I
Ingo1966
Hello DoE Experts,
Let's assume you have 10 Factors and you want to filter the major ones.
Also let's assume there are Interactions.
In my opinion the Blackett Burman (12 Runs) is much more clever for screening as a Fraction Factorial Plan (L16)
Because of the irregular pattern all possible interactions are confounded with all main effects and contaminates nearly equally. (Resolution III)
Enough to see which Factors do have a high Effect.
One Column is the Empty one (Dummy Effect) which gives me indication of existing interactions. --> and all in 12 Runs
Alternatively a Fraction Factorial L16 (10 Factors) has a regular Pattern and a specific Confounding of Main Effects to Interaction effect (Resolution III)
5 empty Columns with confounded 2 way interactions. (defined by Generators)
But 16 Runs.
What do you prefer for Screening ?
Let's assume you have 10 Factors and you want to filter the major ones.
Also let's assume there are Interactions.
In my opinion the Blackett Burman (12 Runs) is much more clever for screening as a Fraction Factorial Plan (L16)
Because of the irregular pattern all possible interactions are confounded with all main effects and contaminates nearly equally. (Resolution III)
Enough to see which Factors do have a high Effect.
One Column is the Empty one (Dummy Effect) which gives me indication of existing interactions. --> and all in 12 Runs
Alternatively a Fraction Factorial L16 (10 Factors) has a regular Pattern and a specific Confounding of Main Effects to Interaction effect (Resolution III)
5 empty Columns with confounded 2 way interactions. (defined by Generators)
But 16 Runs.
What do you prefer for Screening ?
Last edited by a moderator:
