In the dynamic analysis of the nonlinear system with large degrees of freedom, stability of computation is an essential problem. Since noise in high frequency range often triggers instability of the computation, various time integration schemes that have an effect to dampen the noise, such as Wilson method, have been developed. This paper presents a digital filtering time integration (DFTI) scheme, which eliminates high frequency noise by using digital filtering technique. Usage of digital filter enables the implementation of widely adjustable damping. It can implement various frequency characteristics that other conventional time integration schemes can not have. Theory and computation process are described in the paper. The frequency characteristics of the DFTI scheme are also investigated and compared with those of conventional time integration methods. It is shown that the DFTI scheme can eliminate high frequency noise efficiently and that it does not deteriorate the accuracy of the analysis in the frequency range of practical concern. DFTI scheme can work with various conventional time integration methods, and it is applied to a central difference method in this paper. Efficiency of the DFTI scheme is illustrated by the numerical example of a nonlinear dynamic analysis. The results show that the DFTI scheme can eliminate high frequency noise without losing accuracy of the analysis.
展开▼
