Wireless model predictive control applied for dividing wall column control

Abstract

Model Predictive Control (MPC) operation is based on the process model, which can be conveniently used in event-based MPC operation. The process model can estimate process control parameters when a measurement fails or when lab sampling is used as a control parameter. Wireless control is another case in which event-based MPC operation, including both slow sampling and randomness, is required. Wireless actuators connected directly to the MPC outputs or through PID loops with wireless actuators and measurements add a significant sophistication to the system. MPC wireless operation has been validated on a Dividing Wall Column (DWC) process which can provide potentially huge savings in energy and capital cost compared to a conventional column design. In the installation at the University of Texas, WirelessHART transmitters take many of the process measurements used in control. This paper details event driven wireless MPC operation and design and discusses the application challenges illustrated by the test results.