One Variable at a Time Approach

[salah haddad]

hi guys,
i am going to machine aluminum using EDM die sinking to study the MRR, TWR and Surface finish.
i have read many papers where "One Variable at a Time Approach is used" but do not understand how the input parameters are chosen or i can choose them randomly?
thanks

 

[Baluncore]

You should start with a recommended or best guess set of parameters. Then study one parameter at the time by plotting resulting performance against 5 or more different values of that parameter. Then try to progressively optimise the performance by cycling through the parameters making only small changes each time. That is a hill climbing search strategy. There may be other higher hills that you do not climb, so an early crude grid search of unexplored regions is worth the effort.

 

[salah haddad]

thanks for reply Baluncore;
as you said that "i should start with a recommended or best guess of parameters" and for those who like me lack of experience and first time run an experiment, how i know such values because the papers i have read all them machining steal no aluminum or copper using one variable at time. any advice?
thanks

 

[Baluncore]

Use the parameters for steel to start with and try +/– 25% variations. The parameters will be dependent on die electrode material also. The manufacturer of the machine will recommend some best performance specifications. They are advertising so divide them by two.

Your situation remains hopeless, as fear, uncertainty and doubt will paralyse you; until you take the first step.
Only then can you begin to explore the fascinating multidimensional world of EDM.