Search

A high-order locally corrected Nyström discretization for analyzing impressed current cathodic protection systems is presented. Non-linear polarization curves are incorporated using a Newton-Raphson scheme. A Schur complement scheme is introduced to handle large domains with small electrodes. The methods are characterized in terms of error convergence and computation time by comparing to the analytic solution for a sphere and hemisphere. [ABSTRACT FROM AUTHOR]

A broad-band Augmented-Müller (A-Müller) surface integral equation method for scattering from material objects is presented. The formulation incorporates surface electric and magnetic charges into the conventional Müller formulation with added constraints on the normal magnetic and electric fields. A new technique to extract the static fields is introduced which improves accuracy of computing scattered near fields at very low frequencies. The (A-Müller) formulation is discretized using the locally corrected Nyström (LCN) method. Numerical results show that the method is high-order accurate and stable over a broad frequency range from arbitrarily low to high frequencies for simply connected, multiply connected, highly lossy, high contrast and complex material geometries. The proposed formulation does not incorporate line charges, charge continuity constraints, or any frequency scaling of the degrees of freedom. [ABSTRACT FROM AUTHOR]

Local-global solution (LOGOS) modes provide a computationally efficient framework for developing error controllable, fast, direct solution methods for electromagnetic simulations. This communication provides an error control analysis for the non-overlapped localizing LOGOS factorization procedure. The error analysis indicates how to control the error in the factored representation of the system matrix. Numerical examples demonstrate that the proposed error control procedure provides factorizations with controllably small errors for a number of PEC scattering problems while maintaining the efficiency of the LOGOS factorization. [ABSTRACT FROM AUTHOR]

Localizing modes provide an effective basis for developing efficient, error-controlled factorizations of the dense matrices encountered in integral equation formulations of wave phenomena at low to moderate frequencies. An essential component of these factorization algorithms is the numerical determination of the underlying localizing modes. This communication describes the details of an efficient procedure for computing the so-called non-overlapping, localizing modes. The principle component of the procedure is a QR-like factorization of the sparse multilevel data structure used to compress discrete integral operators. Numerical examples demonstrate the performance of the algorithm. [ABSTRACT FROM AUTHOR]

Local-global solution (LOGOS) modes provide a computationally efficient framework for developing fast, direct solution methods for electromagnetic simulations. LOGOS modes provide a single basis for which both electromagnetic operators and their inverses are sparse with a controllable accuracy. For fixed-frequency modeling with h-refinement, efficient application of this observation is shown to provide efficient factorization algorithms for both integral equation and finite element methods. In addition to providing fast, direct solutions, the LOGOS framework is also shown to provide a new strategy for modular electromagnetic analysis. The possibility of developing efficient direct solution strategies at high frequencies using the LOGOS framework is also discussed. [ABSTRACT FROM AUTHOR]

An algorithm is presented for the sparse pseudo inversion of discrete forms of the plane wave transform. The starting point for the sparse pseudo inverse is a new sparse representation of the discrete plane wave transform (DPWT). While other sparse representations of the DPWT exist, the sparse representation developed here is shown to be amenable to sparse factorization. This feature is used to develop an efficient, error-controlled pseudo-inverse of the DPWT. Representative numerical examples are provided to illustrate the general properties of the proposed algorithms. Areas for additional work are identified, and potential applications of the results are discussed. [ABSTRACT FROM AUTHOR]

The details of a Galerkin discretization scheme for a modified form of the electric field integral equation are outlined for smooth, three-dimensional, perfectly conducting scatterers. Limitations of the divergence conforming finite-element bases in pit- serving the self-stabilizing properties of the electric field integral equation operator are indicated. A numerically efficient alternative is outlined which relies on an operator-based Helmholtz decomposition. The condition number of the resulting matrix equation is demonstrated to be frequency independent for scattering from a perfectly conducting sphere at various frequencies. [ABSTRACT FROM AUTHOR]

The properties of different forms of the combined field integral equation (CFIE) formulation of electromagnetic scattering from convex, perfect electrically conducting geometries are considered. Several difficulties encountered in the numerical implementation of the traditional CFIEs of electromagnetic scattering theory are discussed. An alternate form of the CFIE is introduced which is free of many of the deficiencies of traditional formulations for smooth, convex geometries. The improved numerical properties of the new formulation are illustrated with numerical examples. [ABSTRACT FROM AUTHOR]

Presents a study which examined a combined field approach to scattering from infinite elliptical cylinders using the method multiple interactions (MOMI). Techniques used; Equations; Summary of results.

The method of ordered multiple interactions (MOMI) provides a robust iterative series solution to the problem of wave scattering from rough surfaces. In this letter, a rapidly convergent MOMI series is obtained for closed-body scattering problems using a dual-surface magnetic field integral equation. The MOMI series is shown to provide a significant reduction of the iteration count relative to another technique. ©1999 John Wiley & Sons, Inc. Microwave Opt Technol Lett 20: 179–183, 1999. [ABSTRACT FROM AUTHOR]

Adhesive bonding is increasingly being considered by engineers for load-bearing joints to complement, or supplement, traditional mechanical fasteners for a large variety of structural materials, including fibre-reinforced polymers, timber, and steel. However, practitioners are still largely applying analysis and design methods, including codes and standards, that rely heavily on fudge factors. This is not due to the relative novelty of the joining technique, and the complexity of the associated mechanics. This paper guides the reader through some of the most important aspects of adhesively bonded joints, and the reasons why classical mechanics can only be applied in conjunction with fudge factors. Subsequently, the principles of fracture mechanics (FM) are presented; it is shown that, despite its conceptual strengths, FM is rather difficult to implement, as it introduces mechanical concepts practitioners are not familiar with. Then probabilistic methods (PM) are introduced; it is shown that PM can be considered as an extension of classical mechanics, as they can be easily implemented as a post-processing routine of the latter. Additionally, it is shown that PM, unlike FM, neither require specific test procedures for characterisation, nor complicated numerical analyses. Lastly, the paper offers a step-by-step guide for the implementation of PM for the dimensioning of adhesively bonded joints. The second part of this paper illustrates the latter analysis and design of adhesively bonded joints on a series of examples involving FRP, timber, and steel. [ABSTRACT FROM AUTHOR]