VARIABILITY AT VARIASS ‘Various

(1)

VARIABILITY AT VARIASS

‘Various sources of variability ’

Author: Tom Beugelink s1516582 Company: Variass Electronics Cereslaan 4 9641 MJ Veendam Variass supervisor A. de Vries Faculty supervisor Dr. J.A.C Bokhorst Co-assessor

(2)

Variability

Various sources of variability

Author: Tom Beugelink s1516582 Company: Variass Electronics Cereslaan 4 9641 MJ Veendam Variass supervisor A. de Vries Faculty supervisor Dr. J.A.C Bokhorst Co-assessor

Prof. dr. ir. J.Slomp

(3)

T.K. Beugelink Preface

PREFACE

If you asked me ten years ago where I should be, I could have said anything except finishing University. I think even my parents or my tutors should have said the same. Too many reports and tests pointed that “Tom” loved school, but hated learning; loving the breaks, the chat and meeting people, hated the learning, homework and sitting still, something I still have not learned.

But with the years my commitment to school, the homework, the learning grew and even time in the weekend was spend at learning. By the years I realized that with some effort and discipline everything was possible, so after finishing my bachelor Human Technology at the Hanze hogeschool Groningen I registered myself at the rijksuniversity Groningen next to that the student life was I life for the time being I enjoyed.

But on al good things comes an end, so also my life as a student and I’m glad also. After a stirring four and a half year period at the university, with its ups and downs I‘m finally finishing my last ‘product’ as student, in form of a paper lying in front of you.

This paper in front of you describes a research at Variass Electronics B.V at Veendam, which I’m grateful for doing my research at. In the first place I will thank my supervisor Mr. A. de Vries for giving me this research and for the support during the research. Next I want to thank my supervisor of the Rug Mr. J.A.C. Bokhorst for helping me formulating a good research and sharing his expertise during the research but also my secondary supervisor Mr. J. Slomp for giving me useful comments and advise for also structuring my research. Also I want to thank my sister in law A. Schooneman for reviewing my paper. Finally I would thank my girlfriend D. Schooneman and my parents for supporting me during the ‘down’ periods, but also my parents for the financial support.

(4)

T.K. Beugelink 1 Abstract ABSTRACT

This report describes a research to reduce lead time and costs at the Production department of Variass Electronics. Variass is a company which assembles Printed Circuit Boards into subassemblies or into complete products. The problems faced by the management of Variass was that the production process of Variass produces with high costs and creates a lot of waste. Furthermore, products have long lead times.

Following from the exploration phase of this research, high levels of work in process, inventory and indirect hours were indicated as possible factors causing the high lead time and costs. Most of the problems are created due to the organization of the planning and control of the Production

department, leading to the following research question;

Which aspects of planning and control causes the increase in work in process, defects and indirect hours and how can these aspects be changed so eventually lead time and cost will be reduced?

During the next phase of the research, those aspects causing the increase in work in process, defects and indirect hours and how these aspects can be changed are investigated. For the investigation the PAC model of Bertrand et al. is used. From this model three elements are selected for a more in depth analysis, namely capacity, order release and shop floor control.

Capacity- The organization of the Planning department is good and according to the PAC model of

Bertrand et al. it doesn’t lack organization. Planning checks the occupation planning and the material coordination if it is possible to meet total lead time. However, planning does not check all capacity groups. From the planning it could look like there is sufficient capacity at the capacity groups, but during the execution of the orders it may occur there isn’t. This is because not al the processes are recognized and therefore no insight in possible bottlenecks can be given. Therefore, unrealistic delivery dates to customers are given.

Order release –Eventhough planning performs a RCCP while planning the orders, in releasing the orders the planning does not conform to this so precisely. This is mainly caused by the margin of the planned orders. Planned orders are scheduled on a working schedule with a start and end date. Because the production process of Variass has many uncertainties, these end dates oftentimes contain a few days slack. Although planning builds in this slack between production due date and customer due date, it does not have the right effect. This is because the foremen of the teams are aware which orders are on due date and which not, which results in a planning without pressure.

Shop floor control -The Planning department uses the infinite loading principle at the Production department. The capacity and occupation are not taken into account. Due to the irregular workload (by infinite loading), the 100% utilization, the fluctuation of capacity (employees) and the aberrant due dates, the working schedules are soon outdated and only used as a guide. Therefore, foremen spend more time in distributing work among employees. Furthermore WIP and Lead time are also increased due to batch production. Because the transfer batch size equals the production batch size, WIP is high and lead times are long. These large transfer batch sizes also increase lead time and costs by

defects. When a serial fault is made, it results in a very large number of defective products, because the error is detected after a sub process has finished the lot.

(5)

T.K. Beugelink 2 Abstract

Because of this, WIP is high, lead times are long, and the number of defects is high. To reduce WIP and lead times, the Production department should reduce the transfer batch size. However, not every process at the Production department is suitable for the reduction of the transfer batch size due to the setup time regarding the production time. In this research, it is proposed that WP has to split up the orders in batch and hereby they should place the components in numbered or colored bins. As a result that setting up the paternoster only involves placing the bins at the designated place.

The process Assembly and Touch-up have to form small teams with balanced processes enabling a higher transfer rate so buffers between the processes is unnecessary.

Test has to classify which tests can be performed by employees of the teams and should perform QA checks on boards which have to tested.

At last QA has to be placed directly after Touch-up enabling that orders are finished as soon as possible and detection of errors is carried out sooner. Further more capacity at QA should be increased when QA is overloaded.

(6)

T.K. Beugelink 3 Table of contents TABLE OF CONTENTS INTRODUCTION... 5 1 VARIASS ELECTRONICS... 6 1.1 THE COMPANY... 6 1.2 THE ORGANIZATION... 6 1.3 PRODUCTS... 7 1.4 END PRODUCTS... 7 1.5 THE PRODUCTION PROCESS... 8 1.6 RÉSUMÉ... 9 2 PROBLEM DESCRIPTION... 10 2.1 PROBLEM EXPLORATION... 10 2.2 RÉSUMÉ... 11 3 RESEARCH APPROACH... 12 3.1 PROBLEM DEFINITION... 12 3.2 RESEARCH GOAL... 12 3.3 CONCEPTUAL MODEL... 13 3.4 RESEARCH QUESTIONS... 14 3.5 RESEARCH DEMARCATION... 14

3.6 APPLIED DATA SOURCES... 14

3.7 RÉSUMÉ... 15

4 UNDERSTANDING THE PROCESS... 16

4.1 THE PRODUCTION PROCESS... 16

4.2 ROUTING OF ORDERS... 18

4.3 LEAD TIME AND COSTS... 19

5 PLANNING AND CONTROL... 21

5.1 PLANNING... 21

5.2 CAPACITY... 23

5.3 ORDER RELEASE... 25

5.4 SHOP FLOOR PLANNING AND CONTROL... 27

5.5 RÉSUMÉ... 37

6 REDESIGN... 40

6.1 SHOP FLOOR CONTROL... 40

6.2 LOT SIZES... 40

6.3 REDUCING LOT SIZES AT VARIASS... 41

7 RECOMMENDATIONS... 46

7.1 ORDER ENTRY AND MID-TERM CAPACITY PLANNING... 47

7.2 ORDER RELEASE... 47

7.3 OPERATIONS SEQUENCING... 48

REFERENCES... 49

(7)

T.K. Beugelink 4 Glossary

GLOSSARY

EA End Assembly

FTE Full-time Equivalent; is used to measure employee involvement, assuming 1.0 FTE is equivalent to a full-time worker.

Touch-up This means after the components are soldered on the PCB by the machine the PCBs are checked and repaired if necessary and additional components are placed.

PCB Printed Circuit Boards

QA Quality Assurance

RoHS Restriction of Hazardousness Substances

SMD Surface Mounting Device

SMT Surface Mounting Technology

THT Through Hole Technology

TPS Toyota Production System

TQM Total Quality Management

VSM Value Stream Map

WIP Work in Process

WLC Work Load control

(8)

T.K. Beugelink 5 Introduction

INTRODUCTION

The master thesis project described in this research is carried out at Variass Electronics B.V. at Veendam. Variass Electronics is a company which assembles Printed Circuit Boards into subassemblies or into complete products.

Because the assembly of PCB is a time-consuming job and asks much handling time, many companies moved their production to Asia, where the costs per employee are much lower. To compete with the competition of companies from Asia, Variass needs to lower its costs and decrease its lead time to guarantee the continuity of Variass.

In the past the management of Variass tried to increase the efficiency of the Production department by introducing Lean manufacturing. Lean manufacturing should lead to a better performance of the Production department. A way to improve the performance is to decrease the costs in forms of waste and reduce lead time of the production process. This should guarantee the continuity of Variass. Together with an external consultancy group Variass performed a research which contained a Value Stream Map (VSM) and the formulation of a lean strategy. From this VSM and the lean strategy some projects were distracted.

The first project they started did not produce the desired output. Some of the improvements they made during this project were successful and others not. After the lean engineer left the company the

management team decided to stop the lean projects due to lack of time. A year later the management team still wants to implement lean manufacturing, but the desired output and the execution of these projects has many uncertainties. From this perspective the research is carried out. The management team of Variass wants to know how they can improve the performance of their production system desirably by implementing some lean principles, but to what extent can those principles be implemented?

In this research the assembly of PCBs at the Production department is investigated. First a short description of Variass the company shall be given. What the core business is of Variass, the structure, which products they make and how the production process is organized. In the following chapter, chapter 2, a more comprehensive analysis of the problem situation will be given. On basis of this problem description a research proposal will be developed, chapter 3.

(9)

Various Sources Of Variability

T.K. Beugelink 6 Variass Electronics

1 VARIASS ELECTRONICS

Before explaining the research conducted in this research, an introduction of Variass the company, and the products they produce will be given. Because the research is carried out at the Production department a description of this department will also be given.

1.1 The Company

The history of the present Variass company goes back to 1989. The company started by the

management take-over of a family concern which had been involved with various types of electronics assembly works since the seventies called Wortmann Elektronica.

By the increasing need for total order placement the core business shifted from assembly of boards to a broad spectrum of services. Now-a-days Variass can be described as project manager and system supplier which offer a broad variety of products and services; like project management, development and engineering, life cycle management, purchasing, launch, etc. Within these services one main activity can be distinguished, the assembly of Printed Circuit Boards (PCB). The research conducted in this research is held at the Assembly department of Variass, so only this department will be described in this report.

1.2 The organization

The research investigated in this report is held at the Production department of Variass from which the assembly department is a part. The structure of the department is given in the Organization Chart see Figure 1 Hierarchical structure. The production manager controls the head of the department and the foremen of the teams the head of SMT.

Figure 1 Hierarchical structure

The Production Manager is responsible for Assembly, Test department and SMT.

The foreman of Assembly is responsible for the Work Preparation (WP) and team leaders. To increase responsibility by employees the assembly department is divided into three teams, Team A, Team C and Team D. Every team has its own foreman who is responsible for processing the orders before due date. The three teams are the same, possesses the same equipment and can make the same

(10)

targets, time, quality, etc. Within the teams employees are responsible for the different tasks; in paragraph 1.4 this will be further explained.

1.3 Products

Within the assembly process of the assembly of PCBs two main processes can be distinguished which is part of the primary process. These two main processes are assembly by Through Hole Technology (THT) (see Figure 2 THT) and the assembly by Surface Mounting Technology (SMT) (see Figure 3 SMT ) process. The difference between these two processes can best explained by the two different techniques which are used. The first technique THT is a design standard where components are mounted by inserting pins into pre-drilled holes on the PCB, at Variass this is done by hand. Placing THT components by hand will from now on called Assembly.

The second technology called SMT is a technology where components are directly soldered on the PCB by Surface Mounted Devices (SMD). The SMD is a machine which automatically picks and place components on the PCB. SMT is mostly used for complex PCBs with many components. At the SMT department of Variass Electronics PCBs are mounted and optically tested. This testing is done by an Automatic Optical Inspection device.

Figure 2 THT Figure 3 SMT

1.4 End products

The PCBs Variass assemble are applied in many products, some examples of where the PCBs are used for are given below.

Lexguard

The Lexguard is a blind spot detection system. The active blind spot detection system is a plastic strip attached to either the right side of the trailer, the bicycle-guard or directly to the body front and right side. The strip “feels and sees” if anyone is in the blind spot during a right turn and gives a visual and audible warning signal to the driver in the cabin1.

Hearing glasses

The legs of the hearing glasses contain four miniscule microphones which amplify the sound coming from the front and reduce that coming from other directions. This enables the user to listen without strain in a noisy

environment, and hear better from the direction in which the user is looking. In this project Variass focused on the producibility of the printed circuit board in the legs of the spectacles and the development of the battery holder and charger1_.

1_{http://www.variass.com}

(11)

Communication Systems

Some other complete products Variass produces are the communications systems for defense. Those are a sort of TomTom devices for tanks. Variass also makes USB devices which can stand explosions1.

Figure 4 Lex Figure 5 Guauard communication Figure 6 Hearing glasses 1.5 The production process

Within Variass there are two departments which assemble PCBs. The SMT department, which is using machines for assembling PCBs (see paragraph 1.3) and the Assembly department which assemble the boards by hand. About 90% of the boards produced by the SMT department also needs components which cannot be assembled by machines, therefore these boards are also passing the Assembly department. That means that PCBs as well as the PCBs from SMT are assembled at the Assembly bald. At the Assembly department the following process can be distinguished.

1. Order picking 2. Work Preparation 3. Assembly THT 4. Touch-up 5. End Assembly 6. Test 7. Quality Assurance

A more comprehensive description of the production process shall be given in the Diagnoses (see page 21.

To a large extent the production process of Variass can be classified as a Make-To-Order

environment hence the production of PCBs is customer specific. Furthermore the production process of Variass can be classified as a process layout. According to Riezebos, J., Donk van P. and

Bokhorst, J. (2008) ‘in a process layout workplaces with similar activities are placed together, the

machines with the same function are placed together.’

Currently orders are produced in batches. The size of a batch equals the demand of the customer. If a customer wants a batch of 2000, the batch size is 2000. At Order picking, Work Preparation and Quality Assurance batches are processed one by one. At the other processes Assembly, Touch-up/ and Test batches are processed parallel. The parallel processing means that processes can handle more than one batch at a time but the employees handle one.

Most of the times an assembled PCB is the end product, but sometimes customers want their products assembled in end products like the one described in the previous paragraph. Because the testing of the PCB’s is an expensive process, not all the customers want their boards checked. Approximately 25% of the assembled PCBs has to be tested. In the figure below the routing of the orders is represented.

(12)

Figure 7 Routing products

Variass has many small customers. All these small customers cause a high variation within the production processes. For example; on a normal day 20 different products are produced at the same time, all with different routings (see Figure 7 Routing products) and different operation times at every process. Due to this high variation in products, routing and processing time a high degree of control is needed by the management. Most of the time the foremen of the teams (explained in paragraph 1.2) iare occupied with scheduling orders day-by-day. This also was represented during a presentation2 of the manager-director. During this presentation he mentioned that the Production department faces high indirect costs, caused by indirect hours created by control of the Production department. Due to the high variation in products, operations time and routing is not easy to plan.

1.6 Résumé

Variass Electronics is a company which particularly assembles PCBs. During the past years they shifted their core business from PCB Assembly, to a more total package; as project manager and system supplier which offer a broad variety of products and services; like project management, development and engineering, life cycle management, purchasing, launch, etc. Due to this competence, Variass does not have their own products and are dependent on other companies. Because of this Variass has a production process with a high variety of products with various demands. Due to this high variety of products and demands the production process needs a high degree of control.

In this chapter some problems regarding the Production department are already described. In the next chapter a more detailed analysis of the problem situation will be given.

2

(13)

T.K. Beugelink 10 Problem description

2 PROBLEM DESCRIPTION

This chapter will give a more comprehensive description of the problem situation. First the motivation of the research will be described, why Variass has started this research and what the initial problem is. Second an exploration for functional problems is carried out through observations and interviews.

2.1 Problem exploration

During this phase, interviews are held with different employees of the Assembly department.

Interviews are held with the production manager, process manager, and manager of Test, foreman of the teams, employees of WP and Assembly employees. During these interviews several questions are asked regarding the Production department. What are the problems they are facing during at the production process? Which things can be done better? What possible improvements are?, etc. During these interviews several problems were indicated.

Due to the current production method the workload is not evenly spread. As well as in the beginning and at the end of the week the workload is very high which calls for a high level of control.

Every process at the Assembly department has its own schedule of orders they have to make, test or check, etc.

At the beginning of each week the foremen and the employees of Planning are having a meeting to schedule the orders for the coming week. In practice this schedule is caught up with reality very quickly, as result that heads of the different departments are chasing orders to meet delivery time. Next to that, Planning regularly schedules orders with a high priority. These priority orders causes a high workload, but also a high inventory, because other products are put on hold in the production process.

Also Test and QA detect faults which can be found by visual inspection, this means that employees do not check each others work while they have to.

Figure 8 + 9 Subassemblies and end products on shop floor

Next to the interviews also observations are held at the Assembly department. Entering the Assembly department one directly notices the high level of Work in process and the carts with subassemblies. At the Assembly process the batch size is the size of the complete order. The production is organized in a certain manner that production is by batch i.e. a work center finishes the complete batch before it is moved to the next work center. A process only begins with processing when they received the complete batch from the previous process. This kind of production method causes several problems. Next to the costs of the level of inventory (see Figure 8 + 9 Subassemblies and end products on shop floor) ,defects during production are detected by Test while assembly finished the complete order, this can result in that the complete order has to be reworked.

(14)

T.K. Beugelink 11 Problem description

Figure 10,11 and 12 WIP

Next, the management of the Production department thinks that the current layout of the assembly line is insufficient, especially regarding Test. Until the process Touch-up it is clear which products are in production, but at Test it is not clear which products are being tested, which are ready and which products are waiting. During interviews with the management signals about starved processes where reported. Hereby an example about End Assembly was given. End Assembly needs products from Test, but Test is testing other products. Therefore the End Assembly gets starved in other words they cannot produce. From management perspective Test has a disordered layout, because of this it is not clear which batch is ready and which is not. However during the research no solid evidence of the signaled ‘starvation’ could be gathered.

The need for a change is also confirmed by the VSM (see Appendix I Value Stream Map). This VSM is a rough estimate of the average lead time, work in process (WIP) and production time per single product. When all production times will be summed up, from WP to QA (the blue outlined box) a lead time of 16.7 days is the result to a production time of 30 minutes. As a result operations time is approximately 0.6 percent of the lead time. This is also visible in the high work in process (WIP) 1834 hours. The level of inventory, the red triangle in the VSM is also high with 324 products as minimum and 984 single products as a maximum.

According to Nicholas (1998)large batch sizes impede the ability to make necessary changes to production schedules and priorities. Large batch sizes contribute to high level of WIP and long lead times. This example is also illustrated by Little’s Law this law implies (Hopp, W.J., Spearman., M.L. 2001):

[

Lead

Time

] [

Throughput

rate

]

WIP

=

_

*

_

Equation 1 Or

rate

Throughput

WIP

Time

Lead

_

=

Equation 2

Inventory exists in many kinds like inventory at the start of the process, scratch, but also inventory in end products, inventory in subassemblies and in WIP. Looking at equation 2 these high levels of inventory also cause the long lead time.

2.2 Résumé

(15)

T.K. Beugelink 12 Research Approach

3 RESEARCH APPROACH

In this chapter the research approach will be described. From the motivation of the research and the problems found in the previous chapter a research definition and goal will be formulated. On the basis of this problem definition and founded problems a conceptual model will be developed. From this conceptual model a research question and some sub questions will be derived. At last a short description, on how the sub questions will be answered, shall be given.

3.1 Problem definition

For formulating a clear problem description the first step of the DDC model of de Leeuw (2001) will be used. The DDC model consists of three phases, the Diagnose phase, the Design phase and the change Phase. According to Prins (2008) the diagnose phase itself also consist of three steps; the problem defining phase, research design and problem clarification. In this paragraph the problem defining phase is carried out.

According to de Leeuw (2001), only reality problems has to be taken into account and not the goal/ perception problems. The problems found in the previous chapter are summarized below.

• High cost due to indirect hours

• High costs through high level of inventory

• Defects are detected too late

• High lead time due to batch production where the batch size equals the customer order From these problems the following problem statement can be formulated:

Currently the Production department at Variass produces with high costs and a long lead time; during the manufacturing process a lot of waste is created. This waste consists of inventory, defects and indirect hours.

Other forms of Waste are muri and mura, muri focuses on waste created by the management

(Womack, J., Jones, D.T., and Roos, D., 1991), because of poor organization like lifting heavy objects, moving objects but also working significant faster than usual. Mura focuses upon improvement of the “flow” thereby eliminating unevenness.

These two types of ‘waste’ verbalize the other problem viz. the high costs due to indirect hours.

3.2 Research Goal

From the problem definition above the following research goal can be formulated:

To analyze how planning and control causes a high level of work in process, defects and indirect hours resulting in high lead times and costs, and to propose a redesign based on this analysis.

This leads to the following research question?

(16)

3.3 Conceptual Model

The conceptual model displayed below can best be explained as follows. In the conceptual model there are two key variables, costs and lead time. These two key variables are derived from the problem description in the previous paragraph. These two key variables will be explained one by one.

Fr on t en d E n gine B a ck e nd L ong te rm (mo nth ly > ) Me di um t erm (we ekly) Sh or t t er m (d aily <)

Figure 13 Conceptual model

The goal of this research is to decrease lead time and costs. In the conceptual model above some possible key factors, in the researchers his point of view, causing the long lead time and high costs are presented.

Lead time- The high lead time is possibly caused by high levels of WIP and the time it is costing to

repairing defects, which on their turn are caused by batch production.

Currently orders are processed in batches, this means that a process only begins processing when the previous process finished the complete batch. Between the processes subassemblies are stored in a buffer. When the batch is completed the batch is moved to the next process. These buffers cause a high work in process. Little’s Law implies (Hopp et al. 2001)that a higher level of work in process increases the lead time as explained in paragraph 2.1. Next to the direct effect on lead time, WIP also has an indirect effect viz. through defects. In the inventory between processes defects stays

unnoticed, repairing those defects costs time.

Currently orders are sequenced in many different ways. Because of this orders are shifted as a result some orders are occasionally long in process, resulting in high levels of WIP.

The detection of defects is too late, because a defect is only detected when the next process starts the batch. As a result that the complete batch needs to be repaired which results in a higher lead time due to repairing.

Costs - Due to interest over inventory the inventory causes costs. This problem is probably caused by

(17)

costs in indirect hours. Next to the costs over inventory the repairing of the defects mentioned above costs also money.

3.4 Research questions

From the conceptual model the following sub questions can be formulated;

1. To what extent does the Planning department reckon the available capacity during accepting orders? By what time fence is capacity determined? How does the planning deal with the variation of the available capacity? And how does this affect order release?

2. How is the organization of releasing orders into the Production department? When releasing orders into the Production department do they take the workload regarding the different processes into account? How does the order release affect shop floor control?

3. What kind of planning and control activity are carried out on the shop floor level and what is the impact on inventory, defects and indirect hours?

a. How does the allocation of capacity to the orders affect inventory, defects and indirect hours?

b. How does the sequencing of orders affect inventory, defects and indirect hours?

c. How does the batch production of orders affect inventory, defects and indirect hours?

4. Which choices does Variass has to make in order to decrease lead time and to reduce costs?

3.5 Research demarcation

The capacity will first be examined on the highest aggregation level, the front end. It will give insight how Planning of order proceeds and to what extend capacity is taken into account. Next, order release will be investigated. It will examine the planning on a more detailed level, engine. It will give insight on how actual orders are released into the firm and if workload among the different processes is taken into account (workload control).

The last element shop floor scheduling will be analyzed by allocation of capacity, batch production, sequencing of orders and, it will be examined how, on the lowest level, the back end; orders are produced, if shop floor control makes use of sequencing of orders and how the allocation of capacity to the orders proceeds.

Currently the Production department of Variass is transforming to lean manufacturing. During the analysis of the planning and the elements of the conceptual model the applicability to lean manufacturing will be examined. Also the (re-) design will classify certain planning and control characteristics of the current production process and how a perfect lean manufacturing planning and control process should look like.

3.6 Applied data sources

For this research, multiple data sources will be used to provide contradictory and confirming interpretations. Kempen, P. and Keizer, J. (2000) distinguishes five available data sources; participation, interviews, observations, measurement and documents. The last one ‘documents’ is classified as secondary data. Data that is developed for some purpose other than helping to solve the

research question in hand.Coghlan, D. and Brannick, T. (2005) These data will be evaluated during

the action research on basis of their relevance, their availability and accuracy.

(18)

disadvantages, theory will be used, but also actual problems found at Variass by observation will be used.

3.7 Résumé

The motivation of this research is that Variass wants to decrease their costs and lead time. During the problem exploration phase the formulated problems by management were researched. From this exploration phase became clear that Variass faces high cost and long lead time regarding their process time. Some possible causes for these high costs and long lead time are possible high inventory, defects and indirect hours. From this exploration, high costs and long lead time are caused by planning. From this the following research goal can be formulated;

To analyze how planning and control causes a high level of work in process, defects and indirect hours resulting in high lead times and costs, and to propose a redesign based on this analysis which

result in the following research question;

Which aspects of planning and control causes the increase in work in process, defects and indirect hours and how can these aspects be changed so eventually lead time and cost will be reduced?

The work in process, defects and indirect hours are possibly caused by the planning (-system). In the following paragraphs planning will be researched more thoroughly. Because planning is very

extensive, only a few elements will be analyzed, viz. capacity, order release and shop floor control. From this analysis requirements for a (re)-design are formulated. These requirements will finally be used as input for the (re)-design. Before analyzing the elements one by one a better understanding of the production process is required. In the following chapter the production process of Variass will be described and also the routing of an order.

(19)

T.K. Beugelink 16 Understanding the process

4 Understanding the process

Before the diagnoses will be carried out a better understanding of the production process and the routing of the orders is needed. In this chapter, the production process and the routing of the orders are described. First the physical flow of the production process will be given; second the information flow will be given. Also the definitions lead time and cost will be explained more thoroughly.

4.1 The production process

The production process of Variass can be described as a job shop. A job shop produces a wide variety and small quantity of specialized products. Product are generally customized (Evans. J., 1993) Represented in the organization chart different departments can be distinguished. The operations reflected in Figure 7 Routing products (page 9) fall under the different groups within Variass. Because the Production department uses different sorts of definitions a classification by capacity groups will be used Koppenrade (2005). This classification is the same classification the Planning department uses. The capacity groups that can be distinguished are Warehouse, Work Preparation, Team’s A and C, Test and Quality Assurance. In Figure 14 Process lay out the layout of the Production department is depicted.

(20)

The structure of the production process of Variass can be characterized as functional layout / process layout (Riezebos et al., 2008). In a process layout similar types of operations are clustered into functional work areas or departments (Nicholas, 1998). Looking at the clustered functional work areas can be distinguished. The testing department (the yellow tables) is the last but one process of the production process. According to Riezebos et al. in a functional layout there is a coupling of

departments without taken the flow of products into account As far as the Production department of

Variass is concerned this is partly not true. The different departments like WP, Assembly and Touch up are placed in such a way that these are following a line. Due to this dominant product flow the Production department can be further characterized as a flow shop.

4.1.1 Warehouse

The warehouse is the first capacity group that can be distinguished. At the warehouse three different operations can be distinguished namely checking delivered goods, order picking and packaging of produced goods.

Checking delivered goods - Before goods are placed in the warehouse delivered components are checked by quality and quantity. For the checking of components a classification is used. Some components are checked 100% and some components are not. After the check, the components are placed on a trolley so they can be placed in the warehouse.

Order Picking - In the order picking process the start of the production order begins. At order picking

components are picked by a Bill of Material. At the BOM information of the quantity and location is given. Picked components are placed on a trolley and moved to Work Preparation.

Packing – At packing the produced orders are packed and made ready for shipment. 4.1.2 Work Preparation (WP)

At WP the order enters the production floor of Variass (see Appendix II Process layout). At WP three operations can be distinguished WP by machine, WP by hand and model construction

WP by machine and WP by hand -- the two operations are practically the same, only the one contains

the operation by hand and the other by machine. This separation is a result that some suppliers deliver components on tape and others do not. Components on tape can be placed in a machine which automatically cuts the components to the right length. Components which are not delivered on tape has to be cut by hand, one by one.

Model construction -- An extra task of WP is the construction of models which will guide the assembly

employees. Most of the products already has a model which is assembled; these models are stored in a scaffold behind Quality Assurance. Because customers has many product modifications WP often needs to construct a complete new model.

4.1.3 Team A and C

The next capacity group is the teams A and C. The teams A and C can be considered as two parallel cells containing the same kind of operations. These operations are: Assembly, Touch-up and End assembly. Assembly A&C and Touch-up A&C has a fixed position at the facility (Appendix II Process layout), but the allocation of the End Assembly is not fixed, due to the variation of products.

Assembly – The assembly of the PCBs is by hand using Through hole Technology (THT). This is a

(21)

End Assembly – Another operation which falls under the capacity group Team A and C is End

assembly. At end assembly assembled PCBs are assembled into complete products like the products described in the first chapter. End Assembly is not a standard operation of the routing of PCBs because approximately 28% of the orders needs end assembly.

With the formation of teams Variass tries to reduce lead time and faults by repetition of orders but also raise commitment of employees to meet delivery time.

4.1.4 Test

Test is the only capacity group which is not part of the standard routing of PCBs. The testing of PCBs is customer specific. PCBs are only tested if customers want to. At Test also four different operations can be distinguished; functional test, duration test, ICCT and repair.

Functional test – at functional test assembled PCBs are tested by line-up of customers. A line-up is a

test application specified and some times also supplied by customers. Because it is too expensive to check the complete PCBs only critical paths/ components are checked.

Duration test – another kind of test is the duration test which is not used that often, but some

customers do want a duration test. At a duration test a PCB or end products is tested for a longer period, like 24 hours, for its functioning

In Circuit Component Testing -- At ICCT testing not the functioning of the PCB is tested but the correct

positioning of the components is tested. Because the ICCT is a very expensive test it is not applicable for all the PCBs, due to this the ICCT test is also a test which is not used that often. Currently one customer uses the ICCT testing.

Repair – At repair defects like broken components or wrongly placed components are repaired. 4.1.5 Quality Assurance

The last capacity group is quality assurance (QA). This is also the last process of the Production department. At QA the boards undergo one final inspection, if components are placed appropriate. This checking is done randomly. Because at Test specific parts of the PCBs are checked it is possible for QA to find components which are placed incorrectly. In this report QA is considered as a capacity group but at the planning system the QA is not. By drawing up the planning, Planning does not take time for QA into account. Therefore QA follows the planning attentively. Orders with a high defect proportion are checked in front because the repairing of those orders take time. This all is the responsibility of the QA employee.

4.2 Routing of orders

In the previous paragraphs the ‘physical flow’ of orders was described; in the next chapter the

‘information flow’ of the orders will be described. For modeling this process Actor Activity Diagramming (AAD) will be used.

An order undergoes several steps before they are taken into production. The kind of steps is

independent on what kind of order it is. Two kinds of order routings can be distinguished viz. complete new order or repeating orders. By new order the preceding steps of an order is more extensive. (An overview of these steps can be found in Appendix III AAD Order routing)

(22)

order” has a duration of three weeks and is changed into a “production order”. When the scheduled production date starts, Planning checks if the needed materials are present. When they are, Planning gives the order to Order Picking which begins picking the order. After picking the order the order is moved to the Production department. The detailed steps of Production are described in paragraph 1.6. Production communicates with Engineering about problems during production, which on their turn communicates with customers about actions taken.

At the beginning of the week Planning and Production has a meeting whereby the production orders are planned. The goal of this meeting is that Planning can schedule orders with a priority.

After Production produced the order the order is moved to Expedition, which packs and prepares the order for delivery.

A repeating order follows the same routing only the preceding steps like labor and material calculation are already given by the SAP system. Next, the order is handled by Planning and not by Sales. Also the involvement of Engineering is less, because the product already has been produced before.

4.3 Lead time and Costs

In the conceptual model the two main variables: lead time and costs are formulated. Both the variables have many different definitions in theory but also at the company itself. To avoid misunderstanding both the variables will be explained below .

Lead time

Within companies and literature definitions like delivery time, lead time, process time, etc are

confusedly used and mostly delivery time is used as lead time. Because there is a lot ambiguity about these words these definitions will be given below and which definitions will be used.

In literature there are many definitions of lead time; lead time is the amount of time between placing an order and receiving a receipt of the delivered goods (Bertrand et al., 1998 and Blocher et al., 2007). According to Nicholas (1998) lead time “is the time spent at processing a product”. Hopp et al. (2001) defines lead time as the time of a given routing or line is the time allotted for production of a part that routing of line. To prevent misunderstanding of the definition lead time, it will be divided into customer lead time, purchasing lead time and inward lead time i.e. processing lead time.

Customer lead time is the same lead time as the lead time defined by Bertrand et al.(1998) and Blocher et al. (2007) viz.; the amount of time between placing an order and receiving a receipt of the delivered goods. The customer lead time is the sum of purchasing lead time and processing lead time. Purchasing lead time is a period between a decision to purchase an item and its receipt at the

warehouse. Processing lead time is the time required to manufacture an item, including order

preparation time, queue time, setup time, run time, move time, inspection time, and put-away time. For make-to-order products, it is the time taken from the release of an order to production and shipment. The customer lead time of Variass is the Purchasing lead time plus the processing lead time. Because both processes have many uncertainties planning takes a slack of five working days into account. From this the following formula can be formulated:

s

Mlt

Plt

T

=

+

Equation 3

T : Customer lead time Plt : Purchasing lead time Mlt : Processing lead time s : Slack (usually 5 days) Cost

(23)

(24)

T.K. Beugelink 21 Planning and Control

5 Planning and Control

This chapter describes the diagnose phase. During the diagnose phase the conceptual model will be examined with the intend to find those factors which causes the investigated problems with an empirical model as result. During the diagnose phase functional problems will be researched in order to find those instrumental systems attributes which by change causes the functional problems to disappear. Fr on t en d E n gine B a ck e nd L ong te rm (mo nth ly > ) Me di um t erm (we ekly) Sh or t t er m (d aily <)

Figure 15 Conceptual model

5.1 Planning

Variass produces in response of orders from other companies; therefore their production system can be classified as make-to-order. All products are customer specific, therefore components and bald PCBs can only be purchased when the nature of the order is complete. This MTO production system has some constrains regarding inventory, because components and bald PCBs can only be

purchased when the order is definite.

Make-to-order production systems typically produces a larger variety of products in smaller quantities than other production systems like make-to-stock. Because a MTO production system produces customer specific end items, the production process has a high uncertainty how much of companies resources will be required as the engineering is finally completed and exact requirements are determined. Although demand at Variass is irregular it is possible for Planning to calculate the actual required capacity ahead. For the procurement of components and bald PCBs Planning uses a delivery time of five weeks. For the production, Planning uses a process time of two weeks plus an extra week slack. Variass uses this slack because production and the delivery of components has many

uncertainties like delay of delivery, delivery of bad components, product modifications, defects during production, etc.

Currently orders arrive by different ways; some customers give a forecast of four months and some give a forecast of one year. Mostly these are customers with long-term contracts. For example; currently Variass has an order of sixty products over eleven years.

(25)

orders are changed to a production order. Together with Purchase, Planning determines the lead time of new orders.

The variables capacity, order release and shop floor control are elements of planning and control. This is also represented in Figure 15 Conceptual model.

According to Slack et al. (2001) planning is a formalization of what is intended to happen at some time in future and control is the process of coping with changes in this planning. This corresponds

respectively to capacity and shop floor control with order release in between.

As can be derived from the conceptual model planning and control is divided into three key variables viz.; Capacity, Order release and Shop floor control.

By examining the three key variables (Figure 16 Key variables Conceptual model) an answer will be given to the research questions.

Figure 16 Key variables Conceptual model

Bertrand et al. (1998) developed a product activity control system. The three variables assigned above are also represented in this model. By representing the model the relation between the variables and the interaction with other elements become clear. In Appendix IV Planning and control framework the complete framework is represented. In the figure below the production and activity (PAC) framework is also given but adapted to the PAC of Variass.

Ca pacity Or der r elease Shop f loor control Long term (mon thly >) Medium term (w eekly) Short term (d ail y <)

(26)

Most of the elements of the conceptual model are applicable or corresponding with the planning and control activities of Bertrand et al. In the adapted PAC see Figure 17 Planning and control frameworksome alterations are made. The model represented above and in Appendix IV Planning and control framework

also gives a good example how a framework of production and control should look like. By reflecting the execution at Variass of these activities, regarding capacity, order release and shop floor with the model of Bertrand et al. (1998) imperfections should be revealed.

In the following paragraph the elements capacity, order release and shop floor control are examined. This will be done by answering the related research questions but also the specific execution at Variass regarding the planning and control framework of Bertrand et al.

5.2 Capacity

The first variable gives insight in how Planning deals with the capacity of the Production department. In this paragraph an answer to the following questions will be given:

1. To what extend does the Planning department reckon the available capacity during accepting orders? By what time fence is capacity determined? How are orders planned? How does the capacity planning affect order release?

Accepting an order

Before accepting the order Planning consults the Rough-Cut Capacity Planning (RCCP) if capacity is available. According to Evans, J. (1990) a RCCP involves the analysis of the MPS to determine if sufficient capacity is available at critical points in the production process. The RCCP should predict where bottlenecks will occur. The step of the RCCP is to ensure that no departments inadvertently are overloaded (Adam. E., Ebert, R., 1989). RCCP can be applied to all work centers, it is mostly focused on critical ones. The RCCP of Variass does not predict where the bottleneck at the production process occurs because it is not focused on specific operations like assembly, touch-up, End Assembly or QA. The RCCP of Variass is only focused on the capacity groups defined earlier in the research viz. WP, Teams and Test.

Estimating capacity

The first step of the production planning process is to translate demand forecast into planned monthly production levels. In the capacity planning (Figure 18 Capacity planning, for a more complete overview see

Appendix V Capacity planning) the needed capacity over a period of sixteen weeks is given. Variass uses these graphs to present the capacity or needed capacity over the coming weeks.

(27)

Planned occupation End Assembly Max.capacity available

Week 42 0 300 600 900 1200 42 43 44 45 46 47 48 49 50 51 52 1 2 3 4 5 Week 43 0 300 600 900 1200 43 44 45 46 47 48 49 50 51 52 1 2 3 4 5 6

Figure 18 Capacity planning

For estimating the needed capacity/ occupation Planning uses the process time. This process time is estimated by time of previous orders from the customer or by estimation.

When a shortage or abundance of capacity is forehand, Planning change capacity by use of work force changes through hiring or layoffs. Because Variass has many part-time workers they can easily layoff employees. Over the past years Variass changed the workforce several times by hiring and layoffs; therefore they have an extended database to their disposal with many trained workers so they can easily expand capacity by hiring.

From Figure 18 Capacity planning and Appendix V Capacity planning can be derived that the occupation always exceeds the capacity. At the top of the graphs the number of the week when the graph is generated is represented. From these graphs can be concluded that occupation always exceeds capacity in the current week. This is caused by constant backlog of orders.

Planning an order

Planning uses back scheduling. If an order needs to be delivered on the 10th and processing of the orders costs 5 days, production will start on the 5th. Because the production process has many uncertainties, Planning automatically calculates a slack of 5 days so production will start on the 1st. Back scheduling minimizes inventory since products are not produced until they’re needed by the customers. A negative effect of back scheduling is that it results in job lateness, especially for lower-priority jobs (Evans, J. 1990).

The capacity planning in Appendix V Capacity planning is automatically generated by SAP. Planned orders are entered in SAP, which automatically generates a schedule. The capacity is determined by the available employees during the reflected time. Because most of the operations are by hand, machines are not the bottleneck at the assembly department. The biggest bottleneck is caused by the SMT department, but this is left out of consideration in this research.

Planning does not plan at the individual work centers, but by capacity groups defined earlier in the research. Determining the required capacity, the individual work centers at the Teams A an C (viz. Assembly, Touch-up and End Assembly) are not taken into account. Therefore Planning does not give a proper approximation of capacity required at the individual work centers.

(28)

Preliminary conclusion

Because Planning only performs a rough-cut-capacity check on the classified capacity groups no insight is given in the distribution of the occupation/ capacity over the processes within the capacity groups. For some of the processes within the capacity groups employees need specials skills people are not fully exchangeable by each other as result some operations cannot be executed because the qualified worker is not present.

Because QA is not recognized as a capacity group there is no insight in the needed capacity for QA. Because QA is staffed by one person this is extra risky. It is possible that a capacity of 80 hours is required but an occupation of 40 hours is available.

Third, QA is the controlling process so when QA detects a fault the complete batch has to be reworked by the process responsible for the fault and afterwards checked by QA again.

Regularly produced orders are waiting to be checked by QA, but according to QA’s planning the orders have to be checked over two weeks. Planning moved the production forwards because the teams had some free capacity, but QA did not. This problem is caused by the planning system, because Planning does not calculate the capacity of QA.

According to Evans (1990) a capacity planning is not on a detailed level so it does not need detailed product and process data such as bills of materials; it provides only approximations of actual capacity requirements at individual work centers. Planning does give insight in the approximation of the actual capacity requirement but not on at the individual work centre. Because the planning of orders is not specified per process and it does not give complete insight of the occupation per team/ process extra interaction moments with the foremen and the head of the departments. Causing the indirect hours, which on their turn increase costs. In what extend the organization of order release is affected by Planning will be described in the next paragraph.

5.3 Order release

The second key variable which will be investigated is the order release. By investigating this variable an answer will be given to the following research question.

2. How is the organization of releasing orders into the Production department? When releasing orders into the Production department, do they take the workload regarding the different processes into account? How does the order release affect shop floor control?

Considering the PAC framework the ‘Order release’ involves the activities occupation planning, material coordination, workload acceptation and eventually order release.

A week before orders are released into the firm, the Planning department checks ‘material

coordination’ for material availability. In case of Variass this involves PCBs and components. When all the materials are available the planned order is converted into a scheduled receipt. Planning checks if components become available before start date of production. If not, the order is shifted. When materials are delivered sooner than expected and there is free capacity the order is released sooner. When an order is converted in a scheduled receipt, the foremen and the head of the departments can generate a workings schedule from SAP (see Appendix VII Working schedule). In a meeting on

Mondays, with the foremen, heads of the departments and Planning the Working schedule is

(29)

above are also represented in Figure 19 AAD Order release on the next page. In this figure the simplified version of the actor activity diagram (AAD) of order release is represented. In this figure the several capacity groups are put together. In Appendix VIII AAD Order release a more extended version of the AAD is represented, whereby al the capacity groups are identified.

Planning _systemSAP

Scheduled receipt Check materials Warehouse Production Produce orders Picked orders Produced order Communicate problems Pack order Generate working schedule Schedule orders Pick order Production map

Figure 19 AAD Order release

By giving Warehouse a production map, Planning releases the orders. This production map contains information about the order, the quantity, which operations, starting date per operation, process time per operations, batch size, but also information about some special requirements of the order. Like which components have to be placed opposite from each other, etc.

When an order is released by Planning the order is placed in a bin. Order picking picks orders for the Teams A and C by use of kanban cards. By use of these cards WP can control the amount of picked orders for the Teams. As soon as a kanban card for Team A or C is available, Order picking

(Warehouse) begins picking an order. A picked order is moved to WP which prepares the order. After preparing the order the kanban card is send back to Order picking. This is also represented in

Appendix VIII AAD Order release.

(30)

The meetings on Monday can be considered as the ‘Workload acceptance’ (Appendix IV Planning and control framework). As can be derived from the PAC model the ‘Workload acceptance’ is executed by considering the ‘Occupation planning’ and the ‘Execution per capacity group’.

Although the workload always exceeds capacity (see Appendix V Capacity planning) no big alterations in the planning are made due to the flexibility of the workings schedules. During the ‘Workload

acceptance’ meeting the orders which have to be finished are highlighted by the foremen.

Besides the meetings at Monday also a meeting at Thursday is planned. During this meeting problems regarding the running products are discussed. Per capacity group problems regarding different orders are discussed on which Planning adjusts the planning and the Working schedules.

Preliminary conclusion

Due to the absence of a workload control mechanism a sort of ‘Workload acceptation’ meeting is needed on Mondays. During this meeting the orders which have to be finished are pointed out and are highlighted on the working schedule. The orders which are not pointed out do not have to be finished at due date, although it isdesirable it is not necessary. This also becomes visible when we take a look at the working schedule in Appendix VII Working schedule. The working schedule is generated on 1 December 2008. When we take a look at the end date, we can notice that 14 of the 30 orders exceed due date. This number is high but this does not actually mean that process time exceeds customer lead time. This is due to the slack namely;

s

Mlt

Plt

T

=

+

Equation 3

where slack is generally five days. This slack also causes the crux in order release. The foremen and the head of the different department ca not see when the due date on their working schedule equals due date of customer lead time, therefore the meetings on Mondays are necessary to plan those orders which really have to be finished.

Most of these orders become high lighted orders the next Monday, because they also exceeds the five days slack. Planning uses this slack to receive a high delivery performance, theoretically this solution looks fine, but in practice it is not (Bertrand et al., 1998). When a schedule has any latitude it results in a planning without pressure. Because of this, orders are started too late and therefore orders

immediately exceed due date because Planning uses back scheduling. Although Planning does build in this slack between production due date and customer due date it does not have the desired effect. This is because the foremen of the teams are aware of which orders are on due date and which not. For controlling the amount of work the Production department uses kanban, but as described before it only controls the amount of picked orders at WP. Because the loop ends at WP, WP can prepare an infinite amount of orders.

The working schedules with the start and end date offer little support for the shop floor control. But this is mainly caused due to the uncertainty of the process i.e. variety of products, processing times and rate of arrival in combination with the organization of the shop floor control. How this is carried out will be explained in the next paragraph.

5.4 Shop floor planning and control

The last variable which will be examined is the shop floor planning. This variable will be examined by the following three elements; Allocation of capacity, Sequencing and Batch production.

(31)

3. What kinds of planning and control activity are carried out on the shop floor level and what is the impact on inventory, defects and indirect hours?

a. How does the allocation of capacity to the orders affect inventory, defects and indirect hours?

b. How does the sequencing of orders affect inventory, defects and indirect hours?

c. How does the batch production of orders affect inventory, defects and indirect hours?

5.4.1 Allocation of capacity

For controlling the production process Variass uses the foremen and the head of the departments. They use the working schedule to determine which order has to be processed. As described before the orders which are pointed out by Planning during the meeting are leading. If all those orders are in process and capacity becomes available there are different scheduling/ sequencing rules used to determine which order has to be processed first.

To decrease the amount of control needed at the production process the Production department uses self regulating teams (Teams A and C). By use of self regulating teams, management employees will make specific products their ‘own’, with a higher commitment to meet delivery time and ‘zero defects’ production as result. Another advantage of teams is that they will have their specific products. Therefore when a repetitive order arrives it will be processed by the same team. This repetition of orders should let production go faster and reduce defects in production.

Initially the foremen are responsible for the allocation of the capacity on the Workload and vice versa. Next to the foremen the working schedules for determining which orders have to be processed is also used for the distribution of ‘work’ among the employees. Due to the fluctuation of the amount of employees over the day and the high variability in process time this allocation of capacity is not an easy task.

The process time of orders varies a lot; for example during the month October 2008 there was an order with a quantity of 280 products with a production time of 208 hours and an order with a quantity of 500 products with a process time of 44 hours (see Table 1 Variability in process time of complete lots (in hours)). But also an order with a quantity of 9 and process time of 149, see table below.

Order nr. Process WP Ass. Tu EA Test QA Total

Quantity Time (min)

097-B-265-ZZZ 280 16 46 73 3 71 0 208

097-B-285-ZZZ 200 1 13 10 0 40 0 64

1810276 500 5 7 0 32 0 0 44

523117-01-01 9 0 0 0 112 37 0 149

097-B-285-ZZZ 200 1 13 10 0 40 0 64

Table 1 Variability in process time of complete lots (in hours)

* The data of QA are not correct because QA is not considered as a capacity group

(32)

WP Ass. Tu EA Test QA Total

Standard deviation 6.69 10.37 20.95 19.71 26.79 0.19 38.05

Coefficient of Variation 1.55 1.43 2.04 1.44 1.80 12.61 1.67

Average 4.32 7.26 10.26 13.70 14.92 0.001

Table 2 Standard deviation of process time of complete lots (in hours)

The numbers of QA should be left out of consideration because Planning does not take manufacturing for QA into account.

The numbers in table 2 represent the times of complete lot sizes. For a more in depth analysis of the variability also the variability of the process time of a single product is analyzed, see Table 3 Standard deviation of variation in process time (in hours).

WP Ass Tu EA Test QA Total

Standard deviation 0.29 0.42 0.21 0.74 0.64 0.01 5.43

Coefficient of Variation 1.84 1.72 1.29 2.19 3.04 19.87 7.36

Average 0.16 0.24 0.17 0.23 0.21 -

Table 3 Standard deviation of variation in process time (in hours)

From this table can also be concluded that the variability of the process time of one single product varies a lot.

In Appendix VI Process time per lot size examples of the variability in process times is given.

The variability not only occurs in process time between products but also between the processes. For example the order with the quantity of 208 products has a WP time of 16 hours an Assembly time of 46 hours, a Touch-up time of 73 hours, an EA time of 3 hours and a Test time of 71 hours. But also an order with a quantity of 200 products with a WP time of 1 hour, an Assembly time of 13 hours, and Touch-up time of 10 hours and a Test time of 40 hours (see also table 1). This means that the arrival of jobs at the processes also have a high variety.

In Table 4 Capacity by start date the planning for a period of a week is given. For example the process WP needs a capacity of 01-10-2008 of 13 hours and the next day 89 hours. In Appendix IX Process time per start date a complete overview over a period of a month is given. From this table can be concluded that workload has a high variability.

Process WP Ass. Tu Test EA QA Total

Start date (hours)

01-10-2008 13 5 19 11 33 0 81

02-10-2008 89 12 34 23 0 157

03-10-2008 29 39 63 12 2 0 145

06-10-2008 14 55 100 2 125 0 296

07-10-2008 29 96 43 105 31 0 304

Table 4 Capacity by start date

For example workload of End Assembly is 35 hours over three days, but the next day it rises to 125 hours in one day. These figures does not mean that the complete 125 hours need to be finished in one day, but gives an indication on workload that is provided at the different processes. To emphasize this in the table on the next page, the standard deviation of the workload per start date is given.

Process WP Ass. Tu Test EA QA

St.dev 18.01 26.72 45.00 29.53 41.39 0.65

Average 17.92 27.61 46.87 27.63 44.60 0.11