A variable order approach to improve differential quadrature accuracy in dynamic analysis

Differential quadrature (DQ) is a numerical technique which can produce highly accurate results by using a considerably small number of grid points. When it is applied to dynamic equations, however, DQ may exhibit dynamic numerical instability. The present paper analyzed the sources of dynamic numerical instability through a simple example, and the main finding is that dynamic stability is dominated by the grid points near and on boundaries. Based on this, we propose a variable order approach which is characterized by applying different DQ schemes to the grid points near boundaries and grid points far away from boundaries. Numerical examples of both linear and non-linear dynamic equations show that the variable order approach presented in this paper may greatly improve dynamic stability, producing convincing results.