Dataset of Experimental Non-Standard Dynamic Hysteresis Loops

Dataset of Experimental Non-Standard Dynamic Hysteresis Loops C. Delaunaya, L. Solimenea, O. de la Barrièreb, C. Ragusaa,* (a) Politecnico di Torino, Department of Energy, Torino, 10129, Italy (b) Université Paris-Saclay, ENS Paris-Saclay, CNRS, SATIE, Gif-sur-Yvette, 91190, France *Corresponding author : [email protected] Abstract We propose a dataset of experimental hysteresis loops that can be used for dynamic hysteresis model development and testing. This dataset includes standard measurements in sinusoidal conditions, including quasi-static ones, but also non-standard measurements corresponding to magnetisations under Pulse-Width-Modulation voltages, and magnetisation under DC-bias. The investigation are conducted on two different non-oriented FeSi, a thick one in sinus, quasi-static, biased and PWM-like conditions, and a thinner one only under sinus and PWM-like conditions. Keywords : Dynamic hysteresis, magnetic loss, magnetic characterisation, PWM, DC-bias Description The experimental setup is based, as illustrated in fig. 1, on the measurement of the primary current (through a non-inductive shunt resistor) and of the induced secondary voltage. The arbitrary signal generator, power amplifier and oscilloscope are chosen with high bandwidth and/or sampling rate to allow the creation and capture of highly dynamic behaviours. A software monitor the characterisation and can apply shape or amplitude control of the achieved flux density as needed. The samples under investigation were both rings made of 4 non-oriented (NO) FeSi(3 %) laminations, with external and internal diameter of Do= 70mm and Di= 55mm respectively. The first one is 0.35mm thick (table 1a) and extensively characterised, while the second one is thinner with 0.20mm thickness (table 1b) and was mostly studied under PWM-like magnetisation. Processing VgVsec Vsh Ipri Rsh = 100mΩ Shunt Keysight 33200B Generator ARB + – NF HSA4101 Power amplifier Lecroy HDO9104MS Oscilloscope Control Figure 1: Schematic of the characterisation setup. Table 1: Parameters of the samples (a) Sample 0.35mm Property Value Resistivity ρe48.93 ×10−8Ωm Density ρm7632kg/m3 Exact thickness d0.3471mm (b) Sample 0.2mm Property Value Resistivity ρe59.07 ×10−8Ωm Density ρm7568kg/m3 Exact thickness d0.1969mm 1 The experimental data included in this dataset are characterised under standard and non-standard conditions, as described in the following sections. Each test point correspond to a single file containing the measured vectors for time t(s), magnetic field H(A/m), flux density B(T) and its derivative dB/dt(T/s). Magnetic loss are also gathered in a separate table for a more practical access. Magnetisation waveforms Quasi-static (QDC) Quasi-static magnetisation was achieved with imposed 0.5Hz sinusoidal excitation voltage Vg. For this frequency, the difference between triangular and sinusoidal excitation was negligible. Sinusoidal (SIN) Hysteresis loops under sinusoidal excitation were gathered under tight flux density shape control, with an achieved fundamental flux density matching the reference with less than 0.5% error and a total harmonic distortion (THD) of the derivative dB/dtlower than 1%. Biased flux (BIAS) Hysteresis loop with a constant flux component (DC bias) were measured as repeating minor loops inside a given major loop, as illustrated in fig. 2a. Preliminary studies showed that the achieved biased loop depend slightly on the amplitude of the major loop Bp,M but not on its frequency TM, so only the first information is included. The flux density shape control provides the desired loop; in this case, the error on the amplitude Bpand the bias Bdc of the achieved loop was lower than 1%. Pulse-Width-Modulated excitation (PWM) The dataset includes an important set of loops achieved under two-levels Pulse-Width Modulation rectangular voltage. The fig. 2b illustrates this voltage waveform and the definition of the modulation index maand the frequency ratio mf. In this case, no shape control is used and a numerically computed PWM waveform is directly applied through the generator voltage Vg. The low resistances in the primary winding limit the distortion between this applied voltage and the achieved flux density derivative dB/dt. However, the fundamental amplitude of the flux density B1is controlled to be within 0.5% of the set point. Before application, the rectangular voltage pass through a digital Hann filter, to smooth the sharp edges and limit the resonant effects in the circuit, without widening the fronts above 2% of the switching period. -Bp,M Bp,M -Bdc Bdc Flux dernsity B (T) 0 0½Ttot Ttot Time t (ms) Bdc+Bp Bdc-Bp T ½TMajor Major loop Minor loops (a) BIAS -Vdc -maVdc 0 maVdc Vdc Tsw = T / mf ... PWM wave Modulating wave Carrier wave Voltage V (V) Time t (ms) (b) PWM Figure 2: Illustration of the test conditions used for characterisation under (a) DC-biased flux and (b) Pulse-Width-Modulated voltage. 2