Database‐driven practical online adjustment of PID parameters

The database-driven proportional-integral-derivative (PID) control method, which is an online adjustment method of PID parameters for nonlinear systems, has a long calculation time and is difficult to apply to industrial machines. In this paper, a novel practical online adjustment method for PID parameters is proposed. The proposed method can be applied to industrial machines while adopting the database-driven approach. According to the proposed method, the overall calculation cost can be reduced by performing the online adjustment operation of the PID parameters in a slow cycle different from the calculation of the control output. In addition, PID parameters can be learned offline manner using the one-shot experimental data group corresponding to each shot. The learning method can adopt the extended fictitious reference iterative tuning (E-FRIT), which calculates control parameters directly from closed-loop data. Then, the database is updated while maintaining a certain amount of database size. The learned data are added to the existing database, and similar data are discarded using the cluster analysis method. Finally, the effectiveness of the proposed method is verified by a numerical simulation of the time-variant system. (c) 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.