Vector Hysteresis Modeling in Arbitrarily Shaped Objects Using an Energy Minimization Approach

It is known that proper and efficient modeling of vector hysteresis is crucial to the precise design and performance estimation of electric power devices and magnetic recording processes. Recently, discrete Hopfield neural networks have been successfully utilized in the construction of vector hysteresis models. This paper presents a novel energy-minimization Hopfield neural network approach to implement Stoner-Wohlfarthlike vector hysteresis operators in triangular sub-regions. Advantages of the approach stem from the nonrectangular nature of such operators, which could mimic major hysteresis loops as well as their implementation in the most commonly used triangular discretization subdomains. Details of the approach are given in the paper.