Pneumatic Injection Mould Machine Capability and Techno-Economic Study Matthew Paul Keyser Department of Industrial Engineering University of Stellenbosch Study Leader: Theuns Dirkse van Schalkwyk Final year project presented in partial fulfilment of the requirements for the degree of Industrial Engineering at Stellenbosch University B.Eng Industrial
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Pneumatic Injection Mould Machine Capability and
Techno-Economic Study
Matthew Paul Keyser
Department of Industrial Engineering
University of Stellenbosch
Study Leader: Theuns Dirkse van Schalkwyk
Final year project presented in partial fulfilment of the requirements for the degree of
Industrial Engineering at Stellenbosch University
B.Eng Industrial
i
Declaration
I, the undersigned, hereby declare that the work contained in this final year project is my own original
work and that I have not previously in its entirety or in part submitted it at any university for a degree.
The injection moulding process constitutes a vital part of the manufacturing sector and plastic
injection moulding (PIM) is one aspect of this process. The following report discusses the
characteristics and capabilities of a custom made injection moulding machine (CIMM) powered by
pneumatics, with the purpose of moulding plastic parts. These moulded parts will be measured and
analysed statistically in order to determine the optimal operational settings for the CIMM.
For the optimal settings to be determined, a set of experiments needs to be executed. Various
literature was studied to ensure an appropriate project methodology would be implemented to
successfully carry out the experimentation. Consequently, it was determined that a full factorial
design of experiments (DoE) would be executed with the assistance of Taguchi’s Optimisation
Method and various statistical analysis. Twenty-seven experiments were executed with each
experiment consisting of a unique combination of three factors, at three different levels. The three
factors are as follows:
1. The temperature at which the plastic is moulded (MT), in degrees Celsius (oC).
2. The length of time given for the plastic to fill the mould (FT), in seconds (sec).
3. The length of time kept in place before ejecting part (PT), in seconds (sec).
Once all twenty-seven experiments had been completed, each moulded part was inspected and
measured for data collection according to the following four criteria:
1. Visual inspection of part conformance
2. Rework time, measured in seconds (sec)
Rework time for runner
Rework time for finishing
3. Part weight, measured in grams (g)
4. Part thickness, measured in millimetres (mm)
iv
The above analysis provided the data that would be used to determine the optimal operational setting
for the CIMM by conducting the following forms of analysis:
Initial observation analysis
Techno-economic analysis
Statistical analysis
Descriptive statistical analysis
ANOVA analysis
Process control analysis
Waste Analysis
The statistical analysis was executed with the assistance of RSudio which is the interface for the open
source statistical software package R. All forms of analysis shared the same two objectives; total cost
per part (𝑇𝐶𝑝𝑎𝑟𝑡) must be minimised and part thickness must be maximised whilst minimising the
variation of part thickness in moulding process. The brief summary of the three possible operational
settings, based on their performance in the analysis process, are given in the table below.
A break-even analysis was then conducted to determine the single most optimal solution for the
CIMM so that a place for the machine in the business sector can be motivated. This resulted in
Experiment 14 being suggested as the most optimal solution due to its superior monthly profit which
will be the most beneficial factor moving forward, from a business perspective.
The findings of this project recommend that the CIMM would be best suited for small scale
production. This finding can lend itself to small business owners or enthusiasts as it is a compact and
mobile machine that can be operated and stored in a garage.
MT FT PT
14 190 10 5 1,26
23 200 10 5 1,26
26 200 15 5 1,67
Experiment TCpart (R )Operational settings
Table 1: Three possible optimal operational settings for the CIMM.
v
Opsomming
Die spuitgietings-proses speel 'n belangrike rol in die vervaardigingsektor en plastiese spuitvorms
(‘PIM’) is een aspek van hierdie proses. Hierdie dokument bespreek die kenmerke en vermoëns van 'n
doelvervaardigde spuitvormmasjien (‘CIMM’) wat pneumaties aangedryf word deur met die doel om
plastiese onderdele te giet. Hierdie gevormde onderdele sal gemeet en statisties geanaliseer word om
die optimale operasionele verstellings vir die CIMM te bepaal.
Om die optimale verstellings te bepaal, moet 'n stel eksperimente uitgevoer word. Verskeie bronne
uit die literatuur is bestudeer om ‘n toepaslike projek metodologie te identifiseer en te implimenteer
sodat die eksperiment suksesvol uitgevoer kan word. Gevolglik is vasgestel dat 'n volle faktoriale
ontwerp van eksperimente (DoE) uitgevoer sal word met behulp van Taguchi se Optimalisering-
metode asook verskeie statistiese analises. Sewe en twintig eksperimente is uitgevoer waar elke
eksperiment bestaan uit 'n unieke kombinasie van drie faktore, op drie verskillende vlakke. Die drie
faktore is soos volg:
1. Die temperatuur waarteen die plastiek gegiet is (‘MT’), in grade Celsius (oC).
2. Die tydsduur om die gietvorm met plastiek te vul (‘FT’), in sekondes (sek).
3. Die tydsduur wat die onderdeel in die gietvorm in plek gehou word voordat die onderdeel
uitgeset word (‘PT’), in sekondes (sek).
Nadat al sewe en twintig eksperimente afgehandel is, is elke gevormde deel geïnspekteer en gemeet
vir die versameling van data volgens die volgende vier kriteria:
1. Visuele inspeksie om te bepaal of die onderdeel aan die vooraf bepaalde part vereistes
voldoen
2. Herwerk tyd gemeet in sekondes (sek)
Herwerk-tyd vir die ‘loper’
Herwerk-tyd vir die afwerking
3. Die onderdeel se gewig, gemeet in gram (g)
4. Die onderdeel se dikte, gemeet in millimeters (mm)
vi
Die bogenoemde ontleding het die nodige data verskaf om die optimale operassionele verstellings vir
die CIMM vas te stel deur die volgende vorms van analise te onderneem:
Aanvanklike waarnemings analise
Tegno-ekonomiese analise
Statistiese analise
Beskrywende statistiese analise
ANOVA analise
Proses beheer ontleding
Verkwisting (afval) analise
Die statistiese analise is uitgevoer met die hulp van RSudio wat die koppelvlak is vir die “open
source” statistiese sagteware pakket R. Alle vorms van analise het dieselfde twee doelwitte gedeel;
totale koste per onderdeel (‘𝑇𝐶𝑝𝑎𝑟𝑡’) moet tot die minimum beperk word en die onderdeel dikte moet
gemaksimeer word terwyl die variasie van die onderdeel dikte in die gietproses beperk moet word.
Die kort opsomming van die drie moontlike operasionele verstellings, soos gebasseer op hul prestasie
in die analise proses, word in die onderstaande tabel aangebring.
A gelykbreek analise gevolglik uitgevoer om die optimale oplossing vir die CIMM te bepaal sodat 'n
plek vir die masjien in die sakesektor gemotiveer kan word. Daarvolgens word Eksperiment 14
aanbeveel as die optimale oplossing weens sy uitstekende maandelikse wins. Vanuit ‘n
besigheidsperspektief is hierdie wins die mees voordelige faktor om vooruit te gaan.
Die bevindinge van hierdie projek beveel aan dat die CIMM mees geskik sal wees vir kleinskaalse
produksie. Hierdie bevinding leen homself tot kleinsake eienaars of entoesiaste, aangesien dit 'n
kompakte en mobiele masjien is wat in ‘n motorhuis gebruik en gestoor kan word.
Table 2: Drie moontlike optimale operasionele vertsellings vir die CIMM.
MT FT PT
14 190 10 5 1,26
23 200 10 5 1,26
26 200 15 5 1,67
EksperimentOperasionele verstellings
TCpart (R )
vii
Acknowledgements
I would firstly like to thank my family for the love and support they have shown me over my years at
university. Special mention has to go to my parents Min and Randolf Keyser. If it were not for all the
sacrifices they have made for me, I would not have had the privilege of studying Industrial
Engineering at an amazing university.
I would like to thank my good friend Soren Bruce for his time and effort in helping me with the data
collection which was a time consuming and tedious process due to the nature of this project.
I would like to thank my study leader, Theuns Dirkse van Schalkwyk, for his for his guidance and
assistance in the execution of this project.
Lastly, I would like to acknowledge my friends, especially my classmates, who have made the
engineering course just a bit more fun and enjoyable.
Jeremiah 29:11
“I know the plans I have for you,” declares the Lord, “plans to prosper you and not to harm you,
plans to give you hope and a future.”
viii
Contents
Declaration ............................................................................................................................................... i
ECSA Exit Level Outcomes Reference .................................................................................................. ii
Abstract .................................................................................................................................................. iii
Opsomming ............................................................................................................................................. v
Acknowledgements ............................................................................................................................... vii
Contents ............................................................................................................................................... viii
List of Figures ........................................................................................................................................ xi
List of Tables ........................................................................................................................................ xii
Glossary ............................................................................................................................................... xiii
2. Literature Review ................................................................................................................................ 5
3.2.2 Quantitative Analysis of Part Quality ...................................................................................... 25
4. Analysis of Data ................................................................................................................................ 31
Appendix A ........................................................................................................................................... 58
Appendix B ........................................................................................................................................... 59
Classification of Design of Experiments .......................................................................................... 59
Appendix C ........................................................................................................................................... 62
Factors for Control Charts ................................................................................................................ 62
x
Appendix D ........................................................................................................................................... 63
Coding used in RStudio .................................................................................................................... 63
Appendix E ........................................................................................................................................... 65
Digital Watt Meter ............................................................................................................................ 65
Figure 38: Digital watt meter that was used to measure energy consumption (kWh). ....................... 65
xii
List of Tables
Table 1: Three possible optimal operational settings for the CIMM. .................................................... iv
Table 2: Drie moontlike optimale operasionele vertsellings vir die CIMM. .......................................... vi
Table 3: Advantages and disadvantages of PIM (Rosato, 2000). ............................................................ 7
Table 4: Requisites and Tools for Sound Experimentation (Juran & Godfrey, 1998). .......................... 10
Table 5: Advantages and disadvantages of R (Williams, 2012). ........................................................... 15
Table 6: Control factors and their levels for experimentation. ............................................................ 20
Table 7: The L27 Orthogonal Array (OA) ............................................................................................... 21
Table 8: OA with Control Factors and their different levels for Experimentation. ............................... 22
Table 9: Table showing how the standard rework times were determined. ........................................ 27
Table 10: The data collected for experiment 8 is shown here as an example of the data collected for
each experiment. .................................................................................................................................. 30
Table 11: Unit costs for the measured cost parameters. ..................................................................... 33
Table 12: Summary of techno-economic analysis. ............................................................................... 33
Table 13: Optimal based on the variation of total cost per part. ......................................................... 37
Table 14: Optimal settings based on the variation of part thickness. .................................................. 39
Table 15: ANOVA results for total cost per part. .................................................................................. 41
Table 16: ANOVA results for part thickness. ......................................................................................... 43
Table 17: Optimal settings for total cost per part. ............................................................................... 45
Table 18: Optimal settings for part thickness. ...................................................................................... 45
Table 19: Summary of values used for the X-bar chart. ....................................................................... 46
Table 20: Summary of results for process control analysis. ................................................................. 48
Table 21: Experiments that are statistically in control. ........................................................................ 48
Table 22: Table showing wastage as a percentage of the material used for the completed part (i.e.
after rework). ........................................................................................................................................ 50
Table 23: Summary of possible operational settings based on the conducted analyses. .................... 51
Table 24: Remaining possible optimal operational settings. ................................................................ 51
Table 25: Summary of break-even analysis. ......................................................................................... 52
Table 26: Optimal operational settings for the CIMM. ......................................................................... 54
7. For the final step, an experimental confirmation is run using the predicted optimum levels for
the control parameters being studied.
The Taguchi method may not necessarily provide the optimal solution as the experiment does not
contain all the possible combinations of parameters. However, it will provide a clear indication of
which parameters have the greatest effect on quality and cost. In this project, the Taguchi Method will
be implemented in conjunction with a full factorial experimental design to determine the optimal
operating regime of the CIMM.
2.4 RStudio
The software environment R is widely used for statistical computing and constructing graphics. It is
an easy to adopt coding language that allows for a user-created interface designed around a specific
(set of) problem(s) (Le Roux & Lubbe, 2013). RSudio provides the interface for the open source
statistical software package R.
2.4.1 R as a Software Development Platform
R is an open-source software system that is supported by a group of volunteers from many countries
with the central control being in the hands of a group called ‘R-Core’. Its base system provides a
general computer language for performing tasks like organising data, statistical analysis, model-
fitting, data visualisation, building of complex graphs etc. (Chambers, 2008). The R package hosts a
powerful and flexible set of statistical tools which are customisable on the platform to best suit the
required needs of the user. The platform itself facilitates both the handling and storage of data by
making use of a coherent collection of intermediate tools to analyse the data with (Chambers, 2008).
Williams (2012) has analysed R and compiled a list of advantages and disadvantages for the statistical
software system and some of those points are listed in the table on the following page.
15
Table 5: Advantages and disadvantages of R (Williams, 2012).
Advantages Disadvantages
R is the most comprehensive statistical analysis
package available. It incorporates all of the
standard statistical tests, models, and analyses, as
well as providing a comprehensive language for
managing and manipulating data. New technology
and ideas often appear first in R.
R has a steep learning curve (it does take a while
to get used to the power of R) but no steeper than
for other statistical languages. R is not so easy to
use for the novice. There are several simple-to
use graphical user interfaces (GUIs) for R that
encompass point and-click interactions, but they
generally do not have the polish of the commercial
offerings.
R is a programming language and environment
developed for statistical analysis by practising
statisticians and researchers. It reflects well on a
very competent community of computational
statisticians. R is now maintained by a core team
of some 19 developers, including some very senior
statisticians.
There is, in general, no one to complain to if
something doesn’t work. R is a software
application that many people freely devote their
own time to developing. Problems are usually
dealt with quickly on the open mailing lists, and
bugs disappear with lightning speed. Users who do
require it can purchase support from a number of
vendors internationally.
The graphical capabilities of R are outstanding,
providing a fully programmable graphics language
that surpasses most other statistical and graphical
packages. Because R is open source, unlike closed
source software, it has been reviewed by many
internationally renowned statisticians and
computational scientists.
Many R commands give little thought to memory
management, and so R can very quickly consume
all available memory. This can be a restriction
when doing data mining. There are various
solutions, including using 64 bit operating systems
that can access much more memory than 32 bit
ones.
R has over 4800 packages available from multiple
repositories specializing in topics like
econometrics, data mining, spatial analysis, and
bio-informatics.
Documentation is sometimes patchy and terse, and
impenetrable to the non-statistician. However,
some very high-standard books are increasingly
plugging the documentation gaps.
2.4.2 Design of Experiments in R
As a result of being an open-source system, R is exposed to continual scrutiny by the users. This
includes some algorithms for numerical computations and simulation that likewise reflect modern,
open-source computational standards in these fields (Chambers, 2008). This means that users not only
update current algorithms that solve long standing problems, but they also develop algorithm
packages that solve the problems of today. Essentially users can create packages to solve almost any
statistical problem that they can come up with. Looking at the problem related to this project, a
‘Design of Experiments’ is one such package that exists (Cano et al, 2012).
R is the software package that will be used for the statistical analysis of this project due to its
versatility and customisability.
16
2.5 Techno-economic Study
The assessment of the CIMM for its techno-economic feasibility is of utmost importance for the
motivation of the machine to be implemented into the business sector. This section will discuss the
techno-economic factors of the CIMM which consists of two stages; firstly the cost per part followed
by the break-even analysis.
2.5.1 Cost per part
One of the main objectives to this project is to successfully obtain the optimal operating settings for
the CIMM. This can be more accurately achieved by determining the cost per part, at each respective
operational setting combination, as this adds an extra dimension to finding an optimal solution. There
is no existing literature form this section of the report as the cost per part is unique to this project.
Each unique setting combination will be accounted for by producing a number of parts, for that
combination, in an experimental run. Factors that influence the cost per part are:
Energy consumption
Maintenance
Raw material
Compressed air
Labour
The CIMM’s compressed air usage and maintenance costs are not significantly affected by varying
the operational settings, so the incurred cost can be ignored. Therefore, for the purpose of this project,
only the energy, material and labour cost will be considered in determining the cost per part for the
CIMM.
The total labour cost (𝐶𝐿,𝑇𝑜𝑡) will be calculated by taking into account both the cycle and rework time
for each part, in an experimental run, and multiplying it by the labour cost per hour. The cycle time
per part will be obtained by timing the entire run from when the first part starts to mould until the
twenty-fifth part has been moulded. For each experimental run this cost can be represented in the
following equation:
𝐶𝐿,𝑇𝑜𝑡 = (𝑇𝐶 + 𝑇𝑅) × 𝐶𝐿 (2.2)
where 𝑇𝐶 is the cycle time per part, 𝑇𝑅 is the rework time per part and 𝐶𝐿 is the unit labour cost
measured as R/hr. From this result the labour cost per part (𝐶𝐿,𝑝𝑎𝑟𝑡) can be calculated by dividing the
labour cost by the number of conforming parts produced in that run (𝑛):
𝐶𝐿,𝑝𝑎𝑟𝑡 =𝐶𝐿,𝑇𝑜𝑡
𝑛 (2.3)
17
The manner in which the rework time per part shall be obtained is explained in Section 3.2.2.
Material costs forms the second aspect in the cost per part analysis. The total material cost (𝐶𝑀,𝑇𝑜𝑡)
will be calculated by making use of the total part weight (𝑃𝑊) and the raw material cost (𝐶𝑅𝑀) per
kilogram:
𝐶𝑀,𝑇𝑜𝑡 = 𝐶𝑅𝑀 × 𝑃𝑊 (2.4)
The material cost per part (𝐶𝑀,𝑝𝑎𝑟𝑡) can then be calculated by dividing 𝐶𝑀,𝑇𝑜𝑡 by the number of
conforming parts produced in that run (𝑛) as shown in the equation:
𝐶𝑀,𝑝𝑎𝑟𝑡 = 𝐶𝑀,𝑇𝑜𝑡
𝑛 (2.5)
The final cost that will be taken into consideration, for the techno-economic assessment, will be the
energy consumption cost. This consumption will be obtained by using an adaptor device (Appendix
E) which connects to the mains of the machine. This device will read and measure the total energy
consumption (𝐸𝑇𝑜𝑡 ) of the CIMM. The total energy usage cost (𝐶𝐸,𝑇𝑜𝑡) will then be calculated by
taking the energy consumption measurement for the run and multiplying it with the cost per kWh
(𝐶𝑘𝑊ℎ ) given by the local municipality. This can be depicted with the equation:
𝐶𝐸,𝑇𝑜𝑡 = 𝐶𝑘𝑊ℎ × 𝐸𝑇𝑜𝑡 (2.6)
The energy cost per part (𝐶𝐸,𝑝𝑎𝑟𝑡) can then be determined by dividing (𝐶𝐸,𝑇𝑜𝑡) by the number of
conforming parts produced in that run (𝑛):
𝐶𝐸,𝑝𝑎𝑟𝑡 = 𝐶𝐸,𝑇𝑜𝑡
𝑛 (2.7)
The above costs can then be combined to determine the total cost per part (𝑇𝐶𝑝𝑎𝑟𝑡) for each
experimental run with the equation:
𝑇𝐶𝑝𝑎𝑟𝑡 = 𝐶𝐿,𝑝𝑎𝑟𝑡 + 𝐶𝑀,𝑝𝑎𝑟𝑡 + 𝐶𝐸,𝑝𝑎𝑟𝑡 (2.8)
Substituting Equations 2.2 – 2.7 into Equation 2.8 will result in the finalised 𝑇𝐶𝑝𝑎𝑟𝑡 given below in
Equation 2.9:
𝑇𝐶𝑝𝑎𝑟𝑡 = (𝐶𝐿)(𝑇𝐶+ 𝑇𝑅)+(𝐶𝑘𝑊ℎ)(𝐸𝑇𝑜𝑡)+(𝐶𝑅𝑀)(𝑇𝑜𝑡𝑃𝑊)
𝑛 (2.9)
18
2.5.2 Break-even Analysis
Gutierrez and Dalsted (2012) provide a sufficient definition for break-even analysis: “Break- even
analysis is a useful tool to study the relationship between fixed costs, variable costs and returns. A
break-even point (BEP) defines when an investment will generate a positive return”. As its name
implies, this approach determines the sales needed to break even.
From a calculation perspective, the break-even point computes the volume of production at a given
price necessary to cover all costs (Gutierrez and Dalsted, 2012). The formula for this calculation is
given as:
𝐵𝐸𝑃 = 𝐹𝑖𝑥𝑒𝑑 𝐶𝑜𝑠𝑡𝑠
𝑆𝑃−𝑉𝐶 (2.8)
where SP is the selling price per part and VC is the variable cost per part.
In the case of this project, the fixed costs will only involve the cost of the CIMM and the VC will
include the three costs (𝑇𝐶𝑝𝑎𝑟𝑡) mentioned above in Section 2.6.1. From the result obtained in the
break-even analysis, a feasibility case can be provided for purchasing the CIMM.
19
3. Project Methodology
The project methodology will comprise of two sections; the experimental procedure followed by the
data collection and analysis methodology. These two sections will be outlined and discussed in this
chapter.
3.1 Experimental Procedure
In order to determine an optimal operating setting for the CIMM, experiments need to be executed.
These experiments will be performed by using the Taguchi Method in conjunction will a full factorial
experimental design.
The steps of the Taguchi method will now be implemented.
3.1.1 Quality Aspect to be Optimised The quality aspect that is to be optimised is the moulding finish of plastically moulded parts using the
CIMM. The side effects of this optimising process will include moulded parts with varying levels of
quality to the finished part. Each of these parts will be classified as either a conforming or non-
forming part.
3.1.2 Identify the Noise Factors and Test Conditions The experiments will be conducted in the Senrob Lab of the Mechanical and Industrial Engineering
Building. As explained in the Taguchi Optimization Method (Section 2.3), it is important to identify
the noise factors in this experiment as they can have a negative impact on the quality of the moulded
parts. The noise factors that could affect the mould operation on the CIMM are:
Variation in the raw material (plastic granules)
Machine condition
Ambient temperature of the Senrob Lab
Operator skill
3.1.3 Identify the Control Factors and their Alternative Levels As mentioned in step 3 of the Taguchi Method, the control factors (test parameters) are those that can
be set and maintained. Recapitulating from Section 1.2, the control factors are as follows:
The temperature of the molten plastic, or more simply the mould temperature (MT).
The time allocated for the molten plastic to flow into the mould, or more simply the filling
time (FT).
The time allocated for the mould to be kept in place before the part is ejected, or more simply
the packing time (PT).
20
The factors and their levels, for conducting the experiment, were decided upon by moulding a few
parts at random levels until several conforming parts were successfully produced. The control
parameters settings were noted and the levels for the experiment were consequently chosen based on
moderate variations on the noted parameter settings. The control factors and their respective levels for
experimentation are shown in Table 6.
3.1.4 Design Experimental Matrix An appropriate sized orthogonal array (OA) has to be used for conducting the experiments. Given that
a full factorial experimental design is to be executed, Equation 2.1 will be used to determine the size
of the (OA). This equation yields:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑅𝑢𝑛𝑠 = 33 = 27
Therefore the most suitable orthogonal array for experimentation is an L27 array as shown in Table 7
on the next page. This means that a total of twenty seven experiments need to be carried out.
1 2 3
MT 180 190 200
FT 5 10 15
PT 3 5 7
FactorsLevels
Table 6: Control factors and their levels for experimentation.
21
Table 7: The L27 Orthogonal Array (OA)
3.1.5 Conduct Matrix Experiment In accordance with the above OA, experiments were conducted with the factors and their levels as
mentioned in Table 7. The experimental layout with the selected values of the factors is shown on the
following page in Table 8.
Experimental
No.
Control Factors
1 2 3
1 1 1 1
2 1 1 2
3 1 1 3
4 1 2 1
5 1 2 2
6 1 2 3
7 1 3 1
8 1 3 2
9 1 3 3
10 2 1 1
11 2 1 2
12 2 1 3
13 2 2 1
14 2 2 2
15 2 2 3
16 2 3 1
17 2 3 2
18 2 3 3
19 3 1 1
20 3 1 2
21 3 1 3
22 3 2 1
23 3 2 2
24 3 2 3
25 3 3 1
26 3 3 2
27 3 3 3
22
Table 8: OA with Control Factors and their different levels for Experimentation.
Experimental No.
Control Factors
MT FT PT
1 180 5 3
2 180 5 5
3 180 5 7
4 180 10 3
5 180 10 5
6 180 10 7
7 180 15 3
8 180 15 5
9 180 15 7
10 190 5 3
11 190 5 5
12 190 5 7
13 190 10 3
14 190 10 5
15 190 10 7
16 190 15 3
17 190 15 5
18 190 15 7
19 200 5 3
20 200 5 5
21 200 5 7
22 200 10 3
23 200 10 5
24 200 10 7
25 200 15 3
26 200 15 5
27 200 15 7
Each of the above 27 experiments will involve manufacturing 25 parts as to account for the variations
that may occur due to the noise factors. This means that a total of 675 moulds will be manufactured
by the CIMM during the experimentation process. The experiments shall be executed in the order they
are represented in Table 8. One of these experiments will represent the optimal operational settings
for the CIMM to produce the given part.
23
The manner in which each of these 27 experiments will be executed is visually outlined below in
Figure 7.
Figure 7: Flow chart of the experimental procedure.
24
As mentioned in 7, each moulded part will be placed on a sequential numbering chart as it is removed
from the CIMM as they are too hot mark with a permanent marker immediately after their removal.
The reason for numbering system is to track the quality of parts as the experiment is conducted to see
if any trends are identified. An image of the sequential chart is shown in Figure 8 below.
Once an experimental run has been completed and the total cycle time and power usage for that run
have been recorded, each part will be numbered with permanent marker. This number must correlate
with position on the chart.
Following the numbering procedure, the parts for each experimental run will then be placed into a bag
labelled with that experiments number. The data that has to be collected from these experiments and
the manner in which it is to be obtained will be discussed in the succeeding section.
Figure 8: The sequential numbering chart.
25
3.2 Data Collection and Analysis Methodology
This section will look at how the data was obtained from the moulded parts that were produced during
the experiments and the quality inspection procedure that was followed.
3.2.1 Quality Inspection Criteria The operational settings can only be considered optimal if conforming parts are produced and this can
be determined through a quality inspection. The quality inspection of each part was divided into four
inspection criteria:
1. Visual inspection of part conformance
2. Rework time, measured in seconds (sec)
Rework time for runner (𝑇𝑅,𝑟𝑢𝑛𝑛𝑒𝑟)
Rework time for finishing (𝑇𝑅,𝑓𝑖𝑛𝑖𝑠ℎ)
3. Part weight, measured in grams (g)
4. Part thickness, measured in millimetres (mm)
Each of the above criteria will now be discussed and insight will be given as to how each was
measured.
3.2.2 Quantitative Analysis of Part Quality As previously mentioned, visual inspection of part conformance formed the first criteria of the quality
inspection process. This aspect of the inspection process is important as it determines whether a part
was successfully been moulded or not. The result will have significant implications to the cost per part
value for that experimental run as shown in the equations from Section 2.5.1. An example of a
conforming part prior to rework, and non-conforming part can be seen in Figure 9 and Figure 10
respectively on the following page.
26
It is important to note that conformance is based on the circular section of the moulded part as it is the
circular section which constitutes the final part. The longitudinal section is the runner of the moulded
part which will be cut off during the rework cycle.
The second criteria, rework time, was measured for the purpose of determining the labour costs for
the production of the part. For the purposes of this project the rework time will involve two
dimensions; rework time for the runner (𝑇𝑅,𝑟𝑢𝑛𝑛𝑒𝑟), and rework time for the finishing (𝑇𝑅,𝑓𝑖𝑛𝑖𝑠ℎ) of the
part. Rework is a necessary process as it ensures the final part meets the design specifications. This is
a standard time and was determined by measuring the time to rework twenty-five parts which was
repeated three time to get an average. An average rework time per part was obtained from dividing the
average by twenty-five, as there were twenty-five parts produced per experiment.
Figure 9: A visually conforming part, prior to rework.
Figure 10: A visually non-conforming part.
27
The standard rework time of how long it will take to rework a single part was based on the average
rework time per part. The results of this are shown in Table 9 below.
Only conforming parts, as shown in Figure 9, will undergo rework as there is no point wasting time
on reworking non-conforming points as it will not add any value to the part. Figure 11 and 12 below
show a conforming part following rework for the runner and finishing respectively.
The third criteria is the part weight (PW) and it was measured using a very accurate scale which
measures to one-thousandth of a gram (10-3 g) or three decimal places. It is important to mention that
conforming parts were weighed before (PW) and after both rework procedures (PWA) had been
executed in order to get a measurement for the wasted material per part. This is illustrated in Figure
13 and 14 on the next page.
Trial 1 Trial 2 Trial 3
Runner 137 128 134 133 5.32 6
Finishing 368 356 377 367 14.68 15
Rework DimensionTimes (sec) Average Time
per Trial
Average Rework
Time per Part
Standard Rework
Time per Part
Table 9: Table showing how the standard rework times were determined.
Figure 12: Conforming part following rework for finishing.
Figure 11: Conforming part following rework for runner.
28
The fourth and final criteria is the part thickness which was measured using a very accurate digital
vernier calliper which can also measure to one-thousandth of a millimetre (10-3 mm) or three decimal
places. The thickness measurement was chosen at an arbitrary point as the part had no existing
dimensional tolerances.
Non-conforming parts were not given a thickness measurement as they have already been declared as
not being useful. Figure 15 on the following page shows where and how the part thickness was
measured.
Figure 13: Weighing part prior to rework (PW). Figure 14: Weighing part after rework (PWA).
29
All the data that was measured from implementing the four inspection criteria was captured and
entered into excel sheets for each experiment. An example of the data that was captured can be seen
below in Table 10 which displays an excel sheet with the data that was captured for Experiment 8.
This table is displayed on the page that follows.
Figure 15: Illustrating where and how the part thickness was measured using the digital vernier calliper.
30
Table 10: The data collected for experiment 8 is shown here as an example of the data collected for each experiment.
Visually conforming and non-conforming parts are represented with a ‘1’ and ‘0’ respectively in the
‘Conformance’ column. The ‘Waste Material’ column values were achieved by subtracting the PWA
from the PW.
Once the data for all the experiments have been captured into excel sheets, the statistical analysis of
this data can commence. Chapter 4 outlines the steps followed and results obtained from the techno-
economic analysis and statistical analysis.
Experiment No. 8
Temp (oC) 180 Cycle Time (min) 13:54
FT (sec) 15 Power Usage (kWh) 0.0656
PT (sec) 5
Time Started 13:09
1 0 0.607 0.607
2 0 1.589 1.589
3 0 1.751 1.751
4 1 6 15 1.810 0.873 0.937 2.254
5 1 6 15 1.900 0.896 1.004 2.287
6 1 6 15 1.876 0.907 0.969 2.305
7 1 6 15 1.917 0.922 0.995 2.332
8 1 6 15 1.880 0.914 0.966 2.337
9 1 6 15 1.914 0.915 0.999 2.313
10 1 6 15 1.817 0.914 0.903 2.354
11 1 6 15 1.854 0.912 0.942 2.355
12 1 6 15 1.852 0.914 0.938 2.324
13 1 6 15 1.950 0.933 1.017 2.379
14 1 6 15 1.861 0.920 0.941 2.348
15 1 6 15 1.877 0.917 0.960 2.322
16 1 6 15 1.906 0.906 1.000 2.298
17 1 6 15 1.879 0.912 0.967 2.319
18 1 6 15 1.852 0.916 0.936 2.309
19 1 6 15 1.861 0.927 0.934 2.325
20 1 6 15 1.873 0.928 0.945 2.328
21 1 6 15 1.899 0.918 0.981 2.321
22 1 6 15 1.842 0.922 0.920 2.317
23 1 6 15 1.855 0.920 0.935 2.350
24 1 6 15 1.866 0.915 0.951 2.320
25 1 6 15 1.930 0.918 1.012 2.326
Thicknes
s (mm)Part No.
Conformance
(0/1)
Rework Time
(Runner)
Rework Time
(Finish)PW (g) PWA (g)
Waste
Material (g)
31
4. Analysis of Data
To successfully determine the optimal operational settings for the CIMM the data collected in Section
3.2 has to statistically analysed. The following section will review the procedures that were used to
analyse the data that was captured into Excel. Graphs that could not be created in RStudio will be
created in Excel. A copy of the R coding that was used in this project can be found in Appendix D.
Three different analyses will be used:
Initial observation analysis
Techno-economic analysis
Statistical analysis
Descriptive statistical analysis
ANOVA analysis
Process control analysis
Waste Analysis
The two main objectives of the analysis in this section are:
1. Total cost per part (𝑇𝐶𝑝𝑎𝑟𝑡) must be minimised
2. Part thickness must be maximised
These two objectives have to be achieved whilst minimising the variation of part thickness in
moulding process.
4.1 Initial Observation Analysis
Once all the data had been captured into the excel sheets for all the conducted experiments, two initial
observations were noted before any statistical analysis needed to be performed. The first observation
pertained to the number of conforming parts that were produced for each of the conducted
experiments. A summary of these results can be seen in Figure 16 on the following page.
The second observation was that the moulded parts did not automatically eject as expected. This not
only meant that the experimental runs acquired a greater cycle time (𝑇𝐶), it also posed a safety hazard
as the parts now had to be removed by hand as shown in Figure 17 on the following page.
32
From Figure 16 it is evident that only “Experiments; 7, 8, 9, 13, 14, 15, 16, 17, 18, 22, 23, 24, 25, 26,
27” produced conforming parts. Therefore, only these experiments will be considered for the techno-
economic and statistical analyses that follow.
Figure 16: A bar graph showing the number of conforming parts produced per experiment.
Figure 17: Removing moulded parts by hand.
33
4.2 Techno-economic Analysis
The first part of the techno-economic analysis involves determining the cost per part for each
experiment as outlined in Section 2.5. These values are needed for the second stage of the techno-
economic analysis, namely the break-even analysis. Section 2.5.1 stated that only the energy
consumption, labour and raw material costs would be considered for the techno-economic analysis
and their unit costs are summarised in the following table:
Substituting the above unit costs along with the captured data into Equations 2.2, 2.4 and 2.6 will
result in the total costs for labour (𝐶𝐿,𝑇𝑜𝑡), material (𝐶𝑀,𝑇𝑜𝑡) and energy (𝐶𝐸,𝑇𝑜𝑡) respectively. These
values can been seen in Table 12 below:
Cost Parameters Unit Source
Energy (CkWh) R0.88/kWh Stellenbosch Municipality, 2015
Labour (CL) R85/hr Trading Economics Website, 2015
Material (CRM) R20.75/Kg Plastomark PTY LTD Quote, 2015
Table 11: Unit costs for the measured cost parameters.
barplot(NumOfConfParts, names.arg = c("1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27"),xlab = "Experiment",ylab = "Frequency",main ="Number Of Conforming Parts per Experiment" )
#Histogram for Summary Of Total Cost (TC) per Part