Robust multivariable control system design for a fixed-bed reactor

Abstract

The design of multivariable controllers in the face of model uncertainty is addressed. Recently developed structured singular value (µ) based analysis techniques allow a nonconservative assessment of robust stability and robust performance of multivariable systems. On the basis of these techniques a methodology for the design of robust multivariable controllers is outlined and is applied to the design of control systems for a fixed-bed methanation reactor. In one configuration the inlet temperature and the flow rate are optimally combined as manipulated variables for the regulation of the outlet temperature in the face of inlet concentration disturbances. In a second configuration the optimal regulation of the product concentration is investigated when only temperature measurements are available. First, H₂ optimal controllers are designed for the nominal cases. Next, knowledge about the model uncertainty is incorporated explicitly the basis of H₂ or H_∞ performance criteria.

Additional details