Model Predictive Control on a Vacuum Column

In the petroleum industry, process automation plays a significant role in achieving product quality and production goals. Modern automatic control systems are comprised of multi-level hierarchy of control layers. At the lowest layer is a Distributed Control System (DCS) gathering process measurement and information, performing monitoring and PID control of process variables such as temperature, pressure and flow. Model Predictive Control (MPC) is a higher level of control, aimed at driving the plant conditions to make maximum profits by pushing the constraints of the plant while still meeting safety and environmental constraints.

The implementation of an Advanced Process Control (APC) project involves different phases namely, MPC design, pre-step testing and step-testing, model development, controller commissioning and support. Every phase is very important and it should be executed in the order specified above. The success of any APC project implementation also relies on working closely with control and process personnel of the plant and getting information from them to understand conditions and constraints involved in running the plant. In this presentation, the process description of vacuum column is presented first and the various phases of APC implementation are discussed. We will conclude with the results of the case study from this APC project on a vacuum column.