Estrutura de controle plantwide simplificada para a planta de Williams-Otto: otimização descentralizada coordenada com IBMF

Plantwide control systems may have hundreds of variables and this results in a complex optimization problem, requiring effective techniques to solve large scale problems.  This paper investigates the application of an IBMF coordinated decentralized approach to the optimization layer of hierarchical control systems. This layer is based on a decomposition of the original mathematical problem into smaller local optimization problems. The coordination has the task of ensuring that the solutions to the local optimization problems also satisfy the constraints of the other subsystems. This technique is applied to a Williams--Otto plant with IBMF (Interaction Balance Method with Feedback) coordination. The results are satisfactory and produce a reasonable solution to the optimization problem, obtaining a value of the objective function around 95-105% of that for the same problem when solved with a centralized optimization. There is a reduction of up to 30% in the required changes of the variables manipulated in the presence of a process disturbance.