Sei sulla pagina 1di 66

Design of Experiments (DoE)

Pendahuluan (Konsep Definisi) Exp with a Single Factor: ANOVA (Ch: 3) Randomized Bloks (Ch: 4) Latin Square (Ch: 4) Factorial Design (Ch:5) Two level Fractional Factorial Design (Ch:8) Taguchi approach

PENDAHULUAN
Mengapa Eksperimen
1. Menentukan penyebab penyebab variasi dalam response 2. Untuk menentukan kondisi yang optimal (maksimum atau minimum) sehingga optimasi response dapat dicapai 3. Membandingkan response menurut setiap level pada faktor faktor yang dikontrol 4. Mengembangkan model untuk prediksi responses

A Serious Problem...
I want my car to go fast like that one!

24-Dec-13

(c) 2001, Ron S. Kenett, Ph.D.

What Factors Affect the Speed?

Yes

Air Holes
No Slow

Fast

Shape Key Factor is: ___________________

24-Dec-13

(c) 2001, Ron S. Kenett, Ph.D.

Effect of Air Holes


Yes Slow Fast

Air Holes
No Slow

Shape

Key Factor is: _______________________________


24-Dec-13 (c) 2001, Ron S. Kenett, Ph.D. 5

DOE Balanced Effects

Yes Slow Fast

Air Holes
No Slow Slow

Shape

Key Factor is: _______________________________


24-Dec-13 (c) 2001, Ron S. Kenett, Ph.D. 6

KONSEP DEFINISI

Definisi
1. 2. Treatments different combinations of conditions that we wish to test Treatment Levels the relative intensities at which a treatment will be set during the experiment Factor (Treatment factor) one of the controlled conditions of the experiment (these combine to form the treatments) Experimental Unit subject on which a treatment will be applied and from which a response will be elicited also called measurement or response units Responses outcomes that will be elicited from experimental units after treatments have been applied

3.

4.

5.

Perlakuan berbagai kombinasi kosndisi dan situasi yang berbeda yang akan kita uji Level Perlakuan intensitas relatif dimana satu perlakuan akan disetel selama percobaan Faktor (Faktor perlakuan) salah satu kondisi terkontrol dari percobaan (kombinasi ini untuk membentuk perlakuan) Unit percobaan subyek pada yang mana satu perlakuan akan diterapkan dan dari yang mana satu tanggapan akan timbul juga pengukuran dipanggil atau unit tanggapan Tanggapi hasil yang akan timbul dari unit percobaan setelah perlakuan telah diterapkan

6.

Design of Experiments (DOE) also referred to as Experimental Design, this is the study of planning efficient and systematic collection of responses from experimental units 7. Experimental Design rule for assigning treatment levels to experimental units 8. Analysis of Variance (ANOVA) principal statistical means for evaluating potential sources of variation in the responses 9. Replication observing individual responses of multiple experimental units under identical experimental conditions 10. Repeated Measurements observing multiple responses of a single experimental unit under identical experimental conditions

Perancangan Percobaan (KIJANG BETINA) juga dikenal sebagai Desain Percobaan, ini adalah pembahasan dari efisien perencanaan dan koleksi sistematis dari tanggapan dari unit percobaan Desain percobaan ketentuan untuk taraf perlakuan penugasan ke unit percobaan Analisa varians (ANOVA) data statistik terpenting memaksudkan untuk mengevaluasi sumber potensial dari variasi pada tanggapan Replika tanggapan individu pengamatan dari unit percobaan perkalian pada kondisi percobaan serupa Mengulangi Pengukuran tanggapan perkalian pengamatan dari unit percobaan tunggal pada kondisi percobaan serupa

11. Blocking partition the experimental units into groups (or blocks) that are homogeneous in some sense 12. Covariate additional responses collected from the experimental units, usually to be used as predictors (and so are sometimes called predictive responses) these are not part of a designed experiment (why?) 13. Randomization nonsystematic assignment of experimental units to treatments 14. Blinding hiding which experimental units have been assigned to treatments from the analyst 15. Confounding design situation in which the effect of one factor or treatment can not be distinguished from another factor or treatment this is the experimental equivalent of perfect multicolinearity (why?)

Menghalangi sekat unit percobaan ke dalam group (atau halangi) yang homogen di beberapa rasa Covariate tanggapan additional terkumpul dari unit percobaan, biasanya dipergunakan seperti peramal (dan demikian sering menjadi tanggapan bersifat prediksi yang dipanggil) ini bukan bagian dari satu percobaan didisain (kenapa?) Pengacakan tugas nonsystematic dengan unit percobaan ke perlakuan Membutakan sembunyi yang mana unit percobaan telah ditugaskan ke perlakuan dari ahli analisa Mengacaukan keadaan desain darimana akibat dari faktor sesuatu atau perlakuan tidak dapat dicirikan faktor lain atau perlakuan ini adalah padanan percobaan dengan multicolinearity sempurna (kenapa?)

Experimentation
Manipulation of the values (or levels) of one or more (independent) variables or treatments and observation of the corresponding change in the values of one or more (dependent) variables or responses Manipulasi dari nilai (atau tingkat) dari satu atau lebih (bebas tak terikat) variabel atau perlakuan dan observasi dari perubahan sesuai pada nilai dari satu atau lebih (bergantung) variabel atau tanggapan

Basic Principles
1. Replication observing individual responses of multiple experimental units under identical experimental conditions 2. Randomization non systematic assignment of experimental units to treatments 3. Blocking partition the experimental units into groups (or blocks) that are homogeneous in some sense Replika tanggapan individu pengamatan dari unit percobaan perkalian pada kondisi percobaan serupa Pengacakan bukan tugas sistematis dengan unit percobaan ke perlakuan Menghalangi sekat unit percobaan ke dalam group (atau blok ) yang homogen di beberapa rasa

Chap 3. Exp with a Single Factor: ANOVA 1. Simple design: Experimen dengan satu faktor misalnya siswa di kelas A dg kurikulum KTSP vs kelas B dg MBS dalam UAN. 2. Faktor dapat dilakukan lebih dari dua level. 3. Analisis dilakukan dg Anova sederhana.

Strength
Cotton weight Observasi (replikasi) T Rerata Percentage 1 2 3 4 5 15 7 7 15 11 9 49 9.8 20 12 17 12 18 18 77 15.4 25 14 18 18 19 19 88 17.6 30 19 25 22 19 23 108 21.6 35 7 10 11 15 11 54 10.8
Sumber: Table 2.1 hal 62

ANOVA SV Tr E T df 1 n-2 ln - 1 SS SSE SST TJ SSTr G 2 /n nj MS MSE F

SSTr MSTr MSTr/MSE

SST x G /n
2 2

3-1. The tensile strength of Portland cement is being studied. Four different mxing techniques can be used economically. The following data have been collected:
Mixing Technique Tensile Strenght (lb/n2)

1 2 3 4

3129 3200 2800 2600

3000 3300 2900 2700

2865 2975 2985 2600

2890 3150 3050 2765

Test the hypothesis that mixing techniques affect the strength of the comnet. Use = 0,05.

3-4.

And in experiment was run to determine whether four specific firing temperatures affect the density of a certain type of brick. The experiment led to the following data:

Temperature

Density

100 125

21.8 21.7

21.9 21.4

21.7 21.5

21.6 21.4

21.7 21.5

150
175

21.9
21.9

21.8
21.7

21.8
21.8

21.6
21.4

Does the firing temperature affect the density of the brick? Use = 0,05.

3-6. A manufacturer of television sets is interested in the effect on tube conductivity data are obtained:
Coating Type Conductivity

1 2 3 4

143 152 134 129

141 149 136 127

150 137 132 132

146 143 127 129

Is there a difference in conductivity due to coating type? Use = 0,05. Estimate 3-8. An article in the ACI Material Journal (Vol. 84, 1947, pp. 213-216) describe several experiments investigating the rodding of concrete to remove entrapped air. A 3inch X 6-inch cylinder was used, and the number of times this rod was used is the design variable. The resulting compressive strength of the concrete specimen is the reponse. The data are shown in the following table:
Rodding Level Compressive Strenght

10

1530

1530

1440

15
20 25

1610
1560 1500

1650
1730 1490

1500
1530 1510

Is there any difference in compressive strength due the rodding level? Use = 0,05.

Bab 4 Randomized Block, Latin Sq, and Related Designs


1. 2. 3. 4. RCBD (Randomized Complete Block Design) Latin Square Design Graeco Latin Square Design BIBD (Balanced Incomplete Block Design)

1. RCBD (Randomized Complete Block Design)

2. The Latin Square Design

4-1. A chemist wishes to test the effect of four chemical agents on the strength of a particular type of cloth. Because there might be variability from one bolt to another, the chemist decides to use a randomized block design, with the bolts of cloth considered as blocks. She selects five bolt and applies all four chemicals ain random order to each bolt. The resulting tensile strength follow. Analyze the data from this experiment (use = 0.05) and draw appropriate conclusions.
Bolt Chemical 1 2 3 1 73 73 75 2 68 67 68 3 74 75 78 4 71 72 73 5 67 70 68

73

71

75

75

69

4-2. Three different washing solutions are being compared to study their effectiveness in retarding bacteria growth in 5gallon milk containers. The analysis is done in a laboratory, and only three trials can be run on any day. Because days could represent a potential source of variability, the experimenter decides to use a randomized block design. Observations are taken for four days, and the data are shown here. Analyze the data from this experiments (use = 0.05) and draw conclusions.
Days Solution 1 2 3 4

1
2 3

13
16 5

22
24 4

18
17 1

39
44 22

4-5. An article in the Fire Safety Journal (The Effect of Nozzle Design on the Stability and Performance of Turbulent Water Jets, Vol. 4. August 1981) describes an experiment in which a shape factor was determined for several different nozzle design at six levels of jet efflux velocity. Interest focused on potential differences between nozzle design, with velocity considered as a nuisance variable. the data are shown below:
Nozzle Design 1 2 3 4 5 11.73 0.78 0.85 0.93 1.14 0.97 Jet Efflux Velocity (m/s) 14.73 16.59 20.43 23.46 0.80 0.81 0.75 0.77 0.85 0.92 0.86 0.81 0.92 0.95 0.89 0.89 0.97 0.98 0.88 0.86 0.86 0.78 0.76 0.76 28.74 0.78 0.83 0.83 0.83 0.75

4-7. Any alumunium master alloy manufacturer produces grain refiners in ingot form. The comperating produces the product in four furnaces is known to have its own unique operating characteristics, soa any experiment run in the foundry thay involves more than one furnace will consider furnaces as a nuisance variable. The process engineers suspect that stiring rate affects the grain size of the product. Each furnace can be run at four different atirring rates. A randomized block design is run for a particular refiner, and the resulting grain size data is show below:
1. Graph the residuals from this experiment on a normal probability plot. Interpret this plot. 2. Is there any evidence that stirring rate affects grain seze? 3. Plot the residuals versus furnace and stirring rate. Does this plot convey any useful information? 4. What should the process engineers recommend concerning the choice of stirring rate and furnace for this particular grain refiner if small grain significant test.

Stirring Rate (rpm) 5 10 15 20

1 8 14 14 17

Furnace 2 3 4 5 5 6 6 9 9 3

4 4 9 2 6

4-13. An industrial engineer is conducting an experiment on eye focus time. he is interested in the effect of the distance of the object from the eye on the focus time. Four different distances are of interest. He has five subjects available for the experiment. Because there may be differences among individuals, he decides to conduct the experiment in arandomized block design. The data obtained follow. Analyze the data from this experiment (use = 0.05) and draw appropriate conclusions
Distanc e (ft) 1 4 10 6 7 8 5 19 6 2 6 6 3 4 Subject 3 6 6 3 4 4 6 1 2 2 5 6 6 5 3

ANOVA Latin Sq Design


Source RM Operator Latin Error Total DF 4 4 4 12 24 Seq SS 68 150 330 128 676 Adj SS Adj MS 68 17.000 150 37.500 330 82.500 128 10.670 F 1.590 3.520 7.730 P 0.239 0.040 0.003

4-14. The effect of five different ingredients (A, B, C, D, E) on the reaction time of a chemical process is being studied. Each batch of new material is only large enough to permit five runs to be made. Furthermore, each run requires approximately 1 hours, so only five runs can be made in one day. The experimenter decided to run the experiment as a Latin square so that day and batch effects may be systematically controlled. She obtains the data that follow. Analyze the data from this experiment (use = 0.05) and draw conclusions.
Batch 1 2 3 4 5 1 A=8 C = 11 B=4 D=6 E=4 2 B=7 E= 2 A=9 C=8 D=2 Day 3 D=1 A= 7 C = 10 E=6 B=3 4 C=7 D=3 E=1 B=6 A=8 5 E=3 B= 8 D=5 A = 10 C=4

4-15. An industrial engineer is investigating the effect of four assembly methods (A, B, C, D,) on the assembly time for a color television component. Four operators are selected for the study. Furthermore, the engineer knows that each assembly method produces such fatigue that the time required for the last assembly may be greater than the time required for the first, regardless of the method. That is, a trend develops in the required assembly time. To account for this source of variability, the engineer uses the Latin square design shown below. Analyze the data from this experiment (use = 0.05) and drw appropriate conclusions.
Order of Assembly 1 2 3 4 Operator
1 C = 10 B=7 A=5 D = 10 2 D = 14 C = 18 B = 10 A = 10 3 A=7 D = 11 C = 11 B = 12 4 B=8 A=8 D=9 C = 14

3. The Graeco-Latin Sq Design


Graeco Latin Square Design Montgomery tabel 4.9 and Tabel 4.20

RM 1 2 3 4 5

1 A=24 B=17 C=18 D=26 E=22

Operator 2 3 4 B=20 C=19 D=24 A=24 D=30 E=27 B=38 A=26 A=27 C=31 B=26 B=23 D=30 C=20 C=29

5 E=24 A=36 B=21 C=22 D=31

Operator RM 1 2 3 4 5 T 1 24 20 19 24 24 111 2 17 24 30 27 36 134 3 18 38 26 27 21 130 4 26 31 26 23 22 128 5 22 30 20 29 31 132 T 107 143 121 130 134 635

Graeco Latin 1 2 3 4 5 T 1 24 30 26 27 36 143 2 21 17 20 20 23 101 3 29 22 18 24 19 112 4 30 24 18 26 38 149 5 31 26 18 24 22 130 T 135 119 18 121 138 635

ANOVA Graeco-Latin Sq Design


Source DF Seq SS Adj SS Adj MS F P

RM
Operator Latin Greek Error

4
4 4 4 8

68
150 330 62 66

68 17.00
150 37.50 330 82.50 62 15.50 66 8.25

2.06
4.55 10.00 1.88

0.178
0.033 0.003 0.208

Total

24

676

4-22. The yield of a chemical process was measured using five batches of raw material, five acid concentrations, ive satnding times (A, D, C, D, E), and five catalyst concentrations (, , , , ). The GraceoLatin square that follow was used. Analyze the data from this experiment (use = 0.05) and draw conclusions
Acid concentrations 2 3 4 B = 18 C = 14 D = 16 C = 21 D = 18 E = 11 D = 12 E =16 A = 25 E = 15 A = 22 B = 14 A = 24 B = 17 C = 17

Batches 1 2 3 4 5

1 A = B = C = D = E =

26 18 20 15 10

5 E = A= B= C = D =

13 21 13 17 14

4-23. Suppose that in Problem 4-15 the engineer suspects that the workplaces used by the faour operators may represent an additional source of variation. A fourth factor, workplace (, , , ,) may be introduced and another experiment conducted, yielding the Graceo-Latin square that follows, Analyze the data from this experiment (use = 0.05) and draw conclusiona.
Order of Assembly 1 2 3 4 Operator 1 C = 11 B = 8 A = 9 D = 9 2 B = 10 C = 12 D = 11 A = 8 3 D = 14 A = 10 B = 7 C = 18 4 A = 8 D = 12 C = 15 B = 6

4. The BIBD
Cataliyst 1 2 3 4 Tj Table 4.22 Montgomery Raw Material 1 2 3 4 73 74 71 75 67 72 73 75 68 75 72 75 221 224 207 218 Ti 218 214 216 222 870

Perhitungan Anova BIBD sama dengan BCBD

ANOVA BIBD

4-27. An engineer is studying the mileage of performance characteristics of five types of gasoline additives. In the road test he wishes to use cars as blocks; however, because of a time constraint, he must use an incomplete block design. He runs the balanced design with the five blocks that follow. Analyze the data from this experiments (use = 0.05) and draw conclusions.
Additive 1 1 2 3 4 5 14 12 13 11 2 17 14 11 12 Car 3 14 13 11 10 4 13 13 12 12 5 12 10 9 8

4-29. Seven different hardwood concentrations are being studied to determine their effect on the strength of the paper produced. However, the pilot plant can only produce three runs each day. As days may differ, the analyst uses the balanced incomplete block design that follows. Analyze the data from this experiment (use = 0.05) and draw conclusions.
Hardwood Concentration (%) 2 4 6 8 10 12 14 Days 1 114 126 141 145 120 136 2 120 137 3 4 5 120 119 6 7 117

117 129

134
149 150 143 118

123 130

127

Bab 5 Factorial Design


5.1 Basic Definition 5.2 The Advantages 5.3 The Two Factor Factorial Design 5.4 The General Factorial Design 5.5 Fitting Response Curve and Surfaces

5.1 Basic Definition and Principles


a. Main Effect: the change in response produced by the change in the level of the factor. b. Interaction: diffrenece in response between the levels of one factor is not the same at all levels of the other factors. c. The three dimensional graph is called a response surface plot.

Ada interaksi pada eksperiment A. Tetapi ada interaksi pada eksperimen B

A -1 1 -1 1

B -1 -1 1 1

X 20 40 30 52

Y 20 50 40 12

5.2 The Advantages


1. efficient, 2. check the interaction, 3. the effects of a factor to be estimated at several levels of the other factors.

5.3 The Two Factor Factorial Design


An example Statistical Analysis of the fixed Effects Model Model Adequasy Checking Estimating the Model Parameters Choice of Sample Size The Assumption of No Interaction in a two factor model One Observation per cell

Table 5-1 Life (in hours) Data for Battery Design Example

Material Temperature Type 15 70 1 130 155 34 40 74 180 80 75 2 150 188 136 122 159 126 106 115 3 138 110 174 120 168 160 150 139

125 20 82 25 58 96 82 70 58 70 45 104 60

Desain Data untuk MINITAB


No 1 2 3 4 5 6 7 8 9 10 11 12 MT Temp x 1 1 130 1 1 155 1 2 34 1 2 40 1 3 20 1 3 70 1 1 74 1 1 180 1 2 80 1 2 75 1 3 82 1 3 58 No 13 14 15 16 17 18 19 20 21 22 23 24 MT Temp x 2 1 150 2 1 188 2 2 136 2 2 122 2 3 25 2 3 70 2 1 159 2 1 126 2 2 106 2 2 115 2 3 58 2 3 45 No 25 26 27 28 29 30 31 32 33 34 35 36 MT Temp x 3 1 138 3 1 110 3 2 174 3 2 120 3 3 96 3 3 104 3 1 168 3 1 160 3 2 150 3 2 139 3 3 82 3 3 60

Analisis of Variance
Two-way ANOVA: x versus MT, Temp Source MT Temp Interaction Error Total DF 2 2 4 27 35 SS MS F P 10684.0 5341.9 7.91 0.002 39119.0 19559.0 28.97 9613.8 2403.4 3.56 0.019 18231.0 675.2 77647.0

S = 25.98 R-Sq = 76.52% R-Sq(adj) = 69.56%

5.3.7 One Observation per Cell


Temp F 25 5 3 1 Pressure 30 35 40 4 6 3 1 4 2 1 3 1 45 5 3 2

100 125 150

Table 5.10 Impuriry

Hasil Analisis of Variance


Two-way ANOVA: Impurity versus Temp, Press Source DF SS MS F Temp 2 23.33 11.67 46.67 Press 4 11.60 2.90 11.60 Error 8 2.00 0.25 Total 14 36.93 S = 0.5 R-Sq = 94.58% R-Sq(adj) = 90.52%

P 0.00 0.00

5.4 The General Factorial Design


1. Kelanjutan dari two factor factorial design. Faktor A dengan a levels, faktor B dengan b levels, faktor C dengan c levels, dst. 2. Misal nya faktor ABC dengan levels a=3, b=2, dan c=2 tentang response x=fill height deviation.

Table 5.13 Pressure (B) Line Spee (C ) Carbonation (A) 10 12 14

Fill Height Deviation 25 psi 30 psi

200
-3 -1 0 1 5 4

250
-1 0 2 1 7 6

200
-1 0 2 3 7 9

250
1 1 6 5 10 11

Data dalam format MINITAB atau SPSS


No 1 2 3 4 5 6 7 8 9 10 11 12 A 10 10 10 10 10 10 10 10 12 12 12 12 B 25 25 30 30 25 25 30 30 25 25 30 30 C 200 250 200 250 200 250 200 250 200 250 200 250 X -3 -1 -1 1 -1 0 0 1 0 2 2 6 No 13 14 15 16 17 18 19 20 21 22 23 24 A 12 12 12 12 14 14 14 14 14 14 14 14 B 25 25 30 30 25 25 30 30 25 25 30 30 C 200 250 200 250 200 250 200 250 200 250 200 250 X 1 1 3 5 5 7 7 10 4 6 9 11

Analysis of Variance

5.5 Fitting Response Curve and Surfaces

i. Fitting response curve ii. Fitting response surface iii. Usinf regression iv. To predict

Table 5.16 Tool Life cutting speed Angle in/min degree 125 150 175 -2 -3 2 15 -1 0 3 0 1 4 20 2 3 6 -1 5 0 25 0 6 -1

No 1 2 3 4 5 6 7 8 9

A 15 15 15 15 15 15 20 20 20

B 125 150 175 125 150 175 125 150 175

X -2 -3 2 -1 0 3 0 1 4

No 10 11 12 13 14 15 16 17 18

A 20 20 20 25 25 25 25 25 25

B 125 150 175 125 150 175 125 150 175

X 2 3 6 -1 5 0 0 6 -1

5.6 Blocking in a Factorial Design


Untuk lebih efisien dan mengoptimalkan pelaksanaan eksperimen terkadang diperlukan blocking, hal ini untuk menghindari noise yang menggangu proses eksperimen. Blocking didasarkan atas keseragaman dari response yang akan muncul.

Table 5-19 Intensity Level at Target Detection operator block 1 2 3 4 filter type 1 2 1 2 1 2 1 2 Ground cutter low 90 86 96 84 100 92 92 81 medium 102 87 106 90 105 97 6 80 hught 114 93 112 91 108 95 98 83

Tabel 5.21 Radar Detection Experiment


Day 1 1 A 2 C 106 B 96 F 90 A 84 D 112 E 91 B 90 A 114 E 102 D 87 C 93 F 86 Operator 3 4 D 108 F 105 A 95 B 100 E 92 C 97 5 F 81 E 83 D 92 C 96 A 80 B 98 E 90 D 86 C 85 F 110 B 90 A 100 6 88 84 104

2
3 4 5 6

C
B E F D

91
98 92

5-4. An article in Industrial Quality Control (1956, pp. 58) describes an experiment to investigate the effect of the type of glass and the type of phosphor on the brightness of an television tube. The response variable is the current necessary (in microamps) to obtain a specified brightness level. The data are as follow: i.Is there any indication that either factor influences brightness? Use = 0.05. ii.Do the two factors iteract? Use = 0.05. iii.Analyze the residuals from this experiment

5-1 The yield of chemical process is being studied. The two most important variables are thought to be the pressure and the temperature. Three levels of each factor are selected, and a factorial experiment with two replicates is performed. The yield data follow: i.Analyze the data and draw conclusions. Use = 0.05. ii.Prepare appropriate residual plots and comment on the models adequancy. iii.Under what conditions would you operate this process?

5-17. The quality control department of a fabric finishing plant is studying the effect of several factors on the dyeing of cotton-sunthetic cloth used to manufacture mens shirt. Three operators, three cycle times, and two temperatures were selected, and three small specimens of cloth were dyed under each set of conditions. The finished cloth was compared and drw cpnclusions. Comment on the models adequacy.

UTS MK: Design and Analysis of Experiment


1.Dalam rancangan percobaan di kenal konsep: Simple design, RCBD, Latin Square, dan BIBD. Deskripsikan apa yang di maksud dengan empat desain/ rancangan tersebut. 2. Contoh: Anda diminta mengerjakan secara manual analisis rancangan percobaan sederhana (simple anova) dengandata pada soal: 3.6 (Petunjuk lihat mk statistik atau dg exel maupun spss).

3. Saudara diminta membuat rancangan pengolahan SPSS dari tampilan rancangan dalam soal Montgomery no: 3.17 3.30 4.2 4.5 5.7 4.15 4.22 4.27 4.29 Hint: lihat contoh disamping.