Diagnosis Method of Magnetization Distribution in Permanent Magnet Using Approximate Function Based on 1-D Fourier Series Expansion

To manufacture high-quality electrical machinery, the accurate estimation of magnetization is required before the practical design. The early diagnosis of magnetization distribution with high accuracy decreases the number of inferior electrical machineries. Diagnosis methods to identify the magnetization distribution in a permanent magnet are classified into two types. The first method is based on destructive testing (e.g., electron backscatter diffraction patterns, X-ray diffraction). The second method is based on nondestructive testing using measurement data on magnetic flux density in the surrounding region of the magnet. While destructive testing takes several days, the measurement time for nondestructive testing using outside magnetic flux density is much shorter. However, one drawback of the nondestructive testing is that because the number of combinations of magnetization distribution to reconstruct the outside flux is infinite, the measurement error of the measured flux density in the outside leads to infeasible magnetization distribution. Thus, to overcome this difficulty, a diagnosis method for magnetization distribution based on the Fourier series expansion is proposed. The validity of the proposed method is demonstrated in a permanent magnet with multipole orientation.