Optimal modified models for robust predictive controllers

摘要

Predictive controllers suffer from sensitivity and robustness problems in the presence of strongrnill-conditioning such as all model based controllers. In spite of operating on the common robustnessrnparameters of predictive controllers an optimization technique has been developed in order to find thernalternative model to be used in place of the nominal one, which offers the best compromise betweenrnnominal performance and robustness to uncertainty. Such a modification can be undertaken only onrnthe model gains or also on the dynamic parameters according to the process characteristics.rnApplications to badly ill-conditioned distillation columns have shown the great potential of therntechnique.