System model reduction for MBS optimization

Abstract

A disadvantage of optimization of flexible multibody systems (MBS) is a computing time, mainly for large systems, especially designed by FEM. The computing time rises with the complexity of the model significantly. A reduction techniques allow decreasing of degrees of freedom and it contributes to the reduction of the computing time. These techniques can be used for the reduction from thousands and more degrees of freedom to tens, but some limits exist. A reduction degree (ratio between number of DOFs before and after the reduction) is the most important feature because it predicts the final accuracy of the model. The next one is the selection of master and slave degrees of freedom that play an important role in connecting all bodies together within the MBS (e.g. by joints). There are many reduction methods, but they differ in available accuracy, speed, efficiency and suitability for the same reduction degree. A dimension of the original system is decisive for the reduction method suitability, many methods require an inversion matrix from the part of the stiffness matrix. The inversion matrix are than large and the computing time grows up. This paper deals with the reduction techniques, their disadvantages, suitability and applicability.