ME 107B Laboratory B: Process Design for Injection Molding

1
ME 107B Laboratory B Process Design for
Injection Molding

• Spring 2005
• Department of Mechanical Engineering
• University of California at Berkeley

2
Outline

• Administration Issues
• Motivation
• Injection Molding Theory
• Design-of-Experiments (DOE)
• Experiment Details

3
Administration Issues

• Contact Info
• Office 2114 Etcheverry Hall
• E-mail yshon_at_berkeley.edu, xchen_at_berkeley.edu
• Office Hours By appointment/drop in
• Handouts In PDF format on website (see below)
• DeVor, Chang, Sutherland, Statistical Quality
  Design and Control, Chapters 16,17
• Injection molding machine tutorial
• Websites
• ME 107B Adminstrative http//me.berkeley.edu/ME1
  07B/
• Lab Website http//www.me.berkeley.edu/ME107B/mol
  ding.html
• http//islnotes.cps.msu.edu/trp/inj/index.html
• http//www.idsa-mp.org/proc/plastic/injection/inje
  ction.htm

4
Motivation Why Study Injection Molding?

• Molding is one of the most widely used techniques
  for production scale plastics processing.

Metal cutting Plastic injection molds
PSP casings! Source www.mmsonline.com,
www.microsoft.com
5
Injection Molded Products Toys

• 1980s Transformers (Takara) vs. GI Joe (Hasbro)

6
Injection Molding Machine Specifics

• Major Process Parameters
• Temp barrel/nozzle/mold
• Pressure injection, hold
• Time injection, hold, cooling
• Material properties (melt/mold)

Source http//www.idsa-mp.org/proc/plastic/inject
ion/injection_process.htm
7
Typical Pressure/Temp. Cycle
Source http//islnotes.cps.msu.edu/trp/inj/inj_ti
me.html
8
Injection Molds
Haas milling machine
aluminum quad-connector mold
Cybercut process planner
assembled quad-connectors
Prof. Sequins quad-connector
CMOLD simulation
9
Process Defects
Mold fill simulation
Weld lines
short shot
flash
Source http//www.idsa-mp.org/proc/plastic/inject
ion/injection_design_7.htm
10
Introduction to Design of Experiments

• Example injection molding
• Pressure A, B, C
• Temp a, b, c

mass
Gravitational pull of Moon on mold
Response Curves

• Formal method of DOE Taguchi Analysis

11
Purpose of DOE

• To gain understanding of complex (ie.
  multivariable) manufacturing systemsHow can you
  find optimal variables?
• To minimize the experiments (they cost money!)
• Better/more refined/higher quality products. ie.
  more profit
• DOE techniques characterizes the behavior of
  complex systems with many input variables and
  quantifiable outputs DeVor
• Systematic technique for EMPIRICAL modeling
• System can be
• Manufacturing process (injection molding
  optimize fill and surface finish, cost - using
  pressure, various temperatures, etc.)
• Design (motor w/ torque if we use more expensive
  material in windings, can we use less material,
  lowering weight, while maintaining torque?)

12
Full Factorial Analysis Example

• Lets say were trying to make a paper airplane
  fly a long distance
• Two variables are determined to have an effect on
  distance
• Height and width
• These two variables are called FACTORS (B)
• Test each factor at two discrete LEVELS (A),
  High/Low
• (i.e. 20mm and 30mm)
• call them low -1, high 1
• Construct a set of experiments that exposes all
  of the factors to all combinations of levels of
  other factors.
• Plot results
• of experiments 2k where k is the number of
  variables

AB
width
height
13
Conducting Experiments

• Set the experimental table
• Conduct experiments according to table

14
Interpreting Data Main Effects

• The effect of a factor is defined as the change
  in response produced by a change in the level of
  the factor.
• The main effect of factor A is the difference
  between the average response at the high level of
  A and the average response at the low level of A,
  or

Eheight EA
Main Effect Plot (plotting averages)
15
Interpreting Data Interactions

• In some experiments, the difference in response
  between the levels of one factor is not the same
  at all levels of the other factors. When this
  occurs, there is an interaction between factors.

Interaction Plots (plotting each pt.)
16
Mathematical Modeling

• y b0 b1x1 b2x2 b12x1x2
• x1height, x2width
• b0 is an average of all four observed responses,
  25
• Polynomial, y 25 10x1 5x2 0x1x2

17
Statistics/Frequency Histogram

• After finding out optimal condition, with those
  variables, develop frequency histogram
• With large data set of, we can group the data
  into cells and make note of the frequency of
  observations falling within each cell.
• Is the signal robust? What uncontrollable factors
  exist?
• Test of precision of processhow tightly can we
  control?

18
Homework

• A series of experiments where done on an 8am
  class student body, testing average alertness
  as a signal, using factors of coffee intake (2
  cups / 1 cup), donut consumption (1 donut/no
  donut), and sleep (8 hours / four hours).

• Full factorial extraction of the effects (Es, Ec,
  Ed, Esc, Esd, Ecd, Escd),
• Two-way diagrams for the 3 factor interactions
• construct a model.
• Hint Es 18.75, Ec13.25

19
Injection Molding Machine
20
Experiment Details Objectives

• Your Mission (should you decide to accept)
• Get acquainted with injection molding/DOE (not a
  trivial step!).
• Perform a full-factorial DOE
• Characterize precision of measuring equipment.
• Decide which parameters (4) are important
• After you select the parameters, how will you fix
  other parameters value?
• Decide how you measure quality of parts
  (fill/surface quality/flash?)
• Conduct 24 experiments according to full
  factorial design.
• 3. Develop a mathematical model (polynomial).
• 4. Find out optimal parameters to use to
  minimize cost.
• 5. Develop frequency histogramestablish
  precision of process.

21
More Administrivia

• Late report policy
• Group organization
• Group Leader (1)
• Injection Molding Technicians (2)
• Metrology Technicians (3)
• Safety Glasses (obtain from machine shop)
• This W/ThMeet for machine training/walkthrough
• Next week
• Planning Report (due Monday)
• Homework (due Monday)
• Wed/Th, 1st experimental session