Tolerance Limit Optimisation




The simultaneous adjustment of single or multiple process parameter settings is called as ‘tweaking the process’ or ‘tolerance limit optimisation’. Foundry experts do this all the time. Have a look at the example scatter of Zirconium. We want to find whether Top 50% values i.e. values in the range 0.26 to 0.032 are better for response Shrinkage as compared to the bottom 50% region i.e. values from 0.05 to 0.26. If Top 50% region is better, then it means that you are effectively moving the mean slightly upwards. So you are altering the tolerance limit of that parameter.


With penalty matrix approach the tolerance limit optimisation is undertaken with the quartile analysis of factor data that is super imposed by response specific penalty values. The quartile analysis allows to inspect bottom 25%, bottom 50%, middle 50%, top 50% and top 25% regions of the factor data. See data visualisation with p-matrix page for further details.