A system-level model reduction technique for the efficient simulation of flexible multibody systems

作者：Heirman Gert H K^*; Naets Frank; Desmet Wim

来源：International Journal for Numerical Methods in Engineering, 2011, 85(3): 330-354.

DOI：10.1002/nme.2971

• non-linear model reduction
• flexible multibody dynamics
• system-level configurationdependent modal representation
• global modal parameterization
• mode veering/crossing
• real-time simulation

Flexible multibody systems are governed by sets of non-linear differential-algebraic equations (DAE). The number of degrees of freedom (DOFs) required for accurate body flexibility modeling, as well as the presence of the algebraic constraint equations limit the simulation speed. Body-level model reduction and general-purpose system-level model reduction techniques do not result in optimal model dimension reduction and do not necessarily transform the DAE-set into a cheaper to solve set of ordinary differential equations, as opposed to model reduction techniques, such as GlobalModal Parameterization (GMP), which are based on describing the system motion by the contribution of its dominant (configuration-dependent) system-level modal patterns. Their use as a simulation tool has not been thoroughly investigated. This paper investigates the influence of the modal content of the configuration-dependent mode set on the simulation accuracy through numerical experiments. By including the dynamically excited eigenmodes and static deformation patterns for DOFs in which the system is externally loaded, the system can be accurately represented by a very limited number of DOFs. Phenomena such as mode veering and mode crossing, in which the modal content of eigenmodes rapidly varies, can be expected to be problematic for such a description, which is confirmed with numerical experiments.

• 出版日期2011-1-21