Premium
Linear‐non‐linear dynamic substructures
Author(s) -
Leung A. Y. T.,
Fung T. C.
Publication year - 1991
Publication title -
international journal for numerical methods in engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.421
H-Index - 168
eISSN - 1097-0207
pISSN - 0029-5981
DOI - 10.1002/nme.1620310510
Subject(s) - mathematics , fourier series , matrix (chemical analysis) , degrees of freedom (physics and chemistry) , substructure , linear system , nonlinear system , mathematical analysis , displacement (psychology) , physics , psychology , materials science , structural engineering , quantum mechanics , engineering , composite material , psychotherapist
The dynamic substructure method is extended to linear and non‐linear coupling systems. Only those master co‐ordinates with non‐linear nature (non‐linear co‐ordinates) are retained. Other slave co‐ordinates relating to the linear part (linear co‐ordinates) are eliminated by the dynamic substructure method. The dynamic flexibility matrix associated with the linear co‐ordinates is first expanded in terms of the fixed interface natural modes. The condensed dynamic stiffness‐matrix associated with the non‐linear co‐ordinates is formed subsequently. The convergence of the condensed dynamic stiffness matrix with respect to the natural modes can be improved by means of matrix manipulations and Taylor series expansion. To find the steady state solutions, the non‐linear responses are expanded into a Fourier series. Responses of the linear co‐ordinates are related to the non‐linear co‐ordinates analytically. To solve for the unknown Fourier coefficients, the harmonic balance method gives a set of non‐linear algebraic equations relating the vibrating frequency and the nodal displacement coefficients. A Newtonian algorithm is adopted to solve for the unknown Fourier coefficients iteratively. The computational cost of a non‐linear analysis depends heavily on the number of degrees of freedom encountered. In the method, the number of degrees of freedom is kept to a minimum and the computational cost is greatly reduced.