On-line Optimization of a Paraffins Separation Plant

This paper investigates the on-line optimization of an existing petrochemical plant. The process consists of a train of two distillation columns where a group of n-paraffins is separated into kerosene and lights. Because of process heat integration and recycles, a significant interaction between both columns is observed, which makes very complex the adjustments of process operation to the changes in feed conditions. Once developed the steady-state process and economic models for this scenario, the operability and the economic behavior of the process are evaluated by implementing on-line optimization over the dynamic simulation of the process. The dynamic model emulates the plant behavior when it is subject to disturbances with economical influence. This scenario is used to test two on-line optimization strategies: the conventional RTO methodology and RTE (Sequeira et al., 2002). Such validation scheme results very useful for an early identification of real process improvements and early problem identification. Furthermore, the validation procedure also shows the benefits offered by the latter strategy.