Self-Tuning for Decentralized Controller Embedded in a PLC
Multivariable systems, Decentralized control, Decoupler, Industrial automation, Gershgorin's theorem.
Among the possible controller designs for multivariable systems, the decentralized control approach is the most used by control operators. This occurs because the number of required controllers will be equal to the number of inputs and outputs feedback from system. Which leads to fewer controllers needed compared to the centralized control that has a controller for each element of the multivariable process. There is also another approach that a decoupling structure is connected between the process and the decentralized control, in order to minimize the interactions between the system loops. Given this and motivated by industrial demand for automatic controllers, this study aims to develop autotuning methods for multivariable control systems and program them into Programmable Logic Controller (PLC). The functionalities proposed to the PLC are two functional blocks: the first works as a decoupler able to adjust by user demand and; the second is a decentralized controller capable of autotuning the PI controller parameters in order to ensure robustness criteria. The functional blocks estimate the multivariable process model by experiments with the decentralized relay used to evaluate the system's critical points. In order to illustrate the functioning of the methods that will be tried on real plants, when implemented in PLC, numerical simulations were carried out to evaluate the functioning of the autotuning methods.