Abstract:

It is shown that complex dynamical vibratory subsystems can be represented in a numerical simulation by an equivalent boundary condition. This boundary condition is usually written in the form of a time convolution with a known kernel. It is shown that the kernel of this boundary condition can be replaced by a simple approximate kernel when the substructure has a high modal density (a ``fuzzy'' approximation). Depending on the properties of the kernel, the time convolution may be replaced by an equivalent condition that is local in time, but depends on high order time derivatives. The issues related to implementing the boundary condition in both convolution and differential form are discussed. In particular, the effects of the boundary condition on the stability and accuracy of numerical time integration are considered. [Work supported by ONR.]