Automatic tuning of PID controllers using frequency sampling filters

This study proposes new proportional-integral-derivative (PID) controller auto-tuners using frequency sampling filters (FSFs) for the estimation of plant frequency response information under relay feedback control. It differs from the existing auto-tuners in the estimation of plant frequency response information and its utilisation in the frequency domain-based PID controller design. A desired closed-loop time constant is selected by the user as the scaling factor of the estimated open-loop time constant. Simulation examples are used to demonstrate the efficacy of the auto-tuner and its characteristics in disturbance rejection and reference following, indicating that the auto-tuner produces satisfactory closed-loop control performance for the commonly encountered systems in process control. In addition, Monte Carlo simulation studies are used to demonstrate that the FSF model-based algorithm requires not only shorter experimental time to achieve higher accuracy in estimation but also provides much improvement over the traditional Fourier analysis-based approaches when disturbances are encountered during the relay experiments.