Data-driven Temperature Control of Internal Mixers

This paper proposes a model-free adaptive control (MFAC) strategy for the internal mixer temperature (IMT) systems with characteristics of large time lags, large inertia, time varying and disturbances. A dynamic linearization method is applied to reconstruct the IMT system into a linear incremental form to facilitate the consequent controller design. Both input saturation and event-triggering condition are considered in the proposed algorithm, where the former is added to avoid overshooting and the latter is used to reduce the execution number of controller updates. In addition, the unknown parameter in the obtained linear model can be identified by the proposed estimation algorithm using only I/O data. The effectiveness of the proposed MFAC is verified through simulations.