IVL, together with Emerson Process Management, has developed a decision support system (DSS) based on multivariate statistical process models. The system was implemented at Nynas AB's refinery in order to provide real-time TBP curves and to enable the operator to optimise the process with regards to product quality and energy consumption. The project resulted in the following proven benefits at the industrial reference site, Nynas Refinery in Gothenburg: o Increased yield with up to 14 % (relative terms) for the most valuable product o Decreased energy consumption of 8 % Validation of model predictions compared to the laboratory analysis showed that the prediction error lay within 1°C throughout the whole test period.