On Using the Filtered CARMA and CARIMA Models for State-Space Self-Tuning Control

Design procedures of state-space self-tuning controllers are described for both the filtered CARMA and CARIMA models. While closely resembling the generalized approach of Clarke-Kanjilal-Mohtadi (CKM), the proposed methods use the minimum possible dimension for the associated implicit-delay state-space model, and practical issues such as feedforward for linearization, regulation, servo-following, disturbance and offset elimination are considered. The resulting compartmental design allows the adjustment of tuning parameters to affect different control characteristics. For the state-space self-tuners based on the filtered CARIMA model, two different forms can be distinguished: the normal form resembles the controller configuration based on the CARMA model, while the integrating form resembles the CKM self-tuner. Satisfactory results are observed in a simulation study of using the proposed self-tuners to control the steam temperature of a solar central receiver.