Modeling magnetic hysteresis under DC-biased magnetization using the neural network

The excitation conditions of electrical steel are generally sinusoidal but, with the advent of power electronics in recent years, dc-biased excitation is sometimes experienced. The use of an iron core under dc-biased magnetization gives rise to asymmetrical hysteresis loops and the hysteresis loss in the iron core also increases with the value of dc excitation. For iron cores working with dc-biased excitation, accurate modeling of the nonlinear characteristics for the iron core that includes the dc-bias is very important for the computation of the exciting current and the iron loss. In this paper, an efficient approach for simulating the hysteresis loop of iron core under dc-biased excitation using neural-network theory is presented. The proposed method has the merits that a specific hysteresis loop can be identified conveniently and effectively to ensure that accurate electromagnetic-field analysis can be realized .