TPM implementation at Nefit

NEFIT (BOSCH) / UNIVERSITY OF GRONINGEN

TPM implementation at

Nefit

Maintenance the Bosch way

13/08/2009

Supervisor: Prof.Dr.Ir. J. Slomp Secondary supervisor: Dr.Ir. W. Klingenberg

ABSTRACT

This thesis is about a five month action research project on TPM implementation at Nefit. The purpose is to design, and actually start implementation, of a TPM maintenance strategy, complying with Bosch norms and guidelines. Also an accompanying implementation program is to be developed.

The main elements are; a proposed design of a TPM maintenance strategy, a recommended implementation process and two pilot projects conducted at both production facilities during four moths at Nefit. The pilot projects have led to, valuable insights for design and implementation, and to actual implementation of TPM on two stations.

FOREWORD

Five moths of intensive action research at Nefit and some more time spent on writing at home, has eventually lead to this thesis. This is an appropriate moment to thank who ought to be thanked and to reflect on the process of writing a graduation thesis.

I would like to start by thanking everyone at Nefit Buinen and Nefit Deventer for the great five months I had and the pleasure we often had during the pilot projects. Special thanks go to Klaas Oosting and Jos Plasschaert, who were always available when I needed help.

From the University I would like to thank Jannes Slomp for the many hours we spent discussing in his office (Mostly on the research model). Warse Klingenberg I want to thank for his great flexibility and punctuality making it possible for me to finish this thesis within this study year.

Of course, I can not forget Lia, Jesse and Merel who had to withstand my fluctuating moods,but always stayed positive. Sjoerd is thanked for reviewing the slips of my pen. And of course all the people who supported me in one or another way (mum and dad for the money of course), it wouldn’t have worked without you all.

TABLE OF CONTENTS

1.2.1 General production characteristics... 7

1.2.2 Product flow ... 9

1.2.3 Machines and equipment at Nefit ... 13

1.2.4 Logistics ... 14

1.2.5 Planning and Control ... 14

1.3 BPS ... 17 1.4 TPM ... 19 1.4.1 General ... 19 1.4.2 Pillar 1 ... 20 1.4.3 Pillar 2 ... 20 1.4.4 Pillar 3 ... 20 1.4.5 Pillar 4 ... 21 1.5 Maintenance at Nefit ... 21 1.5.1 General ... 21 1.5.2 Pillar 1 ... 21 1.5.3 Pillar 2 ... 25 1.5.4 Pillar 3 ... 26 1.5.5 Pillar 4 ... 26 1.5.6 Training ... 26 2. RESEARCH DEFINITION/METHOD ... 27 2.1 Background ... 27 2.2 Research model ... 27 2.2.1 General ... 27

2.2.2 Elements of the research model ... 28

2.2.3 Relations between the elements ... 29

2.3 Research Questions ... 30 2.3.1 Main question: ... 30 2.3.2 Sub questions: ... 30 2.4 Research Goal ... 30 2.5 Method ... 30 2.5.1 General ... 30 2.5.2 Chapter 3 ... 31 2.5.3 Chapter 4 ... 31 2.5.4 Chapter 5 ... 32

2.6 Restrictions and assumptions ... 32

3. THE PILOT PROJECTS ... 33

3.1 General ... 33

3.2 The Buinen testing station ... 33

3.3 The Deventer burner assembly station ... 34

3.4 The initial workshops ... 35

3.5 Training and evaluation ... 36

3.6 TPM assessment ... 36

4 DESIGN OF THE TPM ORGANIZATION ... 37

4.1 TPM standards and guidelines ... 37

4.1.1 General ... 37

4.1.2 Procedures and standards ... 38

4.1.3 TPM teams ... 40

4.2 Integration of TPM with the production process ... 42

4.2.1 TPM and the takted line ... 42

4.2.2 TPM and 5S ... 42

4.2.3 TPM (AM) and process confirmation ... 43

4.2.4 Integration with existing standards concerning CIP ... 43

4.2.5 Integration with existing training program ... 43

4.3 Effects and overall performance... 44

4.3.1 Plant performance, vision, strategy and targets ... 44

4.3.2 Direct effects of TPM on performance ... 46

4.3.3 Moderating effects ... 48

4.4 The future state of the TPM organization ... 49

5. ACTIONS AND PREREQUISITES FOR IMPLEMENTATION ... 51

5.1 General ... 51

5.2 The BPS on TPM implementation ... 51

5.3 Literature; factors affecting successful implementation ... 52

5.4 TPM implementation so far and recommendations for future ... 54

5.4.1 TPM organization at Nefit ... 54

5.4.2 The TPM assessment and analysis... 54

5.4.3 Project organisation and phased implementation ... 55

6. CONCLUSIONS AND RECOMMENDATIONS ... 56

6.1 Conclusions ... 56

6.2 Recommendations ... 57

REFERENCES ... 58

APPENDIX A: Workshop action sheet ... 59

APPENDIX B: TPM activity list & Proof of execution ... 60

APPENDIX C: TPM stickers visualization ... 62

APPENDIX D: Standardized work instructions ... 63

APPENDIX E: Maintenance schedule ... 65

APPENDIX F: TPM Layout form ... 66

APPENDIX G: Workshop documentation ... 67

APPENDIX H: TPM assessment analyisis ... 68

APPENDIX I: Project documentation ... 72

APPENDIX J: Graph failure data ... 75

1. INTRODUCTION

1.1 The company

This research thesis is about reviewing, and actively changing the organization of maintenance at Nefit. The change in maintenance is directed towards Total Predictive Maintenance, a holistic approach to maintenance often associated with lean production. The maintenance organization of both Nefits production facilities in Deventer and Buinen is considered.

Nefit is a thermo-technology company producing central heating boilers exported to 15 countries worldwide and is market leader in the Netherlands. The company is part of the Bosch group, one of the worlds biggest private industrial corporations with 272.000 employees, of which 3.300 in the Netherlands, and a turnover of 46,3 billion euros in 20071. Originally, Nefit is a Dutch company consisting of two production facilities, 115 km. apart, in Deventer and Buinen, and central research and development, marketing and sales departments in Deventer.

The two production facilities originate from 1967 (Buinen) and 1948 (Deventer). Back then they were separate companies by the names of Fasto B.V., originated in 1897, and Nefit, originated in 1948. In 1980 the first condensing (HR) boiler ever is brought to market by Nefit. In 1992, because of a government subsidy, Fasto and Nefit decided to join forces and continue as Nefit Fasto B.V. In 1993 Buderus Heitztechnik GmbH took over Nefit Fasto with the strategic aim to extend export, which indeed increased to 50% of total turnover. The company name was changed into Nefit Buderus B.V. Recently, in 2004, Buderus in its turn was taken over by the Bosch group, resulting in Bosch Thermotechnik GmbH of which Nefit B.V. is part.2

Over the last decade, both production facilities went through a major change process when the traditionally arranged assembly plants were turned into modern lean plants with takted driven assembly lines and fishbone structures. Both facilities are very similar though slight differences exist. Later in this thesis, these differences will be further elaborated upon. Within the ‘lean manufacturing’ field Nefit is a rather advanced player endorsed by constant top classifications at the yearly Bosch Production System assessment, and by being voted ‘Die Fabrik des Jahres’ by the German journal ‘Produktion’ in 2007 for the Buinen plant.

The Bosch Group stimulates all its subsidiaries to apply lean production elements, and has formulated an overall production strategy called the Bosch Production System (BPS).The BPS contains 8 lean principles, guidelines for implementation, and standards for utilization of these principles and guidelines. Some parts of the BPS are based on Nefits assembly plants, as they are being considered best in practice on some issues.

When it comes to maintenance, lean production, as well as the BPS, consider this to be essential to the system. Both lean production and the BPS prescribe Total Productive Maintenance (TPM) as a way of organizing maintenance. TPM is a holistic maintenance program, aiming to maximize the effectiveness of equipment, by total participation in autonomous small group activities, through preventive maintenance for the equipments entire lifespan (Nakajima, 1988).

For various reasons Nefit chose to implement the TPM program. The redesign and implementation of the maintenance organization is the main subject of this thesis.

11

http://www.bosch.nl/content/language1/html/867.htm

1.2 Production at Nefit

1.2.1 General production characteristics

The production at Nefit consists of two assembly lines, one in Deventer and one in Buinen. Although the two lines visually seem to differ a lot, a closer look into the plants reveal that the lines are in fact rather similar. The reason for the apparent differences is that while the boilers in Deventer are transported through the facility on product carriers hanging on rails suspended from the ceiling, in Buinen the product carrier travels along a conveyor belt on the floor.

Figure 1. Product carriers (left: Buinen, right: Deventer)

Another cosmetic difference is that the paced line in Buinen flows more or less in a straight line from incoming goods to expedition, while the line in Deventer takes a seemingly less logical route, curving through the facility. Some real differences also exist, perhaps the most important being the assortment of boilers produced in the facilities. In Deventer boiler types are produced with higher capacity which are larger and heavier then the lower capacity boilers made in Buinen. The most important consequence for the process is that lifting aids are more often needed in Deventer than in Buinen. Another essential difference is that while the finished goods inventory in Buinen is restricted by the capacity of the automated pallet buffer, in Deventer an automated high-capacity, high-rise warehouse called ‘hoogbouwmagazijn’ (HBM) preceded by a cross-dock, is used for storing raw materials and finished goods. Finished goods from Buinen are also stored in the HBM. Despite the differences it is worth mentioning that Nefit is actively trying to get processes identical in both plants.

The average processing time is approximately 100 minutes on average per boiler. Output in two shift operation, generally varies between 450 and 800 boilers daily, subject to demand. In both facilities production normally is planned in two shifts. The first starting at 6:00h, change of shifts at 14:30h, and ending at 23:00h. At the time of this research, the Buinen plant was operating only in day-shift, starting at 7:00h and ending at 15:30h, caused by falling demand due to the economic recession. Normally the facilities run only in day-shift during holidays. The number of employees for each shift depends on the takt-time the plants are running, ranging from 60 to 180 seconds for Buinen and from 70 to 120 seconds for Deventer. At each shift about 4 mentors are present, each of them responsible for a number of production sections. Mentors are responsible for managing and coordinating of daily events. The mentors report to the production chief, who is responsible for a number of sections as well. All three production chiefs together are responsible for the whole production process. The industrial engineering department, consisting of one employee (manager) for each facility is responsible for changes in production. Collaboration between the different departments is based on informal relations.

At maximum capacity (shortest takt), approximately 90 employees are needed in production, for both Deventer and Buinen. At the lowest takt approximately 40 employees are needed. On average 70% of the production workforce consists of fixed employees, the remaining thirty percent are temporary workers employed through a temporarily employment agency, which has a subsidiary on both locations. Beside the operators assigned to a specific section, also extra universal pool workers are present. As a rule they account for 10% of the total production workforce. The pool workers are assigned to production whenever deviations, e.g. absence, occur. Besides that the pool workers are also assigned to continuous improvement workshops and for cleaning and simple maintenance tasks.

Both production lines are characterized by a driven, takted line, transporting boilers on product carriers, through the factory, in series of four. In Deventer the associates assemble the boiler walking behind the product carrier, in Buinen the product carrier is a platform that carries the associate along with the boiler. In both cases the boiler can be adjusted to the most comfortable height. In Deventer at some points the boiler is automatically adjusted to an average height, manual adjustment can be done but requires a bit of work. In Buinen the height of the boiler on the carrier is adjustable at any time by electrical control.

The overall line-layout is a fishbone structure in which the line is the spine. Sub-assemblies are made in separate sections located next to the line in one-piece flow and delivered directly to the line. Work-cells for sub-assembly production (sections 4 and 5 in Buinen) are equipped with simple universal tools like electrical and pneumatic screwdrivers, torque-wrenches, and some specially designed assembly tools and/or testing gear. The work-cell layout consists of (height adjustable) tables, the line on one side, and pick-to-light kanban racks to the other side. About six people are employed in a cell on a typical day.

In order for the takted line to function efficient, line balancing is an important issue. Since most of the activities within production are manual assembly tasks, the capacity side of the line balancing problem is a relatively easy one. The capacity of each section is mainly determined by the number of employees working in the section, which makes it more ore less continuously adjustable.

because continuous production of the first type will lead to a significant lower degree of utilization than in the case of the second boiler. Capacity demand, defined by the number of workers needed in the section, will change subject to the product mix. This will result in either a changing number of workers per section or frequent idling of section workers, both ill effects. For this reason the average of the processing time of different boilers is kept as uniformly as possible. This is done through both product and process design. Uniform product design will generally lead to uniform processing times. At the time of this research, time studies were done for all assembly tasks, resulting in accurate knowledge of workload per section per boiler type. In the near future the actual daily product mix will be simulated to determine the number of employees needed per section. For process design, sequence independent tasks can be performed in different sections as to level processing times. Important for mix flexibility though is that the total processing time, of different boilers, is as uniform as possible.

Essentially the process is one-piece-flow, though four boilers of the same type are always produced in a row. This is because the boilers are packed and shipped by four on a pallet. Some types of boilers are produced in series of eight, because for these types, eight boilers fit a pallet. The actual batch size is one; after work on a boiler at one section is finished, it flows to the next station without having to wait for the series of four to be completed.

1.2.2 Product flow

Because of the similarity between the both facilities, and because of time restrictions and logistic advantages, this research will mainly be focussing on the Buinen plant. When appropriate, comments will be placed relating to the Deventer situation.

Production starts when the line has an empty product carrier available at the beginning of the line (section 1). First of all a set containing, the product booklet, parts bag, boiler chip and a product guiding form called ‘productbegeleidingsformulier or PBF’, is hung on the eventual wall-mounting bracket on the product carrier. These sets are prepared up front in section 0, and delivered to section 1 using kanban control. These sets are prepared in the sequence determined by planning and dictated by the computer system in section 0.

Figure 2. Product guiding form

While the carrier advances through sections 1 and 2, sections 4 and 5 start preparing the subassemblies in the exact same order as the approaching boiler assemblies. Orders for these sections are released by the computer system and shown on a screen at the beginning of the pick to light racks. Work in progress is restricted by the takt of the system and by the limited space of the conveyor belts between section 4/5 and the line. The associates start by picking parts, to light, for the next order using a shopping cart. Once the order is picked the associate walks to a worktable and assembles the picked parts. When the sub-assembly is ready, it is placed on the conveyor belt by the associate. The associate takes the cart and starts picking parts again as soon as the system releases the next order on the small screen as shown in Figure 3. In section 4 hydro-frames are assembled and in section 5 gas-air modules and electrical wiring assemblies are prepared. Figure 4 shows section 4, section 5 looks similar.

Figure 4. Picture of section 4

The racks on the right are pick-to-light kanban racks, the shopping carts are placed next to the assembly desks. The line is visible on the left, the main product flow is towards the camera. The finished sub-assemblies are placed on the small green conveyors on the left.

While the sub-assemblies are prepared, the carrier progresses to section 6 where an associate boards and starts assembling the boiler. When the carrier passes section 4 the associate picks up a hydro-frame and mounts it to the boiler frame. When the carrier passes section 5 the gas-air module and if necessary the electrical wiring assembly are picked from the conveyor belts and mounted onto the boiler. The associate leaves the carrier and walks back to the beginning of the section and boards on the next free carrier. The carrier moves into section 7 where the next associate boards and performs some last assembly tasks.

At this point the line splits six-ways into the six parallel test-lanes. The associate of section 7, finishes connecting the electrical wiring and performs a high voltage test, before the carrier moves to the final-testing unit (section 8) and stops. Here the operator awaits the boiler and connects the boiler to the testing units through 4 water tubes and a gas tube. The boiler is tested and inspected extensively on several points, e.g. gas/fluid leaks, flow values and operation. The maximum time spent in this section is 6 minutes, if testing takes longer the line is jammed.

After testing the operator disconnects the boiler and releases the carrier to progress into section 9. The cover of the boiler is fitted here and the boiler is boxed and labelled with stickers already printed at section 0. During boxing the boiler is lifted from the carrier with a boom crane operated by an associate, lowered in the box, stapled and placed on a pallet. When a pallet contains 4 boxed boilers the pallet is rolled into the automated pallet transporter (section 10). Here the pallet is strapped and stored in the pallet buffer, all automated, waiting to be shipped.

made up front using kanban control. The reason for this is that some processes require some drying time before the sub-assemblies can be mounted on the boiler. Deventer also has a buffer at the end of the line, this buffer is rarely used.

Figure 6. Lay out production line Deventer

1.2.3 Machines and equipment at Nefit

Generally the MAE of both facilities can be divided into the following groups; the driven lines with product carriers, packaging and handling, forklifts and transport carts, hand tools, specific assembly tools (templates), general testing gear and final testing stands. For the Deventer plant also the gluing robot in the BANS section should be added to this list.

The lines and carriers: A transportation system, consisting of a large conveyor with rolls, on

floor level for Buinen with approximately 70 carriers. For Deventer a rail is suspended from above, also a large number of carriers exist.

Packaging and handling: Automated lines for strapping palletising and buffering of finished

goods. Lifting aids (vacuum lifts and balancers) for lifting the boiler on and from the carrier. For Deventer finished goods are directly transferred to the cross-dock preceding the HBM. Forklifts and transport carts are used for internal transport (milk run). Hand operated pallet carts are used for loading finished goods into the truck (Buinen).

Hand tools: Electrical and pneumatic screwdrivers, hand tools (manual) and torque wrenches Assembly tools: Assembly templates for specific assembly activities for specific boiler types

Testing gear: Ranging from poka-yoke testing gear for electrical wiring, poka-yoke

installations for testing air-tightness, high voltage testing installations to final testing stands.

IS hardware: Kanban scanners, sticker printers, Andon board including help-call and

line-stop buttons. Pick-to-light racks.

BANS (Deventer): This is the only station where parts are actually processed; burners are

glued (sealed) in a gluing robot.

Hot stamp machine: This is a machine for stamping the logos in some of the plastic boiler

covers.

TD workshop: The TD disposes of an all-round workshop, consisting off conventional

machinery for turning, milling, sawing and drilling etc., the machinery is old but still functioning and has very low utilization.

1.2.4 Logistics

As transportation is seen as waste by the BPS lean view, effort is made to keep this to a minimum.

Supply side: Suppliers are pushed to deliver according to Nefits standards, concerning timing,

frequency and packaging. Parts that are outsourced overseas are stored in a hub, by an external supplier and delivered by milk-run 3 times a day. Not all suppliers have conformed to Nefits standards yet. For this reason, in Buinen, a local stocking point exists in the facility called ‘Macro’. This stock is operated by fork-lift. In Deventer the HBM is used for this purpose as well.

Internal logistics: Within both facilities the main transport movement is along the driven line.

The sub assemblies are delivered as close to the line as possible. The parts needed for assembly are picked from FIFO racks, mostly using a pick-to-light system. This means that for a given boiler type red lights indicate which parts to pick. The parts are stored in the racks in standard boxes (bins). When an associate picks the last part from a box, he takes the box inclusive the kanban card, he then puts the empty box on a cart for empty boxes and scans the kanban card. The next box on the rack slides forward. When the kanban card is scanned, a signal is received by the local milk run and by the supplier. The internal mil run loads a box of the specific part type, from incoming goods on the milk run cart, to be delivered on the next round, and to be loaded into the rear side of the concerning rack near the section.

Finished Goods: Buinen; when the boilers are boxed, palletised and strapped, they are

counted as Finished goods. The finished good inventory in Buinen is limited to the size of the pallet buffer, which can hold up to 46 pallets. This is approximately one truckload. Before the pallet buffer is full, a truck will load the pallets and ship the load to Deventer where it will be unloaded in the cross-dock, from where it is either stored in the HBM or shipped to an external party.

Deventer; Like in Buinen, when a boiler is boxed, palletised and strapped it accounts for

finished goods. When strapped, the pallets are automatically transferred to the HBM through a pallet conveyor.

The HBM is a fully automated warehouse, where three unmanned cranes, pick up pallets from the pallet conveyor and assign them to free space in high-racks.

1.2.5 Planning and Control

Scheduling and sequencing: The production schedule is, apart from exceptions, frozen four

on prognoses with a one year sliding horizon. This schedule then, partly manual, levelled subject to takt time, product mix and quantity. In order to take full advantage of schedule levelling over times need to be sufficiently short. In Nefits case there are no change-over times at all. The production schedule is, besides exceptions, frozen three weeks in front.

Shop floor control: The pace maker in the process is the driven line, this takes care of the

movement and timing of products through the process. Provided that nothing goes wrong, the cycle time of the process is fixed by the line. So from the moment an order is started on the line, the finishing time is already fixed. Associates in line-assembly are expected to finish their tasks on the line before a certain mark on the floor. If the carrier passes this mark and the tasks are not yet completed, the associate has to pull the ‘help call' cord, which signals the mentor who will arrange immediate assistance. When the carrier crosses a second mark, and for some reason the tasks are still not completed, the associate pulls the line stop cord, which will stop the entire line until the problem is solved. Help calls and line stops are visualized on large screens in the facility, indicating the section the problem is in, and in case of a line stop, the duration of the stop. This visual system is called andon and also shows the planned production, actual production, and the total duration of all the line stops of that day.

For the sections working on the sub-assemblies, the pace is determined indirectly by the line as well. A screen at the picking racks releases an order to produce a sub-assembly just a few boilers in front paced by the same takt and the same sequence the line is running on (as shown in Figure 3 of paragraph 1.2.2). When for some reason sub-assembly production falls behind, first a ‘help call’ is used and when the problem can not be solved in time a line stop will result. All further actions are authorized by the kanban systems explained earlier at internal logistics.

Figure 7. The Andon board

Quality control: In sections 7 and 8, after most essential assembly steps have been performed,

Throughout manufacturing process quality is controlled through ‘jidoka’and ‘poka-yoke’ systems. When faults are detected every employee has access to ‘help call’ and ‘line stop’ controls. When a problem can not be fixed within time by extra assistance, the line has to be stopped. The source of the problem will ideally be traced, recorded and solved before production can be started. An escalation plan exists classifying problem-types and linking required action for specific problem-types. The escalation plan also shows required action in case of a line stop. The level of action is proportional to the severity of the occurring problem, in case of a line stop action is proportional to line stop duration. Different forms of action can be taken; the least severe classification is a help-call. When the same problem occurs a defined number of times, the action to be taken is: stop the line and fill in and process a standardised problem solving sheet. In the more severe cases (line-stops of over 15 minutes), plant management has to be informed, and in the most severe case the managing director. ‘poka-yoke’ system means that processes are designed in such way that human errors are impossible. An example of a ‘poka-yoke’ is to design an assembly in such a way that parts only fit in one way. At some points in the pick-to-light picking sequence, a part has to be scanned optically for the right size, before an operator can continue picking.

Figure 8. Escalation model Nefit

house. When quality issues occur in production a member from the quality department is engaged in the problem solving team.

Performance Measurement: As mentioned before the Andon screens show some

performance information visible for everyone. The screen shows actual production versus planned production and information about line-stops. This information is also recorded and in combination with other data, formed into the Overall Equipment Effectiveness (OEE), this is an overall performance indicator for production. Performance is graphically displayed throughout the facility on information boards. The OEE is one of the figures displayed on the management dashboard displayed in the info centre in production. The management dashboard also contains the following figures:

Table 1. Summary of performance indicators

Apart from the dashboard centrally displayed, each department has its own specific targets and performance graphically displayed.

1.3 BPS

Bosch provides its subsidiaries with a guideline on how to set-up and run the production process. This is done through the formulation of the Bosch Production System (BPS), a system based on lean principles integrated into a holistic production concept.

BPS documentation is scattered, by subject, over several different places on the Bosch intranet. Most of the documents are in presentation form, and not easy readable, no comprehensive documentation exists. Because most of the documentation is written in German and translated to English, the documentation has to be read in English as well as in German to understand best what is meant. All in all the BPS documentation is far from perfect and is often inconsistent and unclear.

The BPS is built around goals and principles which will now be described in short(Bosch, 2003):

• Pull system: Production and supply only takes place on customer demand. A flow oriented production, with synchronized logistics, eventually leads to minimum lead times and reduced inventory levels’

• Process Orientation: Overall processes are organized and controlled in a holistic manner, to prevent sub-optimisation of individual functions. The objective is to create “simplified and accelerated production processes to benefit our customers, from the first customer request, to order completion”.

• Perfect Quality: Faultless production with zero-errors is what is aimed for. Error prevention is preferred over error discovery. This is to be realized through preventive measures with fast control loops.

• Flexibility: Emphasis is placed on the ability to adapt quickly to the current

requirements of the customer easily. Both in the installation of machines and in the organization of work. Product variants are implemented quickly and at the latest possible time in the value chain. Mechanical equipment is reusable and oriented towards the product life cycle; new and continuously further developed processes and methods can be integrated in the production system at any time.

• Standardization: Whether it involves proven processes, practical methods or reliable equipment, successful solutions are standardized and adopted, as standardization is an important prerequisite for reliable, but flexible, process sequences. Standards are based on ‘best in class’ cases constantly developed further.

• Waste Elimination and Continuous improvement: Standards that are already implemented form a basis for further improvements. The motto is: “If you don’t go forward, you go backward”, and “nothing is too good to be improved”.

• Transparency: Processes are required to be self explanatory, uncomplicated and direct, in order to achieve objectives and to guarantee continuous improvements.

Transparency creates clarity and an all-round, positive appearance. It also directly reveals the deviations in the production processes, everybody is aware of their tasks and objectives while providing a comprehensive overview for optimised orientation in all areas.

• Associate involvement and empowerment: Everybody has the task of contributing to the success of the production processes competently, and with individual

responsibility. In this way the creativity and extensive know-how of all employees can be utilized for the good of the company.

Table 2. Principles and elements correlation

Deployment of these principles, through use of the different elements, makes it possible to have “The right part, in the right quantity, in the right quality, at the right price, at the right time, in the right place”. Ultimately the BPS has two goals; “Customer Satisfaction and Business Success”, and “Associate Satisfaction”. All elements and principles are believed to contribute to aid these goals.

1.4 TPM

1.4.1 General

TPM is a holistic maintenance strategy initially presented by Seiichi Nakajima in 1984 (translated to English in 1988). According to Nakajima, TPM aims to maximize equipment effectiveness, and establishes a thorough system of preventive maintenance for the entire lifespan of Machines And Equipment (MAE). TPM affects every single employee and is executed in autonomous small group activities by various departments (1998). The main characteristic of the change of roles, when transiting to TPM, is depicted in Figure 9.

Figure 9. Division of tasks (adapted from: Aalders et al., 1995)

The BPS also prescribes TPM as maintenance program. In the BPS the content and form of TPM is described as well as an implementation trajectory.

element is represented as a separate fifth pillar while according to the BPS model training is needed throughout all TPM activities.

Other models exist in literature introducing more pillars for more elaborated functions of TPM. An example of such a system is the model suggested by the Japan Institute for Plant Maintenance (JIPM). This model contains 8 pillars which include, besides the five pillars of Yeomans and Millington, ‘Office TPM’, ‘Safety, health and environment’ and ‘Quality Maintenance’ (Ahuja & Khamba, 2008). Within the BPS the scope of TPM is mainly restricted to plant maintenance. The other 3 TPM elements are supposed to be controlled through other BPS elements (BPS 2003). For this thesis the BPS model is used as a basis for TPM. According to the BPS, TPM is modelled as a four pillar temple, shown in Figure 10. Each pillar resembles a TPM element. The basics of the BPS TPM model will now be described:

Figure 10. TPM Temple according to BPS

1.4.2 Pillar 1

The purpose of the first pillar, ‘Eliminating Main Problems’, is to eliminate failures in equipment currently being operated. And to feed the experiences back to improve current MAE and to design better MAE (Nakajima, 1988). This process resembles a continuous improvement cycle making it compliable with the entire BPS as well as with Nefits production system.

1.4.3 Pillar 2

The second pillar, ‘Autonomous Maintenance’ (AM) is based upon maximizing the potential of all associates. The operator, working the machine every day, has valuable knowledge of the behaviour and shortcomings of the MAE. According to this pillar “all routine activities to maintain systems, are done as teamwork, by operating personnel, acting on their own initiative, once they have been trained accordingly” (BPS, 2007). Service activities and minor repair activities are meant by ‘all routine activities’. Through AM the sensitivity of the system increases so that timely service and repair are a matter of course. Faults are recognized quickly and remedied quickly and safely.

1.4.4 Pillar 3

an IT system. Existing maintenance data is analysed and in combination with experience, maintenance activities and intervals are set.

1.4.5 Pillar 4

The fourth pillar is directed at ‘TPM Appropriate MAE Design’. By involving TPM specialists from production, planning, service and quality assessment in systems engineering, the ease of maintenance, accessibility and user-friendliness in new MAE is increased.

For each pillar five implementation steps are defined, paragraph 4.1.1 shows an overview of these steps. At this point the implementation steps will not be elaborated upon any further. Literature states that a typical TPM implementation takes between three to five years (Wang & Lee, 2001). Chapter 5 deals with the elements of implementation in depth.

1.5 Maintenance at Nefit

1.5.1 General

Maintenance at Nefit is done by operators as well as the maintenance staff. Maintenance consists of corrective and preventive maintenance and is coordinated by the Technical Support department (‘Technische Dienst’ TD). Preventive maintenance is done during operation when possible and planned in the lunch break or weekend when necessary.

Technical Support Department; The Technical Support Department consists of one employee

at any time working on fault clearing service called ‘storingsdienst’, and three people working shifts. Apart from fault clearing, the TD is responsible for preventive maintenance, facility maintenance and tool making. Group chief of TD and senior TD are responsible for planning, scheduling of preventive maintenance and procurement of materials.

Organization of maintenance; For the organization of maintenance a Computerized

Maintenance Management System (CMMS), called BopV5 (‘Bedrijfsmiddelen Onderhouds Programma’) is used. All MAE at Nefit are registered in BopV5 as objects. For each object or object group, attributes like manufacturer, location, date of purchase, maintenance required, but also documents like calibration manuals and reports, can be defined. The Bopv5 system also handles order releases of maintenance jobs and other support jobs to be executed by the TD. Some preventive maintenance jobs are executed by production. For these jobs BOPv5 releases orders which are sent to the respective mentor. Whenever a job is finished a record is made in BOPv5.

Corrective Maintenance; Corrective maintenance is carried out by the employee on fault

clearing service. When a problem occurs at production fault clearing is called on the urgency line. The technician on duty comes to fix the problem, if necessary, assisted by a colleague. Afterwards the data is recorded in BopV5.

In the following sub-paragraphs the current situation will be described per pillar in the perspective of the TPM pillars shortly described in 1.4. This situation will be the starting point for analyses and the change process.

1.5.2 Pillar 1

rate. A visual representation of the OEE factors is shown in Figure 11. The OEE figure is then calculated by taking the product of the three factors:

OEE = Availability (%) * Efficiency (%) * Quality rate (%) Where:

Availability = (Loading time – Downtime) / Loading time

Efficiency = Processed amount / (Operating time / Theoretical cycle time) Quality rate = (Processed amount – Defect amount) / Processed amount

Figure 11. OEE factors

In the case of Nefit, efficiency losses are underrated, as every loss not causing exceeding the takt time remains un-revealed because it is not captured in the OEE figures. The takt time is used as theoretical cycle time, it is not sure that shorter cycle times are impossible. As long as the actual cycle time remains within the takt time, performance losses will not be noticed by the OEE measurement. Through MTM time studies (Method Time Measurement analysis) the cycle times of the different product families per section are known. For section 8 the measured cycle times are given in Table 3. This means that actual average cycle times can vary between 311 and 346 seconds subject to product mix. For section 8 (final testing) the takt is always 360 seconds, because the takt of the line is determined by the number of operating final testing stations divided by 360 seconds. Efficiency losses for these stations is always between (360-346)/360 = 4% and (360-311)/360 = 13,6 % subject to product mix. When evenly levelled an efficiency loss of 8,8 % occurs and remains unnoticed by the OEE calculation. These losses, caused by idling, are not captured in OEE figures, and will for this reason not be identified for problem solving when analysing OEE. Looking at OEE history we can see that the efficiency factor is 100% or higher on average. Correcting a 95% OEE with 8,8 % efficiency loss, results in an average OEE of only 87,6 %, which all of a sudden leaves significantly more room for improvement.

Table 3. Cycle time per family for section 8

Machine downtime causing production losses are together with the availability of materials and employees captured by the availability factor. Every time the line is stopped, the cause is recorded. An availability of 98 % is considered world-class and is achievable partly due to the low tech manual assembly character of production and partly because change over times do not exist. Also here a note must be made that in some situations ad hoc solutions are preferred over a line stop, hiding the consequences from the OEE figures. If for example, in section 8, a boiler test fails due to technical reasons the boiler is often lifted from the line and tested on a different testing stand, instead of stopping the line.

OEE figures are discussed weekly by plant management, together with the production chiefs. Disturbances are discussed and when assumed appropriated, standardized problem solving is applied. This goes for availability, efficiency and quality issues. During this weekly meeting all other issues (e.g. supplier issues) that occurred during the week are discussed as well. When found appropriate measures are planned. Possible measures are standardized problem solving techniques or ad hoc problem solving.

Technical management, apart from the weekly meetings, derives opportunities for improvement from e.g. the availability losses graph. From there on further analysis, partly using data, partly common sense is used to identify opportunities for improvement.

Figure 12. Daily OEE graph

Figure 13. Availability loss per department

Figure 14. Availability losses with technical cause categorized per cause

These are examples of a weekly OEE graph, an availability graph divided over departments, and a graph dividing the availability losses of the technical departments responsibility to categories.

Failure data is also recorded. After TD has fixed a breakdown, data is recorded in BopV5, using the breakdown form shown below in Figure 15. Though this data is recorded is not yet systematically used for improvement analysis. Data exists for the last 4 years. These data is recorded separately from OEE data, these systems are not integrated. The BopV5 data is not systematically analyzed. Last year Bram Hoekstra, bachelor student TBK from the University of Groningen conducted research partly concerning failure data analysis (Hoekstra, 2008). He concluded that reliable analysis was hampered by the following factors;

• The data being recorded does not include weather the failure lead to actual downtime of the production line.

• The duration (cost) of the repair caused by a failure is not (accurately) recorded.

• Many objects, to which failures are linked in the BopV5 system, are in fact a composition of many parts or subsystems on which level failures occur. If for example a failure in the main conveyor of the production line occurs, no clear distinction is made in which part of the line or in what kind of part the problem occurred.

maintenance. Before, this or a similar analysis, can be used for determining PM intervals, the issues above should be dealt with. Even then analysis should be used to support decision making (Hoekstra, 2008).

Figure 15. Failure form

Cost of maintenance is not regularly recorded and evaluated in the current situation. Plans exist to do so in the future. Other shortcomings of the current system opposed to the BPS description of TPM is that information is not MAE specific and visualized on the spot, facilitating decentralised problem solving. Suitable, situation specific, performance indicators should be chosen and visualized on the spot. Action taking on the basis of standardized improvement and problem solving methods should be introduced and trained. Procedures and responsibilities in case of deviation from the defined standards should be defined and integrated with overall plant procedures.

1.5.3 Pillar 2

In both facilities 5S programs are in use, so cleaning activities are mainly executed by operating personnel, most of the time these activities are done by the extra pool workers. The basic idea of autonomous maintenance is that the operator services his/her own MAE.

Buinen: In Buinen, apart from the 5S activities, the BopV5 system generates maintenance

orders which are transferred to the mentors, who assign these orders to extra pool workers when available. So it is not really autonomous as meant by the BPS. Often, there is no time to execute the tasks so the orders return to TD without being executed.

Deventer: In Deventer some autonomous maintenance is also done but no formal system

exists to assure execution and monitor results. The weekly changing of the filters on the final testing stands is an example of these tasks. In practice it is the mentor who is responsible of execution assurance and monitoring.

means that the line should be stopped periodically for TPM activities. All actions should be documented and standardized according to BPS standards.

1.5.4 Pillar 3

Planned preventive maintenance is already done within Nefit using the BopV5 system. PM intervals are mainly determined using common sense, no formal analysis exists. For the reasons listed in 1.5.2, the analysis tool developed by Hoekstra is not yet applicable for determining PM intervals. Before, this or a similar analysis, can be used for determining PM intervals, the issues above should be dealt with. Even then analysis should be used to support decision making (Hoekstra, 2008 p.34-35).

The current situation of PM does not comply with some of the BPS requirements for pillar three; no standardized work documentation exist for the planned preventive maintenance tasks, the computerized maintenance management system used, BopV5, does not support all the features demanded by the BPS element description. For this matter the possibilities of the current system regarding these features, should be assessed as well as the need for the extra features demanded by the BPS.

1.5.5 Pillar 4

When acquiring new MAE, a list of demands is drawn up, in the early stages, before manufacturers or suppliers are contracted. In this list also issues of maintainability are included, though no formal TPM directed list exists. There is no formal structure that brings TPM specialists from all relevant departments together. There are no (trained) TPM specialists present in different departments. Although all of the formal structures do not exist, the way things go, partly resembles what the BPS describes for pillar four.

1.5.6 Training

2. RESEARCH DEFINITION/METHOD

2.1 Background

Both the Deventer and Buinen facilities scored very well on the 2008 the BPS assessment, in which Buinen came out best thermo-technical factory and Deventer second best. Despite the audit success both facilities scored below average on their maintenance programs. As mentioned earlier the BPS prescribes TPM as maintenance strategy. During the second half of 2008 Nefit initiated the first steps towards TPM implementation by assigning a colleague student, Bram Hoekstra, from the University of Groningen, to do TPM directed maintenance research. This resulted in a report on TPM and TPM implementation but which focused on historical failure data analysis to determine intervals for preventive maintenance (Hoekstra, 2008). Parts of this study are used as a reference for this research. Still no actual actions directed towards TPM implementation were made. By the end of 2008, Nefit management decided to assign another student to TPM. This time, one of the results had to include actual TPM implementation efforts by means of a pilot case. The primary focus the project was to be on autonomous maintenance, pillar 2, though along the way, it became apparent that focus on the first and second pillar was more appropriate. Results of this project are the application of pillar 2 activities on two pilot areas and this thesis.

Nefits current maintenance organization, does preventive maintenance, to some extent uses autonomous maintenance, and seems to do reasonably well in adapting all MAE to be used effective, given the high OEE figures, so a few of the TPM elements are already applied. The problem is that these TPM elements are not applied according to the Bosch standards and that these elements may not be used to their full potential.

To be able to assess whether TPM, or any maintenance strategy is used to its full potential, first the effects on performance have to be considered with appreciation for the specific circumstances. A circumstance in the case of Nefit might be the low tech character of the production process. Choices will have to be made regarding what, and how far the TPM elements have to be adopted. What adaptations have to be made and what difficulties are to be expected? What prerequisites have to exist or have to be created? These are questions this research aims to answer.

2.2 Research model

2.2.1 General

Figure 16. causal conceptual model

As mentioned in 2.1, Nefit, the customer for this research decided to change its organization of maintenance towards BPS norms, which means implementation of TPM. The purpose of this research is to design this new TPM maintenance organization, focussing on the first and second pillar of the TPM model, and to advice on how to implement this. In the same time also actual implementation had to be started. The actual implementation projects are valuable sources of information for this research. The TPM organization and the implementation process are seen as independent variables in the model. This means that by changing these variables, through the assumed relationships, the other (dependent) variables will change. The independent variables must be designed in such a way that the dependent variables change in the most favourable way possible. In this paragraph the variables and their relations are described.

2.2.2 Elements of the research model

In this sub paragraph the elements in the research model are defined:

TPM

The TPM maintenance strategy is represented by the TPM box in the centre of the model. The TPM maintenance strategy consists of organisation, procedures and standards which have to be determined, divided into pillar one and pillar two. TPM organisation defines what organisational structures are needed for each pillar. The procedures define what has to be done, and the standards define how things are done. The TPM maintenance strategy is based on the BPS TPM model, though it has to be tailored to fit the specific situation at Nefit. Choices regarding what TPM elements have to be incorporated, and how far these elements have to be implemented, have to be made. The current maintenance organization is seen as a starting point for the change towards the future state of the TPM maintenance strategy.

Implementation process

Production system

With production system in the model ‘the way things go around production’ is meant. It consists of existing culture, procedures, organizational structures and processes. An overall description of the production system is already given in paragraph 1.2. The takted line, different maintenance needs between different parts of production, integration with procedures like 5S activities, process confirmation and continuous improvement are elements of the production system, assumed relevant to TPM and TPM implementation as discussed later on.

Overall performance

Eventually the change to the new maintenance organisation has to lead to increased overall plant performance. For this research overall plant performance consists of all elements of the management dashboard that are likely to be influenced by TPM implementation. An overview of the management dashboard is already given in 1.2.3. Elements that are assumed relevant in the research model are: the OEE figures, transparency. Apart from the dashboard also effects on the cost of maintenance and the BPS score are analysed.

2.2.3 Relations between the elements

The relationships assumed in the research model are explained next:

TPM and the Production System

Elements of interaction with the production system are; planned downtime in the takted line, shared standards and procedures with production, integration with the 5S policy.

For this reason, in order to be able to design, we have to analyse, in what way, the TPM organization, interacts with Nefits production processes. Also we have to question what we want to achieve, and what we expect to achieve by implementing TPM.

Production process and overall performance

The question of how the relationship between the production process and the overall plant performance actually works is not thoroughly addressed in this research. More interesting is how this relationship is influenced by the change of maintenance organization. This question is explored further in the moderating effect of TPM on production performance.

The moderating effect of TPM on production performance

Through a decrease of downtime, quality losses and other variability, TPM is believed to enhance overall performance of the production system. The general purpose of maintenance is to support production, maintenance does generally not produce any goods at all. For this reason the main contribution of the TPM maintenance organisation is modelled as a moderating effect on the relation between production and overall production. The most tangible effects TPM promises to brig about is an increase of OEE. Besides the increase of the OEE, TPM is also believed to have effects that are hard to measure. These effects indirectly affect overall performance, though this relationship is unclear and hard to measure. Customer involvement, transparency of processes, learning and growth and continuous improvement are some of these effects. Though these effects are generally believed to contribute to overall performance these relationships are unclear and hard to measure.

Direct effects of TPM on overall performance

Implementation process and TPM

The implementation process is assumed to influence the eventual success of the TPM maintenance organisation. Many companies try to implement TPM but fail, reasons for this can be found in literature, but also company specific elements will be looked into.

2.3 Research Questions

2.3.1 Main question:

How should, given the characteristics of the production process and with possible direct and indirect effects on performance taken into consideration, the future state TPM strategy, as far as pillar 1 and pillar 2 are concerned, be designed, and what should, in order to realize this design in compliance with the Bosch standard on this subject, the implementation program look like?

2.3.2 Sub questions:

1. According to BPS standards, what should the maintenance organization look like? 2. How does the TPM maintenance organization interact with the production system,

what are the consequences, and what measures are desirable?

3. What effects are to be expected from TPM implementation, what effects are desirable and what effect does this have on the design?

4. To what extent does the current situation need to be changed?

5. How should this be implemented? Planning? Organization? Procedures? Measurement of achievements?

2.4 Research Goal

The research goal for this thesis consists of the following parts:

Practical; The customer, Nefit plant management, is helped, through action research, consisting of; the implementation and evaluation of pilot cases, in both facilities, resulting in a proposed, standardized design for application of the first and second TPM pillars on the shop floor. Accordingly a roadmap for future roll-out is proposed.

Economic; Expected effects resulting from TPM implementation are listed and discussed, some quantitative and some qualitative, resulting in a design aiming to increase performance of Nefits production processes in line with company strategy and vision.

2.5 Method

2.5.1 General

the research model. In Figure 17 the research model is shown adapted to show the structure of the thesis and the role of the pilot projects.

Chapter 5 Chapter 4

TPM

Incl. targets and goals

Production system Implementation process Chapter 3 Pilot projects

Figure 17. Strucure of the thesis

2.5.2 Chapter 3

In chapter 3 these pilot projects are introduced; the both pilot areas are described and the main activities conducted during my time at Nefit are described as well as the most relevant results. In the actual analysis in chapters 4 and 5 reference is often made to these projects whenever relevant.

2.5.3 Chapter 4

Chapter 4 deals with analysis regarding the future state of Nefits maintenance strategy, and of the expected effects as a consequence. It aims to answer the first four sub questions formulated in 2.3.2.

In paragraph 4.1, the guidelines and standards supplied by the BPS element description are explored and summarized. Also derived from BPS documentation and literature a proposed TPM organization is presented. Paragraph 4.1 aims to answer the first sub-question. In

paragraph 4.2, the second sub question is treated by looking at the interactions between the

production system and the maintenance system. In the sub paragraphs the interaction of TPM with some characteristics of the production system are described. Accordingly solutions to the anticipated problems are proposed. In paragraph 4.3, relating to the third sub question, respectively the moderating, direct and indirect effects of a change towards TPM are analysed. First the company vision, mission, strategy and targets are described and discussed in relation to TPM. In the following sub paragraphs, the TPM relevant expected effects are described and quantitative or qualitative analysis of their magnitude is presented. In

paragraph 4.4, chapter 4 is concluded with a future state design for the TPM maintenance

2.5.4 Chapter 5

In chapter 5 the implementation trajectory is treated. After an introduction in 5.1, the BPS is consulted in 5.2. A brief review from literature is given in 5.3 followed by a recommended structure in 5.4.

2.6 Restrictions and assumptions

A choice is made to base the TPM model used on the model of Yeomans and Millington (1997) as this version closely resembles the BPS version of TPM as firmly recommended by Bosch.

The study is restricted to the production departments, as this is the responsibility of the customer. Though the direct contact persons within Nefit are the heads of both technical departments, the main customers for this research are both plant managers, as they hold responsibility not only for maintenance performance but for total production performance, and TPM aims to improve this.

As the primary focus of the study is on the production department, and because of time restrictions (typical TPM implementation takes between 3 and 5 years), the implementation focuses and is limited to the first and second TPM pillars. It is assumed that implementation of the third and fourth pillars can be done separately in a later stage of implementation. This is because these pillars are both the responsibility of the production department, and the third and fourth pillars are the responsibility of the maintenance department and top management. The sequence of implementation according to the BPS is also beginning with the first two pillars together followed by the third and fourth separately (BPS, 2007).

Because of the similarity between the production processes in Buinen and Deventer, the outcome of the analysis on the Buinen plant is assumed to be applicable also for the Deventer plant. Because of time restrictions and to minimize travel, the choice is made to base the research mainly on the Buinen plant. In order to keep the TPM implementation in both facilities in pace, actual implementation steps, the pilot projects, are carried out in both facilities.

A major part of this research consists of two cases of TPM implementation at Nefit; one in Buinen and one in Deventer. The time available for this research was four months, in which the pilot project as well as the main research had to be conducted. Because of the time limitation, the run-length of the research was restricted to four months, the main research and the pilot cases had to be conducted simultaneously. As a consequence, some results from the pilot cases were only available near the end of the research. For this reason, the conclusions from the pilot cases are used mainly in the later parts of the research. Results from the first part are checked with the cases and experience is fed back into the design phase.

3. THE PILOT PROJECTS

3.1 General

Basically the pilot projects consisted of; an initial TPM workshop, evaluation of the results, standardisation of the program, a TPM assessment of the plants and formulation of a plan for further roll-out. This thesis can be considered the conclusion of the pilot projects. Because of the differences between the both pilot projects they will be treated separately when necessary. In the later parts of this thesis there will be references to the pilots when appropriate.

For the pilot projects teams were formed. The members of the teams represent all layers of production related employees, in Deventer also the assistant plant manager joined the team, who was relatively new to Nefit. Unfortunately, no group chief was able to be present in the Deventer team. The Buinen TPM pilot team consisted of: The head of the technical management department Buinen (who is my Buinen supervisor), a group chief, a senior TD employee, a TD mechanic, a production mentor, and a section 8 operator. The Deventer TPM pilot team consisted of: The head of the technical management department Deventer (who is my Deventer supervisor), the assistant plant manager, a senior TD employee, a TD mechanic, two production mentors, and a burner assembly station operator.

3.2 The Buinen testing station

The final testing stand: For the Buinen pilot area, one of the six identical final testing stands was chosen. The final testing stand is one of the technically more complex stations in the Buinen production plant. The testing station operates in the takt of the production line. It actually dictates the takt; the number of testing stations in operation determines the takt of the line, as the cycle time of the testing station is fixed at six minutes. A picture of a final testing stand is shown in Figure 18.

The final testing stand is located as on of six parallel stations near the end of the driven line. At this station merely completed boilers are tested in a simulated end user situation. The boiler is connected to a gas lead, cold tap water (supply), hot tap water, hot heating lead and a return heating lead. Also the boiler is connected to 220 V power lead. When the carrier arrives at the testing stand, the boiler is connected, and switched on. The boiler will start bleeding the air from the system before igniting. While the boiler is bleeding the operator performs a serried of tests including visual tests, checking for gas leaks using a detector, checking the connections of the gas system using a torque wrench and sealing the connection, and checking the gas throughput by connecting a testing hose to the boiler. When the boiler starts burning the operator check the values of the meters and signs a checklist. The boiler chipset is reset and the some stickers put on before the carrier is released.

The final testing station consists of:

- A part of the track of the driven line, making the carriers pass the station. - The testing stand itself.

- The surrounding of the testing stand.

The testing stand itself consists of:

- Two metal casings with a number of controls and gauges. The casings contain switches and pumps for the hydraulic system but also high voltage parts. The casings can for this reason only be serviced by authorised personnel.

- Compressed air hose.

- Five hoses with snap on connectors (gas, hot/cold tap water, hot/return heating water). - A water drainage system to discharge used water into the sewage pipe above the

station, consisting of a tank, a pump, a clack valve and pipelines.

Time for the operator to execute the TPM tasks was made through making a pool worker operate the testing stand while the operator performs the required tasks.

Figure 18. Final testing stand Buinen

3.3 The Deventer burner assembly station

Burner assembly gluing station: The burner assembly station was chosen for the Deventer pilot project. At the time of the research some quality issues attracted a lot of attention, from inside Nefit but also from Bosch, to this station, so it was a logical choice to start the TPM implementation on this area.

The burner assembly station consists of:

- The gluing robot, an x-y robot applying a line of glue on a surface. The robot is driven by electric servo motors.

- Glue pump, connected to the robot and to a barrel of glue. The amount of glue is controlled by a pressure regulator.

- The conveyor belt, transferring burner frame assemblies in and out of the robot. - Weighing scales for measuring the quantity of glue applied.

- A frame cutter for removing superfluous pieces of metal, resulting from the production process, from the frame. This cutter is only used for some burner types.

- A pneumatic press for tagging a quality mark into some types of assembled burners. - Inventory racks for, assembly clips raw material and finished, drying, subassemblies.

Figure 19. Overview of pilot area Deventer

3.4 The initial workshops

The initial workshops consisted of three 5 hour sessions in which the following activities were planned. Reference is made to examples of the generated documents in the appendices.:

- A basic TPM introduction was presented; a 60 minutes presentation on TPM theory, OEE calculation at Nefit and the need for TPM.

- Introduction to the pilot area; a TD mechanic shows the team around the pilot area, explaining common maintenance and general workings of the machine to the team. All team members can add issues relevant from their perspective.

- Listing of maintenance needs and possible improvements; after the introduction the team walks around the pilot station, placing numbered red and green stickers. Green stickers for points of maintenance, and red stickers on possible improvement points. - Creating action sheets; for every numbered sticker an action sheet was created,

showing a photographic and/or sketched representation of the point at hand (Appendix A). For the red stickers a solution was made and for the green stickers the parameters of the task was listed. For the initial maintenance frequencies, common sense was used, sometimes supported by historical failure analysis.