*3.5. Control*

Once the causes of the loss in productivity were detected and addressed, an instrumental monitoring scheme was implemented in the plant to detect possible deviations in the process variables, and specially with regards to Premix 1 and Premix 2. This monitoring system was designed in a way that avoids expensive and time-consuming tests (details not disclosed to the project team), the lack of which allowed the actual cause of the issue to go unnoticed before the Six Sigma project was carried out. Currently, the use of multivariate statistical analysis techniques for monitoring purposes supports this instrumental monitoring scheme by quickly signaling out-of-specification batches and delimiting the range of potential causes for such events.

As an example of the achieved improvement, as well as an appropriate Individual-Moving Range (I-MR) control chart to be used for monitoring *y*8 after the project goal was achieved, Figure 13 illustrates the evolution of variable *y*8 before and after improvement, for batches where Premix 2 was fed to the reactor.

**Figure 13.** Individual-Moving Range (I-MR) control chart for variable *y*8, for batches where Premix 2 was fed into the reactor, before (1) and after (2) improvement in accordance to the conclusions of the Six Sigma project.
