Component body modeling and controlled articulated flexible multibody dynamics.

摘要

The modeling and simulation of controlled-articulation flexible multibody dynamic systems tends to require the use of approximating functions, or assumed modes, to represent the structural characteristics of the constituent component bodies. However, clear guidance on appropriate component body modeling techniques is lacking. As a result, researchers and applications engineers encounter severe and unexplained numerical problems when simulating such systems. The objectives of the dissertation are to investigate and explain the numerical problems, and to develop and test modeling guidelines that will provide for predictable accuracy, robustness, and efficiency.; Closed-form solutions to the dynamics of single and multiple link, freely articulating and controlled planar flexible multi-link systems are developed. In the case of freely articulating systems, these solutions are generalized to n-links, are shown to have a recursive form, and are shown to consist of combinations of characteristic expressions associated with the solutions to simple classical beams. Linearized versions of large-articulation models are shown to be analytically equivalent to purely structural systems modeled with a hybrid set of assumed modes. This hybrid set behaves as admissible or quasi-comparison functions, and enables an investigation of numerical phenomena from a purely structural dynamics perspective. The characteristics of both freely articulating and actively controlled systems are investigated from both structural dynamics and articulated multibody dynamics perspectives, and from both analytical and numerical perspectives. The investigations provide clear insight to the dynamic and numerical characteristics of these systems.; Guidelines for modeling the constituent component bodies are formulated. They are tested for their impact on modeling accuracy, numerical robustness, and simulation efficiency. The results show that component models generated with low compliance ratios cause a lack of numerical robustness, and that high-system compliance ratios cause simulation inefficiencies. Otherwise, wide discrepancies between component model parameters and system parameters do not affect low-frequency performance of synthesized system models.