The main economic targets of the unit are related to the capability of keeping the most valuable outputs as close as possible to pre-defined quality targets. Prior to the installation of Advanced Process Control (APC), the outputs were measured only once a day in the plant laboratory, which meant that the control goals had to be turned empirically. The APC inferential model estimates the values of the product qualities, whose accuracy is increased by correcting the neural net predictions with a bias factor. The APC controller, in turn, manages the loop set, in order to control the parameters improving performance while maintaining consistent product quality.