Abstract
Since 100 years, the Epstein tester serves as a compact and simple apparatus for the measurement of magnetic energy losses, in particular for Fe–Si steel. However, drawbacks result from time-consuming cutting and stacking of the high number of just <inline-formula> <tex-math notation="LaTeX">$W =3$ </tex-math></inline-formula> cm wide sample strips of the standard Epstein tester (SET). Furthermore, the effects of cutting result in increased losses. Needs of annealing impede tests of modern materials with domain refinement that may lose its effectiveness. Results of numerical modeling indicated that both problems can be reduced by a strong enlargement of the strip width. Here, we report a novel giant Epstein tester (GET) with <inline-formula> <tex-math notation="LaTeX">$W =10$ </tex-math></inline-formula> cm, in combination with increased strip length <inline-formula> <tex-math notation="LaTeX">$L = 65$ </tex-math></inline-formula> cm. Sufficient averaging over test material is attained with two layers, i.e., from just eight strips. The magnetic field strength is determined by four large 3-D-printed tangential field coils, excluding impact from the four corners of frame. Time-averaged loss <inline-formula> <tex-math notation="LaTeX">$P$ </tex-math></inline-formula> is computed from instantaneous magnetization power values <inline-formula> <tex-math notation="LaTeX">$p$ </tex-math></inline-formula>, offering also maximum loss and orientation power. Results of GET were compared to data from single-sheet tester (SST) data, proving close similarities, without needs of annealing which is significant for laser scribed steel, as tested here. Results from SET prove to be higher. Analysis of power <inline-formula> <tex-math notation="LaTeX">$p$ </tex-math></inline-formula> indicates the impact of supplementary domains in the magnetization process of the scribed material. Compared to SST, the drawbacks of the GET are the need of eight sample strips, as well as lower absolute accuracy, due to higher inhomogeneity. On the other hand, advantages are the absence of a yoke system and simplicity of test apparatus.
Highlights
Since decades, increased world-wide effort is aimed on a reduction of magnetic energy losses of silicon iron steel, as applied for soft magnetic cores of electric machines like transformers, generators and motors
We report about a further enlargement to W = 100 mm, that justifies the designation of a "Giant Epstein Tester" (GET) in even better ways
For grain-oriented steel, the flux tends to follow the rolling direction (RD). It passes into the new RD through almost homogeneous off-plane flux, taking advantage of the whole corner extent. This means that it is impossible to establish a common model for flux take-over. (b) Loss detection has to be restricted to the restricted quasihomogeneous zone of the free limb region which excludes to apply the standardized, so-called current method (e.g. [1,2]). (c) Loss detection has to be based on a consistent detection of induction B and H in the quasi-homogeneous region that can be enlarged by increased strip width W
Summary
Since decades, increased world-wide effort is aimed on a reduction of magnetic energy losses of silicon iron steel, as applied for soft magnetic cores of electric machines like transformers, generators and motors. In order to restrict the impact of corners on the total of frame, standards define very narrow sample strips, according to a width of as little as W = 30 mm This reduces the mass portion of corners in advantageous ways. The further-up listed advantages of Epstein Tester stimulated us to construct enlarged versions of apparatus with increased strip width W, in order to decrease the impact of deteriorated edge zones as a consequence of a reduced ratio WD/W. To predict the effectiveness of this second step, we modeled the to-be-expected modified flux distributions by means of numerical 3D-MACC (Magnetic Anisotropic Circuit Calculation [12,13]) This studies [14] and [15] yielded encouraging results, according to the following findings: (a) The corners of an Epstein Tester are magnetized with complexity that was strongly underestimated so far. This means that it is impossible to establish a common model for flux take-over. (b) Loss detection has to be restricted to the restricted quasihomogeneous zone of the free limb region which excludes to apply the standardized, so-called current method (e.g. [1,2]). (c) Loss detection has to be based on a consistent detection of induction B and H in the quasi-homogeneous region that can be enlarged by increased strip width W
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
