60

CHAPTER 4

CASE IMPLEMENTATIONS

4.1 INTRODUCTION

The Six Sigma approach to quality and process improvement has

been used predominantly by manufacturing organisations since its inception.

Presently, many service organisations are also utilising this methodology

primarily because of its customer-driven basis. Manufacturing organisations

build their Six Sigma efforts on an established foundation of measurable

processes and set quality management programmes. In service organisations,

the Six Sigma programme is introduced to establish and map the key

processes that are critical to customer satisfaction. There are numerous

manufacturing companies applying the Six Sigma to their diverse

non-manufacturing processes, such as human resources, payroll, accounting,

customer relations, supply chain management, safety and hazard engineering,

organisation change and innovation because many of the methods used in Six

Sigma are applicable to both manufacturing and non-manufacturing industries

or services.

All these methods are practiced to minimise not just process

variation in manufacturing but also the variation of expectations to perception

in service organisations. The differences between goods and services lead to

service firms and goods-producing firms having different success factors for

Six Sigma. The extent to which Six Sigma fulfils the quality gaps, leads to the

improvement of the product or service quality. This is dependent on the

61

apparent challenges posed by the very nature and core premises of the

industry. Hence, it is necessary that the performance of the TPE model in both

manufacturing and service sectors be verified before a decision regarding its

suitability is taken.

Two case implementations are conducted to assess the performance

of the TPE model, one in the manufacturing industry and the other in the

service industry. In the manufacturing industry, this TPE model is

implemented to minimise the defect probability of the die casting operation

with the core objective of improving the productivity. In the service industry,

it is used to minimise the gap between customer expectations and perceptions

within an automotive service operation. The details of the studies are

presented in the subsequent sections of this chapter.

4.2 CASE IMPLEMENTATION - 1

4.2.1 About the Company

A South-India based automobile horn manufacturing company is

considered suitable for the application of this TPE model. The company

undertook casting of aluminium components to cater to the needs of various

industrial sectors. The apprehension of the company due to the rejections of

cast products has been tackled through use of the TPE model. A variety of

casting techniques are used by the company including aluminium pressure

casting of automotive components for their domestic as well as international

clientele.

The production process followed a batch-type production with

different lot sizes for the different components. The product range for instance

included governor housing for fuel injection pumps, heat-sink and field

moulds for alternators, oil pumps and pump body covers, fixing brackets for

62

car starters, and pivot housing for wiper motors. The company’s present

production facility is expanded to make castings for the textile and medical

field and created ring holders for ring frames, iron sole plates for electric

irons, and clam shells for surgical interconnect systems.

The major domestic clients of the company are Lakshmi Machine

Works (LMW), Lucus- India, MICO, Philips India, Pricol, TVS Motor

Company Limited and Wipro Infotech; while the international clients

consisted of TRICO, UK and Zinser, Germany to whom nearly 20% of the

overall production is supplied. The present production capacity of the

company is estimated at 920 tonnes per annum with state-of-art techniques in

the present production set-up. The production schedule is prepared against the

client order.

The company prefers to be one of the best suppliers by improving

quality of the product and meeting the order delivery commitments. The

company had its well equipped quality control department to assess the

quality of the castings in terms of dimensional variability, various casting

defects and handling damages. However, they were looking for a systemised

methodology for optimising the casting process to reduce the occurrence of

loss of productivity and the costs incurred in rejections and in payment of

penalties to the concerned customers.

4.2.2 About the Process

The various production processes surrounding components include

die casting, sand casting, permanent mould casting and investment casting.

The most widely practiced casting method is die casting because of its

inherent properties like a high volume of production at a low cost, high

precision rates, and excellent surface finish which eliminate post machining

63

requirements. However, a high cost of die and porosity in the cast product are

the issues that plague this industry.

The vital components of a typical aluminium pressure die casting

process are shown in Figure 4.1. The die casting process involves the use of a

furnace, raw material, die casting machine and die. The metal is melted in the

furnace and then, injected into the dies in the die casting machine. After the

molten metal is injected into the dies, it rapidly cools and solidifies to take its

final form, called the casting. The entire die casting process is pasteurized in

Figure 4.2.

Figure 4.1 Aluminium die casting process

64

Figure 4.2 Die casting process flow chart

The process cycle for die casting consists of five main stages,

which are explained below. The total cycle time is very short, typically

between 2 seconds and 1 minute.

4.2.2.1 Clamping

Preparation and clamping of the two halves of the die constitutes

the first process. Each die half is cleaned initially and lubricated to facilitate

the ejection of the next part. The lubrication time increases with the size of

the component, as well as the number of cavities and side-cores. Lubrication

may be required after 2 or 3 cycles, depending upon the material. Post

lubrication, the two die halves, attached inside the die casting machine, are

closed and securely clamped. Sufficient force must be applied to the die to

Raw Material

Preheating Ingots

Melting the Ingots

Keeping molten metal at set temperature

Furn

ace

Die cleaning

Die lubrication

Die closing

Die

Pouring molten metal in shot

chamber

Injecting molten metal into the die

Solidification

Die opening

Ejecting the casting

Trimming the casting

Final product

Rec

ondi

tioni

ng a

nd

addi

ng w

aste

to m

eltin

g

Die

cas

ting

mac

hine

65

keep it securely closed while the metal is injected. The time required to close

and clamp the die is dependent upon the machine. Larger machines (those

with greater clamping forces) require more time and this can be estimated

from the dry cycle time of the machine.

4.2.2.2 Injection

The molten metal, which is maintained at a set temperature in the

furnace, is subsequently transferred into a chamber where it is injected at high

pressures into the die. Typical injection pressures range from 1,000 to

20,000 psi. This pressure holds the molten metal in the dies during

solidification. The amount of metal that is injected into the die is referred to

as the shot. The injection time is the time required for the molten metal to fill

all the channels and cavities in the die. This time is short, typically less than

0.1 seconds, in order to prevent early solidification of any one part of the

metal. Proper injection time can be determined by the thermodynamic

properties of the material, as well as the wall thickness of the casting.

4.2.2.3 Cooling

The injected molten metal starts to cool and solidify as soon as it

enters the die cavity. The final shape of the casting is formed when the entire

cavity is filled and the molten metal solidifies. The die cannot be opened until

the cooling time has elapsed and the casting is solidified. The cooling time

can be estimated from several thermodynamic properties of the metal, the

maximum wall thickness of the casting, and the complexity of the die. A

greater wall thickness will require a longer cooling time. The geometric

complexity of the die also requires a longer cooling time because the

additional resistance to the flow of heat.

66

4.2.2.4 Ejection

The die halves are opened and an ejection mechanism pushes the

casting out of the die cavity after the predetermined cooling time has elapsed.

The time to open the die can be estimated from the dry cycle time of the

machine and the ejection time is determined by the size of the casting’s

envelope and should include time for the casting to fall free of the die. The

ejection mechanism must apply some force to eject the part since the part may

shrink and adhere to the die during cooling. Once the casting is ejected, the

die can be clamped shut for the next injection.

4.2.2.5 Trimming

The material in the die channel solidifies during cooling along with

the casting. This excess material, along with any flash that has occurred, must

be trimmed from the casting either manually via cutting or sawing, or through

the use of a trimming press. The time required to trim the excess material can

be estimated from the size of the casting’s envelope. The scrap material that

results from this trimming is either discarded or can be reused in the die

casting process. Recycled material may need to be reconditioned to the proper

chemical composition before it can be combined with non recycled metal and

reused in the die casting process.

4.2.3 Scope of the Study

The company being studied manufactures 58 varieties of products

of which invariably there is a higher defective percentage. Problems

persisting consist of the loss in productivity and customer orders not being

met on time. To keep the company on track, the production department used a

strategy of producing more components than the ordered levels to compensate

67

for the rejections. This exercise increased the cycle times for each component

and thereby, a loss in ROI. The nature of casting defects may be of two types:

one, the defect that is noticed immediately after casting the molten metal.

Such defects may be un-filling, gate broken; damage, weld, crack, un-wash,

rib broken, and metal peel off. The second category of defects is not

noticeable until the post machining processes are completed. Those defects

are porosity, blow holes etc. Eventually, these defects cause greater losses

than the former category in terms of money since they involve post machining

processes. Improving the organisational productivity is the foremost objective

for the research model being proposed in this situation. A cross functional

team is therefore, formed by the Head of the Quality Assurance department in

the company. The organisation’s objective is taken as the driver for this study

and iterated using the QFD concept to coincide with what needed

modification / improvement / reduction to achieve the goal. Then, the selected

improvement project is analysed and improved in the subsequent stages of the

research model. The flow chart deployed in Figure 4.3 depicts a summary of

the step-by-step activities undertaken.

4.2.4 TPE Stage 1: QFD Process

QFD technique intends to identify the possible ways to accomplish

the objective through the analysis of the HoQ matrix. The development of the

customer information table (horizontal matrix) is simplified with a single

objective as the requirement. In this study, the CCP analysis has not been

performed since the requirement considered in the matrix is of a unique

nature.

68

Generating solutions using 39 problem parameter and 40 inventive principles

Improve the problem

Control activities End

Is measurement

system available?

Contradiction matrix

Measure the problem

Analyze the problem

Contradiction analysis

Preparation of Innovative situation questionnaire

Function Modeling

Selection of viable project

Define the problem

Objective

Strategies

Strategy

Project plans

Start

Figure 4.3 TPE model in case implementation 1

That is, it would not fetch any useful information because each

organisation could have different objectives to run their business. A cause and

effect (C&E) analysis is carried out to sort the strategies to find out which

would influence the objective. In Figure 4.4, the strategies chosen from the

C&E analysis are cross referred with the objective. In the HoQ matrix shown

in Figure 4.4, the strategies S1 and S4 are found equally important to fulfil the

necessary objective. As a thumb rule, the strategy S1 (minimising the

defective fraction) has been chosen in order to continue.

69

S4 81 27 27

S3 27 9 27

S2 27 9 27

S1 27 27 81

S1 S2 S3 S4

Im

port

ance

Wei

ght

Min

imiz

ing

defe

ctiv

e f

ract

ion

Im

prov

ing

proc

ess

q

ualit

y

Im

prov

ing

empl

oyee

p

rodu

ctiv

ity

Min

imiz

ing

CO

PQ

Objective di Wi Strategies

Improving productivity 9 1 9 3 3 9

Weight Wj 0.375 0.125 0.125 0.375

Priority I II II I

Figure 4.4 QFD matrix – objective Vs strategies

4.2.4.1 Developing project plans to realise the strategy S1

The previous history of records showed that the rate of rejection

ranged from 0.72% to 14.53% of production due to various defects, but the

company target is 1.5%. Nearly 25 casting defects are reported in their record

as reasons explaining the defective products.

With the information in hand, the following project plans have been

formulated to mitigate the occurrence of defects in casting:

1. Process parameter optimisation

2. Die design analysis

3. Component design evaluation

4. Equipment capability analysis

70

Each project is explained in details to the top management as

shown below:

4.2.4.2 Process parameter optimisation

The die casting process handles hot metal. The metal temperature is

the first and foremost important process parameter which has greater

influence on defect formation like un-filling, and flash like the others. Other

operational parameters are injection pressure (first and second stages), die

coat and metal mixing ratio. For the case company, the aluminium alloy is the

principal material used for most of the components at different proportions.

A variation in temperature of the hot molten metal has greater

affinity to cause defects in the final product. Low temperatures result in

improper solidifications, whereas high temperatures cause excess casting

hardness, leading to defects like cracks. Next, an equally important parameter

is the injection pressure, which refers the pressure applied on molten metal

while pushing it into die cavity from shot chamber. Low pressure may result

in a partial solidification of casting due to delay in cavity filling. Excess

injection pressure may however, damage the gate or increase the gate

velocity, which contributes to the casting defect like porosity.

The next parameter of interest is the die coat, which is the medium

used to lubricate the hot die for the purpose of easy ejection after the

solidification process. With respect to the die coat, the attention is paid to

frequency of die coating and die coat material. Last but not the least, the metal

mixing ratio is one other important process parameter, which might contribute

to gas inclusions in the casting due to contamination of the recycled materials.

The ratio in which the scrap or trimmed materials are mixed with the new raw

material in furnace is however, of real interest.

71

The scope of this project proposal is to estimate the optimum

setting of the chosen parameters to obtain high quality casting with

reduced or eliminated defects.

4.2.4.3 Die design analysis

Dies are the custom tools used in this process where the molten

metal is injected to form a casting. The fundamental arrangement of a die is

illustrated in Figure 4.5. It is composed of two halves: the cover die, which is

mounted onto a stationary platen, and the ejector die, which is mounted onto a

movable platen. This design allows the die to open and close along its parting

line.

Figure 4.5 Die structure and design features

Once closed, the two die halves form an internal part cavity. The

cover die allows the molten metal to flow from the injection system, through

an opening, and into the part cavity. The flow of molten metal into the part

cavity requires several channels like venting holes, sprue, runners,

overflow-well and gates.

72

Apart from these hot metal channels, there are also cooling

channels designed in the dies to allow coolant medium water or oil to flow

through the die. These are located adjacent to the cavity and remove heat from

the die. Apart from these structural design parameters, there are other design

issues like the draft angle, and undercuts to ensure easy flow of molten metal

and accommodation of complex casting features. Selecting the material is

another aspect of designing the dies. Dies can be fabricated out of many

different types of metals. High grade tool steel is the most common and is

typically used for 100-150,000 cycles. However, steels with low carbon

content are more resistant to cracking and can be used for 1,000,000 cycles.

Other common materials for dies include chromium, molybdenum, nickel

alloys, tungsten and vanadium.

In a nutshell, the essential design features of dies are:

Hot metal channel; sprue, runners, gates and overflow

well

Air channel, venting holes

Cooling channels, coolant paths

Structural parameters, draft angle and undercut and

Die material.

A project may be formulated to evaluate the influence of

stated die design features on defect occurrences through the Six Sigma

and the TRIZ processes.

4.2.4.4 Component design evaluation

The production of defect free components in the pressure die

casting process solely depends on product design and development factors

73

like die design, operational parameters and materials used. But the probability

of defect occurrence is highly depending on the design complexity of the

product. For instance to cast components with high wall thickness, the

injection pressure should be more than enough to fill the deep cavity. But it

has the adverse effect on casting quality like gas bubble inclusion and

porosity.

Likewise, external rib-like shapes create extra designs on the die to

accommodate external slides, which not only increase the cost of die but also

the operational complexity. Some of design flaws resulting casting defects are

illustrated in Figure 4.6.

Poor part design Good part design Poor part design Good part design

Thick wall design Thin wall design Non-uniform wall

thickness

Uniform wall

thickness

Poor part design Good part design Poor part design Good part design

Sharp corners Round corners No draft angle With draft angle

Figure 4.6 Design flaws causing casting defect

The project may be used to evaluate the design of die for

assessing the proneness of the cavity design to casting defects.

74

4.2.4.5 Equipment capability analysis

This involves an evaluation of the machine’s capability towards

producing defect free castings. In this company, the die casting process uses

cold chamber high pressure die casting machines as shown in Figure 4.7 for

producing aluminium castings. The existing production setup includes 9 such

type of machines with varying capacity from 25ton to 500 ton of clamping

force. The main parts of the machine are: control box panel, fixed plate,

moving plate, back plate, accumulator, injection cylinder, injection rod,

ejector system, oil tank, die regulating valve, pressure regulating valve. The

injection pressure can be varied using the pressure regulating valve. Controls

are provided on the machines to regulate the die opening time, closing time

and the ejector time for the ejection of the cast product.

Figure 4.7 Cold chamber high pressure dies casting machine

The machine can be operated in manual as well as in the automatic

modes. Sample specifications of 80 ton cold chamber die casting machines

75

are given in Table 4.1. In the cold chamber type of high pressure die casting

machine, the molten metal is transferred into the cold chamber cylinder

through a port or pouring slot.

Table 4.1 Specifications of an 80 ton hydraulic pressure die casting

machine

Locking force 80 ton Dist. of centre and bottom injection 85 mm

Injection force 11.5 ton

Ejection force 4 ton Motor capacity 5.5 kw

Die mounting plates 520x520 mm Working pressure

100-135 kg/cm2

Tie bar space 330x330 mm Vane pump 70 ltr/min

Die height – max 400 mm Oil tank capacity

300 ltr

Die height – min 200 mm Machine weight

3.5 ton

Tie bar diameter 60 mm Shot capacity 950 grm

Die opening stroke 200 mm Ejection stroke 50 mm

Injection stroke 250 mm

A hydraulically operated plunger pressurises the molten metal in

the shot chamber and injects it into the die cavity. A second stage injection is

applied to ensure complete packing of the molten metal into the profiles of die

cavity. After the predefined solidification time, the moving die retracts and

the casting is drawn from the machine. Now the hot die halves are cooled by

the application of the die coat and both dies are clamped together. The second

cycle is started and continued as described in Figure 4.8.

76

Figure 4.8 Operations of high pressure die casting machine

The objective of this project is to check whether the machines

are operated in proper manner i.e. according to guidelines issued by the

manufacturers so that casting defects can be eliminated due to

malfunction of machinery and equipments

These proposals were prioritised based on their importance to

accomplish the strategy after briefing the projects to the apex management.

Figure 4.9 illustrates the HoQ in which the projects are weighted and

prioritised. The project proposal “P1 – Optimising the process parameters”

has been chosen to be executed the subsequent stages of this study since it has

higher weight than the rest of the projects and bears a strong co-operative

interrelationship.

Pouring molten metal

Injecting the molten metal

Removing the casting Applying die

coat

77

P4 9 3 3 .,

P3 27 9

P2 27

P1

P1 P2 P3 P4

Impo

rtanc

e

Wei

ght

Opt

imiz

ing

proc

ess

p

aram

eter

s

Ana

lysin

g di

e de

sign

Eva

luat

ing

com

pone

nt

d

esig

n

Mac

hine

cap

abili

ty

ana

lysis

Strategy – S1 di Wi Project plans Minimizing defective fraction 9 1 9 3 3 1

Weight Wj 0.563 0.186 0.186 0.063

Priority I II II III

Figure 4.9 QFD iteration – strategy versus project plans

4.2.5 TPE Stage 2: Six Sigma and TRIZ Processes

The following activities were executed in this process to develop a

feasible solution to the project selected:

Developing the problem statement using Triz - ISQ.

Measuring present defective rate and its Sigma level.

Analysing defect causes and remedies.

Resolving contradictions.

78

4.2.5.1 Defining the problem statement

A literal problem statement was developed to push the study in the

right direction. Since the present operations performed by the organisation are

individually unique, there is no relationship between the components except

for the casting process. But the process parameters are common and the

parameter values are only changed as specified by the production department

for producing each type of component. In this situation, an innovative

approach is needed to guide the team to execute the study. ISQ, problem

modelling / formulation and IFR along with the Triz analytical tools are used

to define the problem by mapping the process. First, the ISQ has been

designed to analyse the present situation, and then, we undertook problem

formulation using problem modelling.

4.2.5.2 ISQ for situation analysis

An ISQ was developed with a set of questions deployed in Table

4.2 to ascertain better understanding of the nature of the present process setup

and the defects. The answers of the ISQ are used to stimulate problem

formulation.

79

Table 4.2 Innovative situation questionnaire

1. What is the purpose of the process parameter optimization?

It is required to optimize the process parameters to minimize the defect probability of the

die casting process for improving the productivity.

2. What are the existing defects?

23 different defects are noticed in the present process

3. What can be done to avoid defects in castings?

Several attempts were made in the area of product design evaluation and die design

evaluation but process optimization is not attempted so for. Hence to resolve the crisis,

process optimization may be practiced.

4. What are the advantages and disadvantages of the known solutions?

The expected benefits of process parameter optimization include reduced defective

fraction, reduced cycle time, improved resource utilisation and improved quality. The

possible disadvantages are high time and cost involvement for process setup

modification, all the defect possibilities may not be eradicated, individual product

requires separate process optimization hence it might be a cumbersome approach.

5. What is the ideal solution to the original problem?

Zero rejections of castings due to defects

4.2.5.3 Problem modelling and formulation

Problem modelling is undertaken after obtaining information from

the ISQ situation analysis. It involved the building of a function diagram by

using the function analysis as shown in Figure 4.10.

80

Figure 4.10 Function modelling using the TRIZ function analysis

The cause-and-effect relationships among the functions are

publicised in this functional diagram. The useful function UF-1 has been

found to be the prerequisite to deliver the UF-2. The UF-2 is mandatory for

achieving UF-3 and thereon, the UF4. However, UF-1 would produce harmful

Function HF-1, which could have a reverse effect on the effectiveness of

UF-1. Problem formulation is examined based on a functional diagram,

subsequently aiding the identification of the following problem statements;

1. Identifying an alternative way to optimise the process

parameters without adding cost and time to it.

2. Formulating a methodology to resolve the contradiction of UF-

1 delivers UF-2, but causes HF-1.

3. Developing an alternative process setup to carry out UF-1

Cau

ses

is required for

HF-1

UF-3 UF-2 UF-1

Influences

is required for

is required for

Optimizing

the die casting process

parameters

Minimizing the defect

occurrence probability

Minimizing the defective

fraction

Cost and time

consumption

Improving productivity

UF-4

81

The most viable problem among the problem statements was

identified based on the following criteria:

Selection of a problem with the best cost / benefit ratio. The

more radical the problem, the greater the potential benefit.

Elimination of the harmful causes than alleviation of results.

Considering the level of difficulty involved in implementing a

solution. Too radical a solution may prove unacceptable,

depending on an organisation’s culture and psychological

inertia.

Finally, problem definition was made by considering the

organisation’s culture and the expectations of the management:

“Determining the optimum parameter settings of the die

casting process by resolving the contradiction of process

parameter optimization for minimizing the defect

probability but it causes time and cost consumption

incurred for modifying present operational setup”

4.2.5.4 Six Sigma measure stage

The information review and data collection was undertaken at this

stage to measure the present performance of the organisation. The current

process is also quantified then. The following metrics are established to make

this stage more systematic:

82

Output – quantity produced

Defective (r) – casting being rejected due to the presence of

defect(s)

Defect – any non-conformity

Yield (y) – output after rejections

Opportunity (m) – chances of being defective (number of

defects)

Defects probability (p(x))– chances of containing one or more

defect in single casting

Defective fraction – ratio of defective to output

Defects per unit (DPU) – ratio of defective to output

Defects per opportunities (DPO) – ratio of DPU to prevailing

defect opportunities

Defects per million opportunities (DPMO) – DPO multiplied

by 1,000,000

Past history on production and rejection is collected as given in

Table 4.3 from the organisation’s database for a period of three months for

further analysis.

83

Table 4.3 Data from past history

Sl. No.

Item name Part Number

Production Qty Production

Total

Rejection Qty

Nov 07 Dec 07 Jan 08 Nov 07 Dec 07 Jan 08 Rejection Total

Rejection %

1 Cover 1622 3661 00 1966 2159 4125 706 583 1289 31.25 2 Governor cover F002 a31 032 2325 2021 104 4450 642 218 11 871 19.57 3 Governor housing F002 a21 342 15251 7769 17522 40542 2452 796 850 4098 10.11 4 Sre bracket 26216315 1294 869 992 3155 169 153 69 391 12.39 5 Rfi shield housing 21420-21310 8140 8140 929 929 11.41 6 De bracket 2625 9934 2230 2230 246 246 11.03 7 Sre bracket 4606 (r7) 8401 21493 5654 35548 798 2163 653 3614 10.17 8 Stop housing 1425 209 028 688 12392 13080 64 250 314 2.40 9 Rocker arm - 33044 14272 47316 2908 221 3129 6.61

10 Sre bracket 4606 (r6) 12581 1999 14580 940 79 1019 6.99 11 Filter housing 1615 7767 02 3486 1414 4900 245 31 276 5.63 12 Foot rest rh 982023 2965 2965 81 81 2.73 13 Foot rest lh 982022 2965 2965 166 166 5.60 14 Tensioning bracket 4-234-50-0220 9152 9152 455 455 4.97 15 Intermediate housing Z 007557 3890 3966 4126 11982 188 89 172 449 3.75 16 Big housing 1615 7668 02 1211 1827 3038 57 35 92 3.03 17 Valve seat Fdopa125-01c 12832 12832 593 593 4.62 18 Head cylinder casting K010039 5244 5244 235 235 4.48 19 Calliper support 597759 1257 1740 2037 5034 56 125 296 477 9.48 20 Pump housing F002 g11 525 20678 6494 27172 885 1534 2419 8.90

84

Table 4.3 (Continued)

Sl. No.

Item name Part Number

Production Qty Production

Total

Rejection Qty

Nov 07 Dec 07 Jan 08 Nov 07 Dec 07 Jan 08 Rejection Total

Rejection %

21 Governor housing F002 a33 002 4356 1008 5364 186 291 477 8.89 22 Port body voss Z011883 4202 4772 13338 22312 141 151 193 485 2.17 23 Closing cover 1415 626 114 13636 27594 41230 452 375 827 2.01 24 Cover oil pump Fdopa 125-02c 4919 884 5803 154 25 179 3.08 25 Corner 1613 9963 00 17244 17244 497 497 2.88 26 Connection pipe 1503 2854 00 1474 1880 2591 5945 39 141 68 248 4.17 27 De bracket 26216314 1243 1243 30 30 2.41 28 Gear housing drw Sw6s 5 1800 2965 2965 70 70 2.36 29 Closing cover 1421 060 013 24452 20784 45236 558 472 1030 2.28 30 F8b housing 2641 2432 12190 6351 5120 23661 230 570 210 1010 4.27 31 Supersonic grill 903063-10 675 6636 3730 11041 11 110 59 180 1.63 32 Mega sonic grill 905032-10 29025 27490 10818 67333 411 837 128 1376 2.04 33 Knuckle 03-7831-0300 1256 1256 17 17 1.35 34 Ce bracket 2625 9888 7100 7100 85 85 1.20 35 Fixing bracket 2625 2922 6519 2305 8824 74 40 114 1.29 36 Delivery pipe 1503 2853 00 2725 640 3365 30 65 95 2.82 37 Regulating holder 1 49112 49112 521 521 1.06 38 Front wheel hub Ii 560284 5158 11627 7198 23983 53 230 93 376 1.57 39 Tension roller 5/0654 708/0 123558 185025 65827 374410 1266 1945 2951 6162 1.65 40 Valve seat 1503 2867 00 583 589 1172 198 83 281 23.98

85

Table 4.3 (Continued)

Sl. No.

Item name Part Number

Production Qty Production

Total

Rejection Qty

Nov 07 Dec 07 Jan 08 Nov 07 Dec 07 Jan 08 Rejection Total

Rejection %

41 Oil pump body 1503 2865 00 4910 4910 1227 1227 24.99 42 Scherenlanger 4 393 21 0001 688 688 83 83 12.06 43 De bracket 2625 9906 1028 1028 117 117 11.38 44 Flange assly 1615 7768 80 1656 1656 150 150 9.06 45 C111 cover f.unloader 1622 3163 00 3885 3885 217 217 5.59 46 Deckel 4 393 17 0008 1714 1714 47 47 2.74 47 Housing clutch N8070219 10128 10128 251 251 2.48 48 Cover outlet 1616 6507 00 7138 3313 10451 163 112 275 2.63 49 Cover l cylinder head N8010280 6684 6684 126 126 1.89 50 Rear bracket 2621 6569 3167 645 3812 34 47 81 2.12 51 Ce bracket 2625 9925 174 174 15 15 8.62 52 C77 housing unloader 1622 1713 05 4180 4180 255 255 6.10 53 Stop cover 1425 520 033 202 202 9 9 4.46 54 Gear case casting K080049 1120 1120 39 39 3.48 55 De bracket 26214564 8310 8310 250 250 3.01 56 Rear bracket 2621 4309 1656 1656 37 37 2.23 57 Cover C40 1616 7265 2370 2370 33 33 1.39 58 Rear bracket 2621 4406 9407 9407 108 108 1.15

Total Production 1049424 Total Rejection 38523

86

The data review is consolidated as:

• The company has produced 58 different castings as batches

• The net production is 10, 49,424 units (output)

• The total rejections were 38,523 units (defective)

• There were 27 casting defects reported (opportunities)

• Out of 58 varieties of castings, only 6 casting variety had

defective fraction less then target of 1.5%

• Overall defective fraction found ranged between 1.06% to

31.25%

Figure 4.11 illustrates the magnitude level of defective fraction

retrieved from the past data.

Figure 4.11 Defective fractions run chart

87

4.2.5.5 Calculating defect probability using the Poisson distribution

Since the subject case contained number of defectives as discrete

data type with defect opportunities more than two, we used the Poisson

distribution to calculate the defects probability (Park, 2003) as tabulated in

Table 4.4. The probability of observing exactly (x) defects in a single casting

is given by the Equation (4.1).

Table 4.4 Defect probability calculations using the Poisson distribution

S.No. Product name Output Defective DPU P(x=0) Defect probability

% Defect probability

1 Cover 4125 1289 0.31248 0.73163 0.26837 26.84

2 Governor cover 4450 871 0.19573 0.82223 0.17777 17.78

3 Governor housing 40542 4098 0.10108 0.90386 0.09614 9.61

4 Sre bracket 3155 391 0.12393 0.88344 0.11656 11.66

5 Rfi shield housing 8140 929 0.11413 0.89214 0.10786 10.79

6 De bracket 2230 246 0.11031 0.89555 0.10445 10.44

7 Sre bracket 35548 3614 0.10167 0.90333 0.09667 9.67

8 Stop housing 13080 314 0.02401 0.97628 0.02372 2.37

9 Rocker arm 47316 3129 0.06613 0.93601 0.06399 6.40

10 Sre bracket 14580 1019 0.06989 0.93250 0.06750 6.75

11 Filter housing 4900 276 0.05633 0.94523 0.05477 5.48

12 Foot rest rh 2965 81 0.02732 0.97305 0.02695 2.69

13 Foot rest lh 2965 166 0.05599 0.94555 0.05445 5.44

14 Tensioning bracket 9152 455 0.04972 0.95150 0.04850 4.85

15 Inter. housing 11982 449 0.03747 0.96322 0.03678 3.68

16 Big housing 3038 92 0.03028 0.97017 0.02983 2.98

17 Valve seat 12832 593 0.04621 0.95484 0.04516 4.52

18 cylinder casting 5244 235 0.04481 0.95618 0.04382 4.38

19 Calliper support 5034 477 0.09476 0.90960 0.09040 9.04

20 Pump housing 27172 2419 0.08903 0.91482 0.08518 8.52

21 Governor housing 5364 477 0.08893 0.91491 0.08509 8.51

22 Port body voss 22312 485 0.02174 0.97850 0.02150 2.15

23 Closing cover 41230 827 0.02006 0.98014 0.01986 1.99

24 Cover oil pump 5803 179 0.03085 0.96962 0.03038 3.04

25 Corner 17244 497 0.02882 0.97159 0.02841 2.84

26 Connection pipe 5945 248 0.04172 0.95914 0.04086 4.09

88

Table 4.4 (Continued)

S.No. Product name Output Defective DPU P(x=0) Defect

probability % Defect

probability

27 De bracket 1243 30 0.02414 0.97615 0.02385 2.38

28 Gear housing drw 2965 70 0.02361 0.97667 0.02333 2.33

29 Closing cover 45236 1030 0.02277 0.97749 0.02251 2.25

30 F8b housing 23661 1010 0.04269 0.95821 0.04179 4.18

31 Supersonic grill 11041 180 0.0163 0.98383 0.01617 1.62

32 Mega sonic grill 67333 1376 0.02044 0.97977 0.02023 2.02

33 Knuckle 1256 17 0.01354 0.98656 0.01344 1.34

34 Ce bracket 7100 85 0.01197 0.98810 0.01190 1.19

35 Fixing bracket 8824 114 0.01292 0.98716 0.01284 1.28

36 Delivery pipe 3365 95 0.02823 0.97216 0.02784 2.78

37 holder 49112 521 0.01061 0.98945 0.01055 1.06

38 Front wheel hub 23983 376 0.01568 0.98444 0.01556 1.56

39 Tension roller 374410 6162 0.01646 0.98368 0.01632 1.63

40 Valve seat 1172 281 0.23976 0.78682 0.21318 21.32

41 Oil pump body 4910 1227 0.2499 0.77888 0.22112 22.11

42 Scherenlanger 688 83 0.12064 0.88635 0.11365 11.36

43 De bracket 1028 117 0.11381 0.89242 0.10758 10.76

44 Flange assly 1656 150 0.09058 0.91340 0.08660 8.66

45 f.unloader 3885 217 0.05586 0.94568 0.05432 5.43

46 Deckel 1714 47 0.02742 0.97295 0.02705 2.70

47 Housing clutch 10128 251 0.02478 0.97552 0.02448 2.45

48 Cover outlet 10451 275 0.02631 0.97403 0.02597 2.60

49 Cover l cylinder head 6684 126 0.01885 0.98133 0.01867 1.87

50 Rear bracket 3812 81 0.02125 0.97898 0.02102 2.10

51 Ce bracket 174 15 0.08621 0.91740 0.08260 8.26

52 f.unloader 4180 255 0.061 0.94082 0.05918 5.92

53 Stop cover 202 9 0.04455 0.95642 0.04358 4.36

54 Gear case casting 1120 39 0.03482 0.96578 0.03422 3.42

55 De bracket 8310 250 0.03008 0.97036 0.02964 2.96

56 Rear bracket 1656 37 0.02234 0.97790 0.02210 2.21

57 Cover 2370 33 0.01392 0.98617 0.01383 1.38

58 Rear bracket 9407 108 0.01148 0.98858 0.01142 1.14

89

xeP(X x) p(x) , x 0, 1, 2, ... m

x!

(4.1)

where

e = a constant equal to 2.71828

λ = defects per unit (DPU)

m = number of independent defect opportunities = 27 for this case

To better relate the Poisson distribution to this case analysis, the Equation

(4.1) is rewritten as Equation (4.2) which can be effectively used when

DPU / m is less than 10% and m is relatively large (m = 27) (Park, 2003):

DPU xe DPUP(X x) p(x) , x 0, 1, 2, ... m

x!

(4.2)

In this case, our interest is focused on minimising the defect

probability which is calculated using the Equation (4.3):

Defect probability p(x) = 1 – probability of having zero defects (4.3)

DPU 0e DPU1

0!

DPU1 e

It is obvious from Figure 4.12 that the concentration of defect

probability of components is noticed as being below 12%. Hence, it is

benchmarked for team members to work-out the future actions towards

minimising it further, below 12% as much as possible.

90

Figure 4.12 Concentration of defect probability

4.2.5.6 Process yield and Sigma quality level calculation

When a given set of data is continuous, the mean and standard

deviation can obtained easily. Also the sigma level can be calculated from the

specification limits. But in this case analysis, the data belongs to discrete type.

Hence it is converted into yield and the corresponding Sigma level is obtained

using the standard normal distribution Equation (4.4) (Park, 2003).

2wZ

21(Z) e dw2

(4.4)

= y

91

The Sigma level Z corresponding to yield ‘y’ is obtained from

standard normal distribution table given in the Appendix -1. Since the value

of Z represented the long term Sigma level, then, the short term Sigma level is

obtained using the Equation (4.5) and summarized in Table 4.5.

Zs = Zl + 1.5 (4.5)

Table 4.5 Yield and sigma level calculation using standard normal

distribution

Sl. No. Product name Output

(n) Defective

(r)

Defective fraction p = r / n

Yield y = 1 - p

Standard normal variant

(Zs)

Sigma level Zs =

Zl +1.5

1 Cover 4125 1289 0.31248 0.68752 2.46 3.96

2 Governor cover 4450 871 0.19573 0.80427 2.405 3.905

3 Governor housing 40542 4098 0.10108 0.89892 2.36 3.86

4 Sre bracket 3155 391 0.12393 0.87607 2.37 3.87

5 Rfi shield housing 8140 929 0.11413 0.88587 2.37 3.87

6 De bracket 2230 246 0.11031 0.88969 2.37 3.87

7 Sre bracket 35548 3614 0.10167 0.89833 2.36 3.86

8 Stop housing 13080 314 0.02401 0.97599 2.33 3.83

9 Rocker arm 47316 3129 0.06613 0.93387 2.35 3.85

10 Sre bracket 14580 1019 0.06989 0.93011 2.35 3.85

11 Filter housing 4900 276 0.05633 0.94367 2.35 3.85

12 Foot rest rh 2965 81 0.02732 0.97268 2.34 3.84

13 Foot rest lh 2965 166 0.05599 0.94401 2.35 3.85

14 Tensioning bracket 9152 455 0.04972 0.95028 2.34 3.84

15 Intermediate housing 11982 449 0.03747 0.96253 2.34 3.84

16 Big housing 3038 92 0.03028 0.96972 2.335 3.835

17 Valve seat 12832 593 0.04621 0.95379 2.335 3.835

18 Head cylinder casting 5244 235 0.04481 0.95519 2.335 3.835

19 Calliper support 5034 477 0.09476 0.90524 2.365 3.865 20 Pump housing 27172 2419 0.08903 0.91097 2.36 3.86

92

Table 4.5 (Continued)

Sl. No. Product name Output

(n) Defective

(r)

Defective fraction p = r / n

Yield y = 1 - p

Standard normal variant

(Zs)

Sigma level Zs =

Zl +1.5

21 Governor housing 5364 477 0.08893 0.91107 2.36 3.86 22 Port body voss 22312 485 0.02174 0.97826 2.335 3.835 23 Closing cover 41230 827 0.02006 0.97994 2.335 3.835 24 Cover oil pump 5803 179 0.03085 0.96915 2.335 3.835 25 Corner 17244 497 0.02882 0.97118 2.335 3.835 26 Connection pipe 5945 248 0.04172 0.95828 2.33 3.83 27 De bracket 1243 30 0.02414 0.97586 2.33 3.83 28 Gear housing drw 2965 70 0.02361 0.97639 2.33 3.83 29 Closing cover 45236 1030 0.02277 0.97723 2.33 3.83 30 F8b housing 23661 1010 0.04269 0.95731 2.34 3.84 31 Supersonic grill 11041 180 0.01630 0.98370 2.335 3.835 32 Mega sonic grill 67333 1376 0.02044 0.97956 2.335 3.835 33 Knuckle 1256 17 0.01354 0.98646 2.33 3.83 34 Ce bracket 7100 85 0.01197 0.98803 2.33 3.83 35 Fixing bracket 8824 114 0.01292 0.98708 2.33 3.83 36 Delivery pipe 3365 95 0.02823 0.97177 2.335 3.835 37 Regulating sleeve 49112 521 0.01061 0.98939 2.33 3.83 38 Front wheel hub 23983 376 0.01568 0.98432 2.335 3.835 39 Tension roller 374410 6162 0.01646 0.98354 2.335 3.835 40 Valve seat 1172 281 0.23976 0.76024 2.43 3.93 41 Oil pump body 4910 1227 0.24990 0.75010 2.43 3.93 42 Scherenlanger 688 83 0.12064 0.87936 2.375 3.875 43 De bracket 1028 117 0.11381 0.88619 2.37 3.87 44 Flange assly 1656 150 0.09058 0.90942 2.36 3.86 45 C111 f.unloader 3885 217 0.05586 0.94414 2.35 3.85 46 Deckel 1714 47 0.02742 0.97258 2.335 3.835 47 Housing clutch 10128 251 0.02478 0.97522 2.335 3.835 48 Cover outlet 10451 275 0.02631 0.97369 2.335 3.835 49 Cover l cylinder head 6684 126 0.01885 0.98115 2.335 3.835 50 Rear bracket 3812 81 0.02125 0.97875 2.335 3.835 51 Ce bracket 174 15 0.08621 0.91379 2.36 3.86 52 C77 housing f.unloader 4180 255 0.06100 0.93900 2.35 3.85 53 Stop cover 202 9 0.04455 0.95545 2.34 3.84 54 Gear case casting 1120 39 0.03482 0.96518 2.34 3.84 55 De bracket 8310 250 0.03008 0.96992 2.34 3.84 56 Rear bracket 1656 37 0.02234 0.97766 2.335 3.835 57 Cover 2370 33 0.01392 0.98608 2.33 3.83 58 Rear bracket 9407 108 0.01148 0.98852 2.33 3.83

93

It is inferred from Table 4.5 that the majority of the component

casting processes found operating with a yield equal to a Sigma level between

3.76 and 3.84 as illustrated in Figure 4.13.

Figure 4.13 Sigma quality level radar chart

4.2.5.7 Defects stratification using the Pareto principle

It becomes imperative to stratify the defects for distinguishing the

“vital few from the trivial many”. To spot the dominant defects in the present

process setup, we applied the Pareto principle. The contribution of each defect

is summarised in Table 4.6. It is noticed from the Pareto chart in Figure 4.14

that the defects 1 to 5 (Un-filling; Blow holes; Gate broken; Damage; and

94

Flash) are found contributing to nearly 78.63% of the overall rejections.

Improving the process by optimising the parameters to reduce these defects is

felt necessary and imperative to improve productivity significantly.

Table 4.6 Summary of defects data under Pareto principle

S.No Defects

Quantity rejected

Tota

l rej

ecte

d

% R

ejec

tions

Cum

ulat

ive

freq

uenc

y

Cum

ulat

ive

%

Month 1 Month 2 Month 3

1 Un-filling 4793 3965 2918 11676 29.21 11676 29.21 2 Blow holes 5896 3618 1333 10847 27.14 22523 56.35 3 Gate broken 1596 1887 801 4284 10.72 26807 67.07 4 Damage 1005 1140 662 2807 7.02 29614 74.09 5 Flash 1104 514 197 1815 4.54 31429 78.63 6 weld 699 771 345 1815 4.54 33244 83.17 7 Bend 974 621 102 1697 4.25 34941 87.42 8 Crack 231 495 237 963 2.41 35904 89.83 9 Un-wash 77 396 251 724 1.81 36628 91.64

10 Shrinkage 238 384 13 635 1.59 37263 93.23 11 White rust 258 242 1 501 1.25 37764 94.48 12 Handling damage 55 208 27 290 0.73 38054 95.21 13 Insert blow holes 124 75 16 215 0.54 38269 95.75 14 Dimension problem 50 95 63 208 0.52 38477 96.27 15 Pin broken 115 84 1 200 0.50 38677 96.77 16 Insert damage 59 84 48 191 0.48 38868 97.25 17 Gate hole 96 46 31 173 0.43 39041 97.68 18 Bubbles 20 99 54 173 0.43 39214 98.11 19 Metal peel off 52 83 35 170 0.43 39384 98.54 20 Dent mark 78 81 0 159 0.40 39543 98.93 21 Ovality 81 25 0 106 0.27 39649 99.20 22 Ej. pin projection 11 15 55 81 0.20 39730 99.40 23 Rib broken 39 32 8 79 0.20 39809 99.60 24 Bush exposer 9 68 2 79 0.20 39888 99.80 25 Insert offset 21 9 12 42 0.11 39930 99.90 26 Insert flash 15 6 4 25 0.06 39955 99.96 27 Without insert 2 9 3 14 0.04 39969 100.00 17698 15052 7219 39969

95

Figure 4.14 Pareto chart of defects

4.2.5.8 Six Sigma analyse stage

Recalling the function modelling depicted in Figure 4.10 at the

define stage; optimising the process parameters is mandatory to produce

another useful function of defect occurrence probability minimisation, while

the UF1 creates the harmful function HF1 and entails cost and time

consumption.

Before attempting to optimise the process parameters, it is needed

to resolve the contradiction for which the TRIZ contradiction algorithm is

executed as exposed in Figure 4.15.

96

Abstract Thought

Analogical Thought

Problem solved

Contradiction Matrix

Generic problem (39 Problem parameters)

2

Specific problem

1

Generic Solution

(40 inventive principles)

3

Specific Solution

4

Specific Problem

Optimizing the

process parameters without cost

and time consumption

Figure 4.15 Technical contradiction algorithms

In the contradiction algorithm, the present contradiction is mapped

as specific problem and abstracted to a generic problem. Figure 4.16

illustrated the process of abstracting using the TRIZ 39 problem parameters.

Specific problem Generic problem parameter

UF Process parameter

optimization

Quantity of substance

(26)

HF Cost and time

consumption

Ease of operation

(33)

Figure 4.16 TRIZ abstracting process

97

The following generic problem parameters are chosen from

Appendix 2 :

No: 26 – Quantity of substance

The number or amount of a system’s materials,

substances, parts or subsystems that might be changed

fully or partially, permanently or temporarily

No: 33 – Ease of operation

The process is NOT easy if it requires a large number of

people, large number of steps in the operation, needs

special tools, etc.

By cross referring the Quantity of substance (26) with Ease of

operation (33) in Triz contradiction matrix shown in Appendix-3, the

inventive principles marked in Figure 4.17 are chosen to develop a general

solution.

HF Ease of operation

(33) UF

Quantity of

substance

(26)

(35) (29)

(25) (10)

Figure 4.17 TRIZ contradiction matrix

The identified solution principles are (see Appendix-4):

Number 35: Parameter change is the principle that focused the

team to change the parameter in one or other way like:

98

Changing the object’s physical state e.g. to a gas, liquid or

solid

Changing the concentration or consistency

Changing the degree of flexibility

Changing the temperature

Number 29: Pneumatics and hydraulics insists on the use of gas

or liquid parts of an object instead of the solid parts.

Number 25: Self service makes an object serve itself by

performing auxiliary helpful functions.

Number 10: Preliminary action entailed the performance, before

it is needed, the required change of an object either fully or partially.

The inventive principle No 10: Preliminary action is chosen to

develop a generic solution because of its relevance to the present problem

solving scenario. As per this principle, it is decided to preset the operation

parameters to the trail level at the start of each batch of production. Hence it is

eliminated to spend extra time to conduct optimisation and thereon the cost

associated.

To determine the parameters, we analysed the defects data

summarized in Table 4.6. Defects No. 3 and 4 (Gate broken and Damage) are

found occurring only after the casting process. These defects occurred due to

mishandling of the castings while trimming and transporting. The rest of the

defects (1-Un-filling, 2-Blow holes and 5-Flash) are noticed to have occurred

during the casting operation. A cause and effect analysis, illustrated in

Figure 4.18 (a, b, c) is performed to identify the parameters that influenced

the occurrence of these process defects.

99

Blow holes

Insufficient degassing

(Degassing frequency)

Non-uniform cooling rate

(Frequency of die coat) High metal temperature

(Metal temperature)

Improper metal mixing (Mixing ratio)

High injection pressure (Injection pressure)

Flash

Poor die halves matching (Die design parameter)

Low clamping force (Die design parameter)

Too high injection pressure (Injection pressure)

Un-filling

Insufficient shot volume

(IInd phase turns)

Slow injection (Injection pressure)

Low pouring temperature

(Metal temperature) (a)

(b)

(c)

Figure 4.18 (a, b, c) Defect cause analysis

The following process parameters are selected from the cause and

effect analysis for further study since these parameters are frequently

modified in each process setup to cast a different variety of components:

Metal temperature (Tm): The present process setup saw the

fresh metal being melted in the furnace at the set temperature

100

before it is poured into the die cavity. The temperature ranges

from 6250 C to 9000C dependent upon the component design.

Injection pressure (Pi): It refers to the pressure at which the

molten metal is pushed into the die cavity from the shot

chamber. The normal working injection pressure ranges from

190 kg/cm2 to 270 kg/cm2

IInd phase turns (Ns): After injecting the molten metal inside

the die, a second injection called “Cushioning” is applied to

ensure complete packing of the metal inside the cavity. This

cushioning effect could be adjusted by means of a lead screw

provided in the shot chamber of the machine.

Degassing frequency (fg): To evaporate the gas content in the

molten metal, at a predefined interval, a degassing agent is

mixed with the molten metal in the furnace. The present

practice entailed degassing at every 200 to 350 shots for

different components.

Metal mixing ratio (Rm): After the trimming operation, the

scrap metal is reused with fresh raw metal in the furnace. The

necessary reconditioning is performed before mixing the scrap

metal with fresh metal. The company followed a ratio of 80:20

(80 of new metal is mixed with 20% of scrap metal) for

casting several component varieties.

The die coating frequency is fixed as it is only applied at the

beginning of each shot and hence, not considered for optimisation. Injection

time and shot volume are not considered because those parameters are

101

associated with the equipment in which they are designed at a predefined

value.

4.2.6 TPE stage 3: Implementing and control processes

Taguchi’s Orthogonal Array (OA)-based Design of Experiment

(DoE) is employed to optimise the parameters in the foregoing analysis. The

reason driving the selection of Taguchi’s DoE approach is to complete the

investigation with minimal trail experiments over the traditional full factorial

experimental approach. It is planned that one component would be selected in

such a way that it would represent the entire range of components given the

company faces defects invariably in all components. Figure 4.12 infers that

the average defect probability is around 12%. The components possessing

higher defect probability are targeted to optimise the process parameters.

Such components and their defect contributions are listed in Table 4.7. The

component “Oil pump body” (No.41) having DPU at the rate of 25% with

occurrence probability near 22% is selected for process optimisation.

Table 4.7 List of components having high defect probability and their

defect contribution

S.No (as on

Table 4.4)

Component name

DPU (as on

Table 4.4)

% defect probability (as on Table 4.4)

Process defects Other defects Un-filling

Blow holes

Flash

1 Cover 0.31248 26.84 32 755 - 502

2 Governor cover

0.19573 17.78 326 420 - 125

40 Valve seat 0.23976 21.32 - 236 - 45

41 Oil pump body 0.24990 22.11 40 1073 2 112

102

4.2.6.1 DoE Step 1: determination of factor levels

All the parameters are considered across three levels to

accommodate the non-linear relationship between factors as shown in

Table 4.8.

Table 4.8 Process parameters with levels

Factor

Notation

Controllable

factors Metric

Level

1

Level

2

Level

3

Tm Metal temperature. Centigrade 665 690 715

Pi Injection pressure. Kg/cm2 220 240 260

fg Degassing frequency. Shots per degassing 320 240 160

Ns II phase turns. Nos. 3 3.25 3.5

Rm Metal mixing ratio. Ratio 80:20 70:30 60:40

4.2.6.2 DoE Step 2: Selection of orthogonal array

It is interested to study the interaction effects on the component

with respect to defect occurrence:

Metal temperature and injection pressure [Tm x Pi]

Injection pressure and degassing frequency [Pi x fg].

Metal temperature and degassing frequency [Tm x fg].

L27, a three level orthogonal array is chosen since it had a greater

degree of freedom (DOF=27) than that of the factors and interactions

(DOF=22) as computed in Table 4.9.

103

Table 4.9 Degrees of freedom

Factors /

Interactions

DOF

(No of level -1)

Tm 2

Pi 2

fg 2

Ns 2

Rm 2

Tm x Pi 2 x 2

Pi x fg 2 x 2

Tm x fg 2 x 2

Total DOF 22

4.2.6.3 DoE Step 3: Arranging factors and interactions in L27 OA

columns

The main factors Tm, Pi, fg, Ns, and Rm are assigned to columns 1,

2, 5, 9 and 10 respectively, and the interactions Tm x Pi, Pi x fg and Tm x fg in

columns 3 and 4, 8 and 11, and 6 and 7 respectively using the linear graphs

and triangular tables (see Appendix - 5) and the resultant trail matrix is

prepared as shown in Table 4.10.

104

Table 4.10 Factor assigned to L27 orthogonal array

Te

st N

o. Columns

Objective function

S/N

rat

io

Tm Pi Tm Pi Tm Pi fg Tm fg Tm fg Pi fg Ns Rm Pi fg * * Run 1

Run 2 1 2 3 4 5 6 7 8 9 10 11 12 13

1 1 1 1 1 1 1 1 1 1 1 1 1 1 y1,1 y2,1 S/N1

2 1 1 1 1 2 2 2 2 2 2 2 2 2 y1,2 y2,2 S/N2

3 1 1 1 1 3 3 3 3 3 3 3 3 3 y1,3 y2,3 S/N3

4 1 2 2 2 1 1 1 2 2 2 3 3 3 y1,4 y2,4 S/N4

5 1 2 2 2 2 2 2 3 3 3 1 1 1 y1,5 y2,5 S/N5

6 1 2 2 2 3 3 3 1 1 1 2 2 2 y1,6 y2,6 S/N6

7 1 3 3 3 1 1 1 3 3 3 2 2 2 y1,7 y2,7 S/N7

8 1 3 3 3 2 2 2 1 1 1 3 3 3 y1,8 y2,8 S/N8

9 1 3 3 3 3 3 3 2 2 2 1 1 1 y1,9 y2,9 S/N9

10 2 1 2 3 1 2 3 1 2 3 1 2 3 y1,10 y2,10 S/N10

11 2 1 2 3 2 3 1 2 3 1 2 3 1 y1,11 y2,11 S/N11

12 2 1 2 3 3 1 2 3 1 2 3 1 2 y1,12 y2,12 S/N12

13 2 2 3 1 1 2 3 2 3 1 3 1 2 y1,13 y2,13 S/N13

14 2 2 3 1 2 3 1 3 1 2 1 2 3 y1,14 y2,14 S/N14

15 2 2 3 1 3 1 2 1 2 3 2 3 1 y1,15 y2,15 S/N15

16 2 3 1 2 1 2 3 3 1 2 2 3 1 y1,16 y2,16 S/N16

17 2 3 1 2 2 3 1 1 2 3 3 1 2 y1,17 y2,17 S/N17

18 2 3 1 2 3 1 2 2 3 1 1 2 3 y1,18 y2,18 S/N18

19 3 1 3 2 1 3 2 1 3 2 1 3 2 y1,19 y2,19 S/N19

20 3 1 3 2 2 1 3 2 1 3 2 1 3 y1,20 y2,20 S/N20

21 3 1 3 2 3 2 1 3 2 1 3 2 1 y1,21 y2,21 S/N21

22 3 2 1 3 1 3 2 2 1 3 3 2 1 y1,22 y2,22 S/N22

23 3 2 1 3 2 1 3 3 2 1 1 3 2 y1,23 y2,23 S/N23

24 3 2 1 3 3 2 1 1 3 2 2 1 3 y1,24 y2,24 S/N24

25 3 3 2 1 1 3 2 3 2 1 2 1 3 y1,25 y2,25 S/N25

26 3 3 2 1 2 1 3 1 3 2 3 2 1 y1,26 y2,26 S/N26

27 3 3 2 1 3 2 1 2 1 3 1 3 2 y1,27 y2,27 S/N27

105

4.2.6.4 DoE Step 4: Execution of experiments

As planned upon the contradiction analysis to conduct the

experimental trails along with usual production run, the operator is given the

parameter levels for each trail run and instructed to carry out the casting

process of oil pump body. The operator used to preset the parameters at the

start of the operations and continued that particular production run until they

stopped. The next trail run is carried out the next day. Likewise, all the 27

trial runs are completed in a span of three and a half months and at the end of

each trail, 500 components are randomly chosen in two replications. Since the

number of good components (yi) is recorded as a response in each test, the

“Higher is Best” S/N ratio characteristic is selected (Ross, 1989) and

calculated using the Equation (4.6) and the results are recorded as shown in

Table 4.11.

Table 4.11 Experiment response with S/N ratio

Test No.

No of Good items out of 500 S/N

ratio

Test No.

No of Good items out of 500 S/N

ratio

Run 1 Run 2 Run 1 Run 2 1 401 452 52.55 15 458 478 53.39

2 462 450 53.17 16 490 486 53.76

3 421 436 52.63 17 369 389 51.56

4 390 401 51.94 18 385 368 51.5

5 369 354 51.15 19 436 455 52.97

6 469 476 53.48 20 395 421 52.19

7 485 468 53.55 21 459 462 53.26

8 359 346 50.93 22 463 475 53.42

9 310 264 49.07 23 401 426 52.31

10 418 431 52.55 24 485 476 53.63

11 479 469 53.51 25 495 486 53.81

12 352 378 51.22 26 425 435 52.66

13 301 320 49.82 27 465 452 53.22

14 329 365 50.77

106

r

2i 1HB i

S 1 110 logN r y

(4.6)

where

r = no of trials

yi = response chosen

4.2.6.5 DoE Step 5: Data analysis

The data collected in two replicates of 27 trials are analysed and the

mean values of the response and S/N ratios for the main factors and

interactions are calculated as shown in Table 4.12. The mean response for

factor Tm at level 1 is calculated by averaging the responses of tests in which

the factor Tm is kept at level 1.

Table 4.12 Mean response and mean S/N ratio

Factors / levels L1 L2 L3 Factors /

levels L1 L2 L3

Tm 406.3 403.6 450.7 Tm 52.06 52.02 53.06 Pi 432.1 413.1 415.4 Pi 52.7 52.22 52.23 fg 436.3 402.4 421.9 fg 52.7 52.03 52.38 Ns 420.8 419.4 420.4 Ns 52.40 52.34 52.39 Rm 419.7 410.5 430.4 Rm 52.36 52.14 52.63 x Pi 442.4 425.5 405.4 Tm x Pi 52.60 52.50 52.00

Pi x fg 412.2 436.1 412.3 Pi x fg 52.20 52.70 52.20 Tm x fg 425.4 420.9 414.2 Tm x fg 52.60 52.40 52.20

Thus, the mean response of factor Tm at L1 =

= (401 462 421 390 369 469 485 359 310) 1(452 450 436 401 354 476 468 346 264) 2 9

= 406.3

107

Similarly, the mean response and the mean S/N ratio values for all

the factors and interactions at each level are also calculated.

4.2.6.6 DoE Step 6: Response curve analysis

The response curves pertaining to the mean response and the mean

S/N ratio values are plotted in Figure 4.19 against each level of the factors to

represent the change in the performance characteristics for the variation in

factor levels.

Figure 4.19 Response curves for process parameters

108

From Figure 4.19, we identified the optimum factor levels based on

the “Higher is Best” S/N ratio characteristic and listed these in Table 4.13.

Table 4.13 Optimum factor settings

Notation Factor Optimum

level Value

Tm Metal Temperature L3 7150 C

Pi Intensifier pressure L1 220 Kg/cm2

fg Degassing frequency L1 320 shots/degas

Ns II phase turns L1 3 no

Rm Metal mixing ratio L3 60:40

4.2.6.7 DoE Step 7: Mean response predicting

The confident interval and the mean response are estimated using

the Equation (4.7) to validate the optimum factor level setting:

µGood = Tm3 + Pi1 + fg1 + Ns1 + Rm3 - 4Mgood (4.7)

where

Tm3 - mean response at level 3 of factor Tm

Pi1 - mean response at level 1 of factor Pi

fg1 - mean response at level 1 of factor fg

Ns1 - mean response at level 1 of factor Ns

Rm3 - mean response at level 3 of factor Rm

Mgood - overall mean response value

109

By using the mean response values from Table 10, we calculated

the estimated means:

µGood = 450.7 + 432.1 + 436.3 + 420.8 + 430.4 - 4(420.18)

µGood = 489.58

The confidence interval for the population is calculated using the

Equation (4.8).

;1;Ve ee

1CI F Vn

(4.8)

where

;1;VeF = F ratio required for α (risk)

ve = error degree of freedom

Ve = error variance

ne = experiment trails = 54

In this study,

α risk is taken as 0.10

Confidence = 1 - risk

ve – degrees of freedom for error variance is 41 from table 5.1

Ve = 1929.65 from table 5.1

Fα (1, 41) = 2.84 (taken from f – ratio table)

110

Hence,

1CI (2.84) (1929.65)54

CI = 10.1

The estimated mean response is 489.58 and at 90% CI, the

predicted optimum output is estimated as:

[µGood - CI] < µGood

< [µGood + CI]

[489.58 – 10.1] < µGood < [489.58 + 10.1]

479.5 < µGood < 499.7

4.3 SIX SIGMA CONTROL STAGE

The real challenge for this research approach lies not in making

improvements to the process but in providing sustained improvement in

organisational productivity through process optimisation. Standardisation,

constant monitoring and control of the optimized process are needed to

maintain the improvements. Process control limits are obtained using the

optimum parameter levels for maintaining the process to free of defects.

Implementation of the aforementioned optimum factor levels resulted in an

improvement of process yield and reduction in defect probability. Moreover,

an iterative fashion in implementing the TPE model is found to be more

indispensable.

An extensive training programme for the personnel connected by

the process changes is conducted within the company to ease the

implementation of the TPE model. It is well known that real improvement

comes only from the shop floor. Process sheets and control charts are made so

111

that the operators could be prepared to take preventive action before the

critical process parameters and critical performance characteristics strayed

outside of the control limits. A complete database is prepared to maintain the

improvements to the results. Proper monitoring of the process helped to detect

and correct out-of-control signals before they resulted in a loss of

productivity.

4.4 CASE IMPLEMENTATION – 2

4.4.1 About the Company

Company ABC Ltd. (real name of the company is not revealed

here as per the Memorandum signed with the management) is an Authorized

Service Centre for TATA Motors’ range of passenger cars since 1995.

Located in a major industrial city in southern India, it is an exclusive five star

rated service centre by TATA Motors to cater to the service needs of the

TATA range of passenger cars—Sumo Grande, Safari, Aria, Indica, Vista,

Manza and Nano—with comprehensive in-house facilities and trained

manpower. The company services about 750 TATA cars every month and

puts in the best effort to offer a delightful service experience to each and

every customer.

The company provides the best possible assistance through a

dedicated and trained staff so that the customer can get more value out of the

car at all times. The primary services offered by this organisation are basic

services, mechanical overhauling, body repair works, painting and polishing,

followed by documenting services like insurance, reimbursement in case of

accidents and maintenance of previous service records. The service centre is

equipped with special tools and TATA original parts to support scheduled

maintenances, the running of repairs, major overhauls, body repairs and

painting jobs.

112

Design data collection medium

Identify target

customers

Selection of customers

and competitors

Customer data

collection

Data analysis

Priorit izing requirements

Ranking and weighing requirements

Choosing rectification

factors

Customer input planning

Customer centric process

Company centric process

Constructing HOQ

Ranking and

weighing

Correlation

Cus

tom

er p

lann

ing

mat

rix

Com

pany

pla

nnin

g m

atrix

Rel

atio

nshi

p m

atrix

Cor

rela

tion

mat

rix

Selection of Customer CTQ

QFD Process

4.4.2 Scope of the Study

The company has to be responsive to all customer issues and react

swiftly when satisfying them. Though the customised approaches are put into

effect at the service centre to address customer problems, the dynamic

behaviour of customer grievances made it difficult to address all of them

successfully. Given this situation, we implemented the TPE model designed

by this research study to improve customer satisfaction.

4.4.3 QFD Process

The activities undertaken in this step are summarised in the flow

chart depicted in Figure 4.20.

Figure 4.20 Activities in QFD process

113

This process started with customer input planning and progressed

with customer voice collection and analysis; rectification factors

identification, ranking and correlating the rectification factors, and ended with

the construction of a final HOQ. To begin with, the TPE model

implementation comprised of a series of meetings has been conducted in the

company with a deployment team being organised to introduce the model to

the company.

4.4.3.1 Customer input planning

Customer input planning involved the design of a data collection

medium, identification of target customers, choosing potential competitors

and examining the volume of data to be collected. There are many data

collection mediums available to collect customer requirements. The most

preferred way, however, is the Survey Design to record the real thoughts of

the customers. An important task in this phase entailed identifying “Who is

the customer?” for which, a brainstorming session is conducted to arrive at a

decision. A conclusion is drawn to collect information from the customers

visit the company to service their cars. It is decided that two potential

competitors C1 and C2 who are the real thrush-hold service providers on par

with company ABC Ltd. would be included. A survey is undertaken for 10

days within premise.

4.4.3.2 Customer centric process

The customers are supplied with the questionnaire format enquiring

about their present complaint/reason for their dissatisfaction, what they

expected, how the aforementioned competitors are performing compared to

ABC Ltd. and the latter’s overall performance.

114

The questionnaire was designed with nine headings: service

initiation, service advisor, in-service experience, service delivery, service

quality, user-friendly service, problems experienced, value added services and

service cost. Each section had three questions with a 10 point scale rating

starting from “1” for unacceptable quality; “5” for average quality and “10”

for outstanding quality. The customers are requested to rate the performance

of ABC Ltd. At the end of day 10, out of 430 customers visiting ABC Ltd.,

398 survey forms are filled and collected by the team.

The reason 32 forms are missing could be attributed to the lack of

interest on the part of the customers to provide information, or because the

customers are new. The survey forms are analysed as per the different aspects

illustrated in Figure 4.21 to assess the reliability of the data collected.

Figure 4.21 Survey contributions

115

It is noted that a major share of the contributions in the survey is

occupied by the most experienced operators. It is understood from the figures

that 42% of the input is from the operators who covered more than 1, 00,000

kilometres driving of their cars. Moreover, 50% of the survey results are

obtained from vehicles aged more than three years.

These two factors are assumed as the criteria for service

complaints. With respect to the levels of satisfaction, we proposed the

“Customer Satisfaction Index” (CSI) metric. If a customer is fully satisfied in

all aspects, then the customer can award rating “10” for all the 27 questions.

Hence, a total of 270 would indicate his CSI level (CSI = 270). All

the 398 results are arranged into four groups with respect to the percentage of

CSI as shown in Figure 4.22. The individual group analysis showed that the

average CSI is 19% for Group 1, 39% for Group 2, 65% for Group 3 and 88%

for Group 4 as given in Figure 4.23.

Figure 4.22 Customer satisfaction index

116

Figure 4.23 Average customer satisfaction index

The Customer-Competitive Performance (CCP) analysis is

performed to compare the performance of ABC Ltd with its competitors. The

statistics have been displayed in Table 4.14.

117

Table 4.14 Competitive performance analysis

Q.No. Question Company

ABC Ltd Competitor

C1 Competitor

C2 CSI Factor – Service Initiation

1 Rate time to speak service advisor 7.2 7.7 7.6 2 Reasonable time to schedule visit 9.0 8.5 8.9 3 Overall service initiation 8.3 8.7 8.6 CSI Factor – Service Advisor (SA)

4 Knowledge expertise 7.3 8.6 8.4 5 Understood problem with vehicle 7.1 8.3 8.9 6 Explanation of service to be done 8.4 8.9 8.1 CSI Factor – In-Service experience

7 Customer lounge amenities 7.3 7.1 7.2 8 Customer lounge comfort 8.5 8.1 8.4 9 Customer lounge cleanliness 9.1 8.9 8.2 CSI Factor – Service delivery

10 Promptness in delivering vehicle 7.4 9.1 8.9 11 Processing for paying for service 8.4 8.1 8.4 12 Time to service vehicle 7.5 8.4 7.9

CSI Factor – Service Quality 13 Thoroughness in fulfilling requests 7.4 8.5 7.2 14 Quality of service performed 8.3 9.3 8.9 15 Availability of spare parts 9.2 9.1 9.3

CSI Factor – User-Friendly Service 16 Convenience of operating hours 9.2 9.1 8.3 17 Fairness of charges 8.4 8.8 8.1 18 Annual maintenance contract 7.6 8.9 8.2

CSI Factor – Problem experienced 19 Trouble free operations 8.4 8.8 8.1 20 Freedom from squeak and rattle 9.3 9.1 7.9 21 Ease of maintenance and repair 8.1 8.9 7.9

CSI Factor – Value added services (VAS) 22 On-road assistance 7.6 8.1 8.8 23 Mishap partnership 7.2 8.4 8.4 24 Provision of float units 8.3 8.2 8.5

CSI Factor – Service cost 25 Repair cost 7.9 8.3 8.1 26 Credit facility 7.2 7.1 7.0 27 Payment channel 9.0 8.9 9.2

118

The CCP analysis shown in Figure 4.24 (a-i) confirmed that the

performance of ABC Ltd. is found to be rated poor for questions 1, 3, 4, 5, 10,

12, 14, 18, 22, 23, 24, and 25 compared to the ratings given to its competitors.

(a) CSI Factor – Service Initiation

(b) CSI Factor – Service Advisor

119

(c) CSI Factor – In-Service Experience

(d) CSI Factor – Service Delivery

120

(e) CSI Factor – Service Quality

(f) CSI Factor – User Friendly Service

121

(g) CSI Factor – Problem Experienced

(h) CSI Factor – Value Added Services

122

(i) CSI Factor – Service Cost

Figure 4.24 (a-i) CSI factor performance analyses

After a meeting with the company’s management, it is decided that

we would concentrate on the expectations of the customers with regard to

those questions. A series of brainstorming sessions are arranged for

inter-department personnel to sort out Customer Requirements (CR) and

possible CRs and their critical nature is projected in Table 4.15. While

examining Customer Requirements, it is noticed that some of them included a

renewal of the Annual Maintenance Contract, time to meet service advisors,

and providing float units, which are directly linked to the management’s

policy decisions. Hence, it is decided that further action in these cases would

be left to the top management. The selected CRs are listed in Table 4.16.

123

Table 4.15 Customer requirements

Q.No Critical Questions Customer

rating Customer Requirement

1 Rate time to speak service advisor 7.2 Less time to meet

3 Overall service initiation 8.3 Improve service initiation

4 Knowledge expertise 7.3 Improve technical expertise

5 Understood problem with vehicle 7.1 Improve technical expertise

10 Promptness in delivering vehicle 7.4 Prompt vehicle delivery

12 Time to service vehicle 7.5 Minimum time to service

14 Quality of service performed 8.3 Improve service quality

18 Annual maintenance contract 7.6 Revise the policy

22 On-road assistance 7.6 Quick on-road assistance

23 Mishap partnership 7.2 Quick and proper assistance

24 Provision of float units 8.3 Provide more units

25 Repair cost 7.9 Nominal cost of service

Table 4.16 Customer centric process summary

No

Customer

Requirements

(CRi)

d i

W(C

Ri)

Rat

ing

ABC

Ltd

Rat

ing

Com

petit

or C

1

Wei

ght C

ompe

titor

C1

Rat

ing

Com

petit

or C

2

Wei

ght C

ompe

titor

C2

Targ

et

IR

CR1 Improve service initiation 4 0.093 7 9 0.173 8 0.174 9 1.286

CR2 Improve technical expertise 8 0.186 6 9 0.173 8 0.174 9 1.500

CR3 Prompt vehicle delivery 9 0.209 6 9 0.173 8 0.174 9 1.500

CR4 Minimum time to service 7 0.163 7 8 0.154 7 0.152 8 1.143

CR5 Improve service quality 8 0.186 7 9 0.173 8 0.174 9 1.286

CR6 Nominal cost of service 7 0.163 7 8 0.154 7 0.152 8 1.142

124

4.4.3.3 Company centric process

Suitable Rectification Factors (RF) for the selected customer

requirements are identified using the brainstorming sessions mentioned

above. Two rectification factors for each customer requirement pertaining to

this specific case study problem have been listed below:

RF1 SA work plan

RF2 Pre-service plan

RF3 Technical training schedule

RF4 On-board practice plan

RF5 Work schedule

RF6 Supervision strategy

RF7 Layout/facility planning

RF8 Work time management

RF9 Spare parts management

RF10 Man power adequacy

RF11 Pricing policy

RF12 Cost of materials

The relationship matrix R with weighted RFj is developed as given

in Table 4.17 and the correlation matrix S is prepared as given in Table 4.18.

Table 4.17 Relationship matrix and weighted RFj

RF1 RF2 RF3 RF4 RF5 RF6 RF7 RF8 RF9 RF10 RF11 RF12 W(CRi)

CR1 9 9 1 0 3 3 1 3 -9 1 -1 -1 0.093

CR2 -1 -9 9 9 0 -1 0 3 -9 -9 1 3 0.186

CR3 3 0 -9 -1 9 9 9 9 3 9 0 0 0.209

CR4 0 3 0 3 9 9 9 9 3 9 0 0 0.163

CR5 3 -9 3 3 1 3 1 9 9 9 0 0 0.186

CR6 0 0 3 -9 0 0 -1 3 0 1 9 9 0.163

W(R

F j)

0.06

853

-0.0

7547

0.03

4825

0.03

9006

0.14

2324

0.14

9267

0.12

9297

0.23

6945

0.01

0414

0.13

4523

0.05

8229

0.07

2114

125

Table 4.18 Correlation matrix

RF1 RF2 RF3 RF4 RF5 RF6 RF7 RF8 RF9 RF10 RF11 RF12

RF1 0 4.185 -4.806 -0.627 8.712 10.014 7.038 12.618 1.044 13.176 -1.023 -1.395

RF2 0 -19.251 -18.621 5.238 3.564 3.564 -13.176 -6.066 5.238 -2.511 -5.859

RF3 0 14.22 -16.092 -16.65 -16.767 -5.139 -16.524 -26.291 5.982 9.33

RF4 0 3.078 2.52 4.545 8.163 -9.204 -8.991 -11.529 -8.181

RF5 0 31.527 30.597 32.643 9.207 32.085 -0.279 -0.279

RF6 0 30.969 35.433 14.229 37.107 -0.465 -0.837

RF7 0 31.596 10.881 31.736 -1.56 -0.56

RF8 0 17.577 40.944 4.68 5.769

RF9 0 39.339 -0.837 -4.185

RF10 0 -0.3 -3.648

RF11 0 13.854

RF12 0

A technical competitive assessment of ABC Ltd. and its rival

organisations i.e. C1 and C2 is estimated and is displayed in Figure 4.25.

RF1 RF2 RF3 RF4 RF5 RF6 RF7 RF8 RF9 RF10 RF11 RF12

6 8 7 8 8 8 6 6 7 6 7 7 ABC Ltd

7 9 8 7 7 7 9 7 8 9 7 8 Company C1

8 8 8 8 9 8 8 9 7 7 9 7 Company C2

Figure 4.25 Technical competitive assessment

4.4.3.4 Constructing HOQ

Finally, all the individual matrices are assembled in a structured

manner and the final form, the HOQ, is constructed as shown in Figure 4.26.

126

Figure 4.26 HOQ Matrix

127

4.4.3.5 Selection of customer CTQ

The CRi rankings are undertaken depending upon the IR values

albeit in descending order as shown in Table 4.19 under each category of

performances. The CRi selected as first priority is in Row 3 in Table 4.19 and

concerns the prompt delivery of the vehicle. Examinations of Row 3 have

showed strong cooperative relations with RF5, RF6, RF7, RF8 and RF10 in

Table 4.17. These are the CTQs to be attended to in the subsequent processes

with regard to customer satisfaction. These are listed in Table 4.20 in

descending order of their weights.

Table 4.19 Prioritizing CRis

No Customer Requirement

(CR) di W(CRi)

Rat

ing

ABC

Ltd

Rat

ing

Com

petit

or C

1 R

atin

g

Com

petit

or C

2

Tar

get

IR

Act

ion

Dec

ision

CR1 Improve service initiation 4 0.093 7 9 8 9 1.286 A 4

CR2 Improve technical expertise 8 0.186 6 9 8 9 1.500 A 2

CR3 Prompt vehicle delivery 9 0.209 6 9 8 9 1.500 A 1

CR4 Minimum time to service 7 0.163 7 8 7 8 1.143 B 5

CR5 Improve service quality 8 0.186 7 9 8 9 1.286 A 3

CR6 Nominal cost of service 7 0.163 7 8 7 8 1.143 B 6

128

Table 4.20 Potential customer CTQs

Priority RFj Rectification Factor W(RFj) Target

1 RF8 Work time management 0.23694 To improve

2 RF6 Supervision strategy 0.14927 To improve

3 RF5 Work schedule 0.14232 To improve

4 RF10 Man power adequacy 0.13452 To be enough

5 RF7 Layout/Facility planning 0.12929 To modify

Since all the selected CTQs have a strong cooperative relation

with CR3, it is required that one be selected from among them for

improvement in the next stage. The relative importance of each CTQ on rows

is expressed in the form of a percentage of the grand total in Table 4.21. It is

inferred that the real CTQ for ensuring the vehicle’s delivery as promised is

RF7 - LAYOUT / FACILITY PLANNING since it had the highest percentage

(51.16%) in terms of the relative importance with respect to CR3. The decision

on CTQ selection is also supported in the Correlation Matrix in which, RF7

had a cooperative correlation with the rest of the CTQs. Hence, any further

improvement in RF7 would result in an improvement in RF5, RF6, RF8 and

RF10.

129

Table 4.21 Relative importance matrix

Column Total Grand

total

Row

total 11.2 25.2 16.1 15.4 0.5 68.4 100

RF8 - Work time management 0.2 1 5 0.2 6.4 9.36

RF6 - Supervision strategy 5 0.1 0.2 0.1 5.4 7.90

RF5 - Work schedule 1 10 0.2 0.1 11.3 16.52

RF10 - Man power adequacy 0.2 5 5 0.1 10.3 15.06

RF7 - Layout / Facility

planning 5 10 10 10 35.0 51.16

CR3

Prompt vehicle delivery

RF 8

- W

ork

time

man

agem

ent

RF 6

- Su

perv

isio

n st

rate

gy

RF 5

- W

ork

sche

dule

RF 1

0 - M

an p

ower

ade

quac

y

RF 7

- La

yout

/ Fa

cilit

y pl

anni

ng

Row

tota

l

Perc

enta

ge

4.4.4 Six Sigma – TRIZ Process

4.4.4.1 Six Sigma Define phase

In this phase, a statement of the problem prevalent at present is

developed using the TRIZ Ideal Final Result (IFR) concept. Since IFR insists

on an ideal state of the problem, it makes problem solving akin to moving

from perfection to practice. In this case, the customer complaint regarding the

timing of vehicle delivery is found to have been greatly affected by the factor

“Layout and facilities planning”. Using IFR principles, the problem statement

is developed as:

130

“To plan and develop trouble-free, stakeholder-friendly

environment to assure customer satisfaction”

4.4.4.2 Six Sigma measure phase

In this phase the Sigma quality levels for customer complaints is

calculated. The objective is to compare the performance improvement after

the study with the past performance to justify the improvements, which could

result through this research approach. The total number of customer

complaints regarding dissatisfaction except those arising from vehicle–related

problems is collected from the company database and the Sigma quality level

is computed as follows:

Total number of complaints [C1]: 10648 Number of customers involved in the complaints [P1]: 6120 Average number of complaints per customers [P2]: C1 / P1

= 10648 / 6120 = 1.739 (Complaints)

Total number of customers visited during the year [P3]: 13600 Total number of possible complaints [C2]: P3 * P2

= 13600 * 1.739 = 23662 (Opportunities) Possible number of complaints per million customers [CPMO]: = (10648/ 23662)*1,000,000 = 450 000

Sigma quality level = 0.8406 + SQRT (29.37 – (2.221 *ln(CPMO)))

= 0.8406 + SQRT (29.37 – (2.221*ln(450 000))

= 1.52

131

Degree of customer

dissatisfaction = actual number of complainers out of total customers

= 6120 / 13600

= 0.45

4.4.4.3 Six Sigma analyse phase

This phase has studied the existing facilities and layout

characteristics using the following criteria:

Ease of future expansion – assessing the flexibility of the

system to change

Flow of movement – assessing the flow of manpower and

materials for any complications

Material handling – assessing the efficient movement of

vehicles in-process

Output needs – assessing the conduciveness of service output

Space utilisation – assessing the usage of existing area

Reception and delivery space – assessing the space allotted for

customer handling

Ease of communication – assessing the easiness and

effectiveness of the channel

Impact of employee morale and satisfaction – assessing

service productivity

Promotional values – assessing the degree of customer

attractiveness

Safety – assessing occupational safety

132

The preliminary analysis discussed in this regard with stated criteria

warranted the need to modify the existing reception and delivery space setup.

But it involved additional investment for which, the management is not

interested and hence, requested the technical crew to work-out the

possibilities of modifying the existing setup with zero investment. The TRIZ

function model analysis is employed to replicate the contradictions in layout

modification at zero investment as shown in Figure 4.27.

Figure 4.27 TRIZ Function Modelling

The research study clarified that the technical contradiction as a

useful function (improving customer satisfaction through faster complaint

solving) is in conflict with the harmful function (cost effective investment in

physical arrangement of resources). The specific problem (minimum cost of

service through layout modification but with zero investment in layout

modification) is abstracted as a generic problem through the use of 39

problem parameters of TRIZ. With the help of the Contradiction Matrix, we

identified generic solutions to the generic problems from 40 inventive

HF

UF UF UF

influences

Cau

ses

is required for is required for Modifying the service setup

Provide quick solution to customer

complaints

Improving customer

satisfaction

Investment on layout and facilities

133

principles. By using analogical thought, we could choose a suitable specific

solution for the specific problems of the case study industry.

The technical contradiction algorithm is illustrated in Figure 4.28.

For stated UF (Improving customer satisfaction) and HF (Investment in layout

and facility design), the problem parameters Productivity (39) and Ease of

Repair (34) respectively, are selected as the generic problem parameters from

among TRIZ’s 39 problem parameters. The pre-defined Contradiction Matrix,

by cross referencing Productivity (39) with Ease of Repair (34) yielded the

inventive principles given in Table 4.22.

Figure 4.28 Technical contradiction algorithm

Abstract Thought

Analogical Thought

Problem solved

Contradiction Matrix

Generic problem (39 Problem parameters)

2

Specific problem

1

Generic Solution

(40 inventive principles)

3

Specific Solution

4

Specific Problem Investment required for

modifying the service setup;

But it improves customer

satisfaction

134

Table 4.22 Selection of specific solutions by technical contradiction

algorithm

ABC’s specific

problem

Generic problem

parameter

HF UF

Ease of repair (34)

UF

Improving customer

satisfaction

Productivity (39)

Productivity (39)

Inventive Principles [1] Segmentation [32] Colour change [10] Preliminary

action [25] Self - service HF

Cost of investment in

layout and facilities

Ease of repair (34)

The selected solution principles are;

Principle 1: Segmentation

Divide an object or system into independent parts.

Make an object or system easy to disassemble.

Increase the degree of fragmentation or segmentation.

Principle 32: Changing the colour

Change the colour of an object or system or its external

environment.

Change the transparency of an object or system or its external

environment.

Principle 10: Prior action

Perform, before it is needed, the required change of an object

or system (either fully or partially).

Selected inventive principles

135

Prearrange objects or systems such that they can come into

action from the most convenient place and without losing time

for their delivery.

Principle 25: Self-service

Make an object or system serve itself by performing auxiliary

helpful functions.

Use resources, energy or substances.

Out of these four inventive principles, except for Principle 1 –

Segmentation, all the others have been considered more appropriate in terms

of generating solutions to this conflict situation. This is because, according to

Principle 1, the system has already been segmented into individual divisions

in order to service customers. The possible generic solutions are generated as

follows;

Principle 32: Changing the colour

[Change the transparency of an object or system or its external

environment]

The existing service system looks like an iron ball in which, the

customers are not given any privileges to access any of the service-related

information while the vehicle is under repair. Moreover, customers are not

given a choice in terms of repair time. As per this principle, the existing

system i.e. the service organisation may provide some degree of transparency

in this in-house operation to the customer in following ways:

Provide a tele-helpline to enable customers to book their

service appointments at their convenience.

136

Communicate the amount of work completed in a day on the car by phone or the Short Message Service (SMS).

Provide a complaint-solving process simulation to clarify their doubts.

Provide an easy communication channel for contacting the top management or appellate authority if customers feel the need to do so.

Principle 10: Prior action

[Prearrange objects or systems such that they can come into action

from the most convenient place and without losing time for their

delivery]

The present service system follows a practice of preparing the

repair invoice only after the customer has visited the company for taking

delivery of his vehicle. The process takes up some time and often, customers

have to stand in a queue. Invoice preparation consumes a lot of time as well.

As per this inventive principle, the company may consider the following

proposals:

Preparing day-end invoices – billing of work completed in a

day and summing up at the end of the day in which the work

is completed.

Providing Performa invoice to the customers one day prior to

delivery of vehicles.

Principle 25: Self-service

[Make an object or system serve itself by performing auxiliary

helpful functions]

The following proposals may be considered generic solutions under

this principle:

137

Allowing the customers to prepare the initial job sheet instead

of waiting for the service advisor.

Providing the customers a touch-screen facility to take out a

print of the mini invoice for swifter payments.

Authorising the service advisors to collect payment directly

while delivering the vehicle.

4.4.4.4 Six Sigma Improve phase

In this case analysis, the customer satisfaction is predominantly

influenced by the layout/facility factor like the wait time to get received or the

time span for which they had to wait to take delivery of their vehicles. When

the customer approaches the company for service-related activity, he has to

approach the service advisor and prepare the job sheet first. The existing

system comprises four divisions with individual service advisors. Depending

upon the nature of service desired the customer is served by the respective

division service advisor. If the service advisor is engaged elsewhere, when the

customer visits the premises, then, the customer has to wait for him.

Similarly, after completion of the requisite job, the preparation of repair

invoices is seen to take some time. Till that is done, the customer has to wait.

If many cars are to be delivered that same time, the time taken for invoice

preparation and delivery will be considerably longer. A brainstorming session

with the top management is conducted to discuss the proposals with regard to

minimising the customer’s wait time.

The feasibilities of implementing this proposal are also analysed.

The following specific solutions are then selected from the list of generic

solutions:

Providing a tele-helpline enabling customers to book their

service appointments at their convenience

138

Allowing the customers to prepare the initial job sheet instead

of waiting for the service advisor

Communicating the amount of work completed in a day on the

car by phone or via a Short Message Service (SMS) to the

customer

Preparing day-end invoices – billing of work completed in a

day and summing up at the end of the day when the work is

fully completed

Providing Performa invoice to the customers a day before

delivery

Authorising the service advisors to collect payment directly

while delivering the vehicle

The overall solution generation process has been illustrated in

Figure 4.29.

Figure 4.29 TRIZ solution generation process

Specific Problem

Improving customer satisfaction

with no or minimum investment in

layout / facility design

Generic Problem

Improving service productivity

with available repair facility

Specific Solution

1. Tele helpline 2. Communication facility 3. Pre-service invoice and Performa

invoice 4. Customer authorization 5. Service advisor authorization

Generic Solution

1. Tele helpline 2. Communication facility 3. Complaint process simulation 4. Pre-service invoice and Performa invoice 5. Customer authorization 6. Service advisor authorization 7. Touch screen to self serve

TRIZ Process

139

4.4.4.5 Six Sigma Control phase

The final phase in this study examined the establishment of certain

control measures for the proposals made in the previous stages. To implement

the ongoing measures and actions to sustain the improvement through

monitoring, standardising, documenting and integrating the new amendments

on a daily basis, we preferred the TRIZ innovative tool and the Anticipatory

Failure Detection (AFD) processes instead of the regular Six Sigma tool like

the control chart or even others because the expectations in this phase are not

within the scope of the Six Sigma tools. The next section briefs the

application of the AFD for failure prediction.

4.4.4.6 Anticipatory Failure Detection (AFD) Application of the AFD method helped improve the understanding

of the functions of various system elements like employees, service advisors,

front line staff and managerial personnel and depicted their interactions with

the service process. The research proposal is submitted to the top management

of ABC Ltd. for perusal and the policy making body came up with the

following questions for clarifications:

Could the customer fill the job sheet technically and correctly?

What could be done if a customer cancelled his appointment?

How could a customer be reached over the phone or the SMS

if the customer is out of the contact circle?

How could cash handling at service advisor’s hand be

controlled?

Would these proposals improve the reliability of the service?

To address all these concerns, a graphic model of the useful and the

harmful functions depicted in Figure 4.30 is prepared to continue further

140

research. In addition to the direct analysis of the conditions related to the

proposed solutions, we applied the key steps of the AFD as illustrated in

Figure 4.31.

Figure 4.30 Function modelling for AFD

By

By

Achieving customer satisfaction

Minimizing the off-service waiting time

Useful

function

Aut

horiz

ing

the

cust

omer

to fi

ll jo

b sh

eet

Prov

idin

g cu

stom

ers t

he

tele

-boo

king

faci

litie

s

Des

igni

ng e

ffec

tive

com

mun

icat

ion

chan

nel t

o se

nd in

form

atio

n vi

a ph

one

or S

MS

Prov

idin

g pr

e-se

rvic

e in

voic

e an

d Pe

rform

a in

voic

es

Aut

horiz

ing

the

serv

ice

advi

sors

to c

olle

ct p

aym

ents

But But But But But

Cus

tom

er la

ngua

ge is

diff

eren

t th

an se

rvic

e la

ngua

ge

Cus

tom

er m

ay w

eak

in

tech

nica

l kno

wle

dge

C

usto

mer

is n

ot in

tere

sted

to

fill t

he jo

b sh

eet

Cus

tom

er m

ay u

nedu

cate

d

It re

quire

s exi

sting

syst

em

mod

ifica

tion

It re

quire

s acc

urat

e in

form

atio

n m

anag

emen

t sy

stem

It re

quire

s tru

st w

orth

on

SA’s

pe

rfor

man

ce

Causes

Causes Causes Causes Causes Causes Causes

Harmful function

Misinterpretation of information and resource misusage

Worsening of customer satisfaction

141

Figure 4.31 AFD algorithm

Step 1 – Inverting the issue

Instead of guessing the possible causes of the issue–worsening of

customer satisfaction–it is inverted in a pro-active way as:

“It is necessary to dissatisfy the customer under the conditions

that when he approaches or leaves the service centre”

By rephrasing thus, the attitude of the service centre peoples to the

issue is changed to “How to dissatisfy a customer?”

Step 2 – Find the method(s) of producing the phenomenon

After the problem is inverted, attention is diverted from “things

that can happen” to “things that can be produced”. The next logical step is:

How to produce customer dissatisfaction?

Worsening of customer satisfaction

Method Method Method

Method

Method

Method R

esou

rces

cau

sing

th

e ac

tion

Actions

Things that can be

produced (Phenomenon)

1 Inverting the

issue

Things that can happen (issues)

Poor information

Wealth of Information

2 Hypothese

s

3 R

esou

rce

caus

e

Method

142

“Identify the areas of science, engineering, or even everyday

life, where the same phenomenon of dissatisfying a customer is

intentionally created”.

Having formulated the inverted problem, by utilising the

possibilities of the Internet and the patent library, and interviewed subject

matter expert (SME) in the corresponding technology, we obtained an

exhaustive set of “standard” ways of producing the desired phenomenon.

Step 3 - Utilise Resources

The utilisation of the TRIZ resources in the AFD process is based

on the following postulate:

“For any issue or drawback, all the necessary components of

the issue mechanism must be present within the system or its immediate

environment as available resources”

Therefore, in order to solve the problem of dissatisfying a

customer, it is sufficient to:

1. Invert the initial problem;

2. Identify all standard ways of producing the desired

phenomenon;

3. Verify that all resources required to produce the concern are

present in the corresponding premise.

Upon listing all the available resources in the system, the most like

hypothesis are carried out. In majority, if all resources within the system or its

environment support the considered hypothesis, the problem is solved. The

143

methods cited in Table 4.23 for dissatisfying a customer have been identified

as Hypotheses. We drew the following conclusions after identifying the

components required for each Hypothesis realisation and comparing them

with system resources:

For almost all methods of dissatisfying a customer when he

approaches or leaves the premise, all resources existed within

the system.

The case of dissatisfying a customer is a unique problem.

It means that all available hypothetical mechanisms are likely

to contribute to the issue.

Table 4.23 AFD hypotheses, components and resources

No Hypotheses Components System resources

H1 Lethargic front line staff Reception in-charge people

H2 Providing instability in

customer management system

Procedures information

H3 Unstable service advisors Absenteeism people

H4 Wrong parts inventory system Shortage of vital item parts

H5 Isolated departmental activities Facility design Space

H6 Poor in and out communication Channel or medium function

H7 Repeated complaints Service frequency function

A verbal study of these hypotheses is carried out and hypothesis H1

and H3 are selected since they could make a critical contribution to the issue.

Hypothesis H1 dealing with a lethargic front staff is a great contributing

phenomenon to create the dissatisfaction. Since the customer approached this

144

resource first as he entered the premises, and since the first impression is the

best impression, it is felt that they could influence the levels of satisfaction by

virtue of their behaviour.

The information obtained from the previous step helped verify the

Hypothesis. Similarly, the hypotheses H3 dealing with the service advisor’s

instability during duty it is felt could create in the customer a sense of

frustration and lead to dissatisfaction even if the work quality is excellent.

Hence, these two resources merit close monitoring if the issue – worsening

customer satisfaction is to be counteracted. It may be accomplished through a

series of motivational programmes and structured customer handling training

modules.

4.5 SUMMARY

Two case implementations were undertaken to investigate the

practical implications for the TPE model. In implementation-1, the TPE

model exertion has been carried out through the iterations of the QFD

process. Process parameter optimisation, Die design analysis, Component

design evaluation and Equipment capability analysis are identified as

improvement projects to accomplish the strategy S1 – Minimising the

defective fraction. Subsequently the ISQ was articulated and function

modelling has been performed.

The problem obstructing organisational productivity has then

generated. Productivity measures such as process yield, Sigma quality level

for defect and defect probability are established on previous production data.

The contradiction of process parameter optimisation which needed cost and

time is resolved and the plan for presetting the operating parameters; metal

temperature [Tm], injection pressure [Pi], degassing frequency [fg], IInd phase

145

turns [Ns] and metal mixing ratio [Rm] at the start of actual production has

been developed. Taguchi experimental design was conducted and optimum

levels for process parameters are identified. The outcome of the process has

been predicted at 95% confidence interval.

In case implementation 2, the TPE model has been put into practice

for improving the levels of satisfaction of customers through service activities

in an automotive service centre. The QFD process had scrutinized the factors

causing customer dissatisfaction and specified that the organisational layout

designs are critical. Further investigations through the Six Sigma and TRIZ

processes resulted in proposals to improve the levels of satisfaction. The

management is given suggestions to control the improvement in a sustainable

way by using the AFD method.