d e f es 
Coronavirus: continuation of our activities
Independently of the coronavirus epidemic... 
more Linie
AICOS Technologies at PCH Meetings 2020 in Lyon
Lyon - 25.-26.11.2020: International business... 
more Linie
Our training schedule for spring / summer 2020 is on-line!
You want to deepen your knowledge in data... 
more Linie
Infos > Design of Experiments (DoE)

Statistical Design of Experiments (DoE) case studies

Different situations - a single methodology: Statistical Design of Experiments (DoE)

In order to give an impression how to work with statistical Design of Experiments (DoE) and STAVEX, we subsequently present some examples with the corresponding graphics. More case studies from various areas of application can be found in our publications data base. bilder/ProductSTAVEX/STAVEX_overlay-e.jpg

  • Process optimisation:
    The overlay plot showing simultaneously yield (blue) and quality deviation (red, to minimize) facilitates the search for a suitable compromise setting. For instance, you could decide for the setting which promises the highest yield, given that the quality deviation does not exceed 1% (orange arrow).

  • Modified release tabletsbilder/ProductSTAVEX/STAVEX_surfaceDesirability-e.jpg:
    The desired dissolution profile was defined by specifying the dissolution rates at certain time points (2h, 4h, 6h, 8h). Then, these response variables were combined into a so-called desirability function in order to obtain an automatic compromise setting. The region where all response variables are within their specifications can be easily identified in the response surface plot as the region where the desirability is > 0. The best compromise setting is at the maximum of the desirability function.

  • Enzyme screening for energy generation from biomass:
    The contour plot matrix shows at a single glance which enzymes lead to a high yield already after a few days.
    bilder/ProductSTAVEX/STAVEX_Enzymes-e.jpgComparing such graphics for different time points, the enzymes can be classified according to their effectiveness (please note that the scale shifts to a generally larger yield with time): 
    • for quick results (2, 5, 9)
    • slower effect, but altogether higher yield (3)
    • effective both at short-term and long-term settings (1, perhaps 8)
    • no effect altogether (4, 6, 7).

  • Biotechnology:
    The 4D-cube illustrates the influence of three parameters on the response variable (colour coded). Ultraviolet represents the highest yield. Another attractive visualisation is obtained by using iso surfaces, which show the region for some fixed yield. Iso surfaces are also valuable for visualising the Design Space: while moving within the "egg"s, no re-validation is necessary. For instance, it is possible to counterbalance a decrease of the protein source content by increasing the amount of saccharose.

    bilder/ProductSTAVEX/STAVEX_4d-plot-e.jpg       bilder/ProductSTAVEX/STAVEX_iso-e.jpg
powered by Swisscom AG