Evaluation of two methods for reconfiguration of multivariable controllers in process control systems

In this paper two methods for the reconfiguration of multivariable decentralized controllers are evaluated for their feasibility in real life applications. As real-life case the air control system of a waste-to-energy plant in the pulp and paper industry is used. The first method is based on the factorization of the closed loop sensitivity transfer matrix and the second on degeneration factors of the closed loop system. The two candidate methods are evaluated by comparing their results with the results obtained from established interaction measures. Thereafter the reconfiguration of the control system is performed and subsequently re-assessed. The results show that both methods can be used to derive reasonable indications on how controllers can be reconfigured to achieve better performance of the closed loop system. It is also estimated that the performance of the air control system can be improved by 3% when a feed forward link is introduced in the current control structure.