Working Mechanics of D.O.E

<<<<<< Page 5 / 5 >>>>>>

 

Conduct Experiment, Collect & Analyse Data

• Having established the layout of the experiment, the runs of the experiment must be conducted in a random order, so as to eliminate bais.
• conducted the experiment, Collect Data
• Apply Mean & Variability anlysis to identify which factors (or interactions thereof) have an effect on mean & on variability of our process
• Focus must be on reducing the process Variability rather than Mean (there is not much point in having a process which has a satisfactory mean but within an unacceptable tolerance band
• Significance of factors can be establish through "Response Curves" as well as ANOVA (Analysis Of Variance)

Optimum Conditions & Confirmation Runs

• Identify optimum level of the significant factors which satisfies your optimisation criteria
• Predict theoretical performance at optimum conditions
• Conduct actual confirmation runs to establish degree of success of the DOE effort
• If satisfactory, then STOP
• If further optimisation required, conduct a 2nd loop (homing-in) DOE