Robustness analysis of indirect adaptive model predictive control supervised by fuzzy logic

In this paper, we consider a diesel generator represented by two models according to its operating points. The first model is an unstable and minimum phase system while the second one is a stable and non-minimum phase system. Knowing that the operating point change can affect the output plant behavior negatively, we want to study two control strategies applied to this plant. Specifically, the control robustness is analyzed regarding the model switching. The first strategy estimates online the plant model parameters while the second one reconfigures the initial tuning parameters of model predictive control. In fact, one adds to the model adaptation a fuzzy logic supervisor which performs the second adaptation regarding measurable performance criteria. Finally, we consider inequality constraints on the control signal, its variation and the output signal to highlight the relevance of our approach.