Simple representation of dynamic hysteresis losses in power transformers

The paper describes a procedure for the representation of hysteresis in the laminations of power transformers in the simulation of electromagnetic transient phenomena. The model is based on the recognition that in today's iron cores the hysteresis loops are narrow and therefore the modeling details are only important in relation to the incurred losses and the associated attenuation effects. The resultant model produces losses proportional to the square of the flux density, as expected from measurement data. It is formulated as a simple, linear relationship between the variation B - B_rev of the magnetic flux density B after a reversal point B_rev and the resulting additional field intensity H_hyst. This idea can be easily implemented in existing transformer models with or without frequency dependent modeling of eddy currents in the laminations. It has been found that in many simulation tests the representation of hysteresis is not necessary and those situations have been described where the modeling of hysteresis appears to be more meaningful.