Automatic Tuning of Augmented PIDs for Active Magnetic Bearings Supporting Turbomachinery

In industry, augmented proportional-integral-derivative controllers are still frequently employed with active magnetic bearings supporting turbomachinery. Despite their simple single-input-single-output (SISO) structure, the process of tuning such controllers remains iterative and manual, demanding considerable time from experienced engineers to ensure the rotodynamic system meets performance requirements. This article introduces an innovative method that facilitates engineers in incorporating design requirements by translating performance criteria into mathematical constraints and objectives. These are then exploited by an advanced nonsmooth optimization algorithm to automatically adjust controller parameters, accounting for system uncertainties and varying operating conditions. The proposed procedure offers a flexible and innovative approach for automatically designing robust SISO-based controllers for turbomachinery equipped with active magnetic bearings. This significantly reduces the time required for manual tuning by experienced engineers while ensuring all performance objectives are met. An implementation of this tuning method on a real turbomachine is presented, and the results are discussed with a particular focus for the application in the oil and gas field.