Enhanced modal approach for model reduction

Abstract

The general applicability of the modal approach for model reduction is restricted by the lack of guaranteed bounds for approximation errors and the lack of a satisfactory modal dominance analysis procedure. Functional and computational enhancements of this approach are proposed. Functional enhancements arise by combining the modal techniques with other methods and by using improved dominance analysis techniques. The computational enhancements are the results of employing numerically reliable algorithms for both dominance analysis as well as for model reduction. The general applicability of the modal approach for model reduction is restricted by the lack of guaranteed bounds for approximation errors and the lack of a satisfactory modal dominance analysis procedure. Functional and computational enhancements of this approach are proposed. Functional enhancements arise by combining the modal techniques with other methods and by using improved dominance analysis techniques. The computational enhancements are the results of employing numerically reliable algorithms for both dominance analysis as well as for model reduction.