Full Paper Isiem Dsg Trisakti

7/28/2019 Full Paper Isiem Dsg Trisakti

1/9

Proceeding, International Seminar on Industrial Engineering and ManagementMenara Peninsula, Jakarta, August 29-30, 2007 ISSN :

TEACHING DESIGN OF EXPERIMENT FOR INDUSTRIALSTATISTICS LABORATORY CLASS

Dedy Sugiarto.

Industrial Statistics Laboratory, Industrial Engineering Department, Trisakti UniversityJl. Kyai Tapa No. 1 Jakarta-Indonesia 11440, telp. 62-21-5663232 Ext. 407

[email protected] , [email protected]

ABSTRACT

Experience in designing and analyzing of experiments is very important for an industrial engineering student as his job later is related to the design, installation, and improvement of an integrated system. Books on design of experiments (DOE) have many exercises at the end of chapters that give students practise in the analysis of completed experiments, but students often receive little experiencein DOE. Generally they only get practice analyzing data that calculated manually or using statistical

software. DOE course also teaches many statistical principles of experimental design and sometimesnot fun for students. The purpose of this paper is to share some ideas about teaching design of experiments by giving examples of simple experiments for laboratory class that integrates practicein designing realistic experiments, running the experiments, and also practice analyzing data in sucha way that is easy to learn, fun, challenging, and memorable. This paper also gives series of simpleexperiment examples from literature that can done by students in laboratory classes or as homeassignments.

Key words : industrial engineering, design of experiments, simple experiments

1. INTRODUCTION

Experience in designing and analyzingof experiments is very important for anindustrial engineering student as his job later is related to the design, installation, andimprovement of an integrated system. Thecourse currently available in design of experiments (DOE) often tend to concentrateon more mathematical or statistical aspects of the subject rather than the techniques whichare more practically useful. Books on design

of experiments (DOE) have many exercisesat the end of chapters that give students practise in the analysis of completedexperiments, but students often receive littleexperience in DOE. Generally they only get

practice analyzing data that calculatedmanually or using statistical software.

The purpose of this paper is to sharesome ideas about teaching design of experiments by giving examples of simpleexperiments for laboratory class that

integrates practice in designing realistic

experiments, running the experiments, andalso practice analyzing data in such a waythat is easy to learn, fun, challenging, andmemorable. This paper also gives series of simple experiment examples from literaturethat can done by students in laboratoryclasses or as home assignments.

In order to keep the experimentsimple, all the control parameters werestudied at 2 or 3-level using single factor or factorial experiments. These simpleexperiments will form a student foundation

for studying the parameters at 3-level andhigher for the wider application in real-lifesituations or using other techniques of DOElike response surface, taguchi or mixtureexperiments.

2. DESIGN OF EXPERIMENTOVERVIEW

An experiment is a test or series of tests in which purposeful changes are made to

the input variables of a process or system so

Title (4 words) 1(surname first author)
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]


2/9

Proceeding, International Seminar on Industrial Engineerig and ManagementMenara Peninsula, Jakarta, August 29-30, 2007 ISSN :

that we may observe and identify the reasonsfor changes that may be observed in theoutput response (Montgomery 2005). DOE isa methodology for systematically applying

statistics to experimentation. DOE letsexperimenters develop a mathematical modelthat predicts how input variables interact tocreate output variables or responses in a

process or system (Lye, 2005).The following steps are useful while

one may be performing an industrialexperiment (Antony, 1998),:

1. Definition of the objective of theexperiment.

2. Selection of the response or output.3. Selection of the process variables or

design parameters (control factors),noise factors and the interactionsamong the process variables of interest. (Noise factors are thosewhich cannot be controlled duringactual production conditions, but mayhave strong influence on the responsevariability. The purpose of anexperimenter is to reduce the effect of these undesirable noise factors bydetermining the best factor levelcombinations of the control factors or design parameters. For example, in aninjection moulding process, humidityand ambient temperature are typicalnoise factors.)

4. Determination of factor levels andrange of factor settings.

5. Choice of appropriate experimentaldesign.

6. Experimental planning.7. Experimental execution.8. Experimental data analysis and

interpretation.

There are many types of DOEtechniques that include general factorial, two-level factorial, fractional factorial, responsesurface designs, mixture designs and taguchidesigns.

General full factorial designs (designs

with more than two-levels) is useful for small

screening experiments. Response surfacemethods are used to examine the relationship

between one or more response variables and aset of quantitative experimental variables or

factors. These methods are often employedafter you have identified a "vital few"controllable factors and we want to find thefactor settings that optimize the response.Designs of this type are usually chosen whenwe suspect curvature in the response surface(Minitab, 2000).

Special class of response surfaceexperiments in which the product under investigations is made up of severalcomponents of ingredients. Response is afunction of the proportions of differentingredients in the mixture. The quantities of components, measured in weights, volumes,or some other units, add up to a commontotal. Mixture experiment example is the tasteof lemonade depends on the proportions of lemon juice, sugar, and water (Minitab,2000).

Taguchi designs provide a powerful andefficient method for designing products thatoperate consistently and optimally over avariety of conditions. In robust parameter design, the primary goal is to find factor settings that minimize response variation,while adjusting (or keeping) the process ontarget (Minitab, 2000).

3. SIMPLE EXPERIMENT EXAMPLESCase 1 : Pop corn experiment

For example, we may want to investigate theinfluence of pop corn brands on the

proportion of un-popped kernels (minimize).We use completely randomize design or without blocking of experimental unit for thissingle factor experiment. There are tree levelsfor brand (A, B and C) and tree replicationsfor each lavel. We use one hundred kernelsfor each trial and 3,5 minutes to make popcorn on stove.

We wish to test hypotheses about thetreatment means, and our conclusion willapply only to the factor levels considered inthe analysis

2 Teaching Design of ExperimentDedy


3/9


Picture 1. Three brands of pop corn

Picture 2. Medium pan with cover

Picture 3. Popped and un-popped kernels

from tree brands

Table 1. The result of pop corn experiment

Run order

Un-poppedkernels

proportionBrand C 0,04Brand C 0,05Brand B 0,11Brand A 0,00Brand C 0,08Brand B 0,13Brand A 0,03Brand A 0,03Brand B 0,08

The small p-values for the brand (p =0.007) that lower than ( 0.05) suggest thereis significant effect of brand on proportion of un-popped kernels. Individual 95%confidence interval for mean of three brandsuggest that brand A has significantlydifference with brand B.

One-way ANOVA:

Source DF SS MS FPBrand 2 0,011356 0,005678 12,46

0,007Error 6 0,002733 0,000456Total 8 0,014089

S = 0,02134 R-Sq = 80,60% R-Sq(adj) =74,13%

Case 2 : Time to boil water experiment

For example, we may want to investigate theinfluence of pan size and cover on the time inminute to boil water. We use general full

factorial design and completely randomizedesign or without blocking of experimentalunit. Single. There are two-level for eachfactor and 3 replications for eachcombination. Dimensions of small pan is 14cm for diameter and 10 cm for height.Dimensions of medium pan is 18 cm for diameter and 11 cm for height. Volume of water is 600 ml.


Ho : 1 = 2 = . = a H1 : i j for at least one pair (i,j)


4/9


Picture 4. Small pan, medium pan and glasscover

Picture 5. Medium pan with cover on stove

Table 2. The result of time to boil water experiment

RunOrder pansize cover

time(minutes)

1 medium no 3.922 small yes 3.073 small yes 3.004 small no 3.535 medium no 3.636 medium yes 3.837 medium yes 3.208 small yes 3.22

9 small no 3.7510 medium yes 3.9811 medium no 3.5012 small no 3.60

The small p-values for the pan size (p= 0.058) and cover (p = 0.093) that lower than (0.10) suggest there is enoughsignificant effect of pan size and cover ontime to boil water. Interaction of pan size andcover is not significant. Mean plot of

rensponse suggests that small and cover (yes)give lower time to boil water.

General Linear Model: time (minutes)versus pan size, cover

Factor Type Levels Values pan size fixed 2 small, medium

cover fixed 2 yes, no

Analysis of Variance for time (minutes), using Adjusted SS for Tests

Source DF Seq SS Adj SS Adj MSF P

pan size 1 0.29768 0.29768 0.297684.90 0.058cover 1 0.22141 0.22141 0.221413.65 0.093

pan size*cover 1 0.20021 0.20021 0.200213.30 0.107Error 8 0.48560 0.48560 0.06070Total 11 1.20489

S = 0.246374 R-Sq = 59.70% R-Sq(adj) =44.58%

Unusual Observations for time (minutes)time

Obs (minutes) Fit SE Fit Residual StResid

7 3.20000 3.67000 0.14224 -0.47000-2.34 R

R denotes an observation with a largestandardized residua

M e a

n o

f t i m e

( m i n u

t e s )

mediumsmall

3.70

3.65

3.60

3.55

3.50

3.45

3.40

3.35noyes

pan size cover

Main Effects Plot ( data means) for time (minutes)

Picture 6. Mean plot of response

Case 3 : Painting experiment

We may want to investigate the influence of painting methods (dipping, spray and brush)



5/9


and brand of paint on visual quality using 10 point scale. As a standard for point 10, we usesample product from PT. Safira TumbuhBerkembang (wooden toys producer).

General full factorial design and completelyrandomize design used in this experiment.There are tree levels for painting methods andtwo levels for paint brands and threereplications for each combination. Thisexperiment is part of ManufacturingIndustrial Design Lab. in IndustrialEngineering Department Trisakti University.The end product is wooden toy train.

Picture 7. Materials and equipments of paintingexperiment

Picture 8. Preparation of experimental unit

Picture 9. Dipping method for painting

Picture 10. Spraying method for painting

The results and anlysis of this experiment arenot written on this paper.

4. CONCLUSION

1. This paper has shown the benefit of employing a systematic approach tosimple experimentation using DOE,rather than utilising a trial and error approach

2. The paper has also illustrated somesimple experiments that can be used as a

powerful teaching and learning tool inindustrial statistics laboratoy.

3. These simple experiments will form astudent foundation for studying DOE for the wider application in real-lifesituations or using other techniques of DOE like response surface, taguchi or mixture experiments.

5. REFERENCES



6/9


(a) Antony J, N Capon (1998). TeachingExperimental Design Techniques toIndustrial Engineers. Int. J. EngngEd. Vol. 14, No. 5, pp. 335-343

(b) Antony J, N Capon. Some key thingsindustrial engineers should knowabout experimental design. Logistics

Information Management Volume 11 Number 6 1998 pp. 386392

(c) Hunter WG. 101 Ways to Design anExperiment, or Some Ideas AboutTeaching Design of Experiments.1975.http://williamghunter.net/articles/101doe.cfm

(d) Lye LM. Tools and toys for teachingdesign of experiments methodology.33rd Annual General Conference of the Canadian Society for Civil

Engineering . 2005(e) Mackisack M (1994). Incorporating

Favourite Experiments: AnAddendum to What is the Use of Experiments Conducted by StatisticsStudents?. Journal of Statistics

Education v.2, n.1(f) Martinez-Dawson R (2003).

Incorporating LaboratoryExperiments in an IntroductoryStatistics Course. Journal of Statistics Education Volume 11

(g) Minitab (2000). Users Guide 2 : Data Analysis and Quality Tools Release 13 for Windows

(h) Montgomery DC (2000). Design andAnalysis of Experiments. New York :

John Wiley & Sons, 2005

http://williamghunter.net/articles/101doe.cfmhttp://williamghunter.net/articles/101doe.cfmhttp://williamghunter.net/articles/101doe.cfmhttp://williamghunter.net/articles/101doe.cfm


7/9


APPENDIX. Some simple ExperimentsExamples.

No Response Factors

1 taste (maximize),un-poppedkernels(minimize) of Microwave

popcorns

brand, time, power,height (on bottom or raised)

2 quality of slidesPhotography

film speed, lighting,shutter speed

3 time to boil water pan type, burner size, cover

4 distance golf ballhit on full

tee height, ball age,club type

5 squash ballrebound distance

ball type, ball age, ball temperature

6 distance footballkicked

kick style, angle togoal, angle of ball

7 number of pinsdown

bowling ball weight,spin, position for throw

8 water evaporation

surface area, sun or shade, volume atstart

9 car acceleration time to gear change,road surface, car type (?)

10 number of baskets thrown

ball size, distancefrom net

11 letter deliverytime

post code, mailboxor post office, town(?)

12 speed running distance, terrain,time of day

13 distance paper aeroplane flew

design, paper weight, angle

14 longjump

distance

runup distance,

shoes on or off, L or R foot

15 time to swim25m

shirt on or off,flippers, goggles

16 time to deliver pizza

weight crustthickness, garlic

bread, Coke17 time to boil water amount of water, lid

on or off, size of pan18 virus scan time RAM cache,

program size,operatingsystem (?)

19 distance cricket ball hit

weight of bat,wearing gloves (?)

20 number of corns popped/100

pot diameter, oil or margarine

21 time to completefixed course on bicycle and pulserate at finish

seat height (26, 30inches), generator (off,on), tire

pressure (40, 55 psi)22 yield of popcorn brand of popcorn

(ordinary, gourmet),size of batch (1/3,2/3cup), popcorn to oilratio (low, high)

23 quality of bread,especially thetotal rise

amount of yeast,amount of sugar,liquid (milk, water),

rise temperature, risetime

24 how well twins,who had colds,slept during thenight

number of pills,amount of coughsyrup, use of vaporizer

25 quality of slidesmade close upwith flashattachment oncamera

speed of film, light(normal, diffused),shutter speed

26 growth rate of algae in saltwater aquarium

hours of illumination, water temperature, specificgravity of water

27 taste of Koolaid temperature, amountof sugar, food prior to drink (water,salted popcorn)

28 strength of radiosignal from

particular AMstation in

Chicago

direction in whichradio is facing,antenna angle,antenna slant

29 blending time for soy beans

blending speed,amount of water,temperature of water, soaking time

before blending30 operation time

for pocketcalculator

charge time, digitsfixed, number of calculations

performed31 time until dryer

stopsclothes dryer (A,B),temperature setting,

load



8/9


32 time to boil water pan (aluminum,iron), burner onstove, cover for pan(no, yes)

33 hours of relief from migraineheadache

aspirin buffered?(no, yes) dose, water temperature

34 taste comparisonof homemadeyogurt andcommercial

brand

amount of milk powder added tomilk, heatingtemperature,incubationtemperature

35 time required torun up steps and

heartbeat at top

pack on back (no,yes), footwear

(tennis shoes, boots),run (7, 14 flights of steps)

36 length of flight of model airplane

width to height ratioof sheet of balsawood, slant angle,dihedral angle,weight added,thickness of wood

37 how much coffeespilled whilewalking

level of coffee incup, devices(nothing, spoon

placed across top of cup facing up),speed of walking

38 cost of knittingscarf, dollars per square foot

type of stitch, yarngauge, needle size

39 time to get steel ball through amaze

type of drink (beer,rum), number of drinks, rate of drinking, hours after last meal

40 cost of order of french fries, incents per ounce

size of order, time of day, sex of server

41 gas mileage for car

brand of gasoline,driving speed,temperature

42 number of daysrequired for letter to be delivered toanother city

stamp (first class, air mail), zip code(used, not used),time of day whenletter mailed

43 length of side of face (left,

whiskers 3 daysafter shaving

right), beard history(shaved once in twoyears0-sideburns,shaved over 600times in two years-

just belowsideburns)

44 number of times(out of 15) thatcorrect gender of

passerby wasdetermined byexperimenter with poor eyesight wearingno glasses

eyes used (both,right), location of observer, distance

45 number of shotthat penetrated aone foot diameter circle on thetarget

distance to target,guns (A,B),

powders(C,D)

46 height of cake oven temperature,length of heating,amount of water

47 density of photographicfilm

strength of developer,temperature, degreeof agitation

48 length of rubber band before it broke

brand of rubber band, size,temperature

49 speed of H.O.scale slot racers

viscosity of oil, typeof pick-up shoes,number of teeth ingear

50 time for car tocover one-quarter mile

type of tire, brand of gas, driver (A,B)

51 time to dissolve

table salt

temperature, stirring

rate, amount of solvent

52 taste of stewedchicken

amounts of cookingwine, oyster sauce,sesame oil

53 angle necessaryto make objectslide

type of surface,object (slide rule,ruler, silver dollar),

pushed? (no,yes)54 number of kicks

necessary to start

motorcycle

ambienttemperature, choke

setting, number of



9/9


charges55 time to bake

biscuitstemperature,location in oven,

biscuits covered

while baking?(no,yes)56 quantity of suds

produced inkitchen blender

temperature of water, amount of grease, amount of water conditioner

57 toys child choseto sleep with

person puttingdaughter to bed(mother, father), bedtime, place (home,grandparents)

58 correct answers

on simplearithmetic test,time required tocomplete test,wordsremembered(from list of 15)

amount of light in

room, type of music played, volume

59 bite, smokingcharacteristics,aroma, and tasteof tobaccomixture

amounts of addedTurkish, Latakia,and Periquetobaccos

60 time to melt ice temperature,humidity, rock salt

61 points in gamesof sheepshead, acard game

number of cardsdealt at one time,

position of picker relative to the dealer

62 taste of pancakes,consensus of group of four living together

amounts of flour,eggs, milk

63 time to get sun

burned

brand of suntan

lotion, altitude, skier 64 soundness of sleep, averagereading from five

persons

amount of sleep thenight before,substantial exerciseduring the day? (no,yes), eat right before

going to bed? (no,yes)

65 clearness andquality of sound,

and absence of noise

brand of tape deck used for playing

music, bass level,treble level,synthesizer? (no,yes)

66 time to filter type of filter paper, beverage to befiltered, volume of

beverage67 time to go down

ski slopetype of ski,temperature, type of wax

68 four quality

characteristics of pizza

ambient temperature

for dough whenrising, amount of vegetable oil,number of onions

69 time for seeds togerminate

amount of fertilizer,location of seeds (3x 3 Latin square)

70 quality of groundmalt for brewing

beer

speed of kitchen blender, batch sizeof malt, blendingtime

71 taste of drink from paper cup

soft drink (A,B),container (can,

bottle), sugar free?(no, yes)

72 number of swingsand duration of these swingsobtained from anautomatic infantswing

child's weight (13,22 pounds),springtension (4, 8 cranks),swing orientation(level, tilted)

Source : Lye LM (2005), Mackisack M (1994)and Hunter WG (1975)

The experiments are listed in the formResponse/factors. For those marked (?) it wasdifficult to control extraneous variation, andsignificant results were not obtained.


Full Paper Isiem Dsg Trisakti

Documents