228.371 Assignment 2 DOE Project
228.371 Statistical Methods for Engineers and Technologists
DOE project assignment
AIM: To design a tw o-level multifactor experiment (> 3 factors), collect data, analyse the data, and derive appropriate, actionable conclusions from the analysis.
You need to use either a 2k full factorial design (Option A) or a 2k-p fractional factorial design (Option B) run in a single block. The report should contain a cover sheet. You must mention which option you selected in the cover sheet.
If you are using a full factorial design (Option A), the experimental strategy must include blocking (i.e. the experiment should run in 2 or more blocks).
Also, your experimental goals must include characterising how the phenomenon of interest (response) relates to the factors you have chosen and how you can manipulate your model to achieve a desired response (i.e. max Y, min Y or a target Y, depending on your experiment).
Brief
Design, perform, analyse (Via Minitab 17) and report on an experiment of your own
d e v i s i n g , based on the material presented on designed experiments. Your
experiment can be the everyday, home type experiment or a lab-based experiment (see note 2). Try to choose something you are interested in or have wondered about. Also try to ensure that your experiment has a purpose (as a common example, boiling water under different conditions is valid enough, but inherently pointless unless you can use the results for some purpose!) .
The phenomenon of interest or the nature of the problem that you investigate must be introduced with some background literature. This literature need not necessarily be of academic nature (newspaper articles and magazines are some examples of nonacademic sources of literature).
Once you decide what experiment you will run, make a guess on what your major conclusions will be and write those down (i.e. specify your key hypotheses in the scope and objectives section of the research). Upon completion of your experiment and analysis, compare your guesses against the actual conclusions drawn from the experiment. If
possible, carry out some confirmation trials to test the accuracy of your model’s
predictions.
Page 1 of 5
228.371 Assignment 2 DOE Project
Your report for Option A must follow the structure below:
Element/Title Description/Content to be included Marks
Introduction State the problem (the phenomenon of interest) and the
associated process that you are going to investigate. P r o v i d
b a c k g r o u n d l i t e r a t u r e ( 2 0 0 w o r d s a p p r o x . ) o n
t h e p h e n o m e n o n o f i n t e r e s t . 15%
The experimental factors and the response variable(s) (you are
free to have more than one response variable if your objectives
requires) should be discussed in this section.
Objectives The objectives, research hypotheses and expected results
should be presented in this section. 10%
Materials and Describe how the factorial experiment was conducted.
Method The description should include, but not be limited to: how
nonexperimental variables — background variables and
nuisance variables — were dealt with; the method(s) of
randomisation; number of replicates; measurement
considerations (accuracy, precision etc.); the experimental factors
and their operating levels (high and low levels); design
issues and challenges must come under a separate subsection
(this carries 10% of the 25% marks assigned for materials and 25%
methods).Design issues and challenges should, among other
things cover the care you took to eliminate potential systematic
error.
You also need to describe how you conducted the confirmation
runs to verify your model’s prediction at the optimum factor
settings.
Include one or two photographs depicting your experimental set
up. Any practical problems encountered (e.g. measurement
problems) and steps taken to tackle these should also be
explicitly stated in this section.
Results and Full analysis of results (you need t o u s e Minitab) providing
Discussion the o u t p u t s and graphs are clearly understandable to
the reviewer / marker.
Use the analysis strategy discussed in study guide/lecture slides.
The lab handouts would also be useful. You should also
show/discuss:
The practical or engineering significance of the significant main 35%
effects and interactions.
The f i n a l s t a t i s t i c a l m o d e l t h a t predicts Y, with
a p p r o p r i a t e consideration of its adequacy.
Include a discussion about what the confirmation runs indicated
about your model.
Conclusion Extent to which aims and objectives were satisfied, your final
model, limitations of your study, and suggested improvements to 15%
your experimentation and/or analysis.
Total Marks 100%
228.371 Assignment 2 DOE Project
Notes
1. Submission:
• This assignment must be uploaded via Stream. In addition to the project report (in MS Word or pdf), the accompanying Minitab project file should also be uploaded. Please do not email directly to teaching staff.
• ‘Turnitin’ plagiarism detection software will be applied to each assignment on upload. This will compare assignments with published sources, and also with previous assignments.
2. Report presentation and length
• The report should be no more than 20 pages (including any appendices).
• Present and discuss your findings, putting each graph, table or analysis that is being referred to into the text at that point (rather than in appendices). This makes it easier for the marker to read. Look at some academic journal papers for guidance on this if necessary.
• A l l t a b l e s a n d f i g u r e s s h o u l d b e n u m b e r e d . All analyses and outputs of
Mi ni tab must be relevant and justified. Thus no table or figure should be left
unexplained.
• If your experiment is a laboratory experiment, sufficient evidence must be provided to convince us that your work is original.
3. Some ideas of suitable experiments (taken from past projects):
• Stationary bike / treadmill (slope, speed, seat height etc. = heart rate, distance).
• Baking bread / muffins / biscuits etc. (ingredients, oven temp, oven time etc. = total rise, sensory quality of product)
• Photography (speed of film, light, shutter speed etc. = picture quality)
• Sustain of a guitar string (gauge, length, age etc. = length of sustain)
• Darts (left hand / right hand, distance, lighting etc. = consistency of hitting bullseye)
4. You are welcome to discuss your experimental ideas with your lecturer (for Albany students in particular, email could be a convenient option).
5. Banned topics include...
• Experiments based on computer / PS / Xbox (etc.) games. While entertaining to play
they are somewhat tedious to read about, and o n l y pseudo-random in the nature of their error variation, which leads to validity issues.
• Experiments that involve consumption of alcohol (or any other drug) as a factor. Such a factor is not controllable in its ‘levels’ (e.g. drunk/sober); nonlinear in its effects; and there are complicated ethical problems and risk management considerations!
Page 4 of 5
228.371 Assignment 2 DOE Project
• Experiments that constitute a danger to yourself, classmates or the public, such as exceeding speed limits or taking undue risks while driving vehicles, making and/or
discharging weapons in public places (although controlled target shooting etc. is OK), cycling while juggling and blindfolded etc.
6. Useful URLs for experimental ideas:
http://curiouscat.com/bill/101doe.cfm (Author: late William Hunter)
http://www.amstat.org/publications/jse/v2n1/mackisack.html (Author: Margaret Mackisack)
7. The Marking Scheme (For both options)
Marks Marks
Component Assessment Dimension Assigned Given
Clarity in the problem statement, the associated process and
Introduction background literature 10
[15 Marks] Factors and the response involved 5
Objectives
[10 Marks] Hypotheses (and/or expected outcomes) 10
Coverage on nuisance factors (controllable/uncontrollable) 3
Strategies for dealing with nuisance factors 3
Randomisation 3
Materials Levels and Factors 3
Design Matters including precautions you took to eliminate
and Method
systematic error, increase reliability and validity. Mentioning the
resolution of the design and confounding (if no confounding you
need to explain why) 5
Description of the experimental setup 6
[25 Marks]
Practical problems that you encountered in designing the
experiment 2
Reporting and discussing normal plot/s of effects 3
ANOVA/factorial Regression 10
Results and Factorial Plots and Discussion Main Effects 4
Discussion
Interaction 3
Cube Plots 3
Finalising the Model and its adequacy (discussion required) 10
[35 Marks] Confirmation (coverage of confirmation runs is adequate) 2
Conclusion A Summary of key findings 5
Limitations 5
[15 Marks] Further Work 5
Total Total 100
Page 5 of 5