Design of an adaptive predictive control strategy for crude oil atmospheric distillation process

Improving performance and reducing the process operational costs represent a priority for the oil refinement industry. The challenge is given by high energy utilization and strict productivity specifications. Automatic control plays an essential role by providing theoretical and practical tools to overcome these challenges. This paper details the design of an Adaptive Predictive (AP) control strategy for an atmospheric distillation process. The strategy uses AP controllers to face the non-linear and time-varying process dynamics and was defined using classical interaction analysis tools such as the Relative Gain Array (RGA). The AP control strategy was simulated using ADEX and MATLAB simulation environments. The process was simulated using an Aspen Dynamic model. The controller performance is evaluated on a simulator of a crude oil atmospheric distillation process operating in a PEMEX refinery. The simulation results are also compared against a PID-based control strategy, showing an improvement of operational stability.