Quantitative methods for optimal regulatory layer selection

An iterative LMI approach to controller design and measurement selection in self-optimizing control

“…The selection of controlled variables for indirect control of the compositions for the distillation example was investigated in [22]. Based on the Exact local method, a Mixed Integer Quadratic Programming (MIQP) approach was used to select the best subsets of the available measurements (T 1 · · · T 41 ).…”

“…Based on the Exact local method, a Mixed Integer Quadratic Programming (MIQP) approach was used to select the best subsets of the available measurements (T 1 · · · T 41 ). This resulted in the following controlled variable when using three measurements: To illustrate, the ease of controlling c ref,2 , the author of [22] implemented two PI controllers, that were tuned using the SIMC method [23]. However, it should be possible to find a different measurement combination together with PI controllers that further improves the transient response, without affecting the steady-state loss.…”

Instrumentação Em Coluna De Destilação Para Fim De Controle

“…A hierarquia de controle se dá em dois níveis, a camada regulatória de cima é baseada nos graus de liberdade de regime permanente e as variáveis controladas primárias pareadas àqueles, e sua função é disponibilizar à camada de baixo setpoints ótimos para operação. A camada de baixo pressupõe análise dinâmica para definir a estrutura do controlador e escolha de variáveis controladas secundárias e seu objetivo é manter as variáveis controladas em seus respectivos setpoints [3,5].…”