An Electromagnetic Load Identification Method Based on the Polynomial Structure Selection Technique

Abstract

Electromagnetic loads can effectively monitor motor health and improve motor design. Considering the weak correlation of the modal shape and Chebyshev orthogonal polynomial in the space-time independent electromagnetic load identification method, a proposed method combining the polynomial structure selection technique together with limited measured displacement responses is presented, in which an error reduction ratio is used to pick out the significant mode shape matrix and the Chebyshev orthogonal polynomial. The time-history function of the electromagnetic load is reconstructed by combining the significant mode shape matrix and the identified concentrated load through modal transformation, and the corresponding spatial distribution function is fitted by the significant Chebyshev orthogonal polynomial. Eventually, a comparative numerical study considering the selection of significant components and measurement noise is carried out to prove the effectiveness of the presented method.