Improved estimation of iron losses for non-sinusoidal voltages

• Valentin Ionita ^[1] ; Lucian Petrescu ^[1] ; Emil Cazacu ^[1]
  1. [1] Department of Electrical Engineering, University Politehnica of Bucharest, Rumanía
• Localización: Compel: International journal for computation and mathematics in electrical and electronic engineering, ISSN 0332-1649, Vol. 37, Nº 5, 2018, págs. 1698-1706
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Purpose The electrical machines connected to modern electric power grids are non-sinusoidal excited, and their augmented losses, including iron losses, limit their working characteristics. This paper aims to propose a prediction method for iron losses in non-oriented grains (NO) FeSi sheets under non-sinusoidal voltage, involving an inverse classical Preisach hysteresis model and the time-integration of each loss component.

    Design/methodology/approach The magnetic history management in inverse Preisach model is optimized and a numerical Everett function is identified from measured symmetrical hysteresis cycles. The experimental data for sinusoidal waveforms obtained by a single sheet tester were also used to identify the parameters involved in Bertotti’ losses separation method. The non-sinusoidal magnetic induction waveform, corresponding to a measured voltage in an industrial electrical grid, was the input for Preisach model, the output magnetic field being accurately computed. The hysteresis, classical and excess losses are calculated by time-integration and the total losses are compared with those obtained for sinusoidal excitation.

    Findings The proposed method allows to estimate the iron losses for non-sinusoidal magnetic induction, using carefully identified parameters of FeSi NO sheets, using experimental data from sinusoidal regimes.

    Originality/value The method accuracy is assured by using a numerical Everett function, a variable Preisach grid step (adapted for the high non-linearity of FeSi sheets) and high-order fitting polynomials for the microscopic parameters involved in the excess loss estimation. The procedure allows a better design of magnetic cores and an improved estimation of the electric machine derating for non-sinusoidal voltages.