DOE evaluation with random factors ends up with unrealistc results. Why and how this can be avoided?

[#1]

Hallo,

we have run a DOE and monitored some DOE factors (Ta_mes, c_mes) and some additional values that supposed to be stable but have not been so (Tb_mes, random test vehicle).

Modelling gives quite different predictions in case random factors are used for the model or not. They are even unrealistic (negative thicknesses s-2, or very high TE (higher than during the trials)).

see for example the Profiler HSS-TE versus HSS_random

Do you have any explanation why this is happening and how to avoid?

General explanation:

The DOE was designed with 8 whole plots and 16 subplots. The whole DOE was run in one tank. The wholeplots represent different plating days, and subplots the daily temperature increase (unfortunately cooling down would take to long - therefore the temperature Ta increase was always positive despite the original design). A third irreversible change (concentration increase c_set) was marked as block during design augmentation.

Results are on the right side of "Y".

[#2 · Accepted solution]

No, it would not make sense to change the design role from Continuous to Noise. That change is unnecessary. That is also not the purpose of the Noise factor type. Please see Help > Books > Design of Experiments for more information.

You are complicating the design (post hoc) and the analysis for no good reason. In the end, the sole purpose of an experiment is to arguably obtain the optimum data to fit a model, in this case, a model that is linear in the parameters. I updated your data table and attached it to one of my previous responses. The additional column properties are correct and assist the regression analysis. The saved model table scripts should also provide a reasonable data analysis.

[#3]

Thank you - (Taguchi designs with control and noise factors are new for me.)

I could not find the attached modell - I'am still curious how you handle the (chategorical) character type responses (LM-void and Peeling).

[#4]

Enclosed please find the modell with responses (data type "Character" for Peeling and LM-void):