Operability and Control of Nonlinear Processes

by

Professor B. Wayne Bequette

Rensselear Polytechnic University

Department of Chemical and Biological Engineering

Abstract - In this talk I will provide an overview of the effect of process design on process operability, and model predictive control, with a focus on challenges due to nonlinear behavior. A classic chemical reactor example is used to illustrate that direct process scale-up can lead to a process design where certain regions of desired operation are infeasible. A case study-based simulation approach may fail to detect possible operating problems, so we develop a bifurcation-based technique to detect when process design/scale-up will lead to infeasible operating regions.

Model predictive control (MPC) has become the standard advanced control technique applied in the chemical process industries. We present two different MPC approaches to handling nonlinear systems. A multiple model predictive control approach is developed and used for drug infusion control in critical care, as well as the control of a chemical reactor with input multiplicity. In addition, an extended Kalman Filter-based nonlinear model predictive control strategy provides the capability of regulating variables that are not directly measured. On-going research involving rotating disk bioreactors and fuel cell systems will also be discussed.

Friday, October 21, 2005

3:30 – 4:30p.m.

Rm. 1500 EECS