Fault diagnosis of active magnetic bearings based on Gaussian GLRT detector

Active magnetic bearings are progressively replacing conventional bearings in many industrial applications, particularly in the energy sector. Magnetic bearings have many advantages such as contactless support and clean operation; however their use also poses some challenges connected to their inherent open loop instability. Occurrence of faults in one or more components of an active magnetic bearing may lead to loss of control of the rotor. Timely detection and isolation of faults in an active magnetic bearing could prevent hazardous system behaviour by enabling proper reconfiguration of the control system. A structural model of the bearing-rotor system is presented and used to perform a detectability and isolability analysis of faults in the magnetic actuator. Structural detectability and group-wise isolability is concluded for single and multiple faults in the actuators. A Gaussian generalized likelihood ratio test is proposed for detecting faults striking the electromagnet. The detector is capable of detecting and isolating the occurrence of faults in e.g. the windings of bearing by tracking changes in the mean value of a Gaussian distribution. The statistical distribution of the residuals in non faulty condition is characterized by experimental data of a full-scale bearing-rotor system. Verification of the detection performance is done through simulated data of a nonlinear model of the magnetic bearing calibrated against the real system.

[1]  Michèle Basseville,et al.  FAULT ISOLATION FOR DIAGNOSIS: NUISANCE REJECTION AND MULTIPLE HYPOTHESES TESTING , 2002 .

[2]  G. Schweitzer,et al.  Magnetic bearings : theory, design, and application to rotating machinery , 2009 .

[3]  Niels Kjølstad Poulsen,et al.  Early Detection of Parametric Roll Resonance on Container Ships , 2013, IEEE Transactions on Control Systems Technology.

[4]  Michel Kinnaert,et al.  Diagnosis and Fault-tolerant Control, 2nd edition , 2006 .

[5]  S. Kay.  Fundamentals of statistical signal processing: estimation theory , 1993 .

[6]  Michèle Basseville,et al.  Detection of abrupt changes: theory and application , 1993 .

[7]  Clifford R. Burrows,et al.  Towards fault-tolerant active control of rotor–magnetic bearing systems , 2004 .

[8]  Nan-Chyuan Tsai,et al.  Fault diagnosis for magnetic bearing systems , 2009 .

[9]  Lars Imsland,et al.  Drillstring Washout Diagnosis Using Friction Estimation and Statistical Change Detection , 2015, IEEE Transactions on Control Systems Technology.

[10]  Ilmar F. Santos,et al.  Design and Calibration of a Full-Scale Active Magnetic Bearing-Based Test Facility for Investigating Rotordynamic Properties of Turbomachinery Seals in Multiphase Flow , 2016 .

[11]  Michèle Basseville,et al.  Fault isolation for diagnosis: Nuisance rejection and multiple hypotheses testing , 2002, Annu. Rev. Control..

[12]  Matthew T. Cole,et al.  Fault-tolerant control of rotor/magnetic bearing systems using reconfigurable control with built-in fault detection , 2000 .

[13]  Active magnetic bearings-chances and limitations , 2002 .

[14]  Mogens Blanke,et al.  Diagnosis of Airspeed Measurement Faults for Unmanned Aerial Vehicles , 2014, IEEE Transactions on Aerospace and Electronic Systems.

[15]  Matthew T. Cole,et al.  Fault detection and tolerance in synchronous vibration control of rotor—magnetic bearing systems , 2001 .

[16]  Mogens Blanke,et al.  Parametric Roll Resonance Monitoring using Signal-based Detection , 2015 .

[17]  Michèle Basseville,et al.  Detection of Abrupt Changes: Theory and Applications. , 1995 .

[18]  Chong-Won Lee,et al.  Diagnosis of sensor faults in active magnetic bearing system equipped with built-in force transducers , 1998 .