Magnetic properties measurement of silicon steel by a high frequency 2-D magnetization structure

2.1. Modeling of the 2-D magnetization structure

The novel 2-D magnetic testing structure consists of four “C-type” cores, B-H sensing box and four multi-layer excitation windings, which are wound around the orthogonal core poles, as shown in Fig. 1. In order to magnetize the specimen to saturation at high frequency, a high-grain-oriented ultra-thin (0.05 mm) silicon steel is applied to the magnetic circuit, which allows to increase the excitation field and reduces the power loss, compared with the common silicon steels [8]. Two pairs of “C-type” cores are mounted in perpendicular. The testing specimen is placed in the center of the 2-D tester and jointed by four core poles, as illustrated in Fig. 1. The multi-layer excitation winding is composed of three layers with different turns, which can be connected in series or parallel to satisfy variable excitation frequency range, flexibly. To sense B signal, B needle method is adopted to avoid the local inhomogeneity of magnetic field distribution caused by drilling holes in B coil method. H signal is measured with traditiona cross H coil method. A combined B - H sensing structure is designed.

OPEN IN VIEWER

Fig. 1.

Structure of the novel 2-D magnetic properties tester.

Figure 2 shows a block diagram of the multi-channel digital signal measurement system. Arbitrary waveforms were generated in a computer through the data acquisition card (DAQ) NI PXIe-6368 and excited the specimen so as to get sinusoidal condition for the flux density components by iterative feedback control. The 2-D magnetic properties measurement system consists of B-H composite sensing structure, data acquisition card, differential amplifier circuit and two power amplifiers, which are used to amplify the excitation signal. In order to obtain reliable measurement results, a frequency domain feedback method is proposed to control the B locus as a pure circular, which greatly diminishes the adverse effects of harmonics and increase the accuracy of the measurement [9]. By using the new designed 2-D tester and control method, alternating B-H loops, various experimental circular B loci, corresponding H loci and power losses are obtained for non-oriented silicon steel sheet (M210-35A) and the loops and the losses were obtained [10].

OPEN IN VIEWER

Fig. 2.

Magnetic properties measurement system.

3.1. Alternating magnetic properties

The alternating magnetic properties measurement of laminated silicon steel specimen is progressed when the B signal is controlled to be sinusoidal alternating waveform by the digital feedback method. In order to examine the measurement accuracy, taking the NO steel M210-35A as an example, the alternating core loss along both RD and TD are measured by the newly constructed measurement system and compared with the results from a Epstein frame of 200 turns, which is used as a reference, as shown in Fig. 3(a) and (b). It can be seen that this system can provide an acceptable measurement accuracy up to 400 Hz. Since the experimental device square is not a high frequency Epstein frame, the experiment has only been carried out to 400 Hz. In Fig. 3(c), the alternating power loss along two directions are compared at 1000 Hz. It can be observed that the alternating power loss of rolling direction is less than that of the transverse direction.

OPEN IN VIEWER

Fig. 3.

Comparison of alternating core losses measured from this measurement system and standard Epstein frame. (a) Rolling direction. (b) Transverse direction. (c) Alternating power loss in the rolling and transverse directions at 1000 Hz.

OPEN IN VIEWER

Fig. 4.

B loci and corresponding H loci under rotating magnetic measurements with harmonic compensation under different excitation currents in clockwise at 1000 Hz: (a) and (d) under small excitation currents, (b) and (e) under middle excitation currents; (c) and (f) under large excitation currents.

For the NO silicon steel M210-35A specimen, the circular rotation excitation is applied to the xoy plane parallel to the whole rolling surface. Figure 4 shows B loci and corresponding H loci under rotational magnetic measurements with harmonic compensation under different excitation currents. It can be seen that the B loci can be well controlled to be circular with a total harmonic distortion (THD) less than 0.8%.

When the excitation current is small, as shown in Fig. 4(a) and (d), the component values of the H vector in the x and y directions are slightly different and the anisotropic magnetic characteristics are insignificant. Note that rolling direction is along y axis, which is relatively easy direction to magnetize; When the excitation current is large, as shown in Fig. 4(c) and (f), the difference between the H _x and H _y components increases during the increase of the excitation current [11].

The reason for the magnetic anisotropy generated under the rotating magnetization conditions is as follows: When the excitation current is small, the anisotropy caused by the rolling stress of the silicon steel sheet plays a leading role; when the excitation current increases, especially when the saturation magnetization approaches, the grain anisotropy of the material plays a dominant role. This is also the reason why the H loci evolves from an ellipse to a saddle-shaped quadrilateral [12].

3.3. Power loss measurement and analysis

Under the condition of rotating flux density, the measurement method of rotational power loss P is based on the orthogonal component of B and H : B _X , B _Y , H _X , and H _Y , respectively. (1) Because: (2) (3) Where 𝜌 is the density of the specimen, f is the magnetizing frequency, and T = 1∕f.

It is well known that in the fieldmetric method, the value of rotational power loss measured in the clockwise (CW) and anticlockwise (ACW) directions can differ significantly. This discrepancy is caused mostly by angular misalignment of the B and H sensors. Therefore, in the fieldmetric method each total value of rotational power loss is typically calculated as an average of the clockwise (P _{rot, CW} ) and anticlockwise (P _{rot, ACW} ) values: (4) Where P _{rot, CW} and P _{rot, ACW} are separately calculated from Eq. ((1)), for measurements under that respective direction of rotation.

OPEN IN VIEWER

Fig. 5.

Rotational power loss components measured for M210-35A at 2000 Hz, under controlled circular B , plotted as: (a) P = f (B) in clockwise and anticlockwise, (b) P = f (B) in clockwise, (c) P = f (B) in anticlockwise.

The results for M210-35A measured at 2000 Hz under controlled circular B are shown in Fig. 5. The material has low anisotropy, so the P _X and P _Y components have comparable numerical values for all exciation levels. They both reach their peak values and then start decreasing at about 1.5 T, as shown in Fig. 5(a). Figure 5b and Fig. 5c show the same data, but plotted separately for x- and y-direction and as P _{rot, CW} = f (B) and P _{rot, ACW} = f (B), respectively.

A series of rotational power loss curves with different magnetic flux density magnitudes and frequencies in average are plotted, as shown in Fig. 6. The rotational power loss rises rapidly with the increasing frequency, owing to intense motions of magnetic domain walls. However, the increasing tendency is slightly different under different magnetization conditions. Besides, it can be observed that the rotational power losses increase to peak value until the magnitude of B reaches about 1.3–1.5 T and then decrease afterwards. When reaching the saturation state, the magnetic domain walls gradually disappear, the classical eddy current loss increases but the hysteresis loss and abnormal loss decrease, so the total rotational power loss also decreases. Thus, in the region close to saturation, the rotational power loss is slightly smaller than the alternating power loss.

OPEN IN VIEWER

Fig. 6.

Rotational power loss curves with different flux density magnitudes and frequencies in average, measured for M210-35A.

OPEN IN VIEWER

Fig. 7.

Comparison of power loss characteristics for M210-35A for alternating and rotational circular magnetization. (a) 50 Hz and 100 Hz. (b) 1000 Hz and 2000 Hz.

Alternating and rotational power losses are compared in 50 Hz, 100 Hz, 1000 Hz, and 2000 Hz, as shown in Fig. 7. When the magnetic flux density is low, the rotational power loss is approximately twice of the alternating loss in the rolling direction. When the magnetic flux density is large, the rotation loss will be drastically reduced and becomes even lower than the alternating loss [13].