Parameter Optimization of Extrusion Machine Producing UPVC Pipes using Taguchi Method: A Case of Amhara Pipe Factory

Parameter Optimization of Extrusion Machine Producing UPVC Pipes using Taguchi Method: A Case of Amhara Pipe Factory

1 Solomon Kerealme, 2 N. Srirangarajalu (Asst. Prof.), 3 Assefa Asmare (PhD)

Faculty of Mechanical and Industrial Engineering, Bahir Dar University,

Bahir Dar, Ethiopia, 2015

Abstract: In this study, an extrusion process that is used for production of UPVC (Un-plasticized Poly Vinyl Chloride) pipes at APF was studied. The study begins: 1st: By identifying possible product quality problems that adversely affect rate of rework or recyclability; 2nd: By prioritizing those quality problems which needs company top management intervention to minimize as much as possible, and 3rd: By performing cause and effect analysis to strengthen and diagnose causes corresponding to the identified, prioritized quality problems. Finally, Taguchi design of experiment is systematically approached from the existing APF extrusion machine temperature set- points to identify the possible Response output values approaching or nearest to the Taguchi Design; L27 OA. Additionally, Taguchis loss function was also discussed in detail.

The study quality characteristics focused on are variation on thickness of pipe. One method of reducing UPVC pipes thickness variation problem is by applying and using Taguchi method. Nominal-The-Better approach is selected for both S/N ratio & loss function analysis. Afterwards Taguchi approach is used to establish optimum parameter combinations that reduce variation on performance of pipe due to asymmetric pipe production. An optimal set of process parameters obtained from a response graph for the main. Moreover, a signal-to-noise ratio computation and analysis of variance (ANOVA) are conducted to evaluate the results of the experiments. Using ANOVA, the significant factors impacting the quality characteristics of UPVC Pipes are obtained at a 90% confidence interval of recorded data.

Furthermore, the extrusion die is initially modeled with Solid Works to demonstrate the distribution of pressure and velocity inside the die to investigate does they have effect on the Quality Characteristics selected in the study using SW Flow simulation.

Key Words: UPVC Pipe, Pareto Diagram, Cause and Effect Diagram, Loss Function Analysis, Taguchi Method, ANOVA, Flow Simulation

1. INTRODUCTION

   Organizations are created to achieve a goal, mission or objective but they will only do so if they satisfy the needs, requirements and expectations of their stakeholders. Their customers, as one of the stakeholders, will be satisfied only if they provide products and services that meet their needs, requirements and expectations. Their other stakeholders (shareholders, employees, suppliers and society) will only

   be satisfied if the products and services provided to customers are produced and supplied in a manner that satisfies their needs, requirements and expectations in other words, it makes a profit [1].

   Improving quality is very often regarded as an activity which is going to increase cost. This view confuses the terms used in industry concerning quality and grade. Improving or raising the grade of products relates to the use of more expensive materials or processes to produce a product and will raise product costs. Improving quality means among other things, making less faulty products with the same amount of effort or cost which usually gives a lower unit cost [2].

   In our country quality assurance is adopted at the stage of inspection after the product leaves production process and /or passed to next step (if it is in the range of tolerance limit of the specification). Otherwise, it might be scrap or rework (if it is not in the range of tolerance limits) which both incur loss for the factory. The purpose of this thesis is to put Quality Engineering (Taguchis Method) into perspective and to highlight its importance in quality improvement at the early stage of product and process design, as well as to present an optimum combination of process parameters setting that yield a sound product with better quality performance characteristics.

   There are a lot of literatures written for optimization of manufacturing process parameters using Taguchi Method. Some of these are:

   Adeel Ikram et al. [3] reports the effect and optimization of eight control factors on material removal rate (MRR), surface roughness and kerf in Wire Electrical Discharge Machining (WEDM) process for tool steel D2. The experimentation is performed under different cutting conditions of wire feed velocity, dielectric pressure, pulse on-time, pulse off-time, open voltage, wire tension and servo voltage by varying the material thickness. Taguchis L18 orthogonal array is employed for experimental design. Analysis of variance (ANOVA) and signal-to- noise (S/N) ratio are used as statistical analyses to identify the significant control factors and to achieve optimum levels respectively.

   Narasimha & Rejikumar [4] presented a systematic approach to find the root causes for the occurrence of defects and wastes in plastic extrusion process. The cause- and effect diagram was implemented to identify the root causes of these defects. The extrusion process parameters such as vacuum pressure, temperature, take-off speed, screw speed of the extrusion process and raw material properties were identified as the major root causes of the defects from the cause and-effect diagram. The quality loss for the current performance variation was calculated using Taguchis principle of loss function and requirement for improvement was verified. In their paper design of experiment (DoE) was applied to optimize the process parameters for the extrusion of high-density polyethylene (HDPE) pipe Ø50mm and plain pipe 25mm. Four independent process parameters involving vacuum pressure, take-off speed, screw speed and temperature were investigated using Taguchi method. Minitab 15 software was used to analyze the result of the experiment. Based on the result of the analysis, optimum process parameters were selected.

   Krishankant et al. [5] studies optimization of turning process by the effects of machining parameters by applying Taguchi methods to improve the quality of manufactured goods. EN24 steel is used as the work piece material for carrying out the experimentation to optimize the Material Removal Rate. they select three machining parameters i.e. Spindle speed, Feed rate, and Depth of cut. Different experiments are done on single point cutting tool made of high speed steel on Lathe by varying one parameter and keeping other two fixed so maximum value of each parameter was obtained. Operating range is found by experimenting with top spindle speed and taking the lower levels of other parameters. Taguchi orthogonal array is designed with three levels of turning parameters with the help of Minitab 15 software i.e., L9.

   Mekonnen L. Nekere et al. [6] took two groups of aluminum blank sand casting processes for comparison of 1st : Single aluminum blank sand casting and 2nd : double aluminum blank sand casting aluminum blank green sand (green) casting process to optimize the process parameter using Taguchis robust design approach. By changing different sittings of the casting process, they attempt to obtain optimal settings of parameters that affect the various quality characteristics of the product made from aluminum. test.

   Cunsheng Zhang et al. [7] applied Taguchis design of experiment and numerical simulation in the optimization of an aluminum profile extrusion process. By means of HyperXtrude simulation software, the extrusion process was simulated and the effects f process parameters on the uniformity of metal flow and on the extrusion force were investigated with the signal to noise ratio and the analysis of variance.

   Alireza Akbarzadeh et al. [8] have studied optimization of parameter in Injection Moulding using Statistical Methods and IWO Algorithm statically. They tried to convince the effect of changing injection moulding parameters on the shrinkage behavior of polypropylene (PP) and polystyrene (PS) plastic materials.

   S. Kamaruddin et al. [9] conduct a study to improve the quality level of an injection molding plastic tray product, made from blends plastic (75% polypropylene (PP) and 25% low density polyethylene (LDPE)) by optimizing the injection molding parameters using the Taguchi method. An orthogonal array (OA), main effect, signal-to-noise (S/N) ratio and analysis of variance (ANOVA) are used to analyze the effect of injection molding parameters on the behavior of the plastic Tray. Their study shows that the optimal combination of parameters that gives a sound product (Plastic Tray) are low melting temperature, high injection pressure, low holding pressure, long holding time and long cooling time.

   Vidal et al. [10] in their study attempt to optimize The Friction Stir Welding (FSW) process which is a solid state mechanical processing technology enabling high quality joints in materials previously considered with low weldability such as most of the aeronautic aluminum alloys. The Taguchi method was used to find the optimal FSW parameters for improvement of mechanical behavior of aluminum alloy. The parameters considered were vertical, downward forging force, travel speed and pin

   length. An orthogonal array of 9 (34) was used; ANOVA analyses were carried out to identify the significant factors affecting tensile strength, bending toughness and hardness field experiments.

2. TAGUCHI METHOD

   DoE is a popular product and process improvement tool for Engineers and scientific professionals that allow easily learning and applying the technique product design optimization and production problem investigation. It provides a method for simultaneously investigating the effect of multiple input factors (process parameters) on the desired output, or response variable. DoE is useful for obtaining information about processes so that critical product and process characteristics can be identified, monitored, and kept on target. The design of experiment is selected to identify the best set of parameter combination among the effective factors by cutting down a number of experiment trials in to smaller number of experiments using orthogonal array.

   As Dr. Taguchi devote his life on searching of quality improvement strategy, he pointed out that by applying design of experiment techniques one could improve the performance of a particular product and process design in such a way that helps by improving consistency of performance & save cost, and building Robustness, or insensitivity of process parameters towards uncontrollable or noise factors.

   Here in this study data were collected for UPVC pipe products of the company and by using Minitab software, process factors mostly affecting the response output i.e., thickness variation in this case would be determined. Minitab provides a lot of statistical function for easily analyzing data and ease to use statistical tools. One of its applications involves using DoE (Taguchi method).

3. DATA COLLECTION AND INTERPRETATION

   Here an assessment on the finished product quality characteristics and the possible defects associated with product was investigated

   (A)Shrinkage

   (B)Scratching & Pin Hole

   1. Thickness variation

   2. Inner diameter variation

      Others (1) Tearing (2) Folding

      (3) Cracking (4) chamfering problem Figure 1: Quality defect of UPVC pipe

      Due to an insertion of Mandrel and Calibrator the outer diameter of the pipe is not changed. The Calibrator keeps the pipe to hold its outer diameter. The Mandrel allows to produce different internal diameter pipe i.e., different thickness as well. The length of the pipe may vary due to inadequate supply of weight of raw material required during production process, and it is not taken as a defect characteristic since its compensated by some other pipe that would be produced with a little longer than the required length or vice versa. So such a length variation happen, the pipe would have been isolated not being rejected. For small pipe dimension mostly seen defect problem is shrinkage due to improper working of rubber hose at the vacuum chamber that sprays water inadequately, or not fully for cooling. Internal diameter variation on pipe is mainly caused by overloading of pipes together and poor material handling and storage. Burnt problems are merely caused by sudden electric power fluctuation, overheating of die and Extruder

      .

      Therefore, the product quality defects of UPVC pipe were tried to be assessed and represent by statistical quality control tools.

      Hence, for UPVC product data about the total frequency of defects was taken and analyzed using Pareto chart to identify the most frequently occurring defect.

      Total customer order = 450 pipe

      OD= 110mm; Nominal pressure = 4 bar; Thickness = 2.2

      2.7 mm; Length= 6m

      Figure 2: Pareto chart

      Among UPVC pipe defects about 44.4% of defects are contributed by shrinkage. Thickness variations, inner diameter variation, screeching are also other contributors of the companys pipe defects. Burnt problems and others (Cracking, Tearing, Color variation etc) are less contributing type of defects on UPVC pipe product type 4.

      Cause and effect diagram constructed using Minitab software as shown in Figure 3 that helps to find out the main and root causes having effect on UPVC pipe thickness variation problem. In Figure 3, some of the causes for thickness variation have also effect on other possible quality defects.

      The main causes for thickness variation problem are identified and put down to measurement, labor, method, machine performance, and material. The table below shows the cause and sub-causes that have an effect on UPVC Pipe thickness variation problem at APF.

      Cause and Effect Diagram

      Cause and Effect Diagram

      Measurements Material

      Personnel

      Measurements Material

      Personnel

      Lack of instrument for

      Die opening

      Trial-and Error

      adjustment

      In-appropriate material composition

      Dryness of powder at the die-opening section

      Agglomeration of material-mix

      Lack of instrument for

      Die opening

      Trial-and Error

      adjustment

      In-appropriate material composition

      Dryness of powder at the die-opening section

      Agglomeration of material-mix

      Thickness Variation

      Thickness Variation

      Sudden stoppage of vacuum

      chamber

      Un-optimized Die-opening Clearance

      Goal-post mentalist approach Improper insertion of Mandrel

      Absence of focus group Improper use of Carvnliator

      Inspection- oriented program

      Lack of team work

      Sudden stoppage of vacuum

      chamber

      Un-optimized Die-opening Clearance

      Goal-post mentalist approach Improper insertion of Mandrel

      Absence of focus group Improper use of Carvnliator

      Inspection- oriented program

      Lack of team work

      Methods

      Methods

      Size and type of powder

      Size and type of powder

      Education background Inadequate training

      Carelessness

      Operator fatigue

      Education background Inadequate training

      Carelessness

      Operator fatigue

      Frequent change in process

      parameter

      Machines

      Frequent change in process

      parameter

      Machines

      Ap plied

      Ap plied

      Ch emistry

      Ch emistry

      Indust rial

      Indust rial

      Che mistry

      Che mistry

      O ve rquali

      O ve rquali

      f ied , B .

      f ied ,B .

      Sc.,

      Sc.,

      Leakag e

      Leakag e

      Loo se of wat er

      Loo se of wat er

      pipe r ings

      pipe r ings

      O verh aul length

      O verh aul length

      of Ma ndrel

      of Ma ndrel

      Uneven wat er

      Uneven wat er

      spry on Pipe

      spry on Pipe

      sur face

      sur face

      Un- tighte ne d

      Un- tighte ne d

      screw knob

      screw knob

      Feed sp eed ( RPM)

      Feed sp eed ( RPM)

      Ma chine speed

      Ma chine speed

      A symmet ric

      A symmet ric

      (Mm/min)

      (Mm/min)

      cent er b etween

      cent er b etween

      Man drel & Un- even heat

      Man drel & Un- even heat

      C arvinlat or dist ribut ion in Die, and

      C arvinlat or dist ribut ion in Die, and

      Ada ptor

      Ada ptor

      Heate r

      Heate r

      Chemistry

      Chemistry

      Temperat ure prof ile

      Temperat ure prof ile

      (C^0 ) at:(HZ 1 , HZ 2 ,

      (C^0 ) at:(HZ 1 , HZ 2 ,

      HZ 3, HZ 4)

      HZ 3, HZ 4)

      Haul- of f speed

      Haul- of f speed

      (Mm/min )

      (Mm/min )

      Figure 3: Cause and Effect Diagram

      As shown in the diagram the main cause of thickness variation problem for UPVC pipe was different machine performance due to frequent change in process parameters by operators to bring the product in to a desired specified dimension through experience. The company runs its production in three shifts. During each shift operators are changed so as possible existence of change in process parameters. Therefore, the effects of these process parameters would be investigated and an optimal setting would be proposed to minimize thickness variation problem and loss ($) associated with desired performance characteristics would be determine by applying Taguchi DoE Approach.

4. LOSS FUNCTION ANALYSIS

   According to Taguchi Loss Function, Society incurs a loss when the product is shipped not on target To demonstrate the impact of deviation from target value, Taguchi introduce quality loss function, L(y) such that:

   L(y) = ( )2; Where:

   L(y) – The loss, $ due to failure, repair, recycling k – The proportionality constant

   y – Actual performance value m – The target value

   Figure 4: Quality loss Function [11].

   Table 1: Raw materials and associated over all purchasing cost used for UPVC pipe production.

   S. no	Raw Material	Purchasing Cost (Birr/kg)
   1	PVC resin	25.33774
   2	Stabilizer	46.77736
   3	Calcium carbonate, ()3	7.35659
   4	Titanium dioxide, ()2	92.16678
   5	Carbon Black	39.62537

   = 480.203 Birr/pc

   As we calculated above the average amount of money the company loss due to quality problem is estimated to be

   480.203 Birr per each UPVC Pipe of product.

   Summary for Thickness Variation

   A nderson-D arling Normality Test A -Squared 0.72

   P-V alue 0.059

   Mean 2.4308

   S tD ev 0.2740

   V ariance 0.0751

   Skew ness 0.72854 Kurtosis 2.45827

   N 156

   Minimum 1.8300 1st Q uartile 2.2450 Median 2.4250 3rd Q uartile 2.5975

   Estimated Raw material cost = 407.3601 = 31.54774 Birr/ kg

   12.9125

   using SWA. Then, the production cost is estimated as

   1.8

   2.1

   2.4

   2.7

   3.0

   3.3

   3.6

   Maximum 3.6600

   95% C onfidence Interv al for Mean 2.3874 2.4741

   95% C onfidence Interv al for Median

   2.4000 2.4500

   95% C onfidence Interv al for S tD ev

   465.104 Birr/pc. By dividing 465.104 Birr/pc to 6, we get

   Mean

   9 5 % Confidence Intervals

   0.2466 0.3083

   77.517 Birr/ m. And, the rate of defect for UPVC pipe is reaching above 17% of the total amount of production.

   Median

   2.40

   2.42

   2.44

   2.46

   2.48

   Since once defect on a product is obtained, it would be recycled and reworked that in turn add an extra production cost and even contribute to the final selling price of the product. This makes an assumption that 17% of pipe goes back to the production process to be recycled.

   Failure cost = production cost (Birr/pc) (mass of a product (kg/pc)

   % of recyclable cost of raw material (Birr/kg))

   = 465.104 / – (12.54 kg/pc* 17%* 31.54774 Birr/kg)

   = 397.851 Birr/pc

   As we calculate above, the failure cost that the company incurred for producing UPVC pipe ( 4) determined to be 397.851 Birr/pc.

   Taguchi investigated in his quality loss function the loss in quality quantified in a parabolic relationship with the deviation of performance characteristics from the target value. From quality control daily report, 156 number of sample data was taken to estimate the mean and standard division.

   Figure 6: Summery Report

5. EXPERIMENTAL PROCEDURE

Histogram of Thickness Variation

Normal

Mean 2.431

StDev 0.2740

Mean 2.431

StDev 0.2740

30

N

156

N

156

25

Frequency

Frequency

20

15

10

5

0

1.8

2.1

2.4

2.7

3.0

3.3

3.6

Thickness Variation

Figure 5: Histogram of Thickness Variation

The average quality loss for a product is calculated using a formula;

= k (2 + (µ )2); since µ = 2.431and s = 0.2740

= 6365.616* ((0.2740 )2 + (2.431 2.45)2)

Figure 7: Flow Chart of Taguchi – DoE for Extrusion Process

Figure 8: APF Extrusion machine components – Die Heater, Adapter and Extrusion Section

The table below summaries the variables and associated levels used in the experiment together with their assignments as shown in Table 2.

Table 2: Extrusion process parameters and their levels

The level should be chosen sufficiently far apart to cover a wide experimental region because sensitivity to noise factors doesnt usually change with small changes in control factor sittings. Although by choosing a wide experimental region, we can identify good regions as well as bad regions, for control factors [12].

In our case there are eight factors each with three levels are selected. From standard table of orthogonal array the preferable OA selected would be 27(38). Minitab software helps to generate Taguchi orthogonal array.

The letters A, B, C, D, E, F, G and H describe the control factors under study.

The number 1, 2, & 3 represent associated levels for each control factor.

Table 3: Experimental setup

UPVC pipe were produced from the material mix on the Jinhu Extrusion machine as per the temperature set- points shown in Table -3. Immediately after manufacturing each pipe, samples are taken and its thickness was measured with the help of Digital Micrometer instrument.

The difference between their targeted thickness and the actual thickness dimension indicated the amount of variation that took place in the pipe. The thickness variation that occurred in the peripheral dimension of the pipe was recorded and that is why thickness variation was considered as a response variable/output characteristic in this study. Less thickness variation means high dimensional stability and vice-versa. It should be noted that the thickness was measured at eight positions with each divided at 45 Degree as it is shown below in Figure 9.

Figure 9: UPVC Pipe Thickness Measuring Positions

To locate measurement positions on pipe cross section, a string of length 42.08mm is used.

Then, it is possible to separate out the effect of each parameter at different levels. By using Minitab software, each values of the response outpu are inserted in the worksheet and evaluated using a special Module to obtain response table for signal to noise ratio and response table for means.

Table 4: Possible factor combinations and their corresponding response outputs

Table 5: The S/N ratio Response Table

Die

Table 6: Response Table for Mean

Main Effects Plot (data means) for SN ratios

Main Effects Plot (data means) for SN ratios

A

B

C

Main Effects Plot (data means) for Means

A

B

C

Main Effects Plot (data means) for Means

40.0

37.5

35.0

A

B

C

40.0

37.5

35.0

A

B

C

162

180

187

179

185

190

162

180

187

179

185

190

130

130

135

D

135

D

139

139

186

186

192

E

192

E

197

197

183

183

188

F

188

F

193

193

40.0

37.5

35.0

40.0

37.5

35.0

175

175

183

G

183

G

189

189

168

168

173

H

173

H

178

178

163

163

175

175

185

185

40.0

37.5

35.0

40.0

37.5

35.0

162

162

180

180

187

187

179

179

185

185

190

190

Signal-to-noise: Nominal is best (10*Log(Ybar**2/s**2))

Signal-to-noise: Nominal is best (10*Log(Ybar**2/s**2))

2.52

2.51

2.50

2.52

2.51

2.50

130

130

135

D

135

D

139

139

186

186

192

E

192

E

197

197

183

183

188

F

188

F

193

193

2.52

2.51

2.50

2.52

2.51

2.50

175

175

183

G

183

G

189

189

168

168

173

H

173

H

178

178

163

163

175

175

185

185

2.52

2.51

2.50

2.52

2.51

2.50

Mean of SN ratios

Mean of SN ratios

Mean of Means

Mean of Means

Figure 10: Main Effects Plot for S/N Ratio

By Looking at the response tables and main effects plots for the signal-to-noise (S/N) ratios to see which factors have the greatest effect on S/N ratio, which in this case is Nominal-is-best is used. In this case, the factor with the biggest impact on the S/N ratio is Feed Drum Heat-4 (Delta

= 7.34, Rank = 1). And the factor with the least effect on S/N ratio is Mold Heat (2) (Delta = 1.35, Rank = 8). In order to reduce the impact of noise on response (i.e., reducing variation around the thickness of pipe), possible

parameter combination settings are the one with higher S/N ratio in each factor. Therefore, from response tables of S/N ratio we can decide that 1 , 3 , 1 , 1 , 2 , 2 , 1 , 1 are the possible parameter combinations for reducing the variation around thickness of pipe.

Figure 11: Main Effects Plot for Means

Here in the response table and main effects plots for mean both show that the factor with the greatest effect on the mean is Mold Heat (3) (Delta = 0.026, Rank = 1). And the factor with the least effect on the mean is core heat, or Adaptor (1) (Delta = 0.006, Rank = 8).Since main effect plots for means is helpful for adjusting the mean on target value, those parameter levels with near or close to the desired target thickness of pipe that means 1 , 2 , 2 , 3 ,

3 , 3 , 1 , 3 would be taken as a possible parameter

combinations for adjusting, or approaching the mean on targeted thickness of pipe to maintain its symmetry as

much as possible.

1. ANALYSIS OF VARIANCE (ANOVA)

   The purpose of the analysis of variance (ANOVA) was to find which parameters significantly affected the quality characteristic. By making analysis of each response values: General Linear Model Noise 1, Noise 2 against factor A, B, C, D, E, F, G and H; ANOVA results indicate whether or not there is statistically significance difference in output characteristics. Since ANOVA analysis involves one of the P values was below 0.05, then we go for validation of the result in either of three cases. I.e., Normality, Constant Variance, and Independence Test.

   Table 7: ANOVA Table

   Table 8: ANOVA Table with Corresponding CL

2. RESULTS AND DISCUSSIONS

   To reduce variation around thickness of pipe (control factor with the biggest impact on the S/N ratio) is identified to be Feed Drum Heat-4 (Delta = 7.34, Rank = 1). And the factor with the least effect on S/N ratio is Mold Heat (2) (Delta = 1.35, Rank = 8).

   In the response table and main effects plots for mean both show that the factor with the greatest effect on the mean is Mold Heat (3) (Delta = 0.026, Rank = 1). And the factor with the least effect on the mean is core heat, or Adaptor (1) (Delta = 0.006, Rank = 8).

   To demonstrate the distribution of pressure and velocity inside the die, the study begins with creating the Solid Works part (APF Extruder Die).

   The extrusion die is initially modeled with Solid Works; Figure 12 illustrates the assembled die in 1800cut section view in order to show all the parts that constitute it.

   Figure 12: Sectional View of APF Extruder Die Machine Showing Melts Flow Channel, And Melt Intake

   General Description

   A 3D model is developed for non-Newtonian materials being processed in the extrusion die based on the cofiguration of Figure 33 that its part dimensions were taken by direct measurement using Tape meter. In this study, an assumption that a homogeneous High Density Polyethylene (HDPE) and solid Epoxy resin melt with a uniform temperature are flowing into the die channel. The temperature of the die wall is kept constant and the volumetric flow rate of the polymer melt is fixed.

   Figure 13: Schematic illustration of solid flow volume inside the die channel

   Figure 14: Pressure Distribution at the Die Channel

   To look out the distribution of pressure inside the channel, and to see which regions of the die opening have gained lower pressure and which one was maximum pressure an equation goal is made by using the formula of pressure drop.

   Figure 15: Fully Open Surface Plot of Pressure Distribution at the Die

   Figure 16: Velocity Profile Showing Points where Velocity Varies

   Figure 17: Fully Open Surface Plot of Velocity Distribution at the Die

   Figure 18: Velocity Variation at Different Points of the Melt

   The velocity of the melt as it is shown in the figure above varies at different points of the melt which plays a vital role for the existence of production of pipes with different thickness even if it is smaller to quantify. To elaborate more what the velocity profile inside the die channel looks like, a plot is made as the melt passes through the channel for a certain prolonged period of time.

   Figure 19: Average velocity distribution via iteration

3. CONCLUSIONS

As part of the general outcome of this study, Taguchi method plays a vital role in any companies (whether engaged in manufacturing of tangible output or service rendering) to satisfy their customer needs, expectations by producing quality product or delivering services in a manner through reduction of unnecessary costs either prior to production or after the item was shipped to customers – meaning TM is helpful for reduction of unit manufacturing cost (UMC) and quality loss ($) respectively through incorporating ideas either at the product development or process design stage and also identifying optimum control parameter combinations. And, rather than directly choosing the response output, pareto analysis and cause and effect diagram needs to be incorporated to make the study visible so as it creates worth for the company profitability etc..

Beyond the factors set in TM, Melt pressure and Melt flow velocity have identified to be one of the contributing factors for production of defected products of Extrusion machine especially at the beginning of machine start up time till its stable.

REFERENCES

Figure 20: Static Pressure inside the Die Channel via Iteration

The pressure inside the die channel via iteration has a profile as it is shown in the figure above. Due to conduction effect of temperature between the melt and Heater pads the pressure distribution varies. At the start of production, most of the Extrudate have smaller injecting velocity via iterations and it have also smaller melt pushing pressure at the die outlet which both contributes highly defective pipe production at the beginning of machine start up.

Figure 21: Static Pressure at the end of Extruder DIE outlet via iteration

1. David Hoyle., Quality Management Essential, Published by Elsevier Limited, ISBN 10: 0-75-066786-9, ISBN 13: 978-0-75-066786-9, Amsterdam, 2007.

2. Jens J. Dahlgaard, Kai Kristensen, Gopal K. Kanji., Fundamentals of Total Quality Management, Process analysis and improvement, Taylor & Francis e-Library, London and new York, 2007.

3. Adeel Ikram et al., Parametric Optimization For Surface Roughness, Kerf And MRR In Wire Electrical Discharge Machining (WEDM) Using Taguchi Design of Experiment, Journal of Mechanical Science and Technology 27 (7) (2013), Pakistan, pp 2133- 2141.

4. M.Narasimha & R.Rejikumar., Plastic Pipe Defects Minimization, International Journal of Innovative Research & Development, ISSN: 2278 0211, Vol 2 Issue 5, Bahir Dar, 2013.

5. Krishankant et al., Application of Taguchi Method for Optimizing Turning Process by the effects of Machining Parameters, International Journal of Engineering and Advanced Technology (IJEAT), Volume-2, Issue-1, October 2012.

6. Mekonnen Liben Nekere, Ajit Pal Singh., Optimization of Aluminium Blank Sand Casting Process by Using Taguchis Robust Design Method, International Journal for Quality research, Vol.6, No. 1, 2012.

7. Cunsheng Zhang et al., Optimization of An Aluminum Profile Extrusion Process Based On Taguchis Method With S/N Analysis, Original Article, International Journal of Advanced Manufacturing Technology (2012) 60:589599, DOI 10.1007/s00170-011-3622-x.

8. Alireza Akbarzadeh & Mohammad Sadeghi., Parameter Study in Plastic Injection Molding Process using Statistical Methods and IWO Algorithm, International Journal of Modeling and Optimization, Vol. 1, No. 2, 2011.

9. S. Kamaruddin, Zahid A. Khan, S. H. Foong., Application of Taguchi Method in the Optimization of Injection Moulding Parameters for Manufacturing Products from Plastic Blend, IACSIT International Journal of Engineering and Technology, Vol.2, No.6, 2010.

10. Vidal et al., Application of Taguchi Method In The Optimization of Friction Stir Welding Parameters of An Aeronautic Aluminum Alloy, Lisbon, Portugal.

11. http://cs.anu.edu.au/courses/ENGN8101/Loss%20functions- lecture%205.pdf

12. Madhav S. Phadke., Quality Engineering Using Robust Design, AT&T Bell laboratories, PTR prentice Hall Inc, New Jersey, 1989.