OPTIMIZATION OF THE LOCATION OF INDUCTION MEASUREMENT POINTS IN THE IDENTIFICATION OF MAGNETIZATION USING THE METHOD OF WEIGHT COEFFICIENTS
Robert V. Arutyunyan, Moscow State Technical University named after N.E. Bauman, Moscow, Russia, rob57@mail.ru
Sergey А. Nekrasov, South Russian state Polytechnic University (NPI), nekrasoff_novoch@mail.ru
Abstract
The article is devoted to the methods of nondestructive control of magnetization of permanent magnets on the basis of solving the inverse problem of magnetostatics. New are the results related to the optimization of the location of the induction measurement points, including taking into account the differences in the measurement error at different measurement points. According to the practice of magnetic flaw detection, the influence of the field of external currents was not taken into account. The presence of a ferromagnetic frame was taken into account. An approach based on the integral equation of magnetostatics is used to identify magnetization. A scalar magnetic potential is introduced for a stationary field. The corresponding numerical method (cells) is applied to solve the integral equation of magnetostatics. The volume of magnetic material is divided into cells (elementary parallelepipeds). Within the cell, the magnetization is considered constant. Integrals over the cell region are calculated analytically, and the strength for points outside the volume of the magnetic material is calculated from the corresponding analytical relations. For Slough regularization, the Tikhonov method based on minimization of a functional with some regularization parameter is used, which leads to SLOUGH with a square matrix. Due to the presence of measurement errors, different measurement points are characterized by different information content. The weighted sum method was used to account for this. The weight factor is taken inversely proportional to the measurement error. The minimized functional in the regularization method in connection with the noted circumstance takes the form of a weighted sum, where the weight coefficients are inversely proportional to the measurement errors. Identification of magnetization in the region of a rectangular permanent magnet located on a ferromagnetic base is considered as a model problem. Optimization of the location of measurement points is considered. In order to improve the adequacy of the model, the weighted sum method was used. The weight of the term was taken inversely proportional to the measurement error. The use of optimization and the method of weight coefficients significantly increases the efficiency of the considered technique.
Keywords:permanent magnets, magnetization, scalar magnetic potential, integral equation of magnetostatics, inverse problem, identification.
References
1. Denisov P.A. (2016). Solution of direct and inverse problems of magnetic field analysis of electrical devices with permanent magnets at their local demagnetization. The dissertation on competition of a scientific degree of candidate of technical Sciences. Novocherkassk. 119 p.
2. Denisov P.A., Chernoivan D.N., Seredina P.B. (2019). Identification of magnetization in the volume of a permanent magnet on a ferromagnetic base by the method of integral equations taking into account the width of the hysteresis loop. T-Comm. Vol. 13. No. 3, pp. 24-31.
3. Arutyunyan R.V., Nekrasov S.A., Seredina P.B. (2018). Identification of magnetization of permanent magnets on the basis of scalar magnetic potential method. Izv. higher educational. Electromechanics. No. 6, pp. 19-25.
4. Denisov P.A. (2018). Description of the hysteresis loop using explicit expressions for the gills-Atherton model of the second level. Izv. higher educational. Electromechanics. Vol. 61. No. 1, pp. 6-12.
5. Identification of magnetization of permanent magnets taking into account the influence of steel frame on the basis of solving the inverse problem of magnetostatics: certificate No. 2019661636 on state registration of computer programs / Nekrasov S.A. application No. 2019618701/69; date of receipt 15.07.2019; date of state registration in the register of computer programs of Russia 04.09.2019.
6. Solution of one nonlinear inverse problem of magnetostatics by regularization method / E.P. Zhidkov, I.V. Kuts, R.V. Polyakova et al. Dubna: JINR, 1988. 10 p. (Prepr. ‘ed. in-t kernels. research.; P11-88-335). https://search.rsl.ru/ru/record/01001422197.
7. Germain-Lacour P., Georges P.L., Pistre F., Bezier P. (1989). Mathematics and CAD. Moscow: Mir. 264 p. (in Russian)
Information about authors:
Robert V. Arutyunyan, Moscow State Technical University named after N.E. Bauman, Moscow, Russia, Ph.D., associate professor of the Department of Computational mathematics and mathematical physics, Moscow, Russia
Sergey А. Nekrasov, South Russian state Polytechnic University (NPI), doctor of engineering, Professor, Moscow, Russia