A Tangential Induction Sensor for 3-D Analyses of Peripheral Flux Distributions in Transformer Cores

This paper concerns the detection of local magnetic induction in peripheral regions of transformer cores. Induction distributions are important for the interpretation of local losses and magnetostriction. Most authors use advanced numerical modeling-like FEM for the computation of local induction values. However, effective estimations tend to be impossible due to non-linearity, extreme anisotropy, and complex effects of hysteresis. As an alternative to FEM, the conventional method for local measurements is to use arrays of single-turn search coils. However, this method is laborious, and it affects the measured values in strong ways due to the thickness of the arranged wires and artifacts from the drilled holes. This paper presents a novel approach for peripheral core regions based on measurements with a tangential induction coil, i.e., a field coil with a ferromagnetic core. The sensor was tested on a model transformer core stacked from three packages of different widths of grain-oriented electrical steel, for the nominal induction of 1.7 T. As an advantageous effect, the sensor does not cause any interlaminar air gaps and detects the amplitude of the induction not only in planar, but also in lateral positions. Results from ~100 measuring positions are presented. They show that the induction distribution is inhomogeneous with variations up to ~15%. High induction values were measured in the widest main package of the core, in contrast to much lower ones in the outer packages. On the other hand, the corners, yokes, and T-joints show much more uniformly flux distribution. The findings of this paper can be assumed to be relevant for industry in order to optimize the construction of cores for the purpose of their uniform utilization.