Design of Experiments

Case Study 1

Data of runs conducted

Case Study 1 Excel

1. Full Factorial Method

I input the values of the data into excel, and started off with analysis of individual factors and evaluated the average of each "HIGH" or "LOW" value of each factor.

After comparing the differences between the averages of the factors when "HIGH" or "LOW", the most significant factor was found to be C (Microwave power), followed by B (Microwave time) and then A (Diameter of bowl).

I proceeded to analyse the data further and determined the interaction of factors and their effects.

In conclusion, using the FULL factorial method, I have concluded that the most significant factor affecting the number of grams of inedible pellets to be microwave power(C), followed by time in microwave(B) and lastly diameter of bowl(A).

2. Fractional Factorial Method

I picked out runs #1, #2, #3 and #6 for analysis and put them into a table below.

Similar to my FULL factorial method, I started off with analysis of individual factors and evaluated the average of each "HIGH" or "LOW" value of each factor.

The results obtained were similar to that of the FULL factorial method, that the most significant factor found was found to be C (Microwave power), followed by B (Microwave time) and then A (Diameter of bowl).

However, I was not able to evaluate the interactions between factors as the FRACTIONAL factorial method provided me with insufficient data; the variations and combinations of factors were not enough to determine interaction effects.

CONCLUSION

From both FULL & FRACTIONAL factorial methods, analysis of both sets of data proved to be the same, with microwave power as the most significant factor affecting grams of inedible unpopped kernels, followed by microwave time and then diameter of bowl. After conducting this data analysis, in order to achieve the most optimal result of the lowest amount of unpopped kernels, the user should set microwave power to HIGH, microwave time to be HIGH and diameter of bowl to be HIGH as well.