Using a pseudorandom binary sequence for rotor-bearing system identification in active magnetic bearing rotor systems

This paper examines usage of a pseudorandom binary sequence (PRBS) in system identification of the rotor- bearing system in active magnetic bearings (AMBs). In the work, PRBS system identification is compared with adaptive amplitude stepped sine identification. PRBS is a deterministic, periodic signal varying between two levels. A stepped sine wave is a type of sine wave that has all the energy on one frequency only. The paper compares the running time, memory consumption and accuracy of the identification methods studied. A five degree-of-freedom (DOF) AMB testrig is used to test the accuracy of the stepped sine and PRBS system identification approaches. Both methods are implemented in a real-time open automation system. Required running time of the PRBS identification was significantly shorter than that of the stepped sine identification, but more memory was needed for data points. The accuracy of both identification methods was similar at lower frequencies, but at higher frequencies there was greater fluctuation with PRBS identification than when using the stepped sine method. The results show the suitability of using a PRBS in AMB rotor-bearing system identification. The PRBS based approach would be the choice for initial fast identifications of the system, while the stepped sine method would remain as the choice when focusing on identification at specific and higher frequencies.