Robust design of discrete-time General Predictive PI controller for plant models subject to variable network time delay

This work shows the advantages of the proposed General Predictive Proportional Integrative (GPPI) controller, quantified by integrated absolute error (IAE), compared to the PPI and PI controllers, established with the same robustness using the M
_s
criterion in discrete time. The controllers are applied to four plants in simulation on a network control system (NCS), three of the plants are FOPTD plants with different self-regulation indices and one corresponds to a higher order non-minimum phase plant. The tests carried out considered fixed and variable delays in a known range for the communication network. The design of the controllers considers a certain degree of robustness using a strategy based on the Nyquist criterion in discrete time. The results show that GPPI has better performance than PPI and PI for FOPTD plants when their delay is greater than their time constant and for the case of the higher order system. The improvement is attributed to the fact that GPPI combines the prediction characteristics of PPI with the smoother transient characteristics of PI, which implies a better behavior in the presence of delay uncertainty.