A Typical Case Study

<<<<<< Page 10 / 10 >>>>>>

 

Having established the optimum levels, the optimal performance of the process can be predicted or projected mathematically (again einther manually or by software). Setting all factors at their optimum level, will result in this case in a projected process average of 189.18g, with a 99% confidence level. This is even larger that the largest value (160.33g) during the experiment (see page 4).

The next step would be to run the process at these optimum levels to CONFIRM the projected performance. This is in effect a measure of success of our DOE effort. In this case, the actuall performance was confirmed to be 180.12g which is pretty good given that 47% of the variations were not explained by the DOE.