Robust H∞ controller based on multi-objective genetic algorithms for active magnetic bearing applied to cryogenic centrifugal compressor

This paper proposes a robust H∞ controller based on Multi-objective genetic algorithms (MOGAs) to control the active magnetic bearings (AMBs) with application to superfluid helium cryogenic centrifugal compressor (CCC). Basic weighting function formulas with seven parameters in all are suggested for H∞ controller, whose physical connections with system performance are clearly explained. Weighting function parameters are defined in a relatively narrow range according to actual operating feature of AMB systems. Then, based on these priori information, such as search-domain and weights expression, tuning and optimizing of the design performance function is carried out applying a MOGA. The control strategy avoids time consuming task in the progress of tuning parameters and provides a complete and versatile method to obtain weighting functions for a H∞ controller. Simulation results demonstrate that the optimized H∞ controller guarantees AMB system better closed-loop behavior performance while retaining low value of control signals.