Validation project

(1)

1

Validation project

Thesis MSc. Technology Management

Faculty of Economics and Business

University of Groningen

April 2013

Name: Zinzy Hordijk

Details: Gedempte Zuiderdiep 37a

9711 HB Groningen +31 6 1364 5368

zinzyhordijk@hotmail.com

Student number: 1685899

Supervisor University of Groningen: Dr. X. Zhu Co-assessor University of Groningen: Dr. J. Riezebos Supervisors Fokker Aerostructures B.V. Rutger van Galen

(2)

2

List of Figures

Figure 1: DMAIC method ... 8

Figure 2: Pareto production phases ... 9

Figure 3: SIPOC map, subs ... 11

Figure 4: NC rate subs, Jan 2012-Aug 2012 ... 12

Figure 5: Input and output variables ... 16

Figure 6: Amount of NC's incl. and excl. TAV ... 16

Figure 7: Actual NC rate... 17

Figure 8: Stratification factors ... 18

Figure 9: Current process flow ... 19

Figure 10: Pareto - process steps ... 20

Figure 11: Pareto - production phases ... 21

Figure 12: Run chart, NC's per week ... 23

Figure 13: I-MR chart, NC's per week ... 24

Figure 14: I-MR chart, no drilling ... 25

Figure 15: Process steps for drilling ... 25

Figure 16: Fishbone diagram ... 26

Figure 17: VSM for drilling ... 27

Figure 18: Pareto – causes... 29

Figure 19: Scatter plot 3.2mm ... 31

Figure 20: Scatter plot 4.8mm ... 31

Figure 21: Regression, 3.2mm without ... 33

Figure 22: Regression, 4.8mm with ... 33

Figure 23: Probability plot, 3.2mm ... 35

Figure 24: Probability plot, 4.8mm ... 35

Figure 25: Probability plot, ovality ... 37

Figure 26: Fishbone diagram, subs ... 43

Figure 27: Main effect plot, subs ... 45

Figure 28: I-MR chart, subs ... 46

Figure 29: I-MR chart, subs ... 49

(5)

5

List of Tables

Table 1: Six Sigma change agents and their characteristics (Pyzdek, 2003) ... 8

Table 2: Short overview of DMAIC method... 9

Table 3: Team members ... 10

Table 4: Schedule, Subs ... 13

Table 5: Potential causes/X's ... 27

Table 6: Potential causes and severity, occurrence and detection ... 28

Table 7: Potential causes and their category ... 28

Table 8: Categories and the total points ... 28

Table 9: Standard deviations 3.2mm and 4.8mm ... 36

Table 10: Seven wastes (Brook, 2010) ... 38

Table 11: Tools 'Improve' phase ... 40

Table 12: Priorisation matrix ... 42

Table 13: PICK chart... 42

Table 14: 5S (Brook, 2010) ... 46

(6)

6

List of abbreviations

ANOVA Analysis of variance

BB Black Belt

BNVA Business non-value added

CEDAC Cause and Effect Diagram with Added Cards CTQ Critical to quality

CpK Process capability index

DMAIC Define, Measure, Analyse, Improve, Control DOE Design of Experiments

DPMO Defect per million opportunities

EAV Own NC’s (translated from Dutch abbreviation) Gage R&R Gage Repeatability & Reproducibility

GB Green Belt

GE General Electrics

HS Horizontal stabilizer I-MR Individual moving range KPI Key performance indicator LCL Lower control limit

LE Leading edge

MSA Measurement system analysis

MR Moving range

NC Non conformance

NVA Non-value added

PpK Process performance index

SC Scrap conformance

SIPOC Supplier, Input, Process, Output, Customer

TAV Suppliers NC’s (translated from Dutch abbreviation) TPD Technical process description (instructions)

TQM Total quality management UCL Upper control limit

VA Value added

VF Vertical fin

VOC Voice of the customer

VSM Value stream map

(7)

7

Chapter 1 Introduction

This document is part of a Master thesis where an investigation is done on how to best execute a Six Sigma project to increase the yield of quality for an assembly process. Below, an introduction to Six Sigma is given and a guide on how to use this document.

1.1 Introduction to Six Sigma

Six Sigma can be defined in many different ways (Eckes, 2003; Furterer & Elshennawy, 2005; Larson, 2003; Pande et al., 2000) However, all sources agree on the fact that Six Sigma is a methodology that provides businesses with the tools to improve the capability of their business processes (Yang & El-Haik, 2003).

Six Sigma oriented in the 1980s at Motorola, where a highly skilled and trained engineer, who knew statistics, began to study the variation in the various processes within the company. He saw that too much variation in any process leads to ineffectiveness (Eckes, 2003). After significant changes were successful within the company, General Electrics (GE), too, adopted the method to improve their company’s performance. By the end of 1995, GE had decided to make Six Sigma a corporate-wide initiative and less than two years after the initial application GE had generated over $320 million in cost savings (Eckes, 2003). These results are one of the reasons that businesses grew interest in the Six Sigma way of improving and why it is becoming more popular.

Now, what is it exactly? A company’s performance can be measured by the Sigma level of their business processes, where Sigma, , is a letter in the Greek alphabet used to measure variability in any process (Pyzdek, 2003). It refers to six standard deviations and means a process performance of no more than 3.4 defects per million opportunities (DPMO) (English, 2004). The purpose of this is to increase performance and decrease performance variation, which will lead to defect reduction and improvement in profits, to employee morale and quality of product, and eventually to business excellence (Yang & El-Haik, 2003).

So, why use Six Sigma methodology to increase quality, while there are so many others ways to do this? The reason is that Six Sigma is relatively simple, unlike other quality improvements programs for example Total Quality Management (TQM), where over 400 tools and techniques can be applied (Pyzdek, 2003). Also, the direct effect of Six Sigma is to simply save money, while TQM aims for relativity more complex effects, for example achieving customer loyalty and improved performance (Andersson, et al. 2006).

(8)

8

Table 1: Six Sigma change agents and their characteristics (Pyzdek, 2003)

Change agents

Characteristics

Champions High-level individuals who understand Six Sigma and are committed to its success

Sponsors/Process owner

Sponsors are owners of processes and systems who help initiate and coordinate Six Sigma improvement activities in their area of responsibility. The two functions can be done by the same person, but also by separate people.

Black Belt (BB) Technically oriented individuals held in high regard by their peers. They should be actively involved in organisational change and development

Green Belt (GB) Project leaders capable of forming and facilitating Six Sigma teams and managing projects from concept to completion. Usually they are assisted by a Black Belt (five to seven per Black Belt)

Master Black Belt The highest level of technical and organisational proficiency. They must be able to assist Black Belts and possess excellent communication skills.

The method by which Six Sigma is executed is the DMAIC method and can be seen below (Fout!

erwijzingsbron niet gevonden.).

Figure 1: DMAIC method

DMAIC is used to improve existing business processes (Schaffer, 2007) and de Mast and Bisgaard (2007) state that the five-step method is the most prevalent Six Sigma method used.

(9)

9 Table 2: Short overview of DMAIC method

Short overview

D

efine

Select the process that needs to be improved.

M

easure

Translate the process into quantifiable forms, collect data, and assess the current

performances.

A

nalyse

Identify the root cause of defects and set goals for performance.

I

mprove

Implement and evaluate changes (solutions) to the process to remove the root

cause of defects.

C

ontrol

Standardize solutions and continuously monitor improvements.

1.2 Introduction to this document

The project aims to reduce the number of NC’s for the G550 empennage assembly process, so that the yield of quality can be increased. During the course of this project, after the ‘Define’, ‘Measure’ and ‘Analyse’ phase, several areas were identified where most NC’s were created. This area was the HS In Jig and VF In Jig, with almost 40% of all NC’s measured. The second largest phase was, surprisingly, the Sub Assembly process (Sub Assy’s or Subs) with 20%. (Figure 2) and this is the area that will be used in this validation. The column in light blue is the number of NC’s that have been solved already, with the help of ordering a new type of tool.

Per phase, only the tools mentioned in the conclusions will be used to execute that phase, to verify whether the conclusions drawn are accurate.

Figure 2: Pareto production phases

(10)

10

Chapter 2 Define

2.1 The Define phase

The ‘Define’ phase starts as soon as a problem is identified in an organisation. This phase is the first in the DMAIC method and helps to clarify the understanding of why it is a problem, before investing time and money in commencing a project (Brook, 2010). Beware, that Six Sigma starts with problems only, and the ‘Define’ phase is one which only focuses on the problem; causes and solutions will come later in the process. The purpose of the ‘Define’ phase is to have the team and its sponsor reach agreement on the scope, goals, and financial and performance targets for the project (George,

et al., 2005). Below a list of tools can be seen that will be executed.

Team members and responsibilities Problem/opportunity statement Process map SIPOC

CTQ/define Y Scope VOC/customer requirements Business case Goal statement Key stakeholders Schedule

2.2 Validation of the ‘Define’ phase

This section will attempt to make a validation for the ‘Define’ phase. To do so, only the top ten tools and activities found in the conclusion of the ‘Define’ phase (in the document “A Standardised Way of Applying Six Sigma to Increase the Yield of Quality of a Production Process”) will be used. Each of these tools will be executed and shown to the so-called review board, existing of experienced Green and Black Belts. They will help determine whether it is possible to pass the ‘Define’ tollgate with the presented tools and verify whether a tool was essential or not. The exact feedback per person can be found in the other document too.

2.2.1 Team members and responsibilities

Table 3: Team members

Role

Function

Days/week

Project leader Responsible for project 4 days/week

Quality engineer Responsible for the quality of products On call basis Program manager Responsible for the G550 empennage On call basis Engineer Involved in the design of the G550 empennage On call basis Production planner Responsible for planning G550 empennage On call basis Team leader Experienced person in charge of assembly workers On call basis

2.2.2 Problem/opportunity statement

(11)

11

this problem is €23.100 per year. This does not yet take into account any rework, repairs, waiting and/or an increase in throughput time which will increase costs even more.

2.2.3 Process map SIPOC

Figure 3: SIPOC map, subs

2.2.4 CTQ/define Y

Y = (#NC + #SC)/1000 production hours/month

2.2.5 Scope

The following aspects of the process will be considered in the scope of the project:

 NC’s/SC’s created in the sub assembly process

 Centerbox

 Feedback to other department on faulty products discovered in the sub-assembly, but not created by sub-assembly

 Improvements useable for the G650 should be communicated

The following aspects of the process will be left out of the scope of the project:

 NC’s/SC’s created by other departments

 Antenna

(12)

12

 Other lead time reduction of assembly processes

 Improvement of the administrative side of the NC/SC processes

2.2.6 VOC/customer requirements

Two people were interviewed, namely the quality engineer of the G550 program and the program manager responsible for the entire G550 empennage program. Both addressed the same reduction of the NC rate to 2,7 (Figure 4).

Figure 4: NC rate subs, Jan 2012-Aug 2012

(13)

13

2.2.8 Goal statement

The primary goal of the project is to empirically test whether the tools and activities found in the document “A Standardised Way of applying Six Sigma to Increase the Yield of Quality of a Production Process”) are the most important tools to execute when passing the different phases in the DMAIC process.

The secondary goal of the project is to reduce the amount of NC’s occurring in the G550 sub-assembly department, so that an NC rate of 2,7 will be achieved.

2.2.9 Key stakeholders

A stakeholder analysis is the process of considering those people that will be involved or affected by the project, or who have some level of control over the project and process (Brook, 2010).

In this project the following stakeholders can be found:

1. The project leader of the entire G550 NC reduction project 2. The program manager of the G550 empennage

3. The production manager of the plant in Papendrecht (also a Black Belt) 4. The assembly workers at the sub’s

5. Other departments that cause faulty products to come to the sub department

The first two parties will be so involved in the process that they are in the team of the subs project. The third party identified will be one of the people assessing whether a tollgate can be passed. The forth party is the assembly workers, which will be involved in each phase of the DMAIC cycle. Especially in the ‘Measure’, ‘Analyse’ and ‘Improve’ phase, the input of experienced employees is needed to make this project a success. The last party will be involved in such a way that feedback will be given when faulty products arrive on the subs shop floor. However, when products are supplied right, no feedback will be given to them.

2.2.10 Schedule

Table 4: Schedule, Subs

Phase

Time

Deadline

(14)

14

2.2.11 Feedback

(15)

15

Chapter 3 Measure

3.1 The Measure phase

The second phase of the DMAIC method is the ‘Measure’ phase. Measurement can be defined as the assignment of numbers to observed phenomena according to certain rules. In theory it is a simple numerical assignment to something, but in reality is problematic, as managers never know the ‘true’ value (Pyzdek, 2003). Even though measuring has its limitations, is it still an important step in the DMAIC process. This step involves trying to collect data to evaluate the current performance level of the process and provide information for the ‘Analyse’ and ‘Improve’ phases (Yang & El-Haik, 2003). The aim is to set a baseline from which a clear measurement plan can be drawn up (Brook, 2010). The tools that can be used to execute this phase can be seen below.

Data collection plan

Inputs, outputs, variables, CTQ’s, KPI’s

Process capabilities, current quality level, CpK, PpK MSA

Data gathering Gage R&R

Stratification factors, brainstorming Sigma level (DPMO), baseline Current process flow

Needed sample size (sampling) Pareto

Operational definition

3.2 Validation of the ‘Measure’ phase

To be able to make an accurate validation of which tools to be used in the ‘Measure’ phase, this section aims to execute all the activities found in the conclusion found in “A Standardised Way of Applying Six Sigma to Increase the Yield of Quality of a Production Process” document. There, a list is presented with which tools are recommended by academic literature, (tool) books and by Green and Black Belts at Fokker Aerostructures.

3.2.1 Data collection plan

The data collection plan is mentioned most in the analysis done in research question 3 and is therefore put first in this set of tools. However, before a data collection plan can be done, information needs to be gathered, so that it is known what should be measured etc. Still, this is mentioned first, so it will be executed first, in order to stay consistent.

To come up with this plan, the brainstorm session that was done in the bigger NC reduction project for the G550 was used. This can be seen in Appendix I.

(16)

16

3.2.2 Inputs, outputs, variables, CTQ’s, KPI’s

The inputs of the process are the variables that influence the output. George et al. (2005) state that it is a measure of how well these variables, predict the output, where output quantifies the overall performance of the process. Yang (2003) also calls the output the CTQ’s or defect counts. Figure 5 below shows a small example on how to execute this phase.

Figure 5: Input and output variables

3.2.3 Process capabilities, current quality level, CpK, PpK

In this paragraph the current process capability or current quality level can be seen (Figure 6).

Figure 6: Amount of NC's incl. and excl. TAV

In this figure (Figure 6), two columns are shown. The blue is the one including TAV’s (supplied NC’s) and green column is the one showing only the EAV’s (own NC’s). The TAV and EAV are, so called, reason codes that are filled in as soon as an NC is written into the system. This distinction is made because the scope of the project clearly states that only the NC’s caused in the subs department are taken into account.

Scrap clean up; & Primer issue

(17)

17

3.2.4 Measurement System Analysis

The Measurement System Analysis consists of two steps. The first one is to assess whether the data collected is accurate. The second part of the MSA entails the creation of a data collection plan. This plan is already shown in the first paragraph of this chapter. To assess whether the data is accurate, second analysis was done on the actual NC rate. Here, the NC’s of the year 2012 have been assessed on whether the TAV and EAV reason codes were filled in correctly. This gave the figure below (Figure 7).

Figure 7: Actual NC rate

3.2.5 Data gathering

This step is included to verify that actual data is being collected. In this way the conclusion that are drawn from this data are based on facts and not on experiences or feelings from people. Even though, it might already be clear where the problems lies collecting data is a good way of verifying this and maybe finding new, unexpected results.

3.2.6 Gage R&R

(18)

18

When measuring continuous data, the Gage R&R can be an extremely helpful tool, though, and this might be taken into account when executing other phases.

3.2.7 Stratification factors, brainstorming

The use of stratification factor is to collect descriptive data that will help identify important patterns in the data. To find these factors, brainstorming can be used. When executing this step, the following figure (Figure 8) can be drawn.

Figure 8: Stratification factors

To find the stratification factors, brainstorming is a method that is often used. It is the search by a group for the solution to a problem and can be extremely useful in a variety of business situations (Ferguson, 2001). It can be used to stimulate the creative thinking process and makes sure that all group members’ ideas are considered (George et al., 2005). When brainstorming some ground rules can be set:

1. Find a group with different types of expertise, but avoid different ranks. A subordinate will always feel restrained in the presence of a superior (Ferguson, 2001).

2. Point out one facilitator.

3. Find the problem and set the goal for the brainstorm session (George et al., 2005). 4. Do not judge ideas (yet!) (Taylor, et al. 1958).

5. Go for quantity, not quality (Taylor, et al. 1958). 6. Build on existing ideas (George et al., 2005).

3.2.8 Sigma level (DPMO), baseline

In this step a baseline is set. This is important as a starting point, form where it is possible to set new targets. The sigma level can be calculated with the defect per million opportunities (DPMO). First, define what a defect is and what an opportunity is. In this case a defect is an NC, an opportunity is calculated in hours. So, on average, there are 1165 hours per month opportunities and 4,25 defects per month. This formula shows how to calculate the DPMO

(19)

19 When filled in the

A conversion table can be used to find the sigma level that corresponds to this DPMO, and here the sigma level is 4.18.

This equation can also be used to set the new target. Here, the formula can be seen below. with a sigma level of 4.29.

3.2.9 Current process flow

Creating a process map can help make a team see how the process works or should work and enables a team to quickly see improvement opportunities within the process (George et al., 2005). The process flow for the subassemblies in the G550 NC reduction project is as follows (Figure 9).

Figure 9: Current process flow

3.2.10 Needed sample size (sampling)

Sampling is taken data on one or more subsets of a larger group in order to make a decision on the whole group (George et al., 2005). The trade-off is between faster data collection versus some uncertainty.

(20)

20

For this project it is, therefore, decided that all the NC’s will be measured from July and August that were created in the subs department, and all NC’s of the past year (2012) can be used to find potential causes in the ‘Analyse’ phase.

3.2.11 Pareto

A Pareto chart can be created once the data has been gathered. It helps the team to find out on which process step or on which production phase to focus. The figure below (Figure 10) shows a Pareto chart on the process step where NC’s occur in the subs and on the production phase which the subs department. As one can see four process steps were mentioned when collecting data. The fifth one (TAV) is NC’s that are from other departments, so they are considered out of the scope of the project. For these four steps, one is most clear, which is drilling. Out of the 28 NC’s that are in the scope, 22 are caused during the process of drilling, which is almost 80%. It is very clear that this is area which will be focused on. Here, drilling can be seen as the z, or searching area, for this project.

Figure 10: Pareto - process steps

The figure below (Figure 11) attempts to focus on a particular area even more. This Pareto looks at the different production phases where NC occurred that were caused during the process of drilling. As one can see almost half of the 22 NC’s created here, are at the subs HS department and this is an area that will be focused on. However, the subs VF production phase is also quite a large percentage of the whole, so both phases will be important in this investigation to reduce the amount of NC’s in the sub assembly department.

22 6 2 2 2 0 5 10 15 20 25

Drilling TAV Assembling Countersinking Rivetting

N C's in t h is p ro ce ss ste p Process step

(21)

21 Figure 11: Pareto - production phases

3.2.12 Operational definition

Operational definitions are developed to provide clear and ambiguous descriptions for each KPI (Brook, 2010). Eckes (2003) also adds that it is a description of something where those affected have a common understanding over what is begin described. It helps ensure common, consistent data collection and interpretation of results (George et al., 2005).Operational definitions are part of the data collection plan, which can be seen in the first paragraph of this chapter.

For this project, the team decided what needed to be measured and this can be seen in Appendix II and Appendix III. It was decided to have an interview with the particular person that either caused the NC in the subs department or was the one that noticed the NC. In this way, the right person was interviewed which had the right information. Every week, all NC’s of that particular week were discussed with the assembly employee so that it was still fresh in the memory. For the interview, questions were answered but no other specific ‘rules’ were set as filling in the form is relatively easy.

3.2.13 Feedback

Feedback was given by a GB a BB and the Director of Operations. This feedback and the conclusions can be seen in “A Standardised Way of Applying Six Sigma to Increase the Yield of Quality of a Production Process” document.

10 9 3 0 2 4 6 8 10 12

Subs HS Subs VF Centerbox

N C's in t h is p ro d cu tion p h ase Production phase

(22)

22

Chapter 4 Analyse

4.1 The Analyse phase

As seen before, the ‘Analyse’ phase comes third in the DMAIC cycle for an improvement process. This phase involves determining the root causes of defective products (Dreachslin, 2007) and the goal is to analyse the problems and process inefficiencies (Furterer & Elsehennawy, 2005). Eckes (2003) states that the ‘Analyse’ step is seen as the most important by many in the DMAIC methodology. This is the case because many project team have preconceived notions of what to improve and after measurement they will want to jump right to the ‘Improve’ phase (Eckes, 2003). To avoid any solving of the ‘wrong’ problems it is essential that a team verifies why the problem exists. George et al. (2004) also mention the challenge of sticking to the data to reach conclusions about the root causes of problems. It is important to look for patterns in the data and target places where there’s a lot of waste. The tools that can help you with this, can be seen below.

Pareto/prioritizing causes Potential causes/X’s

Fishbone/Cause and Effect/CEDAC Process mapping/Flowcharting Brainstorming

Hypothesis testing Scatter plots

Regression/correlation

Value stream mapping/value analysis Run chart

ANOVA

Gage R&R per X Baseline per X Seven wastes Improvement plan Five laws of lean

4.2 Validation of the ‘Analyse’ phase

In this chapter an effort is made on validating the ‘Analyse’ phase as it is presented in “A Standardised Way of Applying Six Sigma to Increase the Yield of Quality of a Production Process” document. There, a list is created with sixteen types of tools that are seen as the vital 80% and should be used when analysing any NC reduction project. To test whether this conclusion is correct, the tools that are listed will be executed. After this, interviews are held with experienced Green and Black Belt to verify whether these tools were essential, can be used optionally or are not useful at all in this particular situation.

(23)

23

4.2.1 Run chart

The purpose of a run chart is to show data points in the order in which they occurred (George et al., 2005). It can help show the current performance. For the subs process the run chart can be seen below (Figure 12). 35 30 25 20 15 10 5 1 3,0 2,5 2,0 1,5 1,0 0,5 0,0 Observation N C s pe r w ee k

Number of runs about median: 17

Expected number of runs: 16,8

Longest run about median: 7

Approx P-Value for Clustering: 0,535

Approx P-Value for Mixtures: 0,465

Number of runs up or down: 20

Expected number of runs: 23,0

Longest run up or down: 4

Approx P-Value for Trends: 0,108

Approx P-Value for Oscillation: 0,892

Run Chart of NCs per week

Figure 12: Run chart, NC's per week

(24)

24 34 31 28 25 22 19 16 13 10 7 4 1 4 2 0 -2 O bserv ation In d iv id u a l V a lu e _ X=0,971 U C L=4,179 LC L=-2,236 34 31 28 25 22 19 16 13 10 7 4 1 4 3 2 1 0 O bserv ation M o v in g R a n g e __ MR=1,206 U C L=3,940 LC L=0 I-MR Chart of NCs per week

Figure 13: I-MR chart, NC's per week

In short, the graph shows that the mean, or average, of the amount of NC’s per week is 0.971. Also Minitab analyses the upper control limit (UCL) and the lower control limit (LCL), which represent the expected variation. The goal of any process is to reduce the variation as much as possible.

(25)

25 64 57 50 43 36 29 22 15 8 1 4 2 0 -2 O bserv ation In d iv id u a l V a lu e _ X=0,171 U C L=0,875 LC L=-0,533 Before A fter 64 57 50 43 36 29 22 15 8 1 4 3 2 1 0 O bserv ation M o v in g R a n g e __ MR=0,265 U C L=0,865 LC L=0 Before A fter 1 1 1 1 1 1 1 1 1 1 1 1 1

I-MR Chart of C3 by Before/after

Figure 14: I-MR chart, no drilling

As one can see, the average has changed drastically from 0,971 NC’s per week to only 0,171 NC’s per week and the variance is now only 1,408 compared to 6.415 of the current state. These two graphs show that the goal of the project is definitely attainable and the project has a lot of potential.

4.2.2 Process mapping/Flowcharting

A process map, or flow chart, is an illustration where the process that was chosen can be seen in details. This map is best constructed with the help of assembly workers that know the process very well and can describe each step in detail. The process map for the subs project can be seen below (Figure 15).

(26)

26

4.2.3 Fishbone/Cause and Effect/CEDAC

A Cause and Effect diagram is a tool that helps a team organize the ideas they have about potential causes of problems. It is sometimes called a fishbone because it resembles the skeleton of a fish (George et al., 2004). CEDAC stands for Cause and Effect diagram with Added Cards and this is a special kind of Cause and Effect diagram, where solution are also given. As we are still in the ‘Analyse’ phase, no solutions are given yet, so a relatively simple fishbone is shown below (Figure 16). To construct the fishbone, a brainstorm session was held where the process map above (Figure 15) was used to find all possible things that can go wrong. With the help of ranking these things on how severe it is, when they occur, the following 5 steps in the process and nine aspects were ranked highest. This ranking helps with a starting point for the measurements of the ‘Analyse’ phase. Also notice that the points under the process step of ‘Drill all holes 2.4mm or 3.2mm’ are applicable to all process steps that involve drilling.

Figure 16: Fishbone diagram

4.2.4 Value stream mapping/value analysis

A value stream map (VSM) is an overview of the value stream in an organisation, from supplier to customer. It is used to identify waste in a process (Bakker et al., 2011). A value analysis is used to distinguish process steps that customers are willing to pay for from those that are not (George et al., 2005). Both can be used to clearly identify all steps of the process, but a VSM is more complicated to construct than a value analysis. Due to the relatively detailed process map an VSM is not necessary, but a value analysis can be helpful. Per step, there are three classification (George et al., 2005):

1. Value added (VA); which is any activity in a process that is essential to deliver the product to the customer

2. Business Non-Value Added (BNVA); these are activities required by the business to executed the VA activities, but are not real value for the customer

(27)

27

In the figure below (Figure 17) each of the three categories have been given a colour to identify VA (green), BNVA (yellow) and NVA activities (red). The aim for the NVA is to eliminate it and the aim for the BNVA is to reduce it to a minimum.

Figure 17: VSM for drilling

4.2.5 Brainstorming

To execute the three phases above, brainstorming is a good tool to use. Sub section 3.2.7 explains brainstorming in more detail and also practical tips and trick are included here on how to facilitate a brainstorming session and make the session as efficient as possible.

4.2.6 Potential causes/X’s

With the use of brainstorming and with the fishbone diagram, a list can be drawn up with the potential causes or X’s. Also, historical data was used, together with an interview with an experienced assemble workers from the subs department was held. Below the list of potential cause can be seen (Table 5). Many more causes for defects can be found, but to make this list more accessible only nine causes are shown.

(28)

28

4.2.7 Pareto/prioritizing causes

Now that the potential causes have been identified, a priority can be given to them. With the help of three aspects each potential cause has been given a rank. The severity, the occurrence and the level of detection is given a grade between 1 thru 10, where 1 is lowest and 10 is highest. To give an example, the potential cause of no grease used is fairly severe when it happens (7), it happens quite often (8) and it can be detected before the mistake is made (6). When multiplying these three factor a total can be given, which helps rank the causes. Below the table for the ranking of the causes for NC’s for subs can be seen (Table 6)

Table 6: Potential causes and severity, occurrence and detection

As one can see, the potential causes fall into one of the three categories of discipline, the use of grease and the drill. In Table 7, the potential causes are show together with the right category. Next to that, the number of point given to each category can be seen in Table 8 and from this the Pareto in Figure 18 was drawn.

Table 7: Potential causes and their category

Table 8: Categories and the total points

Type Total points

Discipline 976

Drill 635

(29)

29 Figure 18: Pareto – causes

With the help of the tables and graphs above, the potential causes are prioritised and the greatest potential cause is a lack of discipline. Assembly workers have the freedom to decide many thing, like the use of a drilling jig or to check their equipment before use. The next category is the drill itself. It is not exactly clear what the ideal drill looks like and there is a lot of variation in the use of the drill. For example, some people use their body weight to drill a hole, while others consider a drill blunt when any type of force needs to be applied. This variation needs to be taken out of the process of drilling a hole, to have a more stable process which is less depend on any human factor. Last, is the use of grease when drilling a hole. If this is not used, a hole can become oval or too big and further research can be done to find the exact amount of grease that should be used. Again, it is a human factor where assembly employees can decide whether or not to use it.

4.2.8 Baseline per X

For this project the x’s that were identified that can influence Y, needs to be measured. This aims to show what drilling actions create NC’s. For this a measurement plan was set up, which can be seen in Appendix IV. Here, a form was filled in by the assembly workers after a product was drilled. Questions were asked about how they knew where the position of each hole is, whether they used a drilling jig and whether grease was applied. The purpose of this form is to collect data on the way that drilling is done and this data can be cross linked to both correct and incorrect holes. So, when an NC is created this form might give insight in that specific drilling actions. However, because only 4 to 5 NC’s occur in this department per month, the measurement might take some time to collect so a parallel measurement is done.

976 635 336 0 200 400 600 800 1000 1200

Discipline Drill Grease

(30)

30

In the second test, an investigation is done to see what the effect grease is, when drilling holes. The test was done on a metal disc, with a thickness of 9.84mm. This is thicker than the usual material that is drilled, but this thickness makes that the effect are more extreme, for example when looking at the wearing of the drill. Four rows of holes were made where two rows were drilled 3.2mm and two were drilled 4.8. Per diameter, one row was drilled with grease, and one without. Per row a new drill was used and each row has 19 holes. This metal disc can be seen in Appendix V. After the drilling, the holes were measured with a two-points meter on 0° and 90°, to check the hole size and ovality. The hypothesis set is that the use of grease gives more accurate holes than the holes drilled without grease.

4.2.9 Gage R&R per X

As described in previous chapters, the Gage R&R measures the reproducibility and the repeatability of a measurement. Its aim is to find out whether more measurements done by the same person, get the same results. Also, it looks at whether the same results are obtained when the same measurement is done by different people. When measuring with continuous data, this is a good system to take all variation out of the measurement system. Continuous data is any variable measured on a continuum or scale that can be indefinitely divided (George et al., 2005). So, in the second experiment a Gage R&R is a useful tools to check whether the data obtained will be reproducible and repeatable. Still, this is a very small and relatively simple test so no Gage R&R was executed. In the case of measuring the potential causes, named above (Baseline per X), the data is not continuous but discrete. Discrete data, or attribute data is ‘data’ with labels; here is no particular order (George et al., 2005). Due to the fact that the data is not continuous but discrete, a Gage R&R cannot be executed in this case. However, in more practical and general terms, the Gage R&R is a tool that makes the project leader think about the measurements that are done. These measurement should somehow give reliable results and show a comprehension of the data that is being obtained.

For the following chapters ‘Scatter plots’ thru ‘ANOVA’ the data from the drilling experiment is used. The data can be seen in Appendix VI.

4.2.10 Scatter plots

(31)

31 3,280 3,275 3,270 3,265 3,260 20 15 10 5 0 3,29 3,28 3,27 3,26 W it h ( 3 .2 m m ) Hole W it h o u t (3 .2 m m )

Matrix Plot of With (3.2mm); Without (3.2mm) vs Hole

Figure 19: Scatter plot 3.2mm

4,83 4,82 4,81 4,80 4,79 20 15 10 5 0 4,800 4,785 4,770 4,755 4,740 W it h ( 4 .8 m m ) Hole W it h o u t (4 .8 m m )

Matrix Plot of With (4.8mm); Without (4.8mm) vs Hole

(32)

32

4.2.11 Regression/correlation

Regression and correlation measures are related to the scatterplot. Only, where the scatterplot shows an overview of the data, regression and correlation also find coefficients related to the data. Correlation indicates whether there is a relationship between the values of different measures (George et al., 2005) the coefficient of the correlation can be measured by the Pearson Coefficient. This is used to measure the degree of linear association between the sets of data (Brook, 2010). Below, Minitab is used to calculate the Pearson correlation and a p-value for each variable (With (3.2mm), Without (3.2mm), With (4.8mm) and Without (4.8mm)).

Pearson correlation of With (3.2mm) and Hole = 0,023 P-Value = 0,925

Pearson correlation of Hole and Without (3.2mm) = 0,871 P-Value = 0,000

Pearson correlation of Hole and With (4.8mm) = -0,427 P-Value = 0,068

Pearson correlation of Hole and Without (4.8mm) = -0,463 P-Value = 0,046

These results give the following information:

 The Pearson coefficient range from +1 (Strong positive correlation), to zero (no correlation), to -1 (strong negative correlation).

 Only the holes without grease of 3.2 mm has a correlation, when looking at the Pearson coefficient of 0.871(blue).

 If the p-value is less than 0.05, a correlation exists. So, for the holes without grease of 3.2mm a correlation exists. However, the holes drilled without grease that are 4.8mm also have a p-value below 0,05 but this is so close to 0,05 that no strong correlation can be identified (green).

(33)

33 20 15 10 5 0 3,295 3,290 3,285 3,280 3,275 3,270 3,265 3,260 Hole W it h o u t (3 .2 m m ) S 0,0043312 R-Sq 75,8% R-Sq(adj) 74,4%

Fitted Line Plot

Without (3.2mm) = 3,260 + 0,001324 Hole

Figure 21: Regression, 3.2mm without

20 15 10 5 0 4,80 4,79 4,78 4,77 4,76 4,75 4,74 4,73 Hole W it h o u t (4 .8 m m ) S 0,0157444 R-Sq 21,5% R-Sq(adj) 16,8%

Fitted Line Plot

Without (4.8mm) = 4,786 - 0,001421 Hole

Figure 22: Regression, 4.8mm with

4.2.12 Hypothesis testing

(34)

34

“Is P is low, H0 has to go”. The p-value is the probability of getting the same results that you got if the H0 was true (Brook, 2010). So if p is low, the probability of getting the same results is also low. For this investigation a normality test, a 2-sample t-test, and a probability plot was drawn.

Normality test

When data is first obtained, the first step is check whether the data is normally distributed. This is important for further steps in the process, as this determines which tools can be used and which cannot. Checking normality can be done with the use of Minitab. Here, the data can be calculated where the p-value will be important. If the p-value is greater than 0.05 the data is normally distributed.

 With grease 3.2mm, the p-value is 0.248

 Without grease 3.2mm, the p-value is 0.313

 With grease 4.8mm, the p-value is 0.374

 Without grease 4.8mm, the p-value is 0.094 This shows that all data is normally distributed.

2-sample t-test

The 2-sample t-test looks at differences in the averages of two different samples (Brook, 2010). The p-value will show whether the samples are different or not. If the p-value is lower than 0.05 the samples are different, otherwise they are not.

The 3.2mm tests are compared with and without the use of grease This give a p-value of 0.325, which means that there is no significant difference in samples.

When looking at the 4.8mm differences in samples of with or without grease, the p-value is 0.00. This means that there is a significant difference between the two samples.

Probability plot

(35)

35 3,295 3,290 3,285 3,280 3,275 3,270 3,265 3,260 99 95 90 80 70 60 50 40 30 20 10 5 1 Data P e rc e n t With (3.2mm) Without (3.2mm) Variable Probability Plot of With (3.2mm); Without (3.2mm)

Normal

Figure 23: Probability plot, 3.2mm

4,84 4,82 4,80 4,78 4,76 4,74 99 95 90 80 70 60 50 40 30 20 10 5 1 Data P e rc e n t With (4.8mm) Without (4.8mm) Variable Probability Plot of With (4.8mm); Without (4.8mm)

Normal

Figure 24: Probability plot, 4.8mm

(36)

36 Table 9: Standard deviations 3.2mm and 4.8mm

Standard deviations

With

Without

3.2mm diameter 0.0058 0.0086

4.8mm diameter 0.0122 0.0173

Here, for both the diameters, the standard deviations are smaller with the use of grease than without the use of grease. This also confirms that the variance is smaller when drilling with grease, than without.

Next, it shows that, for the 3.2mm diameter, none of the holes have a size that is smaller than 3.2mm. The one closest to 3.2mm lies somewhere around 3.261. For the 4.8mm diameter the red and black line are further apart, and the holes drilled without grease are all smaller than 4.8mm. The holes drilled with grease, balance around the 4.8mm and seem to be more accurate, than the other rows with drilled holes. The fact that the holes drilled with grease are bigger than the holes drilled without grease, may be explained through the use of different drills. Each new row, a new drill was used, but it is possible that one drill was slightly bigger than the other one.

(37)

37 0,03 0,02 0,01 0,00 -0,01 -0,02 -0,03 99 95 90 80 70 60 50 40 30 20 10 5 1 Data P e rc e n t Ovality With (3.2mm) Ovality Without (3.2mm) Ovality With (4.8mm) Ovality Without (4.8mm) Variable

Probability Plot of Ovality With; Ovality With; Ovality With; ...

Normal

Figure 25: Probability plot, ovality

4.2.13 ANOVA

ANOVA stands for ANalysis Of VAriance and it compares three or more samples with each other to see if any of the sample means is statistically different from one another (George et al., 2005). ANOVA can only be applied when the data is normally distributed. Because there are only two samples, this tool cannot be used here.

After all the statistical tests, one can conclude that the use of grease improves the holes that are drilled and applied grease with each hole can prevent NC’s from occurring. No relation was found between the use of grease and the ovality of holes, as none of the holes was clearly oval and no sample is significantly better than another sample. Also, we running some quick tests in Minitab, the data did not show any correlation between the holes drilled with and without grease.

4.2.14 Five laws of lean

Adding lean to Six Sigma gives the best of both world. Lean Six Sigma finds its foundation in four important principles (George et al,. 2004):

 Delight customers

 Improve the process

 Use teamwork

 Base decisions on data

(38)

38 1. The Law of the Market

Customers are most important and they need to define quality which will be highest priority to achieved sustained revenue growth.

2. The Law of Flexibility

This law states that the speeds and flexibility of a process are linked. An inflexible process will hinder flow and vice versa (Brook, 2010).

3. The Law of Focus

This law states that 80% of the problems in a process will be caused by 20% of the activities. It should be best to focus on this 20%.

4. The Law of Velocity (Little’s Law)

The speed of a process is inversely related to the amount of Work-in-progress (WIP). If WIP is high, speed is low and vice versa.

5. The Law of Complexity and Cost

In general, the more complex a product is, the more it will cost.

4.2.15 Seven wastes

Seven wastes of a process have been identified by Taiichi Ohno, who was Toyota’s Chief Engineer and these are often used when thinking of ‘lean manufacturing’. When looking at a process it is important to find out what the wastes are, so that it is possible to recognize ways in which to improve. The seven wastes are mentioned below (Table 10) and a short description is given.

Waste Description

1 Overproduction Making more products than the customer requires 2 Waiting Waiting increases lead time and does not add value 3 Transporting Moving things cost money and time without adding value 4 Over processing Adding more value than the customer is willing to pay for 5 Inventory Holding inventory increases lead time and costs money

6 Motion Needless movements at ergonomic level have impact on overall efficiency and can cause health and safety issues

7 Defects Defects cause repairing or replacing which is costly

Table 10: Seven wastes (Brook, 2010)

When trying to think of how to improve the process, use these wastes to see where they occur and can be eliminated. For example, when there is a lot of waiting on a product, see how this time can be shortened.

4.2.16 Improvement plan

The last step in the ‘Analyse’ phase is to start with the improvement plan. This can be a brief and simple description on how to start improving.

(39)

39

4.2.17 Feedback

(40)

40

Chapter 5 Improve

5.1 The Improve phase

The ‘Improve’ phase comes after the ‘Analyse’ phase and is fourth in the DMAIC process. The aim of the improvement phase is to examine the causes which appear during the analysis phase and to generate a set of solutions to improve the performance of the process (Orbak, 2012). Rasis et al. (2002:2) state that the ‘Improve’ phase involves designing experiments to understand the relationship between the Y’s and the vital few X’s … and conducting pilot tests of the action plans. Basically, it aims to make changes in a process that will eliminate defect, waste, costs, etc., that are linked to the customers’ needs identified in the ‘Define’ phase (George et al., 2004). The tools that can be helpful are listed below(Table 11).

Table 11: Tools 'Improve' phase

Generate possible solutions

Brainstorming for solutions Design of Experiments

Benchmarking

Select the best solution

Prioritisation matrix/ Solution selection matrix/Pareto Impact & Effort matrix/Pick chart

Assessment criteria

Assess the risks

Fishbone/Cause and Effect/CEDAC

Pilot & Implement

Pilot studies/Testing

Implementation/Documentation

Main effect plots/Impact measurement

Recalculation of sigma/Process capability

5S

Visual management

5.2 Validation of the ‘Improve’ phase

In this section the ‘Improve’ phase will be validated by executing the tools found in “A Standardised Way of Applying Six Sigma to Increase the Yield of Quality of a Production Process” document in a case study. Four steps identified, each with tools and activities within these steps. To test whether these tools are identified correctly, they will be used in the ‘Improve’ phase for the subs project. Once, this is done, the phase will be discussed with experienced Green and Black Belts at Fokker Aerostructures to test whether they were essential.

5.3 Generate possible solutions

(41)

41

5.3.1 Brainstorming for solutions

To find the possible solutions to the causes, brainstorming can be used. In sub section 3.2.7 some detailed ground rules are set to executed a successful brainstorm session. Here, so ideas are given to improve the use of grease in assembly of the subs:

A. An automatic drill, that only starts when grease is applied B. Make sure the grease is available at all times

C. Train people in explaining the effect of the use of grease

D. Create awareness/hang up signs to help people remember to use grease E. Make it as easy as possible to use the grease

5.3.2 Design of Experiments (DOE)

Design of Experiments aim to identify the important inputs to the process (critical x’s) and to understand their effect on the process output (Brook, 2010). A designed experiment can be defined as an experiment where one or more factors, called independent variable, believed to have an effect on the experimental outcome and manipulated according to a predetermined plan (Pyzdek, 2003). The statistical way in which this can be applied, might be more relevant in the ‘Analyse’ phase. DOE can help verify x’s, but this was already done in the ‘Analyse’ phase with other types of statistical tools.

5.3.3 Benchmarking

Benchmarking can be used in the ‘Improve’ phase which involves identifying and understanding best practices from other processes and organisations (Brook, 2010). It can involve research into the best practices at the industry, firm, or process level (Pyzdek, 2003). And tells you what’s possible so you can set goals for your own operations (George et al., 2005). For the drilling of holes with grease, the benchmarking can be fairly simple and within the process. The causes of drilling wrong holes can be due to the lack of grease on the drill. Here, best practice is to use grease for each hole to eliminate the chance of getting a hole that is outside specifications, in other words, an NC.

5.4 Select the best solution

Now that possible solutions are generated in the previous step, the best solutions needs to be chosen that can be implemented. The selection of the best possible solutions will be done below in three steps.

5.4.1 Prioritisation matrix/solution selection matrix/Pareto

(42)

42 Table 12: Priorisation matrix

From this prioritization matrix it is clear that, with the criteria and its weights, solution D is the best solutions. This solutions entails creating awareness through, for example, hanging up signs. This will not be very expensive, can be implemented quickly, costs less than €1000 and won’t impact the customers.

5.4.2 Impact & Effort matrix/PICK chart

An impact & effect matrix or PICK chart is a tool which can be used to choose one of many solutions. It helps to organise and prioritise the solutions and separates them in four categories, Possible, Implement, Challenge and Kill (George et al., 2004). Below, the matrix can be seen (Table 13) together with the solutions found in chapter 12.2.

Table 13: PICK chart

BIG payoff

SMALL payoff

EASY to implement

_{B D}

Implement Possible

HARD to implement

Challenge

C E

Kill

(43)

43

5.4.3 Assessment criteria

Assessment criteria provide a constant basis of comparison within the solution selection techniques that can be used (Brook, 2010). These solutions techniques can be seen above, where the criteria are already set to complete the selection process. Here, the criteria are:

 Can be implemented quickly

 Will solve the problem fully

 Costs less than €1000

 Won’t impact the customer

5.5 Assess the risks

In section 5.4, the best solutions have been identified, but there is one more step that needs to be taken before the solutions can be implemented. This steps involves assessing the risk of the solution proposed

5.5.1 Fishbone/Cause and Effect/CEDAC

A Cause and Effect diagram is a tool that helps a team organize the ideas they have about potential causes of problems. It is sometimes called a fishbone diagram because it resembles the skeleton of a fish (George et al., 2004). CEDAC stands for Cause and Effect diagram with Added Cards and this is a special kind of Cause and Effect diagram, where solution are also given. As we are still in the ‘Analyse’ phase, no solutions are given yet, so a relatively simple fishbone is shown below (Figure 26). It is filled in with the help of the 6-M model and per ‘M’ the risks are identified. The 6 M’s stand for, Measurement, Machine, Men (people), Method, Milieu (environment) and Materials.

(44)

44

5.6 Pilot & Implement

The fourth and last step of the ‘Improve’ phase is to pilot and implement the solutions. This can be done, because all possible solutions were generated, prioritised and their risks were assessed. It is now safe to take the best solution and first pilot and then implement it, where the following steps can be helpful.

5.6.1 Pilot studies/testing

A pilot study is a localised, controlled trial of a solution in order to test its effectiveness before full implementation (Brook, 2010). It helps identify practical problems and failures in a chosen solution (George et al., 2005). The tests that will be implemented is to create awareness and make sure that the grease is always available.

The awareness can be created by hanging things up, highlighting it in the building descriptions and, in the future, more explicitly teach the new assembly workers the importance of the use of grease when drilling products.

The availability of grease was checked with the suppliers and supplying enough grease was not a problem. However, when looking at the workplace, the availability was not always possible. The grease was there but the workplace was not always clean and well ordered. The concept of 5S can be applied here, where all equipment and other appliances will be given a specific place in the working area and this is cleaned after each shift (for more information on 5s, see below in the paragraph ‘ 5S’). In this way, the grease had its own, logical, place and the availability can be guaranteed.

For this project the actual solutions were not implemented due to a lack of time. However, the following paragraphs represent a proposal for possible improvements, as if the pilot tests were completed successfully.

The availability of grease will not be a big problem , due to the 5S actions that were taken before and the suppliers had no problems delivering the grease. The awareness of the assembly workers was high, as this was right after the experiments with the drilling and all the measurements. In this period, it was therefore a success. However, when moving into the future, this awareness may decrease as less attention will be given to the use of grease For now, the test was completed successfully, but in the ‘ Control’ phase this will need to get some extra attention.

5.6.2 Implementation/documentation

(45)

45

5.6.3 Main effect plots/impact measurement

A main effect plot maps one input against the process output (Brook, 2010).In the example below (Figure 27) the drilling of 3.2mm diameter holes without the use of grease, is plotted against the number of holes. Here, it is possible to see that the more holes are drilled, the bigger the holes get. Knowing this, will help make the assembly workers more aware of the fact that drilling without grease, especially when the frill diameter is relatively small compared to the material thickness, can cause serious problems.

19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 3,295 3,290 3,285 3,280 3,275 3,270 3,265 3,260 Hole M e a n

Main Effects Plot for Without (3.2mm)

Data Means

Figure 27: Main effect plot, subs

Another type of main effect plot exists, but this is related to the tools of DOE. This tools is said to be less helpful in the ‘Improve’ phase and it will, therefore, not be executed here.

5.6.4 Recalculation of sigma/process capability

To recalculate the process capability the following steps were taken.

(46)

46 15 13 11 9 7 5 3 1 7,5 5,0 2,5 0,0 O bserv ation In d iv id u a l V a lu e _ X=2,19 U C L=5,58 LC L=-1,19 Before A fter 15 13 11 9 7 5 3 1 6,0 4,5 3,0 1,5 0,0 O bserv ation M o v in g R a n g e __ MR=1,271 U C L=4,154 LC L=0 Before A fter

I-MR Chart of NC rate per month, real

Figure 28: I-MR chart, subs

5.6.5 5S

The 5s process is the housekeeping method that uses words beginning with the letter S to identify the activities needed to create a workplace with visual control and organized functioning (Martens, 2011). These principles should be kept in mind when piloting the new implementation and it might be possible that applying these rules can help prevent many problems. The table below (Fout!

erwijzingsbron niet gevonden.) shows the 5S’s and a short explanation with them.

Table 14: 5S (Brook, 2010)

Japanese word

English word Explanation

Seiri Sort To ensure that only items that are needed regularly are kept in the immediate workplace.

Seiton Straighten To ensure everything has a defined place and the most frequently used items are most at hand, promoting efficiency.

Seiso Shine Aim to maintain the workplace organised and tidy at all times. Possibly planned clean ups can be scheduled too.

Seiketsu Standardise Develop a system that can be integrated in to organisations’ way of working.

Shitsuke Sustain Keep committed to good housekeeping by giving the good example and ‘leading by doing’. This will also increase safety.

(47)

47

helpful for a process and from there is it possible to see whether the DMAIC project will need to be executed to improve the process even further, or that other areas of the company need more attention, where 5S may have helped improve the process less.

5.6.6 Visual management

Visual management is highly linked to the principle of 5S, described above. It is the use of graphical methods to display and communicate how a workplace or process is managed, controlled and performing (Brook, 2010). There are some simple tools that can support visual management like checklists and notice boards (Martens, 2011). The aim of visual management is to communicate performance measures (George et al., 2005), reduce accidents as part of a safety programme (Brook, 2010), minimise work in progress (Brook, 2010) and helps to quickly identify abnormal situations (George et al., 2005). In the case of the subs project, visual management can also be applied in the form of posters or signs saying to use grease when drilling. This can help the assembly workers to be reminded of the importance of the NC reduction project.

5.6.7 Feedback

(48)

48

Chapter 6 Control

6.1 The Control phase

The ‘Control’ phase aims to ensure that the solutions that have been implemented become embedded into the process, so that the improvement will be sustained after the project has been closed (Brook, 2010). This phase comes with many difficulties, including a lack of resources, lack of coordination between functions and impatience to get results (Gijo & Rao, 2005) while it is so important to finish this phase well, to make the changes and improvements last. To achieve the aim of closing the project in a good manner, several things can help, which may include documenting the new procedures and training everyone (George et al., 2004). Essential and optional tools are seen below (Table 15).

Table 15: Tools 'Control' phase

Implement on-going measurement

Process control plans/chart KPI trees

Standardise the solution

Documentation

Standardised processes

Monitor implementation

5S

Visual management

Training employees/Discipline

Quantify the improvement

Audit results/Financial benefits

Close the project

Communicate/transfer to process owners

Celebrate the success

Lessons learned

New project (continuous improvement)

6.2 Validation of the ‘Control’ phase

The last phase of the Six Sigma DMAIC method will be validated in these sections. In the “A Standardised Way of Applying Six Sigma to Increase the Yield of Quality of a Production Process” document four steps are identified each with tools that can be executed per step to successfully finish the ‘Control’’ phase and the entire improvement project. These tools are executed in a case study, which is a shown below. These results of this case study and the input of several Green and Black Belts, will check which of these tools are essential and which are not.

Validation project