Toward Automated AMB Controller Tuning: Progress in Identification and Synthesis

A significant impediment to low cost commercial deployment of magnetic bearing systems in many applications is the engineering process of tuning. In applications where the dynamics of the platform contribute significantly to overall system dynamics, AMB tuning in the vendor factory may not be sufficient to provide satisfactory performance once installed. In this case, on-site retuning is done ?by hand? by a skilled engineer and the associated cost can be substantial. In addition, later alterations in the plant, the process, or the internal machine dynamics may dictate retuning of the AMB controller, again requiring costly engineering intervention. In this paper, we examine the general characteristics of the plant that govern the requirements of an automated tuning process. The AMB machine tool spindle will serve as an example to illustrate results and outline the approach to the robust control model validation problem.