CHAPTER 4 CASE IMPLEMENTATIONS -...
Transcript of CHAPTER 4 CASE IMPLEMENTATIONS -...
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.