A rigorous homogenization theory of metamaterials—artificial periodic structures judiciously designed to control the propagation of electromagnetic (EM) waves—is developed. The theory is an amalgamation of two concepts: Smith and Pendry’s physical insight into field averaging and the mathematical framework of Whitney-like interpolation. All coarse-grained fields are unambiguously defined and satisfy Maxwell’s equations exactly. Fields with tangential and normal continuity across boundaries are associated with two different kinds of interpolation, which reveals the physical and mathematical origin of “artificial magnetism.” The new approach is illustrated with several examples and agrees well with the established results (e.g., the Maxwell–Garnett formula and the zero cell-size limit) within the range of applicability of the latter. The sources of approximation error and the respective suitable error indicators are clearly identified, along with systematic routes for improving the accuracy further. The proposed methodology should be applicable in areas beyond metamaterials and EM waves (e.g., in acoustics and elasticity).
15 May 2008 in Journal of the Optical Society of America B
Optics InfoBase is the Optical Society's online library for flagship journals, partnered and copublished journals, and recent proceedings from OSA conferences.
9 October 2018 in IEEE Transactions on Industry Applications
The complexity of core loss estimation is an essential limitation for electric machine design engineers. It is critical to estimate the core losses and reduce them in the design stage to improve the machine efficiency. Current estimation methods based on the Steinmetz equation and loss separation are not accurate enough even at the rated conditions. This work describes a loss estimation technique combining finite element analysis (FEA) and actual core loss measurements. First, flux waveforms from various parts of the electric machine are determined using finite element analysis (FEA), then identical flux density waveforms are generated in a toroidal wound core made out of the same material as is used in the machine. The loss is measured per unit mass, and then the total motor core loss is calculated by combining the measured W/kg loss values for predefined sections of the motor. Estimation results are provided and compared with the Bertotti iron loss method. The proposed method is shown to improve the accuracy of loss estimation.
Reflection and refraction of electromagnetic waves by artificial periodic composites (metamaterials) can be accurately modeled by an effective medium theory only if the boundary of the medium is explicitly taken into account and the two effective parameters of the medium (the index of refraction and the impedance) are correctly determined. Theories that consider infinite periodic composites do not satisfy the above condition. As a result, they can not model reflection and transmission by finite samples with the desired accuracy and are not useful for design of metamaterial-based devices. As an instructive case in point, we consider the ``current-driven'' homogenization theory, which has recently gained popularity. We apply this theory to the case of one-dimensional periodic medium wherein both exact and homogenization results can be obtained analytically in closed form. We show that, beyond the well-understood zero-cell limit, the current-driven homogenization result is inconsistent with the exact reflection and transmission characteristics of the slab.
8 August 2014 in Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences
Homogenization of electromagnetic periodic composites is treated as a two-scale problem and solved by approximating the fields on both scales with eigenmodes that satisfy Maxwell's equations and boundary conditions as accurately as possible. Built into this homogenization methodology is an error indicator whose value characterizes the accuracy of homogenization. The proposed theory allows one to define not only bulk, but also position-dependent material parameters (e.g. in proximity to a physical boundary) and to quantify the trade-off between the accuracy of homogenization and its range of applicability to various illumination conditions.
21 November 2011 in Journal of the Optical Society of America B
The proposed homogenization theory—a major extension of the recently published methodology [J. Opt. Soc. Am. B28, 577 (2011)]—yields an extended second-order material tensor consolidating the usual 36 local material parameters and additional ones that rigorously quantify nonlocality. The local part of the tensor relates the mean values of pairs of coarse-grained fields, while the nonlocal part relates the mean values to variations of the fields. The theory is based on a direct analysis of fields in a lattice cell rather than on an indirect retrieval of material parameters from transmission/reflection data. There are no heuristic assumptions and no artificial averaging rules. Nontrivial magnetic behavior, if present, is a logical consequence of the theory. The approximations involved and the respective errors are clearly identified. Illustrative examples of resonant structures with high-permittivity inclusions are given.