Operational optimization of crude oil distillation systems using artificial neural networks

A new methodology for optimizing heat-integrated crude oil distillation systems is proposed in this work. The new procedure considers an artificial neural networks (ANN) model for representing the distillation column. Models of the distillation column and the associated existing heat exchanger network are incorporated in an optimization framework to systematically determine the operating conditions that improve the overall process economics. Of particular interest is the problem of optimizing the net value of the products obtained from the column by increasing the yield of higher-value products at the expense of less valuable products, while taking into account feasibility of the distillation specifications, heat recovery, energy and equipment constraints. A two-stage procedure is applied to first optimize the column operating conditions based on minimum utility requirements. In the second stage the heat exchanger network is designed.