Light on lead-time A RESEARCH TO LEAD-TIME REDUCTION IN A QUARTZ GLASS PRODUCTION COMPANY

(1)

Light on lead-time

A RESEARCH TO LEAD-TIME REDUCTION IN A QUARTZ GLASS PRODUCTION

COMPANY

Philips Lighting Winschoten

(2)

Light on lead-time

-

LEAD-TIME REDUCTION IN A QUARTZ GLASS PRODUCTION COMPANY

- A lead-time research at Philips lighting Winschoten -

J.M. Smit 1495380

Final thesis University of Groningen,

Faculty of Management and Organization, Operations & supply chains

Business Support, Philips lighting Winschoten E.Kral

University of Groningen,

Faculty of Management and Organization, Dr. H.Broekhuis

(3)

Preface

Doing research within Philips Lighting Winschoten was an exciting way to finish my operations and supply chain master. It was not the first time I finished a study but perhaps it was one the most stressfull periods during my studying life. Nevertheless, I am satisfied that I have done it.

First of all I would like to thank Mr. H. Weerkamp. During a period in my youth he was my neighbour and friends with my parents. He arranged the first contact with the Business Support department within Philips Lighting Winschoten.

Here I talked among others with Mr. E. van den Munckhof, manager Business Support and Mrs. E. Kral, head of the planning department. After a meeting of 1,5 hours I went home with the question to summarize the meeting in a problem description. I sent it back to them the same evening. Short after that moment I heard I had the chance to do my final thesis research within Philips Lighting Winschoten. I would like to thank Mr. E. van den Munckhof for this possibility.

Mrs. E. Kral became my day-to-day supervisor and participant in the research. I would like to thank Ellen very much for the constructive cooperation and for the many pleasant moments. Further I would like to thank other frequent participants of the research: Mr. S. Kneten, Mr. W. Perdok, Mr. B. Mein, Mr. J. Niemeyer and Mr. P. van Wijngaarden.

Special thanks to the (temporary) colleagues of the planning department with whom I worked together for a period of 4 months. I hope Hans is still happy with his Sugababes poster, Gonny is as busy as always, Astrid is still showing nice pictures from her little child and Elzo is still looking forward to a glass of beer after his tennis contest.

Besides all the people from Philips, I had meetings at the university. These were sometimes critical but always constructive. I would like to thank Mrs. H. Broekhuis and Mr. D.P. van Donk for their time and the good cooperation.

(4)

Management summary

Philips Lighting Winschoten is in a difficult market that requires flexibility, as well as high quality products and short lead-times. Through diminishing sales to internal Philips customers, third parties around the world become more and more important. Because of heavy competition with countries in the far East on a cost basis, Philips Lighting Winschoten (PLW) sees lead-time reduction as a chance to be competitive.

In this perspective PLW wants to research the possibilities to reduce the lead-times. The following main research question can be obtained:

“How can PLW reduce the lead-time of quartz glass?”

First actual lead-times has to be measured. For measuring the lead-times a group of samples is created which represent the sales for the coming years. Data about lead-times is extracted out of the different computer systems within PLW. The market situation and the way production is organised both contribute to what actual lead-times are. The market as well as the operations are described by using the model of Slack, N. and Lewis, M. (2002). To reconcile both sides of the Slack et al. (2002) model, planning and control (P&C) plays an important role. First a description of the planning process is made and later on influences of P&C on actual performance is investigated.

To obtain basic information about production process, market situation and planning procedures talks with different people in the company are held. Market information of the management is used and general descriptions of planning and production out of the quality system are used. For the analyzing part of the research some brainstorm sessions are organised. In these sessions 1) actual lead-times were tested to the perception of the personnel 2) causes of actual times were discussed 3) lead-time reduction factors are found and 4) possible solutions are created.

The production process of PLW consists of four general production steps. First there is the step of melting the glass in an oven. After the glass is melted and is placed in special containers it goes two the second step, that of acid washing the glass. From here the glass is packed over on special wagons for the fourth step, the stoking of the glass in the vacuum ovens. Stoking takes place on basis of the required stoking quality. After stoking there are two options 1) the glass is ready now and can be shipped to the central stock and after that to the customer and 2) the glass can be shipped to the central stock waiting for furthers processing. In the fourth production step the glass is cut into pieces according to customer requirements, the finishing process.

Two types of lead-time can be measured within PLW. The first type of lead-time is the total lead-time which is the time between the end of the melting process till the end of the finishing process. The second type of time is the time of the long length glass tubes (which is part of the total lead-time) in production which is the time between the end of the melting process till the arrival of the products in the central stock.

The two lead-times extremely differ in comparison to each other. A small part is explainable by the finishing time that rises up to three days. The main part of the this difference is in the time products lie in the central stock. Lead-times of the long lengths also differ in comparison with each other. The main causes for the difference can be found in 1) sorting times 2) cooling times and 3) stoking quality. Sorting times are required by all export products to Japan because these companies require 100% high quality glass without damages. The sorting process itself is not planned but an ad-hoc activity and therefore products are can be on the work floor for long periods which increases lead-time. Cooling time depends on the diameter and density of the glass which differs per product type. The stoking quality is a logical factor that causes lead-time differences, because of the stoking time needed per quality.

(5)

Looking at these performance objectives, range flexibility is guaranteed by offering more then 800 different kinds of quartz glass products. The orders produced are customer specific from the beginning of the process and so PLW adheres the flexibility requirement.

The flexibility objective in some way conflicts with the speed objective. With the position of the decoupling point at the beginning of the production process it seems like PLW is primarily controlling the flexibility objective. In the field of lead-time PLW is not able to fully reconcile market requirements with operations resources.

The planning and control process influences the performance of PLW. P&C can be classified as an three stage planning. First a global MPS planning is made on a monthly basis and for one year in advance (stage 1). Based on this MPS for every week a planning is made (stage 2). A day-to-day planning is made by the production departments their self (stage 3).

In the week plan orders can be loaded till one day before start producing that specific week plan, it contributes here to response flexibility. Only two production processes are planned, the melting process and the finishing process. For the rest, the processes that lie in between (acid washing and stoking the glass) are assumed to have unlimited capacity. This way of thinking results in uncertainty in the process with respect to the order arrival in production.

Planning only two departments in weekly time periods possibly increases lead-time. First the start of an order in the melting process is not know which causes 168 hours (one week) of uncertainty in the start of production of an order. Second the finishing department works with the same procedure, so in total lead-time another week of lead-time uncertainty is created. Besides the above uncertainty there is a planning rule for the finishing process. Long length needed for finishing has to be in the central stock on the moment they are planned. The long length can only be planned in the week after they arrived in the central stock. This rule causes lead-time to rise with a maximum of one week.

Factors that further influence the lead-time are the times between the production processes itself. Lead-time is split up in several times such as handling, waiting, cooling and packaging time. The times are ordered according to the components Johnson, D.J. (2003) comes up with, they are: 1) processing time 2) setup time 3) move time and 4) waiting time.

Processing time and setup time can not be influenced in the scope of this research. Structural changes in the production process are needed which involves high investments. Lead-time can be reduced by decreasing a) move time and b) waiting time in the process.

For reducing both move time and waiting time in the process, two solutions directions are created. To reduce the move time the central stock can be removed out of the total production process in combination with a more detailed planning of the production process. Planning day-to-day further in advance will exclude new lead-times delays caused by planning procedures.

Reducing the waiting time primarily is a solution for reducing long length lead-time. Most of the reduction potential in this part of the process can be found in front of the vacuum ovens. Here products wait for further processing. When combining more sorts of stoking qualities into batches faster processing will be possible and so lead-time reduction can be obtained.

For both of the solutions the lead-time reduction potential is demonstrated by hand of simply excel calculations. Nevertheless hard judgements can not be done about the exact reduction. Therefore this research recommended some further research by using a simulation model of the above situations. A simulation description is made for the production process to analyze the lead-time reduction potential of both solutions in detail. Within the simulation possible disturbances in the production procedure such as malfunctions or under capacity are included.

(6)

1. Description of the problem

This lead-time reduction research in a quartz glass production company is performed within Philips Lighting Winschoten. Before the summer of 2006 the management was struggling with questions about reducing lead-time and keeping flexibility constant. First the company Philips as well as the Winschoten factory are introduced, just as the company structure and its activities.

After the company introduction, the problem is explained by the business case as formulated in June 2006. The business case is formulated in § 1.2 followed by defining the goal and problem statement of the research in § 1.3.

1.1 Philips lighting Winschoten

Philips Lighting Winschoten (PLW) is part of Royal Philips electronics (RPE) which is one of the largest electronic suppliers. Worldwide there are working approximately 126.000 employees (October 2006) and the company has a turnover of 30,4 billion Euros. RPE has production facilities in 28 countries and sells its products in 150 countries worldwide. RPE is divided into four general divisions that are medical systems, domestic appliances and personal care, consumer electronics and lighting. PLW is part of the Philips Lighting division (business group lamps – components). The global structure is represented in figure 1.1.

At PLW quartz glass, special glass types and granules are produced. The quartz glass and the special glass end products are glass tubes that differ in quality, length and diameter. They are used in a wide range of products for water disinfection, for the semiconductor industry, sensors for the chemical industry, for optics and the automotive industry. The granules are a special market that is heavily diminishing.

Figure 1.2: Products made with quartz glass

(9)

Figure 1.4: Achieved customer requirement performance

The organizational chart in figure 1.3 represents the position of the different departments. The Business Support manager and the head of the planning department supervised the research. The planning department falls under the staff department Business Support.

1.2 Business case

PLW is in a situation that it has to make important strategic choices for the future. Competition is heavy and distinguishing itself in the market is necessary. PLW produces quartz glass for two different markets. One is the market of internal customers (for example Philips Turnhout, Philips Aachen). The other market is that of the external customers (worldwide) and can be classified as a market for custom made products. Particular for the internal customers, the glass quantity demanded in the near future is heavily diminishing. This results in an increasing importance of the of custom made products market.

Especially external customers of PLW indicate that short and reliable lead-times are important for them. Figure 1.4 shows the percentage of orders PLW fulfils within customer requirements. PLW considers improving delivery on time as an opportunity and want to compete on this fact, making lead-time an order-winner. Also because of heavy competition with companies in the Far East on a cost basis, PLW has to distinguish on other areas. On of these areas is flexibility, what PLW actually pretends to be. The idea of distinguishing on lead-time is also supported by Tersine & Hummingbird (1995) who stated that lead-time could be a competitive advantage. PLW expects that they could better respond to changing customer demand with shorter lead-times.

At the moment, no clear view on the number of complaints about lead-times can be given. Very little information is available about missed orders through worries from customers about lead-times. However, through frequent contact with customers the sales department knows, orders are missed as a result of long lead-times. The performance indicator (RLIP, Requested Lead-time Indication Performance) used by PLW, which measures the lead-time demanded by the customer against realised lead-time, seldom is above 80%. Approximately PLW scores around 70%, which is a risk in this very competitive market.

As a result of the diminishing quantity of internal products required together with the variety of all the other products (approximately 800 different products), producing all products to stock seems no option. More and more PLW will produce small orders for external customers, so there is many more variety in total production in the near future. Holding all these products in stock will be too costly. This conflicts in some way with the short lead-time objective, that says the more products held in stock, the quicker delivery can take place.

Achieved customer requirement performance

0 20 40 60 80 2004 2005 2006 Year P e rc e n ta g e Plant management Business Support

Finance & Administration Human Resource

Production Innovation

Marketing & Sales

(10)

Because of the reduction in production for the internal market, more free capacity in all of the operations that are a part of the production process will be existent. PLW expects to have more planning possibilities that could contribute to obtaining the target of shorter lead-times. With these facts in mind, now it is a good moment to start a research to shorten lead-times to be more competitive at this point.

1.3 Goal & problem statement

As the previous part described the realized delivery performance to customers is moderate compared to the demanded performance. It is conceivable that customers are not very satisfied with these results. In the new ‘market’ situation, PLW wants among other things to compete on flexibility and low lead-times together. PLW actually pretends to be flexible in the number of different products and the volume of products. So it wants to keep this flexibility at the actual level and to speed up lead-time. The following goal statement is formulated:

“Increase delivery performance by reducing lead-time and holding flexibility constant” Delivery performance is influenced by both speed and reliability (Slack et al., 2002). To influence the delivery performance in a positive direction one or both factors has to improve. The research will focus on speed performance, because quick delivery is an important order-winner in the quartz glass market, especially for the European market in competition with the Asian companies. The research does not concentrate on reliability performance. However, the information obtained about lead-times can possibly be used to analyze this performance further in other research.

The focus on quick delivery leads to the target to reduce lead-times in the process. The following research question could be formulated:

“How can PLW reduce the lead-time of quartz glass?”

The constraint of keeping actual flexibility still has to be taken into account for answering the question. The definition of lead-time in the research question can be defined as; the time between the start of the first operation and the end of the last operation; further explanation follows in § 2.5.

Now the following chapter comes up with a theoretical framework, the scope in which the research is performed. Out of these framework sub questions are formulated and some methodology is setup. Both can be found in chapter 3.

Speed performance

Reliability performance

Delivery performance

(11)

2. Theoretical framework

In this chapter the theoretical framework for the research is introduced. First an overview of the framework is given in § 2.1 followed by a description of the separate building blocks in the subsequent paragraphs.

2.1 Introduction

As introduced in the previous chapter, PLW wants to compete on low lead-times and holding actual flexibility constant. These objectives do not stand alone but are related to each other. To what extend these two objectives relate, depends on market requirements, design of the operations and control elements.

First of all the market requires specific performance of a company. If the company passes or fails on reaching the required performance level or fails, depends on the organisation of the operations. The idea of balancing these two sides is the kernel of the reconciliation idea of Slack et al. (2002). The two blocks in the centre of figure 2.1; ‘market requirements’ (A) and ‘operations resources’ (B); express this idea. The market requirements can be expressed by the performance objectives speed, flexibility, cost, quality and dependability. The operations resources side explains how operations are organised and how they perform. By using this model, the actual situation of PLW can be expressed.

Influencing the lead-time by PLW takes place at the operations side of the model. Strategic choices such as a Make-to-Order (MTO) or Make-to-Stock (MTS) directly affect lead-time as represented by the black arrow (4) in figure 2.1. This strategic choice forms the second block (D) in the conceptual model. The MTO/MTS decision is among other things analyzed by using the Customer Order Decoupling Point theory of Hoekstra & Romme (1992). By using this theory in § 2.2 and § 2.3 the tension between the two objectives lead-time and flexibility becomes more clear.

Planning and control (P&C) is used to reconcile the market requirements with the operations function, (Vollman, 2001). Therefore the P&C box (block C) is drawn under the reconciliation arrows (6) in figure 2.1. Further more planning and control influences the performance indicators lead-time and flexibility directly as the arrows 3 and 4 indicate.

The objectives lead-time and flexibility (blocks E and F) often conflict with each other. Slack et al. (2002) deals with these conflicting objectives by using the trade-off theory. It focuses on the question in what way the performance objectives can be balanced and prioritized. § 2.5 and § 2.6 separately deal with both of the objectives. Flexibility and how it is compliant by among other things the planning is handled in § 2.5 by using a part of the flexibility analysis of van Wezel (2001). Lead-time in relation to speed performance objective is described in § 2.6. Here several definitions of lead-time are analyzed and factors which can possibly influence lead-time are handled by using the throughput time reduction framework of Johnson, D.J. (2003).

All the blocks, also known as concepts, have relations with each other. These relations are represented as black arrows in the conceptual model (figure 2.1). The way how all these concepts

Figure 2.1: Conceptual model

(12)

influence each other, and in this research especially lead-time, is important to give answer to the questions formulated in chapter 3.

2.2 Make-to-Order & Make-to-Stock

The first instrument (block D) in the conceptual model that directly influences lead-times and flexibility, is the MTO/MTO decision of a company. Soman (2005) stated that MTO systems offer a high variety of customer specific products. The competitive priority in this case is shorter lead-time which are realized by MTS systems that offer low variety of producer-specified products. PLW is struggling with combining these two systems as well as combining the different objectives belonging to the systems. For a company deciding to make products to stock or to order, it has to qualify the products. Qualifying can be done by using some simplistic rules such as an ABC classification as Williams (1984) and Carr et al. (1993) suggest or by more detailed methods, for example the demand variability analysis of D’Alessandro and Baveja (2000).

The first method used to classify the type of production of PLW is the product variety concept of Schönsleben (2004). According to Schönsleben (2004) the production type can be defined by the product variety concept in combination with the plant-lay-out. The product variety concept determines the strategy for developing the product and offering it to the customer. It allows the producer to respond to customer requests to varying degrees of variant orientation. This concept is strongly related to the MTO/MTS theory. The higher the influence of the customer is on the product and thus the higher the variety is, the more it is like a MTO situation. The plant lay-out is defined as the physical organization of the production infrastructure, the degree of division of labor among workers, and the course that orders take through work centers.

Figure 2.2 Product variety concept and customer order decoupling point (CODP)

As described by Soman (2005) the CODP concept of Hoekstra & Romme (1992) can be used to solve the MTO/MTS question in a qualitative way. It can directly be related to the product variety concept of

DP1 DP2

DP3

DP4

DP5

Make and ship to stock Make to stock

Assemble to order

Make to order

Purchase and make to order

(13)

figure 2.2). The more variety the products has, the earlier in the process the decoupling point is located. Detailed description of this point and the importance for PLW is described in § 2.3.

To analyze if the product type fits with how production is organized (and so make some categorization according to the product variety concept and the CODP) the demand variability analysis of D’Alessandro and Baveja (2000) is used. They categorize products according to a.) variance in demand and b.) the average demand in a specific time period. According to this, the article come up with four categories of products which are 1. high volume, low variability, 2. high volume, high variability, 3. low volume, low variability, 4. low volume. high variability. The products belonging to category 1 are typically candidates for MTS production. The products belonging to category 2 or 4 are candidates for MTO production.

2.3 Customer Order Decoupling Point

As described chapter 1 the research emphasizes the tension between flexibility and speed (lead-time). The MTO/MTS theory in combination with the decoupling point theory is a good starting-point to make this more clear. Especially the trade-off between flexibility and speed (lead-time) is visualised by the CODP concept, see figure 2.2. The general idea of having the decoupling point (DP) very early in the process indicates a focus on flexibility performance and having the decoupling point (DP) at the end of the process indicates a focus on the speed performance.

The last part of figure 2.2 illustrates the CODP theory of Hoekstra & Romme (1992). In their book they define the decoupling point as the point that separates the customer order part activities (right of the DP) from the activities that are based on forecasting and planning (left of the DP). The customer order penetrates at the decoupling point, and from their goods ordered are supplied to the customer. According to the level of penetration of a customer order, five levels can be distinguished, these are the different decoupling points. In the case of DP1 products are manufactured and distributed to stock points which are spread out and located close to the customer. From DP2 end products are held to stock and from there directly sold to the customer. Using DP3 sub elements of the products are held to stock and are assembled for the customer. DP4 means holding only raw materials in stock and producing each order customer specific. In the case of DP5, even raw materials are bought on the basis of a customer order.

Several authors have studied the CODP and linked the concept to company objectives. For example Van Donk (2001) who linked business characteristics to upward or downward movements of the decoupling point. He stated that short delivery times, high reliability and bad process control requires the CODP to shift downwards. Even so specific products, low stock levels and irregular market demand requires the CODP to shift upwards. Within this perspective, performance objectives of PLW have to be analyzed to determine the DP.

2.4 Planning and control

The second building block (C) in the conceptual model that directly influences performance is planning and control. Planning and control play different roles in the research. At first according to Vollman (2001), matching market requirements with operations resources is a function of planning and control. This system is represented as ‘reconciliation’ by the two opposite arrows (6) between the market and the operations box in figure 2.1.

At second, the conceptual model shows the influence on lead-time and flexibility performance (arrows 3 and 4). Van Wezel (2001) uses planning phases to describe the influence of P&C on flexibility and partly on lead-time. He states that flexibility is related to the time until it is possible to make changes in a schedule. In the case of lead-time performance Bertrand et al. (1998) state that the way of control directly influences the lead-time of a company. To influence this lead-time, they distinguish a set of three control rules:

a) The execution sequence of the orders

In practice planners have the inclination to determine in detail what the production sequence of orders has to be. Because of the many influences of other things in the production process (utilization, malfunctions etcetera) the planning is constantly outdated by the actual situation. b) The application of flexible capacity

This rule deals with the problem of waiting times in the process. It assumes that when having always enough capacity available, no waiting times will occur.

(14)

The start of production of new orders directly influences lead-time of products.

These rules helps to explain the influence of planning and control on the different performance objectives in chapter 7.

P&C also influences flexibility. Van Wezel (2001) classifies planning flexibility by using planning phases and determining at which moment and to which moment changes in the plan are possible. Here he relates flexibility to lead-time by using the time needed to change a plan at different planning phases. The later the plan is changed, the more time it will cost. With this reasoning a planning trade-off exists between speed and flexibility, the later the plan can be changed, the more flexible the company is but the longer the total time of planning and production takes.

2.5 Lead-times

One of the performance objectives of the conceptual model (figure 2.1) and also the central objective in the research is lead-time (E, figure 2.1). At first, a more in depth view of lead-times is needed to define a good definition. In literature a lot of times are defined; there are definitions of flow times, throughput times, cycle times and lead-times. Lead-times can be defined in several ways. Schönsleben (2004) divide lead-times in three parts; the operation time, the interoperation time and the administration time. Lead-time is defined by Hoekstra & Romme (1992) as the period of time between the beginning of the first activity and the end of the last activity of a particular series of activities. Some literature (e.g. Johnson, 2003) defines this period as throughput time or flow time. The operation time of Schönsleben (2004) is defined as the time required carrying out a particular operation. When combining this operation time of Schönsleben (2004) with the activities of Hoekstra & Romme (1992), lead-time can be defined here as the difference in time between the start of the first operation and the end of the last operation. The start of the first operation is the start of the first operation in the production process, the moment physical production starts. This definition includes all operations and interoperation times.

The research concentrates on a reduction in lead-time. As a starting-point for determining factors that influence this time, the framework of Johnson (2003) is used. The model uses Manufacturing Throughput Time per Part (MTTP) as a measure for lead-time. This MTTP is in accordance with the research’ definition of lead-time. The model is layered in four levels, see figure 2.3, where the first level is the throughput time reduction objective. The second level represents four components that can reduce throughput time. The third level represents factors that can influence these components. The fourth level represents actions to change these factors.

In this research, the four main components (level 2) that can influence MTTP are used to influence lead-time. To reduce the components some changes elements (level 3) are given by Johnson (2003). The four components are:

1. Processing time reduction

Processing per part can be reduced by reducing the number of operations required, reducing the processing time per operation, and/or reducing scrap and rework.

2. Setup time reduction

Reducing the setup time can be accomplished by reducing the time per setup and/or the number of setups. Time per setup can be reduced by purchasing equipment with short setup times, improving setup procedures, dedicating workstations to families of parts with similar

Figure 2.3: Levels in the MTTP Framework Objective Components Change Actions

(15)

setup requirements so that common fixes can be developed and used, and/or using family scheduling to group batches that have common setup requirements.

3. Move time reduction

The move time can be reduced by reducing the time per move and/or reducing the number of moves required. The time per move can be reduced by increasing the speed of the material handling equipment, or by reducing the move distance required.

4. Waiting time reduction

Waiting time reduction can be accomplished by reducing setup time, processing time per part, move time, production batch sizes, transfer batch sizes, processing time variability, arrival variability, resource utilization and/or the number of queues.

The level 4 actions suggested by Johnson (2003) are not used in this research. It are some general actions that possibly help to reduce the level 3 factor. In this research some brainstorms sessions are used to find actions to reduce a specific part of the lead-time.

2.6 Flexibility

The other performance objective out of the conceptual model is flexibility. As mentioned there is some tension between both the PLW speed objective and the flexibility objective. Therefore it is important to determine the types of flexibility and how it could be reached. First of all flexibility can be separated in several types according to Slack et al. (2002). The first distinction made is between range flexibility, how much, and response flexibility, how fast the operation can be changed. Further distinctions can made to express total operations flexibility by product/service flexibility, mix flexibility, volume flexibility and delivery flexibility. For the research, the types of flexibility used are plotted and yellow shaded in table 2.1. Schönsleben (2004) links the range flexibility to the MTO/MTS strategy by using the product variety concept. The arrow from MTO/MTS to the performance objectives in figure 2.1 represents this linkage.

Total operations

flexibility Range flexibility Response flexibility

Product/Service flexibility The range of products the company is able to produce

Time necessary to develop and modify products and processes Mix flexibility The range of products the company is able to

produce in a given time period

Volume flexibility Absolute level of output the company can achieve in a given product or service mix Table 2.1: Flexibility types (based on Slack et al.,2002)

Van Wezel (2001) first classifies flexibility at a more aggregate level by distinguishing a) strategic flexibility which deals with reacting to long term changes, b) organizational flexibility to react to medium term changes and c) operational flexibility that deals with reaction on short term changes. Planning influences flexibility and is represented as arrow 2 from the P&C block (C) to the performance objectives in figure 2.1. How P&C exactly influences lead-time and flexibility performance is described in § 2.4.

2.7 Trade-off

The arrow between the performance objectives lead-time and flexibility in the conceptual model represents the tension between this objectives. Slack et al. (2002) deals with the concept of trade-offs between performance objectives. The idea is that a positive shift in one performance in one objective, results in a (negative) shift too in another objective. Schönsleben (2004) uses figure 2.4 to illustrate the trade-off concept.

As figure 2.4 shows, there is a tension between lead-time and flexibility. When having the emphasis on short lead-times (on the right side in figure 2.4), flexibility will be less. Nevertheless it is not an one-to-one relation. Some authors even mentioned a positive relation between these two objectives. As

(16)

Johnson (2003) mentioned; reductions in manufacturing throughput time increase flexibility and reduce the time required to respond to customer orders. In this reasoning it especially influences the response flexibility. Quick throughput times even contribute to improvements in cost, productivity and profitability (James, Garrett, Schmenner (1998)). Van Wezel (2001) stated that the objective of shorter delivery times even causes the need for flexibility.

2.8 Application of the theory

To answer the main research question the theory of this paragraph is used. Sub questions are made based on the conceptual model to create a more in depth view of the problem. First a global overview of the situation within PLW is given by using the reconciliation theory, the planning structure and setting of the performance objectives.

(17)

Figure 3.2: Fit between MEDIC and the DOV-model

3. Research design & Methodology

This chapter describes how the research is set-up. The structure is presented in § 3.1. The theoretical frame presented in the previous chapter leads to the sub questions as formulated in § 3.2. Subsequently § 3.3 deals with the method of research and describes how answer is given to the sub questions.

3.1 Structuring the research

First the type of research and the method of managing the research are described. After this, the research structure becomes clear by using a figure. The last part of the paragraph describes the scope and limitations of the research.

3.1.1 Type of research

In literature research is typified differently. In general there is a distinction in practical research and academic research (De Leeuw, 2001). Cooper & Schindler (2003) call these separation ‘applied research’ with a problem solved emphasis and ‘pure’ or ‘basic’ research, which focus at questions with a theoretical nature.

Cooper & Schindler (2003) distinguish four types of research classified as reporting, descriptive, explanatory and predictive. This research first focused at describing and defining the way of producing and planning followed by explaining the causes of long cycle times. It goes beyond what Cooper & Schindler (2003) called the descriptive research and has many characteristics of the explanatory research. It tries to explain an actual situation by using existing theory.

3.1.2 Managing the research

The research process is guided with help of the Philips project management method, called MEDIC. Within the method, some changes will be made to combine the academic requirements with the practical focused MEDIC method.

The method MEDIC consists out of five phases, which are ‘Map/Measure’, ‘Explore/Evaluate’, ‘Define/Describe’, ‘Implement/Improve’ and ‘Control/Conform’. In this research only the first three phases of the MEDIC method are used. The research is restricted to analyse the current lead-time situation and to do some recommendations for reducing the lead-time. So the implementation and subsequently the control phase are beyond of the project scope. The following figure represented the method in short.

Figure 3.1: MEDIC phases and loops

The steps in the method are not executed in a completely fixed sequence. In every step, there could be extra information out of a preceding phase necessary. Therefore the arrow flowing backwards in figure 3.1 is represented. ‘No one claims that research requires completion of each step before going to the next’, Cooper & Schindler (2003).

The phases can be linked in general to the research methods of De Leeuw and that of Cooper and Schindler (2003) as figure 3.2 shows. De Leeuw (1996) uses the DOV-model1 to have a basic procedure to solve management problems. The M-Phase and partly the E-phase of the MEDIC

1

(18)

method can be fit into the diagnosing phase of the DOV model. Both methods here describe the problem, the actual situation and collect some usable facts. The D-phase of MEDIC stands for ‘describe’ and ‘develop’ and can be seen as the design phase in the method of De Leeuw (1996). 3.1.3 Research structure

The structure of the research is represented in figure 3.3. It helps positioning the different elements in the research and shows the connection with the theoretical framework.

Figure 3.3: Research structure

The model is constructed according to the MEDIC sequence of project management. First in the Pre-MEDIC phase a clear view of the problem is created together with the theoretical backgrounds. In the M-phase in the actual lead-time measurements are done. The block with the operations resources and market requirements describes in what market PLW is operating and how the production process is organized.

In the E-phase all the relations of the conceptual model between the elements are analyzed. On basis of this analysis lead-time could or could not be changed. The final phase, D-phase, comes up with solutions of lead-time reductions within PLW, based on the changing elements of the E-phase.

Problem [H1]

Theoretical framework [H2]

Research design [H3]

Market requirements [H5]

Operations resources [H5] Planning & Control [H6]

Actual lead-times [H4]

Lead-times in a broader perspective [H7] E-Phase Pre-MEDIC

D-Phase Methodology [H3]

M-Phase

(19)

3.1.4 Scope of the project

The project concentrates on the production of quartz glass within PLW in production hall A. Only the vertical melting ovens are taken into account, represented in the picture as SQ-A-B-C-G-H. It further contains the production sections in the yellow colored area, which includes the acid washing department, the vacuum ovens (VO2 till VO4), the finishing department (FD) and the central stock (CS). The PP means (Pre)-Packaging into boxes before the glass tubes are shipped to the central stock. Further information about the production process is given in chapter 5.

Figure 3.4: Research area

Performing the research, some preconditions are taken into account in relation to other performance indications.

• Quality. The glass tubes are very sensible for scratches and dust. The quality of the glass can than be expressed as the total number of handlings with the glass and the time the glass is exposed to open air. The total open-air time and total number of handlings are not allowed to increase.

• Efficiency. The total number of manpower needed for production is not allowed to increase. Re-organization of manpower is possible.

• Costs. The total costs of production are not allowed to increase. Total storage is not allowed to increase. Initial costs for investments are possible, see further in this section.

General restrictions and boundaries

• For volume indications, the quarter 3&4 (2007) forecasts have to be used. As stated in the introduction, volumes are heavily changing in 2007. For the second part of 2007 the changes are largely known.

• The DSO2 is not taken into account. Nearly all of the products produced in the DSO will be phased out in the next year. Remaining DSO products will be handled of in the vacuum ovens. • Within the project, only vacuum ovens 2,3 and 4 (VO2-4) have to be taken into account. Planning restrictions

• The campaign planning for the melting department must be seen as fixed. Personnel restrictions

• The number of shifts in production as well as in the finishing department is a free variable with the preference to have equal shifts.

• Basis operator VAPRO3 and persons with VAPRO A must be able to execute the tasks in the production department.

Equipment restrictions

• Investment in small equipment/material is possible.

• Substantial investment in additional machines/equipment is not possible

2

DSO = Door Schuif Oven (Dutch). Type of vacuum oven where glass flows more or less automatically through the oven.

3

VAPRO = Vakopleiding Procesindustrie, professional education for the process industry

(20)

3.2 Sub questions

The conceptual model presented in the previous chapter comes up with several elements that are in coherence with each other. First organization of the elements within PLW is researched. Filling in the different elements take place in the M-Phase (Map/Measure) of the MEDIC method. This phase is used to create a more specified view at the problem and to collect measurable facts. The following sub questions are constructed.

1. What are actual lead-times at Philips Lightning Winschoten? 2. How can the actual market environment be characterized? 3. How can the actual production process be characterized? 4. How are planning and control organized?

The E-phase (Explore/Evaluate) in MEDIC is created to look at the causes of the identified problems. In this research the phase is primarily used to describe relations between the elements of the conceptual model. The following sub questions can be posed:

5. To what extend is reconciliation achieved between the market requirements and the operations performance with respect to lead-time and flexibility performance?

6. What is the influence of planning and control on lead-time and flexibility performance? 7. What factors further influence lead-time performance?

The D-Phase (Describe/Develop). At the end of the E-phase, there is a clear view of the problems and there is some direction for solutions. Finally there a choice must be made out of the possible solutions based on the pre-set criteria. The last sub-questions deal with this part:

8. What parts of the lead-time can contribute to lead-time reduction? 9. What solutions can help to reduce these contributing factors?

3.3 Methodology

This paragraph describes the research methodology. First there is an overview of which lead times are measured and how they are measured is given. After that there is explained how data is collected in talks with people inside the company. § 3.3.4 describes how the people within PLW are involved by creating solutions for the lead-time problem.

3.3.1 Lead-time measurements

The research starts with measuring actual lead-times. Measuring lead-times of all the Quartz glass products is superfluous, so a selection of samples is made.

Samples

For selecting glass types to be measured a number of criteria and some experience of production and planning personnel are used. For the selection the production facts of 2006 (till November 2006) and the forecasts for Q3-Q4 2007 are used. Out of the different glass ranges that will still being produced in 2007, some samples are selected. The glass ranges are: 300, 304, 361, 370 and 501. The glass types phasing out in 2007 are filtered out of these glass ranges. The aim is the selection is to collect a group of products from where a lot of data is known. So the group of products will represent a large part of the 2007 production.

Selection criteria for the samples: 1. Part of total produced

(21)

2. Frequency ordered

The next measure is the frequency that orders return on monthly basis in the production process. Orders with frequent repetition have high attachment at the production facilities. 3. Diameter variable in the sample

Glass tubes are produced with different diameter sizes. 4. Stoking quality

There are different stoking qualities for glass tubes that influence the time the tubes are in the production process.

5. Different production procedures

Not all the products produced follow the same production route. Some products are directly sold to the customer after the melting department, other products are first acid washed or dried. In the selection a difference is made between products that still have to be finished after the vacuum ovens and products that are ready after the vacuum ovens. Products sold directly after the melting departments are not included. This is < 1% of the products incorporated in this research.

Points to measure

To measure the global ‘ production’ lead-time of products within PLW, there fixed points in time have to be used. To indicate these points figure 3.5 represent the production process. An extensive description of this process can be found in appendix 2. M means the Melting department, A the Acid washing, V the stoking in the vacuum ovens and F the finishing department. The numbers are the start and finishing of each production step. Between 6 and 8 the production is decoupled by the central stock, represented by point 7.

Earlier there is made a distinction between Make-to-Order and Make-to-Stock products, which have different decoupling points (CODP). With respect to these two types of orders, there are also two definitions of lead-time to discern. First there is the lead-time for long lengths, represented by the point 1 till 7 in the figure. Second the lead-time of fully finished products, which represents points 1 till 9. To have a more precise view at the total lead-time and delays, information of the points located in between is useful.

Lead-time data

In the company there is much information about the points illustrated in figure 3.5. The available information is analyzed and checked if it is reliable to use. The extended procedure of searching and selecting the data is described in appendix 1.

During selecting data it appears that not all points are suitable to measure with the information systems. Table 3.1 summarizes the points that can be measured and the information used for measuring.

Point in the process Measuring tool

2 _• MASTERDATA – certifications from APS. 7 _• SAP – Ready for transport notifications. 8 _• PDB – Start times sawing per shift.

• PDB – Start times sorting per shift.

9 _• Certification tool – Date finished products are certificated. • SAP – Ready for transport notifications of finished

products.

F

M

A

V

1 2 3 4 5 6 7 8 9

Figure 3.5: The two lead-times

(22)

3.3.2 General data

To answer sub questions about the market environment, the production process and the planning process, much information is used and a lot of actions are executed.

Information

For collecting solid information about market, production and planning the following sources are used: • Production process

For information about the production process a lot of information is used. Most of the information is derived from production descriptions out of a quality system (flow diagrams). • Market information

Market information is derived using presentations and orientations of the actual and future market situation of the PLW management. Further material about forecasts purchases of internal Philips customers of the PLW sales department are also used.

• Planning process

Few hard information of the planning process can be found. Only a general description of the planning process is available. See ‘Actions’ for further information.

Actions

Beside the above hard information, several information is collected through personal contact with people in the company. Here follows an overview of the actions executed.

• Contact moments

During the research there were several fixed contact moments.

o Once a week operational consultation with head of the planning department E. Kral, production manager P. van Wijngaarden and the head of the finishing department S. Kneten.

o Once a month a short briefing to the planning department personnel about the progress of the project.

• Meetings

o To collect information about the planning process a few (pre)-MPS4 meetings are visited.

o Participation in the weekly departmental meetings of the planning department.

o During the research two meetings with the head of the Business Support department, E. van den Munckhof were held.

o The researched is finished with a presentation for the general management, production personnel and planning personnel.

Production (specific) information

During collecting more specific information about glass types, production facilities and specific procedures in production operators in production and the technical production staff were consulted. For measuring lead-times and day-to-day feedback from production a production employee who was especially trained for helping during this kind of projects was consulted.

3.3.3 Lead-time analyzing

After the lead-time measurements and describing the actual situation at PLW, I organized some multi disciplinarily sessions with production employees, employees of the finishing department, the production manager and planning personnel. Goal of the individual sessions differ, so the sessions are explained here and in the following paragraph.

Session 1: Informative session; causes of lead-times. Input: Measured lead-times.

Participants: Planning personnel, production personnel and production manager.

4

(23)

In this session I informed those present about the facts measured. I tried to found some factors that actually influenced the lead-time within PLW. All the people involved had the opportunity to react on the times measured. Together the most important causes of lead-time and lead-time differences are formulated.

3.3.4 Lead-time changing

In line with the research goal, lead-time has to be changed. So changeable factors in the actual (measured) lead-time have to be discovered. For this purpose, the lead-time was split-up in parts, according to the Johnson (2003) framework (see chapter 2) and is used in the sessions as described here. For increasing legibility of the report the factors as well as how they are measured are both explained in § 8.1.

Session 2: Brainstorm session 1.

Input: Lead-time measurements, causes of lead-times as defined in the first session, lead-time divided in several factors (in line with the conceptual model, among other things factors of the Johnson (2003) framework).

Participants: Head of the planning department, head of the finishing department, MPS-planner, two production employees.

In these session I split-up lead-time in several factors in line with the theory. Together it was decided which of these factors can be influencing lead-time within PLW. Based on this decision the most diverging ideas are created to reduce lead-time. The ideas are used to develop the solutions in chapter 8. At the end, the ideas are ranked by each of the participants. Finally the rankings are collected and put it in a scorecard. For the exact ranking of the ideas the scoring method, see appendix 3.

The ideas are ranked by using four factors: a) the feasibility of the solution, b) how costly the solution is, c) what the consequence for the quality of the glass is and d) the lead-time reduction potential of the solution.

Session 3: Brainstorm session 2.

Input: Same as the first brainstorm session, solutions created in brainstorm session 1, scorecard with the total ranking of the solutions.

Participants: Head of the planning department, head of the finishing department, MPS-planner, production manager, two production employees.

A second brainstorm session is organised to work out some of the ideas partly based on the ranking from the first session.

During this session, more specialised people on the field of planning and logistics and the production processes are involved. I presented the scores on the solutions created in the last session and together it was decided to work out the ideas with the highest scores in more detail.

The above sessions give some more in depth information about actual lead-times, planning and control and the market (first sub questions). It also gives more information in what factors of lead-time can be influenced within the scope of PLW’s production. Among other things based on the scores few solutions were depicted to work out in more detail with some people from the planning department who have a logistical background. Therefore I organized the following session.

Session 4: Lead-time solution session.

Input: Solutions left after brainstorm 2, lead-time reduction factors, total scores of the solution Participants: Head of the planning department, MPS-planner, planner of the finishing department

(24)

(25)

4. Actual lead-times

In this chapter the lead-times of the samples measured are represented. First follows a definition of the optimal time used in this research. Then in § 4.2 an overview is given of the measured lead-times, divided in a) total lead-time and b) long length lead-time. Different statistics are used to find lead-time differences between glass types. In § 4.3 some conclusions of lead-time are formulated. In the research the following samples are used based on the criteria as described in chapter 3. For making reading of the report easier the samples are represented on a page you can unfold placed just before the appendixes.

300 glass Nr. D (mm) L (mm) Fq Q F % of P Code number 1 31,00 1625 7 ED - 2,5% 432211888761 2 10,9 200 10 XD D 1% 432211886771 3 25,8 147 6 XD D 2,5% 432211885531 4 9,0 1220 7 ND - 1 % 432211865011 5 18 83 10 ED D 6,5% 322220265111 304 glass Nr. D mm) L (mm) Fq Q F % of P Code number 6 28 1540 9 ND - 24% 432211886171 7 28 1450 6 MD 22% 432211873851 8 27,1 300 9 ED D 4% 322220263551 361 glass Nr. D (mm) L (mm) Fq Q F % of P Code number 9 32,5 324 10 MD H 19% 432211885291 10 8,0 1495 7 MD - 7% 432211873471 11 9,0 110 6 MD D 9% 432211891841 370 glass Nr. D(mm) L (mm) Fq Q F % of P Code number 12 2 245 10 XD M2 59% 322220269471 13 2 1500 7 SD - 14% 432211877311 14 3,4 1550 3 - - 5% 432211888861 501 glass Nr. D(mm) L (mm) Fq Q F % of P Code number 15 4 1280 9 SD - 26% 432211881841 16 4 220 10 SD M2 22% 322220268451 Table 4.1: Samples

For the selection of the samples, the next abbreviations are used. • D stands for the diameter of the glass tube in mm. • L the length of the tube in mm.

• Fq how often an order is returned in the last 10 months.

• Q is the glass quality after heating in the vacuum oven. The qualities are determined as:

Quality Explanation Drying hours

XD Extreme Dry 60

SD Special Dry 36

ED Extra Dry 20

MD Medium Dry 14

ND Normal Dry 10

(26)

Table 4.4: Acid washing time

• F stands for the type of finishing required for that glass. The types of finishing are determined as:

Finishing type Explanation

H Hand Sawing

D Mechanical Sawing

M2 Scratching and breaking the glass

Table 4.3 Finishing types

• % of P stands for the percentage of the glass in total demand (in kg).

4.1 Optimal lead-time

The optimal lead-time used in this research is defined as ‘the production and handling times a product requires without the loss of any other waiting time in the process’. It is used to give a more or less realistic view at what is possible in production when all goes well.

The optimal lead-time can be divided according to the sequence of the different production steps. Successively this lead-time consists out of the acid washing time (A), handling time before stoking (H1), average standard waiting time before the vacuum oven (SW), the stoking process time (S) itself, the cooling time for some types of glass after stoking (C) and the handling time to pack the glass tubes (H2). For the finishing department no optimal lead-time is calculated.

Optimal lead-time can than be expressed in the formula: OLT = A + H1 + SW + S + C + H2 4.1.1 Acid washing time

For each of the glass types different acid washing programs are used. The program consists of different steps such as acid washing, rinsing and drying. Below the total time of the programs is presented.

4.1.2 Handling

The handling time contains two moments. The first handling time (H1) is the time that the cassettes filled with glass tubes are re-packed to the wagons that are made for the vacuum ovens. The second handling time (H2) is the time required for packing the glass tubes from the wagons into boxes for external transport as well as into boxes for internal use. Both handling times are estimations of employees working in production.

Handling time between point 2 and 3 in production (H1)

Glass type Diameter (mm) Time (hours)

300 > 9 mm 0,33

304 > 27 mm 0,33

361 > 9 mm 0,33

370 2 mm 0,58

501 4 mm 0,58

Table 4.5: Handling time (H1)

Glass type Time (hours)

300 0,92

304 0,83

361 1,50

370 1,63

(27)

Handling between point 6 and 7 in production (H2)

Glass type Diameter (mm) Time (hours)

300 > 9 mm 0,75 300 9 mm 1 304 28mm 0,75 361 external use 9 - 32,5mm 0,75 361 internal use 9 - 32,5 mm 0,33 370 2 mm 1 501 4 mm 8

Table 4.6: Handling time (H2)

The 501 glass type has an extremely long handling time. The glass has to be sorted fully for scratches and little cracks. Even so the 300 9mm and the 370 2 mm have in comparative long handling times, also because of a sorting step.

4.1.3 Standard waiting time before the vacuum oven

Before the glass goes into the vacuum ovens, a charge5 is composite. A charge has some preconditions:

• A charge is put together on basis of the glass quality.

• A charge contains two wagons, which mostly represent eight lots, four lots per wagon.

With these preconditions the glass has to wait substantial time before it will be further processed into the vacuum oven. Johnson (2003) calls this time wait-to-batch time. Based on the information above it is possible to calculate an average standard waiting time before the vacuum oven. For the calculation the following assumptions are made:

• The glass is melted with a speed of 35kg per hour for the 370 and the 501 types and with a speed of 50kg per hour for the rest of the types.

• The acid washing department does not cause a delay more than the production time of the glass per lot.

So with these preconditions and assumptions in mind, the following rule can be setup:

• WT8 = 0. This means the last lot of a charge (lot 8) has a minimum waiting time before the vacuum oven. When the last lot is on the wagon, the charge can directly be loaded into the oven.

• For every other lot (number 1 till 7) the waiting time is longer. It depends on the production time (PT), the time that is required to melt the glass, of a lot for that type of glass. It could be expressed as: WTx = PTx+1 + max PT(8-x)=x

• When taking [ WT1,2,3,4,5,6,7 / 8 ] the average standard waiting time before the vacuum oven is calculated.

The waiting time for one lot of a specific glass type is calculated based on the total weight of a cassette, which depends on the diameter of the glass and the number of tubes that a cassette contains. The more a lot weights, the longer melting time of a lot, so the longer the time the first lots have to wait for the last one. A numerical example:

Glass 300 ND

Melting speed (s) 50 Kg per hour

Weight glass tube (w) 67,5 gram Cassette quantity (q) 1980 pieces

Production time (PT) (w*q) / s ≈ 2,67 hours

Waiting time per lot (1 till 7) 2,67 hours

Waiting time lot 1 (WT1) (7 * 2,67 ) = 18,9 hours Waiting time lot 2 (WT2) (6 * 2,67) = 16 hours Etcetera.

Total standard average WT WT1,2,3,4,5,6,7 / 8 ≈ 9,35 hours

Table 4.7: Wait for batch time calculation

5

(28)

For every glass type used as a sample the above calculation was made, so the following standard waiting times before the vacuum oven occur. Number 9 is the 361 glass with such a great diameter (32,5mm) that a lot is ready very quick. The numbers 10 and 12 have long waiting times because they are produced with anti scratch plates in the cassettes and therefore have to wait for 16 cassettes instead of 8.

Product number WT in days

1 0,24 2 0,66 3 0,38 4 0,39 5 0,52 6 0,20 7 0,20 8 0,21 9 0,15 10 1,38 11 0,60 12 1,73 13 0,86 15 1,07 16 0,67

Table 4.8: Waiting times 4.1.4 Stoking

The stoking times differ per glass quality. In this total stoking time also 6 hours of heating up and cooling down time are included.

Quality Stoking time (hours)

XD 54

SD 30

ED 14

ND 8

MD 4

Table 4.9: Stoking times 4.1.5 Cooling time

Some of the products, especially that with small diameters, require a cooling time down after heated up by the vacuum oven. The table beneath represents the products that require a cooling down period and the period itself.

Product number Cooling time in days

2 0,42 3 0,08 4 0,33 5 0,16 10 0,33 11 0,33 12 1,00 13 1,00 15 1,00 16 1,00

Light on lead-time A RESEARCH TO LEAD-TIME REDUCTION IN A QUARTZ GLASS PRODUCTION COMPANY

Light on lead-time

A RESEARCH TO LEAD-TIME REDUCTION IN A QUARTZ GLASS PRODUCTION

COMPANY

Philips Lighting Winschoten

Light on lead-time

-

LEAD-TIME REDUCTION IN A QUARTZ GLASS PRODUCTION COMPANY

Preface

Management summary

Table of contents

1. Description of the problem

1.1 Philips lighting Winschoten

1.2 Business case

1.3 Goal & problem statement

2. Theoretical framework

2.1 Introduction

2.2 Make-to-Order & Make-to-Stock

2.3 Customer Order Decoupling Point

2.4 Planning and control

2.5 Lead-times

2.6 Flexibility

2.7 Trade-off

2.8 Application of the theory

3. Research design & Methodology

3.1 Structuring the research

3.2 Sub questions

3.3 Methodology

F

M

A

V

4. Actual lead-times

4.1 Optimal lead-time