Intelligent Estimation of Uncertainty Bounds of an Active Magnetic Bearings Using ANFIS

Active magnetic bearings is known to be inherently unstable systems that have a widespread applications and an increased potential research area in today´s technology. So far many controllers including H_∞ robust controller has been developed for the system. However, proper selection of the required weighting functions for robust and non-conservative controller synthesize is still a critical issue that needs more investigations. In this paper the selection of uncertainty weighting function for the robust controller synthesize is automated by intelligent estimation of uncertainty weighting functions using adaptive neuro fuzzy inference system (ANFIS). Then a robust H_∞ controller for the magnetic bearings is designed based on the intelligent estimated uncertainty. v- gap metric is utilized to validate the estimated uncertainty bounds for improved robust stability. Comparison with another neural based estimation method of uncertainty proves the validity of the applied approach.