Tuning Methods for Classical SISO PID Magnetic Levitation Controllers via Analytic Models

Stable operation of magnetically levitated systems requires feedback control. This paper studies tuning methodologies for the popular stabilizing PID controller which generates force/current references for the inner control loop. The pole placement method and frequency design loop shaping methods are compared in terms of their required tuning rules. Since the control system has five parameters, the tuning process is under constrained since the system only has two parameters: rotor mass and displacement unstable stiffness. Therefore, system performance objective metrics are defined and a case study multi-objective optimization is performed to find Pareto optimal tunings. Six tuning regions are identified and the salient characteristics are given and transformed into tuning rules. For example, it is found that high phase margin does not necessarily translate to robust control, nor a highly damped system response. Future magnetic levitation control designers can use the practical rules herein as a guide for creating robust and high-performance feedback PID controllers for magnetic levitation.