Control Structure Design to Achieve Multiple Performance Criteria

The control structure has the dominant effect on control performance. Developing systematic methods for selecting the structure is challenging because of the multi-criteria nature of the problem. Some criteria include dynamic behavior of the controlled and manipulated variables, equipment capacities, and loop integrity. In this paper, a method is presented to address the complex nature of the problem, while using a formulation that can be solved in a reasonable computing time. The primary criterion is controlled variable performance, but additional criteria such as integrity are considered using meaningful metrics, such as Relative Gain Array. The crucial importance of using realistic scenarios and not relying on heuristics is demonstrated using two process examples. Copyright © 2003 IFAC