Scale-Transformations in the Homogenization of Nonlinear Magnetic Processes

A class of nonlinear first-order processes is formulated as a minimization principle. In the presence of oscillating data, a two-scale model is then derived, via Nguetseng’s notion of two-scale convergence. The dependence on the fine-scale variable is eliminated by averaging with respect to the fine-scale (scale-integration or upscaling); conversely, any two-scale solution is retrieved from a coarse-scale one (scale-disintegration or downscaling). These results are first developed in a general functional framework, and are then applied to the homogenization of a relaxation dynamics in magnetic composites: $ \mathcal{A}(x/\varepsilon) {\partial B_\varepsilon\over \partial t} + \alpha(B_\varepsilon, x/\varepsilon) \ni H_\varepsilon $ $ \nabla \cdot B_\varepsilon=0, \quad \nabla \times H_\varepsilon =J(x) \quad \forall\varepsilon > 0. $ Here J is a prescribed current density. ${\mathcal{A}(y)}A class of nonlinear first-order processes is formulated as a minimization principle. In the presence of oscillating data, a two-scale model is then derived, via Nguetseng’s notion of two-scale convergence. The dependence on the fine-scale variable is eliminated by averaging with respect to the fine-scale (scale-integration or upscaling); conversely, any two-scale solution is retrieved from a coarse-scale one (scale-disintegration or downscaling). These results are first developed in a general functional framework, and are then applied to the homogenization of a relaxation dynamics in magnetic composites:

A(x/e) (?Be)/(?t)] + a(Be, x/e) ' He \mathcal{A}(x/\varepsilon) {\partial B_\varepsilon\over \partial t} + \alpha(B_\varepsilon, x/\varepsilon) \ni H_\varepsilon
Keywords:
本文献已被 SpringerLink 等数据库收录！