheersing systeem.
Uit het verdiepende onderzoek bleek dat het huidige planning en productie beheersing systeem niet goed in staat was om de convergente stromen te beheersen. Dit kwam vooral doordat de bewerkingen voorafgaand aan de assemblage bewerking geen inzicht hadden in welke onderdelen allemaal voor welke samenstelling bestemd waren. Verder is er bij de assemblage afdeling een grote variëteit in kunde van de medewerkers. Lang niet alle medewerkers kunnen alle producten maken. Dit aspect zorgt er voor dat de beschikbare capaciteit bij de assemblage afdeling lastig is in te schatten. Door deze twee aspecten is het bij de lasafdeling “hollen of stil staan”. De ene keer moet het wachten op half fabricaten terwijl het later wordt overspoeld met producten die alleen een ervaren lasser kan assembleren.
v Bij eerste deel van het herontwerp is via de werkvrijgave lijst meer informatie verschaft over de volgende bewerking. De datum dat de volgende bewerking zou moeten beginnen aan het product en voor welke samenstelling of klantorder het product bestemd is, is weergegeven op de werkvrijgave lijst. Hierdoor kunnen de voormannen van de verschillende productie afdelingen betere keuzes maken over de volgorde van werkuitgifte aan de werknemers.
Bij het tweede deel van het herontwerp zijn de assemblage werknemers (lassers) gecategoriseerd in drie groepen (A, B en C). Vervolgens zijn de klanten die producten vragen die volgens bepaalde specificaties (ook A, B en C) geïdentificeerd. Het is de bedoeling dat er in plaats van de huidige capaciteitsgroep ‘lassen’, drie capaciteitsgroepen worden gemaakt bij de planning. De planner heeft dan een beter inzicht in de beschikbare capaciteit op de assemblage afdeling.
Tot nu toe is er druk getest met de nieuwe werkuitgave lijst. Deze slaat goed aan bij de voormannen van de productieafdelingen. Er zijn plannen om er nog meer informatie aan toe te voegen zoals een materiaal check (is het benodigde materiaal wel aanwezig) en een vorige bewerking check (is de vorige bewerking al gereed). Uiteindelijk wil het Management Team er naar toe dat de werkuitgifte lijst continu wordt geüpdate door middel van draagbare barcode scanners die de bewerkingen meteen aan- en afmelden en materiaal binnen boeken.
vi
PREFACE
In order to complete my study Technology Management I performed my final research and wrote a thesis at Meijer Plaatbewerking in Sint Jacobiparochie. I planned the research would take four to five months. But once I started at Meijer, I was involved with other activities not directly concerning the thesis. Because of this, the research took almost 6 months at Meijer. I enjoyed it a lot that I could participate in some other projects like the newly installed two bin system at the tube lasers, the idea of re-organizing the assembly department and the restructuring of reports to get the required information out of the ERP-system for different purposes.
I learned a lot at Meijer. Working with the ERP-system, redesigning reports generated by the ERP-system and implementing theories on planning and production control. The big eye opener for me was that not re-designing a system was the hardest part, but to implement (a part of) that system was far more difficult. Although the employees who were going to be affected by the changes of the re-design were involved in a early stage, it was still a difficult task to get the system running. But once it worked, it was a great thing to see that the theories I have learned over the past 5 years in Groningen were put into practice.
I would like to thank my supervisors and all the colleagues at Meijer who always had time to help me if I had questions or ran into difficulties. I learned a lot from their different perspectives on the subjects we encountered, and hopefully, they learned from my inputs and perspectives as well. Furthermore I would like to thank my supervisor from the University of Groningen, J. Riezebos, who guided me through the process of writing the thesis and performing the research very well. He always came with new ideas to tackle problems we encountered at Meijer. Finally I would like to thank my father who taught me many tricks in MS Excel which were very helpful in analyzing the loads of information gathered in this research, and for telling me how he solved similar problems I encountered at Meijer, at his own work.
I enjoyed the past half year at Meijer and I am glad that I can stay until September to further implement the re-design of the planning and production control system.
vii
INDEX
Management samenvatting ... iii
Preface ... vi
1 Introduction ... 2
1.1 Current state ... 2
1.2 Motivation for research ... 4
1.3 Structure of the report ... 4
2 Research design ... 6
2.1 Research goal ... 6
2.2 Boundary conditions ... 7
2.3 Research method ... 8
2.4 Pre-research ... 9
2.4.1 Production process analysis ... 9
2.4.2 Planning process analysis ... 17
2.4.3 Problem statement ... 21
3 Forming a framework for benchmarking ... 22
3.1 Organizations characteristics ... 22
3.2 Planning and production control processes ... 24
3.2.1 Planning and production control concepts ... 25
3.2.2 Selection criteria for planning and production control concepts ... 27
3.3 Theoretical framework ... 29
3.4 Implementation of a re-designed system ... 30
3.5 Conclusion ... 32
4 Diagnose ... 33
4.1 Production process performance ... 33
4.1.1 Quality ... 33
4.1.2 Flexibility ... 34
4.1.3 Work In Progress ... 34
4.1.4 Speed ... 35
4.1.5 Reliability ... 38
4.1.6 Performance of the entire production process ... 39
4.2 Planning process performance ... 40
4.2.1 Quality ... 40
4.2.2 Flexibility ... 40
4.2.3 Speed ... 41
4.2.4 Reliability ... 41
4.2.5 Performance of the entire planning process ... 41
4.3 Conclusion ... 42
5 Design ... 43
5.1 Scheduling process and capacity loading ... 45
5.1.1 Scheduling ... 45
5.1.2 Capacity loading ... 46
5.2 Providing insight in capacity available at assembly ... 47
5.3 Shop floor control system to control the convergent flows... 48
5.3.1 Information required for priority dispatching ... 50
5.3.2 Transparency at shop floor level ... 52
5.4 Integration of the planning process and the shop floor control system ... 54
6 Change ... 57
viii
6.1.1 Implementing the new release list ... 58
6.1.2 Implementing the visualization on the shop floor ... 59
6.1.3 Improvement of scheduling module ... 59
6.2 Required changes in the near future ... 60
6.2.1 Assembly skill classes ... 60
6.2.2 Workload control ... 60
6.3 Points of interest at implementation in the near future ... 61
7 Conclusion and recommendations ... 62
7.1 Conclusion ... 62
7.2 Recommendations ... 63
References ... 65 Appendix 1 Organizational Structure ... A Appendix 2 Layout ... B Appendix 3 Products produced at Meijer ... C Appendix 4 Kanban-system ... D Appendix 5 POLCA-system ... E Appendix 6 Lead time vs. Process time ... F Appendix 7 New production schedule ... G
Used Abbreviations
ConWIP = Constant Work In Progress ERP = Enterprise Resource Planning FIFO = First In First Out
GFS = General Flow Shop
GJS = General Job Shop
LORC = Lean Operation Research Centre
Meijer = Meijer Sheet Processing (Plaatbewerking) MRP = Materials Requirement Planning
MSE = Meijer Special Equipment
MTO = Make To Order
POLCA = Paired-cell Overlapping Loops of Cards with Authorization PPC = Planning and Production Control (system)
RBC = Repeat Business Customizer
TOC = Theory Of Constraints
VMC = Versatile Manufacturing Company
WLC = Workload Control
2
1 INTRODUCTION
The history of Meijer starts in 1921. It was founded as a small family company producing products primarily for the agriculture. In the following years Meijer developed its operations to serve other industries like the building industry and land terracing. In the 80’s the focus grew more to machine building, construction and fork trucks. The fork truck components and especially the sheet processing got a strong impulse. Later on Meijer expanded its sheet processing operations so much that it was decided to split up the fork truck component building and the sheet processing. This resulted in the current situation of Meijer which consist of two factories: Meijer Special Equipment-Forks (MSE-Forks) and Meijer Sheet Processing (Plaatbewerking).
The focus of this report is on Meijer Sheet Processing (which is written as “Meijer” from now on). Meijer produces all kind of sheet metal-, trunk-, tube- and bar-products for various industries. Some examples of these products are frames for cow milking robots, frames for the container lifting device on garbage collection vehicles, parts for bicycle carriers, parts for clothing racks and parts for Meijer Special Equipment-Forks (MSE-Forks). In appendix 3 some pictures of the products produced at Meijer are presented.
The mission statement of Meijer is to be the first choice in the making industry for their customers, with the best “total solution”, with the focus on reliability, quality, logistical services and chain integration.
To produce this high variety of products Meijer uses all kind of machinery. Meijer has metal scissors, tube cutting lasers, sheet cutting lasers (Bystronics), a punch nibbling machine, setters, presses, rolls, folding machines, a winding drum, a grinder, saws, welding robots, welding workspaces and assembly workspaces.
1.1 Current state
3 At Meijer, the machines used for production are placed in a functional organization. The production and material control system is push-based. There are some exceptions though. Some customers have a fairly consistent ordering pattern and send forecasts of products needed, so the construction of those parts is done in a product orientated organization at a separated hall on the production floor. The production and material control system for these products is pull-based and is executed by Kanban cards (see appendix 4).
One of the aspects of a functional organization is that there is a high level of work in progress (WIP) on the shop floor (Nicholas, 1998). At Meijer this is also the case. The WIP is put in racks of the receiving department and the location is written down on a list at that department so it is traceable. Once the receiving department has capacity to work on that particular product it can be retrieved from the rack. The racks have different colours, blue is for raw material, red for work in progress and green for finished products. On the shop floor is also a storage tower for the smaller parts used in production. The internal transport between departments is done by the worker who has finished his operation on the product(s). When the receiving departments is ready to start the next operation of that product, a worker of that department will get the WIP out of the rack and will start working on the part(s).
The products produced at Meijer vary between the kind of operations needed, the number of operations needed, the routing of operations needed, if the product consists of sub parts or not, the number of sub parts which it consists of and the operating time per part. Every product produced is treated as a unique one, no groups are formed. Because of the functional organization of machines, the variability in demand for the products and variety of operations and routings of the different products, the predictability of the capacity of the different departments is low (Suri, 1998). This makes it hard for the planner to plan in the jobs. To correct for the inaccuracy of load prediction, inflated lead times are used in the MRP-system (materials requirement planning) to make sure the department is able to complete the operation on time. Between every operation three days are planned, and if the operation is done externally, one week is planned for this operation. The fact that some products produced consist of several other sub parts which are also produced at Meijer makes it even harder to plan all operations.
4 1.2 Motivation for research
Meijer is, just like its sister company MSE-Forks, member of the Lean Operations Research Centre (LORC) at the University of Groningen. Because Meijer’s goal is to compete with the best in the market they are giving more attention to process improvement lately. In November 2008 a workshop was attended by some employees of Meijer. At this workshop the POLCA system (see appendix 5) was introduced to these employees. They recognized that a pull-based control system could have the potential to improve the processes at Meijer. A pull system could reduce the work load of the planner drastically. The planner can plan on a much higher level then, and the detailed planning is done by the workers on the shop floor.
MSE-Forks has more experience with lean manufacturing. Because MSE-Forks has its own range of products, the shop floor can be organized product-oriented. Meijer does not have products of its own, so a purely product oriented organization is not possible. Meijer was curious how lean manufacturing could improve their operations while they had a wide range of parts to produce with a high variability in customer demand both in frequency and quantity.
Meijer is very dependent of the planner. He is one of the few within Meijer who is able to execute the required tasks. As a consequence, when something happens to him, there is nobody who can take over his tasks.
Furthermore, increasingly more customers are demanding shorter lead times. The three days between every operation is too long. Ever since Meijer is a member of LORC, they are looking differently at their processes in their organization. This leads to the motivation of this research: to investigate possibilities to make Meijer less dependent of the planner.
1.3 Structure of the report
5 Here a new design to improve the current situation is presented. This design is based on the theoretical framework, the research goal and the boundary conditions. Section 6 describes the change phase which involves the implementation of (a part of) the design. In section 7 conclusions and recommendations for further research are presented.
Notes:
End-products are products which do not require an assembly operation. These products are sent to the customer or to an external operation. In this research, if Meijer needs to perform non-assembly operations on the product after the external operation, it is treated as an end-product once Meijer receives it again. if Meijer needs to perform assembly operations on the product after the external operation, it is treated as a half fabricate.
At the welding department most assembly work on products is done. Sometimes assembly operations are performed which do not involve welding. At dispatching convergent flows are present as well. Only here the end-products and / or compositions are picked and kitted. When in this report “welding” or “assembly” is written, it can be treated the same, namely the assembly of several half fabricates into a composition.
A composition can be a product assembled from several half fabricates, but it can also be a customer order which consist of several different kind of products which all have to be shipped at the same time (on the same product carrier).
A Job is defined as; “all the operations performed on the entire customer order until the point it is send to the customer or sub contractor”.
6
2 RESEARCH DESIGN
This section describes the design of the research performed at Meijer. This section starts with the research goal and boundary conditions. Then the research method is described. After that, a pre-research is executed to identify the problem statement. This problem statement is the focus of the main research, which concludes the research design.
2.1 Research goal
The motivation of this research is to investigate possibilities to make Meijer less dependent of the planner. To investigate these possibilities, all aspects involved with the planning process at Meijer have to be analysed. To analyze all aspects involved with the planning process, Bertrand et al. (1998), Nicholas (1998) and de Leeuw (2003) conclude that the link with the production process and the production process itself should be analyzed as well. Furthermore, Kreitner et al. (2002) and Checkland et al. (2005) conclude that the human behavioural aspects in an organization influence the functioning of an organization. Because the dependency on the planner is central in this research, the behavioural aspects of the planner are analyzed as well when looking at the planning process. The aspects of the planning process which this research focuses on are presented in figure 2.1. This is the research system of this report.
Research system:
Figure 2.1 Aspects involving the planning process
7 are planned, work and material is released onto production (3). At production the orders are produced (c) and information on capacity available and progress of production is send to planning (4). The planner’s tacit knowledge and routines (d) influence (5) the planning and capacity loading. Three areas are investigated; the production, the planning and the planner’s tacit knowledge and routines. This leads to the goal of this research:
Research goal:
Analyse, evaluate and re-design if necessary, Meijer’s planning system in order to make Meijer less dependent of the planner.
In order to achieve the goal, the research could target on several aspects of the research system in different directions. In order to limit the scope of this research, more boundary conditions are formulated. These are described in the next paragraph. After that, the research method and the pre-research is described.
2.2 Boundary conditions
In order to complete the research within the targeted four to five months some boundary conditions are set. The main focus of the research is on the planning and the production process on the shop floor. Operations in other departments like sales, work preparation, purchasing and accounting are outside the scope of this research.
The re-design should ease the planning process for the planner while it does not affect the shop floor control performance in a negative way. Also the physical organization of machines on the shop floor has to be kept the same.
Since Meijer is focussing on lean operations, aspects of the lean philosophy have to be integrated into the re-design.
The findings of the research have to be useful both for Meijer, the University of Groningen and the research should meet the specifications required in order to pass my Masters degree in Technology Management.
8 order to find out if aspects are useable in this research, but there is no intention on improving these procedures in this research. The research is finished when Meijer has an improved situation and it is known which aspects to focus on in the future in order to improve the situation further.
2.3 Research method
The method used for this research is the Diagnose Design Change (Evaluate) method (de Leeuw, 2003). This method is used in this report because it takes into account the ambiguous character of management science. The problematic situation is improved in three phases which lead to the improved situation.
In the diagnostic phase the problematic situation is transformed into a management problem. The term problematic situation is used because the situation could also be an opportunity or improvement possibility which has to be implemented. A management problem can be defined as a system which has to be improved according to the vision of management. In this phase, insight has to be gained in the research system. When this insight is gained an assessment is done about the system, based on the performance and its improvement possibilities. This assessment is based on a theoretical framework and the research goal. At the end of the diagnostic phase an evaluation of that phase takes place.
In the design phase the diagnosed management problem is transformed into an improvement proposal. A detailed, clearly specified, re-design of the research system or a sub-system of the research system is made which should improve the current situation. The re-design is based on the theoretical framework, the research goal and the boundary conditions. A feedback loop to the diagnostic phase is performed in order to check the re-design’s improvements to the current situation. This feedback loop is the evaluation of the diagnostic and design phase.
9 the implementation of a new design of a system. After implementation of (a part of) the redesign an evaluation takes place again. Here the change process and the whole improvement process are evaluated in order to find out if the new situation is improved compared to the previous situation.
2.4 Pre-research
The goal of the research is to analyse, evaluate and if necessary, re-design the planning system in order to make Meijer less dependent of the planner. When looking at the planning system in figure 2.1, the cause of the dependability can be located in the functioning of the planning process, the functioning of the production process, the planner in person, or a combination of these factors. That is why in this pre-research, these factors are analysed and evaluated in order to identify the problem statement of the main research. First an analysis of the production process is performed, than the planning process is analysed together with the routines and required knowledge of the planner.
2.4.1 Production process analysis
At Meijer three kind of product types are produced; compositions, half fabricates out of which compositions are assembled and end-products which do not require an assembly operation. In figure 2.2 the percentages of the three types of products of the total number of production orders produced in 2008 are presented. Composition production orders start at assembly once all half fabricates are finished. So at the start of the production process there are only half fabricate production orders and end-product production orders. The ratio between those two types of production orders is 40:60.
Figure 2.2 Product types produced at Meijer (2008)
Product flow analysis
10 the parts produced were recorded every day on random times for a period of two weeks. The foreman of every production department was asked to which department they send the produced parts to. This resulted in the PFA presented in figure 2.3. The blue lines represent the most common routings identified in the PFA.
Figure 2.3 Product flow analysis
Note:
Because in the beginning of 2009 the world economy was in recession1, Meijer had a lot less orders to be produced than usual. That is why data from 2008 is used for further routing and product information. For all other analysis the most recent information is used.
Only the operations performed in the end of 2008 and in the beginning of 2009 at Meijer are presented in the PFA. What can be concluded from this analysis is that if the operations laser, scissors and punch nibbling are required, these operations are always the first. Furthermore can be concluded that if a welding operations is required, it is (almost) always the final operation. The number of operations per routing vary on average from 1 to 6. The most simple products only need a (laser)cutting operation. The more sophisticated products could require (laser)cutting, setting, pressing, drilling, winding and welding operations.
The PFA gives a first impression on the routings at the shop floor. But as said, at the end of the year 2008 and beginning of 2009 fewer orders were produced. So the information from
1
11 the PFA might not be representative for the normal business. In order to get a better understanding of the routings on the shop floor, a from-to-table (Slack et al. 2007) is created from the data obtained from the ERP-system. The routings from all production orders produced in 2008 are converted into the from-to-table (figure 2.4).
Production orders jan-dec 2008
Ass em bly Bys tron ic n on-fe rro Bys tron ic s teel Sci ssor s Dri ll no n-fe rro Drill stee l Grind ingm aste r Gri ndin g no n-fe rro Grind ing stee l Ispe ctio n Pun ch N ibbl e Saw Set ting no n-ferr o Set ting stee l Tube last er Wel ding non -fer ro Wel ding rob ot Wel ding ste el Win ding dru m Exter nal O pera tion Stock out sum in 13 2337 7982 845 19 133 11 3 13 0 1049 260 41 82 2057 454 90 1374 32 86 0 0 16881 Assembly 0 0 0 0 0 0 0 6 0 0 0 0 0 18 0 3 0 63 0 40 10 90 230 Bystronic non-ferro 0 0 0 0 216 0 968 241 0 0 0 0 780 0 0 412 0 0 52 0 0 96 2765 Bystronic steel 0 0 0 0 0 730 0 0 106 0 0 0 0 4171 0 0 104 1525 697 146 283 712 8474 Scissors 0 0 0 0 0 0 10 0 0 0 30 0 27 365 0 17 0 146 0 0 117 61 773 Drill non-ferro 0 0 0 0 0 0 40 21 0 0 0 0 186 0 0 268 0 0 8 26 31 49 629 Drill steel 0 0 0 0 0 0 0 0 34 54 0 0 0 442 0 0 42 959 53 164 119 265 2132 Grindingmaster 0 0 0 0 69 12 0 0 0 0 0 0 833 39 0 144 0 0 0 13 19 40 1169 Grinding non-ferro 0 0 0 0 52 0 14 0 0 0 0 0 176 0 0 113 0 0 0 82 45 133 615 Grinding steel 0 0 0 0 0 17 0 0 0 0 0 0 0 96 0 0 12 131 5 31 16 96 404 Inspection 12 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0 3 0 129 6 72 229 Punch Nibble 0 0 0 0 17 41 20 35 50 0 0 0 91 476 0 0 0 45 14 24 0 210 1023 Saw 0 0 0 0 0 37 0 0 4 0 0 0 0 9 4 58 5 210 0 0 12 17 356 Setting non-ferro 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2089 0 0 0 42 322 755 3208 Setting steel 0 0 0 0 0 90 0 0 0 0 0 0 0 0 0 0 151 3620 358 847 700 1894 7660 Tube laser 67 0 0 0 0 168 0 0 0 0 0 0 0 60 0 969 180 1112 0 40 72 597 3265 Welding non-ferro 32 0 0 0 0 0 0 208 0 0 0 0 0 0 0 0 0 0 0 2087 39 376 2742 Wedling robot 22 0 0 0 0 17 0 0 0 0 0 0 0 0 0 0 0 31 29 312 77 82 570 Welding steel 0 0 0 0 0 168 0 0 87 97 0 0 0 114 0 0 0 0 70 2421 319 2910 6186 Winding drum 0 0 0 0 0 61 0 0 0 0 0 0 0 21 0 112 155 275 0 99 43 952 1718 External Operation 65 0 0 0 2 82 0 5 0 30 0 1 2 21 0 17 4 177 2 0 2175 3699 6282 sum 211 2337 7982 845 375 1556 1063 519 294 181 1079 261 2136 5921 2061 4656 743 9671 1320 6589 4405 13106 67311 123 non-ferro 123 steel
123 most frequent routings
Figure 2.4 From-to-table
12 From the table the following conclusions, with respect to identifying groups, can be drawn:
• A lot of different routings are present but a big part of them do not occur frequently • Bystronic lasers, Setting and Welding have produced the most production orders • Setting and Drilling are operations mostly performed “in the middle” of the routing • Most products received by Welding are send from the Bystronic lasers, Tube lasers
and Setting
• Almost any operation can be the first one of the routing • Almost any operation can be the last one of the routing
From these conclusions a new and more detailed PFA can be performed. In figure 2.5 the shop floor is schematically presented and an explanation of the operations performed is given. When the shop floor processes are understood, an assessment can be made about its performance as a whole. Setting In c o m in g g o o d s Setting Tube lasers Bystronic Lasers Winding drum Welding / Assembly Welding Disp Grinding master Welding Drilling PN / cutting / sawing Setting Nesti ng [1] [2] [3] [4] [5] [6]
Convergent flows towards assembly Flows purchased half- or end -products
Flows towards dispatching wihout an assembly or welding operation
Conv ergen t flow s produ ction -orde rs Conv ergen t flow s custo mer-o rders Nesti ng
Figure 2.5 Material flows on the shop floor
The explanation of operations at the shop floor:
13 [2] At the Bystronics and the tube lasers orders which have to start the current day until the
orders which have to start two days from the current day, may be nested. This is to reduce the material waste. A three day window for nesting is chosen, so the rest of the production process does not have to wait too long for parts. At punch-nibbling, cutting and sawing no nesting is performed.
[3] Setting receives most of its parts from the Bystronics and the tube lasers. Most parts produced at setting have a short operation time. So parts from the previous departments are processed whenever the parts arrive at setting. The schedule is taken into account but is not needed all the time, because of the difference between the short operating time required and the department’s lead time which is given to the department according to the schedule.
[4] Several operations at Meijer like drilling, winding and several others are not frequently performed. The drilling operations are performed at the Tube laser department and the Bystronic department. Most of the other operations are also performed at the Bystronic department, and some at Setting. Each of these operations has a lead time of three days according to the schedule.
[5] At the welding / assembly department several flows come together in order to perform the next operation or to produce compositions. The compositions have assembly-queuing time. The operations can not be performed unless all parts are present. The parts can be received from all production departments, including incoming goods for purchased half fabricates.
[6] At dispatching products are picked together into customer orders. Dispatching also has assembly-queuing time because all products have to be present in order to finish the customer order. Different from the welding / assembly department is that the operations (picking) can start as soon as the first product arrives at dispatching. But the customer order can only be send to the customer if all products are completed.
14 Setting In c o m in g g o o d s Setting Tube lasers Bystronic Lasers Winding drum Welding / Assembly Welding Disp Grinding master Welding Drilling PN / cutting / sawing Setting Nesti ng [1] [2] [3] [4] [5] [6]
Convergent flows towards assembly Flows purchased half- or end -products
Flows towards dispatching wihout an assembly or welding operation
A B C D Conv ergen t flow s produ ction -orde rs Conv ergen t flow s custo mer-o rders Nesti ng
Figure 2.6 Division of shop floor into groups
The first group “A” that is recognized are all initial operations. These are the tube lasers, Bystronic lasers, punch nibbling, cutting and sawing. All raw material received at incoming goods are processed by operations in this group.
The next group “B” are all non-initial-non-assembly operations, mainly supplied by group “A” and for a small part by incoming goods. Other non-assembly operations like pressing and rolling which occur rarely, also have the profile to fit in group “B”. In figure 2.6a group “B” is presented.
15 When a job arrives at group B the number of operations required within group B is as follows: (based on all production orders produced in 2008)
- One operation required within group B: 70% (cumulative: 70%) - Two operations required within group B: 25,6% (cumulative: 95,6%) - More operations required within group B: 4,4% (cumulative: 100%)
The third group “C” are all welding and assembly operations. Here all sub parts come in and are assembled into an end-product. Some preparation operations like grinding and cleaning are performed here as well, but directly afterwards the assembly operations start. Sometimes non-assembly operations are performed here, for example when a product requires a seam to be welded together.
The fourth and last group “D” is dispatching. Here all end-products come in and are picked and kitted into customer orders and shipped to the customer, or end-products are send away to a sub-contractor to have their external operation performed on.
The performance of the operations within these groups is a factor for further diagnosis in the main research. But the control of the flows between the groups should be analysed as well, that is why the shop floor control is analysed.
Shop floor control analysis
The current shop floor control method is straight forward for the employees on the shop floor. The foreman prints the schedule and divides the work over its employees, starting on the production order on top of the list, working his way down. The foreman is free to switch the sequence of some production orders if it fits better at the moment, but the schedule is leading. When a department is very busy parts are delayed, which means receiving departments are waiting for parts or start the production of parts which are not planned to be produced yet to keep the employees busy. The foremen of the production departments know that between every operation required, about three buffer days are scheduled. Because of this knowledge, the foremen sometimes take the liberty to use these three days in production, which messes up the reliability of the production schedule.
16 at Meijer can be divided into the types: compositions, half fabricates and end-products (exclusive compositions). In figure 2.7 the ratio of occurrence of the different routings between the groups is presented.
Figure 2.7 Routings between groups
With the division into these four groups a clear flow can be recognized at the shop floor. Of course there are some exceptions (“Group C Group B”), but the frequency of occurrence is so small that it can be ignored, so these are left out of the research. A remarkable routing is “A B”, which occurs frequently for both compositions and non-compositions.
The operations performed at group D (dispatching) are different than the rest with respect to the flow that is suggested in figure 2.7. Convergent flows can be recognized at group C (welding / assembly) but in this figure the convergent flows present at group D are not drawn. Figure 2.8 presents an example of the production of two customer orders which demonstrates the convergent flows at group C an D.
Figure 2.8 Convergent flows
C A B 60 40 D Half fabricates End-products (no assembly required)
17 At incoming goods the raw material and half fabricates are available. The production in Group A can start. In this example the products can be nested at group A, so all parts from group A arrive at the inventory “location” for group B to D at once. The products are put in racks wherever space is available, that is why it is not a definite location. Group B can start producing once the first parts arrive at the inventory “location”. The ratio of “number of parts in : number of parts out” is 1:1 at group B (non-assembly operations). Parts can be destined for group C and D. Inventory “locations” for group C and D are black boxes for the sending groups. Group A and B know the destination of each part and know which part belongs to which composition according the work card. But there is no insight if all parts for a composition are completed. Group C is only able to start when all parts for the composition are available. Group D can only send away the customer order if all products are completed and kitted.
For the foremen of production departments of non-assembly operations there is no direct insight on which the half fabricates that are produced, are all destined for the same composition. Because of this lack of insight, decisions are made which optimizes the performance of the production department, but not necessarily the performance of the entire production process. This is visible at the assembly department. A high WIP inventory is present which can not be started on because not all parts required for the composition are finished. To get different parts (from different departments) destined for the same composition, simultaneously to assembly or dispatching is difficult to control in the current shop floor control system. The control of the convergent flows is a factor for further diagnosis in the main research.
2.4.2 Planning process analysis
18 The planner estimates if the production of an order can be realized within the due date without any problem. If this is the case, the production order is planned in according to capacity availability, estimated based on the semi-weekly capacity lists received from the production department and the weekly capacity meetings. If the planner estimates that the due date can be problematic, the finishing date of the part with the longest lead time is leading, and all other parts of that order are planned backwards from that date, if the customer agrees. If the customer does not accept a later delivery date (and when it is an important customer or a big order), the order is put on high priority. With this high priority, the planner gets authorization to plan with only one day between the operations or even more operations per day.
The planning is performed at the level of operations. Depending on the composition of the product, the product has to be planned in up to five levels. Most products require three or four levels of planning. The five levels are listed below:
- All operations that have to be performed to finish the sub part - All sub parts have to be finished to complete the compositions
- All end-products which do not require an assembly operation have to be finished to be picked at dispatching
- All end-products which are compositions have to be finished to be picked at dispatching together with the parts which do not require an assembly operation
- Sometimes (seldom) a composition consists of sub parts which consist if sub-sub parts (which contains the fifth level of planning)
In figure 2.9 the five levels of planning are visually presented. Level 0 is the level of dispatching and level 1 is the level of the finished product. The other levels are those of the other operations which send their products to welding and dispatching.
19 Planning operations
In order to analyse the planning process, the planner was interviewed and observed. From this information the planning method is obtained:
A customer places an order. This can be a new order, a repeat order or the calling down of a contract. When work preparation has processed the order into the ERP-system, the planner receives it, together with the due date. The customer order is loaded into the schedule by the planner. The loading is done based on the estimated capacity available. The due date is put in and the ERP-system generates a schedule. The planner then manually alters the schedule. In figure 2.10 the planning process is schematically presented.
Total available production order-pool Total planned production order-pool
Customer Customer Orders
Order intake and Order processing (wp and s*) Production Orders (coupled to cust. orders) Planning Weekly Capacity and production meeting Foreman production department Semi-weekly forecasted capacity per department Weekly extra/less capacity decisions Production schedule (incl. replenishment of contract parts) Production Production orders devided over the workers Mach. Breakdowns Missing employees Delays Shortages * work preparation and sales Input Output Information output/input Function/ operation Capacity availability in coming weeks Calling down of contracts Information about progress customer order Finished customer orders
Figure 2.10 Planning process (Prism model, based on Verweij, 2008)
20 Twice a week the planner receives a list of capacity available for the coming week from the foremen of each production department. This list, the weekly production meeting and the results of the weekly capacity meeting are the base of the capacity loading of the planner. Once the production schedule is send to the shop floor, the foreman divides the work production orders over the employees.
While interviewing and observing the planner, some issues with the current planning operations arose. These issues are described below.
Planning issues encountered
The ERP-system generates the start dates for the different operations required, based on the due date of the customer order and the routing of the product(s). When a product ordered by the customer is a composition, the ERP-system generates incorrect start dates for the sub-parts. Also, the ERP-system can plan the start of operations in the past, so the planner corrects the schedule in these situations.
For every order, the generated schedule is altered in a different manner, depending on the planner’s insights. These insights are gained by the following variables:
1. Priority of the order assigned by the sales department 2. Size of the order
3. Due date tightness of the order, which differs per customer 4. Information on capacity available from the shop floor 5. Foreseen capacity problems in a certain stage of production 6. Whether the products are on stock or have to be made to order
The variables 3 to 5 are mainly based on experience, tacit knowledge and the spreadsheet program for monitoring the available capacity made by the planner. The planner shifts with the start dates of the required operations in the schedule generated by the ERP-system, in order to fit the customer-order into the schedule and load the factory’s capacity. So every customer-order has its own story of how and why it is scheduled the way it is.
21 From the analysis of the planning process the following factors to focus on in the main research can be identified:
- The ERP-system generates an incorrect schedule for compositions and its sub parts.
- Every order is scheduled in a different manner based on insights of the planner which are partly gained by tacit knowledge and experience.
- The monitoring of capacity available.
2.4.3 Problem statement
The pre-research was initiated in order to analyse the planning system in order to find out what causes Meijer’s dependability of the planner. From this analysis several factors for further diagnosis where identified. These factors combined form the problem statement of this research.
The focus for further diagnosis consist of two parts. The first part is the performance of the operations executed and the performance of the shop floor control system described in paragraph 2.3.1. This is to evaluate the control of the flows and to evaluate if the control of convergent flows is a significant problem. The second part for further diagnosis is the performance of the scheduling method and the capacity monitoring in the planning process described in paragraph 2.3.2. Because the planner’s functioning is integrated in the analysis of the planning process, the planner is not discussed separately.
22
3 FORMING A FRAMEWORK FOR BENCHMARKING
In this section, companies similar to Meijer are compared to Meijer based on previous academic researches. By this comparison the performance of the current planning system can be evaluated. Furthermore, the theory can lead to concepts to improve the current planning system. In paragraph 3.1 Meijer’s characteristics are compared to characteristics of organizations in theory in order to find out which type of organizations Meijer can be benchmarked to. Once the characteristics of the planning and production control processes in this type of organizations are identified, they are discussed in paragraph 3.2. The theory described is then combined into a framework in paragraph 3.3. This framework is the base of evaluation of the current planning system. Finally, in paragraph 3.4, theory on implementing changes into a system is described. Designing a new system is only half the work, implementing it is also a challenge.
3.1 Organizations characteristics
Meijer produces products by customer order. In theory organizations like this are called make-to-order (MTO) organizations. Furthermore Meijer produces a high variety of products in low volumes for a high variety of customers. From this characteristic can be concluded that Meijer operates in what in theory is called; the high variety and low volume production environment. In this environment a functional process layout is common (Burbidge et al.,1989). In a MTO organization the number and variety of products is large and constantly expanding. The specifications of the current products produced are frequently changed by the customer (Stevenson et al., 2005). The production process at Meijer can be characterized as a job shop; a manufacturing facility that generates a variety of products in relatively low numbers and in batch lots (Nicolas, 1998). Although the organization is able to produce a great variety of products, the organization is often faced with huge problems with lead time control, as demand for capacity fluctuates over time (Suri, 1998).
23 characteristics. For example, arrival times and process times can be constant or variable, arrivals can consist of single parts or batches and the production process can consist of one single operation to several operations. All these characteristics influence the queuing within a manufacturing system.
Queuing exists in two types; blocking and starving (Riezebos, 2004). Blocking happens when the receiving department is still busy producing a part when the previous department already sends new parts to operate on. Starving happens when the receiving department has to wait for parts from the previous department to arrive.
Meijer has job shop characteristics and the shop floor is functionally organized. A high variety of products is produced in a high volume variability. As a consequence, processing times per operation per job are variable as well. An operation can only start with a production order when the previous operation is finished. These two factors conclude that the production process has dependent processes with variable arrival times between operations (Goldratt, 1994). The expected lead time for a job in such a production process is described by Hopp and Spearman (2000) for a product which requires ‘m’ operations by the following equation (equation 1).
Expected Lead Time = e e
m t t m c + − + + − ) 1 ( 2 ) ² c ² ( a e ( 2( 1) 1) ρ ρ [1] With: e m t m c − + + − ) 1 ( 2 ) ² c ² ( a e ( 2( 1) 1) ρ ρ = Queuing time ρ = µ λ ⋅ m = Utilization
λ = Mean inter arrival rate (1 / average number of arrivals per day)
µ = Mean processing rate (1 / average number of orders processed per day) te = Mean total processing time of a job
ca = Coefficient of variation of arrivals
(variation inter arrival time / mean inter arrival time squared) ce = Coefficient of variation of total processing times
24 At Meijer a high diversity of products is produced. The possibilities of reducing the variation of processing time needs further diagnosis, which is described in section 4. To control the inter arrivals between operations, several planning and production control concepts exist. In paragraph 3.2 planning and production control processes are described and discussed.
3.2 Planning and production control processes
According to Bertrand et al. (1998) planning is performed on two levels; on factory level and on operations level. In figure 3.1 a summarized model of the planning framework developed by Bertrand et al. (1998) is presented. The original framework is very detailed, that is why the author has simplified the model by combining variables into five new variables which cover the information required for this research.
Figure 3.1 Summarized planning framework (Based on planning framework of Bertrand et al., 1998)
25 From the problem statement can be concluded that factors involved with “workload control”, “material- and work release” and “shop floor control” functions should be further diagnosed. Furthermore should the inter arrivals between the operations be controlled. In paragraph 3.2.1 PPC concepts are described involving the factors described above. After that, in paragraph 3.2.2, these concepts are compared and is discussed which concepts are fitted best for Meijer.
3.2.1 Planning and production control concepts
For Meijer, a PPC concept is only relevant if it is applicable in a MTO organization in a low volume, high variety production environment. This means that the PPC concept should not be product specific and should be flexible. Concepts that have these characteristics are POLCA (Suri, 1998), Workload Control (Gaalman et al., 1996) , ConWIP (Hopp and Spearman, 2000), Theory of Constraints (Goldratt, 1990) and the ERP-system.
POLCA stands for Paired-cell Overlapping Loops of Cards with Authorization. For an example of a POLCA see appendix 5. POLCA uses cards to authorize production on the shop floor. POLCA combines elements of push and pull production. Material is pulled through the factory based on available capacity at downstream operations. In one loop (two connected production cells2), no more jobs can be processed or waiting than allowed by the number of cards given out to the production cells. Restricting the number of cards restricts the level of WIP in the factory. POLCA requires a high level MRP-system to coordinate the release of orders on to the shop floor. For convergent flow control, POLCA is a less fitted concept, unless it is accompanied with a release list which gives extra information about the rest of the routing or extra priority rules are implemented (Riezebos, 2008).
Workload control (WLC) is designed especially for MTO organizations. WLC concepts buffer the shop floor against external dynamics by creating a pool of unreleased jobs (Gaalman et al., 1996). Three levels of control are identified which respectively relates to the job entry, job release and priority dispatching. Job release requires three inputs; when to release a job from the pool, how many jobs to release and which jobs to release (Fredendall et al, 2009). Furthermore there has to be taken into account the direct and indirect capacity load of a department (Gaalman et al., 1998). The direct load is the work from jobs queuing at the considered operation and indirect load is work from jobs queuing from operations upstream.
2
26 Also WLC is less fitted for convergent flow control (Henrich et al., 2004; figure 3.2) unless it is accompanied with an other system that is fitted to control convergent flows.
Figure 3.2 Evaluation framework indicating ‘best fit’ for WLC (Henrich et al., 2004)
ConWIP is a shop floor control system that stabilizes the total number of jobs on the shop floor. Once a job is processed completely a new job is allowed to be started. CONWIP does not look at the loading of individual workstations, but limits the loading of the shop floor to a pre-set maximum WIP level. To implement ConWIP the total processing time and the same total capacity requirements of all jobs should not have a high variability, because only the number of jobs within the system are controlled. Furthermore, ConWIP is less fitted for convergent flow control, unless accompanied with an other production control system that is fitted to control convergent flows.
Theory of Constraints (TOC) is developed by Goldratt (1990). Traditionally, management have been dividing organizations into small, manageable parts with the goal to optimize each individual part in order to optimize the whole organization. According to Goldratt (1990; 1994) a change in a part of the system to optimize that part, does only affect the system as a whole for a very small bit. In order to optimize the entire system, a variable needs to be found which will affect the system as a whole. This variable is called a constraint. Sometimes more than one constraint can be found in an organization. Controlling the constraint means controlling the entire production process. This constraint could be an assembly station where different flows come together, so TOC is fitted for controlling convergent flows.
27 modularity of the system. But the core planning and production control functions have developed less rapidly than other modules (Muscatello et al., 2003). The idea of an ERP-system is that all business information is coupled and can be used by the organizations own employees but also that customers and suppliers can access information through an external communication interface. In order to keep the ERP-system reliable, new information needs to be put in and information needs to be updated continuously.
Now the PPC concepts applicable for organizations like Meijer are identified, a comparison and a method for selecting the best fitted concepts for Meijer is discussed in paragraph 3.2.2.
3.2.2 Selection criteria for planning and production control concepts
The comparison of the PPC concepts applicable at MTO organizations and selecting the best fitted concepts, is done in a previous study by Stevenson et al. (2005). The framework that resulted from this research is presented in figure 3.3. The explanation of the framework and the discussion about the performed research of Stevenson et al. (2005) are described below.
Figure 3.3 Framework for selecting PPC concept alternatives (Stevenson et al., 2005)
Lets start with the definitions of the terms in the framework:
GJS General Job Shop
“providing for a multi-directional routing, but with a dominant flow direction, relevant to both versatile manufacturing company (VMC) groups and repeat business customizer (RBC) groups” GFS General Flow
Shop
“work travels in one direction but jobs are allowed to visit a subset of work centres, permitting limited customization, only relevant to repeat business customizer (RBC) groups”
RBC Repeat Business Customizer
“provides customized products on a continues basis over the length of a contract. Products are customized but may be produced more than once, resulting in a small degree of predictability”
VMC
Versatile Manufacturing Company
“market is more complex, each order is competed for
28 A RBC suggests that the same products are produced continuously (with adjustments to specifications of the products allowed) for the length of a contract. When a contract is involved, it is known beforehand how many products from a certain type have to be produced. This means it is quit predictable which products are going to be produced for a period of time. The production process could even be re-organized to a product oriented organization where other PPC concepts like Kanban (see appendix 4) are fitted (like Meijer does for some customers with contracted repeat orders). These concepts are not considered in the MTO framework of Stevenson et al. (2005).
Furthermore, the difference between a general job shop (GJS) and a general flow shop (GFS) is not explained clearly. Based on the definitions presented, it is questionable why certain PPC concepts are applicable at a GJS and not at a GFS. A PPC concept like POLCA is designed especially for multi-directional routing, but is not appointed to the GJS in the matrix. In fact, any PPC concept that is not product specific could be applicable in a GJS as defined in that research. Because of these critical notes, the PPC concepts applicable at Meijer are not blindly copied from the framework of Stevenson et al (2005).
Stevenson et al. (2005) suggest seven needs in a planning process. These needs are: 1. Aggregate control 2. Workload balancing 3. Workload control 4. Material control 5. Order releasing 6. Detailed planning 7. Shop floor control
The PPCs are not specifically evaluated on these characteristics in that research, although the characteristics are the base for evaluation.
29 groups are identified. The second constraint is that if the average total processing time of jobs is highly variable, ConWIP would be less applicable. Therefore the convergent flow control and the processing time variability are included in the theoretical framework comprised in paragraph 3.3.
3.3 Theoretical framework
The analysis of the production process characteristics from Meijer and the analysis and benchmark of PPC systems result in a theoretical framework. This framework is the base for further diagnosis in Section 4. The theoretical framework is presented in figure 3.4.
Figure 3.4 Theoretical framework (based on: Bertrand et al., 1998; Stevenson et al., 2005; Verweij, 2008)
The left hand side of the framework is the summarized planning framework from Bertrand et al. (1998). It is coupled to the seven identified needs in a planning process by Stevenson et al (2005). The evaluation of the planning and production control concepts in the matrix is partly done by Verweij (2008; POLCA, WLC, ConWIP and TOC). The evaluation of the ERP-system on the seven needs is done by the author, based on Muscatello et al. (2003). Also the TOC evaluation is extended by the author based on Goldratt (1990).
30 problematic at Meijer. If this is the case, the PPC should be able to control the convergent flows. These aspects are diagnosed in section 4.
The design of a new control system is only half the work. Implementing it into an organization is a challenging job as well, that is why theory on implementing a re-design is described in the paragraph 3.4.
3.4 Implementation of a re-designed system
According to Checkland and Poulter (2006) every individual performs its tasks as well as possible according to their own worldview. This theory focuses on the aspect of how individuals see the organization they work in. Because of different worldviews within an organization, sub-optima and problematic situations can arise. In order to improve the problematic situation, the worldviews of the different individuals involved should be considered and altered if necessary.
Sometimes companies need to change to survive. It is not easy to implement a change in an organization. Often actors resist organizational change, even when it is occurring for a good reason. Most theories of organizational change originated from the work of social psychologist Kurt Lewin (Kreitner et al., 2002). He introduced the unfreezing-change-refreezing method for organizational change.
People interact in situations, and make judgements about what is good and bad according to their own criteria. Over time, these criteria lead to a relatively stable outlook through which people perceive the world. Checkland and Poulter (2006) call this outlook a worldview. It is the particular way an individual perceives the situation. The criteria used to make judgements about a situation are formed over time through what people experience and learn. For example, an engineer sees a problematic situation different than an employee from the sales department. The worldviews of these two employees are different because of their different experiences and educations. In this report the method of organizational change from Lewin (Kreitner et al. 2002) is used. This method consists of three stages which are discussed later in this paragraph. First the list of assumptions that underlie this method are described.
1. The change process involves learning something new, as well as stopping some current activities or routines
31 4. Resistance to change is found even if the goals of change is highly desirable
5. Effective change requires new types of behaviour, attitudes and practices
Now the underlying assumptions to the unfreeze-change-refreeze method are clear, the three stages can be described. The focus of the unfreezing stage is to create a motivation for change. A compelling reason for changing is needed to create urgency for an organizational change. Also, the management has to begin to disconfirm the usefulness of employee’s current activities and routines. The changing phase involves learning, providing employees with new information so they start looking at the current situation differently. The purpose is to help employees learn new concepts or points of view. During the refreezing stage the change is stabilized by helping employees to integrate the changed activities into new routines.
To implement a change in an organization Tidd et al. (2005) suggest three steps similar to the Diagnose Design Change (Evaluate) model which de Leeuw (2003) describes. Tidd et al. emphasize more on the Change part of the process of implementing change. This phase (which Tidd et al. call Implement) consists of four parts. These parts are Acquire, Execute, Launch and Sustain. The new element introduced here is the “Sustain” part. In order to sustain the change, new routines need to be learned. A routine is an organizational behaviour that is highly patterned, is learned, derived in part from tacit knowledge and with specific goals, and is repetitious (Tidd et al., 2005).
32 3.5 Conclusion
In this section Meijer is classified as a job shop performing in a Make to order environment. Queuing theory is a powerful tool to analyse the performance in these kind of organizations. Earlier research on applicability of different planning and production control systems in job shops was performed by Stevenson et al. (2005). In paragraph 3.2 this research was critically reviewed in order to form a framework to benchmark Meijer’s current PPC with other PPC’s applicable in the MTO environment
From this critical review can be concluded that the current planning and production control system in use at Meijer has to be analysed further on the ability to control the convergent production flows. If the current PPC is not able to control these flows, TOC might be an alternative. Furthermore, the variability of the processing times of the different operations and the different production orders need to be analysed in order to give recommendations on improving the current planning system.
33
4 DIAGNOSE
In order to improve the current situation, the current processes at Meijer are diagnosed. In this section the diagnostic phase is described. From the problem statement and the theoretical framework several aspects for further diagnosis were identified. Aspects that are subjected in the diagnostic phase are, the production process’s performance with the shop floor control method currently in use at Meijer and the planning process’s performance.
4.1 Production process performance
To measure the performance of the production process the following performance dimensions are taken into account: Quality, Flexibility, Work In Progress level, Speed and Reliability (Hopp and Spearman, 2000; Slack, 2007). The production process itself is analysed, together with the shop floor control system. From the theoretical framework presented in paragraph 3.3 two aspects for further analysis were identified. These aspects are whether the processing times are highly variable, and if so, this variability can be reduced and whether the current PPC concept in use at Meijer is able to cope with the convergent flows in the production process. Paragraph 4.1 gives answer to these questions. Because attention is needed at the control of convergent flows, the performance of the half fabricate production is analysed separately in paragraph 4.1.5.
4.1.1 Quality
Quality of the production process is described by the percentage of the products which is produced according to customers’ specifications, the first time right. A prerequisite is that the drawings obtained by the production employees are clear and that the machines are set-up and operated on properly. Every product has a quality check at the end by the quality service.
It is often not checked if the number of products are right and if the previous operation was performed right. At the assembly, several flows come together. If some parts are missing, it shows clearly at this department because compositions can not be completed. Also production failures from previous departments are discovered at assembly, when parts do not fit into a composition.
34 drawing is not clear or is wrong. At Meijer more emphasis is put on this lately, to make sure proper drawings are send to the shop floor. What still does happen sometimes and where still no improvement actions are planned for, is that too little of a product is produced. At welding, dispatching and at the quality check at the end, this shows, not sooner in the production process.
4.1.2 Flexibility
The flexibility of the production process is described by the variety of products which Meijer is able to produce, together with the ability to cope with the variety in order quantities. The ability to produce a variety of different products is present at Meijer. It is in fact the core business to produce a high variety of products. If Meijer is not able to perform a certain operation required in a product, this operation is outsourced. This means most products which a customer orders, Meijer is able to offer.
A lot of workers are contracted via a temporarily employment agency. Because of this, when less customer orders are received, Meijer can decrease the number of workers on the shop floor quite easily. The flexibility in coping with the variety in order quantities is high.
At assembly there is a problem which involves the skills of the workers at the department. Not all workers can produce all products. Some products require a welder with some years of experience and training to be produced. If a lot of parts are send to the assembly which require a skilled worker to assemble it into a composition, the department does not have enough capacity and products are delayed. At assembly the flexibility in being able to produce a high variety of products at all times is lower than that of the rest of the production departments.
4.1.3 Work In Progress
