top of page
Design of Experiment
Full factorial and fractional factorial design data analysis for Case study 2
Full Factorial Analysis for Case Study 2
The table on the right side of the picture above is the table provided to us. Since it is easier to view the data indicated by a '+' or '-', I have made another set of the data using either '+' or '-' to represent each variable.
From my plotted graph, it is seen that Factor C (stirring speed) contributed greatly to the results of the experiment. Factor C has the steepest curve compared to the other variables. Factor B has the steepest gradient and the least amount of pollutant being discharged hence why Factor B contributed the least towards the results collected. Hence, Ranking of single factor: C, A, B.
Interaction
A x B
From the graph of interaction effect A and B, it is seen that there is a minimal difference in the decrease of both gradients. Hence, there is minor interaction between Factor A and Factor B.
Interaction
A x C
From the graph of interaction effect A and C, it is seen that there is a significant interaction between both variables where both gradients are inversely proportional to each other.
Interaction
B x C
From the graph of interaction effect B and C, it is seen that both lines are parallel. Hence, there is no interaction at all.
In conclusion, Factor C which is the stirrer speed, has the most significant effect on the end results for this e
experiment. Since Factor C has an interaction with Factor A, both variables will have to work hand in hand where Factor C is to be at high and Factor A to be at low for effective results.
Fractional Factorial Analysis for Case Study 2
Above, is the same table as the first picture attached for full factorial analysis. However, the rows highlighted in yellow on the right side of the picture attached, is the runs chosen for data analysis for fractional factorial analysis. Runs chosen are 2,4,5 and 8 where the high and low level occur at an equal amount of time giving more accurate results.
Factor C contributes to the most of the amount of pollutants discharged. This is due to it having the steepest curve. Factor A too is similar to that of Factor C but Factor B has the least steep graph and hence shows that it does not affect the overall results compared to that of Factor A and C.
Ranking of single factor: C, A, B
In conclusion, due to the lack of data points used in fractional factorial, a full factorial analysis would be more beneficial and accurate. Despite such, in both analysis Factor C emerged to be the most significant variable that has affected the result of this experiment as in both graphs, it is seen that the gradient of Factor C is the steepest and the overall change is the highest compared to the 2 other variables. Therefore, in this case study, the most efficient way to increase the process of water treatment is by adjusting the stirring speed (RPM)
Hyperlink to excel sheets: DOE CPDD Case Study 2 (full + fractional).xlsx
bottom of page