Subestruturação dinâmica experimental de fundações de máquinas rotativas

Abstract

Dynamic substructuring allows the analysis of the subsystems that constitute a structure as it provides the means for the coupling or decoupling of each of these subsystems. In coupling, the dynamic response originated from the interaction between substructures is determined by knowing their individual behavior. In decoupling, one extracts the response of a component whose modes characterization and influence on the global dynamics is difficult to determine from the measurement of its individual dynamic response. Knowing the behavior of the complete system and its partial configuration, i.e., without the component of interest, it is possible to determine the dynamic characteristics of the latter. The applicability of substructuring was investigated in a rotating machinery foundation of an experimental test rig pertaining to the Laboratory of Rotating Machinery (LAMAR) of the School of Mechanical Engineering (FEM/UNICAMP). The foundation consists of a metal base and two bearing support structures (pedestals) in which decoupling and coupling techniques were employed. Experimental modal analysis, the least squares method for complex exponentials (LSCE) and reconstitution of the frequency response functions by modal parameters were also associated. Four decoupling methods - Standard, Extended Interface, Overdetermined, and Internal - were used to decouple the pedestal from the base in the condition where both are assembled and suspended, i.e., not attached to the test rig. The analysis of the decoupled responses considered the results of the modal analysis performed and the influence of parameters such as the number and location of measurement nodes, the order of approximation of the models, and the driving point location. With the subsequent numerical coupling of one and two pedestals to the base, both suspended and attached to the test rig, and the comparison of the substructuring results with the experimental ones, it was observed that the Overdetermined method is considered as the most suitable for the process. The coupling was performed with different connection nodes and it was concluded that only one point was enough. The work thus demonstrates the advantages of substructuring considering that the different setups of the foundation structure can be tested without the necessity of performing the experimental test of each one of the configurations.